1 EXPLORING PNEUMOCOCCAL VACCINE UPTAKE BARRIERS AMONG AFRICAN AMERICAN ADULTS IN NORTHEAST FLORIDA By CARLA ANN FRY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQ UIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Carla Ann Fry
3 T o my beautiful daughter Olivia. When I began the PhD prog ram you had just turned six years old. Every night you would go to sleep by the light of my laptop computer asking, are you going to sleep Of course, the answer was no. I stayed up late, got up ear ly, and did whatever it took to make sure I kept up with my course work, always with one thing in yo u. Every day I strive to set a good exampl e so you wi ll know that with hard work and perseverance, you can be anything your heart desires You are my highest priority my heart, my life my best friend, and the reason I was put on this earth I love you wi th all of my heart Mi j a!
4 A CKNOWLED G MENTS First, and foremost, I would like to acknowledge my family and friends for their unwavering support throughout this dissertation process. You never lost faith even when I most assuredly had done so. My wonderf ul parents have always bel ieved in me, my daughter has been a constant source of motivation and my friend Brian has served unwittingly as throughout the entire process. My best frien d Teresa has endured right at my side. Together, lau ghed, cried, driven back and forth to Gainesville countless times, and had our car towed when all we wanted to do was get a little sushi after a long hard day in the life of a PhD student. Thank you all for your unique contributions; your love, support, gu idance, and occasional reality checks. I would not be writing this were it not for each and every one of you. Next, I would like to acknowledge Dr. Stacciarini as the Chair of my committee. H er support, organization and hard work have kept me on a timelin e and made me a better writer. I also wish to thank Dr. Cook for his expertise in Epidemiology. He asked the compelling me to think of things differently adding depth and rigor to my research. Next I would like to thank Dr. Neff for her g uidance with my theoretical framework, hypotheses, and APA formatting. I would like to acknowledge Dr. Young for her unique Anthropological perspective. Dr. courses taught me to look closely at the complexities of human behavior and preventative he alth. I wish to thank Mr. and Mrs. Maren for t heir generous support through the Maren Fellowship which lessened the financial burden considerably Finally, I would like to acknowledge Sigma Theta Tau International Nursing Honor Society for the grant award ed to me to carry out this important research.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 Background and Problem Statement ................................ ................................ ...... 16 Incidence and Prevalence ................................ ................................ ................ 17 Vaccine Indications and Administration ................................ ............................ 19 Efforts to Improve Vaccine Uptake ................................ ................................ ... 20 Purpose of the Study ................................ ................................ .............................. 23 Study Aims ................................ ................................ ................................ .............. 23 Hypotheses ................................ ................................ ................................ ............. 24 Significance ................................ ................................ ................................ ............ 24 2 LITERATURE REVIEW ................................ ................................ .......................... 26 Pneumococcal Illness and Vaccine Uptak e ................................ ............................ 27 Diagnosis ................................ ................................ ................................ .......... 27 Vaccine Success Over the Past Decade ................................ .......................... 28 Pre Ex isting Health Conditions ................................ ................................ ......... 30 Potential Barriers to Vaccine Uptake ................................ ................................ ...... 35 Sociodemographic Variables ................................ ................................ .................. 36 Overview ................................ ................................ ................................ .......... 36 Age ................................ ................................ ................................ ................... 36 Gender ................................ ................................ ................................ ............. 37 S ocioeconomic Status ................................ ................................ ...................... 38 Health insurance ................................ ................................ ........................ 39 Education ................................ ................................ ................................ ... 40 Vaccine Awareness and Knowledge ................................ ................................ 40 Awareness Versus Knowledge ................................ ................................ ......... 43 Vaccine Awareness Role of the Healthcare Provider ................................ ..... 44 Provider Recommendation for Vaccine ................................ ............................ 45 Trust in Primary Care Provider ................................ ................................ ......... 47 Perceiv ed Health and Susceptibility ................................ ................................ 50
6 Health ................................ ................................ ................................ ........ 50 Perceived susceptibility ................................ ................................ .............. 53 Prior Negative Experience with Vaccines ................................ ......................... 56 Summary ................................ ................................ ................................ ................ 58 3 METHODS ................................ ................................ ................................ .............. 60 Theoretical Perspective ................................ ................................ .......................... 60 Theory Overview ................................ ................................ .............................. 60 Health Belief Model ................................ ................................ .......................... 60 Precaution Adoption Process Model ................................ ................................ 61 Precaution Adoption Process Model and Pneumococcal Vaccine Uptake ....... 61 Health B elief Model (HBM) and Pneumococcal Vaccine Uptake ...................... 63 HBM Susceptibility, Severity, and Health ................................ ....................... 64 Operationalizing the Theoreti cal Models ................................ .......................... 64 PAPM Vaccine Awareness and Knowledge ................................ ................... 65 HBM Perceived Benefits ................................ ................................ ................ 66 HBM Perceived Barriers ................................ ................................ ................. 66 HBM Cues to Action ................................ ................................ ....................... 67 Research Design ................................ ................................ ................................ .... 69 Sample and Setting ................................ ................................ ................................ 69 Inclusion and Exclusion Criteria ................................ ................................ .............. 70 Eligibility Screening ................................ ................................ ................................ 71 Measures ................................ ................................ ................................ ................ 73 Pneumococcal Vaccine Uptake ................................ ................................ ........ 74 Refusals ................................ ................................ ................................ ..... 74 Unsure ................................ ................................ ................................ ....... 75 Behavioral Risk Factor Surveillance System ................................ .................... 76 Trust in Health Care Provider Wak e Forest Trust Scale ................................ 76 Perceived Health, Susceptibility, and Severity SF ............................. 77 Single Item Questions ................................ ................................ ...................... 79 Prior Negative Experience with Vaccines ................................ ......................... 80 Provider Recommendation for Vaccine ................................ ............................ 80 Data Analysis Plan ................................ ................................ ................................ .. 81 A Priori Analyses ................................ ................................ .............................. 82 Hypothesis Testing ................................ ................................ ........................... 83 Post Hoc Testing ................................ ................................ .............................. 84 Protection of Human Subjects ................................ ................................ ................ 84 4 PILOT STUDY ................................ ................................ ................................ ........ 86 Sample ................................ ................................ ................................ ............. 87 Procedures ................................ ................................ ................................ ....... 87 Establish Trust ................................ ................................ ................................ .. 88 Refine Process and Instrument ................................ ................................ ........ 89 Cultural Sensitivity ................................ ................................ ............................ 91
7 Establish Trust ................................ ................................ ................................ .. 91 Ref ine Process and Instrument ................................ ................................ ........ 93 Cultural Sensitivity ................................ ................................ ............................ 95 Discussion ................................ ................................ ................................ .............. 96 Summary ................................ ................................ ................................ .............. 101 5 RESULTS ................................ ................................ ................................ ............. 103 Univariate Descriptive Statistics Analyses ................................ ............................ 104 Bivariate Analyses ................................ ................................ ................................ 105 Aim One ................................ ................................ ................................ ................ 105 Age ................................ ................................ ................................ ................. 106 Gender ................................ ................................ ................................ ........... 106 Socioeconomic Status ................................ ................................ .................... 107 Insurance ................................ ................................ ................................ 107 Education ................................ ................................ ................................ 107 Aim Two ................................ ................................ ................................ ................ 108 Vaccine Awareness ................................ ................................ ........................ 108 Vaccine Knowledge ................................ ................................ ........................ 109 Trust in Healthcare Provider ................................ ................................ ........... 110 Perceived Health ................................ ................................ ............................ 111 Perceived Susceptibility ................................ ................................ .................. 112 Prior Negative Experience with Vaccines ................................ ....................... 114 Healthcare Provider Recommendation ................................ ........................... 114 Stepwise Logistic Regression Model ................................ ................................ .... 115 A Priori Analyses ................................ ................................ ................................ .. 118 Unsure and Refused ................................ ................................ ...................... 118 Perceived Health Measures ................................ ................................ ........... 11 8 Post Hoc Analyses ................................ ................................ ................................ 120 Age and Perceived Health ................................ ................................ .............. 120 Trust and Racial Concordance ................................ ................................ ....... 120 Trust and Age ................................ ................................ ................................ 121 Smoking and PPSV23 Uptake ................................ ................................ ........ 121 6 DISCUSSION ................................ ................................ ................................ ....... 129 Review ................................ ................................ ................................ .................. 129 Major Findings ................................ ................................ ................................ ...... 130 Interpretation of Results ................................ ................................ ........................ 130 Gender ................................ ................................ ................................ ........... 130 Awareness, Knowledge, and He althcare Provider Recommendation ............ 132 Trust ................................ ................................ ................................ ............... 134 Unsure ................................ ................................ ................................ ............ 140 Refusa ls ................................ ................................ ................................ ......... 141 Age, Perceived Health, and Perceived Susceptibility ................................ ..... 143 Perceived Severity ................................ ................................ ......................... 146
8 Theoretical Framework ................................ ................................ ......................... 147 Lessons Learned ................................ ................................ ................................ .. 148 Study Limitations ................................ ................................ ................................ .. 150 Implications for Practice, Research, and Policy ................................ .................... 153 Implications for Practice ................................ ................................ ................. 154 Implications for Res earch ................................ ................................ ............... 155 Implications for Policy Change ................................ ................................ ....... 158 Conclusion ................................ ................................ ................................ ............ 161 APPENDIX A ADMINISTRATION GUIDELINES ................................ ................................ ........ 163 B PRESCREENING TOOL ................................ ................................ ...................... 164 C VACCINE UPTAKE QUESTIONNAIRE ................................ ................................ 165 D VACCINE UPTAKE DECISION TREE ................................ ................................ .. 175 E WAKE FOREST TRUST INSTRUMENT (FULL INSTRUMENT) .......................... 176 F PERMIS SION TO USE THE WAKE FOREST TRUST SCALE ............................ 185 G SF ................................ ................................ ............. 186 H SF LICENSE AGREEMENT ................................ ................................ .... 189 I IRB PILOT STUDY ................................ ................................ ............................. 194 J IRB FULL STUDY ................................ ................................ ............................... 195 K ELEMENTS OF CONSENT PILOT STUDY ................................ ....................... 196 L ELEMENTS OF CONSENT FULL STUDY ................................ ......................... 198 M INVITATION TO PARTICIPATE PILOT STUDY ................................ ................. 200 N PILOT STUDY QUESTIONS AND SUMMARY RESPONSES ............................. 202 O PILOT STUDY GUIDE ................................ ................................ .......................... 205 LIST OF REFERENCES ................................ ................................ ............................. 207 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 225
9 L IST OF TABLES Table page 5 1 Univariate descriptive statistics ................................ ................................ ......... 122 5 2 Pre existing disease processes ................................ ................................ ........ 124 5 3 Bivariate analyses using chi square ................................ ................................ 125 5 4 Logistic regression model ................................ ................................ ................. 127 5 5 A priori and post hoc analyses using chi square ................................ .............. 128
10 L IST OF FIGURES Figure page 3 1 Health b elief and precaution adaptation substructed model ............................... 68
11 L IST OF ABBREVIATIONS AA African American ABC Active Bacterial Core ACIP Advisory Committee on Immunization Practices AIDS Acquired Immune Deficiency Syndrome BRFSS Behavioral Risk Factor Surveillance System C DC Centers for Disease Control CI Confidence Interval CLAS Culturally and Linguistically Appropriate Services CMS Centers for Medicare Services COPD Chronic Obstructive Pulmonary Disease DF Degrees of Freedom EHR Electronic Health Record HBM Health Belief Model H1N1 Influenza Type A HIV Human Immunodeficiency Virus HPV Human Papilloma Virus IAC Immunization Action Coalition IPD Invasive Pneumococcal Disease LR Logistic Regression MMWR Morbidity and Mortality Weekly Report NCHS National Center for Health Statistics NIS National Immunization Survey
12 OMH Office of Minority Health OR Odds Ratio PAPM Precaution Adoption Process Model PASW Predictive Analytics Software PCV7 Pneumococcal Conjugate Vaccine (7 valent) PI Principal Investigator P PSV23 Pneumococcal Polysaccharide Vaccine (23 valent) RA Research Assistant SAS Statistical Analysis System SES Socio e conomic Status SF Short Form Health Survey SOP Standing Order Pro tocol VUQ Vaccine Uptake Questionnaire SF is a tradema r k of Medical Outcomes Trust and used under license from QualityMetrics Incorporated.
13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EXPLORING PNEUMOCOCCAL VACCINE UPTAKE BARRIERS AMONG AFRICAN AMERICAN ADULTS IN NORTHEAST FLORIDA By Carla Ann Fry December 2012 Chair: Jeanne Marie Stacciarini Major: Nursing Background: Pneumococcus is the bacteria responsible for thre e ma jor invasive diseases: pneumonia, bactere mia, and meningitis. T hese illnesses af fect over one million annually making invasive pneumococcal disease the most prevalent and dangerous of all vaccine preventable illnesses. Despite safety and efficacy, pneu mococcal vaccine (PPSV23) uptake remains challenging particularly among minority populations. The purpose of this study was to examine variables thought to predict PPSV23 uptake such as age, gender, socioeconomic status awareness, knowledge, trust in heal thcare provider, perceived health, perceived susceptibility, perceived severity, prior negative vaccine experiences and healthcare provider recommendation among African American (AA) adults whose age or chronic condition s rendered them PPSV23 eligible M ethods: A substructed version of the H ealth B e lief and P recaution A doption P rocess M odel s served as the theoret ical underpinning for the study A pilot study using focus group methodology was conducted prior t o the full study to ensure the in strument and m ethodologies were culturally sensitive. A quantitative cross sectional exploratory
14 design using convenience sampling was utilized to survey AA adults at two churches in Northeast Florida. A self admi nistered Vaccine Uptake Questionnaire (VUQ) served as the primary instrument. Bivariate analyses w ere conducted using chi square An empirical model with variables found to be statistically significant in bivariate analyses was then eva luated using backward stepwise logistic regression. Findings: Despite eligib ility, only 95 of 295 (32.2%) reported PPSV23 uptake. O lder age, female gender, vaccine awareness, increased knowledge higher trust scores perceived susceptibility and presence of provider recommendation for PPSV23 were significant predi ctors of vaccine uptake in bivariate analyses In the regression model, age, awareness, and provider recommendation remained significant with younger participants four times less likely to be vaccinated, those unaware six times less likely to be vaccinated, and those with out a provider recommendation seven times less likely to have the PPSV23 vaccination. Conclusion: Consistent with existing literature, three d imensions of the HBM ( barriers, cues, and susceptibility ) were significant predictors of PPSV23 upta ke. With 147 ( 47.8%) unaware of PPSV23 existence prior to this study, adding the dimension important finding in the endeavor to increase PPSV23 uptake among AA adults.
15 CHAPTER 1 INTRODUCTION Streptococc us pneumoniae also known as pneumococcus first isolated by Pasteur in 1881, is resp onsible for countless cases of o titis m edia as well as three maj or invasive bacterial diseases including pneumo coccal bacteremia meningitis and pneumonia Together, these illnesses affect over one million Americans annually making I nvasive P neumococcal Disease ( IPD ) the most prevalent and dangerous of all vaccine preventable illnesses (Centers for Disease Control [CDC], 2009). Historically, several turning points have infl uenced the incidence and prevalence of IPD and IPD related deaths including the advent of p enicillin in 1940, licensure of the first p neumococcal vaccine in 1977, and the subsequent licensure of the first p neumococcal c onjugate v accine (PCV) in 2000 recomm ended for every infant and child in the United States. Several iterations of the p neumococcal vaccine have taken place since its inception as new serotypes have been identified and isolated. Current p neumococcal vaccines, PPSV23 for adults and Prevnar 13 for children two months to five years, are considered safe and effective for the prevention of s treptococcal otitis media, pneumonia, bacteremia, and meningitis (Sokos, Skledar, Norwalk, Zimmerman, Fox, & Middleton, 2007). Marked decreases in the incidence and prevalence of these illnesses have been documented over the last decade particularly due to the success of the campaign to vaccinate children under five That said, there are still at least 175,000 hospitalizations annually due to p neumococcal pneumon ia; and more recently, pneumococcal co infections following H1N1 illnesses. Despite the safety and efficacy of PPSV23 in the adult population, vaccine uptake remains challenging
16 particularly among minority populations ( Daniels, Gouveia, Null, Gildengorin, & Winston, 2006; Schweon, 2005). This research sought to identify specific barriers to PPSV23 vaccine uptake among the African American sector of under vaccinated minorities identified by Healthy People 2010 and 2020. This section will explain p neumococc al illnesses, the PPSV23 vaccine, and existing challenges with vaccine uptake. Moreover, the problem statement, study purposes and associated hypotheses are stated, limitations are acknowledged, and the significance of the study will be presented. Backgro und and Problem Statement Invasive pneumococcal disease can lead to three major illnesses including pneumonia, bacteremia, and meningitis. While symptoms and severity differ, each of the illnesses is caused by the same bacteria. The most common of the thre e illnesses and the primary focus of this dissertation, is pneumococcal pneumonia often referred to as s treptococcal pneumonia. It is estimated that pneumococcal pneumonia is directly responsible for 5,000 deaths in the United States annually (Flowers, 200 7). The incubation period for p neumococcal pneumonia is approximately one to three days and the illness manifests itself with a sudden onset of fever, chest pain, coughing, shortness of breath, tachycardia, tachypnea, weakness, and chills (Immunization Act ion Coalition [IAC] 2009). Case fatality rates for pneumococcal pneumonia in 1997 were estimated to be 12% (Feikin, 2000). A decade later, the case fatality rate remains at 10% and it is estimated that the rate is much higher among the elderly who often g o undiagnosed (CDC, 2009). With a large volume of literature supporting the fact that case fatality rates are still high among the elderly, a great deal of effort has been placed on decreasing
17 pneumococcal illnesses in that population. A less studied popul ation vulnerable to invasive pneumococcal illnesses are young and middle aged African Americans. The incidence of pneumococcal illnesses has been found to be higher among African Americans than their Caucasian counterparts in the US by two to four fold (Bu tler & Schuchat, 1999). Pneumococcal diseases are only slightly higher among elderly Blacks than elderly Whites. Although the role socioeconomic status plays in pneumococcal illnesses is poorly understood, it seems to play a confounding role in the inciden ce of pneumococcal illnesses among African Americans who were found to have a 10 fold increase in pneumococcal illness in a study conducted in metropolitan, Atlanta (Butler & Schuchat ). Incidence and Prevalence Pneumococcal bacteremia (blood infection) oc curs in about 50,000 patients with pneumococcal pneumonia, and has an overall case fatality rate of approximately 20%. Meningitis of p neumococcal origin constitutes approximately 15% of all bacterial meningitis cases in the United States annually. Symptoms include headache, fatigue, nausea and vomiting, irritability, fever, seizures, and coma. The case fatality rate for p neumococcal meningitis is approximately 30% (Active Bacterial Core [ABC] Surveillance CDC, 2009). In the past two years, pneumococcal p neumonia and bacteremia have emerged as a significant cause of bacterial co infection among persons with H1N1 influenza, further highlighting the importance of pneumococcal vaccine uptake. Researchers from the Centers for Disease Control ( CDC ) analyzed lun g tissue samples from 77 people who died of H1N1 flu between May and August of 2009 and found that 22 (29%) of the
18 subjects had secondary bacterial pneumococcal infections ( US Department of Health and Human Services, Office of Minority Health [ OMH ] 2009. Of the 22 subjects with bacterial co infections, ten were confirmed to have S. pneumoniae with none of the victims reportedly over 56 years of age. Health histories available on 21 of the subjects revealed that 16 of them had pre existing conditions which may have increased their risk factors for contracting influenza and pneumococcal pneumonia. B y virtue of their pre existing conditions each had indications for PPSV23, though vaccine status on the subjects i s listed as in the literature ( CDC, M orbidity and Mortality Weekly [ MMWR ] 2009). With pneumococcal l pneumonia emerging as a cause of bacterial co infection in patients with influenza, it is more imperative than ever to prevent the spread of opportunisti c bacterial infections such as p neumoc occus Fortunately this is a realistic goal gi ven the proven efficacy of the p neumococcal v accin e The CDC estimates up to half of the pneumococcal related deaths annually could be prevented by administration of PPSV23 to eligible recipients, which may pro ve vital if faced with inadequate supplies of H1N1 vaccine and some measure of public skepticism regarding the newly introduced H1N1 vaccine (CDC, 2009). Statistics regarding invasive pneumococcal disease in the United States are based on active surveilla nce using the CDC Active Bacterial Core Surveillance (ABC) system. In 2008 there were an estimated 43,000 (14.3 per 100,000 population) cases of invasive disease nationally, with 4,400 (1.5 per 100,000 population) deaths. More than half of the pneumococcal related deaths occurred in adults with specific risk factors for severe disease : chronic lung disease and immunosuppression (Muench Herbert, & Rajnik,
19 2010). Invasive Pneumococcal D isease is more common among African Americans, Native Alaskans, and Navaj o and Apache Indians than any other ethnic groups in the United States even when studies have controlled for socioeconomic factors (Muench Herbert, & Rajnik ) The reason for this disparity is yet unclear and more research is sorely needed. Intuitively how ever, the higher rates of pre existing health conditions such as diabetes, heart disease, and chronic kidney disease, may place these groups at an increased risk for illness and death. Vaccine Indications and Administration The pneumococcal vaccine was fi rst licensed in the US in 1977 affording protection from 14 of the 90 known serotypes of pneumococcus. The current form of the pneumococcal vaccine (PPSV23) was licensed in 1983 replacing the former 14 valent vaccine (Sokos, et al., 2007). PPSV23 contain s 25 micrograms of purified capsular polysaccharide from 23 serotypes (1, 2, 3, 4, 5, 6B, 7F, 8, 9N, 9V, 10A, 11A, 12F, 14, 15B,17F, 18C, 19A, 19F, 20, 22F, 23F and 33F) and is reported to be up to 75% effective in preventing pneumonia due to pneumococcus among immunocompetent patients (Sisk, Whang, Butler, Sneller, & Whitney, 200 3 ). The vaccine is administered as a single dose commonly via the deltoid muscle as a 0.5 ml injection. It does not contain live virus and can be given at the same time (via separa te syringe) as any other vaccinations. The PPSV23 vaccine is given throughout the year without regard for seasons (Sokos, et al., 2005). If vaccination status is unknown, it is considered safe to assume that the patient has not received the PPSV23 vaccine and therefore, it should be administered (CDC, 2007). According to the CDC (2007), repeated pneumococcal vaccinations yielded only self limiting local reactions with no other adverse events.
20 PPSV23 vaccination is indicated by the CDC for all persons 65 (Sisk, et al., 2003). It is also indicated for adults 18 64 who are immunological ly suppress ed related to chronic conditions such as congestive heart failure, diabetes mellitus, liver disease, lung disease, renal failure, splenectomy, cancer, cochlear implants, organ transplant, and Human Immunodeficiency Virus (HIV) (Sisk, et al. ). In October 2008, the Advisory Committee on Immunization Practices (ACIP) revised PPSV23 recommendations adding cigarette smoking to the list of indications for immunization confirming smoking as an independent risk factor for p neumococcal p neumonia (CDC, 2008) ( Appendix A ) Efforts to Improve Vaccine Uptake Over the past two decades, multiple efforts have been made to increase vaccine uptake including a governmental regulat ion to allow standing order programs (SOPs) in 2002 (Middleton, et al. 2005). To facilitate the process of pneumococcal vaccination, the CDC provides a generous supply of teaching materials and sample sets of standing orders that can be utilized to improv e outcomes. Development of SOPs gives nurses and pharmacists the autonomy to vaccinate eligible individuals based on a set of predetermined evidence based criterion (Schweon, 2005). In an important ruling, the Centers for Medicare and Medicaid Services rem oved the federal requirement for a order to give each pneumococcal vaccine, paving the way for better compliance by empowering other disciplines such as nurses, assistants, nurse practitioners, and pharmacists to screen and vaccinat e when appropriate without first consulting a physician (Sokos, et al., 2007). Centers for Medicare Services (CMS) reimburses the cost of PPSV23 ranging from $18.50 to $25.00 between states for all Medicare recipients (CMS Guidelines, 2009). S OP s utilize a series of questions to
2 1 establish prior vaccination, identify at risk individuals, obtain consent to administer, and ultimately vaccinate those who agree. Additionally, a standing order protocol ( SOP ) enhances the capability for institutions to document pa tient refusals, medical contraindications, and post vaccination adverse events (CDC, 2006). Florida Medicaid on the other hand, reimburses for the pneumococcal vaccine only for institutionalized recipients who do not have Medicare benefits and despite the recommendation that some persons get a second vaccine after 65; reimbursement is limited to one per lifetime per recipient ( Rosenbaum, Stewart, Cox, & Lee, 2003) With all of the provisions in place, it would appear that the stage has been set for P PSV23 uptake to significantly improve, yet according to Healthy People 2010, only about 50% of those eligible for the vaccine actually receive it, and the number of PPSV23 eligible recipients drops off sharply when examining Black and Hispanic populations. Results of the National Immunization Survey (NIS) in 2007 where 7 055 adults self reported their pneumococcal immunization status revealed PPSV23 uptake rates among those 65 and older at 67.8% for Caucasians, 52.5% among Blacks, and 51.3% for Hispanics (N IS, 2007). The NIS study exposed less promising PPSV23 uptake rates for vaccine eligible adults in the age group 18 to 64 with only 34% of Caucasians, 26% of Blacks, and 19% of Hispanics having ever been vaccinated (National Immunization Survey [ NIS ] 2007 ). While there are a number of studies focusing on immunizations for all ethnicities 65 little attention has been given to the approximately 22 million people aged 18 64 whose chronic health conditions or current smoking habits place them at risk for p n eumococcal disease and thus eligible to receive the PPSV23 vaccine (Cantrell,
22 2009) ; and even fewer studies have focused exclusively on pneumococcal vaccine uptake among African Americans. The most recent statistics on vaccine uptake as well as pneumococca l mortality rates substantiate the pattern of under vaccination and overburden of p neumococcal disease among minority African Americans of all ages; particularly those under the age of 65 with pre existing health conditions such as diabetes and heart disea se, and/ or current smoking habits. Despite efforts to increase awareness and access to immunizations, marked racial disparities persist in pneumococcal vaccination rates Sheppard, Schwartz, & Mandelblatt, 200 4 ). CDC data show that vaccination le vels for non Hispanic African Americans and Hispanics lag behind those for non Hispanic whites. In 2007, pneumococcal vaccination coverage reflected that 62 percent of non Hispanic Whites versus 47 percent of non Hispanic African Americans and 34 percent o f Hispanics had ever received a pneumococcal vaccine ( CDC 2007). Although the percentage of vaccine uptake and pneumococcal disease burden are actually higher among Hispanics, African Americans were chosen as the population of interest in this study prima rily due to the demographic characteristics of the study location. According to the United States Census Bureau, the racial composition of Jacksonville, Florida is 62.3% White, 30.1% African American, 4.2% Hispanic, 0.3% American Indian, 3.5% Asian, 0.1% N ative Hawaiian, and 1.8% non specified. The significant number of African Americans residing in Jacksonville afforded the researcher the opportunity to obtain an adequate sample population in a timely manner. Moreover, subsequent interventional research ai med at increasing PPSV23 vaccine uptake may
23 be more impactful given the large percentage of African Americans residing in Northeast Florida. Purpose of the Study The purpose of this study was to examine factors such as age, gender, SES, vaccine awareness vaccine knowledge, overall trust in healthcare provider, perceived health, perceived susceptibility of contracting pneumococcal illness, perceived severity of pneumococcal illnesses prior negative experiences with vaccines, and healthcare provider recom mendation for vaccination which may influence pneumococcal vaccine uptake among all vaccine eligible African Americans 18 years of age. A better understanding of the characteristics consistent with the vaccinated portion of the population is needed in order to address and eliminate disparities. As understanding increases, nurses and public health professionals will be be tter equipped to develop and implement targeted interventions aimed at increasing vaccination rates among African Americans Thus potentially influenc ing policy at the state level where PPSV23 is currently not covered by Medicaid, bringing us closer to sta ted goals in Healthy People 2020 for a healthier population overall. Study Aims 1. To determine how socio demographic variables such as age, gender, and socioeconomic status, may predict pneumococcal vaccine uptake among African Americans 18 whose age or pre existing health conditions are known to increase the risk of contracting pneumococcal illness such as: diabetes, heart disease, cancer, chronic obstructive pulmonary disease ( COPD ) sickle cell disease, immunological suppression aspl enia, cochlear implants, liver disease and cigarette smoking render them eligible for PPSV23. 2. To determine how vaccine awareness knowledge trust in healthcare provider, perceived health, perceived susceptibility, perceived severity, prior negative expe rience with vaccines, and healthcare provider recommendation may predict pneumococcal vaccine uptake among African Americans 18 whose age or pre existing health conditions known to increase the risk of contracting
24 pneumococcal illness such as: diabetes, heart disease, cancer, COPD, sickle cell disease, immunological suppression asplenia, cochlear implants, liver disease and cigarette smoking render them eligible for PPSV23. Hypotheses H a Sociodemographic variables such as a ge, gender, and socioeconomic status may predict pneumococcal vaccine uptake among eligible African Americans. H a Constructs operationalized from the HBM and PAPM such as awareness knowledge, trust in healthcare provider, perceived health, perceived susceptibility, perceived severity, prior negati ve experience with vaccines and heal thcare provider recommendation may predict pneumococcal vaccine uptake among vaccine eligible African Americans. Significance Over a decade of statistics maintained by the National Health Interview Survey show that African American and Hispanic persons are disproportionately affected by invasive pneumococcal disease (C DC, 2009) In 1998, there were an estimated 62,840 cases and the incidence among African Americans w as 2.6 times higher than that of whites (95% CI [ 2.4 2.8 ] ). Among patients aged 2 to 64 years, 50.6% had a vaccine indication as defined by ACIP. The case fatality rate among patients aged 18 to 64 years with an ACIP indication was 12.1% compared with 5.4% for those without an indication (relative risk, 2.2; 95% CI [ 1.7 2.9 ] ) (ACIP, 2009). An increased understanding of the lack of awareness surrounding pneu mococcal illnesses and the PPSV23 vaccine may assist health care providers in developing better strategies to educate underserved and under vaccinated sectors of the population. A dditional knowledge of the factors influencing decision s to be vaccinated wit h PPSV23 may lead to the development of population based interventions among vaccine eligible African Americans 18 years of age. Through administration and analyses of the V accine U ptake Q uestionnaire (to be detailed in the methods section), the author h oped to identify clinically significant variables responsible for i nfluencing p neumococcal
25 vaccine uptake for use in a future study whereby an intervention designed to specifically address the needs of this underserved population and increase PPSV23 uptake will be implemented.
26 CHAPTER 2 LITERATURE REVIEW This chapter presents a literature review in three major sections relevant to the specific aims of the stud y. First, the broad topic of pneumococcal illness and vaccine uptake in the United States will be a ddressed. Second, the literature review will focus on immunization disparities and the burden of pneumococcal illnesses among African Americans in the Unit ed States. Finally, the description will focus on potential barriers to immunization such as knowledge deficit trust in healthcare provider, perceived health/susceptibility, prior negative experiences with immunizations, and provider recommendation (or lack thereof) for vaccine. R esearch focusing specifically on pneumococcal vaccine uptake was utilized for this l iterature review, though it should be pointed out that there were relatively few articles focusing solely on the pneumococcal vaccine. More commonly, research was found on the influenza vaccine which differs from PPSV23 in a number of significant ways. Whe reas the pneumococcal vaccine is a onetime immunization not associated with a given season, the Influenza vaccine is administered annually and associated with which is often the impetus for uptake. Influenza vaccine campaigns are widespread ; t herefore public awareness is intuitively much higher. And finally, the Influenza vaccine is indicated for nearly everyone in the population leaving little ambiguity about who is eligible to receive it. In spite of these differences, some of the Influenza literature has been cited in this dissertation because it most closely resembles findings one might expect related to beliefs, trust, knowledge, and perceived susceptibility of pneumococcal illnesses and vaccine uptake.
