Race, Rurality and Lifestyle Communication for Diet and Exercise in Primary Care

MISSING IMAGE

Material Information

Title:
Race, Rurality and Lifestyle Communication for Diet and Exercise in Primary Care
Physical Description:
1 online resource (143 p.)
Language:
english
Creator:
Andre Glenn, Rachel
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology, Clinical and Health Psychology
Committee Chair:
Perri, Michael G
Committee Members:
Anton, Stephen Douglas
Janicke, David M
Harman, Jeffrey S
Martin, Anatole D

Subjects

Subjects / Keywords:
care -- counseling -- diet -- exercise -- lifestyle -- obesity -- primary
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre:
Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Provision of weight-related advice in primary care is an important objective toward reducing racial/ethnic and rural-urban disparities in obesity.  However, the combined effect of these factors on obesity-related health communication has not been evaluated exclusively.  The current study examined lifestyle advice—diet, exercise, and combination—in a nationally representative sample of ethnically diverse adults (N = 41,838) using the Medical Expenditures Panel Survey.  Among overweight and obese adults, rural Caucasians had significantly lower odds for receipt of diet,exercise and combination advice compared to urban Caucasians.  Furthermore, obese, rural African Americans had significantly lower odds of receiving any lifestyle advice as compared to urban Caucasians and African Americans. Examined by race/ethnicity, odds of diet advice alone was significantly higher for overweight African Americans. For rurality, among both overweight and obese adults, urban respondents had significantly greater odds of reporting that they received any advice.  Considering obesity-related comorbidity status, obese adults with at least one priority condition had significantly greater odds than all comparison groups (i.e., obese adults without any conditions,non-obese adults with and without at least one condition) to receive any advice.  The number of comorbidities, as well as race-MSA, also influenced odds of intervention to varying degrees.  Both rural Caucasians and African Americans had lower odds of receiving any advice regardless of comorbidity status.  Finally, respondents’ perceptions of provider communication did not differentiate those who received lifestyle advice;however, both urban and rural African Americans were significantly more likely to provide higher scores on provider communication compared to urban Caucasians.  No rural-urban differences were observed among Caucasians. While rates of lifestyle communication were relatively low (28.8% to 37.8%), these findings suggest rural-urban but not racial/ethnic disparities in obesity-related health communication.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Rachel Andre Glenn.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Perri, Michael G.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044611:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 RACE, RURALITY, AND LIFESTYLE COMMUNICATION FOR DIET AND EXERCISE IN PRIMARY CARE By RACHEL ANDR GLENN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMEN TS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Rachel Andr Glenn

PAGE 3

3 To the people God has allowed me to love through Him, e specially my parents, my sister, and my husband

PAGE 4

4 ACKNOWLEDGMENTS I would like to first acknowledge the persons to whom this project is dedicated. How blessed I have been to have my parents, Paul and Chantale Andr, as examples of what it truly means to see a dream through no matter how large and for also instilling in me the knowledge that nothing is impossible with God. I would also like to thank my sister and first love, Patricia Andr, for her unwavering support, unconditional love, and laughter, which helped see me through most of my graduate training. Last, I would like to thank my hu sband, John Glenn, for finding me when and where I least expected it and for sharing this journey with me. Next, I acknowledge those who have provided their time and expertise throughout my graduate studies. First, I would like to thank my advisor and com mittee chair, Dr. Michael G. Perri for his guidance and support in both my clinical and research endeavors. My sincerest thanks to Dr. Jeffrey S. Harman, who served as an invaluable external committee member as evidenced by his generosity with his time a nd expertise in research methods and statistics, as well as the other members of my dissertation committee: Drs. David Janicke, Stephen Anton, and Daniel Martin. Finally, I extend a special thank you to Dr. Jules Harrell who first introduced me to the pos sibilities of a career in clinical psychology and for his continued mentorship from afar. Last, I would like to thank my family and friends whose words of encouragement, love and support I could not have done without.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ 4 LIST OF TABLES ................................ ................................ ................................ ........... 7 LIST OF ABBREVIATIONS ................................ ................................ ............................ 8 ABSTRACT ................................ ................................ ................................ .................. 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ... 12 Overview ................................ ................................ ................................ ............... 12 The Scope and Seriousness of Obesity in the US ................................ ................. 13 Racial and Rural Disparities in Obesity and Associated Disease Burden ............... 15 Racial Dis parities in Obesity ................................ ................................ ............ 15 Rural Urban Disparities in Obesity ................................ ................................ .. 17 Race/Ethnicity, Rurality, and Obesity ................................ .............................. 18 Obesity Management in Primary Care ................................ ................................ ... 20 Current Recommendations for Weight Management in Primary Care Settings ................................ ................................ ................................ ........ 21 Existing Models of Practice: Prevention and Control ................................ ....... 23 Trends in Weight Management Counseling in Primary Care ........................... 27 Race, Rurality, and Regional Contexts ................................ ............................ 33 Methodological Limitations ................................ ................................ .............. 36 Empirical Support for Weight Management Counseling in Prima ry Care ............... 37 Factors Influencing Physician Practices and Patient Provider Communication ...... 4 2 Patient Factors ................................ ................................ ................................ 43 Physician Factors ................................ ................................ ............................ 46 Implicit and Weight Bias in the Primary Care Setting ................................ ....... 48 Aims and Hyp otheses of the Current Study ................................ ........................... 52 Aim #1: Effect of Race/Ethnicity and Rurality on Lifestyle Advice .................... 52 Aim #2: Effect of Obesity Rel ated Comorbidities on Lifestyle Advice ............... 53 Aim #3: Provider Communication Related to Lifestyle Advice ......................... 55 Exploratory Aims ................................ ................................ ............................. 56 2 METHODS ................................ ................................ ................................ ............ 57 Data Source ................................ ................................ ................................ .......... 57 Participants ................................ ................................ ................................ ............ 58 Procedures ................................ ................................ ................................ ............ 59 Dependent Variables ................................ ................................ ....................... 60 Independent Variables ................................ ................................ .................... 62 Statistical Analyses ................................ ................................ ................................ 66

PAGE 6

6 Aim #1: Effect of Race/Ethnicity and Rurality on Lifestyle Advice .................... 67 Aim #2: Effect of Obesity Related Comorbidities on Lifestyle Advice ............... 68 Aim #3: Provider Behaviors Related to Lifestyle Advice ................................ .. 68 Explorato ry Aims ................................ ................................ ............................. 68 3 RESULTS ................................ ................................ ................................ .............. 70 Overview ................................ ................................ ................................ ............... 70 Baseline Characteristics ................................ ................................ ........................ 71 Primary Aim: Effect of Race/Ethnicity and Rurality on Lifestyle Advice .................. 73 Secondary Aim: Effect of Obesity Related Comorbidities on L ifestyle Advice ........ 76 Tertiary Aim: Relationship Between Provider Behaviors and Lifestyle Advice ........ 82 Exploratory Aims: Effect of Lifes tyle Advice on Respondent Behavior ................... 83 4 DISCUSSION ................................ ................................ ................................ ...... 101 Main Findings ................................ ................................ ................................ ...... 101 Comparison to Prior Research ................................ ................................ ............. 107 Strengths and Limitations ................................ ................................ .................... 111 Policy and Clinical Service Implications and Future Direction s ............................ 116 Summary and Conclusions ................................ ................................ .................. 120 APPENDIX : RATES OF LIFESTYLE COUNSELING ................................ ................. 122 LIST OF REFERENCES ................................ ................................ ............................ 126 BIOGRAPHICAL SKETCH ................................ ................................ ......................... 143

PAGE 7

7 LIST OF TABLES Table page 3 1 Participant demographics at baseline ................................ ................................ 86 3 2 Priority conditions, types of lifestyle advice received, and provider communication ................................ ................................ ................................ .. 87 3 3 Receipt of lifestyle advice among overweight adults: Race MSA as predictor .. 88 3 4 Receipt of lifestyle advice among obese adults: Race MSA as predictor .......... 89 3 5 Receipt of lifestyle advice among overweight adults: Race as predictor ........... 90 3 6 Receipt of lifestyle advice among obese adults: Race as predictor ................... 91 3 7 Receipt of lifestyle advice among overweight adults: MSA as predictor ............ 92 3 8 Receipt of lifestyle advice among obese adults: MSA a s predictor ................... 93 3 9 Receipt of lifestyle advice among adults with and without obesity related comorbidities ................................ ................................ ................................ ..... 94 3 10 Receipt of life style advice among adults by number of obesity related comorbidities ................................ ................................ ................................ ..... 95 3 11 Receipt of lifestyle advice among obese adults: Comorbidity status by race MSA as predictors ................................ ................................ ............................. 96 3 12 Receipt of lifestyle advice among obese adults: Provider communication as predictor ................................ ................................ ................................ ............ 97 3 13 Race MSA as predictor of provider communica tion amo ng obese adults .......... 98 3 14 Odds of e ngaging in physical activity at end of survey reference period among adults ................................ ................................ ................................ ..... 99 3 15 Combination a dvice at baseline as predictor of wei ght change among adults 100 A 1 Rates of weight reduction counseling ................................ .............................. 122 A 2 Rates of diet an d nutrition counseling ................................ .............................. 124 A 3 Rates of physical activity counseling ................................ ............................... 125

PAGE 8

8 LIST OF ABBREVIATION S ACSM American College of Sports Medicine ADA A merican Dietetic As sociation AHRQ Agency for Healthcare Re search and Quality ANCOVA analysis of covariance ANOVA analysis of variance BMI body mass index CDC Centers for Disease Control and Prevention FDA Food and Drug Administration HIV/AIDS Human Immunodeficiency Virus /Ac quired Immunodeficiency Syndrome ICD 9 International Classification of Diseases and Related Health Problems Ninth Edition IMI intensive medical intervention treatment condition kg/m 2 kilograms/meters squared LOSS Louisiana Obese Subjects Study MEPS Medica l Expenditures Panel Survey MEPS HC Medical Expenditures Panel Survey Household Component MEPS IC Medical Expenditures Panel Survey Insurance Component MEPS MPC Medical Expenditures Panel Survey Medical Provider Component MSA metropolitan statistical ar ea NAMCS National Ambulatory Medical Care Survey NCHS National Center for Health Statistics NHAMCS National Hospital Ambulatory Medical Care Survey NHIS National Health Interview Survey NHLBI National Heart, Lung, and Blood Institute

PAGE 9

9 OR odds ratio POWER Pr actice based Opportunities for Weight Reduction SES socioeconomic status UCC usual care condition US United States USDHHS United States Department of Health and Human Services WHO World Health Organization

PAGE 10

10 Abstract of Dissertation Presented to the Gra duate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy RACE, RURALITY, AND LIFESTYLE COMMUNICATION FOR DIET AND EXERCISE IN P RIMARY C A R E By Rachel Andr Glenn August 201 2 Chair: Michael G. Perri Major: Psychology Provision of weight related advice in primary care is an important objective toward reducing racial/ethnic and rural urban disparities in obesity. However, the combined effect of these factors on obesity relat ed health communication has not been evaluated exclusively. The current study examined lifestyle advice diet, exercise, and combination in a nationally representative sample of ethnically diverse adults (N = 41,838) using the Medical Expenditures Panel Su rvey. Among overweight and obese adults, rural Caucasians had significantly lower odds for receipt of diet, exercise and combination advice compared to urban Caucasians. Furthermore, obese, rural African Americans had significantly lower odds of receivin g any lifestyle advice as compared to urban Caucasians and African Americans. Examined by race/ethnicity, odds of diet advice alone was significantly higher for overweight African Americans. For rurality, among both overweight and obese adults, urban res pondents had significantly greater odds of reporting that they received any advice. Considering obesity related comorbidity status, obese adults with at least one priority condition had significantly greater odds than all comparison groups (i.e., obese ad ults without any conditions, non obese adults with and without at least one condition) to receive any advice. The

PAGE 11

11 number of comorbidities, as well as race MSA, also influenced odds of intervention to varying degrees. Both rural Caucasians and African Ame ricans had lower odds of of provider communication did not differentiate those who received lifestyle advice ; however, both urban and rural African Americans were sig nificantly more likely to provide higher scores on provider communication compared to urban Caucasians No rural urban differences were observed among Caucasians. While rates of lifestyle communication were relatively low (28.8% to 37.8%), these findings suggest rural urban but not racial/ethnic disparities in obesity related health communication.

PAGE 12

12 CHAPTER 1 INTRODUCTION Overview The challenges faced by rural primary health providers to address the obesity epidemic are significant given the current rates o f overweight and obesity, particularly among racial/ethnic minorities and rural adults. Moreover, there is limited access to obesity related specialty care in rural areas (Jackson, Doescher, Jerant, & Hart, 2005). Provision of weight related counseling i n the rural primary care setting is an important objective toward reducing the racial/ethnic and rural urban disparities in obesity. In general, rates of weight related counseling are low (e.g., Abid et al., 2005), and despite the higher prevalence of obe sity among non Hispanic Blacks and rural residents (e.g., Flegal, Carroll, Kit, & Ogden, 2012; Jackson, Doescher, Jerant et al., 2005), the research examining physician practices (i.e., counseling or advice) related to weight reduction, physical activity, and diet have neglected the possible interaction of race /ethnicity and rurality (i.e., metropolitan statistical area, MSA) Thus, the current study examined lifestyle advice for factors related to weight reduction (i.e., diet and exercise) in a nationally representative sample of ethnically diverse rural and urban adults. Specifically, this study examined (1) the combined and advice among overweight and obese adults, (2 ) the role of obesity related comorbidities MSA analysis), and (3) the effect of provider communication on lifestyle advice received, as well as the racial differences in appraisal o f provider communication. Finally, the study undertook an exploratory examination of changes in respondent behavior (i.e., engaging in

PAGE 13

13 moderate vigorous physical activity and BMI change) as a function of receivin g diet and/or exercise advice. The followin g sections provide a review of the scope and seriousness of obesity with a particular emphasis on the literature regarding racial/ethnic and rural disparities in obesity. This is followed by a discussion of the contribution of obesity to chronic disease a nd a summary of the related trends by race and across various regions of the United States (US). Also discussed is the unique role of primary care in obesity management. The existing models of practice in primary care, including weight reduction, diet, a nd physical activity counseling are reviewed. The final section includes research on patient provider communication and highlights factors influencing physician making and perceptions of patient health beliefs. The Scope and Seriousness of Obesity in the US Obesity, defined as Body Mass Index (BMI) > 30 kg/m 2 continues to be a major public health priority in the United States and abroad. In fact, the World Health Org 4). National survey data all indicate significant increases in the prevalence of adult obesity over the last few decades (e.g., Wang & Kumanyika, 2007). In the United States alone, recent estimates suggest the prevalence of obesity has nearly doubled since 1980 (Hedley et al., 2004) with approximately 35.7% of adults 20 years and older currently classified as obese (Flegal et al., 2012). An additional 33.1% of the population is considered overweight (BMI = 25.0 29.9 kg/m 2 ) (Flegal et al., 2012 ). An increasing body of evidence has linked obesity to a number of adverse health outcomes, including five of the ten leading causes of death: heart disease, cancer,

PAGE 14

14 stroke and other cer ebrovascular diseases, chronic lower respiratory diseases, and diabetes (National Heart, Lung, and Blood Institute (NHLBI), 1998). Obesity has also been linked to other chronic conditions such as hypertension, hyperlipidemia, metabolic syndrome, asthma, g all bladder disease, and osteoarthritis (Dumitrescu & Cotarla, 2005; Freedland, 2005; Gregg et al., 2005; Lowenfels, Sullivan, Fiorianti, & Maisonneuve, 2005; Mokdad et al., 2003; National Task Force on the Prevention and Treatment of Obesity, 2000). The detrimental effects of obesity also include disability (Alley & Chang, 2007; Gregg & Guralnik, 2007) and premature death (NHLBI, 1998), with estimates of obesity related mortality believed to be as high as 400,000 per year (Flegal, Graubard, Williamson, & Gail, 2005; Mokdad, Marks, Stroup, & Gerberding, 2004). Thus, obesity has become the second leading preventable cause of death and disease in the United States (Hoerger, 2006; US Department of Health and Human Services (USDHHS), 2001). The disease burden healthcare system. Indeed, as the prevalence of overweight and obesity increase to include the majority of the United States population, the medical costs of obesity related conditions also increase. Recent national estimates suggest total health costs associated with obesity ranging from $147 billion per year (Finkelstein, Trogdon, Cohen, & Dietz, 2009) to $190.2 billion per year (Cawley & Meyerhoefer, 2012). The direct costs of obesity (associated with diagnosis and treatment of obesity related illness) (Finkelstein et al., 2009 and Cawley & Meyerhoefer, 2012, respectively). Further, aging and morbid obesity have bee n associated with increasing healthcare costs (Wee et al.,

PAGE 15

15 2005). These increases may be attributable to increased utilization of inpatient and outpatient services (i.e., office visits, hospitalizations, nursing home care, rehabilitation), as well as high er costs associated with treatment, medication, and surgical interventions among these subgroups (Arterburn, Maciejewski, & Tsevat, 2005; Colditz & Stein, 2007). Racial and Rural Disparities in Obesity and Associated Disease Burden Racial Disparities in Ob esity The differences in obesity prevalence by race are well documented (Flegal et al., 2012; Hedley et al., 2004; National Center for Health Statistics (NCHS), 2009; Ogden et al., 2006). Recent national survey estimates suggest that the prevalence of obe sity is highest among non Hispanic Blacks (49.5%) and Hispanics (39.1%) as compared to their Caucasian counterparts (34.3%) (Flegal et al., 2012). While some data have shown that women (33.2%) exhibit higher rates of obesity than men (27.6%) (Baskin, Ard, Franklin, & Allison, 2005), these differences do not emerge in more recent estimates (Flegal et al., 2012). However, among women only, there appear to be significant disparities in the prevalence of obesity by race such that non Hispanic Black women (58. 5%) and Hispanic women (41.4%) are significantly more likely to be obese compared to their Caucasian counterparts (32.2%) (Flegal et al., 2012). Recent estimates show no racial differences in prevalence of obesity among men. This represents a continuatio n of previous trends in obesity prevalence by race (Hedley et al., 2004; Must et al., 1999; Ogden et al., 2006). Obesity contributes substantially to the disease burden of chronic health conditions, particularly for type 2 diabetes mellitus, hypertension, and hyperlipidemia (Must et al., 1999; Paeratakul, Lovejoy, Ryan, & Bray, 2002), as well as gall bladder

PAGE 16

16 disease and osteoarthritis (Must et al., 1999). For instance, some studies show an increasing prevalence of these comorbid conditions at higher BMIs ( Must et al., 1999). With few exceptions (e.g., high serum cholesterol), this pattern appears to be consistent across racial/ethnic groups (Must et al., 1999). However, other studies demonstrate racial/ethnic differences in the prevalence of obesity relat ed comorbidities and mortality (e.g., Calle, Thun, Petrelli, Rodriguez, & Heath, 1999; Kumanyika, 2005; NCHS, 2009; Paeratakul et al., 2002; Stevens et al., 1998). While these disparities persist, the differences in overall mortality rates for non Hispani c Blacks and Caucasians decreased by 9% between 1990 and 2006 (NCHS, 2009). When all cause mortality is evaluated across the BMI distribution, non Hispanic Blacks appear to fare better than their Caucasian counterparts. For example, for individuals with B MI > 35, risks for all cause mortality are higher among Caucasians than non Hispanic Blacks (Calle et al., 1999). Specifically, for Caucasians with BMI > 35, risk of death was 75 100% higher compared to Caucasians with BMI < 35. In contrast, the relative risk of death for non Hispanic Blacks was only 20 30% higher. Nonetheless, racial disparities in death rates for obesity related conditions also persist. Among non Hispanic Blacks, age adjusted death rates for 2005 were 46% higher for stroke and other c erebrovascular diseases, 31% higher for heart disease, 22% higher for cancer, and 108% higher for diabetes than among their Caucasian counterparts. Kumanyika (2005) suggests that these racial discrepancies in death rates reflect a variety of factors above and beyond disease incidence, including differences in timing of diagnosis, presence of comorbid conditions, access to and quality of treatment, treatment response (e.g., adherence), and non disease related causes of death, among

PAGE 17

17 other social and environm ental factors. Thus, the variations in disease burden of obesity by race may be attributable to multiple unassessed factors. Rural Urban Disparities in Obesity residency is also doc umented (e.g., Eberhardt et al., 2001; Jackson, Doescher, Jerant et al., 2005; Lucas, Schiller, & Benson, 2004; Patterson, Moore, Probst, & Shinogle, 2004). Recent estimates using measured height and weight show higher rates of obesity among rural (39.6%) as compared to urban (33.4%) adults, even after controlling for important sociodemographic, diet, and physical activity variables, including race/ethnicity (Befort, Nazir, & Perri, 2012). This represents a relative increase of the rural urban disparities in obesity reported in previous studies using self reported height and weight. As rural urban differences have been demonstrated in obesity, care must be taken to account for variations in obesity prevalence across geographic regions. Results from the N ational Health Interview Survey (NHIS) show obesity prevalence to be lower in the Northeast (20.3%) and West (18.9%) as compared to the South (24.3%) and Midwest (24.2%) (Lucas et al., 2004). In addition, NHIS data (Lucas et al., 2004) showed graduated in creases in obesity prevalence by level of urbanization, with the lowest rates of obesity observed in large metropolitan statistical areas (20.0%), followed by small metropolitan statistical areas (23.9%) and non metropolitan statistical areas (25.7%). Reg ional variations appear to be particularly salient for women such that urban communities in the West appear to have the highest rates of obesity nationally. However, obesity is more prevalent in rural than in urban areas in the Northeast and the South (Eb erhardt et al., 2001; Lucas et al., 2004)

PAGE 18

18 The aforementioned regional variations in obesity, particularly in the South, are to describe the largely rural, contiguous cluster of states in the southeast (i.e., Alabama, Arkansas, Georgia, Indiana, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, and Virginia) and Indiana (located on the northern border of this region) that have a markedly high er incidence of stroke deaths as compared to other regions of the country (NHLBI, 1996). The stroke prevalence rates in these states were greater than 10 percent higher than the national average and are particularly salient given the observed association between stroke and obesity, as well as other risk factors including high blood pressure and ci garette smoking (NHLBI, 1996). In addition to stroke, the overall rates of chronic disease are greater in rural than urban areas, which may be attributable to the higher prevalence of obesity in these areas (Eberhardt et al., 2001; Patterson et al., 2004). Furthermore, regional differences in prevalence of modifiable (e.g., diet composition, sedentary lifestyle) and non modifiable (e.g., race/ethnicity, socioecono mic status, access to and utilization of healthcare services) risk factors may contribute to the differences in obesity prevalence in these areas (Martin et al., 2005; Reis et al., 2004). Still, estimates of obesity, obesity related death rates, and all c ause death rates by degree of urbanization must be evaluated carefully given the low population of non Hispanic Blacks in rural areas in particular regions of the US, including the Northeast and the West Race/Ethnicity, Rurality, and Obesity Although the disparities in the prevalence of obesity have been examined independently by race/ethnicity and rurality, research evaluating the association among race/ethnicity, rurality, and obesity has been limited. Interestingly, despite the

PAGE 19

19 overwhelming evidence th at obesity is more prevalent among racial/ethnic minorities and in rural communities, the racial/ethnic differences in obesity prevalence do not emerge when assessed in rural areas. This picture is particularly complicated for non Hispanic Blacks (e.g., H ayward, Pienta, & McLaughlin 1997; Patterson et al., 2004; Sobal, Troiana, & Frongillo, 1996). By contrast, one study examining the adequacy of control of diabetes and hypertension among rural African Americans as compared to urban African Americans and both rural and urban Caucasians found that rural African Americans were not only substantially worse in glycemic control but also in systolic and diastolic blood pressure control (Mainous, King, Garr, & Pearson, 2004). Nonetheless, the absence of racial/e thnic disparities in obesity could be related to the observed homogeneity across sociodemographic domains such as occupation, household income, years of education, and insured status in rural communities. Alternately, the role of risk factors for obesity m ay explain the lack of racial/ethnic differences in obesity prevalence in some rural areas, such that Caucasians and racial/ethnic minorities in rural areas have similar prevalence of certain risk factors, particularly modifiable risk factors as unhealthfu l diets and physical inactivity patterns. Still, the evidence that rural African Americans fare worse than urban African Americans, as well as their rural and urban Caucasian counterparts on these dimensions (e.g., Champagne et al., 2004; Parks, Housemann & Brownson, 2003; Wilcox, Castro, King, Housemann, & Brownson, 2000), draws attention to the growing disparities in disease and healthcare among this cohort. As is suggested regarding a double dose of

PAGE 20

20 rural African Americans, being obese contributes further to the risks associated with being of minority status and of having less access to healthc are as a function of living in a rural area. Obesity Management in Primary Care Given the marked increase in the prevalence of obesity and obesity related health conditions, the healthcare system is increasingly overwhelmed by a number of obese patients re quiring disease management in outpatient medical settings. Indeed, the majority of primary care patients are overweight or obese, with obesity prevalence estimates among primary care populations exceeding those of the general population (Bowerman et al., 2001). In addition, the majority of obese patients have one or more comorbidities (Rippe, Crossley, & Ringer, 1998), and more often obesity is being considered a chronic disease (Perri, Sears, & Clark, 1993; Williamson & Perrin, 1996; Yanovski, 1993). Ev en for patients who have not yet developed overt disease, it may be beneficial to address weight management in the primary care setting (Coeytaux et al., 2004). Indeed, nearly the entire population (>80%) will access the healthcare system in any given yea r (US Census Bureau, 2012) making primary care visits a prime opportunity to address obesity. In recent years, primary care has been identified as an important mechanism for identification, evaluation, and treatment of obesity (Katz & Faridi, 2007; Nawaz & Katz, 2001; NHLBI, 1998; US Preventive Services Task Force, 2003). The current recommendations call for routine screening for obesity utilizing BMI (measured as kg/m 2 high int ensity behavioral interventions targeted at reductions of 5 10% of body weight (Donnelly et al., 2009; Hill & Wyatt, 2002; Seagle, Strain, Makris, & Reeves, 2009; US

PAGE 21

21 Preventive Services Task Force, 2003). Modest weight losses (i.e., of 5 10% of body weigh t) have been shown to produce significant health benefits, including reduction of risk factors such as elevated levels of blood pressure or glucose and abnormal lipid profiles (NHLBI, 1998). Further, clinical guidelines explicitly include assessment of re adiness to change as a prerequisite to development and implementation of a weight management plan (Anderson & Wadden, 1999; NHLBI, 1998). Yet, obesity remains underdiagnosed and undertreated in the primary care setting and the challenges faced by primary care practices to provide quality care continue to prevail (e.g., Ma, Xiao, & Stafford, 2009a; Ma, Xiao, & Stafford, 2009b; Stafford, Farhat, Misra, & Schoenfeld, 2000). The following sections review the current recommendations and approaches to obesity m anagement in primary care practice settings. Current Recommendations for Weight Management in Primary Care Settings As stated previously, the address of the obesity epidemic has become one of the leading public health agendas. The American Dietetic Associ ation (ADA) and the American College of Sports Medicine (ACSM) offer the most recent review of data and recommendations for weight management, particularly in the healthcare context (Seagle et al., 2009 and Donnelly et al., 2009, respectively). Both posit ions offer additions to previous guidelines, which call for routine screening for obesity utilizing BMI and waist circumference, assessment of patient knowledge and readiness for change, and development of a weight management plan targeting reductions of 5 10% of body weight (e.g., Hill & Wyatt, 2002; Lichtenstein et al., 2006; US Preventive Services Task Force, 2003). The ADA provides an evidence based approach for the development of clinical practice guidelines for nutrition care, including assessment, di agnosis, intervention,

PAGE 22

22 monitoring and evaluation (Seagle et al., 2009). Specific recommendations for assessment call for a multidisciplinary approach consisting of clinical measurement of weight status using height, weight, BMI, and waist circumference, a s well as a biopsychosocial evaluation assessing health risk (i.e., weight related comorbidities, other physiologic causes of excess weight, musculoskeletal problems) to establish a diagnosis, guide in the development of weight loss goals, and document out comes. In addition, an assessment of patient goals and barriers to successful weight loss (e.g., depression, eating disorder, food access) should precede establishing realistic expectations for weight loss (i.e., education regarding benefits of modest wei ght loss, de emphasis of cosmetic goals). Next, the ADA recommends nutritional diagnosis to aid in the goal of achieving a negative energy balance required for weight loss. The recommended strategy noted for reduction of energy intake is a low fat, low c alorie diet (though reduced carbohydrate diets are also supported), with use of portion control, meal substitutions, and development of a consistent meal pattern consisting of 4 5 meals/snacks per day (especially breakfast). The ADA notes that successful n utritional intervention should be supplemented by physical activity, which increases the magnitude of weight loss and is implicated in the prevention of weight regain. Specifically, the ADA suggests provider counseling reinforce national recommendations f or reduction of health risk, prevention of weight gain or weight loss, and long term weight maintenance, which call for 30, 60, and 60 90 minutes of moderate intensity physical activity on most days of the week, respectively. The ADA also supports use of behavioral interventions (i.e., cognitive behavioral therapy) to help patients develop behavioral skills for weight loss, including self

PAGE 23

23 monitoring, problem solving, stimulus control, and cognitive restructuring, in addition to nutrition and exercise educa tion and skills. Finally, use of weight loss medication and bariatric surgery are noted as effective lines of intervention for patients who meet criteria for these therapies. While the ADA provides more comprehensive guidelines for weight management in pr imary care settings, the ACSM guidelines focus specifically on the role of physical activity to short term and sustained weight loss (Donnelly et al., 2009). The primary position of the ACSM is that physical activity leads to modest weight losses in the a bsence of diet restriction, while combined with diet modification has an additive effect on weight loss. Specific ACSM guidelines have been improved to account for the different energy expenditure needs for various weight management goals. To prevent sig nificant weight gain and reduce associated chronic disease risk factors, the ACSM recommends at least 150 minutes per week of moderate intensity physical activity for all adults. Recommendations for overweight and obese suggest 150 250 minutes per week of moderate intensity physical activity to achieve modest reductions in weight. Finally, more meaningful weight losses and weight maintenance may be achieved with greater than 250 minutes per week of moderate intensity physical activity. Additionally, the ACSM also recommends strength training as a means of decreasing fat mass and further reducing obesity related health risks. Existing Models of Practice: Prevention and Control Obesity management implies both prevention and control (Katz & Faridi, 2007). In the primary care setting, obesity management entails an array of activities driven by a disease focused or prevention focused approach.

