Exploring Dyadic Concordance in Physician-Patient Dyads Using a Paired Survey Approach

Material Information

Exploring Dyadic Concordance in Physician-Patient Dyads Using a Paired Survey Approach
Coran, Justin
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (109 p.)

Thesis/Dissertation Information

Master's ( M.A.)
Degree Grantor:
University of Florida
Degree Disciplines:
Committee Chair:
Koropeckyj-Cox, Tanya
Committee Members:
Peek, Charles W.
Arnold, Christa
Graduation Date:


Subjects / Keywords:
Diseases ( jstor )
Dyadic relations ( jstor )
Dyadics ( jstor )
Health care industry ( jstor )
News content ( jstor )
Oncology ( jstor )
Pain ( jstor )
Patient education ( jstor )
Physicians ( jstor )
Social interaction ( jstor )
Sociology -- Dissertations, Academic -- UF
agreement, communication, concordance, dyad, interaction, patient, physician
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Sociology thesis, M.A.


The purpose of this thesis is to explore the concept of dyadic concordance in physician-patient dyads using a parallel survey method. Dyadic concordance was conceptualized as a new measure aimed at describing the level of agreement between physician and patient perceptions after their medical encounter. This dimension of medical interaction may help to identify communication barriers or disparities within physician-patient dyads. The project is designed as a preliminary study to test the effectiveness of the concordance instrument with a convenience sample of physicians and patients from family medicine and oncology specialties. Results showed that dyadic concordance can be a useful measurement of physician-patient interaction and can function in a variety of ways. Although results cannot be generalized due to sampling limitations, research findings suggest that male patients, minority patients, patients under the age of 50, and patients with lower levels of education may produce higher discordance levels with their physicians. Levels of dyadic concordance may be influenced by physician communication style and should be explored in future studies. The dyadic concordance instrument may be used in further research to explore factors associated with higher or lower concordance as well as in observational studies to examine how concordance relates to physician-patient communication. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
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.
Thesis (M.A.)--University of Florida, 2009.
Adviser: Koropeckyj-Cox, Tanya.
Statement of Responsibility:
by Justin Coran.

Record Information

Source Institution:
Rights Management:
Copyright Coran, Justin. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
665167570 ( OCLC )
LD1780 2009 ( lcc )


This item has the following downloads:

Full Text




2 2009 Justin J. Coran


3 ACKNOWLEDGMENTS This study would not have been po ssible with out the seven year guidance of my committee chair and close mentor, Dr. Tanya Koropeckyj-C ox, whose endless mentoring helped me grow into the graduate student and re searcher I am today. Without he r I would never have become a sociologist and therefore would never have started my journey to improve the physician-patient relationship. I thank Dr. Charles Peek IV, who has taught me vari ous quantitative method techniques and helped me grow from a young studen t with ideas, to a graduate student with the ability to carry ideas to fruiti on. I especially thank Dr. Christa Arnold, who not only agreed to serve on my thesis committee, but also decided to join me in my pursuit for better health communication, and instead of laughing at my crazy academic theories, helped me explore and develop them in my research. Without her, our dyad-centered theory w ould not exist and this project would have no conceptual framework. To all faculty members and graduate students in the Department of Sociology, I cannot thank you enough for your help throughout the last couple of years.


4 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 3LIST OF TABLES ...........................................................................................................................6LIST OF FIGURES .........................................................................................................................8ABSTRACT ...................................................................................................................... ...............9 CHAP TER 1 AN INTRODUCTION TO DYADI C CONCORDANCE ..................................................... 10Introducing the Physic ian-Patient Dyad .................................................................................10Conceptualizing Dyadic Concordance ................................................................................... 142 BACKGROUND AND SIGNIFICANCE .............................................................................. 18The Evolution of Physician-Patient Interaction ......................................................................18Five Key Dimensions of Patient Centeredness .......................................................................24The Significance and Complexity of the Medical Encounter ................................................. 26Racial and Socioeconomic Influences ............................................................................. 27The Influence of Gender ..................................................................................................31Examining Complexity and Agreem ent in the Medical Encounter ................................. 32Measuring Physician-Patient Interaction ................................................................. 33Influences of Physician-Patient Relationship ...........................................................36Research Questions ............................................................................................................ .....373 RESEARCH DESIGN AND METHODS .............................................................................. 40Physician-Patient Questionnaires ...........................................................................................40Sample ........................................................................................................................ ............41Recruitment and Procedures ................................................................................................... 424 RESULTS ....................................................................................................................... ........45Physician Medical Characteristics ..........................................................................................45Patient Frequencies and Individual Results ............................................................................ 46Dyadic Concordance among 50 Physician-Patient Dyads ...................................................... 47Total Concordance ...........................................................................................................49Concordance and Patients Self Reported Health ...........................................................50Concordance and Patients Confiden ce and Trust in their Physician .............................. 52Concordance and Patien t Education Level ......................................................................54Concordance and Patient Pain Rating .............................................................................55Concordance and Physicians Explanation of Diagnosis ................................................. 56


5 Concordance and Physicians Explanation of Treatm ent ................................................. 57Results Concerning Unanswered Questions ....................................................................58Dyadic Concordance within the Medical Sp ecialties Family Medicine and Oncology ......... 585 DISCUSSION .................................................................................................................... .....76Dyadic Concordance, Perception, and Implications for Practice ........................................... 82Examining Total Concordance ........................................................................................ 82Interpreting Discordance ................................................................................................. 85Comparing Medical Specialties .......................................................................................88Implications for Practice .................................................................................................. 91Suggestions for Instrument Improvement .............................................................................. 91Improving Physician-Patient Interaction: Suggestions for Future Research ..........................92APPENDIX: SURVEY INSTRUMENTS .....................................................................................94LIST OF REFERENCES .............................................................................................................100BIOGRAPHICAL SKETCH .......................................................................................................109