27 Pneumococcal Illness and Vaccine Upt ake Diagnosis Streptococcus p neumoniae is the etiological agent for a number of different diseases including community acquired pneumonia, meningitis, and otitis media. Streptococcus is the leading cause of pneumonia in the US, accounting for 36% of commun ity acquired pneumonia and 50% of hospital acquired pneumonia. Moreover; it has been reported as a major source of bacterial co infections following influenza illnesses. Estimates vary widely, but the Active Bacterial Core Surveillance Report states that 5 70% of adults in the US are asymptomatic carriers of pneumococci. Carriage of the pathogen and how this relates to immunity is not well understood. Those who do become ill typically have a short incubation period of roughly three days, then manifest sympt oms such as chest pain, cough, dyspnea, tachypnea, hypoxia, tachycardia, general malaise, weakness, and productive rust colored sputum (CDC, 2009). Pneumococcal illness may be diagnosed when isolated in fluids which are normally sterile such as blood and cerebrospinal fluid (Robinson, Baughman, Rothrock, et al. 2001). The most accurate way to diagnose pneumococcal illness is through a blood culture, although only 25 30% of the pneumococcal cases have bacteremia so many cases are likely to be missed using only serology (Morrison, Lake, & Crook, et al. 2000 ). Sputum samples can be analyzed for the presence of lancet shaped diplococci by gram stain, however; this diagnostic study must be interpreted carefully because diplococcic are often found in the naso p haryngeal passages and would not necessarily indicate infection. Generally, the presence of at least 25 white blood cells, and a large
28 number of gram positive diplococci along with fewer than 10 epithelial cells in a gram stained sputum specimen are reason to confirm diagnosis of pneumococcal pneumonia (CDC, Active Bacterial Core, 2009). Diagnosis by sputum sample collection is highly problematic because when a patient merely a sputum sample nearly 40% of the time it contains only saliva and bac terial pathogens present in the lung would likely be missed. Among patients who do seek care, the healthcare provider often just prescribes antibiotics as opposed to obtaining a sputum sample or blood culture With diagnostic studies proving to be problema tic, the author believes the full burden of pneumococcal diseases in the US is grossly underestimated particularly when factoring in the large numbers of people who likely fail to seek out any diagnostic confirmation or medical care. Once diagnosed, strept ococcal infections are treated with broad spectrum antibiotics and symptoms begin to abate within 12 to 36 hours (CDC, 2011). As previously stated in the introduction, roughly 50% of all pneumococcal related pneumonia infections could be prevented if perso ns eligible for the PPSV23 vaccine were properly screened and vaccinated. Vaccine S uccess O ver the P ast D ecade Although there is much to be done, the incidence of pneumococcal illnesses has indeed declined over the last decade due in large part to a conce rted effort to vaccinate children, adults with pre existing health conditions, and the elderly. In late 2000, the p neumococcal c onjugate v accine (PCV7) was introduced for children under five. Prior to that time, PPSV23 was available only for adults. Since the introduction of the PCV7 targeting children under five, great strides have been made in the overall reduction of IPD not only among vaccinated children, but also among unvaccinated children and
29 adults thus demonstrating both direct and indirect vaccine effects (Pilishvili, et al. 2010). Efforts to improve vaccine uptake among children under five and adults over the age of 65 have proven highly successful overall. While disease surveillance from 1998 (pre PCV7) to 2007 revealed steadily declining rates of IPD in all age categories, the most significant results were found in children under five where the disease rate per 100,000 decreased from 934 to six (Pilishvili et al. ). The incidence and prevalence of pneumococcal illnesses post PCV7 introduction wer e broken down by ethnicity found that Caucasian children were more likely to be vaccinated than their African American counterparts (82.1%, versus 75.2% respectively), (CDC, 2007). Without vaccination, p neumococcal pneumoniae rates begin to increase dramat ically 65 ; and although persons in this age category only account for 29% of reported cases, they make up 55% of pneumococcal related deaths (Butler & Schuchat, 1999). Often, pneumococcal pneumoniae outbreaks can be found in settings where many people are in cl ose proximity with one another such as nursing homes. With high case fatality rates and the emergence of drug resistant strains of pneumococcal illness among the unvaccinated, the elderly have been heavily targeted for vaccination with the most current ite ration of the pneumococcal vaccine (PPSV23) over the last decade. Since the inception of PCV7, adults 65 were found to have an overall decline in IPD from 638 to 68 per 100,000 (89.34% reduction), and adults between the ages of 18 and 64 had a decrease in disease rates from 930 to 133 per 100,000 (85.69% reduction) (Pilishvili, et al. 2010). Although th e results of this study demonstrate a robust decrease in the incidence of IPD and certainly provide evidence of vaccine efficacy, they should be interpreted with caution because they are not broken down by
30 ethnicity, and as previously discussed ; pneumococc al illness is inherently difficult to diagnose and it is likely that the incidence is underestimated (World Health Organization, 2007). When considering pneumococcal illness among the elderly, it is also important to note that most of the studies reported on age as a demographic variable, but failed to further break down incidence, prevalence, and mortality rates by ethnicity. Pre E xisting H ealth C onditions Adults between the ages of 18 and 64 are typically not targeted for PPSV23 vaccination unless they h ave pre existing health conditions likely to increase their risk of contracting pneumococcal illnesses. Immunocompromising conditions such as leukemia, cystic fibrosis, sickle cell disease, non lymphoma, asplenia, and HIV/AIDS warrant immunizatio n with PPSV23, as do certain heart and lung conditions such as congestive heart failure, cardiomyopathies, COPD and emphysema. Liver disease, kidney failure, cerebrospinal fluid leaks, diabetes mellitus, and alcoholism are also listed by the Centers for D isease Control as pre existing conditions warranting pneumococcal vaccination. In 2010, smoking was added as an independent indicator for the PPSV23 vaccine. A case control study of immunocompetent smokers revealed they had 4.1 times the odds of contractin g pneumococcal pneumonia than their non smoking counterparts (Sisk, 2003). Many studies (Butler, & Schuchat, 1999; Young, Bleyl, Clark, Oderda, & Liou, 2004) examine d special adult populations such as those with HIV/AIDS, and cystic fibrosis to determine t he extent to which they are aware of the vaccine and willing to accept it. Unfortunately, these studies did not present findings specific to ethnicity.
31 In a study ( Young, Bleyl, Clark, Oderda, & Liou, 2004 ) designed to examine PPSV23 vaccine uptake among adults with Cystic Fibrosis (a pre existing condition placing individuals at an increased risk for contracting pneumococcal illnesses, found that only 11.8% of adults between the ages of 18 and 49 and 29.9% of those 50 64 with chronic lung disease had be en vaccinated In this study, a simple reminder system was implemented at the health clinic, and vaccine uptake among patients with Cystic Fibrosis increased from 14.5% to 65% suggesting that awareness may play a key role. While clearly demonstrating a suc cessful intervention among adults between 19 and 64, this study did not categorize by race ; therefore it is unclear whether or not African Americans were represented in the study (Young, et al. ) A study conducted in San Francisco County, California fou nd that patients with HIV had a higher incidence of pneumococcal illness than non infected persons, 54.4% to 45.6% respectively (Nuorti, Butler, & Farley, 2000). N ot only were HIV infected patients more likely to incur pneumococcal infections, but that Afr ican American patients with HIV had a significantly higher rate of pneumococcal infection ( 2,384.8 per 100,000 cases ) The pneumococcal infection rate ratio comparing Black to non Black patients was 4.5 (CI [ 3.1 6.5 ] ), and case fatality rates ( although no t specified ) were reported as (Nuorti, et al.) These findings further substantiate the decision to choose African as the population of interest for this study. Pre existing health conditions such as diabetes, heart disease, cancer, COPD immunological suppression sickle cell disease, asplenia, cochlear implants, liver disease, and current smoking practices place individuals at increased risk of contracting pneumococcal illness. While the literature clearly delineates the ov erall
32 problem with vaccine uptake among individuals 18 64 with pre existing conditions, there are few studies that further examine specific ethnic group demographics thus demonstrating the need for this research to further inform and fill a gap. Immuniza tion Disparities and the Burden of Pneumococcal Illness among African Americans Inadequate PPSV23 vaccine uptake among the general population in the United States has been well established in the literature ( Daniels, et al., 2006; Egede, & Zheng, 2003 ) bu t African Americans and Hispanics reflect the greatest uptake disparity (Healthy People 2010). Vaccine uptake disparities among African Americans reportedly persist independent of access to care, health insurance, and socioeconomic status (Egede et al. ). Even with standing orders in place and healthcare provider recommendations for PPSV23, in a prospective study on vaccine acceptance, African Americans were significantly more likely to refuse the vaccine compared with Caucasians and Asians (19% refused, p = 0.01 ) (Daniel s, et al. ). The decision to focus on African Americans as the population of interest in this dissertation is based largely on the objectives of Healthy People 2010 which highlight s not only disparities in pneumococcal vaccine uptake and IPD, but also a disproportionately high incidence of disease processes among African Americans known to increase the risk of contracting IPD such as diabetes, heart disease, and HIV/AIDS ( US Department of Health and Human Services, [OMH] 2008). Diabetes is am ong the chronic illnesses meeting the criteria for PPSV23 vaccination in the 18 to 64 population. In a study examining Influenza and pneumococcal vaccine uptake in diabetic patients ( n = 1 906), subjects who were
33 White/Non Hispanic had a vaccination rate o f 38% compared to 22% uptake rate in African Americans ( p < 0.005). The study data were adjusted for access to care, SES, and health insurance coverage, and the authors concluded that racial disparities in vaccine coverage were an independent predictor for vaccine uptake. The authors in this study hypothesized that there may be cultural or communication barriers contributing to the disparity, but drew no conclusions. Rather, they suggested that additional research to drill down to the true reasons for the di sparity was sorely needed (Egede & Zheng, 2003 ). Daniels et al. (2006), focused on understanding vaccine acceptance by ethnicity They interviewed a sub set of African Americans who refused the PPSV23 vaccine ( n = 88), and found that reasons for refusal we re related to of (24%), of vaccine related (35%), and not believe it is (35%) (p. 1091). A retrospective cohort study ( Hausmann et al., 2009) focused specifically on racial disparities in the care of patient s with pneumonia and examined quality indicators considered to be treatment for patients diagnosed with pneumonia set forth by Joint Commission and CMS. Of these indicators, African Americans were less likely to receive the pneumococcal vac cine (53.8%) than non Hispanic Whites (67.7%). African Americans were also less likely to receive their first dose of antibiotic within four hours, less likely to be offered smoking cessation counseling, less likely to be given the appropriate antibiotic i nitially and less likely to receive the influenza vaccine. The unadjusted hospital mortality rates were significantly higher ( N = 1,183,753, p < 0.001) for African Americans (4.5%) compared with whites (4.1%). In adjusted models, within hospital effects ind icated that African Americans had significantly higher odds of mortality when compared with whites treated at the same hospital (OR = 1.05, 95% CI
34 [ 1.02 1.09 ] ). Differences in care and mortality rates were explained in two important ways; hospitals with h igh concentrations of minority populations tend ed to be lower performing facilities overall, and the inpatient population tend ed to be sicker on admission (higher number of pre existing health conditions). The comprehensive nature of this study with patien t level information from 4,000 hospitals lends a high degree of generalizability to the findings that African Americans are receiving substandard care when compared with their White counterparts both within and between hospitals across the country (Hausman n, et al., 2009) Whereas the patients in the Hausmann study (2009) all had access to care, this is not always the case. Hypothesizing that access to care might influence immunization uptake among racial minorities, a study ( Rangel, Shoenbach, Wei gle, Hog an, & Bangdiwala, 2005 ) determine d if access to care could account for the racial disparities seen with the influenza vaccine uptake Results showed that although there were significant differences in vaccination between non Hispanic (NH) whites (66%) and Hispanics (50%, p < .001) and between NH whites (66%) and NH blacks (46%, p < .001), they were only partially explained by access to care leaving other factors such as patient attitudes yet to be determined (Rangel et al.) Thus far, the review of literatur e in general has highlighted subpar PPSV23 vaccine uptake among adults in the US in general, and an even larger problem with uptake among African Americans. Efforts to explain the gap in vaccine coverage among African Americans has proven challenging, and the advent of standing order protocols has fallen short of resolving the issue. Unfortunately, there is evidence that African Americans suffer from higher incidence of pre existing health conditions, particularly
35 diabetes; thought to increase their risk of contracting pneumococcal illness. Even more troubling, when faced with IPD, African Americans are less likely to receive care consistent with the quality indicators set forth by The Joint Commission and CMS; and finally, African Americans who do become il l are more likely to die from pneumococcal infections than their White counterparts. Potential Barriers to Vaccine Uptake With substantial evidence to support PPSV23 vaccination, and clearly defined racial disparities with vaccine uptake and pneumococcal disease burden, the next logical step wa s to develop an understanding of what the barriers t o vaccination we re among the population of interest and how they might be overcome. Therefore, a comprehensive review of the literature on PPSV23 vaccine attitudes was conducted M ost studies found did not focus on the attitudes of African American adults ; those that did targeted subjects over 65 years of age. In prior research of a ttitudes among African Americans towards other matters of preventative care such as ma mmography, Human Papilloma Virus ( HPV ) and i nfluenza vaccines five barriers were identified which may be applicable to PPSV23 vaccine uptake ( Ehresmann et al. 2002 ; Santibanez, et al. 2002 ; Young, 2004 ) T hese barriers include: 1) vaccine awareness 2) knowledge 3 ) healthcare provider recommendation, 4 ) trust in healthcare provider, 5 ) perceived health 6) perceived susceptibility, 7) perceived severity, and 8 ) p rior negative experience with vaccines Each of these barriers along with sociodemographic variables thought to influence PPSV23 uptake were explored in the literature and will now be discussed in some detail.
36 Sociodemographic Variables Overview Kamal, Madhavan & Amonkar (2003), concisely measured the sociodemographic variables of interest. U sing data from the Behavioral Risk Factor Surveillance System ( BRFSS ) influenza and PPSV23 vaccine uptake were examined by demographic variables: age, gender, and socioeconomic status (SES). Participants with health insurance were 1.66 times more likely t o receive a pneumonia vaccine ( p = .015). While males were more likely than females to get the influenza vaccine, no gender differences were found for the pneumonia vaccine. Education appeared to predict vaccine uptake, with high school graduates 1.23 time s ( p = .005) more likely to get pneumonia vaccinations than those with less than high school education. Participants with some college were 1.38 times ( p < .0001) more likely than those without high school diplomas to be vaccinated. Each of these sociodemog raphic variables (age, gender, & SES) will now be broken down and discussed independently. Age The incidence of pneumococcal pneumonia is typically high in children and the elderly, but low among young and middle aged adults (Nuorti, et al. 2000). are eligible for the PPSV23 vaccine based on clinical risk factors such as diabetes and heart disease. All adults 65 and older are eligible for PPSV23 by virtue of their age. The National Center for Health Statistics (NCHS) publishes PPSV23 uptake rat es annually based on selected eligibility criteria broken down by age categories. In 2001 PPSV23 vaccine uptake rates were 5.9% for ages 18 49, 15.4% for ages 50 64, 54% for ages 65 74, 50% for ages 65 74, and 58.4% for ages 75 and older. Six years later, in 2007,
37 PPSV23 uptake statistics had improved very little with rates of 5.7%, 18.2%, 52%, and 63% respectively (NCHS, 2007). In 2011, the National Vaccine Advisory Committee (NVAC) published a report on the current state of PPSV23 vaccination in the Unit ed States and concluded that uptake among high risk adults aged 18 64 was only 18% compared with that of adults 65 and older at 61%. These uptake rates fall well below the goals of Healthy People 2020 (58%, and 84% respectively) (NVAC, 2011). Data reflecte d in each report must be interpreted with caution for the purposes of this dissertation work because they were broken down strictly by age or race, not both. Nonetheless, poor PPSV23 uptake rates among African Americans aged 18 64 with existing risk factor s are consistently reported in the literature. Gender There were conflicting data in the literature on the relationship between gender and vaccine uptake. In the aforementioned study by Khamal, et al. (2003), no gender differences for PPSV23 uptake were detected, however; the 2007 NCHS report indicated a slight difference, not statistically significant, in uptake rates (54.3% male, 59.2% female). Qualitative studies pertaining to HIV vaccination and to found that while perceived barrie rs to vaccination differed between genders, overall intent was similar (Kakinami, Newman, Lee, & Duan, 2008). An influenza uptake study among 939 COPD patients yielded no significant uptake differences between genders (Campos, Alazemi, Zhang, Sandhaus, & W anner, 2008). Allen, Kennedy, Wilson Glover, and Gilligan, (2007), conducted focus groups to elucidate perceptions of prostate cancer screening and found that African
38 American men often felt that going to the doctor was a sign of weakness. Two resp ondents in the study called the phenomenon saying men go to the doctor until something is falling (Allen, et al. p. 2194). While gender differences pertaining to vaccine uptake were not different in some of the stud ies, this relationship remains of interest particularly because the pneumococcal study w as composed of all African American subjects and therefore uniquely positioned to observe gender differences within the population of interest. A priori plans examin ing gender as a sub group will be discussed in Chapter 3 of the dissertation. Socioeconomic Status According to the National Center for Educational Statistics (NCES), socio economic status (SES) is defined as combined economic and sociological measures of a person's work experience, economic and social position relative to others. Standard measures are typically based on income, education, and occupation, although wealth is sometimes included as well (NCES, 2008). Measuring income is reportedly problematic because the rate of non response is typically high, particularly among minority groups. Income also fails to capture items such as health insurance and disability benefits (Shavers, 2007). Other measures of SES identified in the literature included insura nce status, geographical location by zip code, and neighborhood conditions such as crime and housing (Moore, 2009). The most common measures of SES found in the literature were education and income combined; however, there was insufficient data to support the use of one combination over another. Hence, for the purpose of this dissertation
39 review, selected SES variables (education and health insurance), will be considered separately as potential demographic predictors of pneumococcal uptake. Based on focus g roup responses, questions regarding income have been deemed intrusive; therefore in the interest of cultural sensitivity, this information will not be collected. Health i nsurance Florida Medicaid reimburses for the PPSV23 vaccine only for institutionalize d recipients who do not have Medicare benefits. Despite the recommendation that certain persons get a second vaccine after 65, reimbursement is limited to one per lifetime per recipient (Rosenbaum et al., 2003 ). The Saint Johns and Duval County Healt h Departments in Florida were contacted for pricing. Duval County stated the fee was $55.00 and Saint Johns County reported $45.00. Sliding scale fee schedules exist, but were not provided ( Garvin, C., personal communication, July 11, 2011). Payment of $45 $55 dollars may be cost prohibitive for some, thus type of insurance and insurance status may indeed be predictive of PPSV23 uptake for some. Data from 6,334 elderly respondents of the National Health Interview Survey (NHIS) in 2005 were analyzed for raci al disparities related to influenza vaccine uptake. Demographic components found to be significant predictors of vaccine uptake were race, education, social support, and insurance status. A greater proportion of Caucasians than African Americans had either private or Medicare coverage (74% and 43% respectively). Overall influenza vaccina tion coverage was 63.2% (95% CI [ 61.7 64.6 ] ) and coverage differences between whites and blacks were 65.6%; (95% CI [ 64.1 67.1 ] ), and 45.9% (95% CI [41.6, 50.3 ] ) respectiv ely. Controlling for education and social support (using marital status as a proxy for social support), the rate of non
40 vaccination only decreased by 1/3, leaving the researchers to conclude that being uninsured was a strong predictor of non vaccination. ( Rangel, et al. 2005). Education Studies have shown that education level is often predictive of vaccine uptake or at least the belief that vaccines are safe (Endrich, Blank, & Szucs, 2009; Galarce, Minsky, & Viswanath, 2011; Rangel, et al. 2005). Those w ith higher levels of education tend to believe vaccines are safe, whereas individuals with less education are more suspect. Concerned with H1N1 as a pandemic virus, Galarce, et al. surveyed 1 569 individuals online, oversampling minorities based on data re flecting persistent disparities. Among other findings such as beliefs about vaccine safety and efficacy, the authors found that less than half of the individuals with less than a high school diploma perceived the vaccine as safe, whereas two thirds of thos e with a or higher degree felt it was. Vaccine Awareness and Knowledge Lack of awareness of the pneumococcal (PPSV23) existence has been written about fairly extensively in the literature. Unlike the influenza vaccine which is given a nnu ally, PPSV23 is typically only given once in an adult s lifetime after the age of 65 unless pre existing health conditions known to increase the risk of pneumococcal illness necessitate that it be given sooner. Given the administration guidelines, it is not surprising that the vaccine could be overlooked. If the target population is largely unaware of the existence of PPSV23, there is little doubt that uptake rates will continue to suffer (Santibanez, et al. 2002).
41 Several studies designed to evaluate and measure the extent to which lack of awareness affects PPSV23 vaccine uptake have identified awareness as problematic irrespective of race, gender, or socio economic status (SES) ( Daniels, et al., 2006; Flowers, 2007; Santibanez, et al. 2002; Winston, Wortley, & Lees, 2006) The following studies represent quantitative and qualitative works on PPSV23 vaccine awareness primarily among the population of interest. Moreover, a distinction has been made between the constructs of and a s they pertain to PPSV23 vaccine uptake. S apsis & Janssen, (2002), conducted qualitative surveys among African American and Hispanics 65, and again found that most were unaware of the pneumococcal vaccine. Questions such as my doctor tell me about were common. Using a quantitative approach, Wahid Nag, Bilous, Marshall and Robinson, (2001), examined 268 diabetic patients (race not reported), 144 of whom had clinical conditions warranting PPSV23 vaccination, and found that only 35% had rec eived the influenza and pneumococcal vaccines (93/144 ). Unvaccinated subjects were largely unaware of the need for either or both vaccines (69% and 91% respectively). In 2002, researchers for the Department of Health studied knowledge, attitudes, and beli efs among the general public regarding the pneumococcal vaccine. Self reported pneumococcal vaccine coverage was 59% (95% CI 54%, 64%). Once again, a lack of awareness contributed significantly to non vaccination (OR 25.3, p < 0.001). Unvaccinated particip ants who had heard of PPSV23 stated they would have accept ed the vaccine had they known i t was safe and played a major role in preventing pneumonia (Ehresmann et al. 2002). While it is unclear whether awareness alone
42 would have increased PPSV23 uptake in the aforementioned studies, or if additional knowledge regarding PPSV23 would first be necessary, the scope of the problem is clearly defined in the literature. A vaccine disparities initiative conducted by Winston, et al. 2006, explored barriers to pneu mococcal and Influenza vaccine uptake using a cross sectional telephone survey stratified by race. V accination awareness was ascertained by asking subjects ( n = 4,577) if they had ever heard of the pneumonia shot and whether or not their healthcare provide r had recommended it. Inclusion criteria for this study were seniors age 65 and older from the Medicare Enrollment Database (EDB) excluding t hose between 18 and 64 whose pre existing health conditions could predispose them to pneumococcal illnesses. Vacci ne uptake rates by race were 70.3% for W hites, 40.8% for B lacks, and 53.2% for Hispanics Importantly, provider recommendation (risk ratio (RR) 52.32, 95% confidence intervals CI [ 52.10 2.57 ] ) and vaccine awareness (RR 51.60, 95% CI [ 51.40 1.82 ] ) were as sociated with greater pneumococcal vaccine uptake (Winston, et al. ). When Daniels et al. (2006) evaluated a sub set of CDC study participants, they found that of 249 patients surveyed, 180 were completely unaware of the existence. The authors of this study noted that was higher among adults 65, but it was not significantly different between Whites and African Americans. In this study alone, 72% of patients were unaware of the PPSV23 vaccine certainly suggesting that healthcare educators have much work to do. The Department of Health and Hu man Services, conducted 18 focus groups in five US cities to examine knowledge, beliefs, and attitudes towards influenza and
43 pneumococcal illnesses. Participants in this particular study were all African American or Hispanic and over the age of 65. While t he study revealed significant knowledge of the influenza vaccine, most of the participants were completely unaware of the existence of the pneumococcal vaccine. The participants were keenly aware of the health risks pneumonia posed to them, but l ack ed know ledge of the vaccine entirely (Department of Health and Human Services, CDC, 2009). Daniels, et al., (2004), conducted a series of four focus groups asking 22 unvaccinated African American and Latino adults eligible for PPSV23 by virtue of age or clinical conditions, what their reasons for non vaccination were Many participants reported that they were altogether unaware of the pneumonia vaccine, or they lacked information about the benefits or potential side effects of PPSV23. They also stated their physi cians had not discussed or recommended the vaccine. The majority expressed willingness to be vaccinated, and felt that community churches might serve as a venue for an immunization campaign. Awareness V ersus K nowledge It is unclear whether awareness alone would have increased PPSV23 uptake in the aforementioned studies, or if additional knowledge regarding PPSV23 would first be necessary. In order to more fully understand the problem, it is necessary to make a distinction between awareness and knowledge. K nowledge implies a higher degree of understanding beyond simple awareness of PPSV23 vaccine existence (Ehresmann, et al. 2002). This distinction was alluded to, although not explicitly stated, in the Daniels study, where subjects felt they would have chos en to be vaccinated had they known the eligibility criteria and benefits of the vaccine. To operationalize and measure the
44 constructs of awareness and knowledge in this dissertation study, two single item questions each of which will be discussed in great er detail in Chapter 4 w ere included on VUQ this study, had you ever heard of the pneumococcal vaccine (sometimes called the pneumonia w as adapted to measure awareness as a potential predictor variable. To measure knowledge, a 5 level Likert style question w as added: have enough information about the pneumococcal vaccine to make an informed decision about whether or not I should receive Vaccine A wareness R ole of the H ealthcare P rovider Without awareness of the PPSV23 va ccine existence, individuals must rely on health care professionals for information, recommendation, and administration. While in many instances health care practitioners appropriately recommend PPSV23 some do not. Time const raints, lack of awareness in p rimary, secondary and tertiary care settings, and the individual belief systems of professionals may be at work as well. In an unpublished thesis entitled Pneumococcal vaccine uptake in the acute care setting Fry ( 2008 ) found significant knowledge gaps a mong nurses related to PPSV23 indications and administration in the acute care setting, and subsequently conducted an education intervention aimed at increasing nursing knowledge of the vaccine. Pre and post intervention analyses demonstrated a statistical ly significant increase ( 2.09, p < .05) in PPSV23 vaccine uptake among subjects ( n = 90) once the nursing staff was educated about the vaccine thus providing one example of how integral the role of the health care provider is in vaccine uptake. Use of ed ucation at the point of service as a means to improving pneumococcal vaccine uptake has proven only moderately successful over the short term No long
45 term improvements in vaccine uptake associated with provider education alone were found in the literature I mplementation of standing order protocols for pneumococcal vaccination among patients at or above 65 years of age was reported to have i mproved compliance from zero to 78% (Shevlin, Summers Bean, Whitney, & Todd, 2002). As these studies demonstrate, no t all health care practitioners are vaccine savvy and when faced with rising costs, staffing challenges, and time constraints preventative care can take a back seat to the acute problem or for in many cases. Armed with this information, con sumers of health in the current delivery model clearly need to educate themselves taking an active in approach to health maintenance and preventative care. The paradigm shift from to will not happen overnight, a nd must be fostered by health care practitioners committed to informing the public about vaccine indications including PPSV23. Upon completion of this dissertation work, the author plans to disseminate findings on knowledge, trust, perceived susceptibilit y perceived health status, and prior history with vaccines among healthcare providers and key administrative leaders in the community. Once healthcare providers are aware of the reasons African Americans opt to accept or decline PPSV23, education for prov ider and interventions aimed at the target population may be designed and implemented. Provider Recommendation for Vaccine The vaccine knowledge and healthcare provider recommendation f or PPSV23 appear to be interrelated. For example, a patien t may be aware of PPSV23, but lack knowledge pertaining to indications, safety, and efficacy largely due to a lack of explanation by the healthcare provider and/or a recommendation. In fact, studies have
46 shown that healthcare providers sometimes lack vacci ne related knowledge themselves, or report being to address vaccine status. Registered nurses have also voiced concerns and reluctance to administer PPSV23 without a order despite existing protocols ( Middleton, et al. 2005). S tudie s on a variety of vaccines including: HPV, varicella, influenza and pneumococcal, support the notion that patients are more inclined to accept a vaccination when a healthcare provider, ( whether a physician, nurse, nurse practitioner or pharmacist ) recomme nds the vaccine (de Courval, de Serres, & Duval, 2003; Telford, & Rogers, 2003, Winston, Wortley, & Lees 2003;Young, 2004). A uthors found that provider recommendation for the v aricella vaccine contributed significantly as a predictor of uptake ( n = 477 ). Winston, Wortley, & Lees (200 6 ), examined barriers to influenza and pneumococcal vaccine uptake among Medicare recipients in five US communities with a particular focus on racial disparities They found that pneumococcal vaccination coverage was 70.3% for W hites, 40.8% for B lacks, and 53.2% for Hispanics Importantly, the proportion of participants who report ed provider recommendation for vaccination differed significantly according to race/ethnicity (RR = 2.32, 95% CI [ 2.10 2.57 ] ). A lthough significantly related to vaccine uptake only half of respondents report ed provider recommendation for influenza vaccination. According to Young (2004) to receive pneumococcal vacc it is unclear why providers often fail to recommend vaccines when they come in contact with patients, especially when they have the ability to influence decisions. Sapsis, and Janssen ( 2002 ) conducted
47 qualitative intervi ews asking elderly African Americans and their physicians why influenza vaccination rates lagged behind Caucasian counterparts, and found that not only were patients unaware; but often healthcare specialists were unfamiliar with vaccine eligibility and gui delines essential for further patient education and decision making. Moreover, physicians did not feel they had time to provide lengthy explanations as to why patients should be vaccinated. Designed to examine the effects of an education cue on vaccine ac ceptance, Ashby Hughes, & Nickerson, 1999 sent two separate brochures, (one for influenza vaccine and another for pneumonia) to of their 463 participants (race not reported). They then followed up with an anonymous questionnaire intended to determine if the pamphlet had prompted participants to be vaccinated. The education intervention was non significant. Instead, healthcare provider recommendation was the strongest predictor for vaccine acceptance. Trust in Primary Care Provider is the belief t hat health care providers or medical institutions will act competently and with the best interests of patients in mind, and is influenced by interpersonal skills, as well as their medical (Mechanic, 1998, p. 342). Several authors have written about the constructs of provider trust, and trust in medical institutions; and how each affects patient satisfaction, health seeking behaviors, and medical outcomes ( Doescher, Saver, Franks, & Fiscella, 2000; Hall, Zheng, Dugan, Camacho, Kidd, & M ischara, 2002; Sohler Fitzpatrick, Lindsay, Anastos, & Cunningham 2007; Harris, Chin, Fiscella, & Humiston, 2006 ) Trust, is an essential tenet of any patient/clinician relationship, yet there are a number of articles ( Anderson,
48 & Dedrick, 1990 ; & Hall, et al. 2002) citing mistrust as a central issue to compliance with general preventative care A lack of trust in the healthcare system as a whole and in primary care providers are two variables which may be considered together or separately. Historical atrocities such as slavery, generations of systematic discrimination, and the Tuskegee incident have left an indelible mark on many African Americans. Although much progress has been made in recent decades, racism and stereotypes persist and undermine effo rts to eliminate health disparities. African Americans continue to suffer significant health disparities some of which have been attributed to discriminatory treatment within the healthcare system (Centers for Disease Control, 2005; Kunitz & Pesis Katz, 2 005; Smedley, Stith, & Nelson, 2002) Reports of verbally dominant physician patient interactions and bias related to beliefs about African Americans lesser intelligence and propensity for non compliance persist (Johnson, Roter, Powe, & Cooper, 2004). Som e of the intrinsic components of patient trust in healthcare provider include confidence in abilities, open communication, compassion, and confidentiality (Anderson & Dedrick, 1990; Mechanic, 1998; Mechanic & Schlesinger, 1996; Pearson & Raeke, 2000; Thom Ribisl, Stewart, & Luke 1999). Given the historical backdrop and lingering inequities in treatment, it is not surprising that African Americans consistently report higher rates of distrust. Bova, Fennie, Watrous, Dieckhaus, and Williams, ( 2006 ), maintain that high rates of mistrust provide partial explanation for persistent healthcare disparities among African Americans. In a community tracking study, Bova, et al. ( 2006 ), found that African Americans residing in cities, and those with lower education
49 lev els and no health insurance were significantly more distrustful. Moreover, African American men were less likely to trust their healthcare provider than women and results varied by geographic region. In other words, the patients who may need us most trust us the least. Sohler et al. (2007) examined the construct of trust in primary care providers with a group of marginally housed African with HIV, and found that at least 50% of the participants had some degree of distrust in their providers. The authors evaluated racial concordance to determine whether or not having a physician of the same ethnicity would decrease distrust, and although not significant, there was less distrust overall when racial concordance was present. When appraising the constr uct of trust in the healthcare system, researchers found that many of the study participants had deeply rooted concerns with respect to their rights to privacy and informed consent (Sohler, et al. ). Participants concerns ranged from thinking that their per sonal information would be shared to beliefs that unauthorized procedures and experiments might be performed on them. These alarming trepidations provide insight to deep seeded mis trust among some portion of the African American community (Corbie Smith, T homas, Williams, & Moody Ayers, 1999; & Freimuth, et al. 2001). When African Americans do have a trusting relationship with their healthcare provider, studies have shown increased adherence behaviors such as vaccine uptake, better follow up, and continui ty of care (Altice, Mostashari, & Friedland, 2001; Thom, et al. ,1999). Measuring trust as an independent variable and possible predictor of outcomes is still in its infancy. Previously when patient provider relationships were examined, the focus was on co mmunication and satisfaction. These concepts may be
50 related to trust, but they are not the same as the often more instinctual nature of trust itself (Hall, et al. 2002). The pneumococcal vaccine may be indicated given the age or risk factors; it may be readily available; the cost covered by insurance, and the patient may be sitting in the exam room, but if the patient does not trust the provider and/or the healthcare system in general, it is quite possible they will refuse the vaccine. Perceived Health and Susceptibility Health Little is known about the construct of perceived health as it relates to African Americans preventative or health seeking behaviors. A review of the literature revealed no findings using key terms: perceived health and im munization or vaccine status, even without factoring in the population of interest. Searching more broadly on perceived health and health seeking behaviors produced two studies (discussed below), only one of which includes the population of interest; there by uncovering a potential knowledge gap to be further explored in this dissertation study. Perceived health or self perceived health, is intuitively defined as individual's perception of his or her (Hunt, et al. 1980), and is increasingly bei ng utilized by healthcare professionals as an adjunctive measure to traditional assessments (Glover, Bellinger, Bae, Rivers, & Singh, 201 0 ). The construct of perceived health status represents a myriad of individual factors such as personal experience, ph ysical and functional disabilities, the presence or absence of diseases, and knowledge of diseases and consequences. In order to operationalize these concepts, they are placed into physical, emotional, and social
51 domains (Krokavcova, et al. 2009). The dec ision to accept or decline a vaccine such as PPSV23 is an In order to better understand the complex nature of the factors which influence this decision, it is necessary to ascertain information on the individuals perceived health status. Whe n comparing perceived health of African Americans with that of Caucasian counterparts, studies have shown African Americans consistently report lower perceived health scores (Cummings, & Jackson, 2008; Hunt, et al. 2011; Jin Sun, Bramlett, Wright, & Poon, 1998). The reason for this finding may be due to complex intersections between dimensions of trust, and socio economic status, among others. Lewis and Reigel, 2010, sampled 1,485 community dwelling adults aged > 60 with hypertension to identify determinan ts of perceived health. Consistent with other studies, perceived health entailed more than the obvious physical components intuitively associated with health. There were striking differences between African American and Caucasian reports of or health: (excellent health: 5% and 14.5% respectively), and (good health 42.6% and 52.6% respectively). Glover, et al. (2010) set out to determine racial variations pertaining to utilization of specialty care (cardiology, nephrology, etc.) among t hose with chronic illnesses such as hypertension. 58% of the respondents ( n = 6,722) self identified as African American. Participants with poor perceived health status were significantly less likely to receive specialty care (OR = 0.528, 95 % CI [ 0.354 0 .788 ] ). Not surprisingly, health insurance was also significantly associated with specialist care utilization, with uninsured participants 13% less likely to utilize specialty care (AOR = 0.138, 95 % CI [ 0.064 0.294 ] ) than privately insured counterparts. A measure of SES, health insurance or lack
52 thereof, may independently predict perceived health. Shields & Shooshtari, (2001) reported that in addition to physical factors, psycho social characteristics and socio economic status were significantly related to perceived health. A broader search of the literature revealed a study in which perceived health was found to influence vaccine uptake in one population. Influenza vaccine uptake is highly encouraged for healthcare workers whose frequent contact with ill patients is an unavoidable job hazard Hubble, Zontek, & Richards, (2011) sought to document vaccination rates of EMS professionals and identify predictors of vaccination uptake and refusals. Using a cross sectional survey designed to capture beliefs about vaccine efficacy, and influenza illness, researchers found that overall vaccine uptake among the 601 participants surveyed was 47.9% ( p = 0.01). Unvaccinated individuals perceived themselves as and thus, did not feel for acquiring infl uenza. Also relevant to this dissertation work, authors noted significant differences in vaccine uptake when the employer recommended it (OR = 3.6, p < 0.01). Perceived health in this study inversely predicted vaccine uptake. In other words, when perceived health was high (the individual believed healthy), vaccine uptake was low. While this study did not target the population of interest, the directional relationship between perceived health and health promoting behavior, provide support for the authors decision to examine how these variables interact in the pneumococcal study. This dissertation s ought not only to examine a potential relationship between perceived health and PPSV23 uptake, but also to evaluate the nature of the relationship betwe en the perceived health and the susceptibility dimension s of the Health Belief Model (HBM).