PAGE 24

24 Disease focused weight management Weight loss, diet, and physical activity counseling often occur as a function of treatment for comorbid conditions (e.g., hypertension, high cholesterol, type 2 diabetes, arthritis) for which obesity is a modifiable risk factor. Studies reporting rates of lifestyle counseling for these specific patient groups show higher rates than t hose observed in studies of general trends in weight management (described in later sections) (e.g., Carlson, Maynard, Fulton, Hootman, & Yoon, 2009; Kreuter, Scharff, Brennan, & Lukwago, 1997; Morrato, Hill, Wyatt, Ghushchyan, & Sullivan, 2006; Wee, McCar thy, Davis, & Phillips, 1999). Specifically, the presence of an obesity related or other chronic conditions appear to increase the likelihood of receiving weight related counseling; however, prevalence of such counseling is still relatively lower than exp ected, ranging from 30.8% (Bell & Kravitz, 2008) to 73.0% (Egede, 2003). For example, among inactive obese adults with arthritis/joint problems or hypertension, 59.1% and 74.0%, respectively, reported ever receiving physical activity or exercise counselin g (Carlson et al., 2009). Similar to findings in previous examinations of counseling practices, provision of physical activity counseling also increased with increases in BMI. These findings are relatively consistent across studies of lifestyle modificat ion for patients with hypertension (Bell & Kravitz, 2008), diabetes (Egede, 2003; Egede & Zheng, 2002), and arthritis (Mehrotra, Naimi, Serdula, Bolen, & Pearson, 2004). The approach to obesity management in the primary care setting that would ideally beco me the dominant model of treatment is the continuous care model. Perri, Nezu, and Viegener (1992) and Perri, Sears, and Clark (1993) were among the first to propose a continuous care model for obesity management that was focused on overcoming the

PAGE 25

25 challeng es of long term weight maintenance (i.e., after initial weight loss efforts). Similarly, the chronic care model, systems based, multidisciplinary approach to chronic disease management as applied to obesity, calls for continued professional contact geared towards engaging patients in self management and practice of nutritional, physical activity, and weight maintenance advice (Ely et al., 2008). Further, the model applicat ion of clinical guidelines and continuing education and training, which have been demonstrated to enhance the long term maintenance of weight loss (Ely et al., 2008). Prevention of obesity Kumanyika (2007) provides an overview of the conceptual framework s for obesity prevention, which in recent years has emerged as its own research area within the broader context of obesity literature. The focus here is on the relevant paradigms, excluding other guiding frameworks that do not apply to the discussion of o besity management in primary care (e.g., WHO framework for population levels of prevention, ecological models, population health model). Particular attention has been drawn to conceptual frameworks for defining prevention of obesity. Here, the primary de termination of prevention is based upon whether obesity prevention necessarily includes treatment of preexisting weight issues as in the primary secondary tertiary prevention continuum. In the chronic disease model, prevention is defined along a continuum that accounts for disease progression in terms of diagnosis and prognosis and emphasizes risk reduction (Kumanyika, 2007). As applied to obesity, primary prevention refers to interventions to reduce the incidence of new cases of obesity by preventing exc ess weight gain in the normal and overweight population; while secondary prevention applies to interventions geared towards reducing the prevalence

PAGE 26

26 of obesity by implementing weight stabilization or weight loss strategies among those who fall within the fi rst two classes of obesity (i.e., Class I or Class II obesity). Finally, tertiary prevention targets the morbidly obese (i.e., Class III obesity), who are at the highest risk for complications associated with weight status. A similar approach can be appl ied to obesity related comorbidities (e.g., obesity treatment as primary prevention for hypertension and diabetes) (Kumanyika, 2007). In an earlier review, Kumanyika (2005) presents two additional prevention paradigms (i.e., in addition to the primary seco ndary tertiary prevention continuum). The first is a health promotion or population health approach, which has the goal of shifting the entire BMI distribution, thereby lowering the mean (Orleans, Gruman, Ulmer, Emont, & Hollendonner, 1999). This approac h actually de emphasizes the role of primary care in obesity prevention, suggesting that weight reduction interventions, particularly those promoting healthy eating and physical activity, can be applied universally and more cost effectively without individ ual screening and counseling (Kumanyika, 2007). Nonetheless, individual treatment remains the dominant paradigm. One approach inherent to individual treatment is behavioral counseling. Independent of the goal (prevention versus treatment), behavioral cou nseling in the primary care setting has the potential to target small behavior changes (e.g., increasing physical activity, improving the quality of diet, reducing caloric intake) that can presumably be maintained over the long run, monitored, and adjusted appropriately as determined in a collaborative manner by the primary care provider and the patient continuum of action model, wherein the emphasis is on individually ori ented and

PAGE 27

27 curative strategies involving behavioral counseling and education (Kumanyika, 2007). The curative nature of these strategies has also been identified as a drawback insofar (Katz & Faridi, 2007, p. 291). Further, these strategies do not allow for introduction and management skills (i.e., problem solving, focus on self monitoring and efficacy, application of skills to real life situat ions) or provider use of reinforcement and behavioral contracting, which have demonstrated efficacy in obesity management (Williamson & Perrin, 1996). Other related frameworks for understanding the complexity of weight management in the primary care settin g include the Transtheoretical Model of Behavior Change, which is often referred to as the Stages of Change Model (Prochaska & DiClemente, 1982) and the Health Beliefs Model (Glanz, Rimer, & Lewis, 2002) The Stages of Change Model illustrates the behavio ral stages that precede health behavior change. These stages are precontemplation, contemplation, preparation, action, and maintenance. They have been applied to a number of behavioral weight loss interventions and models of weight reduction, diet, and e xercise counseling (e.g., Campbell et al., 1994; Kreuter et al., 1997; Macqueen, Brynes, & Frost, 1999; Pinto, Goldstein, & Marcus, 1998; Sutton et al., 2003). Frameworks for understanding health communication and the patient provider r elationship are dis cussed in later sections. Trends in Weight Management Counseling in Primary Care Despite the growing role of primary care in obesity management, much of the research in this area has shown that physicians are missing the mark in terms of the guidelines fo r identification, evaluation and treatment of obesity in their overweight and obese patients. In fact, trends in weight management counseling in the primary care

PAGE 28

28 setting have not changed significantly since the introduction of clinical guidelines for obes ity management (Katz & Faridi, 2007; Jackson, Doescher, Saver, & Hart, 2005). Investigators examining the trends in weight management counseling generally cite a lack of attention to obesity in primary care settings (e.g., Abid et al., 2005; Boardley, She rman, Ambrosetti, & Lewis, 2007; Ely et al., 2006; Flocke, Clark, Schlessman, & Pomiecko, 2005; Huang et al., 2004; Ma, Urizar, Alehegn, & Stafford, 2004; Nawaz, Adams, & Katz, 1999; Nawaz, Katz, & Adams, 2000; Sciamanna, Tate, Lang, & Wing, 2000; Scott et al., 2004; Stafford et al., 2000). in these national and local investigations evaluating counseling for weight reduction and control ranged from 5.8% (lowest) to 48.9% (highest; not including examples from physici an report, which appear to be inflated). Even when the primary reason for the medical encounter was a comorbid condition, rates of counseling for weight loss among obese patients were low (e.g., Galuska, Will, Serdula, & Ford, 1999). The Appendix shows t he rates of weight reduction, diet/nutrition, and physical activity counseling across a number of studies over the last decade. Further, trends in weight management counseling in primary care have been evaluated in a variety of contexts: documentation of obesity during primary care encounters and counseling calling for diet/nutritional changes, increases in physical activity, or use of pharmacologic agents (e.g., appetite suppressants). Finally, the most important contribution to this body of literature ( discussed in detail in a later section) includes empirical support for weight management counseling in the primary care setting (e.g., Appel et al., 2011; Ely et al., 2008; Ryan et al., 2010; Wadden et al., 2011).

PAGE 29

29 Factors related to identification and eva luation of obesity I mplementation of weight management counseling in primary care first requires appropriate identification and evaluation of obesity in primary care patients. Body mass index (BMI) and waist circumference have previously been establishe d as straightforward, easily obtained, cost effective, and validated measures of excess body weight in the clinical setting (Hill & Wyatt, 2002; US Preventive Services Task Force, 2003). National data suggest that rates of measurement of height and weight are suboptimal (i.e., 42.0% for height, 65.0% for weight, and 41.0% for both height and weight) (Ma et al., 2009a), and in some cases inaccurate (Greenwood, Narus, Leiser, & Egger, 2011). In the Ma and colleagues study (2009a), of the total visits where BMI was obtained, only 2 9.0% of cases in which obesity wa s diagnosable had a documented diagnosis of obesity, which is consistent with previous findings regarding poor documentation of obesity (e.g., Bardia, Holtan, Slezak, & Thompson, 2007; Greenwood et a l., 2011; Huang et al., 2004; Ma et al., 2009b; Stafford et al., 2000). However, the proportion of visits during which obesity was diagnosed did increase with weight status (19.0% for BMI = 30.0 34.9; 32.0% for BMI = 35.0 39.9; and 50.0% for BMI > 40.0) ( Ma et al., 2009a), suggesting that physicians reporting may be related to perception of risks associated with graduated increases in obesity status. In addition, there were <2.0% of cases where obesity was misdiagnosed for adults with BMI < 30.0; however, it is not known what proportion of these adults fell in the overweight, normal weight, or underweight categories. This trend was also observed by Boardley and colleagues (2007), as well as Bardia and colleagues (2007), who noted that in addition to highe r BMI, patients with diabetes and obstructive sleep apnea were also more likely to have an obesity diagnosis documented. This is

PAGE 30

30 understandable considering obesity management is the cornerstone therapy for the treatment of diabetes and obstructive sleep a pnea. Despite these findings, inaccuracies in clinical measurement of overweight and obesity are common. For example, Greenwood and colleagues (2011) found that while greater 80% of adult patients had BMI recorded in the electronic medical records availa ble in this multi site study, only 35.4% of the records were accurate based on the study protocol. Following identification and diagnosis of obesity and associated disorders, clinical adiness to change and development and documentation of a weight management plan to include not only diet but also moderate intensity physical activity (Donnelly et al., 2009; Hill & Wyatt, 2002; Lyznicki, Young, Riggs, & Davis, 2001; US Preventive Services Task Force, 2003 ). This includes evaluation of motivation and expectations for weight loss, weight loss history, understanding of personal risks and benefits for weight loss, attitudes toward diet and physical activity, and personal barriers to weight lo ss such as social supports, time restraints, and lack of resources. The available evidence suggests that physicians do not typically engage patients in a discussion of readiness to change even when a weight management plan is documented (e.g., Huang et al ., 2004; Scott et al., 2004) To provide a concrete example, only 2%, 9%, and 10% of encounters including weight loss, diet, or physical activity counseling, respectively, included an assessment of readiness to change (Flocke et al. 2005 ). Further, poor documentation of obesity reduces the likelihood of development of weight management plan or receipt of advice, counseling, or education regarding weight, diet, or exercise, suggesting that

PAGE 31

31 identification of obesity using clinical guidelines may serve as a cue to provide counseling in these domains ( Bardia et al., 2007; Boardley et al., 2007) Diet and exercise Di et and exercise with the goal of creating a negative energy balance are key components of weight management. However, rates of counseling speci fic to these strategies are low. Physicians are more likely to provide education and counseling regarding diet than exercise (e.g., Boardley et al., 2007 ; Ma et al., 2004; Stafford et al. 2000 ). This is likely because caloric restriction is deemed the m ore important part of the energy balance equation (i.e., compared to strategies such as modifications in specific macronutrients). However, exercise has more beneficial effects on long term weight maintenance than initial weight loss (Donnelly et al., 200 9; NHLBI, 1998). While some studies show similar rates of diet and physical activity advice (e.g., Kreuter et al., 1997) or higher rates of physical activity than diet counseling (e.g., 45.0% versus 31.0% in Flocke et al., 2005; and 57% versus 49% in Agen cy for Healthcare Research and Quality (AHRQ), 2012) the relatively low rates of both types of advice are still significant given t he growing problem of obesity. Pharmacological treatments of obesity From the perspective of a chronic disease model, phar macologic treatments may not be the preferred line of therapy for long term management of obesity. Long term (i.e., > 2 years) effectiveness of anti obesity medications have not been established (NHLBI, 2000; Rucker, Padwal, Li, Curioni, & Lau, 2007 ; Yano vski & Yanovski, 2002). Though the same could be said of diet and exercise interventions, physicians appear to be even more reluctant to prescribe weight loss medications for long term use (Foster et al., 2003). Some factors that contribute to this reluc

PAGE 32

32 costs for their patients (Bray, 2002). This may also represent a cautious response to the recent controversy following recall of several popular anti obesity medications such fenfluramine and phentermine) secondary to serious medical consequences (e.g., Wadden et al., 1998). Following recall of sibutramine and t additional anti obesity medications ( lorcaserin and a combination of phentermine and topiramate ) in 2010 (Wright & Aronne, 2011), orlistat is currently the only drug approved by the FDA for long term use (Joy al, 2004; Bray, 2002). Based on the observed literature, one study in this line of research included prescription of anti obesity medications as a dependent variable in an evaluation of related practic es (Stafford et al. 2000). Here, the primary aim was to determine the association between risk status and provision of obesity related counseling or treatment. Results indicated that persons with higher health risks were least likely to be prescribed an ti obesity medication (0.6% very high risk versus 6.3% low risk). Another study included patients who had a documented prescription for a weight loss drug among other drug classes (i.e., for hypertension, hyperlipidemia, or type 2 diabetes) in its sample; however, this was used as an indicator of the presence of associated diseases rather than as an alternate form of weight reduction counseling (Heaton & Frede, 2003). In yet another example, investigators assessing the frequency of 25 weight management st rategies offered as

PAGE 33

33 lari, Darroch, & Wing, 2009). Further, thirty five percent of a sample of obstetrician gyneco logists in the United States reported that they had prescribed weight loss to their obese patients. However, when evaluated from a patient perspective, only 1% of patients recalled ever being offered education or advice regarding pharmacologic options for weight loss ( Huang et al., 2004) Notably, this 1% also included education regarding surgical interventions. According to the literature, there were no studies evaluating racial differences in obesity medications However, data from studies of pain management, HIV/AIDS and cancer treatment, and psychological treatment show that physicians tend to underprescribe to minority patients, especially African Americans ( Baker, 2003; Bogart, Catz, Kelly, & Benotsch, 2001; King, Wong, Shapiro, Landon, & Cunningham, 2004; Opolka, Rascati, Brown, & Gibson, 2004) obesity medication. Taken together, these data represent little consideration and application of a nti obesity medication for management of obesity across a variety of primary care settings. Race, Rurality, and Regional Contexts Across the studies reviewed, a variety of factors have been examined as possible predictors of weight management counseling, i ncluding age, sex, race, socioeconomic status (i.e., income), education level, marital status, degree of obesity, perception of health and actual health status (i.e., presence of comorbidities or other risk factors), among others. The findings are complex and inconsistent making it difficult to draw any substantive conclusions about what factors definitively predict weight related counseling in the primary care setting. Related to the primary variables

PAGE 34

34 of interest only a few studies have considered the r acial, rural urban or geographic variations in receipt of obesity related counseling (e.g., Centers for Disease Control and Prevention (CDC), 1998; Ely et al., 2006; McAlpine & Wilson, 2007). Specifically, research has yet to clearly elucidate how race an d rurality work in concert within this context. It is unclear whether race plays a role in identification of obese in the primary care setting. In some studies, Black patients were more likely than their Caucasian and Hispanic counterparts to be identifie d as obese (e.g., Stafford et al. 2000) In another study (Ferraro & Holland, 2002), Black women were less likely to be classified as obese when compared to white women with comparable anthropomorphic measurements. This is particularly problematic given the evidence regarding poor documentation of obesity and rates of weight related counseling. Significant racial differences do not emerge consistently in rates of counseling across a number of studies (e.g., Jackson, Doescher, Saver et al., 2005; Ma et a l., 2009b ; Sciamanna et al., 2000; Stafford et al., 2000) However, some studies do show significant racial differences in counseling that favor minorities (e.g., Loureiro & Nayga, 2006). Ma and colleagues (2004) reported some racial differences in rates of diet but not physical activity counseling among a sample of patients at risk for cardiovascular disease. Non Hispanic Blacks, Hispanics, and Asian/Pacific Islanders were more likely to receive diet counseling than their Caucasian counterparts (odds ra tios were 1.2, 1.4, and 1.7 times higher, respectively, as compared to the Caucasian reference group). It should be noted, however, that these analyses did not account for patient weight status using BMI, socioeconomic status, or education, as these data were not collected as part of the national surveys

PAGE 35

35 evaluated in this study (National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS)). Similar associations have been observed for rural urban differen ces, as well as regional differences. McAlpine and Wilson (2007) evaluated the association between region and counseling across five time points (i.e., 1995 1996, 1997 1998, 1999 2000, 2001 2002, and 2003 2004). Data showed that a greater proportion of p rimary care visits occurred in the South (e.g., 37.1% of visits in 2003 2004) than any other region. Region was found to be associated with counseling, with visits in the South and Midwest less likely to include weight reduction, diet/nutrition, or exerci se counseling than visits in the Northeast. Further, there were significant negative time trends in the prevalence of all forms of counseling in the Northeast and Midwest across the study period (i.e., rates of counseling decreased significantly from 1995 to 2004). Similarly, Abid and colleagues (2005) and Galuska and colleagues (1999) noted that residents of the South and Midwest were significantly less likely to receive advice to lose weight than those in the Northeast. However, Loureiro and Nayga (200 6) observed respondents in the South and Northeast were more likely than respondents in the Midwest to receive advice to lose weight from their physicians. Other studies examining weight management in the primary care setting include a study (Ely et al., 2 006) of patients in rural Kansas primary care practices, of which management with their physicians. Fewer still reported receiving nutrition (38.0%) or physical activity ( 43.0%) counseling. Also, in an examination of preventive counseling for cardiovascular disease by geographic regions of the United States, Southern

PAGE 36

36 residents were least likely to receive either diet or physical activity participation advice (CDC, 1998). In a recent examination of trends in obesity related counseling, Ma and colleagues (2009b) included geographic region and metropolitan statistical area (i.e., rural urban status as a simple yes/no variable) as possible independent factors in a logistic reg ression model. Neither variable was found to be independently associated with receiving diet, exercise, or weight reduction counseling among patients with BMI > 30.0. Interestingly, region of practice was independently associated with greater chances of lacking complete measurements (i.e., for weight and height) to screen for obesity (for the Northeast and the West) or an obesity diagnosis (for the South). These factors, as discussed earlier, may have significant implications for the development of a wei ght management plan during a primary care encounter ( Ma et al., 2009b) Methodological Limitations One methodological challenge of several of these studies (e.g., Ely et al., 2006; Ma et al., 2009b), which may represent an underestimation of the frequency of weight management counseling in primary care, is the framing of questions related to the timing of weight related counseling. For example, for individuals who did not report receiving counseling during a particular medical encounter, few studies accoun ted for the possibility of having received counseling during a previous medical visit. Similarly, the lack of documentation of weight status does not rule out the possibility that weight management was discussed during a particular encounter (e.g., Bardia et al., 2007). Further, in cases where diet and exercise counseling were provided, there was little or no account of whether that counseling was directly related to a weight reduction goal. Another relevant issue to consider is the recall bias implicit in patient reports. The use of a cross sectional approach is also an important methodological limitation. Taken

PAGE 37

37 together, these methodological issues make it difficult to draw conclusions with regard to role of primary care providers in obesity treatment and management. Empirical Support for Weight Management Counseling in Primary Care Obesity management targeting sustained weight loss has yet to demonstrate efficacy (US Preventive Services Task Force, 2003). Only a few randomized control trials have eva luated the efficacy of weight management counseling in the primary care setting for producing significant weight change (e.g., Appel et al., 2011; Ashley et al., 2001; Ely et al., 2008 ; Kumanyika et al., 2012; Lewis & Lynch, 1993; Logue et al., 2005; Pinto Goldstein, Ashba, Sciamanna, & Jette, 2005; Ryan et al., 2010; Tsai et al., 2010; Wadden et al., 2005; Wadden et al., 2011). Others have focused on implementation of behavioral counseling for exercise, nutrition, and specific obesity related outcomes wi th (e.g., Lewis & Lynch, 1993; Ockene et al., 1999) and without (e.g., Calfas et al., 2002; Goldstein et al., 1999; King et al., 2006) specific attention to weight change. Results of several independent studies (i.e., Christian et al., 2008; Ely et al., 20 08; Kumanyika et al., 2012; Ryan et al., 2010) found that, although participants in the respective intervention arms lost more weight than those in the control arms, differences were not always statistically significant and intervention failed to produce c linically significant weight losses ( > 5% body weight) in most participants. The Think Health! study, a randomized control trial of a behavioral weight loss intervention (adapted from the Diabetes Prevention Program) conducted at five primary care practice s by physicians and auxiliary staff trained as lifestyle coaches, provides one example (Kumanyika et al., 2012). In this trial, a usual care condition (Basic) comprised of quarterly primary care visits with lifestyle counseling was compared to a treatment condition (Basic Plus), which consisted of quarterly primary care visits with counseling

PAGE 38

38 plus individual sessions with a lifestyle coach once per month. Duration of both study conditions was one year. Findings indicated one year weight changes of 1.61 kg for Basic Plus and 0.62 kg for Basic conditions, which did not represent a statistically significant difference ( p = 0.15). In addition, only 23% and 10% of group participants, respectively, lost > 5% of their baseline body weight. Another study, the L ouisiana Obese Subjects Study (LOSS), targeted intervention to extreme obese (BMI = 40 to 60) patients from eight sites (primarily primary care practices), who were randomized to the intensive medical intervention treatment condition (IMI) or usual care co ndition (UCC) (Ryan et al., 2010). Weight change was reported as percentage of weight lost from baseline, which for completers was 9.7% and 0.4% for IMI and UCC, respectively. Given the high attrition noted in the study, percent weight losses were also r eported in baseline observation carried forward and last observation carried forward analyses; however, results were relatively unchanged with IMI losing significantly more than the UCC. Still, only 31% and 9% of the IMI and UCC patients, respectively, ac hieved a 5% or more weight loss at two years, which may in part represent the issue of weight regain and maintenance. Notwithstanding the fact that the vast majority of obese patients in these studies do not successfully achieve or maintain weight losses sufficient to improve their health, these interventions continue to heighten attention to obesity among both participants and physicians, and in some cases offer a significant alternative to current standards of care (i.e., usual care), self directed we igh t loss, or no intervention. The Practice based Opportunities for Weight Reduction (POWER) Trials Collaborative Research Group, funded by the NHLBI, represents the most recent efforts

PAGE 39

3 9 to further research in this area through three comparative effectiveness trials of weight loss interventions for obese patients with cardiovascular risk factors and delivered in primary care settings (Yeh et al., 2010). Results of two of these trials are discussed presenting at one of six primary care practices compared a usual care control condition consisting of education provided at quarterly primary care physician office visits with two intervention groups. The interventions were the brief lifestyle counseling group consisting of usual care component plus monthly, in person counseling visits by trained medical assistants of 15 minute duration and an enhanced lifestyle counseling group consisting of usual care component plus monthly, in person counseling visits b y trained medical assistants of 15 minute duration with addition of meal replacements or weight loss medications. At 24 month follow up, participants in the enhanced lifestyle intervention lost significantly more weight (4.6 kg) than those in the brief li festyle counseling group (2.9 kg) and the usual care group (1.7 kg); however, the latter two groups did not differ significantly. Enhanced lifestyle intervention also resulted in clinically meaningful weight losses (i.e., at least 5% of initial body weigh t lost) in 35% of the enhanced lifestyle counseling participants versus 26% and 22% of the brief lifestyle counseling and usual care group participants, respectively. Appel and colleagues (2011) followed 415 obese patients from six primary care practices w ith at least one cardiovascular risk factor over the 2 year study period. The effectiveness of two behavioral weight loss interventions (remote support only and in person support plus remote support) was compared to a control condition consisting of a sel f directed weight loss (using educational material and resources provided at study

PAGE 40

40 initiation). The remote support condition consis ted of a commercial call center directed group in which office based lifestyle coaches delivered all lifestyle interventions by telephone, internet, and e mail; while the in person condition consisted of in person individual (20 minute) and group (90 minute) sessions plus electronic and telephone contacts delivered by office based lifestyle coaches. Both interventions also inc luded contact with primary care physicians who supported the delivery of the interventions, reviewed participants' weight status, and encouraged engagement (or reengagement) in the study at routine medical visits. At 24 month follow up, participants in bo th the in person and remote support interventions had similar weight losses ( 5.1 kg and 4.5 kg, respectively), which was significantly greater than the weight loss achieved through self directed weight loss (control group, 0.8 kg). In addition, partici pants in the intervention arms were significantly more likely to achieve modest losses of at least 5% of their initial body weight: 41% and 38% of the in person and remote support groups, respectively, versus 19% of the control group. The issue of non sig nificant clinical results may be understood within the context of continuity of care. Repeated communication regarding weight and risk reduction via behavioral activation appears to be necessary to improve adherence, minimize barriers, and improve outcome s (e.g., Katz & Faridi, 2007; Logue et al., 2005). Specifically, patients who are encouraged and monitored with regard to physical activity, diet, and self monitoring of weight and caloric intake appear to have more success in achieving sustainable behavi or change and weight loss. However, one of the greatest limitations to many of the aforementioned trials was attendance, which continues to affect the feasibility and effectiveness of office based behavioral interventions. In addition,

PAGE 41

41 changes in health care delivery systems, particularly with respect to reimbursement, would be required to successfully implement these protocols in primary care settings. Despite the limited evidence for the effectiveness of physician counseling on weight control, studies e valuating the effect of counseling and training specific to weight do show promise. For example, implementation of an obesity counseling skill s training program showed that compared to physicians in a control condition, training improved the frequency and quality of counseling provided to obese patients (Simkin Silverman & Wing, 1997). Specifically, training significantly increased rates of counseling (47% at baseline to 89% after intervention), whereas rates remained stable among those who did not receiv e training (42% at both baseline and after the intervention). Also, p hysician advice in the form of weight management counseling (e.g., weight reduction, diet, or physical activity counseling) has been identified as an important predictor of behavior chan ge (Kreuter, Chheda, & Bull, 2000). Weight reduction counseling increases the likelihood of weight loss attempts (Abid et al., 2005; Bish et al., 2005; Sciamanna et al., 2000; Galuska et al., 1999). For instance, Abid and colleagues (2005) observed a con siderable difference in the prevalence of attempts to lose weight provider (79.8%) compared to those who had not received such advice (58.6%). Of note, among those who were advised and attempted weight loss, only 56% engaged in the recommended strategies for successful weight loss (i.e., diet and exercise) (Galuska in regard to the behavi ors most likely to result in an energy balance conducive to weight loss.