6 LIST OF TABLES Table page 4-1 Frequency of characteristics for physicians (N= 10) and patients (N= 50) ........................62 4-2 Frequency of physician characteristics c oncerning m edical traini ng and experience with patients .......................................................................................................................63 4-3 Patient health descriptors and reported health encounter char acteristics from patient surveys (N= 50) .................................................................................................................64 4-4 Total concordance out of 300 phys ician-patient opportunities by patient characteristics .....................................................................................................................65 4-5 Frequency of concordant/discordant responses between physician-patient dyads on parallel questions ...............................................................................................................66 4-6 Frequency of concordant/discordant responses of parallel questions by patient sociodem ographic variables ...............................................................................................67 4-7 Frequency of concordant/discordant responses on parallel questions by patient incom e, college degree and prior experience with physician ............................................ 68 4-8 Frequency of physicians dont know responses regarding their patients self reported health and characteristics .....................................................................................69 4-9 Frequency of physicians dont know responses regarding their patients confidence and trust in their m edical decisions by patient characteristics ........................ 70 4-10 Frequency of physicians dont know responses by physicians m edical experience (N=50) ........................................................................................................................ ........70 4-11 Illustration of dyadic concordance by com paring physician and patient response when asking about the patient s self reported health ......................................................... 71 4-12 Illustration of dyadic concordance by com paring physician and patient response when asking about the patients confidence and trust in their physician ........................... 71 4-13 Illustration of dyadic concordance by com paring physician and patient response when asking about the patients highest level of education ...............................................72 4-14 Illustration of dyadic concordance by com paring physician and patient response when asking about the patients pain rating .......................................................................72 4-15 Illustration of dyadic concordance by com paring physician and patient response when asking about how well the phy sician explained the diagnosis .................................73


7 4-16 Illustration of dyadic concordance by com paring physician and patient response when asking about how well the phy sician explained the treatment .................................73 4-17 Mean difference between physician and pa tient parallel respons es by the specialties fa mily medicine (n=25) and oncology (n=25) ...................................................................74 4-18 Frequency of concordant/discordant respons es of parallel questions by the specialties fa mily medicine and oncology ........................................................................................... 74 4-19 Frequency of physicians dont know responses by the specialties fam ily medicine and oncology (N=50) .........................................................................................................75


8 LIST OF FIGURES Figure page 2.1 Factors That Effect Satisfaction, Medi cal Decisions, and Relationship between Physicians and Patients ...................................................................................................... 39


9 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts EXPLORING DYADIC CONCORDANCE IN PHYSICIAN-PATIENT DYADS USING A PAIRED SURVEY APPROACH By Justin J. Coran May 2009 Chair: Tanya Koropeckyj-Cox Major: Sociology The purpose of this thesis is to explore th e concept of dyadic c oncordance in physicianpatient dyads using a parallel su rvey method. Dyadic concordance was conceptualized as a new measure aimed at describing the level of agreem ent between physician and patient perceptions after their medical encounter. This dimension of medical interaction may help to identify communication barriers or dispari ties within physician-patient dya ds. The project is designed as a preliminary study to test the e ffectiveness of the concordance instrument with a convenience sample of physicians and patients from fam ily medicine and oncology specialties. Results showed that dyadic concordance can be a useful measurement of physician-patient interaction and can function in a variety of ways. Although results cannot be genera lized due to sampling limitations, research findings suggest that male pa tients, minority patients, patients under the age of 50, and patients with lower levels of educa tion may produce higher discordance levels with their physicians. Levels of dyadic concordance may be influenced by physician communication style and should be explored in future studies. The dyadic concordance instrument may be used in further research to explore factors associated with higher or lower concordance as well as in observational studies to examine how concordance relates to physician-pa tient communication.


10 CHAPTER 1 AN INTRODUCTION TO DYADI C CONCORDANCE Fundamental problems exist between physicia ns and patients. Arguably, research has created more questions than answers when trying to decide how to meet the goals of physicians and patients. Although research has continued in our search for the most efficient medical interaction, we are no closer toda y in instituting a universal pa radigm or training program to meet the numerous expectations of medical part icipants. This chapter will provide a brief overview of the physician-patient relationship an d then suggest a new measure, termed dyadic concordance in order to aid rese archers in exploring the variable perceptions of physicians and patients. Introducing the Physician-Patient Dyad The physician-patient interaction is a cri tical com ponent of the medical encounter. Communication between doctor and patient has been recorded, discussed, and scrutinized for more than fifty years (Parsons, 1951; Szasz & Hollinder. 1956; Balint, 1969; Friedson, 1970; Byrne & Long, 1976; McWhinney, 1985; Hall, Irish, Roter, Ehrlich, & Miller, 1994; Charles, Gafni, & Whelan, 1997; Eggly, 2002; du Pre, 20 05; Street, Gordon, & Haidet, 2007; Teal & Street, 2008). It is surprising that within a world of infinite human interactions to investigate this one medical interaction may have drawn more at tention and produced more literature than any other informal interact ion in the world. Like any human interaction, this specific medical encounter creates a large amount of variability, which is no t surprising considering the unpredictable nature of human behavior; however, after five decades of research we are still struggling to improve facets of physician-patient interaction. Scholars have identified critical themes con cerning the medical encounter. First, patients who describe their relationship with physicians as satisfactory, tend to have higher levels of


11 overall health (du Pre, 2005), listen to their physicians advice more frequently (Klingle & Burgoon, 1995), are more loyal and trusting (Cegal a, Mcgee, & McNeilis, 2008), and engage in healthier lifestyle behaviors (Roter & Hall, 1992). Consequently, physicians who undertake the Hippocratic Oath generally want wh at is best for their patients, and, at the same, to time improve health in their community. When asked, physicians report an interest in seeing their patients satisfied regarding their medical care (Wright Sparks, & OHair, 2008; Talen, Grampp, Tucker, & Schultz, 2008), but also want to benefit from the medical encounter (R oter & Hall, 1992). To facilitate an efficient medical c onsultation, physicians have suggested that patients need to talk to them as well (Halpern, 2007), ask clarifying que stions, and follow medical advice (Bryne & Long, 1976). Taken together, we can identify essentia l themes that both patients and physicians want out of their medical relationship in orde r to be satisfied rega rding medical care. Unfortunately, patients have expressed high leve ls of dissatisfaction, frustration, and lack of confidence in medical care (Heritage & Ma ynard, 2006). An interdisciplinary focus has attempted to rectify fundamental problems in medical care caused by the faltering perceptions of both patients and physicians. Part of the proble m lies in the decline in the centrality of communication during the middle of the 20th century and the co rresponding rise of communication barriers degradi ng the process of car e (Shorter, 1985). Shorter (1985) and Roter (2000) elaborate that physicians wanted to institute a process that increased the accuracy of clinical findings and improve th e science of medicine To this end, physicia ns directed medical inquiry away from the patient and instituted a biomedical approach, which focused on scientific findings and the process of di sease. Although this approach improved medical sciences, diagnostic criteria, and various tr eatment mechanisms, it lost sight of a fundamental aspect of medical care: the patient.