53 In a study examining predictors of influenza and pneumococcal vaccination, Ru Chien Reiber, & Neuzil, (2006) used logistic regression and identified factors such as older age; female sex, higher socioeconomic status, poor perceived health status, and the presence of chronic diseases as positive predictors for vaccination. When health status was perceived as 60% of participants reportedly accepted the PPSV23 vaccine, whereas health status participants had 70.3% vaccine uptake. Of the 39,377 study participants, vaccine acceptance by ethnicity was 65.5% among Caucasians and 41.9% among African Americans. Smokers who participated in the study had a 55.7% v accine uptake rate ; therefore African American ethnicity and current smoking practices were independent negative predictors for receipt of the pneumococcal vaccine (Ru Chien, et al. 2006). Importantly, those with current smoking practices were not furthe r broken down into race or ethnicity. Based on this review of the literature, it would seem that smokers do not consider themselves less healthy or more susceptible to vaccine preventable illnesses, however; much more information is needed to thoroughly ev aluate this notion (Looijmann, Verheij, van Delden, van Essen, & Hak, 2007). In this dissertation, the author include d smoking status in the questionnaire to evaluate what, if any significance there wa s with smoking and PPSV23 uptake specifically among Afr ican Americans. Perceived s usceptibility Perceived susceptibility is defined as degree to which individuals perceive and personalize the risk of acquiring disease or suffering the ill effects from existing (Weissfield Kirscht, & Brock, p. 19 90). Perceived risk or perceived susceptibility reflects an individual's belief about the likelihood of a given health threat such as pneumococcal
54 illness. Works analyzing the construct of perceived susceptibility have sought to determine why certain peopl e access preventative care, while others do not. Similar to what has been described for other variables, s tudies found in the literature examining perceived susceptibility of contracting vaccine preventable illnesses in the adult population focused largely on influenza. Although influenza is not the focus of this study, some of the findings can be helpful to understand beliefs and perceptions of vaccines in general. Gallant, (2007) describe d the personal decision individuals face when it comes to accepting or declining the influenza vaccine and aptly mention ed the fact that this decision may well be influenced by perceived vulnerabiliti es. In some cases, patients described their risks as low because they le d healthy life styles, ha d minimal exposure, and ra rely became ill (Gallant). Methodological problems exist with measuring the construct of perceived susceptibility, particularly when a person may not be familiar with a given disease or illness such as pneumococcal pneumonia. The Health Belief Model serves as a theoretical underpinning guiding the researcher with the task of operationalizing susceptibility. A meta analysis conducted to exam the relationship between risk perception and health behaviors found that although the terms and or are often used interchangeably in the literature; they are not the same ( Brewer, et al 2007) Questions for operationalizing these constructs were proposed: Perceived risk or Likelihood I get immunized, there is a high chance of me getting (Madhavan, Rosenbluth, Amon kar, Fernandes, & Borker, 200 0 ). Perceived susceptibility would be examined with a question pertaining to an
55 susceptibility to a given health hazard and operationalized with a L ikert measure of an item such as this: get sick more easily than other people my (Nexoe, Kragstrup,& Sogaard, 1999). Finally, the construct of perceived severity would be defined as the degree of harm a hazard would cause. In other words, a statem ent such as: pneumonia could lead to would measure the dimension of severity (Nichol, Lofgren, & Gapinski, 1992; Zimmerman et al. 2003). To explain health behaviors and examine predictors of influenza vaccination uptake, Chen, Fox, Cantrell, Stock dale, and Kagawa Singer, ( 2007 ) used key constructs of the HBM including perceived susceptibility, and perceived severity among ethnically diverse parishioners in a faith based congregation. Subjects were stratified by age, race, education, and gender to assess for modifying variables p erceived susceptibility to influenza was test ed using a three level l ikert response question: concerned are you about getting the Perceived severity of getting influenza was measured similarly: would getti ng the flu affect your One open ended question for non vaccinators was included: is the main reason you did not get a flu shot in the past 12 response s were identified, one of which was do not need Caucasian and African Ameri can subjects concerned about getting the flu were significantly more likely to be vaccinated (96% and 91%, respectively), compared with those who were not concerned (45% and 33%). Forty five percent of African Americans reported being not at all concerned about getting influenza compared with 35% of Caucasians ( p < 0.01) (Chen, et al. ). Other responses relevant to this dissertation work, not in the category of perceived health were: 1) influenza vaccine causes influenza, 2)
56 access and cost issues, 3) lack o f knowledge of vaccine and 4) healthcare professionals did not recommend. Acceptance or declination of a vaccine such as PPSV23 is a personal decision. In order to better understand the complex nature of the factors which influence this decision, it is ne cessary to ascertain information on the individuals perceived health status. The constructs of perceived health and susceptibility represent a myriad of individual factors such as personal experience, physical and functional disabilities, the presence or a bsence of diseases, social support, and knowledge of diseases and consequences (Shields, & Shooshtari, 200 1 ; Lewis, & Riegel, 20 10 ). The construct of perceived susceptibility (a component of the HBM) w as therefore measured in two ways: first with two si ngle item questions asking participants if they got sick easily or considered themselves more vulnerable to illness, and second with the proxy variable of perceived health The reasons for this were twofold. First, the author believe d perceived susceptibil ity wa s inextricably tied to perceived health. In other words, if one perceives themselves as healthy, they are less likely to feel susceptible to contracting a given illness Second, it was based on the premise that an adult would have to believe they wer e susceptible to pneumococcal illnesses such as pneumonia for the preventative health behavior (immunization) to take place (Becker, et al., 197 7 ; Rosenstock, 1974). Prior Negative Experience with Vaccines Review of the literature revealed a portion of vaccine refusals, stemming from some prior negative experience with immunization either personally or through an acquaintance (Rosenthal, 2007). Many elderly participants refused the influenza
57 vaccine because they felt that they had become ill soon after they received it in years prior (Telford, 2003). In the same study another group of individuals refused the vaccine because people they knew reported similar experiences and they felt it best not to take any chances. Prior negative experience with immuniz ations, even if indirect, could impact decisions and lead to refusals. A study pertaining to the challenges with the Human Papilloma Virus (HPV) gave further credence to this notion (Rosenthal, 2007) Five focus groups with women of varied age, ethnicity, and background revealed a few participants who knew of others who had significant side effects from a vaccine, and even one who reported that she knew a person who developed syndrome after receiving a vaccine Another participant vividly recalled a story her g randmother told about how her two brothers had died of diphtheria back when no vaccine was available. She was clearly a proponent of vaccines and passed those feelings along to her offspring (Rosenthal). As health professionals, it might be easy to discount these prior experiences and accounts; however, the impact these stories have on others (regardless of their factual basis) cannot be denied. Harris et al. ( 2006 ) used semi structured qualitat ive survey questions and asked 20 African American s 65 reasons why they had not received pneumococcal and influenza vaccinations. The majority unvaccinated participants felt that had become ill or gotten the flue after an influenza vaccine. Subjects stated that many African Americans get their vacci nes because they are worried they will get sick instead of actually prevent ing an illness. Moreover, negative vaccine experiences extended beyond
58 concerns that vaccines could cause illness. A 70 year old male participant recounted his traumatic experience with vaccines as a child: I remember when I was a boy in the South; we had to take shots for everything right until the fifth grade. And the nurse down there treated you like you were an animal. They did not care. They were not sensitive. They would just jab you in your arm like you were an animal. That's how they treated us, you see. So I don't want any shots. I still have those memories. (p. 1680). Summary Literature clearly supports pneumococcal vaccine uptake for adults age 65 and over as well as thos e between the ages of 18 and 64 with existing pre existing health conditions such as diabetes, heart disease, lung disease, liver disease, organ transplant, cochlear implants, asplenia, smoking, and compromised immunity. The literature also highlights a ga p in vaccine uptake despite a decade worth of efforts to improve and meet Healthy People 2010 goals ( CDC, 2010; Harris, et al., 2006 ) Although studies have been conducted to examine vaccine uptake, there have been very few studies specific to pneumococcal vaccine uptake among African Americans as a subset of the population. It was therefore not surprising that the author did not find a central theme lending itself to a parsimonious approach to the problem of insufficient PPSV23 uptake among vaccine eligibl e minorities. Factors such as knowledge deficit s trust in healthcare provider, perceived susceptibility of contracting pneumococcal illness, perceived health, and prior experience with vaccines, are potential barriers to vaccine uptake overall. The decisi on to look carefully at PPSV23 vaccine uptake specifically is largely due to the fact that pneumococcal illnesses among African Americans occur at nearly twice the rate as their White counterparts (24 per 100,000 and 12.2 per 100,000 respectively) (CDC, 20 08).
59 The novelty of this study lies in the integration of the five aforementioned variables believed to influence PPSV23 vaccine uptake. While some of the studies relevant to this dissertation work examined one or more of the potential predictor variables, none have examined all of them. Much of the data collected in pneumococcal vaccine uptake studies has focused on awareness, and while many have found significance with this variable, none have made a distinction between awareness and knowledge. Trust in h ealthcare provider and PPSV23 vaccine uptake had not been previously explored. Perceived health established with one question from the BRFSS was asked in two studies found in the literatur e; however, perceived health has been defined by experts as a comple x construct and therefore may not lend itself well to a single item question. No studies using SF as a measure of perceived health to examine the potential relationship between perceived health and vaccine uptake were found. Components of the Healt h Belief Model have been utilized to examine vaccine uptake in a number of studies. In the pneumococcal study, the author has placed insurance status in the model as a potential barrier to vaccine uptake based on prior studies where SES was deemed signific ant. Those wi th Florida Medicaid benefits were of particular interest with this variable since vaccine cost is not covered for them. Disparities in PPSV23 uptake are well documented and must be addressed. A parsimonious approach to the problem may not be f easible given the complexities of the decision making process. Through exploration of the aforementi oned variables, the author sought to contribute to what is already known, and build new knowledge with an eye toward targeted interventions and a reduction in health disparities.
60 CHAPTER 3 METHODS Theoretical Perspective Theory Overview In order to investigate the relationships between the dependent variable ( pneumococcal vaccine uptake or willingness to accept) and the independent variables: knowledge, trust in hea lthcare provider, perceived susceptibility perceived health status, and prior experience with vaccines; two theoretical models; the Health Belief Model (HBM) and the Precaution Adoption Process Model were evaluated and subsequently substructed to serve as the underpinnings for this study. Health Belief Model First, the HBM developed by Hochbaum, Leventhal, Kegeles, and Rosenstock in 1950, will be used to evaluate and explain individual differences in preventative health behavior. Perceived susceptibility, perceived seriousness, perceived benefits, perceived barriers, and cues to action comprise the key components of the HBM The belief that a specific health action such as pneumococcal vaccine uptake might prevent illness can be interpreted and explained t hrough the lens of the HBM. The HBM proposes that action for prevention such as pneumococcal vaccine uptake will occur if the individual perceives themselves as susceptible to the potentially serious illness. When the perception of susceptibility exists, t he individual must also weigh potential consequences determining whether a particular action is beneficial in decreasing susceptibility or severity of the condition, and if the benefits for the action outweigh the barriers ( Hochbaum et al. 1958 ).
61 Precaut ion Adoption Process Model The second theoretical framework to be utilized in this dissertation is the Precaution Adoption Process Model (PAPM). Adoption of a new precaution (healthy behavior) serves as the construct for PAPM which was developed by Weinste in in 198 9 The five stages of this model are as follows: "unaware of the issue," "aware of the issue but not personally engaged," "engaged and deciding what to do," "planning to act but not yet having acted," and "acting" (Weinstein ). For the purpose of t his study, the most important distinction between the HBM and Precaution models lie in the precontemplation component; specifically, the inclusion of of the healthy behavior and its implications. Although the underlying assumptions of the PAPM ar e not explicitly stated, review of the seminal works (Weinstein, & Sandman, 1 989 suggest that the parallels between the PAPM and the Transtheoretical model are many, thus for the purpose of this dissertation the underlying assumptions are considered to be as follows: b ehavioral change is a process that unfolds over time through a sequence of stages; stages are both stable and open to change; there are a common set of change processes that people apply across a broad range of behaviors; and without planned interventions, individuals or communities may remain stuck in the early stages (Edberg, 200 6 ). Review of the literature for pneumococcal vaccine uptake utilizing the PAPM and HBM frameworks will now be presented. P recaution Adoption Process Model and Pneum ococcal Vaccine Uptake While there were no findings in the literature linking PP S V23 vaccination to the PAPM, there were findings for other preventative behaviors. Weinstein and Sandman ( 1989 ) first presented PAPM as a theoretical model for examining the likelihood that
62 individuals would test for Radon in their homes. The initial stages of the theory, U naware of, or unengaged by the issue played a crucial role in the same research findings. Since the Radon studies, there have been a number of studies wh ere PAPM has been tested including a study on Hepatitis B vaccination (de Vet, 2008). In summary, PAPM is an attractive possibility as a theoretical framework for the pneumococcal research for two reasons: 1) the variable, and 2) the unidirection al nature of the model. Further support for the PAPM theory is illustrated when examining a qualitative study conducted by the Office of Health Communications and the Centers for Disease Control and Prevention whereby a comprehensive summary of what Afric an Americans in five US cities think about the p neumococcal v accine was explored ( Winston, Pascale Wortley, & Lees, 2006) In this study, participants perceived pneumonia to be dangerous and in some cases life threatening ; however most respondents were un aware of the existence of the pneumococcal polysaccharide vaccine. Comments such as, my doctor tell me about were recorded during the focus groups. While many of the participants knew little or nothing about the vaccine, there were some wh o were aware of the basic indications for the vaccine and many knew that it was not administered annually like the flu shot. A few of the respondents believed that the vaccine could protect against the flu as well. None of the respondents felt that the vac cine itself would cause illness. This was a favorable response to researchers who felt that with proper introduction, the PPSV23 vaccine might be well received within the target population. Reasons for non vaccination mirrored the excuses for flu shot refu sal
63 in that most participants said they were healthy and would not get pneumonia (CDC, 2003). H ealth Belief Model (HBM) and Pneumococcal Vaccine Uptake While few studies examining pneumococcal vaccine uptake with the HBM as a theoretical underpinning were found in the literature, there were two poignant studies on influenza vaccine uptake with relevance to this research. In an effort to better understand barriers to influenza vaccination, Nexoe et al. (1999) constructed and validated a questionnaire using the HBM as a theoretical underpinning. The 46 item instrument was sent to 2 147 participants 65 and revealed that the HBM positively predicted influenza vaccine acceptance with perceived benefits, perceived barriers, and perceived severity constructs all being significant in the model. Th e HBM was also utilized in a study examining five ethnic g roups and their vaccine determinants. A telephone survey of aged 50 75 was conducted in Los Angeles and Honolulu specifically addressing the constructs of susceptibility, perceived severity, and self reported barriers to vaccination. Findings revealed that Whites and African Americans who were concerned about getting the flu (perceived susceptibility) were significantly more likely to be vaccinated (96% and 91%, respectively),compared with those who were not concerned (45% and 33%) (Chen, et a l. 2006). Authors in this particular study chose to explore only the constructs of perceived susceptibility, perceived severity, and barriers to vaccination, omitting the constructs of perceived benefits, cues to action, and self efficacy. They rationaliz ed the decision to amend the model based on empirical evidence and the uni directional nature of vaccine uptake.
64 HBM S usceptibility Severity, and H ealth The first construct of the HBM is perceived susceptibility which is defined as a subjective percept ion of the risk of an illness (Glanz, & Schwartz, 2008). belief regarding their individual risk of being diagnosed with a specific disease process or illness in the immediate or long term future may be based on personal characteristics or behaviors. Perceived severity is the second construct of the HBM. Perceived severity is belief about the seriousness of a medical condition and the ramifications of a given diagnosis. Consequences could be minimal such as a brief disruption of daily activities or quite severe including loss of work, long term disability, chronic pain, or death. The concept of a perceived threat merges the construct of susceptibility with that of severity in the HBM. Perceived severity, formerly known as perceived seriousness, i s defined as perceived morbidity and mortality (Glanz, 2008). For this dissertation study, perceived health will be addressed using the SF to be discussed in detail under the methods section. Operationalizing the Theoretical Model s Methodological pr oblems exist with measuring the construct of perceived susceptibility, particularly when a person may not be familiar with a given disease or illness such as pneumococcal pneumonia. The literature on health seeking behaviors such as vaccination and the con structs of perceived risk, susceptibility and severity were therefore examined closely. Questions for operationalizing these construct were proposed by Nexoe, Kragstrup, and Sogaard (1 999) Within this study, a series of questions were used to determine p erceived risk susceptibility, and severity. Questions
65 such as: I get immunized, there is a high chance of me getting were used to evaluate risk Thought to be a different construct, p erceived susceptibility w as examined with a questio n pertaining to an susceptibility to a given health hazard and operationalized with a Likert measure of an item such as: get sick more easily than other people my Finally, the construct of perceived severity which was defined as the d egree of harm a hazard would cause was evaluated with a question such as: Pn eumonia could lead to (Nichol, et al. 1992; & Zimmerman et al. 2003). These studies provided an excellent blueprint for evaluation of perceived susceptibility and severit y in the pneumococcal study. decision to accept or decline a vaccine such as PPSV23 i nvolves a complex multifactorial process, therefore, it is important to ascertain information on the individuals perceived health status. The construct o f perceived health represent s a myriad of individual factors such as personal experience, physical and functional disabilities, the presence or absence of diseases, social support, and knowledge of diseases and consequences (Shields, et al. 2002; Lewis, & Riegel 2009). Examination of relevant works on perceived health along with instruments such as the SF which are designed to measure self reported perceived health provided a framework to guide the pneumococcal study. PAPM V accine Awareness and K n owledge The construct of awareness is not addressed in the HBM which is why the PAPM has been included in the theoretical framework. Using the P APM model, the author ascertain ed participants awareness of the PPSV23 existence. Knowledge of the pn eumococcal vaccine w as addressed using a single item on the questionnaire
66 whereby participants w ere asked, H ave you ever heard of the pneumococcal Scoring for this item is discussed in the measures section. HBM P erceived B enefits According to the HBM, perceived benefits refer to belief in the various disease reducing effectiveness. Perceived benefits are belief in the efficacy of the advised action to reduce health risk Also termed as perceived benefits of taking health a ction, the attitudes of health behavior changes are reliant on view of the health benefits for performing a health action (Glanz & Schwartz, 2008 In this study perceived benefits w ere addressed indirectly by examining the constructs of perceived h ealth, susceptibility, and severity. These constructs w ere measured using the SF and single item questions all of which are discussed in detail in the methods section of th is dissertation. HBM Pe rceived B arriers Perceived barriers refer to the pot ential negative aspects of or obstructions to taking a recommended health action (Glanz, & Schwartz, 2008). This can also be conceptualized as individual belief s about the physical and psychological costs of taking health action A n internal cost benefit analysis occurs, weighing the health expected effectiveness against perceptions that it may become an obstacle. Potential barriers may include financial expense, fear/ perceived danger, pain, difficulty, inconvenience, and time consumption Percei ved barriers to PPSV23 uptake could be emotional (a fear of shots), social (trust or lack thereof), or physical (concern over a past experience with a vaccine) (Glanz & Schwartz ). In this study, perceived barriers w ere identified by asking participants a single item question such as whether they fear ed
67 needles or had a prior negative experience with vaccines, and having them complete the Wake Forest Trust Scale which w ere discussed in detail in the research design section. HBM C ues to A ction Cues to act ion are the strategies or internal incentives that initiate readiness to take a health action such as vaccination. Considered difficult to study, Glanz & Schwartz, (2008) suggest that a cue to action could range from media announcements to a person coughing or sneezing. Cues to action w ere operationalized in this dissertation as a recommendation by a healthcare provider to receive the pneumococcal vaccine. This cue to action may or may not be present; therefore the researcher asked one question on t he VUQ to ascertain the existence of this cue : your healthcare pr ovider ever suggested you get the pneumococcal In summary, each of the independent variables in the pneumococcal study were informed by the theoretical model, categorized acco rding to the constructs awareness, perceptions and modifying factors, and assessments and then operationalized in the VUQ through a series of questions designed to predict PPSV23 uptake Figure 3 1 provides a visual representation of the substructed theor etical framework used for this study.
68 Figure 3 1. Health b elief and p recaution a daptation substructed m odel Assessments Assessed Sum of Benefits minus Perceived Barriers for Preventative Action (prior negative experience with vaccines, insurance ) Perceived Threat of Illness or Injury (Severity) PAPM : Knowledge/Awareness of PPSV23 V accine Likeliho od of Action L ikelihood of Preventative Action ( PPSV23 Uptake ) Perceptions and Modifying Factors Perceived Seriousness & Susceptibility ( p erceived h ealth ) Demographic Variables (a ge g ender & education) S ociopsychological Variables (t rust in healthcare pr ovider ) S tructural Variables Cues to Action (provider recommendation)
69 Research Design This study used a quantitative cross sectional exploratory design Based on the literature review concerning pneumococcal vaccine uptake among individuals who self identify as African Am erican, th e proposed study s ought to fill in the knowledge gap by performing a n exploratory study to examine pneumococcal vaccine uptake among A A adults in Northeast Florida I ndependent variables of interest were : vaccine awareness, vaccine knowledge, trust in heal thcare provider, perceived health, perceived susceptibility of pneumococcal illness, perceived severity prior experience with vaccines and the impact of healthcare provider recommendation. Demographic information w as also be obtained to further detail th e characteristics of this population and examine possible relationship s between PPSV23 uptake and ag e, gender type of insurance, and education. Insurance and education served as proxy variables for socio economic status in this study. In order to ensure t hat the instrument and research procedures were culturally sensitive, a pilot study using focus group methodology was conducted prior to the full study. T he pilot study is discussed in C hapter 4 of this dissertation. Sample and Setting The population being investigated in this study includes African Americans 65 and those 18 64 with pre existing conditions making them eligible for PPSV23 vaccination residing in Northeast Florida. To determine an adequate sample size, two predictor variables were considere d: awareness of vaccine, and cueing. From the literature it is estimated that 60 % in this population are aware of the vaccine and 50 % are cued by their health care provider to receive the PPSV23 vaccine (CDC, 2011 ;
70 Winston et al. 200 6) It was therefore d etermined that a sample size of 300 would provide adequate power (i.e., power greater than .80). The power analysis was conduc t ed using the POWER procedure in Statistical Analysis System ( SAS ) for logistic regression (version 9.2, Cary, N.C.). The level of significance was set at 0.05. Assuming an unvaccinated rate of .60 and a dichotomous predictor with prevalence between .30 and .70 (for example, between 30% and 70% of the sample report that they are aware of the vaccine), the detectable odds ratio for va riable significance is at least 2.1 for a sample size of 300 with power at least .80 (Faul, Erdfelder, Buchner, & Lang, 2009) This study use d convenience and snowball s ampling and participants were recruited at two predominantly African American churches in Jacksonville/Northeast Florida. Participants were recruited utilizing a pre screening tool ( Appendix B ). Recognizing the fact that recruiting African Americans in the community m ight be challenging, a n AA community informant actively involved in the ch urches was hired as a research assistant (RA) to facilitate recruitment and participation Inclusion and Exclusion Criteria Participants inclusion criteria were : a) adults ages 18 and up who self identified as AA and whose age or pre existin g conditions rendered them PPSV23 eligible ; b) those who spoke fluent English ; c) individuals who were capable of completing a questionnaire either independently or with the assistance of the principal investigator ( PI ) or RA Eligibility was determined u sing a pre screening tool which is discussed in detail in the next section. Exclusion criteria included : a) a ge less than 18 ; b) p regnant or breastfeeding women Using the suggested vaccination guidelines published annually by the Centers for Disease Contr ol, children under the age of 18 would be eligible for
71 the PCV7 vaccine, rather tha n the PPSV23, and were therefore excluded from this study. The decision to vaccinate women who are pregnant or breastfeeding is a highly personal choice fraught with controv ersy. Tillet (2004) published a paper compiling the basic guidelines for immunization during pregnancy and breastfeeding along with recommendations that the practitioner have a discussion with the patient assessing risks and benefits carefully befor e making a decision. PPSV23 is considered safe for administration during the second and third trimesters of pregnancy as well as during breastfeeding, and is thought to confer some immunity to the newborn. That said, pregnant mothers without pre existing m edical conditions such as diabetes, heart disease, cancer, COPD immunological suppression sickle cell disease, or current smoking habits, would not be eligible for PPSV23. Expect ant mothers with one or more of the pre existing conditions are not only eli gible, but highly encouraged to be vaccinated with PPSV23 (Tillet). While this dissertation examined knowledge, trust, perceived susceptibility perceived health, and prior exp erience with vaccines; it did not address the variable perceived susceptibility for harm to fetus/newborn which would implicitly be present with an expectant or breastfeeding mother. Given the relatively low number of pregnant women who would typically be eligible for PPSV23, the controversial nature of vaccine uptake while pregnan t, and the additional variable of perceived susceptibility to fetus/newborn, pregna nt and breastfeeding mothers were excluded from this study. Eligibility Screening The populat ion of interest for this study wa s African Americans whose age or pre existing conditions render them eligible to receive the PPSV23 vaccine ; thus, a pre
72 screening t ool was designed to determine eligibility. Studies in the literature where participants were not pre screened for vaccine eligibility, found that roughly 44% were not eli gible for the vaccine (Allen, 2010). If this were true in the pneumococcal study, nearly 600 participants would be needed, necessitating a much larger budget and additional settings. Following IRB approval, the on e page screening tool was distributed among members of the congregation at each of the chur ches. This tool was not collected; it was merely a tool for potential participants to determine if they were eligible. The tool ask ed the participants for their age, gender, ethnicity, and whether or not they had any of the following conditions: heart disease, lung disease, kidney disease, liver disease, cancer, immunological suppression asplenia, cochlear implants, sickle cell disease, or diabetes. Ad ditionally participants were asked if they smoke d At the end of this brief questionn aire, a statement was included telling the participant that if they have answered to any of the conditions above, OR they were 65 years of age, they we re eligible to participate in a questionnair e based research study that would take approximately 15 minutes of their time If willing to participate, info rmation on when and how they could obtain the questionnaire was be provided on the lower half of the pre screening tool In addition to the pre screening instrument, the study was advertised on bulletin boards at the churches and via word of mouth. The total estimated population of the first church was 350 and the estimated popula tion of the second church was 870. These estimates were provided by support staff for each of the churches and may not be reflective of the actual number o f attendees on a given Sunday. Of the 1,220 possible subjects, 329 self identified as eligible based on the screening criteria and
73 ultimately participa ted. Again, screening tools were merely posted, not collected; therefore no one ac tually refused to participate. Rather, those who fell into the eligibility criteria listed and felt inclined to participate approached the PI or RA following services to obtain and complete a questionnaire. Measures The primary instrument for this study known as the Vaccine Uptake Questionnaire (VUQ) was a self administered paper and pencil questionnaire written in English a t a 5th grade education level for ease of understanding ( Appendix C ). The PI and RA were available on Sunday s following church services for the duration of the study to assist participants who experienced any difficulty completing the questionnaire. The qu estionnaire took approximately 15 20 minutes to complete, and participants were asked to place it in a sealed envelope prior to turning it in. The VUQ included demographic data such as age, ethnicity, education, insurance, chronic disease diagnoses, smoki ng history, alcohol intake, and vaccination history. Additionally, the VUQ asked questions related to the research variables of interest utilizing a compilation of three validated instruments whose authors granted expressed written permission for utilizati on in this study : 1) Behavioral Risk Factor Surveillance System (BRFSS) 2) The Wake Forest Trust Scale and 3) SF 12 version two Single item questions addressing knowledge, perceived susceptibility perceived severity, and healthcare recommendation were al so be included in the VUQ. Each of the instruments as well as the construction of single item questions will now be discussed in detail. The dependent variable PPSV23 uptake is dichotomous categorical variable and wa s reported as 0 = unvaccinated, 1 = vacci nated, and 2 = unsure. Trust in healthcare
74 provider is an independent scale level variable with scores ranging from 1 to 5. Perceived health susceptibility and severity were independent scale level variable s with Likert based scor ing Prior negative experi ence with vaccines was an independent dichotomous categorical variable where 0 = no past negative experience and 1 = past negative experience. Univariate descriptions of the scale variables: trust, perceived health susceptibility and severity were analyzed to obtain a mean, SD range, minimum and maximum. Awareness, prior negative experience and PPSV23 vaccine uptake were categorical variables ; consequently, frequency and variance data were obtained and reported Pneumococcal Vaccine Uptake The outcome var iab le for th is dissertation study was determined by self reported compliance with pneumococcal vaccination. Intended to be a dichotomous response, the question: you ever received the pneumococcal vaccine (sometimes known as the pneumonia wa s asked of all participants. Answer choices were 0 = no, 1 = yes, and 2 = unsure. From a clinical standpoint, when a patient is unsure of his/her vaccine status, the recommendation is to administer PPSV23 (CDC, 2009). Given these empirical guidelines, subj treated as unvaccinated (0 = no) for analyses. Refusals Respondents who answered to PPSV23 vaccine uptake, were given an opportunity to inform the researcher whether or not they had refused the vaccine in the past. Those who did refuse could then check boxes or write in an explanation for their decision on the questionnaire. Refusals have often been eliminated from prior studies
75 found in the literature Arguably, are more important th an anyone else in the study to better inform healthcare professionals of their thought processes and rationale for the decision. For the purposes of this dissertation and data management, those who refused the vaccine were be treated as 0 = no. Prior to co llapsing the category however a priori plans to examine this group separately for unique characteristics and responses were carried out. Unsure According to Mieczkowski, and Wilson (2002), a large number of patients may not recall whether or not they have been vaccinated. ACIP recommends that if a status is they should be re vaccinated. Moreover, they may not remember whether or not their healthcare provider has recommended the vaccine. This is highly problematic for researchers attempt ing to accurately uncover the full extent of the pneumococcal vaccine uptake disparity. Methods for handling know/unsure responses varied in the literature. In two separate studies conducted by the CDC respondents who reported unknown pneumococcal va ccination status were excluded from the analysis. The 2000 study reported 1,037 adults 65 with unknown PPSV23 status as having been excluded compared with only 212 who reported unknown influenza vaccine status (NIH, 2000). Believed to be an important co mponent o f the study, a decision was made not to eliminate these subjects from data analyse s. Instead, this sub group was analyzed separately, for unique characteristics, then ult imately subjects were placed into the category since the recommendation from ACIP and the CDC is that they be treated as unvaccinated. These patients were not eliminated. Instead, to was
76 measured in a priori analyses prior to collapsing the variable by asking subjects you answered uns ure would you b e willing accept the vaccine if it were offered to you today ? Behavioral Risk Factor Surveillance System The Behavioral Risk Factor Surveillance System ( BRFSS ) is a state based collection of health surveys established by the Centers for Disease Control i n 1984 as a means of collecting relevant health data in the United States, the US Virgin Islands, and Guam (CDC, 2011) Each year, the BRFSS collects information on more than 350,000 adults making this instrument the largest telephone based survey in the w orld. Questions from the BRFSS were developed by a federal agency and are considered to be public domain ; accordingly they may be utilized without permission provided they have been properly referenced and the core component questions are administered wit hout modification. Core module topics include such items as seatbelt use, hypertension, arthritis burden, health status, tobacco use, and immunizations. Optional m odule topics are variable by year and have included survey items such as smokeless tobacco, o ral health, cardiovascular disease, and firearms (CDC BRFSS, 2011) For this dissertation, the BRFSS primarily informed the demographic questions ; particularly the verbiage utilized to ask questions regarding health seeking behaviors, age, education status, insurance status, smoking, and immunization uptake. Trust in Health Care Provider Wake Forest Trust Scale trust in their healthcare provider w as measured by using a 10 item instrument known as the Wake Forest Physician Trust Scale (WFPTS) Fi rst developed by Hall, et al. in 1992, the WFPTS measures trust through a multidimensional model
77 conceptualized through review of seminal works on trust with high internal consistency and reliability scores 0.93). The authors of the trust scale used the theoretical and empirical works of others (Luhmann, 1973; Goold, 1998; Mechanic, 1998) to conceptualize and test five domains: 1) fidelity, 2) competence, 3) honesty, 4) confidentiality, and 5) global trust. With the theoretical model and constructs in pl ace, Hall, et al. (2002) developed questions, pilot tested them, revised based on factor analyses, then field tested 26 candidate items nationally via telephone survey ( n = 959). Items demonstrating a high degree of reliability were retained and ultimately the authors published long and short versions of the Trust instrument ( Appendix E ). Permission to utilize the WFPTS was obtained from the author (Appendix F). In a national sample of 1 045 adults with primary care relationships including physicians and practitioners, mean scores of the WFPTS were 20.4 (77.0 on a scale of 100), SD = 3.1 (15.5), alpha = .87. The 10 item instrument is scored 5 1, for strongly agree, agree, neutral, disagree, strongly disagree. Items 2, 3, and 8 are reverse scored. A score of 50 (or 100 on a scale of 100) indicate s complete trust in healthcare provider, whereas a score of 10 (20 on a scale of 100) indicates complete distrust. Sample questions from this instrument (not in same numbering order) are: 1) [Your doctor/healthcare pro vider] is extremely thorough and careful, and 2). You completely trust [your doctor's] decisions about which medical treatments are best for you. In the VUQ, questions 22 31 are drawn from the Wake Forest Trust Scale questions. Perceived Health, Susceptib ility, and Severity SF When perceived health is higher (the patient believes he/she is very healthy), it was hypothesized that there w ould be an inverse relationship with vaccine uptake.
78 Several studies: (Glover, 2010; Hunt, 1980; Ru Chien, et al., 2006) have emphasized the importance of obtaining subjective health status information from individuals to examine the relationship between perceived health and susceptibility and accessing preventative services. For this study, perceived health was exami ned using the SF ( Appendix G ). A well known instrument for evaluating gener al health status utilizing a 12 item short form self reported questionnaire, the SF has demonstrated high degree of validity and reliability. Permission to utilize the i nstrument and scoring software were obtained from Quality Metrics ( Appendix H ). The SF involves a complex theoretical scoring system to be conducted using software. Raw scores using a summative approach are also possible with 56 being the highest pos sible score for perceived health and 10 being the lowest. Resnick and Parker, 2001 proposed scoring the SF 12 by reverse weighting four items and using cumulative scores from each category, and found a high degree of internal consistency on a pha > .70. Given the precedence set by this and other researchers ( Magery, Johnsen, & Moen, 2007) for utilizing raw scores in lieu of the more complex analyses requiring training and specialized software the simplified method using cumulative scores was selected for analyses of the SF 12 data withi n the VUQ Sample questions from this instrument (not in same order) included items such as: I n general you would say your health very good, good, fair, Indirect measures of health status w ere examined with questions such as, your health now limit you in these cleaner, climbing several flights of In the VUQ, questions 58 64 made up the SF component of the survey.