PAGE 42

42 weight on specific behavioral outcomes such as changes in dietary intake and physical activity. Pinto and colleagues (2005) and Lewis and Lynch (1993) reported that should be noted that the increases in physical activity in the former study included extended counseling outside of the primary care setting Loureiro and Nayga (2006) fewer calories and fat to lose weight and engaging in physical activity to lose weight. The tendency to u tilize these strategies increased with counseling. Additionally, weight reduction counseling increases awareness and understanding of health risks (Ely et al., 2008; Huang et al., 2004) and increases the likelihood that individuals will consequently respo nd to cues to engage in weight management (Kumanyika, 2007). This may be particularly important for minorities and rural residents, who are disproportionately affected by obesity and obesity related disease. However, even fewer studies have evaluated the effectiveness of these primary care based interventions for minority groups (e.g., Kumanyika et al., 2012; Martin et al., 2008; Martin et al., 2006). Factors Influencing Physician Practices and Patient Provider Communication Factors such as time, resource s (i.e., validated tools and materials), training regarding delivering obesity interventions, and availability of brief and effective counseling techniques have been identified as barriers to successful obesity treatment in primary care (e.g., Bowerman et al., 2001; Katz & Faridi, 2007; Kushner, 1995; Simkin practices are not well understood. This is particularly true for rural communities (Ely et al., 2006). The 2002 Institut e of Medicine (IOM) report draws attention to patient and

PAGE 43

43 system level factors that affect quality and equality of care, particularly among racial and ethnic minorities (Smedley, Stith, & Nelson, 2002). Specifically, the IOM argues that factors such as th e inability to gather a relevant medical history, cultural insensitivity, and specificity in assessing symptoms of disease effect treatment of chronic medical conditions. Moreover, issues of patient adherence to medical recommendations play a significant role in assessment and diagnosis. In addition, provider attitudes regarding the ineffectiveness of behavioral counseling as a first line of intervention for obesity management may affect rates of use (Foster et al., 2003; Katz & Faridi, 2007; Lyznicki et al., 2001). Thus, the literature on patient provider communication may help to elucidate factors that facilitate effective communication and discussion of obesity management in the primary care setting (e.g., Durant, Bartman, Person, Collins, & Austin, 20 09; Glasgow, Eakin, Fisher, Bacak, & Brownson, 2001; Irving & Dickson, 2004; Makoul, 2003; Vahabi, 2007 ). Patient Factors A variety of patient factors are known to influence patient provider communication regarding weight reduction, diet, and exercise coun seling in the primary care setting. Overweight and obese persons access and utilize healthcare services to a greater extent than all other patient groups. However, much of the care received in this context is for the acute care of related disorders (e.g. hypertension, type 2 diabetes, hyperlipidemia). Despite the demand to address these issues first, patients appear to want weight management advice from their physicians, including specific directives on how to achieve weight loss goals and support in do ing so (e.g., Potter, Vu, & Croughan Minihane, 2001). For patients who are ready to change, the manner in which obesity is

PAGE 44

44 willingness to engage in weight control efforts (An derson & Wadden, 1999; Serdula, Khan, & Dietz, 2003). This is particularly true for African American patients who reported that doctors typically do not dedicate substantial time to discuss weight or readiness to change. Doctors also often neglected disc ussions of their specific expectations for weight loss or the specific health consequences to the individual patient (War d, Gray, & Paranjape, 2009). Physicians have a significant influence on the discussion of weight in the primary care context. In fact, research has shown that physicians miss opportunities to initiate this process even when the need is obvious or there are valid reasons to address weight (i.e., obesity is an exacerbating factor of a comorbid disease) ( Scott et al., 2004) Patients repor only in the context of related medical conditions. The strategy appears to be effective; however, it remains a sensitive subject for both patients and providers, who as demonstrat ed by Scott and colleagues (2004), may avoid discussions related to weight altogether. In one study that examined patient and physician perceptions of weight agreement on p. 366). Some variables assessed were motivation for weight loss, patient comfort and preferences regarding discussion of weight, and frequency of weight loss discussions in t he primary care setting. Overall, patients in this rural primary care setting felt the discussion of weight control was delayed and infrequent. The cultural context is relevant to the discussion of delimiting factors for weight related counseling at the i ndividual patient level. Race and ethnicity are salient

PAGE 45

45 concerns, particularly for non Hispanic Blacks and Hispanics, for whom racial and ethnic differences act as barriers to partnership and effective communication in the medical encounter (e.g., Cooper Patrick et al., 1999; Kaplan, Gandek, Greenfield, Rogers, & Ware, 1995; Saha, Komaromy, Koepsell, & Bindman, 1999 ). Overweight and obesity in African Americans pose a unique challenge for primary care physicians, particularly in regards to the role of hea lth beliefs and participation in health decision making. A number of studies have linked race concordance with patient satisfaction and engagement in care, as well as continuity of care (e.g., Cooper Patrick et al., 1999). Of note, these associations are evident independent of patient centered communication, suggesting that factors such as patient and physician beliefs or attitudes may mediate these relationships (Cooper et al., 2003). For instance, despite the greater prevalence of obesity and obesity re lated comorbidities among African Americans (both actual and self reported), African Americans generally have a more optimistic view of their overall health, as well as their body image (i.e., weight status) (Burroughs et al., 2008). This has significant implications regarding both primary constructs (i.e., perceived susceptibility, perceived severity) and secondary constructs (i.e., cues to action) of the Health Beliefs Model (Glanz et al., 2002). Similar concerns are noted for Hispanics. In one study ( Durant et al., 2009), both non Hispanic Blacks and Hispanics were less likely to perceive their weight as damaging to their health than their Caucasian counterparts. However, those who were subsequently identified by their physician as overweight were mor e likely to have a less optimistic perspective of their health. These findings have significant implications regarding the role of patient provider communication in

PAGE 46

46 influencing patient perceptions regarding the deleterious effects of obesity and their beh avior. Namely, failure to communicate effectively the risks associated with obesity status may hinder their ability to actively engage in decision making and self management behaviors necessary for weight control and management (Vahabi, 2007). Physician F actors Physicians have the responsibility to ensure that their patients are well informed regarding their overall health. Given the growing impact of obesity on the health of adults in the United States, discussion of weight status is a necessary componen t of the primary care encounter. However, more often than not, the literature points to explain the relatively low rates of counseling observed in the literature. Fo r instance, efficacy for and practice of healthy lifestyle has been related to lower rates of obesity prevention counseling (Abramson, Stein, Schaufele, Frates, & Rog an, 2000; Crawford et al., 2004 ). Other studies suggest thi s inverse relationship is bidirectional such that physicians who report success in managing their weight using a variety of strategies such as engaging in recommended levels of physical activity are more likely to counsel their patients to do the same (Abr amson et al., 2000; Phelan et counseling targeting weight loss appear to be a salient factor contributing to the low prevalence of related counseling. For example, in a s ample of physicians in Louisiana, an overwhelming majority reported the effectiveness of weight reduction (80%), diet (77%) and exercise (70%) counseling to be fair or poor (Martin, Rhode, Howe, & Brantley, 2003). Given these findings, rates of counseling reported by physicians seem to be inflated as compared to those reported by patients and from national cross

PAGE 47

47 sectional surveys such as the NAMCS. For example, in a recent study (Phelan et al., 2009) physicians reported that they addressed weight loss wit h 75.5% of their patients. Similarly, high rates of counseling were reported by obstetrician gynecologists, who reported counseling 80% of all their patients about weight control (Power Cogswell, & Schulkin, 2006). Phelan and colleagues (2009) also descr ibed the criterion applied by most Physicians reported that an equivalent of 21.5% and 10.6% weight loss would be imilar study, Foster and colleagues respectively. This is significant given the consensus of national guidelines suggesting that 5 10% weight losses are typical, if not the most o ptimistic, outcomes during initial heuristics, or decision rules, to determine who to counsel regarding weight reduction (Kreuter et al., 1997). For example, a physician may opt to forego weight reduction counseling for an African American patient with diabetes and focus his/her efforts on medication adherence. This decision may be made in part because of assumptions about or experiences with African American patients. Thus a heuristic based approach may contribute to the lower rates of advising observed and possibly represent systematic exclusion of groups who would benefit from such counseling. Specifically, those who were classified as having therapeutic risks (i.e., ri sks related to existing comorbidities) were more likely to receive counseling than those for which prevention was suggested (42.9% versus 16.5% for diet advice and 35.2% versus 20.5% for

PAGE 48

48 physical activity advice, respectively). All in all, these findings suggest that not only may physicians have similar unrealistic expectations as their patients (e.g., Foster, Wadden, (Kreuter et al., 1997, p. 831). Implicit and W eight Bias in the Primary Care S etting There is now substantial evidence that discrimination against obese persons is widespread and occurs across a number of contexts and settings, including employment, education, adoption, media, interpersonal relationships, and healthcare (Andreyeva, Puhl, & Brownell, 2008; Puhl & Brownell, 2001; Puhl & Brownell, 2006; Puhl & Heuer, 2009). The latter is most relevant to the current review. In studies evaluating the attitudes, beliefs, and practices of physicians, obese pati ents were overwhelmingly described as lacking self control, lazy, noncompliant, unattractive, weak willed, and dishonest, among other negative behavioral and character descriptions (Foster et al., 2003; Friedman, 2008; Price, Desmond, Krol, Snyder & O'Con nell 1987; Schwartz, Chambliss, Brownell, Blair, & Billington, 2003). In some instances physicians judgments based on weight included beliefs that heavier patients were less healthy, worse at taking care of themselves, and less likely to comply with medi cal advice or to benefit from counseling (Hebl & Xu, 2001). Physicians and other change, reflecting the belief that overweight and obese patients may be reluctant to impl ement suggested behavioral strategies for weight reduction (Sussman, Williams,

PAGE 49

49 reporte perceptions of their motivation for weight management (Befort et al., 2006). also play a role in the quality and type of healthcare patients receive. In fact, literature on the role of provider bias in health disparities suggests that physicians use information and to mak e medical decisions when thorough clinical judgments are necessary (Burgess, Fu, & van Ryn, 2004, p. 1155). In regard to weight control, educated patients may be perceived as more likely to undertake weight loss than their less educated counterparts, as d ifferential rates of counseling to these groups sometimes suggest (e.g., Abid et al., 2005; Jackson, Doescher, Saver, et al., 2005; Loureiro & Nayga, 2006; Sciamanna et al., 2000). Similar findings are noted by the income status of obese patients. Taken together, these factors do appear to negatively making, including whether to broach the subject of weight during medical encounters, the duration of clinical contact with obese versus non obese patients, and the amount o f time devoted to health education (Bertakis & satisfaction with care (Fong, Bertakis, & Franks, 2006; Hebl et al., 2003; Wee et al., 2002). The results regarding patient sa tisfaction are complex, likely owing to inconsistent findings and methodological variations across studies. For example, Wadden and and obesity specific healthcare and fou nd that patients reported lower overall

PAGE 50

50 satisfaction with their obesity specific healthcare as compared to overall healthcare. Another study found that obese patients demonstrated lower satisfaction with the care they received during their most recent med ical visit; however, adjusting for health status attenuated this finding (Wee et al., 2002). Further, no BMI differences were noted for overall quality of the provider or practice. In contrast, Fong and colleagues (2006) found that obese patients reporte d greater satisfaction with their providers than did their normal weight counterparts even after adjusting for health status. Finally, at least one investigation found no association between weight and positivity of care; however, a weight by gender inter action effect was observed, where differences in patient satisfaction with care were noted for women but not men (Hebl et al., 2003). Specifically, while overweight men did not perceive their care to be significantly worse than nonoverweight men, overweig ht women reported significantly more satisfaction with their care than did nonoverweight women. Race is also an important factor in how healthcare providers identify their patients and make medical decisions (e.g., Coeutaux et al., 2004; van Ryn & Burke, 2 000). Findings related to racial/ethnic differences in patient adherence to medical examination of hypertension control among rural and urban Caucasians and African Ameri cans, rural African Americans were most likely to be prescribed hypertension medication and to have a larger discrepancy between recommendation and use (i.e., adherence) (Mainous et al., 2004). Related to lifestyle interventions, African Americans have bee n noted to consume more unhealthful diets and have substantial difficulty adhering to dietary prescriptions

PAGE 51

51 than their Caucasian counterparts (Ard, Skinner, Chen, Aickin, & Svetkey 2005; Gary et al. 2004). They are also more likely to be perceived as le ss intelligent and less likely to adhere to physician advice than their Caucasian counterparts (van Ryn & Burke, 2000). Furthermore, regardless of whether advice to lose weight is provided, non Hispanic Blacks and Hispanics are less likely to use caloric restriction and reduction of fats as a behavioral strategy to lose weight (Loureiro & Naygo, 2006). Collectively, these findings may underlie implicit biases held by physicians regarding the effectiveness of their recommendations to select patient populat ions and potentially contribute to the disparities observed for preventive care (e.g., screening and behavioral interventions) (Schneider, Zasloversusky, & Epstein, 2002; Taira, Safran, Seto, Rogers, & Tarlov, 1997). While there are a variety of implicatio ns regarding the impact of negative attitudes towards obese persons, less attention is given to the combined impact of race and weight on stigmatization. Puhl and colleagues (2008) compared prevalence of weight/height discrimination to known rates of race and discrimination and found that weight/height discrimination among adults in the United States was relatively high (ranging from 4.9% for men to 10.3% for women) and increased as a function of weight status. Rates also increased when age, education, an d race were examined independently with perceived weight/height discrimination. Strikingly, African American men and woman reported significantly higher rates of weight/height discrimination (23.9% and 12.7%, respectively); however, in regression analyses race was not an independent predictor of weight/height discrimination for African Americans. Notwithstanding, Kumanyika suggests that stigmatization of obese persons of color,

PAGE 52

52 particularly Black women, is enhanced by the attention drawn to the racial di sparities in obesity (Kumanyika, 2007; Kumanyika, 2005). Aims and Hypotheses of the Current Study Aim #1: Effect of Race/Ethnicity and Rurality on Lifestyle Advice The primary aim of the proposed study was to examine the influence of race/ethnicity and rur lifestyle advice among overweight and obese adults. Lifestyle advice was evaluated in three dimensions: dietary advice, exercise advice, and combination (i.e., both diet and exercise). Inter action of race/ethnicity and rurality on lifestyle advice First, the study evaluated the combined effect of race and MSA status on overweight and obese advice It was hypothesized that race MSA group assignment w advice. Specifically, African Americans in both rural and urban settings would have lower odds than their Caucasian counterparts to receive dietary advice, exercise advice, and c ombination advice; whereas, rural Caucasians would have lower odds of receiving all forms of lifestyle counseling than their urban counterparts. Last, it was hypothesized that among African Americans there would be no rural urban differences for all types of advice. Main effect of race/ethnicity on lifestyle advice The study also evaluated the advice among overweight and obese adults. It was hypothesized that race/ethnicity would ha ve an independent influence on reports of counseling such that dietary, exercise, and combination advice would be less common (as reflected by lower odds ratios) among

PAGE 53

53 African American individuals (regardless of rurality) than in their Caucasian counterpar ts. Main effect of rurality on lifestyle advice Third, the current study evaluated the advice among overweight and obese adults. It is hypothesized that MSA status would be associated with lifes tyle counseling Namely, report of all forms of lifestyle advice would be less common (as reflected by lower odds ratios) among rural respondents than urban respondents. This was evaluated at the person the location of the primary care office). Aim #2: Effect of Obesity Related Comorbidities on Lifestyle Advice The secondary aim of this study was to examine the influence of obesity related pertension and advice for diet and/or exercise. The study evaluated the role of these priority conditions on lifestyle counseling independently, as well as by race MSA status. Comparison of lif estyle advice among obese and non obese respondents by presence of priority condition. The effect of obesity related comorbidities on advice was also examined. It was hypothesized that the presence of a priority condition(s) would significantly influence obese and non obese respondents. That is, primary care providers would be more likely to provide all forms of lifestyle counseling to patient s with priority conditions compared to those patients without priority conditions. Specifically, as compared to obese adults with priority conditions (reference group), non obese respondents with one or more priority

PAGE 54

54 conditions would have lower odds of re ceiving advice followed by obese respondents without a priority condition and non obese respondents without a priority condition. The rationale is that providers weigh risk first by presence of comorbidities followed by weight status. Number of priority c onditions. In addition, it was hypothesized that the odds of lifestyle counseling would increase with the number of priority conditions (with some variations based on weight status). Obese adults with all three priority conditions were assigned as the re ference group, which was compared to seven comparison groups (explicitly, obese adults with one condition, obese adults with two conditions, obese adults without a priority condition, non obese adults with three conditions, non obese adults with two condit ions, non obese adults with one condition, and non obese adults obese adults with one two, or three priority conditions) would have greater odds of receiving lifestyle advice groups (i.e., obese adults without comorbidity and non obese adults without comorbidity) would have lower odds of receiving lifestyle advice compared to obese adults with three priority conditions. Comparison of lifestyle advice among obese patients with and without obesity related comorbidities by race MSA status. In addition to the precedin g aim, the present study evaluated the impact of race MSA status on lifestyle advice among obese respondents with and without at least one priority condition. Our hypotheses mimicked those for our primary aim such that when stratified by race MSA status, obese

PAGE 55

55 African Americans in both rural and urban settings would have lower odds than their Caucasian counterparts to receive dietary, exercise, and combination advice regardless of whether an obesity related condition is present. Similarly, rural Caucasian s would have lower odds of receiving all forms of advice than their urban counterparts regardless of whether a priority condition is present. Aim #3: Provider Communication Related to Lifestyle Advice The tertiary aim of the present study evaluated provide r communication behaviors (as reported by the respondents) that might be related to the delivery of lifestyle advice perceptions of the quality (e.g., respectfulness of explain treatment related issues in a manner that is understandable to the patient) and duration of clinical contact (i.e., whether the patient perceived that the provide r spent enough time with them). Effect of provider communication on lifestyle advice Using a summative indicator of provider communication (i.e., pooling scores on each of four individual indicators of provider communication as a summary variable), the current study also examined the influence o f provider communication on lifestyle advice (controlling for race advice such that higher pr ovider communication scores would be related to greater odds of receiving all forms of lifestyle counseling and vice versa (i.e., poor provider communication would be related to lower odds of receiving lifestyle advice ). Influence of race MSA status on pro vider communication. Additionally, the current study examined the differences in perceptions of provider communication (i.e.,

PAGE 56

56 using the summary indicator) by race MSA status among obese adults. It was hypothesized that there would be a significant differ ence in perception of provider communication by race MSA status such that obese African Americans in both rural and urban settings would have less favorable perceptions of provider communication ( as indicated by the summary indicator of provider communicat ion behaviors ) than their Caucasian counterparts. While we did not expect to see rural urban differences among African Americans, we did expect to observe significant differences in perceptions of provider communication between rural and urban Caucasian r espondents with less favorable scores being reported by rural respondents. Exploratory Aims Using MEPS longitudinal data, exploratory aims examined changes in weight reduction outcomes (i.e., engaging in moderate vigorous physical activity and BMI change) as a function of receiving diet and/or exercise advice. First, we evaluated whether physical activity increased over time as a function of lifestyle advice for exercise among a subgroup of respondents. Second, the present study evaluated change in weight status (i.e., change in BMI from baseline to end of survey participation) as a function of receiving combination advice among a subgroup of respondents. As these aims were exploratory, no a priori hypotheses were offered.

PAGE 57

57 CHAPTER 2 METHODS Data Source The Medical Expenditures Panel Survey (MEPS) is a nationally representative annual survey of the US non institutionalized civilian population, their healthcare utilization and expenditures, health insurance coverage, and health status. Beginning in 1996 a nd sponsored by the Agency for Healthcare Research and Quality (AHRQ), MEPS utilizes an overlapping sample design to summarize information across a number of large scale, comprehensive surveys conducted over a two and a half year period ( Machlin, Yu, & Zod et, 2005 ). This includes a baseline interview and five follow up interviews, which enables both cross sectional and panel (i.e., longitudinal) data analyses. The MEPS is comprised of two major components: the Household Component (MEPS HC) and the Insuranc e Component (MEPS IC). MEPS HC was used for the current investigation. MEPS HC data represents a subsample of households pooled year consolidated data file of MEPS HC consists o f person unique identifiers, including demographic characteristics, geographic variables, and health conditions and status. Also, MEPS HC files contain information on healthcare utilization and expenditures, access to care, health insurance, and employment and income variables, as well as event level files (e.g., prescription medications, outpatient visits, office based medical provider visits) (AHRQ, n.d.). Of note, the MEPS HC is supplemented by data supplied by individua

PAGE 58

58 MPC). These are used to validate self reported data (i.e., person level data) regarding medical conditions and healthcare utilization. However, these data are not publicly available for reason s of confidentiality. Data from the MEPS MPC are not designed to produce national estimates and are not available for all household reported events (AHRQ, n.d.). The MEPS is uniquely designed to enable annual estimates using weight and variance estimation variables ( Machlin et al., 2005 ). Sampling of the NHIS data ensures national representation of the US non institutionalized civilian population, including representation of African Americans (i.e., non Hispanic Blacks) and Hispanics using oversampling an d weighting procedures (Cohen, DiGaetano, & Goksel, 1999). The complex sampling procedure of MEPS data also include utilization of stratification, clustering, multiple stages of selection, and d isproportionate sampling to account for the complexity of the study design (described in further detail by Cohen et al., 1999). In light of aims to determine the unique contribution of race/ethn icity and rurality to physician s advice related to lifestyle modification and potentially weight management, the current s tudy combined data from the 2007 and 2008 MEPS HC (sampled from the 2006 2007 and 2007 2008 NHIS data). The response rates for 2007 and 2008 full year files were 56.9% and 59.3%, respectively. Participants Participants were 41,838 adults (>17 years) who r eported receiving routine care during the sampling years 2007 and 2008. Respondents were grouped into one of six race MSA categories based on self identification as non Hispanic Caucasian, non Hispanic Black (i.e., African American), or Other race (i.e., Hispanic, Asians, Native Americans, and Pacific Islanders) as well as a urban or rural residence. Though Other

PAGE 59

59 categories were included for analytic purposes, all discussions centered on race MSA differences were limited to non Hispanic Caucasians and A frican Americans living in rural and urban areas only (urban Caucasians, n = 16,919; rural Caucasians, n = 4,453; urban African Americans, n = 6,254; and rural African Americans, n = 933). As previously described, MEPS respondents include non institutiona lized, US civilians in this study, only adults with complete demographic, independent or outcome variables data were in cluded in the estimation sample Where indicated, su bpopulations of interest were examined and relevant sample sizes are reported. A smaller sample taken from the MEPS Panel 12 Longitudinal Data File were used to address exploratory aims. Participants were 7,706 adults who reported receiving routine care d uring the sampling years 2007 and 2008 of MEPS and completed all measures assessing weight status (i.e., BMI), receipt of diet and exercise advice, and exercise habits. Procedures All survey items were captured using computer assisted personal interviewing technology, which enables collection of information about each household member, and builds on this information from interview to interview across the two and a half year study period or five rounds. Items are round specific (e.g., may be asked every rou nd or only in select rounds) and include unique person identifiers and survey administration variables, demographic variables, income and employment related variables, person level priority condition variables, health status and disability related variable s, access to care variables (including data on health insurance, utilization, expenditure, and source

PAGE 60

60 of payment), and weight and variance estimation variables. It should be noted that all data for a given household was reported by a single household desi gnee. Analytic files were created by selecting variables of interest, including survey administration items, weight and variance estimation variables, demographic variables, as well as dependent and predictor variables from consolidated full year files for 2007 and 2008. In addition, a number of variables were constructed (e.g., race MSA dummy variables) to fulfill analytic needs. Full year consolidated files were merged with corresponding Medical Conditions files, which were formatted to person level var iables for conditions of interest, including diabetes, hypertension, and hyperlipidemia. Merged files were then stacked to create an analytic file that could be used to run survey procedures. Dependent Variables Physician lifestyle counseling/advice. The primary outcomes of the current study were used to determine the differential odds of receipt of lifestyle advice among rural and urban overweight and obese persons. Lifestyle counseling was evaluated on three dimensions: dietary advice, exercise advice and combination advice (i.e., diet and exercise). MEPS HC included two variables that were used to define receipt of diet and exercise advice items was constructed to identify participants with affirmative responses on both dietary and exercise advice items.