12 Shortly after this new medical direction in the 20th century, scholars wanted to find a new medical paradigm to address medical dialogue a nd create a more efficient medical encounter. This paradigm would focus on re lationship-centered care between phys icians and patients, and in its first iteration became patient-centered communication (Balint, 1969). The concept of patient-centered communication (i.e. patient-centere dness) has existed in medical practice and health communication for over three decades and many scholars and clinicians have advocated its use within medical consultation over the traditional biomedical or paternalistic approach (Balint, 1969; Byrne & Long, 1976; Henbest & Stewart, 1989; Mead, & Bower, 2000; Zandbelt, Smets, Oort, & de H aes, 2005). This communication approach was designed to shift the locus of c ontrol partly away from physicians toward the patient in hope of improving medical interactions by increasing levels of patient satisfaction, trust, and confidence. Patient-centered consultations reflect recognition of patie nts needs and preferences, characterized by behaviors such as encouraging the patient to voice id eas, listening, reflecting and offering collaboration (Mead & Bower, 2000). Presently, a clear and concise definition for pa tient centered interactions that could be taught to physicians still does not exist within the medical field. Zanbe lt et al. (2005) recently argued that a one size fits a ll approach or definition of patient-centeredness may not be possible if we wish to create a universal defini tion. On the other hand, a universal definition of this communication approach may be crucial fo r instructing new medical students and accurately measuring its effectiveness in cl inical application. Arnold and Coran (2008) further speculate on the difficulties of teaching a term that is mi ssing a universal definition. They note that if researchers and physicians can not agree on a revised definition and approach for patientcenteredness, then how are medical professors su pposed to teach a patient-centered approach to


13 new medical students? Arguably, a universal approach is needed in order to decrease the inconsistency in medical dialogue and physic ian communication styles, while improving the perception of medical care. A strong emphasis has been placed on this concept due to a documented positive effect on the physician-patient dyad, which was the original goal that sparked the medical shift toward patient-centeredness. Heritage and Maynard (200 6) discuss a trend wher e, patients receiving patient-centered care tend to show greater primary care attendance, smoking cessation, adherence to healthy lifestyle choices, more positiv e attitudes toward physical and mental health, and improvement in their physiological symptoms (p. 354). Further positive effects have been docume nted (Teutsch, 2003; Mast, 2007), however recent concerns have surfaced that the patient-ce ntered approach may no longer be effective for all types of patients (Swenson, Zettler, & L o, 2006; Teal & Street 2008) Today, physicianpatient disparities are still occurring and dependi ng on the extent of these disparities, have a negative impact on health communication and th e medical encounter. These disparities range from a patients lack of conf idence or trust in th eir physician, to overa ll dissatisfaction among minority patients (Commonwealth Foundation, 2001; Ngo-Metzger, Telfair, & Sorkin, 2006). Medical dissatisfaction among minor ity patients is highly reported and may be indicative of unequal treatment associated with social and economic inequali ty, prejudice, or bias (Johnson, 2004). Physicians, patients and scholar s still want to achieve the best outcomes for the medical encounter. To accomplish this today, perhaps it is time for a new medical approach and new research ideas. Many researchers are still tr ying to improve the notion of patient-centered communication, but Arnold and Coran (2008) have proposed a new dyad-centered approach that


14 shifts the focus away from just the physic ian or patient. Instead, the physician-patient relationship is defined as a co -constructed, actively negotiating dyad. A dyad-centered approach institutes a more comprehensive framework. Furthermore, to accurately address the historic problems between physicians and patients, th e dyad-centered approach emphasizes an interdisciplinary combination of two paradigm s: relational communication theory and patientcenteredness (Arnold & Coran, 2008). Arguably, these two communication approaches have not been u tilized in unison within the medical field and furthermore, are not ade quate individually, because they are missing components necessary for consistently successf ul medical encounters. Additionally, neither approach suggests training both the physician and patient in ne gotiating medical consultations. Finally, the dyad-centered approach integrates the need for effective communication between physicians and patients in order to address di ssatisfaction among minorities, male and female patients, and communication disparities comm on in patients from lower income households. Perhaps this is the necessary next step in order to fix many problems in the medical encounter. Examining the physician-patient relati onship, as a co-constructed encounter, is a new direction that could help both physicians and patients account for their differences so each participant is more satisfied with the outcome However, a potential problem remains, how do you measure those differences? Th is study aims to introduce a ne w measure that can examine both physician and patient perceptions of the me dical interaction by de scribing the level of agreement between physician and patient percep tions after the medical consultation. This measure is called, dyadic concordance. Conceptualizing Dyadic Concordance The idea to m easure concordance or agreem ent between physicians and patients stems from a challenge initiated by the accomplished sc holar Debra Roter (2000) in an essay on the


15 evolving nature of the patient-physician relationshi p. In this essay, she put forth the challenge to explain the social context of the medical encoun ter with regard to both physician and patient characteristics. Furthermore, she emphasized the need to create new models of measurement to further the research area. Finding the appropriate measure and instrument to evaluate physicianpatient interaction has always been problematic. Two issues need to be examined at the same time: how physician communicatio n style impacts the medical encounter and how medical barriers facilitate or inhibit positive consulta tion outcomes. In order to identify the potential links be tween communication styl e and physician-patient barriers (i.e. dissatisfaction with care, patient perceptions of trust and medical competency, physician and patient expectations concerning medi cal visits), a comprehensive measurement is needed. Past instruments have only attempted to identify biomedical or patient-centered behaviors in the medical interview (Mead & Bower, 2000; Walker, Ar nold, Miller-Day, & Webb, 2002; Zandbelt et al. 2005) and have rare ly combined physician communication style with the identification of poten tial barriers involving a physicians or patients race, class or gender. Furthermore, studies trying to devel op valid and reliable measures concerning physician-patient communication (e .g. patient-centeredness) have been constrained by lack of theoretical clarity and the inevitable difficultie s of measuring relationship processes (Mead & Bower, 2000, p. 1091). Despite problems in measuring physician-pati ent interaction, anothe r component seems to be missing that can help explain why patient a nd physician characteristic s cause high variability in the medical encounter and ultimately lo wer satisfaction levels Past communication approaches have been strictly one dimensional, and therefore the focus has usually strayed to either the physician or patient. We need to re-conceptualize our ideas in order to more accurately