79 Single Item Questions Typically used to measure global constructs such as attitudes, health status (DeSalvo, Fisher, Tran, Bloser, Merrill, & Peabody, 2006), readiness to change (Williams, Horton, Samet, & Saitz, 2007), and symptom severity; single item questions have often proven to be nearly as effective as multiple item questionnaires (Bergkvist, & Rossiter, 2007; Youngblut, & Casper, 1993). Single item questions were created to measure awareness, knowledge and theoretical components of the HBM including perceived susceptibility perceived se verity perceived barriers, and cues to action ; specifically healthcare provider recommendation Awareness was measured with the question: Before this study had you ever heard of the pneumococcal vaccine sometimes known as the pneumonia vaccine? Responden t options for this question were yes/no. Knowledge was measured using two separate questions the first of which was: Do you know who is supposed to get the pneumococcal (pneumonia) vaccine? Respondent options for this question were yes, no, and unsure. Tho se who answered unsure were ultimately collapsed into no for the purposes of data analyses. The second question designed to measure knowledge had five item likert response options ranging from strongly agree to strongly disagree. have enough information to decide whether or not to accept the pneumococcal (pneumonia) vaccine. Constructs derived directly from the HBM were also measured with single item questions. Perceived susceptibility was measured with two single item questions: 1) I get sick more easily than others, and 2) I am at risk for getting pneumonia. Each of these questions had five point likert response options ranging from strongly agree to
80 strongly disagree. Perceived severity was measured using two additional questio ns: 1) I believe pneumonia is serious, and 2) I believe pneumonia can cause death. Again, five point likert responses ranging from strongly agree to strongly disagree were offered on the VUQ. Finally, p erceived health ( also measured with the SF ) was evaluated using a single item question: consider myself to be with a five level l ikert response option from strongly agree to strongly disagre e While the single item questions for perceived susceptibility, severity, and health were constructed explicitly for the purposes of the pneumococcal study, there were many other studies using the HBM to examine health seeking behaviors with s imilarly constructed questions found to perform reliably. Prior N egative E xperience with V accines Using the HBM a rior negative experience with va ccines was considered a barrier to vaccination. This construct was measured using two question s : 1) Have you ever gotten sick after getting any type of vaccine, and 2) Has anyo ne you know, other than yourself, ever gotten sick after getting any type of vaccine. Res ponse choices for these question s w ere dichotomous yes/no and although the study itself was quantitative, a free text option was provided for subjects to explain any symptoms, side effects, or complications they might have experienced. Provider R ecommendation for V accine Another component of the HBM known as cue to action served as the framework provider recommendation for PPSV23 vaccine This cons truct was measured by asking : your healthcare provider recommended that you receive the
81 pneumococcal vaccine sometimes known as PPSV23 or the pneumonia Response choices for this question were yes, no, and unsure. Ultimately, those who answe red unsure, were treated as no for statistical analyses since intuitively it would seem that a subject would remember if their healthcare provider had recommended something to them. Data Analysis Plan Analyses were conducted using the Statistical Package for Social Sciences software (PASW Statistics version 20.0). Univariate descriptive statistics were analyzed and reported. Next, Ch i square, analyses were used to for bivariate measurement between PPSV23 uptake (DV) and each of the IVs Once bivariate anal yses and a priori comparisons (discussed below) were completed, a logistic regression model was built and run using a backward stepwise method to assess the overall pr edictive value of the model, assess model fit and to interpret the coefficie nts. This met hod was chosen as the primary statistical technique because it allowed exploration of the influence and predictive nature of all variables whether scale level or categorical on the outcome variable. A 2LL. Nagelkerke pseudo R 2 cautiously predicted the pe rcen t of variance the model predicted and the Hosmer and Lemeshow test showed whether the predicted probabilities match ed the observed probabilities. Normality of distribution is not required for logistic regression analysis; however a histogram was scree ned to determine if data were severely skewed. Additionally, b ivariate descr iptions of each IV to IV were run using correlations to prescreen for multicoll inearity. The regression was assessed with a model in which variables were ordered and entered by the ir level of significance in the bivariate analyses.
82 Outliers were identified by using ZRE (standardized resi dual) scores. ZRE scores were examined and 99% of the cases had values < 2.5 and 95% of the cases should had values < 2. Outliers were evaluated to see if they we re influential cases using dfbetas. Any in fluential cases (dfbeta >1) were examined. Two influential cases we re found and determined to be miscoded. Examination of the original surveys revealed the entry errors, and corrections were made acc ordingl y. No exceptional cases were identified Assumptions were t ested. Multicollinearity was prescreened using IV IV bivariate analyses to obtain correlations and evaluate shared variance. Independence of error terms were examined with a null plot and a Durbin Wat son statistic. Crosstabs were run for each dichotomous pair to check f or empty cells. If empty cells were present, the data were re categorized. Collinearity diagnostics were run and VIF values checked. VIF values >10 indicate multicollinearity If the a ssumption of multicollinearity was violated, eigenvalues were reviewed. Only the variables awareness and knowledge were identified as offenders with Eigenvalues > .5 Ultimately, a decision was made to remove the knowledge variable from the model given the fact that a person cannot have knowledge of something if they are unaware of its existence. With the knowledge variable removed from the model the regression was rerun without further violations. A P riori Analyses Upon completion of data collect ion, there were a number of variables well suited for comparison in a priori analyses prior to running the logistic regression model. Among the variables of interest, gender and age categories w ere evaluated to determine if there were differences in refusa l rates between the groups.
83 Hypothesis Testing The hypotheses tested in this analysis we re as follows: Hypothesis 1 Sociodemographic variables such as a ge, ge nder, and insurance type, and education may predict pneumococcal vaccine uptake a mong eligible African American adults. Using the HBM, demographic questions tak en primarily from the BRFSS were reported and analyzed for their ability to predict vaccine uptake. This was accomplish ed by conducting bivariate analyses between each predictor variable and the outcome variable using the chi square statistic, and subsequently entering the statistically significant sociodemographic variables into the logistic regression model Hypothesis 2 Constructs operationalized from the HBM and PAPM such as awareness kn owledge, trust in healthcare provider, percei ved health, perceived susceptibility perceived severity, prior negative expe rience with vaccines and hea lthcare provider recommendation may predict and pneumococcal vaccine uptake among vaccine eligible African Americans. Using a substructed model of the PAPM and HBM as theoretical underpinning s this hypothesis was tested via survey in strument. Responses were coded and analyzed using PASW. Bivariate analyses were conducte d using Chi s quare analyses Scale level variables (perceived health su sceptibility, severity, and trust) were scored cumulatively then dichotomized with theoretically driven cut points before being en ter ing each into the model. Ultimately, the variables of interest from each hypothesis found t o be significant in bivariate analyses were entered using backward LR to determine the pr edictive nature of the variables as a whole
84 Post Hoc Testing After each of the hypotheses were tested, a number of bivariate analyses revealed statistically signific ant relationships which piqued further interest for post hoc evaluation. Among the variables of interest, the construct of trust was evaluated between age groups of study participants. Trust in healthcare provider was also compared with the construct of ra cial concordance between healthcare provider and participant. Perceived health scores from the SF were evaluated to determine whether or not age had from the SF 12v were compared with the single item question created to measure health in order to ascertain whether or not a single item question could reliably capture perceived health as effectively as a more complex instrument. Protection of Human Subjects Approval for this study was obtained from the University of Florida Health Science Center Institutional Review Board (IRB 01) prior to any subject recruitment or data collection. Two different IRB approvals were sought and obtained: 1) the pilot study to validate t he VUQ instrument, (Appendix I) and 2) the eligibility screening tool and primary data collection for the full study (Appendix J) Pa rticipation in the focus group of the pilot study and full study w ere str ictly voluntary and subjects could opt to withdra w from the study at any time without fear of repercussion. A waiver of informed consent was obtained from the IRB at the University of Florida in order to avoid collection of any protected health information or personal identifiers. Elements of informed co nsent were reviewed with all subjects ( Appendic es K and L) whose participation/survey completion served as tacit consent for the study. The
8 5 advisor, and IRB staff, was g iven to each participant for their personal records. Strict confidentiality was maintained by use of a coding sys tem whereby participants were assigned a number from 1 300 entered into PASW strictly for analyses. All data were de identified, file s were kep t in a locked file cabinet in the private o ffice, and no one else had access to the data.
86 CHAPTER 4 PILOT STUDY Exploring Pneumococcal Vaccine Uptake Rates among African Americans in Northeast Florida is a dissertation proposal intended to develop a b etter understanding of persistent vaccine uptake disparit ies between African Americans and their Caucasian counterparts despite access to care (Daniels, et al. 2006). Though health disparities are well documented in the literature (CDC, 2009; Smedley, et al. 2002) relatively little is known about how attitudes and beliefs among African Americans might influence preventative health decisions such as vaccine uptake. A number of studies cited in the literature emphasized challenges with s urveying African A mericans, given the fact that this population has been historically difficult to rea ch. Explanations for accessibility issues when attempting to survey African Americans do not suggest a parsimonious answer. Rather, it would seem that historical exploitati ons have led to mistrust and hesitance to allow to be privy to private health matters. Lack of cultural competence among typically Caucasian researchers may also explain some of the challenges with African American research participation (Shave rs Hornaday, Lynch, Burmeister, & Torner, 1997). Cognizant of the challenges with surveying African American subjects, a pilot study using focus group methodology was designed to address the aforementioned barriers to participation, and facilitate collabor ation and problem solving to ensure the development of a culturally appropriate data gathering instrument Hence, pilot study objectives were to: 1) establish face to face communication with key community stakeholders to begin building trust; 2) refine the data collection process and instrument to ensure cultural sensitivity, visual appeal, understanding, and to include any additional questions deemed important to key
87 stakeholders; and 3) give the Principal Investigator (PI) and research assistant (RA) an o pportunity to operationalize the construct of cultural sensitivity Chapter 4 details the methodology, procedures, and results of the pilot study conducted to address barriers to African American research participation. Integration of pilot study findings and current literature are also discussed. Methods Sample Upon receipt of IRB approval, the RA in collaboration with the PI identified 12 individuals willing to participate in the pilot study. P ilot study participants were chosen based on their representa tiveness of the population of interest ( n = 12), and for their roles within the congregation as key community gatekeepers. Typically, pilot study participants do not constitute a random sample. Nonetheless, they should be representative of the population o f interest (Portney & Watkins, 2009). A few weeks before the pilot study, an invitation letter exp laining the purpose of the focus group and the r ole of the participants was given to each individual in person, along with the date, time, and locatio n of th e meeting ( Appendix M ). Six participants who self identified as African American from each of the two church es were invited t o participate in the focus group Procedures T he VUQ was developed from an extensive review of the literature and the integration of the aforementioned instruments and was thought to be culturally and developmentally appropriate, theoretically grounded, and a fairly self explanatory self administered instrument. That said it had not been tested among the intended audience
88 for cultura l sensitivity, understandability, and visual appeal, thus necessitating a pilot study This was accomplished in four steps. First, group members were apprised of the study and consented ; second, they were all asked to complete the survey just as participan ts in the full study would do ; and third, focus group members were then provided with a second copy of the VUQ on which to make comments and suggestions. Finally, focus group members were asked a series of questions ( Appendix N ) about the consent, the VUQ, recruiting, incentives, and cultural sensitivity. Establish Trust To operationalize pilot study objectives, a n African American Research Assistant (RA) whose longstanding relationship with church leaders placed her in a unique role to serve as a key infor mant, was sought out, apprised of the study, and hired to assist with recruiting and data collection. E xisting literature (Alvarez, 2011; Dein, 2006; & Shellman & Mokel, 2010) supports this strategy citing the notion of commonalities in he ritage and experi ences as a bridge to the population of interest, and an opportunity to enhance cultural sensitivity and foster a trust ing relationship. The focus group was held at th e church on Sunday afternoon following services, so participants were familiar and comfor table with the setting. On the day of the focus group lun ch was provided for pa rticipants. The PI and RA greet ed each participant welcoming them and thanking them for their ti me. The PI and RA introduce d themselves, provide d a brief biography, then discu ss ed the background, purpose, and significance of the study as previously mentioned to assuage concerns pertaining to study legitimacy, privacy, and procedures. In order to avoid leaving out important details, a ba sic guide was designed as a referen ce tool for the PI and RA ( Appendix O ).
89 The PI then review ed the consent process with the group as a whole, encouraging partic ipants to interject if they had questions or concerns about wording, privacy, or otherwise. Informed consent forms were not signed, nor were they collected. Instead, participants were informed that they should read the consent s and keep it for their records. Since the overarching goal of informed consents is to ensure that everyone clearly understands the study and what is being asked of them, best practices for consenting subjects were reviewed prior to conducting the focus group ( Fl ory & Emanuel, 2004). Although methods for carrying out the informed consent process varied in the literature good communication and taking the time to addre ss questions were key components in the process. The informed consent process for the pneumococcal focus group was congruent with recommendations from the literature, followed the stated guidelines of the IRB, sought to elicit participants concerns, and a fforded the PI an opportunity to fine tune any issues pertaining to the process prior to the full study. Refine Process and Instrument Following the study introduction and informed consent, focus group participants were asked to complete the VUQ ( Appendi x C ) on their own thus providing the PI with an average completion time. Because the focus group was small and participants were not randomly selected, protecting their health information was of paramount importance. For this reason, VUQ completion was use d only to assess completion time and generate discussion. Participants were instructed not to place their names anywhere on the survey, and completed VUQs were not collected. Instead, participants were
90 instructed to take the questionnaire home with them ei ther to keep, or dispose of as they saw fit. In the third step of the focus group, participants were given a clean copy of the VUQ and asked to either work on their own or in pairs to carefully examine each question for understandability, readability, and cultural sensitivity. Participants were instructed not to place their names on the VUQ, and not to actually answer any of the questions since this copy would indeed be collected. P ens and highlighters were provided and focus group participants were encour aged to circle, highlight, or put notes b eside any item they either understand or thought wa s worded im properly When the group completed scrutinizing the VUQ, a series of questions was asked of focus group participants to ascertain the ir opinions on wording, ease of understanding, and cultural sensitivity Any questions deemed difficult to understand, vague, or intrusive were discussed with participants, soliciting their recommendations for revising or possibly delet ing the item. Time spent d iscuss ing the instrument was framed as an idea generating session where all thoughts and sugg estions could be heard. The PI also solicit ed opinions from the group on recruiting strategies, incentives, data collection venues and cultural sensitivity. Additiona lly, the VUQ questions about income and responses of with respect to PPSV23 uptake were specifically addressed. P articipants in the full study would be provided with a modest incentive of a $10.00 gift card During th e pilot study, the PI ask ed th e group what type of gift card would be best received by study participants
91 Cultural Sensitivity Participants were asked specifically about the use of an African American research assistant for the full study. The researcher hoped to learn more about wh ether or not racial concordance with the RA would enhance cultural sensitivity and participation ; therefore focus group participants were specifically questioned with regard to the use of an African American research assistant and how it would impact buy in and willingness to participate. Findings The pilot study examining p neumococcal vaccine uptake serve d as a for the population o f interest. Establishing comfort with the researcher and engaging community members in the research proces s as mutual experts has documented successes in the realm of Community Based Participatory Research (CBPR) (Christopher, Watts, Knows, McCormick, & Young, 2008). While the pneumococcal study was not designed as a CBPR study, the intuitively humanistic elem ents of this methodology such as mutual respect and open communication ar e seen as essential and served as key components of the pilot study. Establish Trust Six (five female and one male) of the twelve individuals invited to participate attended the f ocus group held on November 20, 2011 at a church in Jacksonville, Florida. Step one of the process was to go over the study background and purpose and complete the informed consent. The group had no questions about the study background, purpose or signifi cance. The consent process was reviewed with the group as a whole by the PI. One participant asked where to sign, at which time the PI
92 emphasized that no signatures or other personal identifiers would be collected at any time during the study. The PI furth er explained that by reading the consent and subsequently completing the survey, participants consent was implicitly understood. Importantly, subjects in the Pneumococcal pilot study had questions about whether or not they would sign the consent form. Thre e of the six participants tried to return the consent form at the end of the focus group. At which time, the PI explained that the consent form was theirs to keep and that it contained important contact information should questions or concerns arise. The only other question pertaining to the informed consent process was, else do we have to The PI explained that the only thing being asked of participants for the focus group was completion of the survey and input regarding content and the research process for the full study Further, the PI explained that in the full study, participants would only be asked to complete and turn in the survey. Moreover, the paramount importance of info rmed consent and privacy were discussed step by step, and all conc erns were addressed by the PI. Explanation of the study purpose highlighting vaccine uptake disparities and the disproportionate number of African Americans affected b y pneumococcal illnesses was emphasized as well. Questions and feedback were encouraged, and the contribution participants are making towards enhancing nursing knowledge of vaccine beliefs and practices among African Americans were ac knowledged. The researcher explain ed how results may be used to develop targeted interventions aimed at increas ing vaccine uptake, thereby preventing pneumococcal illness among African Americans.
93 Refine Process and Instrument In step two of the focus group, participants completed the VUQ individually to provide the PI with an average completion time. Completion ti mes ranged from eight to 19 minutes with a mean completion time of 13.5 minutes. The following questions arose during the VUQ completion: do you mean by current smoker, does that include cigarettes, and I finish my degree, which education box do I have the sickle cell trait, but not sickle cell disease. Should I still check the sickle cell disease mother is black, but my father is white. Can I still check African American for you going to ask us t o get the vaccine does immunological suppression if I leave something had cancer, but cancer free now. Do I still check the When everyone had completed the VUQ for the purpose of establishing completion tim es, step three of the focus group process where participants were asked to provide feedback on the instrument began. All but one of the participants paired up for the survey completion component of the focus group. The group engaged in lively discussion as they completed the survey but had few questions for the PI and RA. When the group completed the editorial component of the process, a series of questions were asked of the participants to ascertain the opinions on wording, ease of understandi ng, and cultural sensitivity ( Appendix N ). Participants expressed no concerns about the wording of questions. One participant commented that the survey
94 was too long and asked if it could be shortened. Another participant echoed this sentiment. The PI expla ined that although the survey was long, each of the questions was felt to be of great importance ; however their recommendations would certainly be taken under advisement. Participants were asked for their opinions about particular wording choices and incl usion versus exclusion of a question regarding income. Although no income questions were included on the VUQ at the time of the focus group, the PI inquired about the potential inclusion of such an item. Participants overwhelmingly responded that they woul d not answer such a question. A male participant commented, private, we B lack folk like to talk about that kinda stuff A female participant added taught at an early age not to tell people about our money troubles The survey was o riginally designed with several questions having three possible answers: yes, no, and unsure. Pertaining to refining the instrument, the PI asked participants about a response of to determine whether or not it should be retained since it would lik ely complicate the statistical analyses. Participants unanimously agreed that the category should be retained because they always know if they had been vaccinated, or if their healthcare provider had rec ommended the pneumonia vaccine. Focu s group participants were then asked about recruiting, data collection times and locations, and incentives. The majority felt that plenty of participants could be obtained through word of mouth alone, but agreed that flyers could also be posted. One female participant stated, If giving out a $10 grocery card, you have any trouble getting people to fill [the survey] All participants agreed that survey
95 completion could be done immediately following church services on Sunday afternoons in e ither the Community Room or the chapel itself. When asked about the use of a barbershop in order to recruit male participants, the male in the focus group said, are plenty of men here at church who will fill [the survey] The PI asked the group if the $10 gift card should be a gas or grocery card. All, but one participant felt that the grocery card would be better with one individual saying, everyone has a car, so the gas card always Cultural Sensitivity The final series of questions for focus group participants pertained to the cultural sensitivity of the VUQ, racial concordance between the RA and subjects, and the research process as a whole. P articipants expressed no concern about the instrument or the study as a whole from a cultural sensitivity standpoint The PI asked the group to think about the instrument and the process in terms of respect for differences in culture, respect, and trust. The researcher asked the group if they felt the congregation would be comfortab le having a Caucasian female asking them to complete a survey about their health care practices and beliefs. Further, the PI asked if having an African American RA would help pave the way for buy in and participation. One woman spoke up and said, mean, as long as you treat everyone with respect, it will be fine. We are private about a lot of things, but doing this to help Black people understand vaccines The participant also stated, think great that Sylvia [RA] is helping with the research. We need more Black people to be involved in No further discussion followed.
96 Discussion Consistent with the literature which suggests that challenges with recruiting, engaging, and enrolling minority subjects in research studies includ es lack of knowledge concerning the research process, cultural differences between t he researcher and participants, the use of technical in the consent process, and logistic al barriers to participation such as time, transportation, and incentives ( Jones, Steeves, & Williams, 2009; King, et al. 2010; Loftin, Bunn, & Sullivan, 2005 ; Moreno John, Forte, Rangel Lugo, & Perez Stable, 20 00; Staffileno & Coke, 2006), we found that focus group members needed clarification on each of these items. Focus gro up subjects clarified the anonymous nature of the study, made sure they were not going to be asked to do anything beyond survey completion, provided their opinions about data collection times and incentives, and emphasized their beliefs that measures of in come should not be included in the study largely because of cultural beliefs about the private nature of the subject matter and concern for privacy When Kosoko Lasaki, et al. (2005), conducted focus groups among African Americans to refine a data collec tion instrument they learned valuable lessons pertaining to the manner in which questions were worded. For example, study participants sure how to interpret the term, thus resulting in terminology changes. Open ended questions originally on the survey were found to be ambiguous and consequently replaced with check boxes and a series of choices based on participant feedback. When we asked about the open ended questions pertaining to refusal reasons, focus group participants recommended that t hey remain in the VUQ. Although this is inconsistent with the findings of Kosoko Lasaki et al. post hoc analyses
97 of the full study revealed that these items were indeed problematic. This component of the research will be discussed in further detail in Cha pter 6 but is certainly worth noting. Race and ethnicity in the Kosoko Lasaki et al. study were originally two separate questions but researchers collapsed the question based on their focus group finding that participants felt they were one and the same. Participants in the pneumococcal focus group did not recommend substantive changes to the VUQ, but did have questions about which ethnicity to select if their parents were not racially concordant. Use of monetary incentives and attention to logistical is sues such as time, location, and transportation have been cited in the literature by Paskett, DeGraffinreid, & Tatum, 1994, and others as important considerations for conducting research successfully among African American communities. Monetary incentives, where the burden of study participation is significant, have been said to help off set costs such as transportation and child care. Although i n the pneumococcal study there wa s only a one time survey to complete, monetary incentives were also reported in the literature as a means to make participants feel that researchers value their time and opinion. Word of mouth communication about monetary incentives associated with the study may also aid subject recruitment (Paskett, et al. 1994; Souder, 1992). Our f ocus group participants validated the body of literature supporting the use of incentives, informing the PI that the gift card would greatly aid in recruitment and participation in the full study. Data collection locations and times were viewed as importan t components for recruiting and enrolling subjects; particularly African American men who are often underrepresented in research studies (Plowden, John, Vasquez, & Kimani, 2006). In a
98 series of focus groups conducted among African American men about potent ial interventions for preventi on of prostate cancer, a number of relevant suggestions for reaching the population of interest were revealed Among the suggestions, participants recommended churches, barbershops, fraternal organizations, and sporting events to reach African American males. where the people as opposed to asking them to come to you, was a key message in the focus groups (Allen, et al. 2007). Subjects in the pneumococcal focus group did not support the use of organizations outside of the church for the full study, primarily because they believed there would be plenty of willing participants within the churches themselves. Studies emphasized that time is an important resource not to be overlooked, citing busy schedules and time commit ment s as barriers to participation ( Flaskerud & Nyamathi, 2000; Rettig, 2000 ) S et appointment times for the consent process and distribution/collection of questionnaires have also been said to aid in the success of research studies among African Americans (McNeely & Clements, 1994).Rather than assuming these decision points focus group participants were specifically questioned With respect to data collection times and appointments, focus group members recommended that we collect data immediately followin g services on Wednesdays and Sundays and that we make sure the survey take too long to complete. They did not feel that appointment times were necessary, nor did the feel they would be adhered to. Word of mouth recruitment strategies originating fr om key community gatekeepers have been shown to decrease anxiety and distrust among potential participants, particularly among populations whose cultural values show a high affinity for informal
99 means of communication. W ord of mouth strategies have been ci ted as effective among African American communities as both a recruitment tool and a cost saving measure; decreasing or eliminating the need for advertisement of the study ( Hooks, et al. 1988; Jones, et al. 2009) Our focus group participants validated t his finding by telling us that they did not see the need for any additional advertising beyond word of mouth and possibly some flyers. Participants felt that members of the congregation would be much more likely to participate knowing that the key stakehol ders had already the study and felt it was worth their time and efforts. Becoming culturally competent is a life long process, not something that can be achieved overnight. Fortunately, many years of working in the healthcare profession with diverse patient populations affords the PI a solid foundation from which to build. In addition to establishing trust and laying the ground work for effective recruiting, the PI needed to be cognizant of cultural differences which could impede good communi cation and data collection. A few exemplars in the literature provide d guidance to operationaliz e the process. Recognizing the need to facilitate culture competence in health care, in 1999 the US Department of Health and Human Services Office of Minority Health (OMH) designated 14 specific standards for providers, policy makers, and researchers to facilitate culturally appropriate health care standards. While the Culturally and Linguistically Appropriate Services (CLAS) standards are slanted towards the c linical aspect of healthcare, emphasis on the development of participatory and collaborative partnerships to achieve better outcomes is certainly applicable in the research set ting ( US Department of Health and Human Services, [OMH] 200 1 ).
100 Taking into acc ount what is known about the population, the CLAS recommendations, and the experiences of other researchers, this pilot study was undertaken as a means by which to the and fine tune aspects of the pneumococcal study prior to implementation. M oreover, the pilot study served as a method by which the PI could evaluate and heighten understanding of the unique aspects of the population of interest pertaining to their cultural heritage, communication preferences, and attitudes towards research and p reventative health services such as vaccinations. Each of the CLAS standards was reviewed and incorporated into the methodology for conducting the focus group. None of the participants had concerns about the VUQ or the research process with r espect to cultural sensitivity, which led the PI to believe that the extensive preparation for the pilot study was well worth the effort. Purported to be an effective strategy for African American recruitment and research participation, use of a racially matched research assistant (insider) (Gerrard, 1995; Pattillo McCoy, 2000; Smetana and Gaines, 1999) was a key component of the pneumococcal study. When the researcher and the potential subjects have different ethnic and cultural characteristics mistrust and reservations may be exacerbated. When racial concordance exists between subject and researcher, similarities in culture and shared life experiences have been shown to enhance trust and facilitate entre into the community (Blumenthal, Sung, Coates, Wil liams, & Liff, 1995; Wasserman, Flannery, & Clair, 2007; Witten, 2005). Participants in the pneumococcal focus group echoed these sentiments saying that they felt more comfortable since many of them knew the RA and had seen her at church. The pneumococcal focus group did not
101 express any reservations with the study despite the fact that the PI was not racially concordant. Though no findings in the literature linked racial concordance between RA and subjects to study limitations, the PI was aware of the fact the RA in the pneumococcal study could introduce bias into the research process. It is impossible to know if the outcome would have been different had there not been a racially concordant RA ; however, future studies with multip le data collection sites might present an opportunity to further explore the presence or absence of a racially concordant researcher. Summary This pilot study was designed as a precursor to the full study to lay the groundwork for a trusting relationship, establish open communication, and aid in the recruitment of subjects through cultural sensitivity, use of incentives, and the presence of a racially matched research assistant. Considering the historical and contemporary barriers to minority research parti cipation, the d evelopment of a culturally appropria te process and data collection instrument wa s an integral component of the proposed pilot. Supported in a recent study by Kneipp, Lutz, & Means, (2009) these findings were congruent with existing literatu re Kneipp, et al. used a descriptive survey ( n = 35), and found that low income, predominantly African American women were more likely to enroll in research studies when: 1) they fe lt the researcher recognized an unmet health need (91%); 2) nurses were co nducting the study (57%); and 3) researchers could be trusted to follow the procedures explained in the consent (100%). This work inform ed the pneumococcal pilot study of the critical role trust, good communication, and careful adherence to procedures cove red in the informed consent play in the recruitment and retention of
102 underrepresented groups. While explaining the full study rationale to the focus group took time, this step wa s of paramount importance in fostering trust and open communication among Afri can American subjects, and is reportedly often overlooked by novice researchers ( Alegra 2009). Upon completion of the focus group, the PI reviewed field notes and as well as each of the VUQ instruments collected from participants. Notes placed on the VU Q were minimal, with no major recommendations for revision Although no significant concerns arose from the Pneumococcal Vaccine Uptake pilot study, responses were consistent with much of the current literature concerning the recruitment and participation of African Americans in research studies. Moreover, questions that arose in the focus group provided the researcher with a better understanding of the two churches in which the full study would be conducted. Satisfied that while the survey wa s lengthy, it was easily understood, free of errors, and visually appealing, the PI opted to retain the survey without changes for the full study. With respect to methodology, pilot study participants provided valuable information about recruitment and par advertising for the study was performed, nor did the PI seek other venues previously believed to be necessary to reach the number of participants needed for analyses. Questions asked by the partici pants on topics such as smoking, and prior history of cancer did not result in changes to the VUQ, but helped the researcher frame a consistent introductory message for all subsequent data collection points in the full study.
103 CHAPTER 5 RESULTS In Chapter 5 results will be presented by addressing each of the study aims. M e thods for data cleaning will be described briefly followed by univariate descriptive statistics bivariate analyses and the logistic re gression mod el itself Moreover, findings from a priori a naly ses conducted to examine secondary resear ch questions concerning vaccine refusers and trust, as well as vaccine refusers and age categories will be detailed Post hoc analyses including trust and age, trust and racial concordance, perceived health and age, and smoking and PPSV23 were analyzed and will be reported. Data collection for the full study took place between December 18, 2011 and May 6 2012 at two churches in Jacksonville Florida one of which was Catholic and the other Baptist Each of the chu rches has been in the community for a number of years with well established congregations and community ties. The PI and RA were on hand after church services on eight Sunday mornings and two Wednesday mornings for recruitment and survey collection. In tot al, 32 9 study participants answered the survey; however, 26 were eliminated because subjects did not meet aforementioned eligibility criteria for the PPSV23 vaccine Upon further examination of the 301 remaining surveys six subjects did not answer the DV you ever gotten the pneumonia Consequently, they were also eliminated yielding a final samp le size 295 for data analysis Data were entered into PASW version 20, and double checked for accuracy. The dat a set was cleaned using the PASW Stat istical Procedures Companion Guidelines (Norusis, 2008). Each of the labels and definitions were re con firmed, m issing data were evaluated for patterns using histograms and scatterplots, and
104 extreme values were explored using the descriptives and explore t abs as well as dfbeta s As mentioned in the methods section, o utliers identified on scatterplots and histograms were examined for accuracy and retained. The dependent variable (PPSV23 vaccine status) had three possible responses: yes, no, or unsure. Clini cally speaking, patients who do not know their PPSV23 vaccine status are treated as tho ugh they are unvaccinated. As previously discussed, there are no known harmful effects of re vaccination other than localized inflammation at the injection site; therefo re all persons with unknown status are given the PPSV23 vaccine. Given these clinical guidelines, the researcher sought the advice of a statistician ( p ersonal communication, August 1, 2012) and made a theoretical decision to collapse the categories and into one category in PASW for analyses. A separate column was created in PASW and after listing the number of participants who were unsure of their status ( n = 24, 8% ) in the descriptive section, future analyses were based upon the collapsed category. Univariate Descriptive Statistics Analyses Each of the study participants by virtue of inclusion criteria were eligible for the pneumococcal vaccine ( N = 295) From this population, 1 76 ( 59.7 5%) were unvaccinated despite eligibility with 95 (32 .2%) vaccinated and 24 (8.1%) unsure These results mirror the nationally reported Healthy People 2020 statistics and underscore the need for intervention in the African American community to increase vaccine uptake through a better understanding of the va ri ables influencing vaccination. Ap proximately 2/3 of the total participants (N = 295 ) were female ( n = 196 66. 4 %) yielding 99 ( 33.6 %) male s Age categories were utilized in the survey in lieu of actual
105 age. These categories were consistent with those ut ilized in the BRFSS instrument, with the highest number ( n = 81, 27.5 % ) of participants falling into category three (46 55 years old ). With health insurance and education status serving as proxy variables for SES, the majority of participants had private h ealth insurance ( n = 126 42.7%) and at least a high school diploma 93 (31.5%) Table 5 1 provides a complete description of demographics for the study population. In order to be eligible for the pneumococcal vaccine (PPSV23), study participants needed to have at least one of the aforementioned conditions or be 65 years of age. Only 47 of 29 5 (15. 9 %) subjects self reported as current smokers an independent indicator for PPSV23 vaccination since 2010. Among the chronic diseases diabetes ( n = 10 3 34.9 %) and hypertension ( n = 19 4 65.8 % ) rates were the highest in study participants Eligibility criteria and demographics for PPSV23 vaccination are listed in T able 5 2 Bivariate Analyses Aim One Aim one of the study was to determine how sociodemographic variables such as age, gender, and socioeconomic status, in fluenced pneumococcal vaccine uptake among African Americans 18 whose age or pre existing health conditions including diabetes, heart disease, cancer, COPD, sickle cell disease, immunological suppression asplenia, cochlear implants, liver disease, and cigarette smoking were known to increase the risk of contrac ting pneumococcal illness thus rendering them eligible for PPSV23. Chi square analyses were used to examine each of the categorical sociodemographic variables : age, gender, type of insurance, and education.
106 Age Subjects age d 18 and over were included in this study provided they met the aforementioned eligibility criteria for PPSV23. Ages were grouped according to BRFSS categories rather than obtaining an exact age for each subject. The younger the patient was, the less often they were vaccinated. For exam ple, there were 19 subjects between the ages of 18 and 25 of which only one (0.05%) had received the PPSV23 vaccine despite eligibility. Conversely, 29 of the 52 subjects in the 65 75 age category were vaccinated (55.76%) Following completion of univari ate descriptive measures, some of the age categories had very f ew subjects. For example, there were only two subjects who reported themselves 90+. In an effort to report findings in a succinct manner and build a more stable regression model, the original s even age categories were collapsed into three ca tegories: 18 44 ( n = 79), 45 64 ( n = 129), and 65+ ( n = 87). These categories were then examined using chi square analyses and found to be statistically significant for PPSV23 uptake whereby those who were ol der were more likely to be vaccinated than their younger counterparts 2 (1, n = 295) = 40.794, p < .001 ( T able 5 3 ). Gender Evaluation of gender and vaccination rates revealed that 75 of 99 males ( 75.75% ) were unvaccinated whereas 125 of 196 females ( 63. 77 % ) were unvaccinated. Female gender proved a statistically significant predictor for PPSV23 uptake with 12% more unvaccinated males than females 2 ( 1, N = 294 ) = 4.445, p < .05 ( T able 5 3 ).