PAGE 61

61 Provider communication. A summative indicator of provider communication was constructed by pooling respondent sc ores on each of four individual indicators assessing provider communication behaviors related to quality (e.g., respectfulness of related issues in a manner that is understandable to the patient) and duration of clinical contact (i.e., whether the patient perceived that the provider spent enough time with them). This variable summed the scores of each item, scored on a 4 point Likert scale, resulting in summary scores from 1 to 16. This summary variable was used as a predictor in a multiple regression model assessing the influence of provider communication on lifestyle counseling (controlling for race MSA status), as well as a dependent variable in a logistic regression model assessing its relati onship to race MSA. Physical activity participation. For the first exploratory aim, respondents were asked about their physical activity habits at baseline and at the end of survey moderate to vigorous physical activity three times per week was used as a measu re of exercise participation. Weight status and change in weight status. MEPS HC calculates body mass index (BMI, defined as weight in kg divided by height in m 2 ) from partic reported height and weight. Due to confidentiality concerns (i.e., to protect the identity of MEPS participants), weight and height data are not publicly available in MEPS public m into groups by weight status (i.e., obese, overweight, normal weight, etc. based on clinical guidelines). Also, change in weight status was assessed using BMI as a crude measure of weight and

PAGE 62

62 height. BMI change was calculated as the difference from bas eline to survey endpoint with negative values reflecting weight loss and positive values reflecting weight gain across the study period. Independent Variables A number of independent variables were used in the current analyses. In some cases these were us ed as predictor variables while in others they served as covariates. Age. Age is strongly associated with obesity and level of urbanization (i.e., rural urban status) (Befort et al., 2012; Lewis et al., 2000; Lucas et al., 2004) Age was measured using self reported date of birth. For 2007 and 2008 data, MEPS HC measured age as the difference in years from the last calendar day of the representative year (e.g., for 2007, 12/31/2007 ). Sex. Recent estimates suggest that the gender gap in rates of obesity for men and women is closing; however, factors such as race and rurality show different effects for men and women (Flegal et al, 2012; Lucas et al., 2004; Ogden et al., 2006). MEPS report as male or female Marital st atus. A commonly used sociodemographic factor and proxy for social support in behavioral research, marital status was measured in MEPS HC as married, separated, divorced, widowed, or single. Years of e ducation. The association between obesity and socioec onomic status (SES) is complicated (with notable differences noted in the literature by race and gender) and also highly dependent on the SES indicator used. When years of education are used as a proxy, education has been shown to be inversely related to obesity such that lower education is associated with higher rates of obesity (e.g., Mokdad et al., 2003; Zhang & Wang, 2004). Notably, for Black women, the prevalence

PAGE 63

63 of obesity and education is an inverted U shaped association in some datasets (i.e., 199 9 2000 National Health and Nutrition Examination Survey ( NHANES), Zhang & Wang, 2004), while it remained a simple inverse relationship in others (i.e., 2001 Behavioral Risk Factor Surveillance System (BRFSS), Mokdad et al., 2003). Thus, it is important to statistically control for variations in education status. MEPS HC measured educational attainment as self reported years of education (continuous) when respondent first enter ed MEPS. Annual household income. Similarly, income is a readily used indicator of SES. Regarding weight status, lower income has generally been associated with higher rates of obesity ( Zhang & Wang, 2004 ). However, while it appears that this association has weakened over time, particularly with regard to gender and race, African A merican women continue to demonstrate higher rates of obesity than Caucasian women regardless of SES. Annual household income was collected in MEPS HC using definitions of income and family developed by the Current Population Survey. Family income, or an nual household income, was derived using the sum of person level total income (excluding income from tax refunds and capital gains) for each member of a given household at the end of reference period (for 2007, 12/31/2007) and is also a continuous variable In some analyses income is represented as the percentage of poverty: poor, near poor low income, middle income, or high income. Insurance status. Regarding healthcare, the increasing healthcare costs attributable to obesity have been the focus of much research and policy initiatives (USDHHS, 2001). While the association between health insurance and body weight has been virtually neglected in the literature (likely due to its endogeneity or ambiguity

PAGE 64

64 relative to education, employment, and income), it i s an important consideration when evaluating health disparities and obesity. MEPS HC provides a measure of insurance type held as reported by survey respondents. Each respondent identified type of insurance held during reference period, which were catego rized into five groups as any private insurance, Medicare (included respondents covered jointly by Medicaid), Medicaid, uninsured, and other public assistance for the p urpose of this investigation. Race/ethnicity. Racial/ethnic differences in obesity have consistently been demonstrated in the literature with minorities having higher rates of obesity than their Caucasian counterparts (Flegal et al., 2012; Hedley et al., 2004; NCHS, 2009; Ogden et al., 2006) MEPS HC collects race and ethnicity information in several ways, including race alone, ethnicity alone, race/ethnicity as a single variable, and a number of additional variables created to account for respondents identifying multiple races. For the purpose of the current study, race and ethnicity were collapsed into three groups: non Hispanic white, non Hispanic African American, and Other, which includes individuals identifying as Hispanic regardless of race, Asians, Native Americans, and Pacific Islanders Metropolitan statistical area (MSA). Rural u rban differences in obesity are also well documented (e.g., Befort et al., 2012; Jackson, Doescher, Jerant et al., 2005; Patterson et al., 2004) with higher rates of obesity observed among rural adults. In the current study, MSA status was used as a proxy for rurality, which is defined by the United States Census Bureau (2005) based on population density. As confidentiality restrictions of the MEPS HC limited use of certain identifying variables, rural and urban residence was differentiated using metropol itan statistical area (MSA). This was

PAGE 65

65 defined using the classification scheme outlined by the Office of Management and Budget, which collapses US counties across the rural urban continuum as either MSA signations were based on total population and proximity to and integration with large cities (see Ricketts, Johnson Webb, & Taylor, 1998; US Department of Agriculture and Economic Research Services, 2007). We expected approximately a fifth of the US popul ation to be rural based on recent estimations (e.g., Larson, Machlin, Nixon, & Zodet, 2004; US Census Bureau, 2005). Race/rurality term. A race/rurality variable, referred to as race MSA, was constructed using the MEPS HC race/ethnicity variable and MSA s tatus variable, resulting in six race MSA categories: rural Caucasians, urban Caucasians, rural African Americans, urban African Americans, and urban and rural Other. Census region. The aforementioned reports of differences in obesity by level of urbaniza tion (i.e., obesity rates greater in rural, non metropolitan areas compared to urban areas) also show variations in this relationship by region of the country (Lucas et al., 2004). Therefore, where relevant, analyses accounted for differential effects of geographic region. MEPS HC designated region of the country based on Census region criteria (Northeast, Midwest, South, and West), zip code. Obesity related comorbidities or priority conditions. A number of conditions, inclu ding hypertension, hyperlipidem ia, and type 2 diabetes mellitus, have been associated with obesity (Gregg et al., 2005; Mokdad et al., 2003). Respondents with these conditions represent a subpopulation with increasing need for lifestyle counseling

PAGE 66

66 based o n their risk profiles. MEPS HC consisted of a number of variables to determine the presence of these diseases (e.g., priority conditions, conditions enumeration, and medical provider visits sections). For hypertension and diabetes, condition enumeration (as indicted by ICD 9 codes found in MEPS Medical Conditions files) was determined to be a valid measure for determining diagnosis. While for hyperlipidemia, priority whether they had ever been diagnosed with the condition by a doctor) was utilized. It should be noted that for hyperlipidemia, priority conditions were used in lieu of conditions enumeration due to observed discrepancies in prevalence rates attained usin g the latter method (i.e., for condition enumeration method, rates fell substantially below what would be exp ected in the adult population). In addition, a number of variables were constructed to evaluate our secondary aims. These included yes/no variable s used to identify comparison groups by weight (obese versus non obese adults) and obesity related comorbidity status (i.e., presence of priority condition in one analysis and number of priority conditions in another). Variables were also constructed for adults with and without at least one priority condition for each of the six race MSA groups. Statistical Analyses Preliminary analyses using t test and analysis of variance (ANOVA) procedures were carried out to determine whether race MSA groups differed w ith respect to baseline demographic (e.g., age, BMI, education) characteristics. Those factors for which differences were noted were included as covariates in the subsequent analyses.

PAGE 67

67 Aim #1: Effect of Race/Ethnicity and Rurality on Lifestyle Advice The p rimary aim was to examine the role of race/ethnicity and rurality (i.e., MSA advice among overweight and obese adults. Given the dichotomous nature of the dependent variables, we conducted a series of logistica l regression analyses. Logistic regression analyses allow for evaluation of the relative effects of race MSA status on provision of diet, exercise and combination (i.e., diet and exercise) advice. These logistic regressions evaluated the odds of reportin g receipt of dietary, exercise, and both diet and exercise advice for each of the race MSA status groups. Results were interpreted using odds ratios (OR), which represent the change in odds as a result of a unit change in the predictor. Here, the focus w ill be the odds of dependent variables (i.e., diet and/or exercise advice) occurring or not occurring in each of the race MSA status groups. An odds ratio of one suggests equal likelihood of the dependent variable for a given race MSA group as compared to reference group (i.e., urban Caucasians). That is, the event in question is equally likely to occur in both groups. An odds ratio value greater than one indicated higher odds of the outcome variable occurring in a particular group (i.e., rural Caucasian s, urban African Americans, rural African Americans) than the odds of the outcome variable occurring in the reference group (i.e., urban Caucasians). Also, post hoc analyses (t tests) were conducted to examine the differences between specific race MSA gro ups for each form of lifestyle counseling. Logistic regression analyses were also used to evaluate the independent role of advice among overweight and obese adults. Separate logistic regression analyses were conducted for dietary, exercise, and combination advice to determine the predictive capacity of race. Similarly, the

PAGE 68

68 advice was evaluated using MSA status as predictor in logistical regression analyses for each form of lifestyle advice Aim #2: Effect of Obesity Related Comorb idities on Lifestyle Advice For the secondary aims examining the influence of obesity related comorbidities advice logistic regression models were used to test the effect of having at least one or none of the priority conditions among obese and non obese adults. Additional logistic regression models assessed the differences in receiving lifestyle advice using the com bination of number of priority conditions present and weight status as predictors, as well as the combination of race MSA and co morbidity status as predictors. Aim #3: Provider Behaviors Related to Lifestyle Advice Tertiary aims evaluated provider communic ation behaviors (as reported by the respondents) that may be related to the receipt of lifestyle advice Provider communication was examined as a predictor in one logistic regression model and as an outcome in a multiple regression model consi dering race MSA as predictors. Exploratory Aims The exploratory aims allowed for assessment of changes in weight related behavior change or outcomes using a longitudinal design. First, an examination of the differential impact of receiving exercise advice on physical activity habits of respondents was analyzed using a logistic regression model. Second, the effect of receipt of diet and exercise (i.e., combination) advice on changes in BMI was examined using a logistic regression analyses and confirmed in an analysis of covariance (ANCOVA),

PAGE 69

69 including the following covariates: age, sex, years of education, annual family income, insurance status, race, MSA status, region, and baseline BMI Data were analyzed at the person level for main aims and at the longitudinal level for exploratory aims. Data were weighted appropriately to produce national estimates. All of the above analyses were conducted using the survey procedures of Stata 10.1 statistical software (StataCorp, 2009).

PAGE 70

70 CHAPTER 3 RESULTS Overview The results of the current study are presented as follows. First, the demographic and socioeconomic characteristics of the sample are described. Respondents were grouped into one of six rac e MSA categories based on self identification as non Hispanic Caucasian, non His panic Black (i.e., African Americ an), or Other race, as we ll as urban or rural residence. For the purpose of the present study, results are presented for non Hispanic Caucasian and African Americans living in rural and urban areas only (Other racial cat egory omitted). All sections are organized by the following outcome measures: diet advice, exercise advice, and combination (i.e., diet and exercise advice) and reported, where relevant, by subgroup of interest (e.g., overweight adults, obese adults). Re levant sample sizes are reported in each corresponding section. First, the logistic regression analyses for the differential impact of race MSA, race/ethnicity, and MSA status (i.e., rurality) on overweight and obese adults are presented. Next, results of logistic regression analyses evaluating the differential impact of having versus not having at least one obesity related comorbidity (i.e., priority condition) for both obese and non obese respondents are presented. In addition, the differences in number of obesity related conditions from 0 3 are reported, as well as a model including dummy variables for race MSA by comorbidity status as predictors. Third, the results of logistical regression analyses for the differential impact of provider communication (as indicated by a summary score) on each form of lifestyle advice are described. This tertiary aim also includes a multiple regression analyses examining the role of race MSA on provider communication among obese adults Finally, exploratory

PAGE 71

71 analyses w ere performed to examine changes in respondent behavior, namely exercise habits and weight change, as a function of receiving diet and/or exercise advice. Following a descriptive summary of this sample, results of analyses examining the differential impac t of receiving exercise advice on physical activity habits of respondents (logistic regression) and receipt of combination advice on change in BMI (logistic commands of Stata 10.1 (StataCorp, 2009) were employed for all statistical inference tests undertaken in the present investigation. Baseline Characteristics Sample characteristics are summarized in Table 3 1 for respondents meeting eligibility criteria for the present study. Th is included a total of 41,838 adults who reported receiving routine care during the sampling years 2007 and 2008 of the Medical Expenditure Panel Survey. As indicated by our statistical plan, adults whose race/ethnicity was identified by Other are include d in analyses for estimation purposes but not described herein. Table 3 1 displays the demographic characteristics of the urban Caucasian ( n = 16,919), rural Caucasian ( n = 4,453), urban African American ( n = 6,254), and rural African American ( n = 933) r espondents. Overall, race MSA groups varied significantly across all demographic characteristics, including age, sex, marital status, annual household income, education, insurance status, as well as weight status. Specifically, urban Caucasian participan ts (referent, M + SE 47.72 + 0.27 years) were significantly younger than their rural counterparts (rural Caucasians, M + SE 49.18 + 0.62 years, p < .05) and significantly older than both urban and rural African Americans ( M + SE 43.18 + 0.32 years, p < .001, and 44.11 + 1.05 years, p < .01, respectively). There were slightly more females represented in the urban and rural African American

PAGE 72

72 subgroups and fewer of these respondents were married as compared to urban Caucasians ( ps < .05). However, more rur al Caucasians were married than their urban counterparts ( p < .001). In addition, urban Caucasians ( M + SE 13.69 + 0.04 years) were generally more educated than all other race MSA subgroups ( M + SE rural Caucasians, 12.78 + 0.07; urban African Americans 12.75 + 0.06; and rural African Americans, 11.85 + 0.15 years; ps < .001). Similar findings were observed for socioeconomic status as indicated by annual household income and type of health insurance coverage ( ps < .001; see Table 3 1). Furthermore, ur ban Caucasians had lower BMIs ( M + SE 27.15 + 0.07 kg/m 2 ) than rural Caucasians, as well as urban and rural African Americans (in order, M + SE 28.11 + 0.15, 29.01 + 0.13, and 30.32 + 0.30 kg/m 2 ; ps < .001). While tabular presentation (Table 3 1) describ es the subgroups of interest, the full sample (i.e., including Other race categories) were majority middle aged ( M + SD 45.00 + 17.58 years), Caucasian (51.1%) women (53.6%), who were at least high school educated ( M + SD 12.50 + 3.14 years), married (54 .1%), of relatively high socioeconomic status ( M + SD average annual income of $60,072 + $53,552), and privately insured (51.7%). Additionally, the majority of participants resided in metropolitan areas (84.5%) with the largest proportion representing th e South (38.1%). Average BMI for the full sample was ( M + SD ) 27.79 + 6.25 kg/m 2 For obesity related comorbidities (also referred to as priority conditions), 4,057 (9.7%) individuals reported having diabetes; 10,315 (24.7%) reported having hypertension; and 11,639 (27.8%) reported having hyperlipidemia. With respect to variables related to lifestyle advice, 14,271 (34.1%) individuals reported being advised to diet (i.e., restrict high fat and high

PAGE 73

73 cholesterol foods), 15,822 (37.8%) reported being advise d to exercise, and 12,052 (28.8%) reported receiving both diet and exercise advice (i.e., combination advice) from a doctor. Provider communication scores (N = 24,529) ranged from 4 to 16 points with a mean of 13.81 + 2.47 points. The aforementioned medi cal comorbidities and types of lifestyle advice received are summarized by race MSA in Table 3 2. Primary Aim: Effect of Race/Ethnicity and Rurality on Lifestyle Advice The effect of race MSA, race/ethnicity, and MSA on receipt of lifestyle advice was exam ined among both overweight (BMI = 25.0 29.9 kg/m 2 ; n = 14,004) and obese (BMI > 30 kg/m 2 ; n = 11,967) respondents. Race and rurality (race MSA) First, the current investigation examined the of receiving diet, exercise, and combination (i.e., diet and exercise) advice. Lifestyle advice for overweight adults. As summarized in Table 3 3, a logistic regression analysis showed that among a subgroup of overweight respondents, rural Caucasians ( OR = .68 p < .001 ) h ad 32 % lower odds of receiving diet advice than urban Caucasians. However, both urban African Americans ( OR = 1.13, p = .078) and rural African Americans ( OR = 1.02, p = .915) did not differ from urban Caucasians with respect to odds of receiving diet advice. Further, no group differences were observed for rural Caucasians as compared to rural African Americans ( OR = .67, p = .066) or for urban African Americans as compared to rural African Americans ( OR = 1.11, p = .615) in post hoc ana lyses. Parallel findings were observed for exercise advice Specifically, rural Caucasians ( OR = .68 p < .001 ) had 32 % lower odds of receiving exercise advice than urban Caucasians Again we observed no differences between urban or rural African

PAGE 74

74 America ns ( OR = 1.00, p = .980 and OR = .81, p = .223, respectively) as compared to their urban Caucasian counterparts for receipt of exercise advice. Follow up tests showed no differences between rural Caucasians and rural African Americans ( OR = .84, p = .398) or between urban and rural African Americans ( OR = 1.23, p = .232 ). Last, following the previous independent findings for diet and exercise, differences in combination advice were evident for rural Caucasians ( OR =.67 p < .001 ), who had 33 % lower odds of reporting that they received combination advice than urban Caucasians. Urban African Americans ( OR = 1.10, p = .187 ) and rural African Americans ( OR = .98, p = .917 ) did not differ from urban Caucasians for receipt of combination advice. Post hoc tests revealed no differences between rural Caucasians and rural African Americans ( OR = .68, p = .087) or between urban and rural African Americans for combination advice ( OR = 1.12, p = 550 ). Lifestyle advice for obese adults. The combined effect of race and rurality was also evaluated among obese respondents Similarly, rural Caucasians were significantly lower odds of receiving diet, exercise, and combination advice ( OR = .73, p < .00 1; OR = .68, p < .001; and OR = .73, p < .01, respectively) as compared to urban Caucasians (Table 3 4). This corresponds to 27 % of the odds for diet advice, 32 % of the odds for exercise advice, and 27 % of the odds for combination advice. In addition, obese rural African Americans had 24%, 43%, and 32 % lower odds of receiving diet, exercise, and combination advice from their doctors ( OR = .76, p < .05; OR = .57, p < .001; and OR = .68, p < .01, respectively) relative to urban Caucasians. However, urban African Americans did not differ from urban Caucasians for receipt of all forms of

PAGE 75

75 lifestyle advice ( diet advice OR = 1.05, p = .510; exercise advice OR = 1.02, p = .823; and combination advice OR = 1.09, p = .264). For post hoc tests, w hile no group differences were observed between rural Caucasian and rural African Americ an respondents for all types of advice ( diet advice OR = .96, p = .780; exercise advice OR = 1.19, p = .370; and combination advice OR = 1.07, p = .664 ), there were significant differences observed between urban and rural African Americans. Specificall y, urban African American respondents had greate r odds of receiving all types of advice than their rural counterparts For diet advice (OR = 1.38, p < .05) urban African Americans had 38% greater odds of receiving advice than rural African Americans Whi le for exercise advice ( OR = 1.79 p < .001 ) and combination advice (OR = 1.60, p < .01), they had 79% and 60% greater odds, respectively, of receiving advice compared to rural African Americans Race/ethnicity. Additional logistic regression analyses w ere conducted to evaluate the independent effect of race/ethnicity on receipt of diet, exercise, and combination advice. Among overweight respondents, African Americans ( OR = 1.16 p < .05 ) had 16% greater odds of report ing that they received diet advice than Caucasians (Table 3 5). However, these differences were not observed for exercise or combination advice ( OR = 1.01, p = 848 and OR = 1.13, p = .086, respectively ). Among obese respondents (Table 3 6), no significant differences were apparent for Af rican American respondents as compared to Caucasians for all types of advice ( diet advice OR = 1.05, p = .456; exercise advice OR = .99, p = .916; and combination, OR = 1.07, p = .342 ). Rurality. With respect to the independent effect of MSA status, log istic regression analyses showed that among the overweight subgroup, urban respondents had

PAGE 76

76 significantly greater odds of report ing recei pt of diet, exercise, and combination advice ( OR = 1.44, OR = 1.45, and OR = 1.45, all ps < .001, respectively) as compa red to rural respondents (Table 3 7). That is, the odds were 44%, 45%, and 45% greater for diet, exercise, and combination advice, respectively among urban respondents Similar findings emerged in the logistic regression analyses for the subsample of ob ese adults with urban respondents having significantly higher odds of receiving all forms of advice ( diet OR = 1.36 p < .001 ; exercise OR = 1.47 p < .001 ; and combination OR = 1.38 p < .001 ) than rural respondents (Table 3 8). This corresponds to 36 % greater odds for diet advice, 47% greater odds for exercise advice, and 38% greater odds for combination advice. Secondary Aim: Effect of Obesity Related Comorbidities on Lifestyle Advice Comparison of lifestyle advice among obese and non obese responde nts by presence of priority condition. The secondary aim of the current study undertook an examination of the role of obesity receiving diet, exercise, and combination advice. Explicitly, we aimed to determ ine whether the presence of at least one priority condition (i.e., hypertension, diabetes, lifestyle advice for adults (N = 41,838). Diet advice. Compared to obese r espondents with at least one comorbidity, obese respondents without a priority condition ( OR = .26 p < .001 ) had significantly l ower odds of report ing that they recei ved diet advice (Table 3 9). That is obese adults with a comorbidity had almost four ti mes greater odds of receiving diet advice. Non obese respondents with and without priority conditions ( OR = .72 p < .001 and OR = .08, p < .001, respectively) also had significantly lower odds of report ing that they receiv ed diet

PAGE 77

77 advice than obese adults with a priority condition. Explicitly, non obese adults with comorbidity had 28% lower odds of receiving diet advice than obese adults with comorbidity; while obese adults with comorbidity had 12.5 times the odds of receiving diet advice than non obese a dults without comorbidity. Exercise advice Similar findings were observed for exercise advice such that obese respondents without comorbidity ( OR = .38 p < .001 ), n on obese respondents with comorbidity ( OR = .51 p < .001 ), and non obese respondents without comorbidity ( OR = .11 p < .001 ) all had significantly lower odds of report ing that they receiv ed exercise advice than obese adults with a priority condition. Stated differently, obese adults with comorbidity had 2.6 times greater odds than health y obese adults to have received exercise advice. However, compared to non ob ese adults with comorbidity, obese adults with comorbidity had nearly two times the odds of receiving exercise advice; while they had 9.1 times the odds of receiving exercise adv ice compared to non obese adults without comorbidity. Combination advice A similar pattern was observed for combination advice. That is, obese adults with comorbidity had 3.6 times the odds compared to obese adults without comorbidity ( OR = .28, p < .001 ) and 12.5 times the odds compared to non obese without comorbidity ( OR = .08, p < .001) of receiving combination advice. However, non obese adults with comorbidity ( OR = .59, p < .001) had 41% lower odds of receiving combination advice compared to o bese adults with comorbidity. Post hoc tests examining the differences between the comparison groups for each form of counseling were all significant Results indicated that non obese respondents with a priority condition had significantly greater odds th an obese respondents without

PAGE 78

78 comorbidity to report receiving diet advice ( OR = 2 77 p < 001 ), exercise advice ( OR = 1. 34 p < .001 ), and combination advice ( OR = 2 11 p < 0 01) That is, non obese respondents with a priority condition had nearly three times the odds of receiving diet advice and over two times the odds of receiving combination advice, while they had only 34% greater odds of receiving exercise advice as compared to obese adults without comorbidity. However, non ob ese respondents without comorbidity had significantly lower odds th an obese adults without a priority condition to report receiving diet advice ( OR = .31, p < .001 or 3.2 times the odds ), exercise advice ( OR = .29, p < .001 o r 3.4 times the odds ), and combination advice ( OR = .29 p < .001 or 3.4 times the odds ). Finally, n on obese respondents with an obesity related comorbidity had greater odds than those without an obesity related comorbidity to report receiving all forms of lifestyle advice : diet advice OR = 9.00, p < .001 ; e xercise advice OR = 4.64, p < .001; and combination advice OR = 7.38, p < .001 That is, non obese adults with a comorbid condition had nine times greater odds of receiving diet advice, 4.6 times the odds of receiving exercise advice, and 7.4 times the odds of receiving combination advice compared to non obese adults without comorbidity. Number of obesity related comorbidities. Further, it was hypothesized that lifestyle advice would increase with the number of priority conditions with the greatest od ds of receiving advice among obese followed by non obese respondents with comorbidity and the lower odds of receiving advice among obese and non obese respondents without comorbidity. Results varied relative to our hypotheses (see Table 3 10).