16 measure a co-constructed relationship. Does the physicians perception influence patient behavior or vise versa? How do perceptions shape the medical interaction? These questions are important and should be addressed when develo ping a measure to explore either perception. Instruments attempting to explore physician and patient perception individually may not the best option. Instead, a measure is needed that can account for both physic ians and patients perceptions at the same time and assess those perceptions for effective medical dialogue. Furthermore, we need to understand how medical perceptions change and why they may differ depending on participant characteri stics or medical specialties. To accomplish this task, the idea of concor dance was re-conceptualized into a useful interactional measure. Concorda nce is simply defined as a state of agreement or harmony; however, this definition can also be applied to medical perception. Utilizing the basic definition of the word concordance, the concept of dyadic concordance will describe the level of agreement between physician and patient perceptions after their medical encounter and provide a numerical measure for comparison. This dimension of medical interaction may allow us to further identify communication barriers or dispar ities between physician-patien t dyads. Therefore, the main objective of this thesis is to measure communication indirectly by focusing on the perception of each medical participant. The extent to which dyadic concordance can be applied using a pa ired-survey technique will be examined. Dyadic concordance will investigate dimensions of trust and medical competency, satisfaction with care, and expectati ons concerning the medical visit. To assess the feasibility of dyadic concordance this study will create a new inst rument, using six parallel items on separate physician and patient questionnaires aimed at iden tifying the level of agreement between physicians and patients perceptions of the medical encounter. This instrument will


17 contain the following six parallel questions: (1) assessment of the patients self-reported health, (2) amount of confidence and trust the patient has in their physicia ns medical decisions, (3) the patients education level, (4) assessment of patient pain rati ng, (5) perception of how well the physician explained the diagnosis, and (6) perception of how well the physician explained treatment options. Each item will function as a sample of an important clinical dimension that may influence successful medical encounters. Additio nally, the instrument will be utilized in two different medical specialties in order to identify any relational differences stemming from different medical disciplines. Th is level of analysis has not be en attempted in past research studies. If successful, dyadic concordance can add valu able information to the literature and for researchers when attempting to evaluate communicat ion disparities or relational differences. (e.g. physician and patient perceptions of trust and medical competency, satisfaction of care, or expectations concerning the medical visit). To summarize, the follo wing specific aims guide this study: To test if a paired-survey tec hnique can be used to explore dyadic concordance within the medical encounter. To explore and document relational differen ces within physician-patient dyads in two different medical specialties. Accomplishing these specific aims will help valid ate this new measure and explore patterns of disparities in perceptions that may affect physicia n-patient dyads. This could be considered the next step in understandi ng perceptions of the medi cal interaction and help describe the variability inherent in physician-patient dya ds. Furthermore, results from th is measure and from the pairedsurvey approach could provide healthcare professi onals with necessary in formation about factors that inhibit or facilitate communication behaviors.


18 CHAPTER 2 BACKGROUND AND SIGNIFICANCE This chapter provides the conceptual fr amework, literature on physician-patient interaction, and the rationale that guides this project. Chapter 2 will draw on a wealth of information pertinent to the physician-patient dyad and discuss how this relationship evolved into a viable research domain. A number of thes e sources are drawn across disciplines, including sociology, communication studies, and medicine. The synthesis of these fields creates a comprehensive understanding of what we current ly know about the physician-patient dyad and examines why patient dissatisfaction is incr easing within todays medical environment. Additionally, this chapter will ou tline the complexity of the medical encounter, methological techniques of past studies, and finally, will conclude with a set of research questions that guide this study The Evolution of Physician-Patient Interaction Medical historians have found ear ly depictions of the m edical encounter during the time of ancient Greek societies. The co mposition of the physician-patient relationship, along with its accompanying medical dialogue was first alluded to in some of Platos famous dialogues (Hamilton & Cairns, 1961). Since then, the field of medicine has grown exponentially, and with rapid growth has become a need to differen tiate and specialize. However, early medical interaction was quite nurturing, with the focus completely on th e patient (Shorter, 1985). This mentality started to change in the middle of the 20th century, when the locus of control shifted away from the patient and toward a more biomedi cal or scientific approach to medicine (Short, 1985). This new evolution in medicine emphasized efficiency and productivity, while also curing disease and illness. If patients were healthy as a result of doc tors work, then physicians were


19 meeting their primary goals. Prior to the 1970s, physicians followed a paternalistic model of communication, or a more biomedical model, in treating patients and teaching new physicians about the medical interview. Patient satisfaction regarding medical treatment was rarely considered as long as individuals were treated for the symptom or illness they brought to their physicians attention. Patients were viewed as obj ects of science, where a differential diagnosis and treatment would cure the patients medi cal problem. Although this approach improved medical sciences, diagnostic criteria, and vari ous treatment mechanisms, it lost sight of a fundamental aspect of medical care: th e patient (Balint, 1969; Roter, 2000). This conceptualization was first described in Parsons (1951) theoretical book, The Social System, where he portrayed the institution of medicine as a social systems normative mechanism for assisting patients who fall ill to retu rn them to their regular work-related capacity (Heritage & Maynard, 2006). Pars ons analysis described medi cal treatment as a set of generalized technical standards or a universalism approach, rath er than standards that are adjusted to the social charac teristics of the patient. Additio nally, Parsons (1951) went on to elaborate that physicians enact a specific technical focus on medical care rather than a general counselor type role. They treat patients withou t extensive emotional invo lvement and with their best welfare in mind regardless of the patients personal interests. Although Parsons intended to describe the f unction of illness, it also provided a new viewpoint of the medical field that illustrate d fundamental problems with medical paradigms during that era. Many scholars believed that medical care was suffering, but argued for decades regarding the exact cause (Szasz & Hollender, 1956; Foucault, 1975; Freidson, 1988; Heritage & Maynard, 2006). Shortly after this new medical direction in the 20th century, scholars began


20 looking for a new medical paradigm to serve as a solution and began by examining the element physicians had forgotten was a critical medical instrument; medical dialogue. Parsons ideas regarding the actions and mind set of physicians led to two ground breaking studies in the 1970s, which consequently, brought popularity to physician-patient interaction as a viable research domain and argue d the need of vast improvements. First, Korsch & Negrete (1972) conducted a study in Los A ngeles that observed 800 pediatri c cases and recorded parental and patient perceptions after the medical interaction. One fifth of the parents left the clinic without a clear statement or idea of what was wrong with their child, an alarming and confusing situation for any parent. Half of the parents reported that th ey did not even know what had caused their childs illness. Furthe rmore, a quarter of the parents re ported that they did not have the opportunity to ask thei r greatest concerns and additionally, were never encouraged to do so (Korsch & Negrete, 1972). Half of the parents left the medical visit with the opinion that physicians did not meet expectations and t hose physicians were grossly noncompliant (Heritage & Maynard, 2006, p. 351). This became one of the first studies to uncover major communication breakdowns in the medical consultation. The second study conducted in the 1970s anal yzed a physicians communication behaviors by examining 2500 audio recordings of primary care encounters (Byrne & Long, 1976). Utilizing Parsons description of the physician and Balin ts (1957) label for Parsons analysis as paternalistic (i.e. doctor-centered) behaviors, Byrne and Long set out to identify if these doctorcentered behaviors exist in medi cal consultations. Results showed that a majority of these behaviors were overwhelmingly doctor-centered, and therefore patients were not encouraged to provide their opinions or fully participate in medical discussions. Cassel (1985) and Cockerham (2007) both provide examples of doctor-centere d communication in their work and emphasize