107 Socioeconomic Status The most common measures of socioeconomic status ( SES ) found in the literature were education and income combined; however, income was deemed to o invasive and therefore not asked in the VUQ instrument. Insurance and education status and were used as proxy variables for SES and will now be report ed separately in bivariate analyses. Insurance Among 246 subjects who answered the question regarding which type of insurance they had 12 6 ( 42.7 %) were privately insured. Once univariate descriptive statistics were reported, the original number of categor categories were self pay, Medicaid, and all others. These decision points were based on the fact that self pay and Medicaid are unique in that the burden of PPSV23 falls solely to the individual, where private insurers including VA and Tricare cover the vaccine for those who are eligible. To be certain nothing was missed, bivariate analyses were conducted before and after collapsing this variable ; however, type of insurance still did not appear to predict vaccine status 2 (1, n = 246) = 1.638, p > .05 ( T able 5 3 ). Education Study participants were asked to indicate their highest level of education. Response choices ranged from grade school to gr aduate degree. Of the 294 valid responses, 13 ( 4.4 %) had completed only grade school, 93 ( 31.5 %) completed high school or the equivalent, 109 ( 37 %) had some college or an degree, and 78 ( 26.4 %) had a degree or higher Upon completion of univariate descriptive
108 statistics, education level responses were collapsed into two categories: 1) no college, and 2) college, in order to provide data in a more succinct manner. Bivariate analyses were then conducted using a c hi s quare statistic wher eby education status did not appear to significantly predict PPSV23 uptake 2 (1, n = 293) = 0.008, p > .05 (Table 5 3) Aim Two A im two of the study was to determine how constructs operationalized from the HBM and PAPM such as vaccine awareness knowledg e, trust in healthcare provider, perceived health, perceived susceptibility, perceived severity, prior negative experience with vaccines and healthcare provider recommendation predicted pneumococcal vaccine uptake among African Americans 18 whose age or pre existing health conditions such as diabetes, heart disease, cancer, COPD, sickle cell disease, immunological suppression asplenia, cochlear implants, liver disease, and cigarette smoking were known to increase the risk of contracting pneumococcal illness thus render ing them eligible for PPSV23. Each of these relationships was examined in bivariate analyses prior to entering them into the logistic regression model. Vaccine Awareness The construct of vaccine awareness was measured util izing one question in the survey. Participants were asked, this study have you ever heard of the pneumonia Response choices f or the question were yes, no, or not sure. Of 295 respondents 129 (43.7%) reportedly never heard of PPSV23 prio r to the study 12 ( 4.1 % ) of participants were unsure if they had ever heard of PPSV23, and 154 ( 52.2% ) of subjects had at least heard of the pneumonia vaccin e at some point When
109 yses. After collapsing the categories no and unsure, 141 ( 47.8 %) had never heard of PPSV23. Of the 141 who had never heard of PPSV23, 129 (91.48%) were unvaccinated. Of the 154 subjects who had heard of PPSV23 71 remained unvaccinated (46.1%). A relations hip between vaccination status and vaccine awareness was consequently found to be significant 2 (1, n = 2 95 ) = 69.44 p < .001 ( T able 5 3 ). Vaccine Knowledge The construct of vaccine knowledge differed from vaccine awareness based on findings in the litera ture review and accordingly, was considered independently for this dissertation. Vaccine knowledge was measured utilizing two questions in the survey. First, participants were given a statement pneumonia vaccine to make a decision about whether or not to accept it, with five possible responses on a l ikert scale ranging from strongly agree to strongly disagree Of the 143 subjects who agreed that they lacked sufficient knowledge to make a decision about PPSV23, 124 (86.71%) were unvaccinated. Among those who felt they did have sufficient knowledge of PPSV23 ( n = 76), only 23 (30.26%) were unvaccinated. Using chi square analyses, lack of knowledge among study participants was highly significant 2 (1, n = 219) = 71.6 60 p < .001 ( T able 5 3 ) Participants were then asked, you know who is supposed to get the pneumonia Of the 29 3 valid cases, 146 (49.5 % ) did not know who should be vaccinated, 52 (17.6%) were unsure, and 95 ( 32.42% ) felt they did know who s hould receive PPSV23 Of the 198 subjects who did not know who should be vaccinated 80.8% ( n = 160 ) remained unvaccinated whereas those who felt they did know who
110 should be vaccinated had a 41.05% ( n = 95 ) non vaccination rate. A relationship between vac cination status and vaccine awareness was therefore found to be significant with this item on the questionnaire as well 2 (1, n = 2 93 ) = 46.56 p < .001 ( T able 5 3 ). Trust in Healthcare Provider Trust in h ealthcare provider was measured in this study us ing the Wake Forest Trust Scale which consisted of ten Likert Scale questions ranging from to ( Appendix E). Cumulative trust scores ranged from 10 (very low trust) to 50 (very high trust). The mean cumulative trust scor e was 29.8 ( SD 8.258, n = 288). The 90 subjects who reportedly received the PPSV23 vaccine had a mean trust score of 41.29 indicating a high degree of trust in their healthcare providers ( n = 90, SD 8.566) whereas the 193 subjects who were unvaccinated ha d a mean trust score of 38.33 ( n = 193, SD 7.527). While the initial data analyses plan was to analyze the trust variable using a t test for independent samples, assumption s for scale level DV and normality were violated Given these violations, a decision was therefore made to dichotomize the trust variable using a cut point of 35 whereby those with scores of 35 and below were considered to have low trust, and those with scores 36 and above were considered to have high trust. After dichotomizing the trust variable, 67 (81.70%) of those with low trust ( n = 82) were unvaccinated compared with 126 (62.69%) of those reporting high trust ( n = 201). Chi square analyses were conducted and higher trust in healthcare provider scores were found to be significantly pr edictive of PPSV23 uptake 2 (1, n = 283) = 9.715 p < .05 ( T able 5 3 )
111 Perceived Health When perceived health is higher (the patient believe s he/she is very healthy), it has been hypothesized that there could be an inverse relationship with preventative h ealth behaviors such as vaccine uptake To examine this more carefully, p erceived health was measured us ing the SF ( Appendix G ) instrument which consisted of 12 Likert Scale questions. Raw scores were calculated using a summa tive approach with 56 bei ng the highest possible score (very healthy) for perceived health and 10 being the lowest (very unhealthy) The SF 12 portion of the VUQ instrument was the final component of the survey which may explain the attrition rate with only 260 of 295 subjects co mpleting this section. Participants who answered some but not all of the SF 12 questions could not receive a cumulative score and were consequently eliminated from this portion of the data analysis With a range of 18 56, the mean score was 42. 68 ( n = 260, SD 8.464). Consistent with the trust in healthcare provider variable, assumptions of normal distribution and continuous DV were violated with the perceived health variable as well, thus precluding the use of a t test for independent samples. A decision w as therefore made to dichotomize perceived health scores using a cut point of 33 whereby those 33 and below were considered to have low perceived health and those 34 and above were considered to have high perceived health. Of those who reported low perceiv ed health ( n = 42), 26 (61.9%) remained unvaccinated. Those with high perceived health scores ( n = 218) had similar rates of non vaccination with 156 (71.56%) never having received PPSV23. Though participants with lower perceived health scores were slightl y more likely to be vaccinated than those with high perceived health scores ( n = 16, 38.1%, and
112 n = 62, 28.44% respectively), chi square analyses using dichotomized perceived health scores did not reveal a statistically significant difference between the g roups with respect to PPSV23 uptake overall 2 (1, n = 259) = 1.563 p > .05 ( T able 5 3 ). Perceived Susceptibility Perceived susceptibility and perceived severity are core components of the HBM and were measured using a series of four single item questions in the VUQ to measure this theoretical component of the model and determine their impact on PPSV23 uptake. Perceived susceptibility was measured using two single item questions each with a five point Likert Scale response option Question 39 get sick more easily than others had a mean score of 2.17 ( n = 291) Q uestion 40 am at risk for getting pneumonia had a mean score of 2.62 ( n = 292). The two susceptibility questions were added together for the purpose of data analysis and cumulative scores were considered. A cumul ative susceptibility score of 10 would indicate high perceived susceptibility whereas a cumulative score of two would indicate low perceived susceptibility. Evaluation of a histogram revealed that perceived susceptibility was not normally distributed C onsequently, susceptibility scores were dichotomized using a cut point of five, whereby those with scores of five and belo w were categorized not susceptible and those with scores above five, susceptible Among those who did not bel ieve they w ere susceptible to pneumonia ( n = 185), 136 (73.51%) were unvaccinated and 49 (26.48%) reported PPSV23 uptake. For those who did perceived themselves as being at risk for contracting pneumonia, 59 (58.41%) were unvaccinated, and 42 (41.58%) were vaccinated. Chi square analyses detected a significant relationship between perceived
113 susceptibility and PPSV23 uptake with those in the high perceived susceptibility cohort more likely to be vaccinated 2 (1, n = 286) = 6.864, p < .05 ( T able 5 3 ). Percei ved Severity Perceived severity questions w ere treated in the same manner as perceived susceptibility and i ndividual scores for each question were evaluated I tem 37 I believe pneumonia is serious had a mean score 4.48 ( n = 294), and item 3 8 I belie ve pneumonia could lead to death had a mean score of 4.39 ( n = 289) A cumulati ve score for the two severity questions was then utilized for statistical analyses with a score of 10 indicat ing the individual perceived pneumonia as severe and a score of t wo indicating low or no perceived severity. Of the participants who answered the question, believe pneumonia is serious 251 ( 85.1 % ) either strongly agreed or agreed that pneumonia was a serious illness ( n = 290, mean 4.47, SD .767) Those who answered the question, believe pneumonia could lead to death strongly agreed or agreed with this statement 82.86% of the time ( n = 286, mean 4.39, SD .882). The histogram for p erceived severity was skewed a nd the DV was dichotomous. With the assumptions of n ormality and scale level DV violated, a t test for independent samples could not be utilized The rel ationship between perceived severity and PPSV23 uptake was accordingly examined using two separate nonparametric tests including the Mann Whitney U test a nd the Wilcoxon Rank Sum test. In each instance, results demonstrat ed higher mean scores for perceived severity among the vaccinated group than the non vaccinated group using a 95% significance level and p value of 0.05 or less That said, the difference i n means between groups appeared negligible and neither the Mann Whitney nor Wilcoxon Rank Sum produced more than a suggestion to
114 reject the null hypothesis. To assuage these concerns, perceived severity scores were therefore transformed into a dichotomized variable where a score of zero indicated low perceived severity and a score of 1 indicated high perceived severity For the purposes of analysis and reporting the cut point on the perceived severity scale which ranged from 2 10 was five. Fortunately, all respondents ( n = 285 ) fell squarely into one category or the other, with no neutral responses. Using the dichotomized perceived sev erity score and PPSV23 uptake, chi square analyses were not found to be significant 2 (1, n = 2 85 ) = 1.903 p > 0.05 ( T able 5 3 ). Prior Negative Experience with Vaccines Prior experience with va ccines was measured using two single item question s : you ever gotten sick after getting any type of vaccine And anyone you know, other than yourself ever gotten sick afte r getting any type of Response options for each question w ere dichotomous yes or no. Subjects who responded that they had become ill after a vaccine or knew someone who did, were considered to have had a prior negative experience with a vaccine. Bivariate analyses were conducted using c hi square. Of th e 2 90 subjects who answered th e question s about prior negative experiences with vaccines, 113 ( 38.3 % ) indica ted they or someone else had become il l following a vaccine whereas 177 ( 61.03 % ) had not B ivariate analyses between prior negative experience and PPSV23 uptake were not statistically significant 2 (1, n = 2 90 ) = .039 p > .05 ( T able 5 3 ) Healthcare Provider Recommendation In o rder to ascertain whether or not participants had a healthcare prov ider recommending the PPSV23 vaccine to these subjects, all of whom met eligibility criteria
115 set forth by the CDC, subjects were asked, your healthcare provider ever recommended that you get the pneumonia There were four missing responses fo r this question leaving 291 valid cases for analyses. Response options for this question were no (139 47.1%), yes (103 34.9%), and unsure (49 16.6%). Participants who answered unsure for provider recommendation were treated as though they did not have a recommendation since they could not recall one. With the categories unsure an d no collapsed, 188 ( 64.6% ) of the respondents had not received a provider recommendation for PPSV23. The relationship between healthcare provider recommendation and PPSV23 upt ake was measured using c hi s quare analyses given the fact that each of the variables was dichotomous. Of the 188 respondents who did not recall receiving a healthcare provider recommendation, 165 (87.76%) were unvaccinated whereas 23 (12.23%) reported havi ng gotten the PPSV23 vaccine by some means despite the lack of recommendation. Worth noting the VUQ did not ask participants to delineate where and when they had been vaccinated. When healthcare provider recommendation was present, 71 (68.93%) of 103 subj ects did receive the PPSV23 vaccine while 32 (31.06%) went unvaccinated When comparing the rate of vaccine uptake between those whose providers had recommended the vaccine with those who had not received a recommendation, there was a statistically signif icant difference 2 (1, n = 291 ) = 97.820 p < .001 ( T able 5 3 ). Stepwise Logistic Regression Model Using stepwise backward logistic regression, variables of interest found statistically significant in bivariate analyses were en tered into the model beginnin g with
116 sociodemographic s age and gender Next statistically significant variables operationalizing the HBM and PAPM were entered i nto the model to determine their ability to predict pneumococcal vaccine uptake. These variables included awareness, knowled ge, trust, perceived susceptibility, and provider recommendation all of which were significant in bivariate analyses. The whole model was then considered to determine how well age gender, vaccine awareness, vaccine knowledge, trust in healthc are provider perceived susceptibility and prov ider recommendation predicted PPSV23 uptake among African Americans whose age or pre existing disease processes rendered them eligible to receive it. In total, 19 subjects had missing data in some portion of the questio nnaire ; therefore the case processing s ummary indicated that 276 valid subjects w ould be included in the whole model of regression analyses. Using logistic regression, the initial model with no variables in the equation (step 0) predicted vaccine uptake c orrectly 68.8 % of the time. In Block one, a ge and gender were entered into the model using bac kward hierarchical methodology Age was found to be a significant predictor for PPSV23 uptake ( = 1.422 p < .0 5 ) whereby those ages 18 44 ( n = 79) were four time s less likely to be vaccinated than those 65+ ( n = 76) (referent group), and those ages 45 64 ( n = 124) were roughly two times less likely to be vaccinated than the referent group Though gender was significant in bivariate analyses, this was no longer the case when the variable was entered into the LR model along with age (Table 5 4) In Block one the 2LL was 307.509 The Nagelkerke R square can be loosely interpre ted that age and gender accounted for 16.7 % of the variance with PPSV23 uptake adding
117 stren gth to the model. The Hosmer and Lemeshow test statistic ( 2 = 5. 804 DF = 7 p > .05 ) confirms that the predicted probabilities match the observed probabilities. In Block two, predictor variables derived from the substructed theoretical model found to be si gnificant in bivariate analyses were entered using backward hierarchical methodology. Variables included in this step were vaccine awareness trust in healthc are provider, perceived susceptibility and healthcare provider recommendation Though the knowled ge variable was highly significant in bivariate analyses it was left out of the model due to multicollinearity with the awareness variable as evide nced by VIF values > 5. A composite variable using awareness and knowledge could have bee n created to resolv e this issue; however, theoretically and empirically the se constructs had clearly delineated differences which is why the most intercorrelated variable; knowledge, was ultimately excluded from the LR model. In Block one, vaccine awareness was a signif ican t predictor ( = 1.868 p < 001 ) of PPSV23 vaccine uptake Those who were unaware of the existence of PPSV23 prior to this study ( n = 132) were 6.5 times less likely to be vaccinated than participants who (referent gr oup) ( n = 147) (Table 5 4) Provider recommendation was also statistically significant in the LR model ( = 1.989 p < .0 01 ) with participants who did not have a recommendation for PPSV23 uptake from their provider ( n = 179) roughly seven times less likel y to be vaccinated than those whose provider had recommended the vaccine ( n = 100) (Table 5 4) With age, gender, awareness, and provider recommendation in the model, trust in healthcare provider and perceived susceptibility were no longer statistically si gnificant (Table 5 4). By adding the variables in block two the 2LL decreased from 307.509 to 177.215 indicating a
118 redu ction in unexplained variance The Negelkerke R square statistic indicated that the model with all the variables now predict ed 63.4% o f the variance and the Hosmer and Lemeshow statistic remained without significance ( 2 = 8.214 DF = 8 p > .05 ) With all significant variables in the model, the final iteration of the backward hierarchical logistic regression revealed improved PPSV23 vac cine uptake pr edictability from 68.6% baseline to 8 6.6 %. A P riori Analyses Unsure and Refused A priori plans to examine the unique characteristics of participants who refused PPSV23 ( n = 29 ) were conducted using chi square analyses Reasons for refusal of PPSV23 vaccine in this study included: fear of needles ( n = 39), out of pocket expenses ( n = 22), general lack of knowledge ( n = 40), mistrust of vaccines ( n = 33), and lack of healthcare provider recommendation ( n = 43). Self reported PPSV23 refusal s on the VUQ numbered 29; however, a total of 46 refusal reasons were provided which could indicate measurement bias, or perhaps a refusal then subsequent acceptance of PPSV23 at a later date. Relationships between refusals and the sociodemographic variables ag e, 2 (1, n = 294 ) = 11.233 p > 0.05 and gender, 2 (1, n = 293 ) = 2.621 p > 0.05 were not significant I nsurance type and education were also examined among refusers using c hi square analyses and not found to be statistically significant 2 (1, n = 24 5) = 6.608 p > 0.05 and 2 (1, n = 293) = 14.708 p > 0.05 r espectively (Table 5 5) Perceived Health Measures Perceived health was measured with the widely used SF 12 item instrument with demonstrated validity and reliability in the literature and ease of
119 completion. As previously discussed, t he relationship between perceived health using dichotomized SF 12 scores and PP SV23 uptake was not found to be statistically significantly ( T able 5 3 ).To ascertain whether or not perceived health could be mea sured more simply, the author of this dissertation also included a single item question from the BRFSS with a five level Likert Scale response: consider myself to be (5 = strongly agree, 4 = somewhat agree, 3 = neutral, 2 = somewhat disagree, a nd 1 = strongly disagree) ( n = 291, mean 3.87, SD 1.080) This single item question was dichotomized into yes and no, and reexamined. With the variable dichotomized, descriptive statistics revealed that 41 (13.9%) considered themselves unhealthy, whereas 216 (73.2 % ) considered themselves healthy. Moreover, 153 (70.83%) of those who considered themselves to be healthy were unvaccinated while 24 (58.53%) of those who did not consider themselves to be healthy were unvaccinated. Chi square analyses w ere then conducted to examine the relationship between PPSV23 uptake and the single item for perceived health and consistent with the SF measure, the relationship was not found to be significant 2 (1, n = 257) = 2.430 p > .05 ( T able 5 5 ) Finally, when comparing single item perceived health with the more complex measure of health found in the SF findings were strikingly similar. In the single item measure ( n = 256), 67 reported themselves as unhealthy and 189 reported themselves as healthy. U sing the cumulative SF scores ( n = 259), 73 reported themselves as unhealthy whereas 186 repor ted themselves as healthy. Chi square analyses of these two measures of perceived health were statistically significant 2 (1, n
120 = 256) = 32.967, p < .001, suggesting that a single item question may be as effective in some instances as a more complex instrument to measure perceived health. Post Hoc Analyses Age and Perceived Health Intuitively, it would seem that younger adults would perceive themselves as healthier than older adults; however, because all participants in this study had one or more preexisting condition s such as diabetes, heart disease or COPD, these variables were examined in post hoc analyses to determine if this belief held true. Only 13 (17.1%) of those aged 18 45 reported themselves as unhealthy, compared with 32 (28.07%) of those aged 46 64, and 27 (39.13%) in the 65+ age group. Findings in this post hoc analysis therefore revealed a statistically significant difference in perceived hea lth between age groups with younger age predictive of better perceived health scores 2 (1, n = 259) = .8.748, p < .05 (Table 5 5) Trust and Racial Concordance In post hoc analyses, r acial concordance with healthcare provider as a means of enhancing trust was examined using data from the VUQ. A total of 63 (24.51%) participants reported racial concordance with their healthcare provider in the pneumococcal study. The construct of racial concordance was examined to determine if it had an impact on the variab did not have racially concordant providers, 53 (27.31%) reported low trust, and 141 (71.68%) reported high trust. Findings with the racially concordant group were very similar with 18 (28.57%) repo rting low trust, and 45 (71.42%) reporting high trust. Thus,
121 the relationship between racial concordance and trust in healthcare provider was not statistically significant in this study 2 (1, n = 257) = .037, p > .05 ( T able 5 5 ) Trust and Age The constr uct of trust in healthcare provider was also examined with respect to age categories of study participants. In this post hoc analysis, younger subjects were significantly less trusting than their older counterparts 2 (1, n = 283) = 15.106, p =.001 with th ose in the age group 18 45 ( n = 79) reporting low trust 41.78% of the time compared with those in the 46 64 ( n = 125) reporting low trust 30.4% of the time, and those 65+ ( n = 79) reporting low trust only 13.92% of the time (Table 5 5) Smoking and PPSV23 Uptake In 2008 smoking was added to the list of independent indicators for PPSV23 uptake when the CDC and others ( Breiman, 2000; Grua 2005; Nuorti, et al. 2000) found that current smokers with known immunological suppression were 4.1 times more likely to contract pneumococcal illness than their non smoking counterparts (95%, CI 2.4 7.3, p < 0.001) Accordingly, p articipants in the pneumococcal study were asked you currently smoke Of the 289 respondents 47 (16.26%) were smokers. Non s mokers had a PPSV23 uptake rate of 34.29% ( n = 83) whereas smokers had an uptake rate of 19.14% ( n = 9). Thus, s mokers were found to be significantly less likely to be vaccinated than non smokers 2 (1, n = 289 ) = 4.162, p < .05 ( T able 5 5 ).
122 Table 5 1 Univariate d escriptive s tatistics Variable Response n % Vaccinat ion s tatus Unvaccinated 175 59.7 Vaccinated 95 32.2 Unsure 24 8.1 Refused 29 10.0 Missing 6 eliminated Gender Male 99 33.6 Fema le 196 66.1 Missing 0 0 Age 18 25 19 6.3 26 35 27 9.0 36 45 33 11.2 46 55 81 26.9 56 64 48 16.3 65 75 52 17.6 76 89 33 11.0 90 and over 2 0.7 Missing 0 0 Primary l anguage English No 0 0 Yes 295 100 .0 Missing 0 0 Vete ran No 241 80.1 Yes 12 4.0 Missing 42 14.23
123 Table 5 1 Continued Variable Response n % Education l evel Grade s chool 13 4.4 High s chool 82 27.8 GED 11 3.7 Some c ollege 79 26.8 d egree 30 10.2 d egr ee 54 18.3 Graduate d egree 24 8.0 Missing 1 0.3 Type of i nsurance Self p ay 27 9. 2 Private i nsurance 12 6 42. 7 Medicare 4 2 14. 2 Medicaid 22 7. 5 VA b enefits 7 2. 4 Tricare 18 6. 1 Other 4 1.4 Missing 49 16.6
124 Table 5 2 Chr onic d isease p rocesses among participants Variable Response n % Current s moker No 242 82.0 Yes 47 15.9 Missing 6 2.0 Diabetes No 188 63.7 Yes 103 34.9 Missing 4 1.4 Heart d isease No 262 88.8 Yes 29 9.8 Missing 4 1.4 Hyperten sion No 97 33.9 Yes 194 65.8 Missing 4 1.4 Alcoholism No 277 93.7 Yes 14 4.7 Missing 4 1.4 Liver d isease No 283 95.9 Yes 8 2.7 Missing 4 1.4 Kidney d isease No 276 93.6 Yes 15 5.1 Missing 4 1.4 Cancer No 272 92.2 Ye s 19 6.4 Missing 4 1.4 Sickle c ell d isease No 281 95.3 Yes 10 3.4 Missing 4 1.4 Leukemia No 290 98.3 Yes 1 0.3 Missing 4 1.4
125 Table 5 2. Continued Variable Response N % Asplenia No 290 98.3 Yes 1 0.3 Miss ing 4 1.4 Immune System Disease No 288 97.6 Yes 3 1.0 Missing 4 1.4 Cerebro spinal Fluid Leak No 291 98.6 Yes 0 0 Missing 4 1.4 Cochlear Implants No 290 98.3 Yes 1 0.3 Missing 4 1.4 Organ Transplant No 290 98.3 Yes 1 0 .3 Missing 4 1.4 Currently on Dialysis No 287 97.3 Yes 4 1.4 Missing 4 1.4 Receiving Chemotherapy or Radiation No 286 98.3 Yes 5 1.7 Missing 4 1.4
126 Table 5 3. Bivariate a nalyses using chi square Variable Vaccinated P value No Yes n (%) n (%) Age 18 45 ( n = 79) 46 64 ( n = 129) 65+ ( n = 87) Total ( n = 295) 69 94 37 200 (87.34%) (72.87%) (42.52%) (67.79%) 10 35 50 95 (12.66%) (27.13%) (57.48%) (32.21%) p < .001 Ge nder Male ( n = 99) Female ( n = 196) Total ( n = 295) 75 125 200 (75.75%) (63.77%) (67.69%) 24 71 95 (24.25%) (36.23%) (32.21%) p < .05 Insurance Self Pay ( n = 27) Medicaid ( n = 22) All Others ( n = 197) Total ( n = 246) 22 17 139 178 (81.48%) (77.27%) (70.56%) (72.35%) 5 5 58 68 (18.52%) (22.73%) (29.44%) (27.65%) p > .05 Education No College Degree ( n = 185) College Degree ( n = 108) Total ( n = 293) 127 73 200 (68.65%) (67.59%) (68.26%) 58 35 93 (31.35%) (32.41%) (31.74%) p > .05 Awareness (Before this research stu dy, had you ever heard of the pneumonia vaccine?) No ( n = 141) Yes ( n = 154) Total ( n = 295) 129 71 200 (91.49%) (46.10%) (67.80%) 12 83 95 (8.51%) (53.90%) (32.30%) p <.001 Knowledge (Do you know who is supposed to get the pneumonia vaccine?) No ( n = 198) Yes ( n = 98) Total ( n = 293) 160 39 199 (80.81%) (39.80%) (67.92%) 38 56 94 (19.19%) (57.14%) (32.08%) p <.001
127 Table 5 4. Logistic regression m odel for PPSV23 non vaccination 95% C.I for Exp(B) B Sig. Exp(B) Lower Upper Gender Male ( n = 93) Female ( n = 186) 0 .024 ref. .955 0 .976 0 .419 2.274 Age 18 45 ( n = 79) 46 64 ( n = 124) 65+ ( n = 76) 1.422 0 .730 ref. .006 .075 4.147 2.075 1.492 0 .930 11.525 4.628 Awareness No ( n = 132) Yes ( n = 147) 1.868 ref. .000 6.478 2.807 14.952 Trust Low ( n = 79) High ( n = 200) 0 .834 ref. .080 2.303 0 .906 5.855 Susceptible No ( n = 182) Yes ( n = 97) 0 .725 ref. .068 2.065 0 .948 4.500 Provider r ecommendation No ( n = 179) Yes ( n = 100) 1.989 ref. .000 7.30 3.6 11 14.785 Odds ratios indicate the probability of non vaccination. For example, those without provider recommendation were 7.3 times less likely to be vaccinated than those with a recommendation.
128 Table 5 5. A p riori and post h oc a naly ses using chi square Independent variable Dependent variable P value Refusal Gender No Yes Male ( n = 98) 87 (88.78%) 11 (11.22%) Female ( n = 193) 175 (90.67%) 18 (9.33%) Total ( n = 291) 262 (90.03%) 29 (9.97%) p > .05 Insurance No Yes Self p ay ( n = 27) 25 (92.59%) 2 (7.4%) Medicaid ( n = 22) 21 (95.45%) 1 (4.55%) All Others ( n = 193) 171 (88.6%) 22 (11.4%) Total ( n = 282) 217 (89.67%) 25 (10.33%) p > .05 Trust in health care provider Age Low High 18 45 ( n = 79) 33 (41.78% ) 46 (58.22%) 46 64 ( n = 125) 38 (30.4%) 87 (69.6%) 65+ ( n = 79) 11 (13.92%) 68 (86.07%) Total ( n = 283) 82 (28.98%) 201 (71.02%) p =.001 Racial concordance Low High No ( n = 194) 53 (27.31%) 141 (72.42%) Yes ( n = 63) 18 (28.57%) 45 (71.42%) Total ( n = 257) 71 (27.63%) 186 (72.37%) p > .05 Perceived health ( SF ) Age Unhealthy Healthy 18 45 ( n = 76) 13 (17.1%) 63 (82.9%) 46 64 ( n = 114) 32 (28.07%) 82 (71.93%) 65+ ( n = 69) 27 (39.13%) 42 (60.87%) Total ( n = 259) 72 (27.8%) 187 (72.2%) p < .05 PPSV23 uptake Smoker No Yes No ( n = 242) 159 (65.7%) 83 (34.3%) Yes ( n = 47) 38 (80.85%) 9 (19.15%) Total ( n = 289) 197 (68.17%) 92 (31.83%) p < .05
129 CHAPTER 6 DISCUSSION Following a brief review, Chapter 6 will present an interpretation of statistical analyses of the major findings with i n the context of the literature and theoretical underpinnings In addition to answering the study aims several find ings extraneous to the study ai ms but relev ant to the overall purpose will also be discussed. Moreover, this section will address study limitations and implications for future practice as well as research and policy recommendations Review It is estimated that pneumococcal pneumonia i s directly responsible for 5,000 deaths in the United States annually (Flowers, 2007). Case fatality rates for pneumococcal pneumonia in 1997 were estimated to be 12% (Feikin, 2000). A decade later the case fatality rate remained 10% and it is estimated t hat the rate is much higher among the elderly who often go undiagnosed (CDC, 2009). The incidence of pneumococcal illnesses has been found to be higher among African Americans than their Caucasian counterparts in the US by two to four fold (Butler & Schuch at, 1999). A better understanding of the characteristics consistent with the vaccinated portion of the population is needed in order to address and eliminate dispari ties ; therefore the purpose of this study was to examine factors such as age, gender, SES, vaccine awareness, knowledge, trust in healthcare provider, perceived health perceived susceptibility perceived severity prior negative experiences with vaccines, and healthcare provider recommendation for vaccination to predict pneumococcal vaccine up take among eligible African Americans 18 years of age.
130 Major Findings This study revealed several statistically significant factors predicting PPSV23 vaccine uptake. Significant findings in bivariate analyses were the relationships between age, gender, aware ness knowledge trust in healthc are provider, perceived susceptibility, and healthcare provider recommendation and PPSV23 uptake Using backward logistic regression, age vaccine awareness, and healthcare provider recommendation significantly predicted PPSV23 vaccine uptake in the whole model Consequently, t he alternate hyp otheses were retained in each instance. H a Sociodemographic variables such as age gender, and socioeco nomic status may predict pneumococc al l vaccine uptake am ong eligible African Americans The alternate hypothesis is retained given the significant impact age and gender had on PPSV23 uptake in this study. H a Constructs operationalized from the HBM and PAPM such as awareness knowledge, trust in healthcare provider, perceived health, perceived susceptibility perce ived severity, prior negative experience with vaccines, and healthcare provider recommendation may predict pneumococcal vaccine uptake among vaccine eligible African Americans. The alternate hypothesis is retained given the significant impact awareness, kn owledge, trust, perceived susceptibility, and healthcare provider recommendation had on PPSV23 uptake in this study. Interpretation of Results Gender In bivariate analyses, gender was found to significantly predict PPSV23 uptake. With males less likely to be vaccinated than females, gender remained statistically
131 significant in the regression model until the final iteration highlighting the importance of this demographic variable for further exploration and possible targeted interventions. Historically, Afr ican American males have been less likely to seek out preventative care (Byrd, et al. 2011). Moreover, males come into contact with healthcare providers less often than their female counterparts (Byrd), thus decreasing the likelihood that their healthcare provider would have the opportunity to recommend the PPSV23 vaccine. Unfortunately, this population suffers a disparate burden of chronic illnesses such as hypertension, diabetes, and renal disease all of which place African American males at a higher ri sk for pneumococcal illnesses. A prospective exploratory analysis conducted by Klag, et al. (1997) found that African American males had a 44.22 per 100,000 case incidence of end stage renal disease (ESRD) compared with a rate of 13.90 per 100,000 with th eir Caucasian counterparts. Diabetes, poorly controlled hypertension, and smoking were cited as some of the reasons for the marked increase in ESRD among African Americans. In another study HIV prevalence between genders and ethnicities was evaluated using a multisite survey ( n = 3,316) and African American males were found to have the highest prevalence of HIV (16%) compared to only 6.9% for Latino males and 3.3% for Caucasian males ( Harawa et al. 2004). Given the increased risk factors for pneumococcal illnesses, underutilization of healthcare systems, and well documented under vaccination rates among African American males, more emphasis on the health of this population and their rationale for accessing or refusing preventative care is sorely needed.