PAGE 79

79 Diet advice Obese respondents with one ( OR = 1.29 p < .01 ) or two ( OR = 2.60 p < .001 ) priority conditions had greater odds than obese respondents with three priority conditions (reference group) to report receiving diet advice (i.e., 29% greater odds for obese a dults with one condition and 2.6 times the odds for obese adults with two conditions). Similarly, non obese respondents with two ( OR = 1.53 p < .001 ) or three ( OR = 2.39 p < .001 ) priority conditions had 53 % of the odds and 2.4 times the odds of receiv ing diet advice from their physician as compared to obese respondents with three priority conditions. Finally, compared to obese respondents with three priority conditions, being obese without the presence of illness significantly lessened the odds of rec eiving diet advice ( OR = .36 p < .001 or 2.8 times the odds ), as did being non obese with one priority condition ( OR = .83, p < .05 or 17% of the odds) or non obese without any priority conditions ( OR = .11 p < .001 o r 9.1 times the odds ). A similar patt ern emerged for exercise and combination advice (also described in Table 3 10). Exercise advice The odds of receiving exercise advice (as compared to obese adults with three priority conditions) was as follows for obese adults with two conditions OR = 3 .05 p < .001 ; one condition OR = 1.83 p < .001 ; or none OR = .65 p < .001 and for non obese adults with three conditions OR = 1.83 p < .001 ; two conditions OR = 1.21 p < .05 ; one condition OR = .74 p < .001 ; or none OR = .18 p < .001 Specif ically, obese adults with two conditions had 3.1 times the odds of receiving exercise advice than obese adults with three priority conditions, while obese adults with one condition had 83% greater odds of receiving exercise advice than obese adults with th ree priority conditions. However, obese adults with no comorbidities had 35% lower

PAGE 80

80 odds of receiving exercise advice as compared to obese adults with three priority conditions. Non obese adults with three or two conditions had 83% and 21% higher odds, res pectively, of receiving exercise advice than obese adults with three priority conditions. However, non obese adults with one priority condition had 26% lower odds of receiving exercise advice. Finally, obese adults with three priority conditions had 5.6 times greater odds of exercise advice than non obese adults without comorbidity. Combination advice. T he odds of receiving combination advice (as compared to obese adults with three priority conditions) was as follows for obese adults with two conditi ons OR = 2.50 p < .001 ; one condition OR = 1.37 p < .001 ; or none OR = .41 p < .001 and non obese adults with three conditions OR = 1.94 p < .001 ; two conditions OR = 1.27 p < .01 ; one condition OR = .69 p < .001 ; or none OR = .11 p < .001 Thus, obese adults with two conditions had 2.5 times the odds of receiving combination advice than obese adults with three priority conditions, while obese adults with one condition had 37% greater odds of receiving combination advice than obese adults wi th three priority conditions Non obese adults with three or two conditions had 94 % and 27 % higher odds, respectively, of receiving combination advice than obese adults with three priority conditions. However, non obese adults with one priority condition had 31 % lower odds of receiving combination advice. Finally, obese adults with three priority conditions had 2.4 times greater odds of combination advice than obese adults without any conditions and 9.1 times the odds of receiving advice than non obese ad ults without comorbidity. Role of race MSA and comorbidity on receipt of lifestyle advice. As indicated in Table 3 11, the current aim also examined impact of race MSA status on lifestyle

PAGE 81

81 advice among obese respondents with and without at least one prior ity condition ( N = 11,967) These logistic regression analyses showed that compared to urban Caucasians with at least one priority condition, all race MSA groups with and without at least one priority condition had significantly lower odds of receiving di et, exercise, and combination advice ( ps < .05) with one exception. Namely, no significant differences were noted for urban African Americans with at least one priority condition for receiving diet advice ( OR = .91 p = 336 ), exercise advice ( OR = .99 p = .899 ), or combination advice ( OR = .97 p = .774 ) as compared to urban Caucasians with at least one priority condition. Regarding the magnitude of observed race MSA differences, urban Caucasians and urban African Americans without comorbidity had 5.6 times and 5.3 times lower odds, respectively, of receiving diet advice ; 4.2 times and 4.5 times lower odds, respectively, of receiving exercise advice ; and 5.0 times and 4.8 times lower odds, respectively, of receiving combination advice than urban Caucasi ans with comorbidity R ural Caucasians w ith comorbidity had 27% lower odds of receiving diet advice 35 % lower odds of receiving exercise advice and 29 % lower odds of receiving combination advice than their urban counterparts. Similarly, rural African A mericans with comorbidity had 30% lower odds of receiving diet advice 50% lower odds of receiving exercise advice and 37% lower odds of receiving combination advice than urban Caucasians with comorbidity. However, rural Caucasians and rural African Ameri cans without comorbidity each had 7.7 times l ower odds of receiving diet advice ; 5.6 times and 7.1 times lower odds respectively, of receiving exercise advice ; and 6.3 times and

PAGE 82

82 7.7 times the odds respectively, of receiving combination advice than urban Caucasians with comorbidity. Tertiary Aim: Relation ship Between Provider Behaviors and Lifestyle Advice Effect of provider communication on lifestyle advice. Using a summative indicator of provider communication as scored by respondents, the present inv estigation also examined the influence of provider communication on receipt of all forms of lifestyle advice among obese respondents who completed the Self Administered Questionnaire of MEPS ( N = 7,872) The differential impact of this summative indicator of provider communication on advice received was analyzed in a logistic regression model. Results are reported in Table 3 12. Controlling for race/ethnicity and MSA status, provider communication was not a significant predictor of receipt of diet advice ( OR = 1.01, p = 329) exercise advice ( OR = 1.01, p = .335) or combination advice ( OR = 1.01, p = 641 ). Results did not differ when provider communication was evaluated as a categorical variable (as upper and lower bound in terms of N, number of parti cipants, or when comparing scores from 1 to 8 versus scores from 9 to 16 ) Influence of race MSA status on provider communication. As shown in Table 3 13, a multiple regression analysis was conducted to assess the influence of race MSA group assignment on provider communication. Compared to urban Caucasian respondents, both urban and rural African Americans were significantly more likely to provide higher scores on provider communication, b = .46, t (430) = 4.35, p < .001 and b = .69, t (430) = 3.89, p < 001. The full model accounted for only 3% of the variance in provider communication scores ( R 2 = .028, F (11,430) = 15.37, p < .001). Additionally, there were no rural urban difference s among Caucasians ( b = .11, t (430) = .90, p = .367). P ost hoc tests showed significant differences between rural Caucasian and rural

PAGE 83

83 African American respondents such that rural Caucasians had significantly lower odds of providing higher scores than rural African Americans ( OR = .45, p < .001 ), However urban and rural A frican Americans did not differ significantly from one another on their appraisal of provider communication ( OR = .79, p = .196) In analyses evaluating provider communication as a categorical variable comparing split sample (i.e., upper and lower bound o n provider communication scores), results were unchanged. However, when examining differences between participants with scores less than or equal to eight versus those with scores greater than or equal to nine, race MSA differences in appraisal of provide r communication were not demonstrated. Exploratory Aims: Effect of Lifestyle Advice on Respondent Behavior Supplementary analyses were conducted to address changes in respondent behavior as a function of receiving diet and/or exercise advice. As these are questions of a longitudinal nature, a sample of 12,440 respondents were analyzed using the MEPS Panel 12 Longitudinal Data File, which was designed to analyze changes over the two year period 2007 and 2008. The following is a summary of characteristics fo r respondents meeting eligibility criteria for the exploratory study. Our sample included a total of 7,706 adults who reported receiving routine care during the sampling years 2007 and 2008 of MEPS and completed all measures assessing weight status (i.e., BMI), receipt of diet and exercise advice, and exercise habits. Similar to the study sample, the sample of the exploratory study consisted of majority middle aged ( M + SD 45.83 + 17.49 years), Caucasian women (55.9% and 53.1%, respectively). Further, the majority of the sample was at least high school educated ( M + SD 12.67 + 3.09 years), married (57.1%), of relatively high

PAGE 84

84 socioeconomic status ( M + SD average yearly income of $60,858 + $53,523), and privately insured (53.5%). The majority of participa nts resided in metropolitan areas (84.7%) with the largest proportion representing the South (37.6%). Baseline BMI ( M + SD ) for the subsample was 27.73 + 6.07 kg/m 2 with an average BMI change ( M + SD ) of .123 + 2.81 kg/m 2 Regarding other predictor varia bles for this sample, 2,676 (34.7%) respondents reported being advised to diet at baseline, 2,839 (36.8%) reported being advised to exercise at baseline, and 2,186 (28.4%) reported receiving both diet and exercise advice (i.e., combination advice) at basel ine. Further, 4,250 (55.2%) of the respondents reported engaging in moderate to vigorous activity at least three times per week at baseline. Change in physical activity as a function of receiving exercise advice. A logistic regression model evaluated the differential impact of receiving exercise advice from a physician at baseline on the level of activity reported at endpoint of survey participation. For those who reported receiving exercise advice at baseline, the odds of engaging in moderate to vigorou s physical activity decreased significantly ( OR = .64, p < .001) even after controlling for baseline physical activity (in addition to other covariates). That is, those who received exercise advice had 36% lower odds of reporting that they engaged in mode rate to vigorous physical activity at least three times per week. Results are presented in Table 3 14. Change in weight status as a function of receiving combination advice. The final analyses evaluated the impact of combination advice on weight change, as indicated by change in BMI. First, a multiple regression model assessed whether receiving lifestyle advice at baseline predicted weight change at study conclusion. The

PAGE 85

85 results were significant, indicating that receipt of combination advice at baseline significantly predicted BMI increase ( b = .37, t (7 706) = 4.48, p < .001). The full model including our predictor and all covariates accounted for 5% of the variance in BMI change scores ( R 2 = .054, F (10,177) = 16.87, p < .001). However, an ANCOVA was us ed to assess the impact of receiving both diet and exercise advice on weight change, controlling for a variety of covariates including age, sex, years of education, annual family income, insurance status, race, MSA status, region, and baseline BMI. The ov erall model was significant, F (32, 7673) = 15.44, p < .001. As seen in the previous logistical regression model, the ANCOVA confirmed that receiving combination advice at baseline was a significant predictor of BMI change, F (1, 7673) = 31.30, p < .001. H owever, the independent contribution of combination advice was weak ( R 2 = .004), suggesting that lifestyle advice has no impact on weight change.

PAGE 86

86 Table 3 1. Participant demographics at baseline Characteristic Urban Caucasians (N = 16,919) Mean + SE or n (%) Rural Caucasians (N = 4,453) Mean + SE or n (%) Urban African Americans (N = 6,254) Mean + SE or n (%) Rural African Americans (N = 933) Mean + SE or n (%) Age (years) 47.72 + 0.27 49.18 + 0.62 c 43.18 + 0.32 a 44.11 + 1.05 b Sex Male Female (referent) 8,128 (48.9) 8,791 (51.1) 2,099 (48.4) 2,354 (51.6) 2,616 (45.1) 3,638 (54.9) a 382 (46.3) 551 (53.7) a Marital Status Married (referent) Widowed Divorced or Separated Never Married 9,961 (57.0) 1,070 (6.7) 2,312 (13.7) 3,576 (22.5) 2, 732 (61.5) a 386 (7.8) 583 (12.4) 752 (18.4) 2,197 (35.1) a 462 (6.5) 1,108 (18.0) 2,487 (40.4) 322 (37.1) a 94 (9.1) 170 (16.6) 347 (37.2) Annual Family Income ($) 76,117 + 1,142 58,905 + 1,809 a 49,821 + 1,386 a 36,253 + 1,712 a Education (years) 13 .69 + 0.04 12.78 + 0.07 a 12.75 + 0.06 a 11.85 + 0.15 a Type of Insurance Private (referent) Medicare Medicaid Uninsured Other Public Assistance 10,475 (64.2) 3,089 (18.7) 1,261 (5.8) 1,888 (10.0) 206 (1.2) 2,206 (55.6) a 1,014 (21.2) 522 (8.7) 664 (13. 4) 47 (1.1) 2,874 (52.3) a 743 (10.7) 1,407 (18.4) 1,157 (17.4) 73 (1.2) 361 (44.5) a 114 (11.4) 239 (21.6) 203 (21.1) 16 (1.4) Body Mass Index (kg/m 2 ) 27.15 + 0.07 28.11 + 0.15 a 29.01 + 0.13 a 30.32 + 0.30 a a p < .001, b p < .01, and c p < .05 for betw een group differences (urban Caucasians as reference group) are reported. Also, individuals with both Medicare and Medicaid were classif ied as Medicare.

PAGE 87

87 Table 3 2. Priority conditions, types of lifestyle advice received, and provider communication Variable Urban Caucasians (N = 16,919) n (%) Rural Caucasians (N = 4,453) n (%) Urban African Americans (N = 6,254) n (%) Rural Afri can Americans (N = 933) n (%) Priority Conditions Diabetes Hypertension Hyperlipidemia 1,420 (8.2) 4,236 (25.1) 5,339 (32.2) 471 (9.9) a 1,307 (27.6) a 1,438 (30.7) 774 (11.4) a 2,036 (30.9) a 1,622 (25.3) a 155 (14.7) a 345 (35.9) a 253 (25. 8) c Types of Lifestyle Advice Received Diet Exercise Combination 5,981 (35.5) 6,591 (39.2) 4,939 (29.4) 1,401 (30.9) a 1,547 (34.5) a 1,135 (25.4) a 2,279 (36.3) 2,570 (40.5) c 1,951 (31.1) c 305 (33.2) 320 (34.3) 247 (27.0) (N = 11,310 ) Mean + SE (N = 2,942 ) Mean + SE (N = 3,612 ) Mean + SE (N = 517 ) Mean + SE Provider Communication 13.89 + 0.03 13.93 + 0.06 14.07 + 0.06 b 14.10 + 0.12 a p < .001, b p < .01, and c p < .05 for between group differences (urban Caucasians as reference group ) Medicare and Medicaid were classified as Medicare.

PAGE 88

88 Table 3 3. Receipt of lifestyle advice among overweight adults: R ace MSA as predictor Diet Advice (N = 14,004) Exercise Advice (N = 14,004) Combination Advice (N = 14,004) Variable OR 95% CI p OR 95% CI p OR 95% CI p Race MSA group Urban Caucasians (referent) ---------Rural Caucasians .68 .56 .83 <.001 a .68 .56 .82 <.001 a .67 .56 .83 <.001 a Urban African Americans 1.13 .99 1.30 .078 1.00 .87 1.15 .980 1.10 .99 1.30 .078 Rural African Americans 1.02 .69 1.51 .915 .81 .57 1.14 .223 .98 .69 1.51 .915 Covariates Age (year s) 1.04 1.03 1.04 <.001 a 1.03 1.03 1.04 <.001 a 1.03 1.03 1.03 <.001 a Sex (male=1, female =2) 1.05 .96 1.15 306 1.43 .96 1.15 <.001 a 1.14 1.05 1.25 .003 b Education (years) 1.02 1.00 1.04 .020 c 1.04 1.00 1.04 <.001 a 1.02 1.01 1.04 .011 c Income (% of pove rty) 1.06 1.01 1.10 .017 c 1.05 1.01 1.10 .013 c 1.08 1.03 1.13 .002 b Type of Insurance .84 .80 .89 <.001 a .83 .80 .89 <.001 a .84 .80 .89 <.001 a Region 1.02 .96 1.07 .565 1.01 .96 1.07 .581 1.01 .96 1.06 .734 a p < .001, b p < .01, c p < .05 Income (% of po verty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 89

89 Table 3 4. Receipt of lifestyle ad vice among obese adul ts: R ace MSA as predictor Diet Advice (N = 11,967) Exercise Advice (N = 11,967) Combination Advice (N = 11,967) Variable OR 95% CI p OR 95% CI p OR 95% CI p Race MSA group Urban Caucasians (referent) ---------Rural Caucasians .73 .61 .86 <.001 a .68 .55 .83 <.001 a .73 .61 .88 001 b Urban African Americans 1.05 .91 1.22 .510 1.02 .89 1.16 .823 1.09 .94 1.25 .264 Rural African Americans .76 .59 .97 025 c .57 .43 .77 <.001 a .68 .53 .89 .005 b Cov ariates Age (years) 1.03 1.03 1.04 <.001 a 1.02 1.02 1.03 <.001 a 1.03 1.02 1.03 <.001 a Sex (male=1, female=2) 1.03 .93 1.14 .530 1.40 1.27 1.55 <.001 a 1.08 .98 1.20 .116 Education (years) 1.02 1.01 1.04 .005 b 1.03 1.01 1.05 001 b 1.02 1.01 1.04 .012 c Income (% of poverty) 1.01 .96 1.05 .821 1.01 .97 1.06 .536 1.03 .98 1.07 .240 Type of Insurance .88 .84 .92 <.001 a .85 .81 .88 <.001 a .87 .83 .91 <.001 a Region .98 .93 1.04 .512 .98 .92 1.05 .617 .98 .93 1.04 .608 a p < .001, b p < .01, c p < .05 Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 90

90 Table 3 5. Receipt of lifestyle ad vice among overweight adults: R ace as predictor Diet Advice (N = 14,004) Exercise Advice (N = 14,004) Combination Advice (N = 14,004) Variable OR 95% CI p OR 95% CI p OR 95% CI p Racial/ethnic group Caucasians (referent) --------African Americans 1.16 1.02 1.32 .026 c 1.01 .89 1.16 .848 1.13 .98 1.30 .086 Covariates Age (years) 1.04 1.03 1.04 <.001 a 1.03 1.03 1.03 <.001 a 1.03 1.03 1.03 <.001 a Sex (male=1, female=2) 1.05 .96 1.15 .310 1.43 1.32 1.5 5 <.001 a 1.14 1.05 1.25 .003 b Education (years) 1.02 1.00 1.04 .019 c 1.04 1.02 1.05 <. 001 a 1.02 1.01 1.04 .011 c Income (% of poverty) 1.06 1.01 1.10 .017 c 1.05 1.01 1.10 .013 c 1.08 1.03 1.13 .002 b Type of Insurance .84 .80 .89 <.001 a .83 .80 .87 <.001 a .84 .80 .89 <.001 a MSA (rural=0, urban=1) 1.43 1.21 1.69 <.001 a 1.46 1.23 1.72 <.001 a 1.46 1.22 1.73 <.001 a Region 1.02 .96 1.07 .543 1.01 .96 1.06 .566 1.01 .96 1.06 .712 a p < .001, b p < .01, c p < .05 Income (% of poverty): Poor=1, Near Poor=2, Low I ncome=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsure d=4, Other Public Assistance=5 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 91

91 Table 3 6. Receipt of lifestyle advice among obese adults: R ace as p redictor Diet Advice (N = 11,967) Exercise Advice (N = 11,967) Combination Advice (N = 11,967) Variable OR 95% CI p OR 95% CI p OR 95% CI p Racial/ethnic group Caucasians (referent) ---------African Americans 1.05 .92 1.20 .456 .99 .87 1.13 .916 1.07 .94 1.21 .342 Covariates Age (years) 1.03 1.03 1.04 <.001 a 1.02 1.02 1.03 <.001 a 1.03 1.02 1.03 <.001 a Sex (male=1, female=2) 1.03 .93 1.14 .530 1.40 1.27 1.55 <.001 a 1.08 .98 1.20 .113 Education (years) 1.02 1.01 1 .04 .005 b 1.03 1.01 1.05 001 b 1.02 1.01 1.04 .011 c Income (% of poverty) 1.01 .96 1.05 .813 1.01 .97 1.06 .528 1.03 .98 1.07 .238 Type of Insurance .88 .84 .92 <.001 a .85 .81 .88 <.001 a .87 .83 .91 <.001 a MSA (rural=0, urban=1) 1.36 1.17 1.57 <.001 a 1. 47 1.23 1.76 <.001 a 1.38 1.17 1.62 <.001 a Region .98 .93 1.04 .507 .98 .92 1.05 .582 .98 .93 1.04 .580 a p < .001, b p < .01, c p < .05 Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 92

92 Table 3 7. Receipt of lifestyle advice among overweight adults: MSA as predictor Diet Advice (N = 14,004) Exercise Advice (N = 14,004) Combination Advice (N = 14,004) Variable OR 95% CI p OR 95% CI p OR 95% CI p MSA status Rural (referent) ---------Urban 1.44 1.22 1.69 <.001 a 1.45 1.23 1.71 <.001 a 1.45 1.22 1.73 <.001 a Covariates Age (years) 1.04 1 .03 1.04 <.001 a 1.03 1.03 1.03 <.001 a 1.03 1.03 1.03 <.001 a Sex (male=1, female=2) 1.05 .96 1.16 .298 1.43 1.31 1.55 <.001 a 1.14 1.05 1.25 .003 b Education (years) 1.02 1.00 1.04 .018 c 1.04 1.02 1.05 <. 001 a 1.02 1.01 1.04 .011 c Income (% of poverty) 1.05 1.01 1.10 .020 c 1.06 1.01 1.10 .010 b 1.08 1.03 1.13 .002 b Type of Insurance .84 .80 .89 <.001 a .83 .80 .87 <.001 a .84 .80 .89 <.001 a Race 1.10 1.04 1.17 <.001 a 1.10 1.04 1.17 .001 a 1.14 1.07 1.22 <.001 a Region 1.02 .96 1.07 .566 1.02 .97 1.06 .522 1.0 1 .96 1.06 .707 a p < .001, b p < .01, c p < .05 Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Race: Caucasian=1, African American=2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 93

93 Table 3 8. Receipt of lifestyle advice among obese adults: MSA as predictor Diet Advice (N = 11,967) Exercise Advice (N = 11,967) Combination Advice (N = 11,967) Variable OR 95 % CI p OR 95% CI p OR 95% CI p MSA status Rural (referent) ---------Urban 1.36 1.17 1.57 <.001 a 1.47 1.23 1.76 <.001 a 1.38 1.17 1.62 <.001 a Covariates Age (years) 1.03 1.03 1.04 <.001 a 1.02 1.02 1.03 <.001 a 1.03 1.02 1.03 <.001 a Sex (male=1, female=2) 1.03 .93 1.14 .556 1.40 1.27 1.54 <.001 a 1.08 .98 1.19 .120 Education (years) 1.02 1.01 1.04 .005 b 1.03 1.01 1.05 001 b 1.02 1.00 1.04 .012 c Income (% of poverty) 1.01 .96 1.05 .763 1.01 .97 1.06 .503 1.03 .98 1.07 214 Type of Insurance .88 .84 .92 <.001 a .85 .81 .88 <.001 a .87 .83 .91 <.001 a Race 1.09 1.02 1.17 .010 b 1.02 .95 1.09 .655 1.10 1.03 1.18 .004 b Region .98 .93 1.04 .529 .98 .92 1.05 .595 .98 .93 1.04 .601 a p < .001, b p < .01, c p < .05 Income (% of po verty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Race: Caucasian=1, African American=2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 94

94 Table 3 9. Receipt of lifestyle advice among adults with and without obesity related comorbidities Diet Advice (N = 41,838) Exercise Advice (N = 41,838) Combination Advice (N = 41,838) Variable OR 95% CI p OR 95% CI p OR 95% CI p Weight co morbidity status OB1PLCOM (referent) ---------NO1PLCOM .72 66 .78 <.001 a .51 .47 .56 <.001 a .59 .53 .64 <.001 a OB0COM .26 .23 .28 <.001 a .38 .34 .42 <.001 a .28 .25 .31 <.001 a NO0COM .08 .07 .09 <.001 a .11 .10 .11 <. 001 a .08 .07 .09 <.001 a Covariates Age (years) 1.01 1.01 1.01 <.001 a 1.01 1.01 1.01 <.001 a 1.01 1.00 1.01 <.001 a Sex (male=1, female =2) .91 .86 .96 .001 a 1.19 1.12 1.26 <.001 a .97 .92 1.03 .277 Education (years) 1.02 1.00 1.03 .005 b 1.03 1.01 1.04 <.001 a 1.01 1.00 1.03 .028 c Income (% of poverty) 1.06 1.02 1.09 <.001 a 1.05 1.02 1.08 <.001 a 1.07 1.04 1.10 <.001 a Type of Insurance .90 .87 .93 <.001 a .88 .85 .90 <.001 a .89 .86 .92 <.001 a Race 1.15 1.10 1.20 <.001 a 1.12 1.07 1.17 <.001 a 1.18 1. 13 1.23 <.001 a MSA (rural=0, urban=1) 1.35 1.19 1.52 <.001 a 1.32 1.15 1.51 <.001 a 1.31 1.14 1.50 <.001 a Region .98 .94 1.02 .375 1.01 .97 1.05 .686 .99 .95 1.03 .721 a p < .001, b p < .01, c p < .05 Weight comorbidity status: OB1PLCOM=obese adult with at least one comorbidity, NO1PLCOM=non obese adult with at least one comorbidity, OB0COM=obese adult with no comorbidities, NO0COM=non obese adult with no comorbidities Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Race: Caucasian=1, African American=2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 95

95 Table 3 10. Receipt of lifestyle advice among adults by number of obesity related comorbidities Diet Advice (N = 41,838) Exercise Advice (N = 41,838) Combination Advice (N = 41,838) Variable OR 95% CI p OR 95% CI p OR 95% CI p Weight comorbidity status OB3COM (referent) ---------OB2COM 2.60 2.21 3.05 <.001 a 3.05 2.55 3.63 <.001 a 2.50 2.13 2.94 <.001 a OB1COM 1.29 1.10 1.50 .001 a 1.83 1.56 2.15 <.001 a 1.37 1.18 1.60 <.001 a OB0COM .36 .31 .41 <.001 a .65 .56 .75 <.001 a .41 .35 .47 <.001 a NO3COM 2.39 1.88 3.05 <.001 a 1.83 1.47 2.27 <.001 a 1.94 1.58 2.37 <.001 a NO2COM 1.53 1.31 1.78 <.001 a 1.21 1.04 1.42 .015 c 1.27 1.09 1.48 .003 b NO1COM .83 .72 .96 .011 c .74 .64 .86 <.001 a .69 .60 .80 <.001 a NO0COM .11 .10 .13 <.001 a .18 .16 .21 <.001 a .11 .10 .13 <.001 a Covariates Age (years) 1.01 1.00 1.01 <.001 a 1.01 1.00 1.01 <.001 a 1.00 1.00 1.01 <.001 a Sex (male=1, female =2) .92 .87 .97 .003 b 1.21 1.14 1.28 <.001 a .99 .93 1.04 .046 c Education (years) 1.02 1.01 1.03 <.001 a 1.03 1.02 1.04 <.001 a 1.02 1.01 1.03 .613 Income (% of poverty) 1.06 1.03 1.09 <.001 a 1.05 1.03 1.08 <.001 a 1.08 1.05 1.11 <.001 a Type of Insurance .90 .87 .93 <.001 a .88 .85 .91 <.001 a .89 .86 .92 <.001 a Race 1.15 1.10 1.21 <.001 a 1.12 1.07 1.17 <.001 a 1.18 1.13 1.23 <.001 a MSA (rural=0, urban=1) 1.34 1.19 1.52 <.001 a 1.32 1.15 1.52 <.001 a 1.31 1.14 1.50 <.001 a Region .98 .94 1.03 .455 1.01 .97 1.05 .580 1.00 .96 1.04 .842 a p < .001, b p < .01, c p < .05 Weight comorbidity status: OB3COM=obese adult with three priority conditions, OB2COM= obese adult w ith two priority conditions; OB1COM=obese adult with one priority condition, OB0COM= obese adult with no priority conditions, NO3COM=non obese adult with three priority conditions, NO2COM=non obese adult with two priority conditions, NO1COM=non obese adult with 1 priority condition, NO0COM=non obese adult with no priority conditions Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assista nce=5 Race: Caucasian=1, African American=2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 96