21 that many physicians will start off conversations with closed-ended questions, or do not allow the patient to interrupt and ask for further explanations. Furthermore, Cassel (1985) transcribes situations where physicians refused to explain me dical conditions or anomalies because they felt it was beyond patients comprehension, thus resulting in patient misunderstanding and uncertainty concerning the medical consultation Freund, McGuire, and Podhurst (2003) believe a social distance was created between the physician and patient as a result of doctor-centere d behaviors. In this atmosphere, the physician was assumed to be the knowledgeable expert and the patient a recipient of the physicians advice. Information primarily flowed one way in th is interaction (from doctor to patient) instead of a mutual dialogue where the patients ag enda was fully discussed and encouraged. Furthermore, the greater the social distance be tween the physician and pa tient, the less likely either would be able to identify with the other, and the less sensitivity either was likely to have toward the others problems (F reund, McGuire, & Podhurst, 2003). Research by both scholars and clinicians has clearly demonstrated those patients who have satisfactory relationships with thei r physicians have better health outcomes (du Pre, 2005; Street et al. 2007) and thus an increas ed likelihood of complying with their physicians advice (Klingle & Burgoon, 1995). Eggly (2002) explains, eff ective communication during the medical interview plays an essential role in determini ng doctor-patient rapport; patient and physician satisfaction; patient adherence to the treatment plan, and, freque ntly the course of the illness itself (p. 339). For the first time, studies have suggested that doctor-cen tered behaviors are not conducive to a satisfactory health care relationship, and if physic ians want to regain patient satisfaction, they need to find a more comprehensive and satisf actory model of communication.


22 Research following Parsons description of th e physician-patient inte raction has added to our understanding of the experience, and in the late 1950s, the mu tual participation model was documented, where the management of chronic illn ess allowed the patient full participation in working with the physician to control the dis ease (Cockerham, 2007). One could argue that it was this model that led to the evolution of the relationship-centered approach to medical communication and the concurrent research to find the best style for communicating with patients. Arguably, the key to building a strong patie nt-doctor relationship and delivering high quality patient care starts with effective co mmunication. Although a plethora of communication models have been debated th roughout the literature to replac e paternalistic communication styles, researchers have advocated one sp ecific communication mode l, patient-centered communication, as the preferred replacement and a pproach to attaining effective communication (Irish & Hall, 1995; Mead & Bower, 2000; Teutsch, 2003; Swenson, Zettler, & Lo, 2006; Mast, Hall & Roter, 2007). The goals of patient-centered theory have provided a good starting point for discussing how to meet the needs of physicians and the expect ations of patients in th e medical field. Patientcentered consultations reflect r ecognition of patients needs and preferences, characterized by behaviors such as encouraging patients to voice ideas, listening, reflecting and offering collaboration (Mead & Bower, 2000, p. 1090). In ot her words, the patient becomes the central figure in the medical interview, and the physicia n allows a more interpersonal collaboration regarding the pati ents medical condition and history. Past reviews of patient-centered communicati on have depicted elements of the patientcentered approach, which focus on six similar com ponents: (1) exploring disease and illness as


23 an experience, (2) understanding the patient, (3 ) negotiating management of the illness, (4) incorporating prevention and a h ealthy lifestyle, (5) enhancing th e doctor-patient relationship and finally, (6) being realistic about time and re source constraints (Mead & Bower, 2000, p. 10871088). However, researchers have not been in agreement concerning the most important components of patient-centered communication. A co mpilation of studies (Cassel, 1985; Arora, 2003; Teutsch, 2003; Back & Hutchinson, 2006; Swen son et al. 2006; Mast, 2007; Haskard et al. 2008) have argued about other comp onents that should be the primary focus such as the patients concerns (along with fears and symptoms) or obj ectives for the medical visit, the patients perspectives, personal values (what is important to them), health beliefs, empathy for the patient, and finally non-verbal behavi ors displayed by the patient. Taken together it is hard to distinguish wh ich components should be taught to physicians or medical school students. These studies beca me ambiguous when trying to help the medical community truly grasp what patient-centeredness means, how it can be applied, or why it is a viable approach for physicians (Zandbelt et al. 2006). For example, although the original intention of this line of research was to address fundamental problems concerning medical communication, patient satisfaction and decreasing health outcomes, we are still searching for a definition that adequately and comprehensively e xplains what a patient-ce ntered approach is. The definition of patient-centered interaction has been reconceptualized repeatedly since the 1960s, ranging from Balints (1969) descrip tion of understanding the patient as a unique human being to McWhinneys (1989) approach of having the physic ian place himself in the patients world and attempting to view the dis ease through the patients eyes. Recently, Mead and Bower (2000) outlined their conceptual defi nition of patient-centeredness by explaining five key dimensions related to fifty year s of research. It is quite possible, this concept is too extensive


24 to sum up in one sentence or conceptual design and therefore, physicians take it upon themselves to interpret this framework and apply it in to their own practice with mixed results. Even if medical professionals and academi c researchers can not agree on a simple definition regarding patient-center edness, the overall concept can be explained in order to teach old and new physicians how to move away from u tilizing a traditional bi omedical approach in their practice. Today, many re searchers have accepted Mead and Bowers (2000) five dimensions as a viable framework to base future conceptualizations. Five Key Dimensions of Patient Centeredness The five key dim ensions of patient-centere dness, as identified by Mead and Bower are: the biopsychosocial approach, the patient-as -person, sharing power and responsibility, therapeutic alliance, and the doctor-as-person. These were orig inally conceived by Mead and Bower (2000) in their pursuit of a universal de finition for patient-centeredness. Although their goal of a simplistic definition did not reach fruiti on, they were able to explain the differences between the biomedical model and patient-cen tered model and provide direction for anyone wishing to improve health communication skills. First, Mead and Bower (2000) explained that physicians need to broaden their understanding of illness and how illness is percei ved by patients. Many illnesses presented in hospital teaching sessions can not adequately be classified by conventional illness behavior (Morrell, 1972). In these cases, a lack of sympto ms and a physicians reassurance that there is nothing medically wrong with them may compo und rather than relieve a patients suffering. Additionally, if a physician continue s to utilize a more biomedical approach in these cases, the patient may leave the office visit with a skewed perception ranging from physician incompetence to lack of faith in their physicians diagnostic skills. Conversely, patients who do not feel ill and report to the physician that they are fine may actually have so me classifiable disorder, for