132 A wareness, Knowledge and Healthcare Provider Recommendation The constructs of awareness and knowledge are by definition, different from one another Accordingly, they were examined separately for the purposes of this res earch With each variable statisti cally significant in the regression model after controlling for the demographic variables age and gender, this prospective study demonstrates that PPSV23 acceptance may depend largely on awareness and knowledge. From a common sense perspective, one could c onclude that if a person was unaware of the vaccine existence, he or she would remain unvaccinated. Moreover, lack of knowledge concerning risks, benefits, eligibility, and availability would likely produce similar findings. Though each of the participan ts in the pneumococcal study met eligibility criteria for PPSV23, 141 (47.8%) had never heard of vaccine before study 198 (67.1%) did not know who should receive the vaccine, and 143 (48.47%) felt they did not have enough information about the vaccine to make an informed decision Moreover, 67 (22.7%) participants answered t he knowledge question with a neu tral response implying they also lack ed adequate information to accept PPSV23 T hese findings are congruent with a large body of literature suggesting th at vaccine uptake could be improved significantly with educational strategies aimed at enhancing knowledge and addressing the unique needs of a given population ( Berkley Patton, et al., 2010; Daniels, Juarbe, Rangel Lugo, Moreno John, & Prez Stable 2004; Thomas, Ray, & Morton 2003) Focus groups conducted in faith based communities in San Francisco ( n = 22) found that participants were not routinely informed of vaccine indications, benefits, and side effects when visiting their healthcare providers (Dani els, et al. 2004) thus
133 highlighting missed opportunities to educate and subsequently vaccinate eligible adults. Reasons for these missed opportunities to educate the lay public and enhance vaccine uptake are not entirely clear; however, a number of st udi es ( Fry, 2008; Rushton, Ganguly, Sinnott, & Banerji, 1994 ) have identified deficient knowledge among healthcare provider s themselves leading the author to conclude that awareness, knowledge, and healthcare provider recommendation are interwoven. In 2008, Fry conducted a prospective quantitative survey among registered nurses in an acute care setting and found that the primary reason nurses did not follow protocol for encouraging and administering PPSV23 to inpatients w as their own lack of knowledge (unpubl ished thesis) Similarly, a study examining internal medicine knowledge of PPSV23 indications, safety, and efficacy ( n = 33) found a sizable number 69.7% of participants were unaware of vaccine guidelines, and 60.6% had an exaggerated fear of ad verse reactions thus presenting a significant barrier to provider recommendation (Rushton, et al. 1994). If physicians and nurses lack the proper information on PPSV23, it stands to reason that they will not educate their patients or otherwise endorse va ccine uptake. While variables such as age and gender may be non modifiable barrier s to vaccination, awareness, knowledge and recommendation can most assuredly be changed with a concerted effort among healthcare providers and public health officials. Gaps in awareness and knowledge identified in this dissertation study and the work of others provide healthcare professionals and public health officials with a number of actionable item s on which to frame interventions Interventions such as targeted message s for underserved at risk populations faith based outreach programs, and
134 healthcare provider education modules may prove beneficial and have been detailed in the recommendations for future research practice, and policy section Trust The relevance betw een provider trust and vaccine uptake was examined carefully in this dissertation study and findings in bivariate analyses suggested that participants with higher trust score s were more likely to be vaccinated. When controlling for age, gender, awareness, and knowledge in the regression model, trust did not remain significant perhaps because participants in the pneumococcal study had fairly high trust scores overall ( n = 295, mean trust score 39.27, range 10 50). Another possible explanation for these findi ngs is the difference in trus t scores between age and gender categories. Younger participants had lower trust scores than older participants, and m ales had lower trust scores than their female counterparts Findings between trust and PPSV23 uptake in the p neumococcal study are congruent with a sizable body of evidence supporting the relationshi p between trust in healthcare providers and increased adherence to preventative health behaviors such as vaccine uptake ( Altice et al. 2001; Sohler, et al., 2007; Th om et al. 1999 ) Further, although not measured in the pneumococcal study, the literature overwhelmingly supports the notion that African Americans are less trusting of healthcare providers than their Caucasian counterparts ( Boulware, Cooper Ratner, LaVe ist, & Powe 2003; Doescher, et al. 2000 ; Saha, Komaromy, Koepsell, & Bindman 1999 ). Rationale cited in the literature for this phenomena include s h istorical influences such as slavery, the Tuskegee incident, and discrimination which persists even today l eaving a lasting impression on many members of the African American community, and leading them to
135 feel exploited distrustful, and reluctant to follow prescribed treatment regimens, or preventative health int erventions (Thomas, & Quinn, 1991). Undoubtedly there are many African Americans whose parents or grandparents can still prov ide first hand accounts of the Tuskegee atrocity. Given the historical backdrop, it would seem intuitive that older males would be the least trusting of all; however, this did n ot prove entirely true in the pneumococcal study. While males were less trusting than females, younger participants were less trusting than their older counterparts. Much of the lite ratur e ( Allen, et al., 2007; Armstrong, Ravenell, McMurphy, & Putt 2007; Jones, et al., 2009 ) found on trust focused on patterns between races as opposed to within and between genders and age categories A few s tudi es ( Allen, et al., 2007; Benkert, Hollie, Nordstrom, Wickson, & Bins Emerick 2006 ) did cite g ender barriers to t rust within the African American population as a central issue of concern but it does seem as though there is a gap warranting further exploration. An unexpected finding in the pneumococcal study was the statistically significant difference in trust betwe en younger and older African Americans. P articipants age 18 25 ( n = 19) had a mean trust score of 37.79 whereas those ages 56 64 ( n = 48) had a mean score of 44. No studies examining African Americans, age, and trust in healthcare provider were identifie d in the literature, again highlighting a potential gap in the state of the science. That said, as a n experienced nurse the researcher believes one possible explanation for lower trust scores among younger participants is the changing healthcare delivery system. Anecdotally, older patients are much more likely to blindly trust their healthcare provider taking medications and having procedures they often understand because the physician told them it was necessary. The paternalistic
136 nature of the physi cian patient relationship ha s begun to evolve over the past two decades into more of a partnership where the patient is now a client and the physician or extender engages in a conversation about treatment options rather than a dictatorial relationship Y et challenges do remain with one author calling the current state of care crisis in communication between physician and patient and in the trust in health care professionals in ( Hoffman, 2002 p 89 ) A nother plausible explanation for the differi ng trust scores between age groups in the pneumococcal study is the premise that a trusting relationship between patient and provider develops over time through multiple encounters and familiarity. Though the dynamics of trust between provider and patient cannot be explained in a simplistic manner, a study using semi structured interviews in the UK ( n = 20) found that trust was fragile, easily undermined, and developed over time as a series of positive interactions unfolded. Though the authors in this study did not break trust out into age categories, they emphasized continuity over time as a crucial component to trust (Tarrant, Dixon Woods, Colman, & Stokes, 2010). If this holds true, it would stand to reason that older patients who have likely been seeing a given provider for a longer period of time would be more trusting. In another s tud y ( Grande, Shea, & Armstrong, 2011 ) authors examine d how trusting relationships were affected when patients believed their healthcare provider received gifts from pharmac eutical companies. In a telephone based interview participants were asked if they believed their personal physicians and/or all physicians accepted gifts from pharmaceutical companie s. African American ( n = 762) and Caucasian ( n = 1297 ) participants who b elieved gifts were accepted by their physicians
137 were far less likely to trust their providers and the healthcare system in general. Analyses revealed that 55% of respondents believed their personal physician did accept gifts. Further, 31% believed all phys icians accepted gifts routinely. Similar to our pneumococcal study, younger respondents were less trusting of their healthcare providers. In this instance, lack of trust among younger patients was related to the widespread belief that physicians accepted p harmaceutical gifts with 64.9% of age d 40 64 believ ing gifts were accepted and only 26.4% of those ages 65 and above h olding this belief (Grande, et al. ). Though the Grande study concluded that lower trust was related to acceptance of pharma ceutical gifts, additional variables not included in their study could also be contributory. There was no obvious explanation for the lower trust scores among younger participants in the pneumococcal study, which is why more research on this subject is rec ommended. Findings in the pneumococcal study yielding lower trust scores among males than female participants were found to be congruent with existing literature in related areas Bass, et al. (2011) used focus groups to examine perceptions of colorectal screening among African Americans and found that males were less likely to be screened than their female counterparts ( n = 22), in part due to distrust and actual or perceived racism Moreover, numerous studies have reported that African American males go to the doctor less often than their female counterparts ( Allen, et al., 2007), which may provide a parsimonious explanation for a less developed rapport with their healthcare provider, hence lower trust. While the pneumococ cal study findings are consisten t with the colorectal screening study in as much as males were less often vaccinated and had
138 lower trust scores overall more research is needed to better understand differences found between trust and gender among African Americans. In post hoc analyses, one final variable r acial concordance with healthcare provider as a means of enhancing trust was examined using data from the VUQ. In total, 69 participants in the pneumococcal study reported being racially concordant with their healthcare provider. Of th ese 69 participants, 27 (39.13%) were vaccinated whereas among those without racially concordant providers ( n = 198), 55 (27.77%) were vaccinated with PPSV23. Though not statistically significant, 2 (1, n = 293) = 2.058, p > .05 a difference in vaccine u ptake between groups of 11.36% may wel l be clinically significant and findings are consistent w ith literature citing racial concordance between patient and provider as a means by which to enhance trust and facilitate adherence to preventative measur es and treatment modalities (van Wieringen, 2002 ) According to Cooper & Powe (2004), African Americans are in the healthcare profession with fewer than six percent of physicians and nine percent of nurses in the United States self identifyi ng as African American. Although no studies were found linking racial concordance with improved patient outcomes overall, a number of studies linked concordance with higher patient satisfaction, better communication, and higher overall trust scores (Cooper 200 4 ; Malat, 2001; & Oliver 2001). A study conducted in 16 urban primary care clinics in 2003 used pre and post visit surveys to examine the relationship between racial concordance and patient satisfaction. Authors of this study found that patients with racially concordant providers had longer visits (+2.15 minutes, 95% CI 0.60 3.71), higher satisfaction scores, and a more collaborative relationship (+8.42 points, 95% CI 3.23 13.60) (Cooper). With a large
139 body of evidence linking trust in healthcare prov ider with provider concordance, it is recommended that future research on concordance focus specifically on patient outcomes in order to strengthen the argument for policy change encouraging recruitment of African Americans into the healthcare field. These recommendations are discussed in further detail in the policy section of this chapter. Despite the non modifiable variables historical backdrop and possibly racial concordance there are a number of actionable items which could influence and improve trust in healthcare provider among African Americans. Although not all inclusive, reasons attributing to distrust of healthcare providers in contemporary society included excessive wait times, poor quality of care, bias and stereotyping on the part of the healt h care provider perceived racism, lack of continuity of care, perceived acceptance of pharmaceutical gifts, and poor communication ( Bass, et al. 2011; Grande, et al. 2011; & Smedley, et al. 200 2) In the past decade, trust has been increasingly recogni zed a s a crucial component of health care; however, challenges with measuring the construct and the development of targeted interventions leading to improved trust in healthcare provider remain elusive given the multi dimensional nature of the construct (H all, Dugan, & Zheng 2001) In order to develop a successful vaccination program for PPSV23, providers must not only be educated on vaccine indications and provide recommendation for the vaccine they must also develop and nurture a trusting relationship with their patients or under vaccination particularly among African Americans will likely continue.
140 Unsure Using the VUQ, a priori analyses found that 24 subjects reported unsure/unknown vaccine status This problem is not unique to the pneumococcal stu dy. Difficulty determining PPSV23 vaccination status is a persistent problem in the literature among healthcare professionals with approximately 57% of physicians reporting unknown vaccination status as a significant barrier to th is important preventative health measure (Mieczkowski & Wilson 2002). Practice guidelines suggest that anyone with unknown PPSV23 vaccine status be revaccinated (CDC, 2010); however, this can result in localized pain and inflammation at the injection site which, although not serio us, would be uncomfortable for the person and potentially lead to a negative attitude towards future vaccines. Revaccination among adults age 18 64 is relatively uncommon; however, the same cannot be said of the population aged 65 and older. Currently, gui delines state that PPSV23 should be given at the age of 65 even if a person has been previously vaccinated. That said, there are no definitive guidelines for the interval between these vaccinations ; therefore it is feasible that a person could receive PPS V23 at age 64 due to a pre existing condition such as diabetes, then be revaccinated at the age of 65 just by virtue of their age. As many as 18% of adults aged 65 and over have reportedly been revaccinated often with in an interval of five years or less ( J ackson, Baxter, Naleway, Belongia, & Baggs, 2009) U nlike most vaccines currently offered to adults, PPSV23 eligibility criteria are somewhat confusing even to healthcare providers (Fry, 2008; Mieczkowski, et al. 2002; & Shevlin, et al. 2002), which cou ld explain why some
141 remain unvaccinated while others receive needlessly repeated PPSV23 immunizations. Moreover, there is a financial burden associated with revaccination. T he burden of revaccination cost must be assumed by someone whether the individual or the insurance company, thus contributing to rising costs and wasteful spending in healthcare. A careful examination of this persistent oversight is therefore one of the recommendations of this dissertation study. With the advent of electronic persona l health records, it may be possible for individuals to maintain better records of immunization moving forward. A multidisciplinary task force whose purpose was to identify barriers to PPSV23 uptake found that incomplete or inaccessible documentation was a significant problem (Rehm, File, Metersky, Nichol, & Schaffner 2012 ) Although it is yet unclear whether or not electronic health records (EHR) would impact personal r ecord keeping in a positive way, an integrated EHR with interface capabilities between h ealthcare facilities would unquestionably reduce duplication and eliminate some of the widespread repetition of diagnostic procedures, treatments, and preventative measures such as immunization currently seen in both acute and primary care settings Reco mmendations for handling unknown/unsure vaccine status are discussed further in the practice section. Refusals PPSV23 refusers were also examined in a priori analyses. A single item question on the VUQ asked participants you ever refused the pneumo nia While 29 (9.8%) reported vaccine refusal, nine of them also reported h aving received PPSV23. Upon further consideration, the question as it was worded could not be considered mutually exclusive of actual vaccination and is therefore discussed in the
142 limitations section Despite the poor performance of the refusal question, refusal reasons provided by participants such as fear of needles and cost, ( F igure 5 1) were congruent with the literature such as the study focusing on refusal reasons wh ereby authors examined 500 medical records investigating PPSV23 and found that healthcare provider gender discordance, and the absence of a healthcare provider were statistically significant (OR 2.09, CI 1.07 4.09, p < .05, and OR 2.26, CI 1.13 4.49, p < .005 respectively) (Miller, Kourbatova, Goodman, & Ray, 2005). G ender discordance between provider and subject was not addressed on the VUQ ; therefore results from the Miller, et al. st udy cannot be compared with this variable. Although not ident ical to the Miller study, p articipants in the pneumococcal study were asked re do you typically seek with response options as follows: office, emergency department, urgent care clinic, health department, free clinic, or other. The majo rity of participants reportedly received care in a office ( n = 237, 80.3%) with 14 (4.7%) responding that they used the ED for care. PPSV23 r efusals and care seeking locations were evaluated in post hoc analyses using chi square analyses and did not prove significant. N evertheless those who typically received care at a office refused at a rate of 10%, whereas those who received care in the emergency department or a free clinic refused at rates of 14% and 33% respectively. Based on prior discussions reporting higher vaccine acceptance rates when a long standing trusting relationship exists between p atient and provider, it is reasonable to assert that those who seek care in facilities such as the ED where continuity of care with the same provider could not be expected, would be more inclined to refuse. With
143 the exception of the Miller et al. study, no other research was found linking vaccine refusals to primary care locations. In order to more fully evaluate the demographics and characteri stics of vaccine refusers, a larger study with more rigor than a medical record review is therefore suggested and detailed in the future research section Age Perceived Health and Perceived Susceptibility A number of studies have asserted that percei ved health and perceived susceptibility influence the likelihood of an individual engaging in preventative health behaviors ( Dassow, 2005 ; Rimple, Weiss, Brett, & Ernst, 2006; Santibanez, et al., 2002 ). or her well being as opposed to an objective assessment which might consider factors such as the presence of a disease process or disability ( Cummings & Jackson, 2004 ). When an individual perceives themselves as healthy, they in turn tend to perceive them selves less vulnerable. This inverse relationship was evident in the pneumococcal study and congruent with existing research. Because perceived health and susceptibility are subjective, the mere presence of a disease process such as diabetes or COPD may or may not lead the individual to re port themselves as unhealthy or vulnerable. Several studies ( Glover, 2010; Hunt, 1980; Ru Chien, et al., 2006 ) have emphasized the importance of obtaining subjective health status information from individuals to examine th e relationship between perceived health and susceptibility and accessing preventative services. Based on results from the SF portion of the VUQ, perceived health was not significantly predictive of PPSV23 uptake in bivariate analyses or the overall l ogistic regression model. Upon further examination of perceived health and susceptibility
144 analyses, a patter n was identified between age groups. Y ounger participants perceived themselves as healthier than older participants with a m ean perceived health sco res for those age d 18 25 at 47, compared with a mean score of 41.41 for those age d 65 75 and 37.09 for subjects over 75. P ost hoc bivariate analyses identified statistically significant perceived health and perceived susceptibility scores between age grou ps While this may not be surprising, the concern lies in the fact that all of the subjects in the pneumococcal study (even those under 65) met eligibility criteria for PPSV23 based on their preexisting disease processes Consequently despite their subjec tive interpretations of health and vulnerability, they may well be at risk for pneumococcal illnesses. Prior to this study, research on PPSV23 uptake has largely focused on persons aged 65 and older likely because of automatic eligibility and because elder ly are at increased risk of pneumococcal illnesses (Albanese, 2002 ; Butler, 1999). Nevertheless, if the pneumococcal study ( n = 295) is representative of the African American population at large, a great percentage of persons under 65 are indeed PPSV23 eli gible and potentially falling through the cracks. Based on self reported age on the VUQ 87 of the 295 participants were 65 or older highlighting the importance of including the 70.5 % of eligible adults been largely excluded in prior PPSV23 researc h. Intuitively, a healthy 65 year old person is far less likely to get pneumonia than a 45 year old smoker with immunological suppression though this would be very difficult to measure Perhaps youth creates a false sense of security with respect to perce ived health and susceptibility, thus contributing to under vaccination in this particular population
145 Though findin gs in the pneumococcal study were intuitively as expected, there is conflicting data in the literature on perceived health and age Some stud ies report ed that younger participants were more likely to perceive themselves in better health ( Cott, Gign ac, & Badley 1999; Denton & Walters 1999 ) while others reported that perceived health actually improves with age ( Damian, Ruigomez, Pastor, & Mar tin Moreno, 1999; Smith Shelley, & Dennerstein, 1994) Perhaps some of the variance is due to the wide range of instruments designed to measure perceived health. Though as previously discussed, the SF used in the pneumococcal study is widely accepte d as a valid reliable measure of perceived health, there is also a great deal of support for a single item question such as the one in the BRFSS despite its simplicity (Benyamini & Idler, 1999; Bosworth, 1999). Examination of the literature also illumin ates the fact that perce ived health is a highly complex and multi dimensional factor with each respondent assessing his or her own health a little differently. For example, in a secondary analysis of 1485 older adults with hypertension, some adults with h ypertension perceived themselves as very healthy, while still others perceive d themselves as unhealthy by virtue of the fact that eat enough fruit or exerci se as often as they would like. When the authors examined non modifiable and modifiable factors thought to be predictive of lower perceived health they found that age, SES (using education as a proxy variable), and the presence of one or more chronic illnesses accounted for much of the variance in the overall model in this older population ( Lewis & Reigel, 2010) Based on the diverse measurement options conflicting reports in the literature, an d individualized beliefs that seem to influence perceived health, it is difficult to draw
146 any conclusions from the findings in the pneumococcal study in regards perceived health. H owever, given the descriptive statistics which clearly show that younger people perceiving themselves as healthier, and less likely to be vaccinated than older participants despite universal eligibility, more research on the impact perceived health has on preventative care is sorely needed. Further, the studies supporting the use of a single item measure for perceived health ( Chandola & Jenkinson 2000; Mossey & Shapiro 1982; Ross & Wu 1995), as opposed to a questionnaire such as the SF create an opportunity for a comparative study between instruments. Each of these possibilit ies is discussed further in the recommendations for future research section Perceived Severity Cumulative severity scores for pneumonia were quite high with 98.6% of respondents ( n = 281) believing that pneumonia was very serious and could lead to death. This perception did not translate to vaccination however, which is some what inconsistent with findings in the literature supporting an increase in preventative health behaviors when perceived severity is high (Dassow, 2005 ). In an endeavor to better understand this divergent finding, the concept of fatalism was explored in the literature. Two definitions for the term fatalism were found: on one hand, fatalism is defined as the belief that a given disease ( e.g. cancer) is a death sentence and therefore screening ( e.g mammogram) would not be of any benefit. The second, more relevant definition for the purposes of the pneumococcal study holds that contr acting an illness or a disease is Nakagawa, Gregorich, & Kuppermann, 2010; Spurlock & Cullins, 2006).
147 According to the literature, persons who embrace this premise ar e far less likely to engage in preventative health behaviors because they are not seen as making a difference (Spurlock & Cullins, 2006). Moreover, studies have found that African Americans, particularly women, are more likely to embrace the perception of fatalism (Mayo, Ureda, & Parker, 2001; Facione, Miaskowski, Dodd, & Paul, 2002). All of the studies on the construct of fatalism pertained to cancer screening and pregnancy, each supporting the notion that when perceived fatalism was high, preventative hea lth behaviors were low. No studies were found on the topic of fatalism and vaccine uptake, thus highlighting a potential gap in the literature. Theoretical Framework The Health Belief Model (HBM) has been used extensively in research to explain or analyze preventative health behaviors including vaccine uptake ( Becker et al. 1977; Mirotznik, Feldman, & Stein, 1995). According to the Health Belief Model, motivation to engage in preventative health measures may stem from perceived disease risk ( susceptibili ty ), perceived disease severity, and overall perceived benefits (Thanavaro Moore, Anthony, Narsavage, & Delicath 2006). The Health Belief Model does not however, take preventative behavior awareness or knowledge into account. For example, a person may pe rceive pneumonia to be a very serious illness which could lead to death; however, they may be completely unaware of the availability of the PPSV23 existence. The Health Belief Model and Precaution Adaptation Model were consequently substructed fo r this study based on the clinical experience whereby lack of awareness has been shown to impact vaccine uptake repeatedly. The
148 model ( F igure 4 1) provided a framework on which to pin all of the constructs of interest in a manner that was logical and easy to follow. Each dimension of the model (perceived barriers, cues, and perceived health) functioned as predicted with the exception of perceived severity which was not statistically significant in bivariate analyses or the logistic regression mode l Though beliefs about severity were quite high in this study, it is quite possible that perceived susceptibility mediated perceived severity particularly given the fact that very few ( n = 24) considered themselves highly susceptible to pneumonia. It is also possible that the aforementioned construct of fatalism played a role in perceived severity versus PPSV23 uptake although this was not measured on the VUQ. One of the unique aspects of this dissertation is the fact that to our knowledge no research has been performed on the relationship of pneumococcal illness to Health Belief Model constructs f or vaccine uptake as a preventative health behavior. Additionally, no works with substructed models from the HBM and PAPM were found despite the fact that awaren ess appears to play a critical role in vaccine uptake. Lessons Learned A number of important lessons worthy of discussion took place during the dissertation process, particularly in the data collection phase of the study. Some of the decision points and ch allenges will now be discussed in detail. As the researcher prepared to begin data collection, nine churches were contacted via telephone, letter, or e mail requesting permi ssion to conduct the pneumococcal study at their facility. Although none of the chu rches refused the study, seven of the nine did not follow up despite repeated attempts to reach an administrator able to consent to the process. The
149 two churches who granted permission for the study to take place had very different demographics with the Ca tholic Church having a much older population than the Baptist church. While this could likely be seen as an advantage for representativeness, the of each congregation was distinctly different and worthy of some discussion. Each Sunday over t he course of d ata collection the PI and RA went to the church together. After only two Sundays, all members of the congregation at the Catholic Church who were willing to participate had done so ( n = 78) Though prescreening flyers had been circulated and an announcement had been made after services two weeks prior to our arrival, parishioners seemed hesitant to participate. As the PI and RA stood in the lobby after services to encourage individuals to participate, the PI was greeted, but largely ignored. The RA had a much higher success rate when she approached parishioners to complete t he survey. While no data were gathered to substantiate this observation anecdotally it was noteworthy and theoretically it seems consistent with much of the research sugg esting that might have a difficult time recruiting subjects in the African American community (Rivera Goba, et al. 2011; Smith, et al. 2007). In this instance, c hoosing an African American RA to assist with the study provided a vital link to the community of interest and seemed to enhance recruitment and participation. Data collection at the Baptist Church was a different exp erience, likely due in part to the fact that the Pastor made a compelling speech to the congregation asking them to supp ort the research and b ecause the congregation was larger and had a younger demographic than the Catholic Church Members of the congregation were very willing
150 to participate, and after only two visits to Wednesday Bible study and four visits after Sunday c hurch services 247 surveys had been completed. On one occasion, consent and data collection were quite challenging given the large number of participants ( n = 79). Although the PI and RA were each on hand, answering questions as they arose was slightly mo re challenging than it would have been in a smaller setting. Though no significant concerns or problems arose, for future studies, the number of participants on a given day might be restricted Study Limitations Results from this descriptive study had seve ral limitations thus results should be interpreted with caution. Sample size, instrument validity and reliability, and the descrip tive nature of the study were all study limitations. The study population may not be representative of the population i n gene ral Correlation as a statistical measure of a relationship between two or more variables informs the nature of the relationship between the variables, but does not imply causation. Receipt of pneumococcal vaccination was based on self report and not valid ated. Each of these limitations will now be discussed in detail. To determine an adequate sample size, two predictor variables were considered: awareness of vaccine and provider recommendation From the literature it is estimated that 60% in this populatio n are aware of the vaccine and 50% are cued by their health care provider to receive the PPSV23 vaccine (CDC, 2011 ; Winston et al. 200 6 ). It was determined that a sample size of 300 would provide adequate power (i.e. power greater than .80). The power ana lysis was conducted using the POWER procedure in SAS for logistic regression (version 9.2, Cary, N.C.). The level of significance was set at 0.05.
151 Assuming an unvaccinated rate of .60 and a dichotomous predictor with prevalence between .30 and .70 (for exa mple, between 30% and 70% of the sample report that they are aware of the vaccine), the detectable odds ratio for variable significance is at least 2.1 for a sample size of 300 with power at least .80 (Faul et al. 2009) Thus the sample size was still sl ightly small ( n = 295 ) and shou ld be considered a limitation. Although subject s from two churches were recruited, over sampling occurred at one of the c hurch es due in part to the larger congregation and because of the willingness of the Pastor to allow th e PI greater access. Self selection bias was likely a limitation as well, because although subjects read a document with eligibility criteria, the y were not truly pre screened. Essentially, subjects themselves pre screened and decided they would or would n ot participate. Upon examination of the survey, 26 subjects did not me et the eligibility criteria and were eliminated from the analyses. Monetary compensation may have served as a motivator for participation despite ineligibility in some cases. There was n o evidence of self selection bias immediately obvious in the study other tha n those who proved ineligible. In other words, age, gender, vaccine knowledge, etc. appeared fairly randomly distributed. There were no obvious patterns in the findings s uggesting that one group or another was more likely to participate other than the fact that 2/3 of the subjects were female. It is well documented in the literature however that African American females are more likely to participate in research than mal es (Rich & Ro 2002; Satcher, 2003) In order to obtain IRB approval for the pneumococcal study, subject s had to be completely anonymous with no signature on the consent form and no identifiers on the survey itself. While the anonymous nature of the survey may have protected sensitive
152 health information and enc ouraged respondents to be more forthcoming, it also created several limitations. With no identifiers on the survey and data collection taking place at two locations (one Catholic and one Baptist churc h) over the course of several weeks, there was no way to guarantee that respondents t ake the survey more than once although there was no evidence that this occurred. The VUQ survey relied entirely upon self reported data creating a significant threa t to validity Although there were no incentives to be dishonest, inaccurate or carelessly reported data remains a distinct possibility with self reporting ; therefore results shoul d be interpreted with caution. The VUQ although constructed using three p reviously validated instruments had not been tested independently prior to utilization in this study. The pilot study served as a venue to explore readability, functionality, and cultural sensitivity, however, questions were not validated using wide ly acc epted statistical methodology such as factor analysis M any of the questions in the VUQ were created specifically for this study and not previously tested for validity and reliability. Questions for trust and perceived health were taken entirely from valid ated instruments and were therefore of no concern. Questions to evaluate awareness, knowledge, perceived susceptibility perceived severity, and healthcare provider recommendation were created using the HBM framework, but wording varied from study to stu dy, so these questions were essentially untested prior to this research. When analyzing the data, it became clear that the survey would provide valuable informatio n for those who had received the vaccine, but limited information on those who had refused i t. Although subjects had the opportunity to fill in a response when they had refused, many left this section blank O thers who had not refused the vaccine
153 completed one or more check boxes indicating refusal reasons hence a measurement bias for this quest ion is possible More in depth data collection focused specifically on respondents who refused PPSV23 might have been useful in determining how best to reach this population and will likely be considered in future studies Finally, the question pertainin g to education level appeared problematic with a higher than expected number of respondents reportedly having completed graduate school. Though this question was taken directly from the BRFSS, a measurement bias is suspected, but cannot be confirmed. Impl ications for Pra ctice, Research and Policy At this juncture, PPSV23 uptake disparities among African American adults in the US have been reported in the literature and confirmed in this dissertation study. Although researchers could continue to collect in formation on PPSV23 uptake across the country, findings appear to be consistent with those reported by the CDC and Healthy People 2020. Results of this dissertation offer an array of explanations and complexities associated with PPSV23 uptake among African American adults including acceptance issues with age, gender, awareness, knowledge, trust, perceived susceptibility, and healthcare provider recommendation. Bearing this in mind, the author recommends t hat the aforementioned findings from this and other s tudies be utilized to design targeted intervention s for practice development of future research programs, and to inform policy changes all designed to increase PPSV23 uptake in the African American community. Recommendations will now be discussed in detai l.
154 Implications for Practice A major emphasis on information and education surrounding pneumonia prevention, risk factors, and PPSV23 vaccine uptake is essential for prevention efforts to be effective. Findings in this research clearly delineate deficits i n awareness, knowledge, and healthcare provider recommendation for the vaccine. Moreover, awareness does not necessarily indicate knowledge of vaccine eligibility, risks, benefits, and immunization guidelines. It is therefore the recommendation that the confusion surrounding eligibility be assuaged not only by targeting the African American community, but the healthcare community itself. Nurses, e xtenders, and p hysicians have annual requirements for continuing education. The PI created an educati on module highlighting the importance of PPSV23 uptake in 2008. This module could be tailored to each audience and offered as a continuing education credit to enhance healthcare provider knowledge of PPSV23, thus producing a sense of urgency with those in a position to make an impact on the problem. Data from this research provides elements to recommend that such an education module be extended to healthcare providers on a larger scale. Integration of PPSV23 vaccine education into continuing education modul es for physicians, extenders, and registered nurses in order to bolster healthcare provider recommendations is therefore a primary recommendation for practice Based on findings from this and other st udi es ( Staffileno & Coke, 2006 ) African Americans ha ve shown a high affinity for face to face communication and an emphasis on trusting relationships. With this in mind, another recommendation for future practice for public health nurses is to e nlist key stakeholders such as community leaders, parish
155 nurses and pastors to deliver messages to African American communities about the importance of vaccination in an effort to assuage existing fear and mistrust of outsiders Moreover, dialogue about cultural sensitivity and the impact it may have on trusting rela tionships between he althcare provider and patient should begin among healthcare providers with an eye toward decreasing disparities in the African American community. Given the body of literature that supports the development of trust over time, emphasis o n the importance of having a regular healthcare provider with whom the individual can build a rapport, maintain a complete and accurate health history, and become a partner in his/her preventative care and treatment regimens is also recommended Though smo king cessation efforts are already in place at a high level, many continue to smoke ( n = 47 in this study). Develop ment of education and smoking cessation programs targeting the African American community with an emphasis on increased risk factors for many disease processes including pneumococcal illnesses should be considered Finally, with evidence to support deficient health care among African American men, this study suggests that practitioners strive to i dentify the unique charact eristics of this po pulation pertaining to preventative health and develop targeted interventions to better meet their needs. Implications for Research A number of important findings in this dissertation study highlight the need for continued efforts to understand healthcare disparities among African Americans. First, it is recommended that researchers c ontinue to study the construct of trust in healthcare provider perhaps using CBPR methodology to ask the African American community
156 how best to enhance trust between patient and health care providers Rather than assuming historical influences have shaped trust/distrust, the use of semi structured qualitative interviews to ask African Americans what specifically leads them to feel trust or distrust in their healthcare provide r and the healthcare system as a whole could prove quite useful. Additionally, exploration of patterns in trust between genders and age groups in the African American community to determine if needs and priorities are different could lead to more finely tu ned interventions to enhance trust. While there have been studies examining construct of trust in the African American community, a qualitative approach or perhaps CBPR methodology could uncover new information particularly between genders and age groups. Healthcare provider education and public awareness campaigns would address the three of the significant variables (awareness, knowledge, and provider recommendation) found in this dissertation; however, the issue of trust in healthcare provider as a deter minant of PPSV23 uptake would require a much more complex myriad of interventions. More research would likely be needed to better understand the construct of trust by examin ing whether the issue is with trust in the healthcare provider, pharmaceutical comp anies producing the vaccine, or the healthcare system at large. In aforementioned studies, racial concordance appeared to enhance trust and, in turn, compliance with prescribed treatment regimens and preventative health interventions. Studies examining ra cial concordance and vaccine uptake were not evident in the literature; therefore it is recommended that future research be conducted to e xamine the relationship between racially matched healthcare providers and PPSV23 vaccine uptake
157 Within the VUQ, soci oeconomic status was measured using two proxy variables : education, and insurance type. Though as mentioned previously, education has often been utilized as a measure of SES, neither variable performed as expected in this dissertation study. Exploration of variables which might be better suited to measure SES could prove beneficial moving forward. Use of straightforward SES measures such as income may not be readily accessible given the cultural propensity of African Americans to keep such matters private; therefore, it is recommended that discrete measures such as possession of certain durable goods, or perhaps a more detailed question pertaining to education status be explored among African Americans. A study comparing known SES status of a group of Africa n American adults with a series of self reported proxy variables thought to accurately reflect SES could help inform existing methods of SES assessment moving forward. Despite high levels of perceived severity for pneumonia, many remain ed unvaccinated. The construct of fatalism was reviewed briefly in the literature, and to our knowledge no studies focusing on fatalism and vaccine uptake were identified A closer l ook at perc eived severity and how this variable might interact with the construct of fatalism and affect vaccine uptake among African American adults is therefore recommended. If individuals believe illnesses are out of their control, preventative behaviors such as vaccination may not be considered. A mixed methods study designed to measure the deg ree to which fatalism influences preventative health seeking behaviors such as vaccination could highlight important cultural barriers and better prepare clinicians for discussion and communication strategies to overcome these barriers.
158 Florida is one of t he few states where Medicaid does not cover the PPSV23 vaccine. Costs associated with a single case of pneumonia, could be measured through examination of Medicaid data. These costs would likely vary from fairly simple cases of pneumonia, to very complex cases where the patient ends up on the ventilator in the intensive care unit for a number of days. Providing policy makers with factual data on the costs associated with caring for patients with vaccine preventable pneumococcal pneumonia could lead to impo rtant changes in vaccine coverage for Medicaid recipients and will therefore be a high priority next step in this program of research. Finally, predictors of vaccine refusal should be examined more carefully. Using the HBM, the author recommends predictor variables; susceptibility, severity, and perceived health be examined more closely with instruments that may be more sensitive than those included on the VUQ and with a larger sample size Although this dissertation examined vaccine refusals, it was not a major focus, and more research is needed to determine how HBM constructs differentially predict PPSV23 refusals. Perhaps a focus group approach with vaccine refusers would provide more insight for targeted interventions in the African American community. Implications for Policy Change Findings from this dissertation may add to the body of existing knowledge on PPSV23 uptake in the African American community and strengthen the argument for policy change pertaining to vaccination guidelines, recruiting Afric an Americans into the health care professions, public service messag ing, electronic health record interfaces and cost, Each of these items will now be discussed briefly.
159 With multiple studies emphasizing confusion surrounding eligibility and vaccination criteria, the first recommenda tion of for policy change is the simplification o f PPSV23 uptake recommendation guidelines for ease of understanding and compliance among healthcare practitioners whose recommendation for vaccine is vital. PPSV23 vaccine guid elines in other developed nations such as Canada differ slightly in that everyone over 50 is vaccine eligible Before undertaking what could be a very costly endeavor, efforts should be made to compare incidence and prevalence of pneumococcal illnesses in countries where guidelines differ. If there is a significant difference in pneumococcal illnesses, we might consider revising guidelines to vaccinate all adults over 50 rather than 65 as well. As previously mentioned, African Americans remain underrepresen ted in the healthcare professions. It is therefore recommended that policies be examined with respect to r ecruit ment of qualified African Americans into healthcare professions in order to increase the overall percentage of African American physicians, nurs es, and extenders thus increasing the likelihood of racial concordance between patient and provider Cultural sensitivity among healthcare providers has been emphasized in practice for a number of years; howe ver, there is more to be done. Inevitably, pro viders will sometimes be different from cultural and ethnic perspectives than the patients they serve. While cultural competence is a life long process, certainly strides can be made to enhance communication across cultures. It is therefore recommended tha t policies mandating initial and ongoing training on cultural sensitivity be implemented across the healthcare disciplines.