96 Table 3 11. Receipt of lifestyle advice among obese adults: Comorbidity status by race MSA as predictors Diet Advice (N = 11,967) Exercise Advice ( N = 11,967) Combination Advice (N = 11,967) Variable OR 95% CI p OR 95% CI p OR 95% CI p Comorbidity status by race MSA category WHTURB1 (referent) ---------WHTURB0 .18 .16 .22 <.001 a .24 .21 .29 <.001 a .20 .17 .23 <. 001 a WHTRUR1 .73 .58 .92 .009 b .65 .47 .90 .010 b .71 .55 .92 .009 b WHTRUR0 .13 .11 .17 <.001 a .18 .14 .23 <.001 a .16 .12 .20 <.001 a BLKURB1 .91 .75 1.11 .336 .99 .81 1.20 .899 .97 .80 1.18 .774 BLKURB0 .19 .15 .24 <.001 a .22 .18 .27 <.001 a .21 .17 .26 <.001 a BLKRUR1 .70 .51 .97 .031 c .50 .36 .71 <.001 a .63 .48 .83 .001 a BLKRUR0 .13 .08 .22 <.001 a .14 .08 .23 <.001 a .13 .07 .22 <.001 a Covariates Age (years) 1.01 1.00 1.01 .003 b 1.00 1.00 1.00 .853 1.00 1.00 1.00 .965 Sex (male=1, female =2) 1.06 .96 1.18 .263 1.47 1.32 1.63 <.001 a 1.12 1.01 1.24 .032 c Education (years) 1.03 1.01 1.05 .003 b 1.04 1.02 1.06 <.001 a 1.03 1.01 1.05 .008 b Income (% of poverty) 1.03 .98 1.08 .213 1.03 .99 1.08 .127 1.05 1.01 1.10 .029 c Type of Insurance .90 .86 94 <.001 a .86 .82 .89 <.001 a .88 .84 .92 <.001 a Region .97 .91 1.03 .263 .97 .91 1.04 .436 .97 .91 1.03 .358 a p < .001, b p < .01, c p < .05 Comorbidity status by race MSA category: WHTURB1=urban Caucasian adult with at least one priority condition, WHTUR B0= urban Caucasian adult without any priority conditions, WHTRUR1=rural Caucasian adult with at least one priority condition, WHTRUR0= rural Caucasian adult without any priority conditions, BLKURB1=urban African American adult with at least one priority c ondition, BLKURB0= urban African American adult without any priority conditions, BLKRUR1=rural African American adult with at least one priority condition, BLKRUR0= rural African American adult without any priority conditions Income (% of poverty): Poor=1 Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 97

97 Table 3 12. Receipt of lifestyle advice among obes e adults: P rovider communication as predictor Diet Advice (N = 7,872) Exercise Advice (N = 7,872) Combination Advice (N = 7,872) Variable OR 95% CI p OR 95% CI p OR 95% CI p Provider communication 1.01 .99 1.04 .329 1.01 .99 1.04 335 1.01 .98 1.03 .641 Covariates Age (years) 1.03 1.03 1.03 <.001 a 1.02 1.01 1.02 <.001 a 1.02 1.02 1.02 <.001 a Sex (male=1, female =2) .84 .75 .95 .005 b 1.09 .96 1.24 .165 .89 .79 1.00 .042 c Education (years) 1.01 .99 1.03 .272 1.03 1.00 1.05 .021 c 1.01 .99 1 .03 .314 Income (% of poverty) 1.01 .95 1.07 .752 1.03 .98 1.08 .309 1.04 .98 1.09 .193 Type of Insurance .90 .85 .96 <.001 a .86 .81 .91 <.001 a .88 .83 .93 <.001 a Race 1.16 1.07 1.25 <.001 a 1.07 .98 1.16 .134 1.16 1.07 1.25 <.001 a MSA (rural=0, urban=1 ) 1.23 1.03 1.47 .021 c 1.40 1.14 1.73 .002 b 1.27 1.05 1.53 <.012 b Region 1.01 .95 1.08 .682 1.02 .94 1.09 .683 1.01 .95 1.08 .661 a p < .001, b p < .01, c p < .05 Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Typ e of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Race: Caucasian=1, African American=2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 98

98 Table 3 13. Race MSA as predictor of provider communication among obese adults (N=7,872) Predictors Unstandardized Coefficient (b) Standard Error t p Race MSA Constant 13.05 .27 47.55 <.001 a Urban Caucasians (referent) ----Rural Caucasians .11 .13 .90 .367 Urban African Americans .46 .11 4.35 < .001 a Rural African Americans .69 .18 3.89 <.001 a Covariates Age (years) .01 .002 6.28 <.001 a Sex (male=1, female =2) .06 .07 .92 .357 Education (years) .02 .01 1.34 .180 Income (% of poverty) .10 .03 3.02 .003 b Type of Insurance .20 .04 5.31 <.001 a Region .14 .04 3.48 .001 a a p < .001, b p < .01 Income (% of poverty): Poor=1, Near Poor=2, Low Income=3, Middle Income=4, High Income=5 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Region: No rtheast=1, Midwest=2, South=3, West=4

PAGE 99

99 Table 3 14. Odds of engaging in physical activity at end of survey reference period among adults (N=7,706) Predictors OR 95% CI p Exercise Advice (0=no, 1=yes) .64 .56 .73 <.001 a Covariates Age (years) .99 .99 1.00 .002 b Sex (male=1, female =2) .81 .73 .90 <.001 a Education (years) 1.06 1.04 1.08 <.001 a Income (dollars) 1.00 1.00 1.00 .180 Type of Insurance .94 .89 1.00 .039 c Race .90 .82 .99 .023 c MSA (rural=0, urban=1) 1.11 .88 1.41 .363 Region 1.02 .95 1.10 .550 Physical activity habits at baseline (0=no, 1=yes) 5.37 4.73 6.11 <.001 a a p < .001, b p < .01, c p < .05 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Race: Caucasian=1, African American =2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 100

100 Table 3 15. Combination advice at baseline as predictor of weigh t change among adults (N=7,706) Predictors Unstandardized Coefficient (b) Standard Error t p Constant 3.81 .42 9.04 <.0 01 a Combination advice received (no=0, yes=1) .37 .08 4.48 <.001 a Covariates Age (years) .01 .002 2.98 .003 b Sex (male=1, female =2) .02 .07 .25 .801 Education (years) .02 .01 1.30 .180 Income ($) 1.22 6.36 1.91 .196 Type of Insurance .004 .04 .11 .057 Race .04 .05 .74 .909 MSA (rural=0, urban=1) .32 .08 3.83 <.001 a Region .03 .04 .77 .440 Baseline BMI (kg/m 2 ) .11 .01 12.14 <.001 a a p < .001, b p < .01 Type of Insurance: Private=1, Medicare=2, Medicaid=3, Uninsured=4, Other Public Assistance=5 Race: Caucasian=1, African American=2, Other=3 Region: Northeast=1, Midwest=2, South=3, West=4

PAGE 101

101 CHAPTER 4 DISCUSSION Main Findings The primary aim of this investigation was to examine the influence of race/ethnicity and rurality ( among overweight and obese adults. Lifestyle advice was operationalized in three dimensions: dietary advice, exercise advice, and diet and exercise advice or combination advice. C onsistent with our hypotheses, in independent analyses of overweight and obese adults, rural Caucasians had significantly lower odds of receiv ing diet, exercise and combination advice from their doctors than urban Caucasians T his is clinica lly meaningful and represents substantially lower odds of receiving needed advice The relationship was more complex among African American respondents. While the expected differences between both urban and rural African Americans and their Caucasian counterparts were not observed among overweight respondents, rural African American respondents who were obese had lower odds of receiving any form of lifestyle advice as compared to urban Caucasians. Interestingly, for obese rural African Americans the magnitude of the odds for receiving lifestyle advice was notably lower for exercise advice (43% lower odds ) than for diet advice (24% lower odds ) and both diet and exercise advice (32% lower odds ) than their urban Caucasian counterparts. That is these differences were m ore l ikely to be clinically meaningful for exercise advice; while they were trivial for diet and combination advice Finally, while there were no rural urban differences among overweight African Americans, among obese African American respondents, rural a dults had significantly lower odds than urban adults to report receiving all forms of

PAGE 102

102 lifestyle advice. This suggests that for rural African Americans in particular, higher BMI may be a barrier to receiving appropriate weight related advice. Results were mixed (with respect to our hypotheses) when race/ethnicity and rurality were examined as independent predictors. Specifically, while we hypothesized that diet, exercise, and combination advice would be less common among African Americans as compared to Ca ucasian respondents, results showed that there were generally no racial/ethnic differences with one exception. Namely, among overweight respondents only, African Americans had more (not less) of the odds of report ing that they had ever receiv ed diet advic e, which was not true for either exercise or combination advice. However, 16% greater odds may be considered a clinically trivial effect. W ith respect to rural urban differences, as we expected rural respondents (who had significantly higher BMI than the ir urban counterparts) reported receiving diet, exercise, and comb ination advice at lower rates. Indeed, u rban respondents had greater odds than rural respondents to report receiving all forms of advice (i.e., among overweight adults odds were 44% higher for diet advice and 45% higher for both exercise and combination advice; while for obese adults odds were 36% higher for diet advice, 47% higher for exercise advice, and 38% higher for combination advice) From a clinical perspective urban residence subs tantially increased odds of receiving all forms of lifestyle related communication. Thus, while higher rates of obesity in a particular group should lead to higher (or at least equal) rates of counseling for that group, our results generally suggest that disparities in obesity par allel disparities in (i.e., receipt of lifestyle advice).

PAGE 103

103 The secondary aim of this study was to examine the influence of obesity related lifestyle advice. Having at least one obesity related comorbidity appeared to significantly influence the odds of receiving of all types of advice. Specifically, obese adults without any comorbidities had substantially lower odds of receiving all forms of lifestyle advice than obese adults with at least one priority condition and these represent clinically meaningful differences (i.e., ranging from greater than two fold to four times greater odds for having a comorbidity versus not having a comorbidity) Non obese adults with at least one priority condition and non obese adults without any conditions also had substantially lower odds of receiving all forms of advice than obese adults with at least one comorbidity. Further, the magnitude of these differ ences were higher for non obese adults without comorbidity (e.g., compared to obese adults with comorbidity who had 12.5 times the odds of receiving diet advice which is clinically meaningful ) than for non obese adults without comorbidity ( i.e ., who had 2 8% lower odds of receiving diet advice compared to obese adults with comorbidity which represents a clinically trivial difference ). However, it was noted that non obese adults with a priority condition had greater odds than obese adults without a priorit y condition to report receiving all forms of lifestyle advice (e.g., nearly three times the odds for diet advice) Thus, having a priority condition appears to be driving the relationship to receipt of advice (as compared to weight status). The number o f priority conditions reported was also hypothesized to increase the odds of receiving lifestyle advice. Obese respondents with one or two priority conditions had greater odds than obese respondents with three priority conditions to report

PAGE 104

104 receiving any f orm of advice. Similarly, non obese adults with two or three priority conditions had greater odds tha n obese respondents with three priority conditions to report receiving any form of advice. In contrast, obese adults without any comorbidities had signif icantly lower odds than obese adults with three priority conditions to receive any advice. Finally, non obese adults with one priority condition and non obese adults without any priority condition also had lower odds than o bese adults with three priority conditions to receive any advice. The aforementioned differences were clinically meaningful with few except ions (e.g., for both obese and non obese respondents with one comorbidity differences were trivial for diet advice; for non obese respondents with on e or two comorbidities differences were trivial for exercise advice; and for obese respondents with one comorbidity and non obese respondents with two comorbidities differences were trivial for combination advice) Nonetheless, t hese variations are still relevant to medical decision making for obese versus non obese patients. Assuming the reference group (i.e., obese with presence of all priority conditions) represents the highest level of risk, how then can we explain why they exhibited lower odds of rec eiving lifestyle advice than obese adults two or three comorbidities)? One consideration is that for high risk groups, primary care providers must prioritize types of intervention approaches used, which likely favor use of medication versus behavioral interventions for those with the highest risk (e.g., obese adults with hypertension, diabetes and hyperlipidemia). Additionally, lifestyle counseling attempts may have al ready been exhausted with these adults (but

PAGE 105

105 lifestyle advice). Additional analyses examined comorbidity status and the effect of race MSA group identification. Rural C aucasians and rural African Americans with and without at least one comorbidity all exhibited lower odds of receiving each form of advice than urban Caucasians with disease. However, there were no differences noted between urban Caucasians and urban Afric an Americans. Thus, our hypothesis that all comparison groups would have significantly lower odds of receiving all forms of counseling was partially supported. As might be expected in rural populations, primary care physicians likely face the same dilemm a regarding prioritizing disease management using medication over behavioral interventions and prevention. (as indicated by quality and duration of clinical contact) as a predict or of diet, exercise, and combination advice. First, contrary to our hypothesis that there would be a positive relationship between provider communication scores and odds of receiving all forms of lifestyle advice, provider communication was not a signifi cant predictor of diet, exercise, or combination advice. We also examined the differences in perceptions of provider communication by race MSA status among obese adults. We found that both urban and rural African Americans had more favorable perceptions of their providers as compared to urban Caucasians. In addition, there were notable differences between rural Caucasian and rural African American respondents with higher appraisals reported by African Americans but no rural urban differences among this g roup. While these differences are statistically significant, the provider communication scores for all

PAGE 106

106 respondents were relatively high and the amount of variance that race MSA accounted for in the model for predicting provider communication was very low (3%); therefore, the differences observed may not be clinically meaningful. Finally, exploratory aims designed to assess changes in respondent behavior as a function of receiving diet and/or exercise advice showed that receipt of exercise advice at baselin e was a significant predictor of physical activity at the end of MEPS survey year. Those who reported receiving exercise advice had 36% lower odds of engaging in a physical activity at the end of the survey period. However, this must be interpreted in li ght of the fact that the majority of participants (55.2%) reported engaging in physical activity at the beginning of study period (i.e., selection bias) and that those with pre established exercise habits had five times the odds of report ing continued enga gement in moderate to vigorous activity (at least three times per week) at the end of survey period. Thus, receiving exercise advice may not have an influence in changes in physical activity. Receipt of diet and exercise (i.e., combination) advice also si gnificantly predicted change in BMI over course of MEPS participation such that receipt of lifestyle advice was associated with an increase in BMI (i.e., weight gain). While we have little with which to weigh these findings, including lacking information regarding whether weight loss was being attempted by a particular respondent and the time between receipt of lifestyle advice and weight loss attempts, we do know that the independent contribution of lifestyle advice to the full model was negligible, accou nting for less than 1% of the variance of the full model and that selection bias may also contribute to these results

PAGE 107

107 Comparison to Prior Research The current findings were consistent with previous research (e.g., Abid et al., 2005) with respect to gener al rates of counseling. For the current study, it appears that among adults, rates of lifestyle advice are relatively low, ranging from 28.8% to 37.8%. However, a general trend was observed showing tha t rates of counseling increased as weight increased, with the highest rates observed among obese adults (for diet, 51.4%; exercise, 59.5%; and combination, 46.6%). This is consistent with findings of the most recent National Health Disparities Report, which demonstrated that 49% of obese adults reported rec eiving advice from a health provider to maintain a healthy diet, while 57% reported receiving advice to exercise more (AHRQ, 2012). Three notable trends are reported in the National Health Disparities Report. First, there were no statistically significan t changes by race/ethnicity in the percentage of obese adults who received advice to diet or exercise. Second, between the survey years 2002 2008, there were no racial/ethnic differences in receipt of diet or exercise advice between non Hispanic Caucasian s and non Hispanic Blacks; however, differences were noted for Hispanics (as compared to non Hispanic Caucasians) for six of these years. Finally, while the National Health Disparities Report does not provide specific details regarding the findings by MSA status, rural residents were identified as a priority population in this review. Thirty measures of healthcare quality and access were examined for changes over time (i.e., in rural urban disparities). Findings demonstrated a growing gap between obese, rural and urban adults who reported ever receiving diet advice from a health provider. Further, it has been previously established that diet advice is provided more often than exercise advice (Boardley et al., 2007 ; Ma et al., 2004; Stafford et al. 2000 ).

PAGE 108

108 While the current study did not test the difference in rates of exercise and diet advice, rates observed for exercise advice were higher than those for diet advice (37.8% and 34.1%, respectively), which affirms other findings showing variations across st udies (i.e., similar rates of diet and physical activity advice observed by Kreuter and colleagues (1997) and higher rates of physical activity than diet counseling as noted by Flocke and colleagues (2005) and AHRQ (2012)) Regardless of these variations, the relatively low rates of both types of advice are still significant. Racial/ethnic and rural urban disparities in receipt of lifestyle counseling have been examined previously as independent predictors. Significant racial differences in rates of couns eling have not emerged consistently across a number of studies (e.g., Jackson, Doescher, Saver et al., 2005; Ma et al., 2009b ; Sciamanna et al., 2000; Stafford et al., 2000) However, some studies do show significant racial differences in counseling that favor minorities, namely showing that they are more likely to receive lifestyle advice (e.g., Loureiro & Nayga, 2006, Ma et al., 2004). Our findings were partially consistent with those noted by Ma and colleagues (2004) for patients at risk for cardiovasc ular disease, which found racial differences in rates of diet (but not exercise) advice for, non Hispanic Blacks, as well as Hispanics and Asian/Pacific Islanders, who were more likely to receive diet counseling. In our study, this was only the case for o verweight (not obese) African Americans. For rural urban differences, however, MSA was not previously found to be independently associated with receiving diet, exercise, or weight reduction counseling among patients with BMI > 30.0 (Ma et al., 2004). In c ontrast, urban residents in the present study had unequivocally higher odds of reporting that they received all forms of

PAGE 109

109 lifestyle advice It should be noted that region was not found to be a significant covariate in any of our models, which parallels rec ent literature suggesting that previous observations showing disparities in the South with respect to weight reduction, diet/nutrition, or exercise counseling have not persisted (CDC, 1998). As noted earlier, the majority of obese patients have one or more comorbidities (Rippe et al., 1998). Our study elucidates arguments regarding the prominence of disease focused versus prevention focused approaches to obesity management in the primary care setting (e.g., Katz & Faridi, 2007). As confirmed by our findin gs, weight loss, diet, and physical activity counseling often occur as a function of treatment for comorbid conditions (e.g., hypertension, high cholesterol, type 2 diabetes) for which obesity is a modifiable risk factor. Similar to prior studies reportin g higher rates of lifestyle counseling for high risk patient groups (e.g., Carlson et al,, 2009; Kreuter et al., 1997; Morrato et al., 2006; Wee et al., 1999), our findings indicated that the presence of an obesity related condition increased the odds of r eceiving weight related advice regardless of weight status; while being in a lower risk category was related to lower odds of receiving the same advice. Previous research suggests that doctors typically do not dedicate substantial time to discuss weight or readiness to change with their patients (e.g., Anderson & Wadden, 1999; Serdula et al., 2003), and this appears to be especially significant for African American patients (Ward et al., 2009). Our results demonstrate that p rovider communication scores wer e relatively high among all obese respondents independent of race and rurality but these appraisals may not account for weight specific health communication.

PAGE 110

110 Finally, with respect to our exploratory aims, only a few randomized control trials have evaluate d the efficacy of weight management counseling in the primary care setting for producing significant weight change (e.g., Appel et al., 2011; Ashley et al., 2001; Ely et al., 2008 ; Kumanyika et al., 2012; Lewis & Lynch, 1993; Logue et al., 2005; Pinto et a l., 2005; Ryan et al., 2010; Tsai et al., 2010; Wadden et al., 2005; Wadden et al., 2011) or for other obesity related outcomes such as increased engagement in physical activity, improvements in diet quality, or reductions in caloric intake without specifi c attention to weight change (e.g., Calfas et al., 2002; Goldstein et al., 1999; King et al., 2006; Lewis & Lynch, 1993; Ockene et al., 1999). Though the longitudinal design (of our exploratory analyses) offered an opportunity to assess changes in physica l activity and BMI over course of survey participation, the retrospective nature of the study (specifically our research aims) has its limitations. Nonetheless, contrary to previously published reports, in our study receipt of lifestyle advice was related to increases (not decreases) in weight. However, it is important to note that weight losses reported in intervention studies were not clinically significant (<5% of body weight reductions) for the majority of intervention participants. Similarly, for beh avioral changes such as exercise participation, the current study showed that receiving exercise advice early on did not increase odds of exercise participation by the end of study period, which is in direct contrast to previous findings suggesting that ph active (Lewis & Lynch, 1993 and Pinto et al., 2005 ). One possibility is the lack of repeated communication regarding weight and risk reduction via behavioral activation (e.g., engaging in ph ysical exercise) may thwart robust findings related to improved

PAGE 111

111 adherence, minimization of barriers, and improvement on tangible outcomes such as weight loss (e.g., Katz & Faridi, 2007; Logue et al., 2005). Strengths and Limitations There are a number of f actors to be weighed when considering the implications of the present study. These include a number of limitations specific to the research design and methodology. To begin, a fundamental limitation of many national surveys such as the Medical Expenditur es Panel Survey are that they are limited by their cross sectional design. While cross sectional analyses allow us to examine the associations between disease or other health related characteristics such as receipt of lifestyle advice and other variables of interest (e.g., race MSA), these findings are limited to a point in time (i.e., they speak only to the association between variables in a defined population at one particular time). Thus, for our primary aim, while we can conclude that rural residents have differential odds of receiving lifestyle advice (race and weight dependent), we cannot establish causality for the lower odds observed or determine whether these associations persist over time. A minor limitation, the external validity or generalizab ility of these results is limited to the US non institutionalized civilian population. However, given we were interested in sociodemographic differences in receipt of lifestyle advice (rather than actual prevalence rates) and our assumption that measureme nt errors are distributed evenly across groups, findings should not be affected. In addition, cross sectional designs often rely on self report, which are susceptible to measurement error such as recall bias. While an experimental design would address th is issue, for survey research it has been previously reported that self report may be superior to provider report for lifestyle related health communication given inflated rates shown in studies where physicians report on their own behavior (e.g., Phelan e t al., 2009; Power et al.,

PAGE 112

112 2006 ). But it is important to note that the same could also be true of patient reports (i.e., self report bias). Methodological challenges are also inherent in survey designs. For this study, the primary variables of interest, lifestyle advice (diet, exercise, and both) are subject to some critique regarding variable operationalization. Respondents were asked yes/no exercise advice, respectively. Responses from these items were used to construct a variable for lifestyle advice, which established whether respondents had received both diet and exercise advice. The primary limitation of this method of defining lifestyle counseling is that the nature of advice (i.e., counseling) is unknown. While the literature establishes that behavioral counseling, the gold standard for targeting small behavior or lifestyle change s (e.g., increasing physical activity, improving the quality of diet, reducing caloric intake), can presumably be maintained over the long run, monitored, and adjusted appropriately as determined in a collaborative manner by the primary care provider and t he patient (Kumanyika, 2007), these survey items might not reflect the totality of lifestyle counseling and may have also been interpreted and/or experienced differentially by respondents. As Katz and Faridi (2007) suggest, an affirmative response to thes 291) and may not account for elements of behavioral counseling that have demonstrated efficacy in obesity management (i.e., self management skills such as problem solving, focus on self mo nitoring and efficacy, application of skills to real life situations; provider use of reinforcement and behavioral contracting).