25 example, hypertension. Thus, these examples provi de a physician with th e first limitations of utilizing the biomedical approach and the advantages of adapting a wider explanatory framework. Combining biological, ps ychological, and social perspectives is regarded necessary to account for the full range of problems pres ented in primary care (Mead & Bower, 2000). The second dimension, patient-as-person, emphas izes that the biopsychosocial approach alone is not sufficient for a full understanding of the patients experience of illness (Armstrong, 1979). For example, a torn anterior cruciate li gament (ACL) in a patients knee may not be experienced in the same way by two different people. An administrative assistant may experience far less distress than a constructi on worker, who relies on that knee to function efficiently in his/her occupati on. Decisions in medical treatmen t may not alleviate suffering for all patients, and some patients will need to weigh the costs and benefits regarding medical options (i.e. pain tolerance, cancer treatment). Mead and Bower (2000) emphasize that in order to understand illness and alleviate suffering, medicine must first understand the personal meaning of illness for the pati ent (p. 1089). Personal meaning is significant in understanding illness, and patient-centered techniques accentuat e the patient as an experiencing individual rather than the object of some di sease entity (Mead & Bower, 2000). Sharing power and responsibility, the third dimension in patient-centeredness, is fairly straightforward. Early critiques of medicine, as discussed in the previous section, illustrate a more paternalistic relationship be tween the physician and patient. However, this does not always lead to satisfying experiences within health care (May & Mead, 1999). Authors of the patientcentered framework advocate that both the physician and patient should participate equally in the general consultation. Lay knowledge of the patient is just as important as the professional knowledge gathered by the physician. Reports of pa tient non-compliance and dissatisfaction with


26 care were attributable to some failure on part of the doctors, not because of a mistake in medical practice, but because of a lack of mutual co mmunication (Tuckett, Boulton, Olsen & Williams, 1985). A therapeutic alliance, used consistently in psychology practices, empha sizes that the core of a therapist should include attitudes of empa thy, congruence, and unconditional positive regard if any change is expected in clients (Rogers, 1967). In the biomedical model, developing this notion is not as pertinent, but if the patient is to be treated like a real person and not just an object, inclusion of the therapeutic alliance is es sential. Since the patient-centered approach highlights the patient as an expe riencing individual, physicians s hould be aware of their verbal and nonverbal behaviors and consequently unde rstand the effect thos e behaviors have on patients. The last key dimension in the evolution of patient-centeredness, the doctor-as-person, concerns the influence of the personal qualities of the doctor. Balint, C ourtenay, Elder, Hull and Julian (1993) describe the biomedical model as one person medicine, whereas Mead and Bower (2000) describe patient-centered medicine as two-person medici ne. The physician and patient are both integral within their medical relationship and are always influencing each other throughout the medical interview. Mead and Bower (2000) further point out that physicians need to be aware of their emotional responses to pa tients and that subjectivit y, alone, could heavily influence patient perception in consultations. Mo dern theoretical frameworks including aspects of patient-centered care also emphasize how th ese subjective responses may shape positive or negative outcomes. The Significance and Complexity of the Medical Encounter The m edical consultation has dramatically evolved from the traditional days of in-home medical visits by physicians, to modern medical visits in state-of-the-art hospitals and offices.


27 Physicians have seen a rise in medical malpract ice lawsuits in recent years, many citing a failure in communication, as a primary reason for lit igation (Brenner & Bartholomew, 2005). Though patient-centered communication aids the physicia n in finding out more relevant information about their patients and helps promote patient sa tisfaction, it is not efficient in all medical specialties or with ev ery type of patient. Arguably, patient-centered frameworks have become increasingly hard to teach and evaluate (Zandbelt et al. 2005). A dditionally, the lack of a central definition makes this concept almost impossible to teach in medical schools throughout the United States. Furthermore, the idea of patient-centered communication as a rela tionship building strategy between physicians and patients is a good idea, but still focu ses on the physicians communication style and therefore, continues to focus more on one participant than the othe r. This could lead to additional communication barriers or health di sparities in the medical encounter. Health communication is only one ingredient that complicates medical interaction. The influence and perception of health disparities regarding a patients gender, race and socioeconomic status also need to be accounted for. Each demographic component has been found to significantly affect not only the medical encounter, but also a pa tients overall health. Many physicians do not have sociological training and therefore, ma y not be aware of the current literature on disparities of disadva ntaged groups or how to interp ret a patients perception during the medical encounter. The following subsections present literature on patient sociodemographic characteristics and their importance to physician-patient interaction. Racial and Socioeconomic Influences Differences in overall health throughout the Un ited States stem from social inequality when comparing the health profile s of various racial groups. In general, Asian Americans have typically enjoyed a high quality of health, while African American s are especially disadvantaged.


28 Hispanics and Native Americans also have health disadvantages when compared to white, nonHispanic individuals (Cockerham, 2007). Recent studies have found that blacks, Hispanics, Native Americans, and Asians receive less adequate and less intensive health care than whites and additionally, such disparitie s persist even after taking into account health insurance status, age, sex, income, and education (Johnson, Saha, & Arbelaez, 2004). A comparison of the life expectancies of Af rican Americans and whites shows that black males are especially disadvantaged with respec t to longevity (68.8 years in 2002, and on average, 6.3 years less than white males). Black females follo w a similar pattern with a life expectancy of 75.6 years, which is 4.7 years less than white fema les (Cockerham, 2007). Underlying issues that result in this disparity have b een traced to higher rates of cance r, heart disease, hypertension and AIDS (Cockerham, 2007). Socioeconomic factors such as poverty, marginal employment, low incomes, segregated living conditions, and inadequate education are more common among blacks than whites, and furthermore, contribute to poorer health and widen the gap concerning health disparities (Link & Phelan, 1995). In regards to the availability of medical treatment, race and socioeconomic status are closely linked, and minorities are more disadvant aged compared to whites when more complex forms of medical treatment are needed. Cockerham (2007) explains this di sparity in regards to African Americans: Blacks, for example, are much less likely than whites to have heart by-pass surgery, appendectomies, and other surgical care, a nd they receive fewer mammograms and tests and drugs for heart disease and diabetes. The availability of physicians and hospitals providing quality care where blacks live is a ma jor reason for the differences in care. For example, the few cardiac surgeons in predomin antly black communities, especially in rural areas, helps explain why blacks receive fewe r coronary artery bypass operations than whites (p. 57). Additionally, a lack of financial resources and information about h ealth are important barriers to health care for low-income minorities.