160 P ublic health endeavors for influenza vaccination which are already largely in place could also be utilized as a spring board to en hance PPSV23 uptake. Strengthening individual and community knowledge in communities of color is critical to reduce the nearly 33% deficit in vaccine uptake. Although a public policy and messaging would likely have some impact, social networks, faith based interventions, and grassroots efforts may prove more successful given the African American cultural propensity towards word of mouth information. It is therefore recommended that public health officials u tilize existing public servi ce announc ements and efforts designed to enhance influenza vaccine uptake as a platform to increase PPSV23 awareness and knowledge Across the country, primary, secondary, and tertiary care facilities have undertaken efforts to implement electronic health records. T hough this is an important first step, there are a wide variety of EHR systems which often lack the interface capability necessary for seamless transfer of information. While programs exist to facilitate data exchange between hosts, they can be costly and have not yet been widely adopted. In order to enhance record keeping, cut costs, and avoid duplication of services such as vaccination, it is therefore recommended that policies facilitating and mandating interfaces between systems to be explored and imple mented. Finally, a lthough not a major focus of this study, the cost of PPSV23 vaccination and how it impacts vaccine uptake should also be examined more carefully. Lau, et al. (2009) investigated H1N1 vaccine acceptance among age 18 60 in 2009 pandemic and found that 45% were likely to receive a free H1N1 vaccination, but this intention
161 decreased as the hypothetical cost of vaccination increased, thus hypothesizing that cost was a potential barrier. Conclusion A look back at statistics on vaccination reveal that i n 2009, only 68.5% of patients surveyed reported had received PPSV23 in their lifetime (BRFSS, 2010). Moreover, according to Healthy People 2020, African Americans are roughly 20 30% less likely to receive the PPSV23 vaccine than their Caucasian cou nterparts. Given these disparities in vaccination uptake, this study sought to better understand and describe the decisions to accept or decline PPSV23. Descriptive statist ics in the pneumococcal study were strikingly similar to the Healthy People report with 119 (40.34%) vaccinated and 176 (59.66%) reportedly unvaccinated despite eligibility highlighting a persistent problem and cause for concern Several conclusions were reached as a result of this dissertation work including the significant relationship s identified between PPSV23 uptake and demographics such as age and gender, as well as the significant relationship s between PPSV23 uptake and the variables; awareness, kno wledge, trust, perceived susceptibility, and provider recommendation thus supporting the substructed theoretical framework and strengthening the overall model D emographics intertwine d with beliefs created a logistic regression model that predicted PPSV23 uptake 83 .9 % of the time. It is clear that the associations between demographics, health beliefs, and vaccine uptake are complex, thus creating a major challenge for researchers and healthcare pro fessionals is to disentangle decision maki ng process regarding immunization into a
162 series of actionable items/interventions leading to increased PPSV23 uptake and decreased health disparities for pneumococcal illnesses among the African American population. Consistent with prior research on preve ntative behaviors such as vaccine uptake, three dimensions of the HBM (perceived barriers, cues to action, and perceived susceptibility) were significant pred ictors of PPSV23 uptake. With 147 (47.8%) of respondents reporting they were unaware of the existe nce of PPSV23 prior to this overall model and may reflect and important finding in the endeavor to increase PPSV23 uptake and narrow the uptake disparity between African America ns and their Caucasian counterparts. Despite the aforementioned limitations from this descriptive research, this study was the first of its kind to examine age, gender, SES, awareness, knowledge, trust, perceived health, perceived susceptibility, perceive d severity, prior negative experiences with vaccines, and healthcare provider recommendation for the PPSV23 vaccine among eligible African American adults of all ages The sample size of 295 revealed a number of statistically significant predictors for vac cine uptake and may indeed serve as a springboard for future studies where the research can focus specifically on variables with the greatest influence and predictability for this preventative health behavior
163 APPENDIX A A DMINISTRATION GUIDEL INES
164 APPENDIX B P RESCREENING TO OL You are invited to complete this brief screening paper to find out if you are eligible to participate in a research study being conducted by Carla Fry, MSN, RN. Carla is a PhD student at the University of Florida. She has been a nurse for 23 years and her research focuses on keeping people healthy. She wants to learn more about vaccine beliefs and practices among African American Adults. As soon as you complete this screening paper, you will know if you are eligible, and you will receive information on when and where you can fill out the main survey. If you are eligible, you will be given a survey that will take 15 minutes to complete. As a way of saying thanks, you will be given a $10.00 gift card as soon as you turn in the survey. The survey is the only thing Carla will be asking you to complete. All of the information you list is Completing this one page screening tool does not mean you have to participate. You can change your mind a t any time. If you have any questions, please contact Carla at the number or e mail listed below. Thank you very much! Carla Fry c email@example.com 904 485 0718 Are you 65 or older? YES NO Do you curre ntly smoke cigarettes? YES NO Do you have heart disease? YES NO Do you have lung disease? YES NO Are you diabetic? YES NO Do you have liver disease? YES NO Do you have sickle cell disease? YES NO Do you have any type of cancer? YES NO Do you have any type of immune disorder? YES NO Do you have kidney disease? YES NO Has your spleen been removed? YES NO Are you currently on dialysis? YES NO Have you had an organ transplant? YES NO Do you have cochlear implants? YES NO Are you currently taking radiation or chemotherapy? YES NO African American, you are eligible to complete the qu estionnaire about vaccine beliefs and get the $10.00 gift card
165 APPENDIX C V ACCINE UPTAKE QUESTI ONNAIRE
17 5 APPENDIX D V ACCINE UPTAKE DECISI ON TREE Have you ever received the pneumonia vaccine? Unsure Are you willing to be vaccinated today? No Yes Yes No Unsure Refer for Vaccine Stop Were you over 65? Yes No Are you over 65 now? Yes No Do you need additional information? Yes Provide Education No
176 APPENDIX E WAKE FOREST TRUST IN STRUMENT (FULL INSTR UMENT) Wake Forest University Trust Scales Mark A. Hall, J.D. Wak e Forest University School of Medicine Department of Public Health Sciences Winston Salem, NC 27157 1063 (336) 716 9807 9213 firstname.lastname@example.org email@example.com Elizabeth Dugan, Ph.D. University of Massachusetts Medical School Division of Geriatric Med icine Worcester MA 01605 (508) 856 3493 Elizabeth.Dugan@umassmed.edu Rajesh Balkrishnan, Ph.D. Merrell Dow Professor Ohio State University, College of Pharmacy 500 West 12th Avenue Columbus, Ohio 43210 (614) 292 6415 firstname.lastname@example.org See ge nerally: Hall M.A. Researching Medical Trust in the United States. J. Health Organization & Mgt. 20(5): 456 67; 2006. Hall, M. A., E. Dugan, B. Zheng, and A. Mishra. Trust in Physicians and Medical Institutions: What Is It, Can It Be Measured, and Does It Matter? Milbank Q. 79(4): 613 639; 2001. Balkrishnan, R., E. Dugan, F. Camacho, and M.A. Hall. Trust and Satisfaction with Physicians, Insurers, and the Medical Profession. Medical Care 41(9):1058 1064; 2003.
177 Interpersonal Trust in a Physician 10 ite ms scored 5 1 for Strongly Agree, Agree, Neutral, Disagree, and Strongly Disagree. Negatively worded items (2, 3, and 8 ) are reverse scored. In a national sample of 959 adults with established primary care relationships (including non physicians), alpha = .93, mea n = 40.8 (77.0 on a scale of 100), SD = 6.2 (15.5) 1. [Your doctor] will do whatever it takes to get you all the care you need. 2. Sometimes [your doctor] cares more about what is convenient for [him/her] than about your medical needs. 3. [Yo ur d octor] 's medical skills are not as good as they should be. 4. [Your doctor] is extremely thorough and careful. 5. You completely trust [your doctor's] decisions about which medical treatments are best for you. 6. [Your doctor] is totally honest in telling you about all of the different treatment options available for your condition. 7. [Your doctor] only thinks about what is best for you. 8. Sometimes [your doctor] does not pay full attention to what you are trying to tell [him/her] 9. You ha ve no worries about putting your life in [your doctor]'s hands. 10. All in all, you have complete trust in [your doctor]. Hall, M. A., B. Zheng, E. Dugan, K. E. Kidd, A. Mishra, R. Balkrishnan, and F. Camacho. 2001. Measuring Patients' Trust in Their Primary Care Providers. Medical Care Research & Review, 2002 Sep;59(3):293 318.
178 Interpersonal Trust in a Physician -Short Form 5 items scored 5 1 for Strongly Agree, Agree, Neutral, Disagree, and Strongly Disagree. Item 1 is reverse scored. In a na tional sample of 1045 adults with established primary care relationships (including non physicians), alpha = .87, mea n = 20.4 (77.0 on a scale of 100), SD = 3.1 (15.5). 1. Sometimes [your doctor] cares more about what is convenient for [him/her] than abo ut your medical needs. 2. [Your doctor] is extremely thorough and careful. 3. You completely trust [your doctor's] decisions about which medical treatments are best for you. 4. [Your doctor] is totally honest in telling you about all of the different treatment options available for your condition. 5. All in all, you have complete trust in [your doctor]. Dugan, E., M. Hall, F. Trachtenberg. Development of Abbreviated Measures to Assess Patient Trust in a Physician, a Health Insurer, and the Medic al Profession, BMC Health Services Research 2005, 5:64, http://www.biomedcentral.com/1472 6963/5/64
179 Trust in a Health Insurer 11 items scored 5 1 for Strongly Agree, Agree, Neutral, Disagree, a nd Strongly Disagree. Negatively worded items (2, 6, 7, and 9 ) are reverse scored. In a national sample of 410 adults with public or private health insurance, alpha = .92, range = 11 55, mea n = 36.5 (58.0 on scale of 100), SD = 7.8 (17.7). 1. You think t he people at XXX are completely honest 2. XXX cares more about saving money than about getting you the treatment you need 3. As far as you know, the people at XXX are very good at what they do 4. XXX would pay for you to see any specialist you might ne ed 5. If you asked XXX about what treatments your insurance covers, you think XXX would be totally honest with you. 6. If someone at XXX made a serious mistake, you think they would try to hide it 7. You worry there are a lot of loopholes in what XXX co vers that you know about 8. You believe XXX will pay for everything it is supposed to, even really expensive treatments. 9. If you got really sick, you are afraid XXX might try to stop covering you altogether 10. If you have a question, you think X XX will give a straight answer 11. All in all, you have complete trust in XXX. Zheng B, Hall MA, Dugan E, Kidd KE, Levine D. Development of a scale to measure patients' trust in health insurers. Health Services Research. 2002;37(1):187 202.
180 Trust in a Health Insurer Short Form 5 items scored 5 1 for Strongly Agree, Agree, Neutral, Disagree, and Strongly Disagree. Items 1 and 2 are reverse scored. In a national sample of 410 adults with public or private health insurance, alpha = .84, mean = 16.6 (58.0 on scale of 100), SD = 3.9 (19.5). 1. XXX cares more about saving money than about getting you the treatment you need 2. You feel like you need to double check everything XXX does. 3. You believe XXX will pay for everything it is supposed to, eve n really expensive treatments. 4. If you have a question, you think XXX will give a straight answer 5. All in all, you have complete trust in XXX. Dugan, E., M. Hall, F. Trachtenberg. Development of Abbreviated Measures to Assess Patient Trust in a Physician, a Health Insurer, and the Medical Profession, BMC Health Services Research 2005, 5:64, http://www.biomedcentral.com/1472 6963/5/64
181 Trust in Doctors Generally 11 items scored 5 1 for Strongly Agree, Agree, Neutral, Disagree, and Strongly Disagree. Negatively worded items (2 7 ) are reverse scored. In a national sample of 502 adults, alpha = .89, range = 11 54, mean = 33.5 (51.1 on scale of 100), SD = 6.9 (15.7). 1. Doctors in gen eral care about their patients' health just as much or more as their patients do. 2. Sometimes doctors care more about what is convenient for them than about their medical needs. 3. Doctors are extremely thorough and careful. 4. You complet ely trust doctors' decisions about which medical treatments are best. 5. Doctors are totally honest in telling their patients about all of the different treatment options available for their conditions. 6. Doctors think only about what is best for thei r patients. 7. Sometimes doctors do not pay full attention to what patients are trying to tell them 8. Doctors always use their very best skill and effort on behalf of their patients. 9. You have no worries about putting your life in the hands of doct ors. 10. A doctor would never mislead you about anything. 11. All in all, you trust doctors completely. Hall, MA, Fabian Camacho, Elizabeth Dugan & Rajesh Balkrishnan, Trust in the Medical Profession: Conceptual and Measurement Issues. Health Serv. Res. 2002; 37: 1436 39.
182 Trust in Doctors Generally Short Form 5 items scored 5 1 for Strongly Agree, Agree, Neutral, Disagree, and Strongly Disagree. Item 1 is reverse scored. In a national sample of 502 adults, alpha = .77, mean = 15.0 (50.0 on a scal e of 100), SD = 3.4 (6.8). 1. Sometimes doctors care more about what is convenient for them than about their medical needs. 2. Doctors are extremely thorough and careful. 3. You completely trust decisions about which medical treatm ents are best. 4. A doctor would never mislead you about anything. 5. All in all, you trust doctors completely. Dugan, E., M. Hall, F. Trachtenberg. Development of Abbreviated Measures to Assess Patient Trust in a Physician, a Health Insurer, and the Medical Profession, BMC Health Services Research 2005, 5:64, http://www.biomedcentral.com/1472 6963/5/64
183 Trust in Medical Researchers Generally Scored 5 1, strongly agree to strongly disag ree, with negative items (3, 5, 8, 11) reverse coded so that higher score indicates more trust. In pilot testing with 124 adults, overall scale mean was 36.25 (50.52 on a 0 to 100 scale), with a standard deviation of 7.83 (16.32) a lpha = .87. Factor model consists of one factor. 1. Doctors who do medical research care only about what is best for each patient. 2. Medical researchers have no selfish reasons for doing research studies. 3. There are some things about medical research that I do not trust a t all. 4. A doctor would never ask me to be in a medical research study if the doctor thought there was any chance it might harm me. 5. Medical researchers do not tell people everything they really need to know about being in a research study. 6. The only reason doctors do medical research is to help people. 7. safe to be in a medical research study. 8. Some doctors do medical research for selfish reasons. 9. A doctor would never recommend something that is not the best treatment, just so he or she can study how it works. 10. Doctors tell their patients everything they need to know about being in a research study. 11. Medical researchers treat people like 12. I completely trust doctors who do medical research. Mark A. Ha ll, Fabian Camacho, Janice S. Lawlor, Venita DePuy, Jeremy Sugarman, Kevin Weinfurt. Measuring Trust in Medical Researchers. Medical Care. 44(11):1048 53; 2006.
184 Trust in Medical Researchers Generally Short Form Scored 5 1, strongly agree to strongly d isagree, with negative item (3) reverse coded so that higher score indicates more trust. In a U.S. national survey of 3623 adults, overall scale mean was 12.4 (52.5 on a 0 to 100 scale), with a standard deviation of 2.7 (16.8) a lpha = .72. Factor model co nsists of one factor. 1. Doctors who do medical research care only about what is best for each patient. 2. Doctors tell their patients everything they need to know about being in a research study. 3. Medical researchers treat people like pigs. 4. I completely trust doctors who do medical research. Mark A. Hall, Fabian Camacho, Janice S. Lawlor, Venita DePuy, Jeremy Sugarman, Kevin Weinfurt. Measuring Trust in Medical Researchers. Medical Care. 44(11):1048 53; 2006.
185 APPENDIX F P ERMISSION TO US E THE WAKE FOREST TR UST SCALE
SF 12v2 Health Survey 1994, 2002 Medical Outcomes Trust and QualityMetric Incorporated. All rights reserved. SF 12 is a registered trademark of Medical Outcomes Trust. (SF 12v2 Health Survey Standard, United States (English)) 186 APPENDIX G SF QUESTIONNAIRE Your Health and Well Being This survey asks for your views about your health. This information will help keep track of how you feel and how well you are able to do your usual activities. Thank y ou for completing this survey! For each of the following questions, please mark an in the one box that best describes your answer. 1. In general, would you say your health is: Excellent Very good Good Fair Poor 1 2 3 4 5 2. The following questions are about activities you might do during a typical day. Does your health now limit you in these activities? If so, how much? Yes, limited a lot Yes, limited a little No, not limited at all a Moderate activities such as moving a table, pushing a vacuum cleaner, bowling, or playing golf ............................ 1 ............. 2 ............. 3 b Climbing several flights of stairs ................................ ............. 1 ............. 2 ............. 3
SF 12v2 Health Survey 1994, 2002 Medical Outcomes Trust and QualityMetric Incorporated. All rights reserved. SF 12 is a registered trademark of Medical Outcomes Trust. (SF 12v2 Health Survey Standard, United States (English)) 187 3. During the past 4 weeks how much of the time have you had any of the following problems with your work or other regular daily activities as a result of your physical health ? All of the time Most of the time Some of the time A little of the time None of the time a Accomplished less than you would like ................................ ...... 1 .............. 2 .............. 3 ............... 4 ............. 5 b Were limited in the kind of work or other activities .................. 1 .............. 2 .............. 3 ............... 4 ............. 5 4. During the past 4 weeks, how much of the time have you had any of the following problems with your work or other regular daily activities as a result of any emotional problems (such as feeling depressed or anxious)? All of the time Most of the time Some of the time A little of the time None of the time a Accomplished less than you would like ................................ ..... 1 ............. 2 ............. 3 .............. 4 ............ 5 b Did work or other activities less carefully than usual ................ 1 ............. 2 ............. 3 .............. 4 ............ 5 5. During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)? Not at all A little bit Moderately Quite a bit Extremely 1 2 3 4 5
SF 12v2 Health Survey 1994, 2002 Medical Outcomes Trust and QualityMetric Incorporated. All rights reserved. SF 12 is a registered trademark of Medical Outcomes Trust. (SF 12v2 Health Survey Standard, United States (English)) 188 6. These questions are about how you feel and how things have been with you during the past 4 weeks For each question, please give the one answ er that comes closest to the way you have been feeling. How much of the time during the past 4 weeks All of the time Most of the time Some of the time A little of the time None of the time a Have you felt calm and peaceful? ................................ ....... 1 .............. 2 ............. 3 .............. 4 ............. 5 b Did you have a lot of energy? ....... 1 .............. 2 ............. 3 .............. 4 ............. 5 c Have you felt downhearted and d epressed? .............................. 1 .............. 2 ............. 3 .............. 4 ............. 5 7. During the past 4 weeks, how much of the time has your physical health or emotional problems interfered with your social activities (like visit ing with friends, relatives, etc.)? All of the time Most of the time Some of the time A little of the time None of the time 1 2 3 4 5 Thank you for completing these questions!
189 APPENDIX H SF LICENSE AGREEMENT
194 APPENDIX I I RB PILOT STUDY
195 APPENDIX J I RB FULL STUDY
196 APPENDIX K E LEMENTS OF CONSENT PILOT STUDY
198 APPENDIX L E LEMENTS OF CONSENT FULL STUDY
200 APPENDIX M I NVITATION TO PARTICI PATE PILOT STUDY Dear Mrs. X, I am a nurs e and a doctoral student at the University of Florida conducting a study beliefs about the pneumococcal vaccine (sometimes known as the pneumonia vaccine) To learn about vaccine beliefs, I will be asking volunteers to complete a survey that will take about 15 minutes. They will not put their name on the survey and all of the information will be private. When they finish, they will be given a $10.00 gift card for their time. Before I start the study, I need to get 1 2 members of the congregation to take the survey and give me their opinion to make sure it is easy to understand and nothing is missing. You have been selected to participate because you are a key me mber of the church community and I value your opinion. M y research assistant, Sylvia Roberts, and I will be holding a meeting right after church services on Sunday, Month Date, at 12:30 pm. During the meeting, I will explain the study, ask you sign a consent form, and ask you to fill out the survey. After you c questions about it to see what you think might need to be changed before it is given to others.
201 We will p rovide lunch and give you a $20 .00 gift car d for your participation. If you are able to attend, please call me (Carla Fry) at 904 485 0718 by Month Date. Thank you for considering attending the meeting. Your input will help me learn more about vaccine beliefs among African Americans and my goal is to help provide services to the community based on the survey results. After the study is over if you woul d like a summary of the results, I will be happy to provide them to you. Sincerely, Carla A. Fry MSN, RN Principal Investigator PhD Student, Unive rsity of Florida School of Nursing
202 APPENDIX N P ILOT STUDY QUESTIONS AND SUMMARY RESPONSE S Questions for participants Summary Responses quotes from participants What do you think about the appearance of the survey? Is there anything you would change? The survey looks good. Can you make it in color? Was the font size (writing) large enough to read easily? Yes, the font size is fine. Was did you think about the length of the survey and the time it took to complete ? It was long. Is there something you can take out? Did any of the questions make you feel uncomfortable? None of the questions made [us] uncomfortable. Did you have trouble understanding any of the questions? If so, which ones? See above. In text. Were any of the questions too personal? If so, wh ich ones? None of the questions were too personal as long as the survey is anonymous. Were there any questions that you felt you needed to explain your answer? If so, which ones, and what type of information would you have written? Only the ones where you already have blanks need to be explained. Knowing the purpose of the study, is there anything else you would ask? If so, what, and why? No. There is nothing else to ask. Do you think members of the congregation would be willing to complete the survey? Why or why not? Yes. Members of the congregation will be willing especially with the gift card. There are a couple of questions whereby you have the choice of answering yes, no, or unsure. One example is when you are asked, you ever received the p neumonia Would you keep the check box or just stick with yes and no? In other words, do you think everyone will know for sure if been vaccinated? Keep it [unsure]. Not everyone will remember if they have had the vaccine. A $10. 00 gift card will be provided to participants who complete the survey. Do you think this is a fair amount of compensation for their time? Given the choice, which type of gift card (gas, food, or cash) would you prefer? a fair amount. A card from Winn Dixie would be good.
203 Questions for participants Summary Responses quotes from participants If we were to ask you questions about your income (individual or household), would you answer these questions? Why or why not? No. This is too personal. When the study is complete, the PI and RA will offer a health screening even t as a way to thank the congregation for their participation. What types of services would you most like to see (ie. Blood pressure checks, blood sugar checks, influenza and pneumococcal vaccine offerings, pamphlets on stroke prevention, etc.). Blood pres sure and blood sugar checks would be good. In addition to flyers, and word of mouth, can you think of any other good recruitment strategies? In addition to the two churches, are there other areas where you would recruit participants? If so, where? Word of mouth will be enough if the pastor announces it. What would you like the researcher to know about your community? What makes your congregation stand out, feel proud, etc.? We have a lot of people been coming to this church for years. We are a very tight community. What role does your faith play in health related matters? Do you think we should ask any questions about this in the survey? We do pray to God about our illnesses and that it be his will to keep us healthy. Do you think having a researcher who is not African American will hinder the study or make it difficult to get participants? No, not as long as she is respectful and kind. Does having a research assistant who is African American make you more inclined to participate? If so, why or why not ? It will be nice to have her [RA] because she knows us and knows the church. What do you think the major health problems are in your community? Diabetes and high blood pressure. When is the best time to reach the most people with the survey? Right after church services and bible study on Wednesdays and Sundays. Should we schedule appointments for survey completion? No. People keep their appointments. Just come after church and get the whole group.
204 Questions for participants Summary Responses quotes from participants A few of the questions have a place for open ended responses. Would you retain these in the survey or do you think they might be confusing? Keep them so we can explain our answers. Do you think people will be worried about their privacy? If so, what can we do to make sure they fee l comfortable? No, not as long as we sign the surveys. What comes to mind when we mention the term Do you have any concerns about cultural differences between the researcher and the participants? If so, what? about und erstanding how different and how the same. Is there anything else you would add to the survey? No. Do you have any other ideas or suggestions for the study? Just to tell us about the results.
205 APPENDIX O P ILOT STUDY GUIDE Welcome and Introduction: PI and RA to introduce and speak briefly on background. Background: The pneumococcal vaccine (sometimes known as the pneumonia vaccine) is recommended for all adults 65 and older AND people under 65 with certain chronic illnesses, like diabetes and heart disease. Many African Americans are eligible for the vaccine, but do not get it for reasons not sure of. Unfortunately, more African Americans become sick and die from pneumonia every year than their Caucasian counterparts. Purpose : The nurse rese archer wants to find out more about whether or not African Americans know about the vaccine and how they decide whether or not to accept it. By learning more about the reasons African Americans choose to be vaccinated, the researcher hopes to reach out t o those who have not yet been vaccinated and at risk for getting pneumonia through education and trusting relationships between the healthcare provider and the client. Procedure : The study will begin on Month/Date/Year and end on Month/Date/Year. During that time, African American adults over the age of 18 can complete the study: Sundays immediately following church service from 12:30 to 2:00 pm Wednesday mornings after Bible study Individuals will be asked to complete the questionnaire which will ta ke approximately 15 minutes. If they agree to participate, they will sign the consent form and receive the instructions and questionnaire. When they have finished the questionnaire, they will place it in a sealed envelope and then place it in a locked bo x in the community room where the researchers will then provide a $10.00 gift card to the participant thanking him or her for their time. Meeting Objectives: meeting with a small group of people today to explain the study, review the consent proces s, have everyone complete the survey, and then get your feedback on the questions and the process. I believe it is important to make sure asking you questions that are respectful, easy to read, easy to understand, and not too time consuming. I also wan t to make sure I get good quality data so the study is meaningful. RA provides copy of meeting objectives to all present. I want to encourage you to speak your mind, tell me if you think there are any issues with the consent process or the questionnaire so I can fix them before the main study
206 begins. You know much more about the members of the congregation than I ever will, so your opinion is very valuable. While having your lunch, pass out the consent form to have you read and sign it. Th e consent form will not be turned in and your name will never be associated with your answers. The same will be true in the main study. If, after reading the consent, you decide not to do the questionnaire, not a problem at all. You can enjoy your lunch and stay as long as you like. RA hands out consent form _________PURPOSE COMPLETED __________ All consent forms distributed and read ( instruct participants to keep ). Provide all participants with a copy ( not signed or collected ). After everyo ne finishes the questionnaire, we will have a brief discussion about the questions, recruiting people to take the survey, the type of gift card most preferred, and whether or not you think members of the congregation would like a health fair after the stud y is complete. If you have any questions at all while completing the questionnaire, please hesitate to raise your hand or come up and Sylvia or I will help you. RA hands out questionnaire All questions are fielded _________CONSENT COMPLETED __________ RA times completion of questionnaire and documents results PI and RA to assist, answer questions. _________QUESTIONNAIRE COMPLETED __________ PI perfor m s respondent interview RA records conversation digitally if permission granted and on pap er RESPONDENT INTERVIEW COMPLETED THANK PARTICIPANTS AND PROVIDE GIFT CARDS
207 L IST OF REFERENCES Albanese B., Roche, J., Pass, M., Whitney, C., McEllistrem, M., & Harrison, L. (2002). Geographic, demographic, and seasonal differences in penicillin resis tant streptococcus pneumoniae in Baltimore Clinical Infectious Diseases 34, 15 21. Alegra, M. (2009). Training for research in mental health and HIV/AIDS among racial and ethnic minority populations: meeting the needs of new investigators. American Jo urnal of Public Health 99S, 26 30. Allen, C. (2010). CDC says immunizations reduce deaths from influenza and pneumococcal disease among older adults. National Association of Chronic Diseases. Atlanta: GA. Allen, J., Kennedy, M., Wilson Glover, A., & Gil lig an, T. (2007). African American men's perceptions about prostate cancer: Implications for designing educational interventions. Social Science & Medicine 64(11), 2189 2200. Altice, F., Mostashari, F., & Friedland, G. (2001). Trust and the acceptance of and adherence to antiretroviral therapy. Journal of Acquired Immune Deficiency Syndromes 28(1), 47 58. Anderson, L. & Dedrick, R., (1990). Development of the Trust in Physician scale: a measure to assess interpersonal trust in patient physician relationsh ips. Psychological R eports, 67(3), 1091 100 Armstrong, K., Ravenell, K., McMurphy, S., & Putt, M. (2007). Racial/Ethnic differences in physician distrust in the United States. American Journal of Public Health, 97(7), 1283 1289. Ashby Hughes, B., & Nicker so n, N. (1999). Provider endorsement: the strongest cue in prompting high risk adults to receive influenza and pneumococcal immunizations. Clinical Excellence for Nurse Practitioners 3(2), 97 104. Baldwin, D. (1996). A model for describing low income Afr ican American women's participation in breast and cervical cancer early detection and screening. Advances in Nursing Science 19(2), 27 42. Bass, S., Gordon, T., Ruzek, S., Wolak, C., Ward, S., Paranjape, A., et al. (2011). Perceptions of colorectal cance r screening in urban African American clinic patients: differences by gender and screening status. Journal of Cancer Education 26(1), 121 128. Benkert, R., Hollie, B., Nordstrom, C., Wickson, B., & Bins Emerick, L. (2009). Trust, mistrust, racial identit y and patient satisfaction in urban African American primary care patients of nurse practitioners. Journal of Nursing Scholarship 41(2), 211 219
208 Benyamini Y, & Idler E. (1999). Community studies reporting association between self rated health and mortal ity additional studies, 1995 1998. Res earch on Aging 21 392 401. Bergkvist, L., & Rossiter, J. R. (2007). The predictive validity of multiple item versus single item measures of the same constructs. Journal of Marketing Research 44, 175 184. Berkley Pat ton, J., Bowe Thompson, C., Bradley Ewing, A., Hawes, S., Moore, E., Williams, E., Martinez, D., & Goggin K. (2010). Taking it to the pews: A CBPR guided HIV awareness and screening project with Black churches. AIDS Education and Prevention 22(3), 218 23 7. Betancourt, H., Flynn, P. M., Riggs, M., & Garberoglio, C. (2010). A cultural research approach to instrument development: the case of breast and cervical cancer screening among Latino and Anglo women. Health Education Research 25(6), 991 1007. Blum enthal D., Sung J., Coates R., Williams J., & Liff J. (1995). Recruitment and retention of subjects for a longitudinal cancer prevention study in an inner city black community. Health Service Research 30 197 205. Butler J., & Schuchat A. (1999). Epidemio logy of pneumococcal infections in the elderly. Drugs and Aging 15(1), 11 19. Bosworth H. (1999). The relationships between self rated health and health status among coronary artery. J ournal Aging Health 11 565 5 84. Boulware L Cooper L Ratner L L aVeist T & Powe N. (2003). and t rust in the h ealth c are s Public Health Reports 118(4) 358 3 65. Bova, C., Fennie, K., Watrous, E., Dieckhaus, K., & Williams, A., (2006). The health care relationship (HCR) trust scale: Development and psych ometric evaluation. Research in Nursing & Health 29, 477 488. Brewer, N., Chapman, G., Gibbons, F., Gerrard, M., McCaul, K., & Weinstein, N. (2007). Meta analysis of the relationship between risk perception and health behavior: the example of vaccination. Health Psychology 26(2), 136 145. Breiman R Keller D & P helan M. (Year). Evaluation of effectiveness of the 23 valent pneumococcal capsular p olysacch aride vaccine for HIV infected p atients. Arch ives Intern al Med icine 160(17) 2633 2638. Bryant, A., Nakagawa, S., Gregorich, S., & Kuppermann, M. (2010). Race/ethnicity and pregnancy decision making: the role of fatalism and subjective social standing. Journal Of Women's Health 19 (6), 1195 1200.
209 Butler, J., & Schuchat, A. (1999). Epidemiology of p neumo coccal i nfections in the e lderly. Drugs & Aging 15(1), 11 19. Byrd, G., Edwards, C., Kelkar, V., Phillips, R., Byrd, J., Pim Pong, D., et al. (2011). Recruiting intergenerational African American males for biomedical research studies: A major research cha llenge. Journal of the National Medical Association 103, 480 487. Cantrell, 2009. Infection control update. Pneumonia vaccine may help limit H1N1 flu deaths. Healthcare Purchasing News 33 (9), 26. Centers for Disease Control and Prevention (CDC) (1998) Active bacterial core surveillance (ABCs) report, Emerging Infections Program Network, Streptococcus pneumoniae, 1998. Atlanta, G eorgia : Emerging Infections Program Network Centers for Disease Control and Prevention (CDC) (2000). MMWR W ee kly Rep ort: Pr evention and control of influenza: recommendations of the Advisory Committee on Immunization Practices (ACIP) 49:1 38. Centers for Disease Control and Prevention (CDC) (2003). Public health and aging report: Influenza vaccination coverage among adults age d >50 years and p neumococcal vaccination coverage among adults aged >65 years. Retrieved Feb 1, 2011 from: http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5241a3.htm Centers for Disease Control and Prevention (CDC) (2005). MMWR Weekly Report: Health disparit ies experienced by black or African Americans United States. 54(01), 1 3 Centers for Disease Control and Prevention (CDC) (2006). MMWR Weekly Report: Using standing orders to improve adult vaccination rates. Retrieved Feb 1, 2011 from: http://www.cdc.gov /MMWR/preview/mmwrhtml/rr4901a2.htm Centers for Disease Control and Prevention (CDC). (2007). Epidemiology and prevention of vaccine preventable diseases. Retrieved January 10, 20 10 from: http://www.vaccineinformation.org Centers for Disease Control an d Prevention (CDC). (2007). Vaccine information statement: pneumococcal polysaccharide vaccine: what you need to know. Rockville (MD): Retrieved January 2011 from: http://www.cdc.gov/vaccines/pubs/vis/downloads/vis ppv.pdf. Centers for Disease Control and Prevention (CDC) ( 2008 ) Active Bacterial Core (ABC) Surveillance Report, Emerging I nfections Program Network. Streptococcus pneumoniae, 2007. Retrieved June, 2010 from: http:www.cdc.gov/ncidod/dbmd/abcs/surveyreports/spneu07.pdf
210 Centers for Disease Cont rol and Prevention (CDC) (2008). MMWR Weekly Report: Prevention of pneumococcal disease: recommendations of the Advisory Committee on Immunization Practices (ACIP). 46 1 24. Centers for Disease Control and Prevention (CDC) (2010). MMWR Weekly Report: Up dated recommendations for prevention of invasive pneumococcal disease among adults using the 23 valent pneumococcal polysaccharide vaccine (PPSV23). 59(34), 1102 1106 Centers for Disease Control and Prevention (CDC). (2011). Behavioral Risk Factor Surveil lance System Survey Questionnaire Atlanta, Georgia: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention Centers for Medicare and Medicaid (CMS). (2009). question & answer guide to Medicare coverage of sea sonal influenza and pneumococcal vaccinations steps to promoting wellness adult immunizations. Chen, J., Fox, S., Cantrell, C., Stockdale, S., & Kagawa Singer, M. (2007). Health disparities and prevention: racial/ethnic barriers to flu vaccinations. Journ al of Community Health 32(1), 5 20. Chandola, T. Crispin, J. ( 2000 ) self rated health in different ethnic g Ethnicity & Health 5 151 15 9. Chen, J., Fox, S., Cantrell, C., Sto ckdale, S., & Kagawa Singer, M. (2007). Health disparitie s and prevention: racial/ethnic barriers to flu vaccinations. Journal of Community Health 32(1), 1 7. Chi, R., Jackson, L ., Neuzil, K. (2006). Characteristics and outcomes of older adults with community acquired Pneumococcal bacteremia. Journal of America n Geriatrics Society 54 115 120. Cooper L., Roter D., Johnson R., Ford D., Steinwachs D., & Powe N. (2003). centered communication, ratings of care, and concordance of patient and physician Annals of Internal Medicine 2 139(11) 907 15. Cooper & Powe (2004) Disparities in patient experiences, health care processes, and outcomes: The role of patient provider racial, ethnic, and language concordance. Commonwealth Fund Publication. Retrieved, June 2012 at: https://www.cmwf.org Corbie Smit h, G., Thomas, S., Williams, M., & Moody Ayers, S. (1999). Attitudes and beliefs of African Americans toward participation in medical research. Journal of General Internal Medicine 14 (9), 537 546. Cott CA, Gignac MAM, Badley EM (1999) Determinants of sel f rated health for Canadians with chronic disease and disability. J ournal of Epidemiol ogy and Community Health 53 731 73 6.