PAGE 113

113 Further, the framing of questions related to weight related counseling was not time sensitive. Whereas in previous studies que stions were addressed for a specific possibility of having received counseling or advice during a previous year or medical the lack of specificity does not rule out the possibility that weight management was discussed at a point where it may not have been relevant (and not the same timeframe is referenced by all respondents). In addition, the lack of follow up questions also fails to account for whether the reported provider communication was specific to patient relationship or determine whether the diet and exercise advice was directly related to a wei ght reduction goal. Taken together, the lack of a clear definition and parameters for evaluating lifestyle counseling in the primary care setting represents optimistic estimates of obesity related counseling. Additional methodological concerns include ope rationalization of weight status, First, the use of self reported weight status (assessed using BMI as height and weight data are not readily available in public use fi les due to confidentiality concerns) was imperfect. As respondents were not weighed, it is possible that a fraction of the respondents were miscategorized based on their self report. Further, for each of the two full year files used there was not a consi stent method available to exclude respondents who were pregnant at time of their MEPS participation. For example, while only ICD 9 codes were available for 2007 data, 2008 full year file included respondent self report of whether they were pregnant at any point during reference period in

PAGE 114

114 addition to clinical classification data. In addition, for confidentiality reasons, MEPS of whether they met the criteria for pregnancy. Despite these limitations, where applicable, analyses were conducted with and without excluding respondents known to have been pregnant during survey period, which did not alter results. Results reported herein include the full sample for the following reasons: a) given the items in question not have been pregnant when diet or exercise advice was received; and b) assuming measurement error is evenly distributed across r espondents, results should be unaffected. Next, the priority conditions examined in the current investigation were also derived from self report. Unlike data collected for weight status, responses regarding diagnosis of hypertension, diabetes, and high ch format, as well as conditions enumeration, which was subject to a verification process (i.e., repeated in multiple rounds, verbatim responses coded to ICD 9 codes and verified using MEPS Medical Provider Component, which ties conditions to events and expenditures). As noted in the study methods, for hyperlipidemia only, self report was used in lieu of condition enumeration method. Also, for diabetes, ICD included respondents with diagnoses of both Type 1 and Type 2 diabetes Finally, it is worth noting that representative scores on provider communication were not asked in the context of weight management. While this may represent a limitation to our use of the provider communication variable, it may also explain the racial differences observed when evaluating differences in appraisals of providers by

PAGE 115

115 Caucasian versus African American respondents. Similar to previous research where African Americans noted poor provider communication specific to weight ( War d et al., 2009), our results may have differed if health communication variables specifically nd 83) namely, race/ethnicity comparisons, which are achieved by the current study. Notwithstanding the aforementioned limitations, the current investigation also adds to the literature by examining the important yet understudied interaction between race and influence on lifestyle communication (e.g., CDC, 1998; Ely et al., 2006; Loureiro & Nayga, 2006; Ma et al., 2004; Ma et al., 2009b ), to our knowledge ours is the only study to exam ine the race MSA interaction. Other strengths include the use of a nationally representative survey design that employs oversampling of minorities and the ability to pool together multiple years of data, both which provided a large enough sample to examin e both racial and rural urban differences in provider lifestyle communication in primary care settings. In addition to overall low rates of lifestyle counseling observed, the study contributes to our knowledge of how race MSA, weight status, and obesity r elated comorbidity influence rates of lifestyle counseling. By observing these trends, we can determine where disparities lie and focus on subgroups that may benefit from more attention in this area. For example, while obese adults with no comorbidities and

PAGE 116

116 non obese adults with one comorbidity had lower odds of receiving all forms of lifestyle advice these adults would likely benefit from this advice in order to reduce their overall risk of increased weight, as well as morbidity and/or mortality attribu table to their current weight or comorbidity status. Finally, while the main aims of the study were addressed using cross sectional data, our exploratory aims utilized MEPS longitudinal file for Panel 12 to examine weight and behavioral changes as a funct ion of receipt of lifestyle advice. Policy and Clinical Service Implications and Future Directions The findings of this study have several implications for both policy and clinical service. Though weight status appeared to play a secondary role to obesity related disease management, race MSA differences were magnified among obese as compared to overweight subsamples suggesting that BMI is still a significant factor with respect to receipt of lifestyle advice. Additionally, in several models rural residenc e, not race, appeared to drive any race MSA differences in odds of receipt of lifestyle advice. This was true even in analyses examining the role of comorbidities, where rural Caucasians and rural African Americans with and without at least one comorbidit y all had lower odds of receiving any form of advice than urban Caucasians with disease. Given this information, we might consider evaluating our conceptualization of health disparities with respect to the paradigms in which we view obesity treatment in t he clinical setting (e.g., chronic disease model), as well as factors that influence medical decision making. While it seems intuitive that targeting groups who are less likely to receive lifestyle counseling would reduce gaps in evaluation and treatment of obesity, the disparities in obesity do not always follow treatment related disparities. For example, although obesity rates are substantially higher among African American respondents, odds of counseling were not significantly different from the refere nt group

PAGE 117

117 with only one exception that was also contrary to our expectation (i.e., overweight African Americans had higher odds of receiving diet advice only). This suggests that targeting African American patients on the basis of race alone may not elimin ate these discrepancies. With respect to other risk factors, such as the presence of comorbidity, our results indicate that acting on clinical data alone also fails to reduce disparities. Thus, physicians bear the responsibility of ensuring that all pati ents who would benefit from lifestyle advice (i.e., overweight and obese adults regardless of race or residence and independent of medical risk profile) receive timely and appropriate advice. Though this implication essentially models the clinical guidelin es set forth by various governmental entities (e.g., NHLBI, 2000; National Task Force on the Prevention and Treatment of Obesity, 2000; and the US Preventive Services Task Force, 2003), there are several documented barriers to their implementation, includi ng lack of reimbursement, physician training and self efficacy, and limited time during clinical contacts. While we may be far from mandates and requirements for evaluation and treatment of overweight and obesity, as these conditions are pushed into the r ealm of diagnosable and reimbursable disorders and framed as a public health priority (in part because of the increasing healthcare costs attributable to obesity), policy may reflect the need for specific directives for ensuring implementation of these rec ommendations. As an example, public health concerns regarding HIV/AIDS in the 1980s led to numerous policy changes geared towards increasing public awareness, reducing the incidence of HIV/AIDS, and improving health outcomes. These include state level HI V recommendations for the routine offering of HIV screening for pregnant women and

PAGE 118

118 their newborns as part of prenatal care programs (includes an opt out option) with some states (e .g., Maryland, New York) enacting statutes for specific aspects of these clinical guidelines; state specific partner notification and disclosure laws; and mandatory HIV screening of blood, blood products, bodily fluids, and organs used for blood transfusio ns, donation, and transfer/transplantation (e.g., artificial insemination, grafts) among numerous other HIV specific policies and laws (Gable, Gamharter, Gostin, Hodge, & Puymbroeck, 2007 ). As previously noted, to our knowledge, ours is the first study to examine race MSA (specifically, the unique combination of race/ethnicity and MSA status) in the context of lifestyle communication in primary care settings. Our primary findings indicate that race MSA is a significant predictor of lower odds of receipt of all forms of lifestyle counseling for both overweight and obese rural Caucasian residents, as well as for obese rural African American residents compared to the reference group (urban Caucasians). Further, there were rural urban differences observed fo r obese African Americans. These findings are not consistent with what we would expect given the higher rates of overweight and obesity among rural residents and among non Hispanic Blacks (and thus a greater disease burden associated with obesity for rura l African Americans). Similar findings emerged when evaluated in the context of obesity related comorbidities. Namely, regardless of whether a comorbidity was present, both Caucasian and African American respondents identifying as rural had lower odds th an urban Caucasians to receive any form of lifestyle advice. With respect to future directions in research, prospective surveys and experimental efforts are needed to support and expand these findings. While MEPS

PAGE 119

119 and other national surveys have their stre ngths, their use for secondary analyses often fail to address all components of a given research aim. In the present case, we would have benefited from more specificity in our operationalization of lifestyle advice for diet, exercise, and combination advi ce (the latter which we constructed for the purpose of the current investigation). Specifically, revisions would have addressed whether the who did, as well as the t ype of provider), the number of times lifestyle was addressed (within a reasonable timeframe such as one to two years, and an estimate of duration of (perhaps with sp ecific time markers), and whether the advice was offered specifically for the purpose of weight control or weight loss. In addition, it would be important to determine whether any additional weight specific advice was offered (e.g., wearing a pedometer, r educing or increasing consumption of other macronutrients, setting goals), whether weight, height, and BMI were assessed (and documented), whether motivation to change was assessed, and whether any education (e.g., regarding expectations, common barriers, non behavioral weight loss options) and resources (e.g., referral to a dietician, nutritionist, health psychologist, physical therapist, or behavioral weight loss program) were offered. Including more researchers in survey design panels might facilitate i mprovement of large scale surveys and thus improve our ability to address race MSA differences in lifestyle communication in primary care settings, as would any increase in the number of smaller scale, multi site studies designed to specifically address th ese issues.

PAGE 120

120 Last, empirical research examining outcomes by race MSA when clinical guidelines are utilized appropriately might determine the extent to which lifestyle communication reduces disparities in overweight and obesity, as well as obesity related il lnesses. Further, minor revisions or additions to the existing clinical guidelines that have proven efficacious for improving patient provider communication and that attend to cultural and geographical factors which may impede lifestyle change should also be incorporated. For example, a potential research question might assess whether activation (e.g., doctor advised diet and exercise with and without specific directiv es and direct referral to resources). Alternately, our exploratory aims might be extended with an intervention providing comprehensive, guideline driven lifestyle advice (compared to standard of care) to determine whether it directly influences weight, ex ercise and dietary habits, and other weight control behaviors. All these research questions should include analyses of racial/ethnic, rural urban, and race MSA differences. Summary and Conclusions Despite clinical guidelines calling for increased attentio n to obesity in the primary care setting, our study found that rates of lifestyle communication among adults were low, ranging from 28.8% to 37.8%. While rates for overweight and obese adults were relatively higher across race MSA groups, these findings s uggest rural urban but not racial/ethnic disparities in obesity related health communication. Given primary care providers are often the first and only line of intervention available for the treatment of obesity, more policy driven support, including adeq uate reimbursement, training and resources for providers to address obesity, is warranted. This is particularly true for those serving rural communities. Finally, these findings demonstrate a need for more

PAGE 121

121 research examining racial/ethnic and rural urban disparities in weight related lifestyle communication, as well as research evaluating the extent to which providing appropriate weight management advice may reduce disparities in obesity.

PAGE 122

122 APPENDIX RATES OF LIFESTYLE COUNSELING Table A 1. Rates of weight r eduction c ounseling Authors Data Source Prevalence Rates Phelan et al., 2009 Data from a cross sectional survey of New England based physicians examining physician reported rates of weight loss recommendations 75.5% McAlpine & Wilson, 2007 Trend anal ysis of office based primary care visits of 9,583 to 14,071 adults from annual NAMCS surveys between 1995 2004 Significant negative trend from 1995 to 2005 of 8.5% (1995 1996) to 5.8% (2003 2004) Boardley et al., 2007 Data from a small cross sectional ob servational study of two large family medicine practices located in a Midwestern city; data from the charts of 405 adults described here 48.9% (among adults only) Loureiro & Nayga, 2006 Retrospective analysis of association between patient characteristics receipt of advice, and diet and physical activity behaviors of adults (N = 64,388) from BRFSS 2001 2003 18.0% (of all adults regardless of weight status); 15.2% of overweight sample and 40.3% of obese sample Abid et al., 2005 Trend analysis of obese adu lts (N = 63,957) from BRFSS 1994, 1996, 1998, and 2000 Decreasing trend from 1994 (42.3%) to 2000 (40.3%) Scott et al., 2004 Comparative case study design using both qualitative and quantitative methods to describe patient physician communication around weight control (N = 327 encounters with adults) 11.0% of adult encounters included weight loss counseling Kuppersmith, Smoot, Williams, & Cliburn, 2006 Data from a small cross sectional observational study of two urban community based clinics in Kentucky 30.5%

PAGE 123

123 Table A 1. Continued Authors Data Source Prevalence Rates Flocke et al., 2005 Cross sectional direct observation study of family medicine outpatient visits (N = 300) 33.0% Potter et al., 2001 Survey analysis of adults (N = 366) from a divers e primary care population from the University of California San Francisco 48.0% of obese adults reported receiving advice compared to 24.0% of overweight and 12.0% of normal weight adults Stafford et al., 2000 Serial cross sectional analysis of 55858 adu lt physician office visits from the NAMCS, 1995 1996; also using NHANES III 1988 1994 to adjust for underreporting of obesity 35.5% Sciamanna et al., 2000 BRFSS, 1996; representative sample of adults (N = 13,288) from 10 states 14.4% of all respondents w ith increasing rates by BMI status (9.8% 41.7%) and risk (as indicated by presence of an obesity related comorbidity Galuska et al., 1999 BRFSS, 1996; national study of adults (N = 12,835) classified as obese 42.0% Jackson, Doescher, Saver et al., 2005 Trend analysis of obese adults (N = 69,775) from BRFSS 1994, 1996, 1998, and 2000 Decreasing trend from 1994 (44.0%) to 2000 (40.0%)

PAGE 124

124 Table A 2. Rates of diet and nutrition c ounseling Authors Data Source Prevalence Rates McAlpine & Wilson, 2007 Tren d analysis of office based primary care visits of 9,583 to 14,071 adults from annual NAMCS surveys between 1995 2004 Stable trend from 1995 to 2005 of 20.3% (1995 1996) to 19.6% (2003 2004) Boardley et al., 2007 Data from a small cross sectional observat ional study of two large family medicine practices located in a Midwestern city; data from the charts of 405 adults described here 50.2% (among adults only) Flocke et al., 2005 Cross sectional direct observation study of family medicine outpatient visits (N = 300) 31.0% Ma et al., 2004 Cross sectional analysis of diet and physical activity counseling for adults (sample size not reported) with cardiovascular risk factors using NAMCS and NHAMCS, 1992 2000 Significant increases in counseling among at risk adults, with some fluctuations noted in intermediate years. Rates in 1992 were 30% compared to 39% in 2000. Among obese respondents, 48% received diet advice. Potter et al., 2001 Survey analysis of adults (N = 366) from a diverse primary care populati on from the University of California San Francisco 27.0% of obese adults (rates not reported for overweight and normal weight adults) Wadden, Anderson, Foster, Bennett, Steinberg, & Sarwer, 2000 Cross sectional analysis of treatment seeking obese women (N = 259) and their perception of physician recommendations regarding weight 56.2% of patients report receiving advice to engage in any of 10 common weight control strategies (e.g., diet plan, commercial weight loss program, medication, readings, exercis e) Stafford et al., 2000 Serial cross sectional analysis of 55858 adult physician office visits from the NAMCS, 1995 1996; also using NHANES III 1988 1994 to adjust for underreporting of obesity 41.5%

PAGE 125

125 Table A 3. Rates of physical activity c ounseling Authors Data Source Prevalence Rates McAlpine & Wilson, 2007 Trend analysis of office based primary care visits of 9,583 to 14,071 adults from annual NAMCS surveys between 1995 2004 Stable trend from 1995 to 2005 of 15.2% (1995 1996) to 14.2% (2003 2004 ) Boardley et al., 2007 Data from a small cross sectional observational study of two large family medicine practices located in a Midwestern city; data from the charts of 405 adults described here 41.0% (among adults only) Flocke et al., 2005 Cross secti onal direct observation study of family medicine outpatient visits (N = 300) 45.0% Ma et al., 2004 Cross sectional analysis of adults (sample size not reported) with cardiovascular risk factors using NAMCS, 1992 2000 Significant increases in counseling among at risk adults, with some fluctuations noted in intermediate years. Rates in 1992 were 17% compared to 26% in 2000. Among obese respondents, 35% received diet advice Potter et al., 2001 Survey analysis of adults (N = 366) from a diverse primary car e population from the University of California San Francisco 30.0% of obese adults (rates not reported for overweight and normal weight adults) Stafford et al., 2000 Serial cross sectional analysis of 55858 adult physician office visits from the NAMCS, 1 995 1996; also using NHANES III 1988 1994 to adjust for underreporting of obesity 32.8%

PAGE 126

126 LIST OF REFERENCES Abid, A., Ga luska, D. A., Khan, L. K., Gill e s p ie, C., Ford, E. S., & Serdula, M. K. (2005). Are healthcare professionals advising obese patients to lose weight? A trend analysis. Medscape General Medicine, 7 10. Abramson, S., Stein, J., Schaufele, M., Frates, E., & Rogan, S. (2000). Personal exercise habits and counseling practices of primary care physicians: A national survey. Clinical Journal of Sport Medicine 10 40 48. Agency for Healthcare Research and Quality (AHRQ). (n.d.). MEPS HC sample design and data collection process. Retrieved from http://meps.ahrq.gov/mepsweb/survey_comp/hc_data_collection.jsp Agency for Healthcare Research and Quality (AHRQ). (2012). National healthcare disparities report, 2011. Retrieved from http://www.ahrq.gov/qual/nhdr11/nhdr11.pdf Alley, D. E., & Chang, V. W. (2007). The changing relationship of obesity and disability, 1988 2004. Journal of the Amer ican Medical Association, 298 2020 2027. Anderson, D. A., & Wadden, T. A (1999). Treating the obese patient: Suggestions for primary care practice. Archives of Family Medicine, 8 156 167. Andreyeva, T., Puhl, R. M., & Brownell, K. D. (2008). Chang es in perceived weight discrimination among Americans, 1995 1996 through 2004 2006. Obesity, 16 1129 1134. Appel, L. J., Clark, J. M., Yeh, H., Wang, N., Coughlin, J. W., Daumit, G., Brancati, F. L. (2011). Comparative effectiveness of weight lo ss interventions in clinical practice. New England Journal of Medicine, 365 1959 1968. Ard, J. D., Skinner, C. S., Chen, C., Aickin, M., & Svetkey, L. P. (2005). Informing cancer prevention strategies for African Americans: The relationship of African American acculturation to fruit, vegetable, and fat intake Journal of Behavioral Medicine, 28 239 247. Arterburn, D. E., Maciejewski, M. L., & Tsevat, J. (2005). Impact of morbid obesity on medical expenditures in adults. International Journal of Obe sity and Related Metabolic Disorders 29 334 339. Ashley, J. M., St Jeor, S. T., Schrage, J. P., Perumean Chaney, S. E., Gilbertson, M. C., McCall, N. L., & Bovee, V. (2001). Archives of Internal Medicine, 161 1599 1604. Baker, D. W. (2003). Racial and ethnic disparities in emergency department analgesic prescription. American Journal of Public Health, 93 2067 2073.

PAGE 127

127 Baskin, M. L., Ard, J., Franklin, F., & Allison, D. B. (2005). Prevalence of obesity in th e United States. Obesity Reviews, 6 5 7. Bardia, A., Holtan, S. G., Slezak, J. M., & Thompson, W. G. (2007). Diagnosis of obesity by primary care physicians and impact on obesity management. Mayo Clinic Proceedings, 82 927 932. Befort, C. A., Greiner K. A., Hall, S., Pulvers, K. M., Nollen, N. L., Charbonneau, A., Ahluwalia, J. S. (2006). Weight related perceptions among patients and Journal of General Internal Medi cine, 21 1086 1090. Befort, C. A., Nazir, N., & Perri, M. G. (2012). Prevalence of obesity among adults from rural and urban areas of the United States: Findings from NHANES (2005 2008) Journal of Rural Health, 00 1 6. doi: 10.1111/j.1748 0361.2012.00 411.x Bell, R. A., & Kravitz, R. L. (2008). Physician counseling for hypertension: What do doctors really do? Patient Education and Counseling, 72 115 121. Bertakis, K. D., & Azari, R. (2005). The impact of obesity on primary care visits. Obesity Res earch, 13 1615 1622. Bish, C., Blanck, H. M., Serdula, M., Marcus, M., Kohl III, H. W., & Khan, L. K. (2005). Diet and physical activity behaviors among Americans trying to lose weight: 2000 Behavioral Risk Factor Surveillance System. Obesity Research, 13 596 607. Boardley, D., Sherman, C., Ambrosetti, L., & Lewis, J. (2007). Obesity evaluation and intervention during family medicine well visits. American Board of Family Medicine, 20 252 257. Bogart, L. M., Catz, S. L., Kelly, J. A., & Benotsch, E. G. (2001). Factors influencing HIV disease. Medical Decision Making, 21 28 36. Bowerman, S., Bellman, M., Saltsman, P., Garvey, D. Pimstone, K. Skootsky, S ., Heber, D. (2001). Implementation of a primary care physician network obesity management program. Obesity Res earch, 9 321S 325S. Bray, G. A. (2002). Drug treatment of obesity. In T. A. Wadden & A. J. Stunkard (Eds.), Handbook of Obesity Treatment (pp. 317 338). New York: Guilford Press. Burgess, D. J., Fu, S. S., & van Ryn, M. (2004). Why do providers contribute to disparities and what can be done about it? Journal of General Internal Medicine, 19 1154 1159.

PAGE 128

128 Burroughs, V. J., Nonas, C., Sweeney, C. T., Roha y, J. M., Harkins, A. M., Kyle, T. K., & Burton, S. L. (2008). Self reported comorbidities among self described overweight African American and Hispanic adults in the United States: Results of a national survey. Obesity, 16 1400 1406. Calfas, K. J., Sa llis, J. F., Zabinski, M. F., Wilfley, D. E., Rupp, J., Prochaska, J. J., Patrick, K. (2002). Preliminary evaluation of a multicomponent program for nutrition and physical activity change in primary care: PACE+ for adults American Journal of Preven tive Medicine 34 153 161. Calle, E. E., Thun, M. J., Petrelli, J. M., Rodriguez, C., & Heath Jr., C. W. (1999). Body mass index and mortality in a prospective cohort of US adults. New England Journal of Medicine, 341 1097 1105. Campbell, M. K., DeVe llis, B. M., Strecher, V. J., Ammerman, A. S., DeVellis, R. E., & Sandler, R. S. (1994). Improving dietary behavior: The effectiveness of tailored messages in primary care settings. American Journal of Public Health, 4 783 787. Carlson, S. A., Maynard, L. M., Fulton, J. E., Hootman, J. M., & Yoon, P. W. (2009). Physical activity advice to manage chronic conditions for adults with arthritis or hypertension, 2007. American Journal of Preventive Medicine, 49 209 212. Cawley, J., & Meyerhoefer, C. (2012 ). The medical care costs of obesity: An instrumental variables approach. Journal of Health Economics 31 219 230. Centers for Disease Control and Prevention (CDC). (1998). Missed opportunities in preventive counseling for cardiovascular disease: Unite d States, 1995. MMWR Morbidity and Mortality Weekly Report, 47 91 95. Champagne, C. M., Bogle, M. L., McGee, B. B., Yadrick, K., Allen, H. R., Kramer, T. R., Lower Mississippi Delta Nutrition Intervention Research Initiative. (2004). Dietary inta ke in the lower Mississippi delta region: Results from the Foods of our Delta Study. Journal of the American Dietetic Association, 104 199 207. Christian, J. G., Bessesen, D. H., Byers, T. E., Christian, K. K., Goldstein, M. G., & Bock, B. C. (2008). Cl inic based support to help overweight patients with type 2 diabetes increase physical activity and lose weight. Archives of Internal Medicine, 168 141 146. Coeytaux, R. R., Gillespie, H. M., Callahan, L. F., Kaufman, J. S., Plescia, M., Williams, C., & Sloane, P. D. (2004). Racial disparities in health between white and African American family practice patients: Clinical implications. North Carolina Medical Journal, 65 330 334.

PAGE 129

129 Cohen, S. B., DiGaetano, R., & Goksel, H. (1999). MEPS Methodology Repor t #5: Estimation procedures in the 1996 Medical Expenditure Panel Survey Household Component. Retrieved from http://meps.ahrq.gov/mepsweb/data_files/publications/mr5/mr5.shtml Colditz, G. A., & Stein, C. (2007). Costs of obesity. In S. Ku manyika & R. C. Brownson (Eds.), Handbook of obesity prevention: A resource for health professionals (pp. 73 83). New York: Springer. Cooper Patrick, L., Gallo, J. J., Gonzales, J. J., Vu, H. T., Powe, N. R., Nelson, C., & Ford, D. E. (1999). Race, gender, and partner ship in the patient physician relationship. Journal of the American Medical Association, 282 583 589. Cooper, L. A., Roter, D. L., Johnson, R. L., Ford, D. E., Steinwachs, D. M., & Powe, N. R. (2003). Patient centered communication, ratings of care, and concordance of patient and physician race. Annals of Internal Medicine, 139 907 915. Crawford, P. B., Gosliner, W., Strode, P., Samuels, S. E., Burnett, C., Craypo, L., & Yancey A. K. (2004). Walking the talk: Fit WIC wellness programs improve self eff icacy in pediatric obesity prevention counseling. American Journal of Public Health 94 1480 1485. Donnelly, J. E., Blair, S. N., Jakicic, J. M., Manore, M. M., Rankin, J. W., & Smith, B. K. (2009). American College of Sports Medicine position stand: Ap propriate physical activity intervention strategies for weight loss and prevention of weight regain for adults. Medicine and Science in Sports and Exercise, 41 459 471. Durant, N. H., Bartman, B., Person, S. D., Collins, F., & Austin, S. B. (2009). Pati ent provider communication about the health effects of obesity. Patient Education and Counseling, 75 53 57. Dumitrescu, R. G., & Cotarla, I. (2005). Understanding breast cancer risk: Where do we stand in 2005? Journal of Cellular and Molecular Medici ne 9 208 221. Eberhardt, M. S., Ingram, D. D., Makuc, D. M., Pamuk, E. R, Freid, V. M, Harper, S. B., Xia, H. (2001). Urban and rural health chartbook. Health, United States. Hyattsville, Maryland: National Center for Health Statistics. Eged e, L. E., & Zheng, D. (2002). Modifiable cardiovascular risk factors in adults with diabetes: Prevalence and missed opportunities for physician counseling. Archives of Internal Medicine, 162 427 433. Egede, L. E. (2003). Lifestyle modification to impro ve blood pressure control in individuals with diabetes: Is physician advice effective? Diabetes Care, 26 602 607.

PAGE 130

130 Ely, A. C. Greiner, K. A. Born, W. Hall, S. Rhode, P. C. James, A. S. Ahluwalia, J. S. (200 6). Concordance of patient physici an obesity diagnosis and treatment beliefs in rural practice settings. Journal of Rural Health, 22 364 366. Ely, A. C. Banitt, A. Befort, C. Hou, Q. Rhode, P. C ., Grund, C. Ellerbeck, E. (2008). Kansas primary care weighs in: A pilot randomiz ed trial of a chronic care model program for obesity in 3 rural Kansas primary care practices. Journal of Rural Health, 24 125 132. Ferraro, K. F., & Holland, K. B. (2002). Physician evaluation of obesity in health So cial Science and Medicine, 55 1401 1413. Finkelstein, E. A., Trogdon, J. G., Cohen, J. W., & Dietz, W. (2009). Annual medical spending attributable to obesity: Payer and service specific estimates. Health Affairs, 28 w822 w831. Flegal, K. M., Carroll M. D., Kit, B. K., & Ogden, C. L. (2012). Prevalence of obesity and trends in the distribution of body mass index among U.S. adults, 1999 2010. Journal of the American Medical Association, 307 491 497. Flegal, K. M., Graubard, B. I., Williamson, D. F. & Gail, M. H. (2005). Excess deaths associated with underweight, overweight, and obesity. Journal of the American Medical Association, 293 1861 1867. Flocke, S. A. Clark, A. Schlessman, K. & Pomiecko, G ( 2005). Exercise, diet, and weight loss adv ice in the family medicine outpatient setting. Family Medicine, 37 415 421. Fong, R. L., Bertakis, K. D., & Franks, P. (2006). Association between obesity and patient satisfaction. Obesity, 14 1402 1411. Foster, G. D., Wadden, T. A., Makris, A. P., Da vidson, D., Sanderson R. S., Allison, D. B., & Kessler, A. (2003). Primary care physicians' attitudes about obesity and its treatment. Obesity Research, 11 1168 1177. Foster, G. D., Wadden, T. A., Phelan, S., Sarwer, D. B., & Sanderson, R. S. (2001). O them. Archives of Internal Medicine, 161 2133 2139. Freedland, S. J. (2005). Obesity and prostate cancer: A growing problem. Clinical Cancer Research, 11 6763 6766. Friedm an, R. R. (2008). Weight bias: The need for public policy. Rudd Report. New Haven, CT: Rudd Center for Food Policy and Obesity.