29 Unfortunately, health data on Hispanics in the United States is limited because, until 1976, federal, state, and local agencies grouped Hi spanics with the non-Hispanic white category (Cockerham, 2007). Although, Hispanic s relative overall health prof ile remains slightly better than blacks, they are overrepresented among poor or near poor Americans and this culminates in higher rates of diabetes, hypertensi on, tuberculosis, STDs, alcoholism, and cirrhosis of the liver (Rogers, Hummer, & Nam, 2000). A report on unequal treatment issued by the 2002 Institute of Medicine suggested that health disparities are not explained entirely by diffe rences in services or access, but instead exist in the broader context of social and economic inequality, prejudice, and bias (Johnson, 2004). Researchers at John Hopkins University reported that blacks, Hisp anics, and Asians were more likely than whites to agree that they would rece ive better medical care if they belonged to a different race; and that medical staff judged them unfairly or treated them with disrespect because of their race or how well they s poke English (Johnson, 2004). Within the published findings, Johnson noted that, the se differences were somewh at diminishedbut not fully explainedby differences between r acial groups in demographics, he alth status, source of care, or patientprovider communicati on. We conclude that, overall, racial and ethnic minorities perceive bias and a lack of cultural competence in the health care system as a whole. Cultural competence can be defined broadly as the need for health care profe ssionals to recognize and respond to their own and their pati ents' cultures (Teal & Street, 2008). Cultural competency is important in medical communication, and sensitivity or adaptation of different communication styles, by the physician, may be needed for interactions with patients across ethnicities. Some evidence suggests that bias and stereotyping by clinicians may contribute to differences in pati ent care and racial disparities in health outcomes (Teutsch, 2003).


30 Physicians lacking cultural competency or not recognizing the n eeds of their minority patient, may become frustrated and less effective in their communication efforts during the medical interview (Teutsch, 2003). Sympto m reporting by patients may reflect their own distinct cultural values, and if not recognized by physicians, co uld lead to less accura te diagnoses. A couple examples of this stem from regional culture diffe rences among ethnicities in the United States. In southern California, a commonly held belief by Latinas is that breast cancer is attributed to sinful behavior, thus many of these indi viduals will either seek out re peated voluntary breast screenings or lack thereof based on their own perceptions or beliefs if they discover a breast lump (Chavez, Hubbell, McMullin, Martinez, & Mishra, 1995). Another example involves the stigma of depression among Asian Americans, which is generally unreported, yet many Asian patients report fatigue and weight loss in the medical interview similar to depressed patients. Finally, among Hispanics in New York, many will lead phys icians away from recommending surgery to treat breast cancer because they think surger y leads to metastases (Kundhal & Kundhal, 2003). Minority patients also report lo wer satisfaction levels when as ked about their medical care. A study conducted by Ngo-Metzger and colleagues (2006) reported that Asian Americans were less likely than white patients to report that they were very satisfied with their care overall and less likely to have a great deal of trust in their doctor. Furthermore, minority patients and individuals from lower socioeconomic backgrou nds tended to receive less health-related information from their physicians compared with non-minorities and individuals from higher socioeconomic backgrounds (Ngo-Metzger et al. 2006). Another study (Schoenbaum, Collins, Tenney, Hughes, & Audet, 2002) found that across all racial and ethnic groups, respondents who perceived that their doctors did not understand th eir background and values, or did not listen to everything they had to say, were less likel y to be very satisfied with their care.


31 The Influence of Gender Like race and incom e characteristics, gender can also influence perceptions in the medical encounter and may start with a ma n or womans perception of overa ll health or their quality of life. Today, males exceed female death rates at all ag es and also lead females in the top causes of death: heart disease, cancer, ce rebral vascular diseases, a ccidents, and even, pneumonia (Cockerham, 2007). Although men may die sooner, women fall ill at much higher rates, and thus may not live the best quality of life at older ages. The National Center for Health Statistics (2004) reported that women have higher rates of acute illness, including parasitic disease, digestive, and respiratory conditi ons. Women also have higher rates of chronic illn ess, including hypertension, thyroid conditions, chronic enteritis a nd colitis, diabetes, and arthritis (Cockerham, 2007). All this adds up to a lif etime of medical intervention for both genders, but how each gender perceives that interaction may be di fferent for the patien t and the physician. Gender equity within healthcare has also been a major concern in medicine. Research in the 1980s and 1990s focused on females as a underprivileged group, but in recent years there has been a shift away from talking about women as underprivileged, to talking about gender (Doyal, 2000). Current research examining the two genders in health interactions emphasizes the social construction of gender identitie s (Doyal, 2000) and the nature of the relationship between different combinations of male -female physician-patient dyads. Today, researchers identify gender as one sour ce of systematic variation in the medical encounter (Doyal, 2000; Street, 2002; Gorman & Read, 2006). Issues stemming from gender differences are not just limited to the patient but some research does show that patients, regardless of sex, often vary their medical resp onses depending on the physicians gender (Roter & Hall, 1993).


32 Female patients generally receive more info rmation, ask more questions, and have more partnership building with physicians than male patients (Cooper, 2002). Le ss is known about the effects of physician gender, but similar studies ha ve shown that female physicians exhibit more empathy, engage in more positive talk, partne rship building, question asking and information giving compared to their male counterparts (Coop er, 2002). In other words, female physicians and patients are more likely to engage in pati ent-centered behaviors du ring the consultation. However, research does suggest that male and female physicians may approach the patient interview differently, which could lead to a variation in clinical characteristics. In the past, physicians expressed difficulty in diagnosing women with heart disease, due to a common stigma that coronary heart disease was an illness f ound primarily in older men (Bakalar, 2007). Adams et al. (2007) examined 112 male and female primar y care physicians while they watched video of professional actors portraying patien ts who all had medically apparent symptoms of heart disease (Adams et al. 2007). Findings suggest that both male and female physicians picked up more psychological cues from female patients than from male patients. Additionally, even though older aged patients are more at risk for heart disease, female physicians paid significantly less attention to female patients ages than those of males (Adams et al. 2007). Examining Complexity and Agreem ent in the Medical En counter The previous subsections outlined potential pr oblems with demographic disparities that may influence the medical encounter. Numerous approaches have tried to examine how these disparities influence physician-pa tient interaction and why they are important. Strong evidence has reported that cultural differen ces between physicians and patient s are serving as a barrier to effective communication (Betancourt, Green, Ca rillo, & Park, 2005). Dissa tisfied patients, are complaining about poor communication and are quite frustrated, and as a result are currently filing malpractice claims (Brenner & Bartho lomew, 2005). Breakdowns are occurring during