211 Cummings, J., & Jackson, P. (2008). Race, gender, and SES disparities in self assessed health, 1974 2004. Research on Aging 30(2), 137 168. Damian J, Ruigomez A, Pastor V, & Martin Moreno JM (1999) Determinants of self assessed health among Spanish older people living at home. J ournal of Epidemiol ogy and Community Health 53 412 41 6. Daniels, Juarbe, Rangel Lugo, Moreno John, & P rez Stable (2004). Focus group interviews on racial and ethnic attitudes regarding adult vaccinations. Journal of the National Medical Association, 96(11), 1455 1461. Daniels, N., Gouveia, S., Null, D., Gildengorin, G., & Winston, C. (2006). Acceptance of pneumococcal vaccine under standing orders by race and ethnicity. J ournal of the Nat iona l Med ical Assoc iation 98(7) 1089 1094. Dassow, P. (2005). Setting educational priorities for women's preventive health: measuring beliefs about screening across dise ase states. Journal of Women's Health 14(4), 324 330. Davis, M., Ndiaye, S., Freed, G., Kim, C., & Clark, S. (2003). Influence of insurance status and vaccine cost on physicians' administration of pneumococcal conjugate vaccine. Pediatrics 112(3), 521 5 26. Davis, T., Fredrickson, D., Bocchini, C., Arnold, C., Green, K., & Humiston, S (2002). Improving vaccine risk/benefit communication with an immunization education package: A pilot study. Ambulatory Pediatrics 2 (3), 193 200. de Courval, F., De Serre s, G., & Duval, B. (2003). Varicella vaccine: factors influencing uptake. Canadian Journal of Public Health 94(4), 268 271. Dein, S. (2006). Race, culture and ethnicity in minority research: a critical discussion. Journal of Cultural Diversity 13(2), 68 75. Denton M, & Walters V. (1999). Gender differences in structural and behavioral determinants of health: an analysis of the social production of health Soc ial Sci ence and Med icine 48 1221 12 35. DeSalvo, K., Fisher, W., Tran, K., Bloser, N., Merrill, W., & Peabody, J. (2006). Assessing measurement properties of two single item general health measures. Quality of Life Research 15(2), 191 201. Doescher, M., Saver B., Franks, P., & Fiscella, K. (2000). Racial and ethnic disparities in perceptions of ph ysician style and trust Arch ives of Fam ily Med icine 9(10) 1156 1163. Edberg, M. (2006). Essentials of health behavior: Social and behavioral theory in public heal th. Retrieved June, 2012 from: http://samples.jbpub.com/9780763737962/37968_FMxx_Edberg.pd f.
212 Egede, L., & Zheng, D. (2003). Racial/ethnic differences in adult vaccination among individuals with diabetes. American Journal of Public Health 93 (2), 324 329. Ehresmann, K., Ramesh, A., Como Sabetti, K., Peters on, D., Whitney C., & Moore, K. (2002). Factors associated with self reported pneumococcal immunization among adults 65 years of age or older in Minneapolis St. Paul metropolitan area. Preventative Medicine. 32(5), 409 415. Endrich, M., Blank, P., & Szucs, T. (2009). Influenza vaccination upta ke and socioeconomic determinants in 11 European countries. Vaccine 27, 4018 4024 Facione, N., Miaskowski, C, Dodd, M., & Paul, S. (2002). The self reported likelihood of patient delay in breast cancer: New thoughts for early detection. Preventive Medici ne, 34 (4), 397 407 Faul, F., Erdfelder, E., Buchner, A., & Lang, A (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41 1149 1160. Feikin, D., Schuchat, A., Kolczak, M., Barr ett, N., Harrison, L., Lefkowitz, L., McGeer, Al, Farley, M., Vugia, D., Lexau, C., Stefonek, K., RN, Patterson, J., Jorgensen, J. (1997). Mortality from invasive pneumococcal pneumonia in the era of antibiotic resistance. Public Health Reports Flannery, B., Heffernan, R., Harrison, L., Ray, S., Reingold, A., Hadler, J., et al. (2006). Changes in invasive pneumococcal disease among HIV infected adults living in the era of childhood pneumococcal immunization. Annals of Internal Medicine 144 (1), 1 9. Flan nery, B., Schrag, S., Bennett, N., Lynfield, R., Harrison, L., Reingold, A., et al. (2004). Impact of childhood vaccination on racial disparities in invasive Streptococcus pneumoniae infections. JAMA: Journal of the American Medical Association 291 (18), 2 197 2203. Flaskerud, J., & Nyamathi, A. (2000). Attaining gender and ethnic diversity in health intervention research: Cultural responsiveness versus resource provision. Advances in Nursing Science, 22(4), 11. Flowers, L. (2007). Racial and ethnic dispa rities in influenza and pneumococcal immunization rates among Medicare beneficiaries. AARP Public Policy Institute: Issue Brief. Retrieved: May, 2011 from: http://www.aarp.org/ppi Freimuth, V., Quinn, S., Thomas, S., Cole, G., Zook, E., & Duncan, T. (200 1). African Social Science and Medicine 52 (5), 797 808. Fry, C. (2008). Pneumococcal vaccine compliance rates in the acute care setting. ( Unpublished thesis ) Jacksonville University, Jacksonvi lle FL.
213 Gallant, D. (2007). Influenza vaccination: a personal decision. University of Calgary (Canada), AAT NR25730 1 191. Gallarce, E., Minsky, S., Viswanath, K. (2011). Socioeconomic status, demographics, beliefs and A(H1N1) vaccine uptake in the Unite d States. Vaccine doi:10.1016/j.vaccine.2011.05.014. Glanz, K., & Schwartz, M. (2008). Stress, coping, and health behavior. In K. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health Behavior and Health Education. Theory, research, and practice. (4th ed., pp 211 236). San Francisco: Wiley & Sons. Glover, S., Bellinger, J., B ae, S., Rivers, P., & Singh, K. (2010). Perceived health status and utilization of specialty care: racial and ethnic disparities in patients with chronic diseases. Health Education Journ al 69(1), 95 106. Grande, D., Shea, J., & Armstrong, K. (2011). Pharmaceutical industry gifts to physicians: patient beliefs and trust in physicians and the health care system. Journal of Gen eral Intern al Med icine, 27(3), 274 279. Hall, M. Dugan, B. Z heng, & Mishra A (2001). Trust in p hysicians and m edical i nstitutions: What i s i t, c an i t b e m easured, and d oes i t m atter? Milbank Q uarterly, 79(4) 613 639. Hall, M., Zheng, B., Dugan, E., Camacho, F., Kidd, K. Mishra, A., & Balkrishnan, R. (2002). Me asuring patients' trust in their primary care providers. Medical Care Research & Review 59(3), 293 318. Harawa, N., Greenland, S., Bingham, T., Johnson, D., Cochran, S., Cunningham, W., et al. (2004). Associations of race/ethnicity with HIV prevalence an d HIV related behaviors among young men who have sex with men in 7 urban centers in the United States. Journal of Acquired Immune Deficiency Syndromes 35(5), 526 536. Harris L, Chin N, Fiscella K, & Humiston S. (2006). Barriers to pneumococcal and influen za vaccinations in Black elderly communities: mistrust. J ournal of Nat iona l Med ical Assoc iation 98(10) 1678 84. Hausmann, L., Ibrahim, S., Mehrotra, A., Nsa, W., Bratzler, D., Mor, M., & Fine, M. (2009). Racial and ethnic disparities in pneumonia treatme nt and mortality. Medical Care (47) 9. Hicks, L., Harrison, L., Flannery, B., Hadler, J., Schaffner, W., Craig, A., et al. (2007). Incidence of pneumococcal disease due to non pneumococcal conjugate vaccine (PCV7) serotypes in the United States during the era of widespread PCV7 vaccination, 1998 2004. Journal of Infectious Diseases 196 (9), 1346 1354.
214 Hochbaum, G. (1958). Public participation in medical screening programs: a socio psychological study. Washington D.C. US Department of Health, Education, and Welfare. Hoffman, B. (2002). Respect for patients dignity in primary health care: a critical appraisal. Scandinavian Journal of Primary Health Care 20, 88 91. Holt, C., Roberts, C., Scarinci, I., Wiley, S., Eloubeidi, M., Crowther, M., & Coughlin, S. (20 09). Development of a spiritually based educational program to increase colorectal cancer screening among African American men and women. Health Communication 24(5), 400 412. Hooks, P., Tsong, Y., Baranowski, T.,Hen ske, J., Nader, P., & Levin, J. (1988). Recruitment strategies for multiethnic family and community health research. Family & Community Health 11(1). Hubble, M., Zontek, T., & Richards, M. (2011). Predictors of influenza vaccination among emergency medical services personnel. Prehospital Emer gency Care 15(2), 175 183. Immunization Action Coalition (2009). Ask the experts about pneumococcal polysaccharide vaccinati on. Retrieved June, 2010 from: http://www.immunize.org/search/?cx=017386553933306082099%3A5o7di6h q e&cof=FORID%3A10&ie=UTF 8&q=pn eumococcal Infection control update. (2009). Pneumonia vaccine may help limit H1N1 flu deaths. Healthcare Purchasing News 33 (9), 26. Retrieved from CINAHL database. http://www.cdc.gov/vaccines/vpd vac/flu/downloads/influ pneu disease 2008.pdf Jackson, L. Baxter, R., Naleway, A., Belongia, E., & Baggs, J. (2009). Patterns of pneumococcal vaccination and revaccination in elderly and non elderly adults: a vaccine safety datalink study. B iomedical Central Infectious Diseases. 9 ( 37 ) Janz, N., Wren, P., Schot tenfeld, D., & Guire, K. (2003). Colorectal cancer screening attitudes and behavior: a population based study. Preventive Medicine 37(6), 627 634. Johnson, R., Ro ter, D., Powe, R., & Cooper, L. (2004). Patient race/ethnicity and quality of patient physic ian communication during medical visits. American Journal of Public Health 94, 2084 2090. Jones, R., Steeves, R., & Williams, I. (2009). Strategies for recruiting African American men into prostate cancer screening studies. Nursing Research 58(6), 452 4 56. Kakinami, L., Newman, P., Lee, S., & Duan, N. (2008). Differences in HIV vaccine acceptability between genders. AIDS Care 20(5), 542 546.
215 Kamal, K., Madhavan, S., & Amonkar, M. (2003). Determinants of adult influenza and pneumonia immunization rate s. Journal of the American Pharmaceutical Association 43(3). Kegeles, S., Johnson, M., Strauss, R., Ralston, B., Hays, R., Metzger, D., et al. (2006). How should HIV vaccine efficacy trials be conducted? Diverse U.S. communities speak out. AIDS Education & Prevention 18 (6), 560 572. Keller, T. (2008). Mexican American perceptions of culturally congruent interpersonal processes of care during childhood immunization episodes a pilot study. Journal of R ural N ursing and H ealthcare ( 8 ) 2. Kennedy B., Mathis, C., & Woods, A. (2007). African Amer icans and their distrust of the health care system: healthcare for diverse populations. Journal of Cultural Diversity 14(2), 56 60. Kim, J., Bramlette, M., Wright, L., & Poon, L. (1998). Racial difference s in health status and health behaviors of older adults. Nursing Research 47(4), 195 257. Kimble, L., Dunbar, S., McGuire, D., De, A., Fazio, S., & Strickland, O. (2001). Cardiac instrument development in a low literacy population: the revised Chest Disc omfort Diary. Heart & Lung 30(4), 312 320. King, D., Duello, T., Miranda, P., Hodges, K., Shelton, A., Chukelu, P., & Jones, L. (2010). Strategies for recruitment of healthy premenopausal women into the African American Nutrition for Life (A NULIFE) Stud y. Journal of Women's Health 19(5), 855 862. Klag, M., Whelton, P., Randall, B., Neaton, J., Brancati, F., & Stamler, J. (1997). End stage renal disease in African American and White men. J ournal of American Medical Association, 277(16), 1293 1298. Knei pp, S., Lutz, B., & Means, D. (2009). Reasons for enrollment, the informed consent process, and trust among low income women participating in a community based participatory research study. Public Health Nursing 26(4), 362 369. Krokavcova, M., van Dijk, J., Nagyova, I., Rosenberger, J., Gavelova, M., Gdovinova, Z., & Groothoff, J. (2009). Perceived health status as measured by the SF 36 in patients with multiple sclerosis: a review. Scandinavian Journal of Caring Sciences 23(3), 529 538. Kunitz, S. J., & Pesis Katz, I. (2005). Mortality of white Americans, African Americans, and Canadians: The cause and consequences for health of welfare state institutions and policies. Milbank Quarterly 83, 5 39. Lancaster, G., Dodd, S. & Williams on, P. (2004) Design and analysis of pilot studies: recommendations for good practice. Journal of Evaluation in Clinical Practice, 10 307 312.
216 Lau, Yeung, N., Choi, K., Cheng, M., Tsui, H. & Griffiths, S. (2009). Acceptability of A/H1N1 vaccination during pandemic phase of i nfluenza A/H1N1 in Hong Kong: population based cross sectional survey. British Medical Journal 1 6. Lewis, L., & Riegel, B. (2010). Determinants of perceived health in older adults with hypertension. Heart & Lung : The Journal of Acute and Critical Care 39(1), 41 49. Loftin, W., Barnett, S., Bunn, P., & Sullivan, P. (2005). Recruitment and retention of rural African Americans in diabetes research: lessons learned. Diabetes Educator 31(2), 251 259. Looijmann, I., Verheij, T., van Delden, J., van Essen, G., & Hak, E. (2007). No intention to comply with influenza and pneumococcal vaccination: behavioral determinants among smokers and non smokers. Preventive Medicine 45(5), 380 385. Madhavan S., Rosenbluth S., & Amonkar M. (2000). Predictors of influenza and pneumonia vaccinations in rural adults 18 64 years of age. Presented at: National Immunization Conference; July 5 8; Washington, DC. Magery, N., Riise, T., Johnsen, B., & Moen, B. (2007). Health related quality of life in the Royal Norwegian Navy: do es officer rank matter?. Military Medicine 172 (8), 835 842. Malat J. (2001). e althcare p Journal of Health and Social Behavior 42(4) 360 3 72. Mayo, R., Ureda, J., & Parker, V. (2001). Importance of fa talism in understanding mammography screening in rural elderly women. Journal of Women and Aging, 13(1), 57 72. McNeely, E., & Clements, S. (1994). Recruitment and retention of the older adult into research studies J ournal of Neurosurg ical Nurs ing 26(1) 57 61. Mechanic, D. (1998). The functions and limitations of trust in the provision of medical care. Journal of Health Politics, Policy & Law 23(4), 661 686. Mechanic D, Schlesinger M. (1996). The impact of managed care on patients' trust in medical c are and their physicians. The Journal of the American Medical Association, 275(21):1693 1697. Middleton, D., Fox, D., Nowalk, M., Skledar, S., Sokos, D., Zimmerma n, R., et al. (2005 ). Overcoming barriers to establishing an inpatient vaccination program fo r pneumococcus using standing orders. Infection Control and Hospital Epidemiolog y 26(11), 874 881. Mieczkowski, T. & Wilson, S. (2002). Adult pneumococcal vaccination: a review of physician and patient barriers. Vaccine 20, 1383 1392.
217 Miller, L., Kourbat ova, E., Goodman, S., & Ray, S. (2005). Brief report: Risk factors for pneumococcal vaccine refusal in adults. Journal of Geriatric Internal Medicine. 20 650 652. Moore, T. (2009). Unveiling the various methods to measuring socioeconomic status (SES) amo ng minorities. 2009 Southern Nursing Research Society Conference. Southern Online Journal of Nursing Research 9(2). Moreno John, G., Forte, D., Rangel Lugo, M., & Perez Stable, E. (2000). Clinical Research with older African Americans and Latinos: perspec tives from the community. Research on Aging, 22(6), 668 692. Morrison, K., Lake, D., Crook, J., Carlone, G., Ades, E., Facklam, R., & Sampson, J. (2000). Confirmation of psaA in all 90 serotypes of streptococcus pneumoniae by PCR and potential of this assa y for identification and diagnosis. Journal of Clinical Microbiology 38(1), 434 437. Mossey, J. rated health: A predictor of m ortality among the e American Journal of Public Health 72 800 80 8. Muench, D., Herbert, F ., Rajnik, M. (2010). Pneumococcal infections: multifaceted approach and its effectiveness. J ournal of the National Med ical Assoc iation http://emedicine.medscape.com/article/225811 National Center for Health Statistics Health, United States (2007). Char t book on trends in the health of Americans: Hyattsville, MD. National Immunization Survey Vaccination coverage among U.S. adults (2007). Retrieved June, 2011 from: http://www.cdc.gov/vaccines/stats surv/nis/downloads/nis adult summer 2007.pdf National Institutes of Health. (2000). Outreach notebook for the NIH guidelines on inclusion of women and minorities as subjects in clinical research. NIH Publication No.97 4160. National Vaccine Advisory Committee. Report of the Subcommittee on Adult Immunization. 2011. Available from: http://www.hhs.gov/nvpo/nvac/meetings/upcomingmeetings/02162011agenda.ht mlhttp://www.hhs.gov/nvpo/nvac/Adult%20Immunization%202 90.pdf. Nexoe, J ., Kragstrup, J., & Sogaard, J. (1999). Decision on influenza vaccination among the elde rly: a questionnaire study based on the Health Belief Model and the Multidimensional Locus of Control Theory. Scandinavian Journal of Primary Health Care 17(2), 105 110. Nichol, K., Lofgren, R., Gapinski, J. (1992). Influenza vaccination. Knowledge, attit udes, and behavior among high risk outpatients. Archives of Internal Medicine 152(1), 106 110.
218 Nichol e K. (1992). Long term success with the national health objective for influenza immunization: an institution wide model. Journal of General Internal Medic ine. 7 595 600. Norusis, M. (2008). SPSS16.0 Guide to data analysis. Prentice Hall, Upper Saddle River, NJ. Nowalk, M., Tabbarah, M., Hart, J., Fox, D., Raymund, M., Wilson, S., & Zimmerman, R. (2009). Office manager and nurse perspectives on facilitator s of adult immunization. American Journal of Managed Care 15(10), 755 760. Nowalk, M., Zimmerman, R., Shen, S., Jewell, I., & Raymund, M. (2004). Barriers to pneumococcal and influenza vaccination in older community dwelling adults (2000 2001). Journal o f the American Geriatrics Society 52(1), 25 30. Nu orti J., Butler J., & Farley M. (2000). Cigarette smoking and invasive pneumococcal disease. N ew Eng land J ournal of Med icine 342, 681 89. Oliver M Goodwin M Gotler R ., Gregory, P., & Stange, K. (2 001) clinical encounters: a re Afr ican American patients treated d Journal of the National Medical Association 93(10) 380 3 85. A., & Forrest, C. (2006). Immunization disparities in older Americans. American Journal of P reventive Medicine 31(2), 150 158. O'Malley, A., Sheppard, V., Schwartz, M., & Mandelblatt, J. (2004). The role of trust in use of preventive services among low income African American women. Preventive Medicine 38(6), 777 785. Owen Smith, A., Sterk, C ., McCarty, F., Hankerson Dyson, D., & DiClemente, R. (2010). Development and evaluation of a complementary and alternative medicine use survey in African Americans with acquired immune deficiency syndrome. Journal of Alternative & Complementary Medicine 16(5), 569 577. Paskett E ., DeGraffinreid C., & Tatum C. (1994). The recruitment of African Americans in cancer prevention and control studies. Proceedings of the American Association for Cancer Research 35, 665 666. Pearson, S., & Raeke, L. (2000). Pat measures, and little data. Journal of General Internal Medicine 15(7), 509 513. Peek, M., Odoms Young, A., Quinn, M., Gorawara Bhat, R., Wilson, S., & Chin, M. (2010). Race and shared decision making: perspe ctives of African Americans with diabetes. Social Science & Medicine 71 (1), 1 9. Pilishvili, T., Lexau, C., Farley, M., Hadler, J., Harrison, L., Bennett, N., et al. (2010). Sustained reductions in invasive pneumococcal disease in the era of conjugate va ccine. Journal of Infectious Diseases 201 (1), 32 41.
219 Plowden, K., John, W., Vasquez, E., & Kimani, J. (2006). Reaching African American men: a qualitative analysis. Journal of Community Health Nursing 23(3), 147 158. Podolsky, S. (2005). Public health then and now. The changing fate of pneumonia as a public health concern in 20th century America and beyond. American Journal of Public Health 95 (12), 2144 2154 Portney, L., & Watkins, M. (2008). Foundations of clinical research: applications to practice Upper Saddle River, N.J.: Pearson/Prentice Hall. Prescott, P., Soeken, K. (1989). Methodology corner: The potential uses of pilot work. Nursing Research 38(1), 60 62. Rangel, M., Shoenbach, V., Weigle, K., Hogan, V., Strauss, R., & Bangdiwala, S. (200 5). Racial and ethnic disparities in influenza vaccination among elderly adults Journal of General Internal Medicine 20, 426 431. Reaves, P., Weaver, R., Gaines, F., Carson, E., Williams, R., McCray, R., & Holder, M. (2009). Culturally sensitive wellnes s challenge interventions in predominantly African American adults in the southern region of Florida: an observational study. Internet Journal of Health 8(2), 21. Rehm, S., File, T., Metersky, M., Nichol, K., & Schaffner, W. (2012). Identifying barriers to adult pneumococcal vaccination: an NFID task force meeting. Postgraduate Medicine 124 (3), 71 79. Resnick, B., & Parker, R. (2001). Simplified scoring and psychometrics of the revised 12 item Short Form Health Survey. Outcomes Management For Nursing P ractice 5 (4), 161 166. Rettig, R. (2000). Are patients a scarce resource for academic clinical research? Heatth Affairs, 19(6), 195 205. Battle Creek, MI: WK Kellog Foundation Report. Richard, K., Tammy, A., Janosky, J., Michael, J., Mahlon, R., Stephen, A., et al. (2003). What affects influenza vaccination rates among older patients? An analysis from inner city, suburban, rural, and veterans affairs practices. The A merican journal of medicine 114(1) 31 38. Rimple, D., Weiss, S., Brett, M., & Ernst, A. (2006). An emergency department based vaccination program: overcoming the barriers for adults at high risk for vaccine preventable diseases. Academic Emergency Medici ne 13(9), 922 930. Robinson KA, Baughman W, Roth rock G, et al. (2001). Epidemiology of invasive streptococcus pneumoniae i nfections in the United States 1995 1998:
220 Opportunities for prevention in the conjugate vaccine e ra. J AM A 285(13), 1729 1735. Rosen baum, S., Stewart, A., Cox, M., & Lee, A. (2003). The epidemiology of U.S. immunization law: a national study for the national immunization program, centers for disease control and prevention. Medicaid coverage of immunizations for non institutionalized ad ults. The George Washington University Medical Center: Washington, D.C. Rosenthal, D. (2007). Challenges to accepting a human papilloma virus (HPV) vaccine: a qualitative study of Australian women. Women & Health 45 (2), 59 73. Ross, C., & Mirowsky J ( 1 995 Journal of Health and Social Behavior 36 230 2 43. Ross C., & Wu C. (1995). The links between education and health. American Sociological Review 60, 719 7 45. Ru Chien, C., Reiber, G., & Neuzil, K. (2006). Influenza and pneumococcal vaccination in older veterans: results from the Behavioral Risk Factor Surveillance System. Journal of the American Geriatrics Society 54(2), 217 223. Rushton, T., Ganguly, R., Sinnott, J., & Banerji, M. (1994). Barriers to immunization an examination of factors that influence the application of pneumococcal vaccine by house staff. Vaccine 12(13) 1173 1179. Saha S Komaromy M Koepsell T & Bindman A (1999) p hysician r acial c oncordance and the p erceived q uality and u se o f h ealth c Archives of Internal Medicine 159(9), 997 1004. Santibanez, T., Nowalk, M., Zimmerman, R., Jewell, I., Bardel la, I., Wilson, S., & Terry, M. (2002). Knowledge and beliefs about influenza, pneumococcal disease, and immunizations among older people. Journal of the American Geriatrics Society 50, 1711 1716. Sapsis, K., & Janssen, A. (2002). Immunization: A qualitative assessment of the beliefs of physicians and older Hispanic and African Americans national immunization program. 37 th National Immunization Conference. Atlanta, GA. Satcher, D. (2003). Overlooked and underserved: improving the health of men of color. American Journal of Public Health. 93(5) 707 709. Schweon, S. (2005 ). Do your part to stave off strep RN, 68 (10), 34 38. Shavers, V. (2007). Measurement of socioeconomic status in health disparities research. J ournal of the Nat iona l Med ical Assoc iation 99(9), 1013 1323.
221 Shevlin, J., Summers Bean, C., Whitney, T., & Todd D. (2002). A systematic approach for increasing pneumococcal vaccination rates at an inner city public hospital. American Journal of Preventive Medicine 92 92 97. Shields, M., & Shooshtari, S. (2001). Determinants of self perceived health. Health Reports 13 (1), 35 52. Sisk, J., Whang, W., Butler, J., Sneller, V. & Whitney, C. (2003). Cost effectiveness of vaccination against invasive pneumococcal disease among people 50 through 64 years of age: Role of co morbid conditions and race. Annals of Internal Medicine. 12 (138) 960 968 Sme dley, B., Stith, A., Nelson, A. (2002). Unequal treatment: confronting racial and ethnic disparities in health care minorities more likely to receive lower quality health care, regardless of income and insurance coverage. The National Academies. Executive Summary. Retrieved November, 2 011 from: http://www.allhealth.org/briefingmaterials/UnequalTreatment 56.pdf Smith A Shelley J ., & Dennerstein L. (1994). Self rated health: biological continuum or social discontinuity. Soc ial Sci ence and Med icine 39 77 83. Smith, L., & Harrison, M. (2009). Framework for planning and conducting pilot studies. Ostomy Wound Management 55(12), 34. Smith, Y., Johnson, A., Newman, L., Greene, A., Johnson, T., & Rogers, J. (2007). Perceptions of clinical research participation among African Americ an women. Journal of Women's Health 16(3), 423 428. Sohler, N., Fitzpatrick, L., Lindsay, R., Anastos, K., & Cunningham, C. (2007). Does patient provider racial/ethnic concordance influence ratings of trust in people with HIV infection? AIDS & Behavior 11(6), 884 896. Sokos, D. (2005 ). vaccination. American Journal of Health System Pharmacists 62, 367 377. Sokos, D., Skledar, S., Ervin, K., Norwalk, M., Zimmerma n, R., Fox, D., & Middleton, D. (2007). Designing and implementing a hospital based vaccine standing orders program. American Journal of Health Systems Pharmacists 6 4, 1096 1102. Spurlock, W., & Cullins, L. (2006). Cancer fatalism and breast cancer screening in African American women. A ssociation of Black Nursing Faculty Journal 17 (1), 38 43. Staffileno, B., & Coke, L. (2006). Recruiting and retaining young, sedentary, hypertension prone African American women in a physical activity intervention study. Journal of Cardiovascular Nursing 21(3), 208 216.
222 Stanton, B., Black, M., Feigelman, S., Ricardo, I., Galbraith, J., Li, X., & Nesbitt, R. (1995). Development of a culturally, theoretically and developmentally based survey instrument for assessing risk behaviors among African American e arly adolescents living in urban low income neighborhoods. AIDS Education & Prevention 7(2), 160 177. Stokley, S., Shaw, K., Barker, L., Santoli, J., & Shefer, A. (2006). Impact of state vaccine financing policy on uptake of heptavalent pneumococcal conj ugate vaccine. American Journal of Public Health 96(7), 1308 1313. Tarrant, C., Dixon Woods, M., Colman, A., & Stokes, T. (2010). Continuity and trust in primary care: a qualitative study informed by game theory. Ann als of Fam ily Med icine 8, 440 446. T elford, R. (2003). What influences elderly peoples' decisions about whether to accept the influenza vaccination? A qualitative study. Health Education Research 18 (6), 743 753. Thanavaro, J., Moore, S., Anthony, M., Narsavage, G., & Delicath, T. (2006). P redictors of health promotion behavior in women without prior history of coronary heart disease Applied Nursing Research, 19 (3), 149 155. Thom, D., Ribi sl, K., Stewart, A., & Luke, D. (1999). Further validation and reliability testing of the trust in phys ician scale. Medical Care 37(5), 510 517. Thomas D Ray S & Morton F ( 2003 ). Patient education strategies to improve pneumococcal vaccination rates: randomized trial. J ournal of Investig ative Med icine ( 51(3), 141 148. Thomas, S., & Quinn, S. (1991 ) The Tuskegee Syphilis Study, 1932 to 1972: Implications for HIV education and AIDS risk education programs in the Black community American Journal of Public Health 81 (11). 1498 1504. Tillett, J. (2004). The use of vaccines in pregnancy. The Journal o f Perinatal & Neonatal Nursing 18(3), 216 229 US Department of Health and Human Services : National Center for Health Statistics. (2008). Heyman K., Schiller J., & Barnes P. Early release of selected estimates based on data from the 2007 National Health I nterview Survey. US Department of Health and Human Services: Office of Minority Health (2001). National standards on culturally and linguistically appropriate services (CLAS) Retrieved August, 2011 from: http://minorityhealth.hhs.gov/templates/browse.aspx ?lvl=2&lvlid=15 US Department of Health and Human Services: Office of Minority Research (2008) Retrieved April, 2011 f rom: http://minorityhealth.hhs.gov/templates/content.aspx?ID=2139
223 US Department of Health and Human Services: Office of Minority Healt h (20 09 ). HHS a ction p lan to r educe r acial and e thnic h ealth d isparities. A nation free of dis parities in health and health care http://minorityhealth.hhs.gov/npa/files/Plans/HHS/HHS_Plan_complete.pdf US Department of Labor. U.S. Bureau of Labor Statist ics. August 2010. Report 1026. Retrieved September, 2010 from: http://www.bls.gov/cps/cpsrace2009.pdf van Wieringen J Harmsen J & Bruijnzeels M. ( 2002 ) Intercultural Communication in General Practice. European Journal of Public Health 12(1), 63 68. Wahid, S., Nag, S., Bilous, R ., Marshall, A., & Robinson, J. (2001). Audit of influenza and pneumococcal vaccination uptake in diabetic patients attending secondary care in the northern region. Diabetic Medicine 18(7), 599 603. Wasserman, J., Flannery, M ., & Clair, J. (2007). Raising the ivory tower: the production of knowledge and distrust of medicine among African Americans. Journal of Medical Ethics 33(3), 177 180. Watts, R. (2003). Race consciousness and the health of African Americans. Online Journ al of Issues in Nursing 8(1). Weinstein N., Sandman, P, & Roberts, N. (1989). Public response to the risk from radon, 1988 1989 New Brunswick, NJ: Rutgers University, Environmental Communication Research Program. Weissfeld, J. Kirscht, J., & Brock, B. (1990). Health beliefs in a population: The Michigan blood pressure survey. Health Education Quarterly, 17(2), 141 155 Williams, E., Horton, N., Samet, J., & Saitz, R. (2007) Do brief measures of readiness to change predict alcohol consumption and consequ ences in primary care patients with unhealthy alcohol use? Alcoholism Clinical and Experimental Research 31(3), 428 35. Williams, I., Utz, S., Jones, R., Hinton, I., Steeves, R., & Alexander, G. (2011). Recruitment of rural African Americans for research projects: lessons learned. Southern Online Journal of Nursing Research 11(1) Williams, S., Hanson, L., Boyd, C., Green, M., Goldmon, M., Wright, G., et al. (2008). Communication, decision making, and cancer: what African Americans want physicians to kno w. Journal of Palliative Medicine 11 (9), 1221 1226. Winston, C., Pascale, Wortley, P., & Lees, K. (2006). Factors associated with vaccination of Medicare beneficiaries in five U.S. communities: results from the Racial and Ethnic Adult Disparities in Immu nization Initiative Survey, 2003. Journal of the American Geriatrics Society 54(2), 303 310.
224 Young, D., Bleyl, K., Cl ark, T., Oderda, G., & Liou, T. (2004). Increasing pneumococcal vaccination rates among adults with cystic fibrosis. American Journal of Health Care Pharmacology 5(88), 232 238. Youngblut, J., & Casper, G. (1993). Single item indicators in nursing research. Research in Nursing and Health 16, 459 465. Zimmerman, R., Nowalk, M., Tabbarah, M., Hart, J., Fox, D., & Raymund, M. (2009). Underst anding adult vaccination in urban, lower socioeconomic settings: influence of physician and prevention systems. Annals of Family Medicine 7(6), 534 541.
225 B IOGRAPHICAL SKETCH Carla A nn Fry began her career in the medical field in 1985 when she joined th e United States Army as a Medic. During her time in the military, Carla worked as a medic with the fourth infantry division prior to attend ing West Point Preparatory Academy and the US Army School of Practical Nursing. Upon completion of six years in the m ilitary, Carla went back to school at Kent State University where she would complete her a ssociate then subsequently her b degree to become a r egistered n urse. During her time as an LPN and RN, Carla worked at the bedside as a critical care nurse in the Coronary Care Unit, Medical and Surgical Intensive Care Units, Trauma Centers, a Burn Unit, and the Cardiac Cath eterization Lab. In 2004 when Carla relocated to Jacksonville, Florida, she had the privilege of assisting with the opening of a new stat e of the art all electronic medical center. At that time, she returned to school to complete her m degree in Nursing with a minor in Education. In 2008, Carla was accepted into the PhD program at the University of Florida as a Maren Fellow. She lef t her role at Baptist Medical Center South as the Education Coordinator and began a career at Jacksonville University as an Assistant Professor of Nursing where she remains today. Presently Carla teaches critical care theory, research, information manageme nt, and a variety of clinicals. Carla still loves direct patient care so she maintains her skills by Carla enjoys spending time with her te n year old daughter and friend s swimming, snorkeling, fishing sailing, travelling, and exploring new restaurants.