PAGE 131

131 Gable, L., Gamharter, K., Gostin, L. O., Hodge Jr., J. G., & Puymbroeck, R. V. V. (2007). Legal aspects of HIV/AIDS: A guide for policy and law reform. Washington, D.C.: The International Bank for Reconstruction and Development/The World Bank. Galuska, D. A., Will, J. C., Serdula, M. K., & Ford, E. S. (1999). Are healthcare professionals advising obese patients to lose weig ht? Journal of the American Medical Association 282 1576 1578. Gary, T. L., Baptiste Roberts, K., Gregg, E. W., Williams, D. E., Beckles, G. L., Miller III, E. J., & Engelgau, M. M. (2004). Fruit, vegetable and fat intake in a population based sample o f African Americans. Journal of the National Medical Association, 96 1599 1605. Glanz, K., Rimer, B. K., & Lewis, F. M. (2002). Health Behavior and Health Education. Theory, Research and Practice. San Francisco CA: Wiley & Sons. Glasgow, R. E., Eakin, E. G., Fisher, E. B., Bacak, S. J., & Brownson, R. C. (2001). Physician advice and support for physical activity: Results from a national survey. American Journal of Preventive Medicine 21 189 196. Goldstein, M. G., Pinto, B. M., Marcus, B. H., Lynn, H., Jette, A. M., Rakowski, W., Tennstedt, S. (1999). Physician based physical activity counseling for middle aged and older adults: A randomized trial. Annals of Behavioral Medicine 21 40 47. Greenwood, J. L., Narus, S. P., Leiser, J., & Egger, M. J. (2011). Measuring body mass index according to protocol : How are height and weight obtained? Journal of Healthcare Quality, 33 28 36. Gregg, E. W., Cheng, Y. J., Cadwell, B. L., Imperatore, G., Williams, D. E., Flegal, K. M., Williamson, D S. (2005). Secular trends in cardiovascular disease risk factors according to body mass index in US adults. Journal of the American Medical Association 293 1868 1874. Journal of the American Medical Association, 298 2066 2067. Hayward, M. D., Pienta, A. M., & McLaughlin, D. K. (1997). Inequality in men's mortality: The socio economic status gradient and geographic context. Journal of Health and Social Behavior, 38 313 330. exercise, and smoking cessation: Results from the National Ambulatory Medical Care Survey. Journal of the American Pharmacological Association, 46 364 369.

PAGE 132

132 He patient. International Journal of Obesity, 25 1246 1252. physician care as a f unction of gender and weight. International Journal of Obesity, 27 269 275. Hedley, A. A., Ogden, C. L., Johnson, C. L., Carroll, M. D., Curtin, L. R., & Flegal, K. M. (2004). Prevalence of overweight and obesity among US children, adolescents, and adul ts, 1999 2002. Journal of the American Medical Association, 291 2846 2850. Hill, J. O., & Wyatt, H. (2002). Outpatient management of obesity: A primary care perspective. Obesity Research, 10 124S 130S. Hoerger, T. J. (2006). Controversies in obesit y mortality: A tale of two studies. Health Promotion Economics Issue Brief, 1 1 4. Huang, J., Yu, H., Marin, E. Brock, S., Carden, D., & Davis, T. (2004). weight loss counseling in two public hospital primary care clinics. Academic Medicine 79 156 161. Irving, P., & Dickson, D. (2004). Empathy: Towards a conceptual framework for health professionals. International Journal of Healthcare Quality Assurance Incorporating Leadership in Health Services 17 212 220. Jackson, J. E., Doescher, M. P., Jerant, A. F., & Hart, L. G. (2005). National study of obesity prevalence and trends by type of rural county. Journal of Rural Health, 21 140 148. Jackson, J. E., Doescher, M. P., Saver, B. G., & Hart, L. G. (2005). Trends in professional advic e to lose weight among obese adults, 1994 to 2000. Journal of General Internal Medicine 20, 814 818. Joyal, S. V. (2004). A perspective on the current strategies for the treatment of obesity. Current Drug Targets CNS Neurological Disorders, 3 341 356. Kaplan, S. H., Gandek, B., Greenfield, S., Rogers, W., & Ware, J. E. (1995). Patient and making style: Results from the Medical Outcomes Study. Medical Care, 33 1176 1187. Katz, D. L ., & Faridi, Z. (2007). Healthcare system approaches to obesity prevention and control. In S. Ku manyika & R. C. Brownson (Eds.), Handbook of obesity prevention: A resource for health professionals (pp. 285 316). New York: Springer.

PAGE 133

133 King, A. C., Marcus, B ., Ahn, D., Dunn, A. L., Rejeski, W. J., Sallis, J. F., Activity Counseling Trial Research Group (2006). Identifying subgroups that succeed or fail with three levels of physical activity intervention: The Activity Counseling Trial. Health Psychology 25 336 347. King, W. D., Wong, M. D., Shapiro, M. F., Landon, B. E., & Cunningham, W. E. (2004). Does racial concordance between HIV positive patients and their physicians affect the time to receipt of protease inhibitors? Journal of General Internal Medicine, 19 1146 1153. Kreuter, M. W., Chheda, S.G., & Bull, F. C. (2000). How does physician advice influence patient behavior? Evidence for a priming effect. Archives of Family Medicine 9 426 433. Kreuter, M. W., Scharff, D. P., Brennan, L. K., & Lukwago, S. N. (1997). Physician recommendations for diet and physical activity: Which patients get advised to change? American Journal of Preventive Medicine, 26 825 833. Kumanyika, S. (2005). Obesity, health disparities, and prevention paradigms: Har d questions and hard choices. Preventing Chronic Disease, 2 1 9. Kumanyika, S. (2007). Obesity prevention, concepts, and frameworks. In S. Ku manyika & R. C. Brownson (Eds.), Handbook of obesity prevention: A resource for health professionals (pp. 85 1 14 ). New York: Springer. Kumanyika, S., Fassbender, J., Sarwer, D., Phipps., E., Allison, K. C., Localio, R., Wadden, T. A. (2012). One year results of the Think Health! study of weight management in primary care practice. Obesity, 20 1249 1257. Kuppersmith, N. C., Smoot, T. M., Williams, E. D., & Cliburn, E. A. (2006). Increasing awareness of obesity as a treatable medical condition. Journal of the Kentucky Medical Association, 104 141 145. Kushner, R. F. (1995). Barriers to providing nutrit ion counseling by physicians: A survey of primary care practitioners. American Journal of Preventive Medicine, 24 546 552. Larson, S. L., Machlin, S. R., Nixon, A., & Zodet, M. (2004). Chartbook #13: Healthcare in urban and rural areas, combined years 1998 2000. Rockville, MD: Agency for Healthcare Research and Quality. Lewis, B. S., & Lynch, W. D. (1993). The effect of physician advice on exercise behavior. American Journal of Preventive Medicine, 22 110 121.

PAGE 134

134 Lewis, C. E., Jacobs Jr., D. R., Mc Creath, H., Kiefe, C. I., Schreiner, P. J., Smith, D. E., & Williams, O. D. (2000). Weight gain continues in the 1990s: 10 year trends in weight and overweight from the CARDIA study. American Journal of Epidemiology, 151 1172 1181. Lichtenstein, A. H., Appel, L. J., Brands, M., Carnetho, M., Daniels, S., Franch, H. A., Wylie Rosett, J. (2006). Diet and lifestyle recommendations revision 2006: A scientific statement from the American Heart Association Nutrition Committee. Circulation 114 82 96. Logue, E., Sutton, K., Jarjoura, D., Smucker, W., Baughman, K., & Capers, C. (2005). Transtheoretical model chronic disease care for obesity in primary care: A randomized trial. Obesity Research 13 917 927. Loureiro, M. L., & Nayga Jr., R. M. (2006). Social Science and Medicine, 62 2458 2468 Lowenfels, A. B., Sullivan, T., Fiorianti, J., & Maisonneuve, P. (2005). The epidemiology and impact of pancreatic diseases in the United States. Current Gastroente rology Reports, 7 90 95. Lucas, J. W., Schiller, J. S., & Benson, V. (2004). Summary health statistics for US adults: National Health Interview Survey, 2001. Vital & Health Statistics 10 n.p. Lyznicki, J. M., Young, D. C., Riggs, J. A., & Davis, R. M (2001). Obesity: Assessment and management in primary care. American Family Physician, 63 2185 2196. Ma, J., Urizar Jr., G. G., Alehegn, T., & Stafford, R. S. (2004). Diet and physical activity counseling during ambulatory care visits in the United S tates. American Journal of P reventive Medicine, 39 815 822. Ma, J., Xiao L., & Stafford, R. S. (2009a). Under diagnosis of obesity in adults in outpatient settings [Research Letter]. Archives of Internal Medicine, 169 313 314. Ma, J., Xiao, L., Staff ord, R. S. (2009b). Adult obesity and office based quality of care in the United States. Obesity, 17 1077 1085. Machlin, S., Yu, W., & Zodet, M. (2005). Computing standard errors for MEPS estimates. Rockville, MD: Agency for Healthcare Research and Qua lity (AHRQ). Retrieved from http://www.meps.ahrq.gov/survey_comp/standard_errors.jsp Macqueen, C. E., Brynes, A. E., & Frost, G. S. (1999). Treating obesity: Can the stages of change model help predict outcome measures? Journal of Human Nutrition and Di etetics, 12 229 236.

PAGE 135

135 Mainous, A. G., King, D. E., Garr, D. R., & Pearson, W. S. (2004). Race, rural residence, and control of diabetes and hypertension. Annals of Family Medicine, 2 563 568. Makoul, G. (2003). The interplay between education and rese arch about patient provider communication. Patient Education & Counseling 50 79 84. Martin P. D., Dutton G. R., Rhode P. C., Horswell R. L., Ryan D. H., & Brantley P. J. (2008). Weight loss maintenance following a primary care intervention for l ow income minority women. Obesity, 16 2462 2467. Martin P. D., Rhode P. C., Dutton G. R., Redmann, S. M., Ryan D. H., & Brantley P. J. (2006). A primary care weight management intervention for low income African American women. Obesity, 14 1412 142 0. Martin, P. D., Rhode, P. C., Howe, J. T., & Brantley, P. J. (2003). Primary care weight management counseling: Physician and patient perspectives. Journal of the Louisiana State Medical Society, 155 52 56. Martin, S. L., Kirkner, G. J., Mayo, K., M atthews, C. E., Durstine, J. L., & Hebert, J. R. (2005). Urban, rural and regional variations in physical activity. Journal of Rural Health, 21 239 244. McAlpine, D. D., & Wilson, A. R. (2007). Trends in obesity related counseling in primary care: 1995 2004. Medical Care, 45 322 329. Mehrotra, C., Naimi, T .S., Serdula, M., Bolen, J., & Pearson, K. (2004). Arthritis, body mass index, and professional advice to lose weight: Implications for clinical medicine and public health. American Journal of Pre ventive Medicine, 27 6 21. Mokdad, A. H., Ford, E. S., Bowman, B. A., Dietz, W. H., Vinicor, F., Bales, V. S., & Marks, J. S. (2003). Prevalence of obesity, diabetes, and obesity related health risk factors, 2001. Journal of the American Medical Associa tion 289 76 79. Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes of death in the United States, 2000. Journal of the American Medical Association, 291 1238 1245. Morrato, E. H., Hill, J. O., Wyatt, H. R., Ghushch yan, V., & Sullivan, P. W. (2006). Are healthcare professionals advising patients with diabetes or at risk for developing diabetes to exercise more? Diabetes Care, 29 543 548. Must, A., Spadano, J., Coakley, E. H., Field, A. E., Colditz, G., & Dietz, W. H. (1999). The disease burden associated with overweight and obesity. Journal of the American Medical Association, 282 1523 1529.

PAGE 136

136 National Center for Health Statistics (NCHS). (2009). Health, United States, 2008 with special feature on the health of y oung adults chartbook. Hyattsville, MD: NCHS. National Heart, Lung, and Blood Institute (NHLBI) (1996). Stroke belt initiative: Project accomplishments and lessons learned. Retrieved from http://www.nhlbi.nih.gov/health/prof/heart/other/sb_spec.pdf Nat ional Heart, Lung, and Blood Institute (NHLBI). (1998). Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: The evidence report. Obesity Research 6 51S 210S. National Heart, Lung, and Blood Instit ute (NHLBI). (2000). Practical guide to the identification, evaluation and treatment of overweight and obesity in adults Bethesda, MD: US Department of Health and Human Services. National Task Force on the Prevention and Treatment of Obesity. (2000). Overweight, obesity, and health risk. Archives of Internal Medicine 160 898 904. Nawaz, H., Adams, M., & Katz, D. (1999). Weight loss counseling by healthcare providers. American Journal of Public Health 89 764 767. Nawaz, H., Katz, D., & Adams, M. (2000). Physician patient interactions regarding diet, exercise and smoking. American Journal of Preventive Medicine 31 652 657. Nawaz, H., & Katz, D. (2001). American college of preventive medicine practice policy statement: Weight management counse ling of overweight adults. American Journal of Preventive Medicine 21 73 78. Ockene, I. S., Hebert, J. R., Ockene, J. K., Saperia, G. M., Stanek, E., Nicolosi, L., Hurley, T. G. (1999). Effect of physician delivered nutrition counseling training and an office support program on saturated fat intake, weight, and serum lipid measurements in a hyperlipidemic population: Worcester Area Trial for Counseling in Hyperlipidemia (WATCH). Archives of Internal Medicine, 159 725 731. Ogden, C. L., Carroll, M. D., Curtin, L. R., McDowell, M. A., Tabak, C. J., & Flegal K.M. (2006). Prevalence of overweight and obesity in the United States, 1999 2004. Journal of the American Medical Association, 295 1549 1555. Opolka, J. L., Rascati, K. L., Brown, C. M., & G ibson, P. J. (2004). Ethnicity and prescription patterns for haloperidol, risperidone, and olanzapine. Psychiatric Services, 55 151 156.

PAGE 137

137 Orleans, C. T., Gruman, J., Ulmer, C., Emont, S. L, & Hollendonner, J. K. (1999). Rating our progress in populat ion health promotion: report card on six behaviors. American Journal of Health Promotion, 14 75 82. Paeratakul, S., Lovejoy, J. C., Ryan, D. H., & Bray, G. A. (2002). The relation of gender, race, and socioeconomic status to obesity and obesity co morb idities in a sample of US adults. International Journal of Obesity Related Metabolic Disorders, 26 1205 1210. Parks, S. E., Housemann, R. A., & Brownson, R. C. (2003). Differential correlates of physical activity in urban and rural adults of various soc ioeconomic backgrounds in the United States. J ournal of Epidemiology and Community Health 57 29 35. Patterson, P. D., Moore, C. G., Probst, J. C., & Shinogle, J. A. (2004). Obesity and physical inactivity in rural America. Journal of Rural Health 20 151 159. Perri, M. G., Nezu, A. M., & Viegener, B. J. (1992). Improving the long term management of obesity: Theory, research, and clinical guidelines. New York: Wiley. Perri, M. G., Sears Jr., S. F., & Clark, J. E. (1993). Strategies for improving ma intenance of weight loss: Toward a continuous care model of obesity management. Diabetes Care 16 200 209. Phelan, S., Nallari, M., Darroch, F., & Wing, R. (2009). What do physicians recommend to their overweight and obese patients? Journal of the Ameri can Board of Family Medicine, 22 ,115 122. Pinto, B.M, Goldstein, M.G., Ashba, J., Sciamanna, C.N., & Jette, A. (2005). Randomized control trial of physical activity counseling for older primary care patients. American Journal of Preventive Medicine, 29, 247 255. Pinto, B. M, Goldstein, M. G., & Marcus, B. H. (1998). Activity counseling by primary care physicians. American Journal of Preventive Medicine, 3 506 513. Potter, M. B., Vu, J. D., & Croughan Minihane, M. (2001). Weight management: What pati ents want from their primary care physicians. Journal of Family Practice, 50 513 518. Power, M. L., Cogswell, M. E., & Schulkin, J. (2006). Obesity prevention and treatment practices of U.S. obstetrician gynecologists. Obstetrics and Gynecology, 108 96 1 968. Price, J. H., Desmond, S. M., Krol, R. A., Snyder, F. F., & O'Connell, J. K. (1987). American Journal of Preventive Medicine, 3 339 345.

PAGE 138

138 Prochaska, J. O., & DiClemen te, C. C. (1982). Transtheoretical therapy: Toward a more integrative model of change. Psychotherapy Theory, Research and Practice, 9 276 288. Puhl, R., Andreyeva, T., & Brownell, K. D. (2008). Perceptions of weight discrimination: Prevalence and compa rison to race and gender discrimination in America. International Journal of Obesity, 32 992 1000. Puhl, R., & Brownell, K. D. (2001). Bias, discrimination, and obesity. Obesity Research 9 788 805. Puhl, R. M., & Brownell, K. D. (2006). Confrontin g and coping with weight stigma: An investigation of overweight and obese adults. Obesity, 14 1802 1815. Puhl, R. M., & Heuer, C. A. (2009). The stigma of obesity: A review and update. Obesity, 17 941 964. Reis, J. P., Bowles, H. R., Ainsworth, B. E., Dubose, K. D., Smith, S., & Laditka, J. N. (2004). Nonoccupational physical activity by degree of urbanization and US geographic region. Medicine and Science in Sports and Exercise, 36 2093 2098. Ricketts, T. C., Johnson Webb, K. D., & Taylor, P. ( 1998). Definitions of rural: A handbook for health policy makers and researchers. Bethesda, Maryland: Federal Office of Rural Health Policy. Rippe, J. M., Crossley, S., & Ringer, R. (1998). Obesity as a chronic disease: Modern medical and lifestyle man agement. Journal of the American Dietetic Association, 98 S9 S15. Rucker, D., Padwal, R., Li, S. K., Curioni, C., & Lau, D. C. (2007). Long term pharmacotherapy for obesity and overweight: Updated meta analysis. British Medical Journal, 335 1194 1199. Ryan, D. H.. Johnson, W. D., Myers, V. H., Prather, T. L., McGlone, M. M., Rood, J., Sjstrm, L. V. (2010). Archives of Internal Medicine, 170 146 154. Saha, S., Komaromy, M., Koepsell, T. D., & Bindman, A. B. (1999). Patient physician racial concordance and the perceived quality and use of healthcare. Archives of Internal Medicine, 159, 997 1004. Schwartz, M. B., Chambliss, H. O., Brownell, K. D., Blair, S. N., & Billington, C. (2003). Weight bias among health professionals specializing in obesity. Obesity Research, 11 1033 1039

PAGE 139

139 Sciamanna, C. N., Tate, D. F., Lang, W., & Wing, R. R. (2000). Who reports receiving advice to lose weight? Results from a multistate survey. Archives of Internal Medicine 160 2334 2339. Scott, J. G., Cohen, D., DiCicco Bloom, B., Orzano, A. J., Gregory, P., Flocke, S. A., Crabtree, B. (2004). Speaking of weight: How patients and primary care clinicians initiate weight loss counseling. American Journal of Preventive Medicine, 38 819 827. Schneider, E C., Zaslaversusky, A. M., & Epstein, A. M. (2002). Racial disparities in the quality of care for enrollees in Medicare managed care. Journal of the American Medical Association, 287 1288 1294. Seagle, H. M., Strain, G. W., Makris, A., & Reeves, R. S. (2009). Position of the American Dietetic Association: weight management. Journal of the American Dietetic Association, 109 330 346. Serdula, M. K., Khan, L. K., & Dietz, W. H. (2003). Weight loss counseling revisited. Journal of the American Medical A ssociation, 289 1747 1750. Simkin Silverman, L. R., Conroy. M. B., & King, W. C. (2008). Treatment of overweight and obese in primary care practice: Current evidence and future directions. American Journal of Lifestyle Medicine, 2 296 304. Simkin Silv erman, L. R., & Wing, R. R. (1997). Management of obesity in primary care. Obesity Research, 5 603 612. Smedley, B. D., Stith, A. Y., & Nelson, A. R. (Eds.). (2002). Unequal treatment: Confronting racial and ethnic disparities in healthcare Washingto n, DC: National Academies Press. Sobal J., Troiana R. P., & Frongillo Jr. E. A. (1996). Rural urban differences in obesity. Rural Sociology 2 289 305. Stafford, R. S., Farhat, J. H., Misra, B., & Schoenfeld, D. A. (2000). National patterns of phy sician activities related to obesity management. Archives of Family Medicine, 9 631 638. StataCorp. (2009). Stata Statistical Software: Release 10.1 [Software]. College Station, TX: StataCorp LP. Stevens, J. Plankey, M. W. Williamson, D. F. Thun, M J. Rust, P. F. Palesch, Y., & (1998). The body mass index mortality relationship in white and African American women. Obesity Res earch 6 268 277.

PAGE 140

140 Sussman, A. L., Williams, R. L., Leverence, R., Gloyd Jr., P. W., & Crabtree, B. F. (2 decisions: Obesity as a case study. Annals of Family Medicine, 4 327 333. Sutton, K., Logue, E., Jarjoura, D., Baughman, K., Smucker, W., & Capers, C. (2003). Assessing dieta ry and exercise stage of change to optimize weight loss interventions. Obesity Research, 11 641 652. Taira, D. A., Safran, D. G., Seto, T. B., Rogers, W. H., & Tarlov, A. R. (1997). The relationship between patient income and physician discussion of hea lthcare risk behaviors. Journal of the American Medical Association, 278 1412 1417. Tsai, A. G., Wadden, T. A., Rogers, T. A., Day, S. C., Moore, R. H., & Islam, B. J. (2010) A primary care intervention for weight loss: Results of a randomized controlle d pilot study. Obesity, 18 1614 1618. US Department of Agriculture and Economic Research Services. (2007). Measuring rurality: Rural urban continuum codes. Retrieved from http://www.ers.usda.gov/Briefing/Rurality/RuralUrbCon/ US Census Bureau. (2005) Population change in metropolitan and micropolitan statistical areas: 1990 2003. Washington, D.C.: US Department of Commerce. US Census Bureau. (2012). [Table 166, Percent distribution of the number of visits to healthcare professionals by selected cha racteristics: 2000 and 2009]. U.S. Census Bureau, Statistical Abstract of the United States: 2012 Retrieved from http://www.census.gov/compendia/statab/2012/tables/12s0166.pdf US Department of Health and Human Services, Public Health Service. (2001). Th e Rockville, MD: Office of the Surgeon General. US Preventive Services Task Force. (2003). Screening for obesity in adults. Retrieved from http://www.ahrq.gov/clinic/uspstf /uspsobes.htm Vahabi, M. (2007). The impact of health communication on health related decision making: A review of the evidence. Health Education, 107 27 41. van Ryn, M., & Burke, J. (2000). The effect of patient race and socio economic status on phys Social Science and Medicine, 50 813 828. Wadden, T. A., Andersen, D. A., Foster, G. D., Bennett, A., Steinberg, C., & Sarwer, D. attitudes and pr actices. Archives of Family Medicine, 9 854 860.

PAGE 141

141 Wadden, T. A., Berkowitz, R. I., Silvestry, F., Vogt, R. A., St John Sutton, M. G., Stunkard, A. J., Aber, J. L. (1998). The fen phen finale: A study of weight loss and valvular heart disease. Obes ity Research, 6 278 284. Wadden, T. A., Berkowitz, R. I., Womble, L. G., Sarwer, D. B., Phelan, S., Cato, R. K., Stunkard, A. J. (2005). Randomized trial of lifestyle modification and pharmacotherapy for obesity. New England Journal of Medicine, 353 2111 2120. Wadden, T. A., Volger, S., Sarwer, D. B. Vetter, M. L., Tsai, A. G., Berkowitz, R. I., Moore, R. H. (2011). A two year randomized trial of o besity t reatment in primary care p ractice New England Journal of Medicine, 365 1969 1979 Wang, Y., & Kumanyika, S. (2007). Descriptive epidemiology of obesity in the United States. In S. Ku manyika & R. C. Brownson (Eds.), Handbook of obesity prevention: A resource for health professionals (pp. 45 71 ). New York: Springer. Ward, S. H., Gr physician attempts to address obesity in the primary care setting. Journal of General Internal Medicine, 24 579 584. Wee, C. C., McCarthy, E. P., Davis, R. B., & Phillips, R. S. (1999) Physician counseling about exercise. Journal of the American Medical Association 282 1583 1588. Wee, C. C., Phillips, R. S., Cook, E. F., Haas, J. S., Puopolo, A. L., Brennan, T. A., & isfaction with ambulatory care. Journal of General Internal Medicine, 17 155 159. Wee, C. C., Phillips, R. S., Legedza, A. T. R., Davis, R. B., Soukup, J. R., Colditz, G. A., & Hamel, M. B. (2005). Healthcare expenditures associated with overweight and obesity among US adults: Importance of age and race. American Journal of Public Health 95 159 165. Wilcox, S., Castro, C., King, A. C., Housemann, R., & Brownson, R. C. (2000). Determinants of leisure time physical activity in rural compared with urba n older and ethnically diverse women in the United States. Journal of Epidemiology and Community Health, 54 667 672. Williamson, D., & Perrin, L. (1996). Behavioral therapy for obesity. Endocrinology and Metabolism Clinics of North America, 25 943 954 World Health Organization (WHO) (1998). Obesity: Preventing and managing the global epidemic. Geneva: World Health Organization. Wright, S. M., & Aronne, L. J. (2011). Nature Reviews: Endocrinology, 7 69 71.

PAGE 142

142 Yanovski, S. Z. (1993). A practical a pproach to treatment of the obese patient. Archives of Family Medicine 2 309 316. Yanovski, S. Z., & Yanovski, J. A. (2002). Obesity: Drug therapy. New England Journal of Medicine, 346 591 602. Yeh, H., Clark, J. M., Emmons, K. E., Mooree, R. H., B ennett, G. G., Warner, E. T., Appel, L. J. (2010). Independent but coordinated trials: Insights from the Practice based Opportunities for Weight Reduction Trials Collaborative Research Group. Clinical Trials, 7 322 332. Zhang, Q., & Wang, Y. (200 4). Trends in the association between obesity and socioeconomic status in U.S. adults: 1971 to 2000. Obesity Research 12 1622 1632.

PAGE 143

143 BIOGRAPHICAL SKETCH Rachel Andr Glenn is a native of Miami, FL. She received her Bachelor of Science in p sychology at Howard Uni versity (2005, Phi Beta Kappa, magna cum l aude). In 2005, she matriculated to the doctoral program in the Department of Clinical and Health Psychology at the University of Florida (h ealth p sychology track), where she has since earned a master (2007). M r include obesity and health promotion, as well as health disparities impacting women and racial/ethnic minorities. She completed her predoctoral psychology internship at the Bruce W. Carter V A Medical Center in Miami, FL in 2011 and will be returning the Miami VA for her postdoctoral residency with an emphasis in behavioral m edicine