33 patient handoffs to medical resi dents and these residents may not be consistently notifying their attending physicians of critical post-operative ev ents (Greenberg et al. 2007). Patients are not telling physicians what is important to them and instead may just display non verbal hints (Halpern, 2007); however, non verbal behaviors are often misinterpreted by physicians (Mast, 2007). Employing a patient-centered approach may not solve all negative outcomes. Some patients may prefer a different style of communication, and theref ore no single approach may be adequate for all patients. For instance, many older patients prefer a more doctor-centered relationship and would rather leav e important decisions at the so le discretion of their physician (Irish, 1997). Other patients have adopted a consumers approach to the medical encounter, seeking as much information as possible, and u tilizing their physicians as consultants (Haug & Ory, 1987). Accounting for each negative outcome and attempting to identify why the medical encounter drastically varies from one situation to another is a difficu lt task. Concerns about variations in standards of medical care throughout medical sp ecialties have cr eated a surge in research to quantify key components regard ing quality of care (Roland, 1999). To my knowledge, no study has been able to account fo r the wide variation in potential outcomes involving the physician-patient dyad. This has le d to a number of studies only isolating one variable and trying to explain as much variati on as possible in that one component of medical interaction. Measuring Physician-Patient Interaction Due to the com plexity of the medical encount er and the variability between physicians and patients, scholars have struggled to reliably m easure physician-patient inte raction. At the time of publication, Mead and Bower (2000) found twenty-t hree different studies that attempted to


34 measure the medical interaction utilizing a plet hora of methodologies. However, most fell short in some aspect of reliability and/or validity, and thus have not developed into a universal measuring instrument for other researchers to apply. Finally, none of th e twenty-three studies were able to effectively measure all five of Mead and Bowers dimensions regarding patientcentered communication. It has been suggested that a pa tient-centered approach to medi cal care is contingent on the physician possessing certain attit udes and values (Grol, de Maeseneer, Whitfield, & Mokkink, 1990), personality traits (Balint, 1964) or ev en specific communication styles (McWhinney, 1985). Methodologies incorporating these elements into statis tical scales that measure physicians relationship-centered at titudes or values are usually from self-report questionnaires that examine a myriad of psychol ogical attributes. Arguably, there are a number of reliability and validity issues associated with these types of scales and self-report questionnaires. For example, Cronbachs alpha scores may be low, and thus the internal consistency of these scales may come into question (Mead & Bower, 2000). Deleting specifi c items from scales in order to increase their internal consistency may re strict the measurement of physic ian-patient interactions due to high variability in human behaviors and communication styles. Self-reported scales are not the only methodologies currently being implemented when attempting to measure the complex social environment between physicianpatient dyads. Since the paradigm of patient-centeredness can be concep tualized clinically, another research approach involves the observation of consu ltation behaviors. Such observati onal studies have contributed to different approaches: first, a rating scale that focuses on how much or how well a specific physician behavior was performe d, and second, a verbal behavior coding system that involves


35 categorizing and coding specific el ements of physician and patient dialogue. These approaches have also been used individually and in combination with mixed results. Verbal behavior coding or other interaction coding instrument s appear to be frequently used and adapted throughout the literature. The most reliable in clude Bales (1950) Interaction Process Analysis (IPA), Stiles (1978) Verb al Response modes (VRM) and Roters (1977) Interaction Analysis System (R IAS). More recently, a new verbal coding instrument has entered the literature, Zanbelt an d colleagues (2005) patient-centered behavior coding instrument or PBCI. The PBCI was developed to determine the occurrence of physicians facilitating and inhibiting behaviors and also ut ilizes the Eurocommunication scal e with the goal of explaining more physician-patient variation th an other conceptual approaches. All of these coding schemes share the same function of sorting me dical dialogue into mutually exclusive categories. Most often, these instruments us e videotape or audiotapes to analyze the frequencies and proportions of speech and then focus on discovering the intent or effect of the chosen words/phrases rather then th e literal meaning of what is actually said. This type of methodology is effective if the researchers are trying to analyze a specific physician behavior or its effect on patient satisfaction; however, they struggle to explain why multiple physician or patient behaviors may cause one outcome over another. If that outcome is negative, it becomes difficult to assess a solution to the presented problem. Problems measuring patient-centered interact ions highlight its conceptual problems. Multiple definitions have caused too much variab ility in the types of measures utilized to evaluate medical interactions. If physicians are not taught th e same communication style or scheme and are left to navigate the medical en counter individually, than it becomes too difficult to explain varying results. Schol ars are in agreement that both physicians and patients want a


36 satisfactory and efficient medi cal consultation, while at the same time improving patients overall health, but the question remains: how do we achieve this ultimate goal? Influences of Physician-Patient Relationship Further exploration of the m edical encounter may need to include an examination of physician and patient perceptions concerning the medical consultation. Arguably, examining perception is a complicated procedure due to the unpredictability of human behavior. Figure 2.1 shows how potential physician-pa tient factors may influence thei r medical satisfaction, medical decisions, and the relationship between physicians and patients. These factors are further divided into micro and macro level construc ts in order to show that both theoretical divisions need to be accounted for when investigating the physician-patient relationship. First, the status of the physician and patient must be reported, which requires the recognition that sociodemographic variables will shape persona l values of both physicians and patients. In turn, e ach participants status and associated values will impact thei r own perception of the medical encounter along with their knowledge of communication skills. Ot her influences would in clude the physicians communication style and their level of cultural competency. The idea behind Figure 2.1 is to show that if we examine the patients percep tion and the physicians perception regarding the medical encounter, then we may be able to explain how those perceptions influence other individual characteristics in medical interaction. Physician-patient agreement has been examined in end-of-life care discussions (Desharnais, Carter, Hennessy, Kurent, & Carter, 2007), within the validation of measuring instruments (Ahlena, Mattsson, & Gunnarsson, 2007), and in examining whether providing feedback to the doctor can improve physician-patient concordance (Liaw, Young & Farish, 1996). Although, these recent articles have star ted examining physician-patient agreement and assessing what that means, a large amount of work remains. For instance, the effect of patient or


37 physician demographics on concordance levels was rarely discussed, only one study linked communication style to variati on in concordance, and to my knowledge, no study has attempted to measure and describe concor dance variation between medical specialties. Therefore, this appears to be viable research domain in need of attention. Few studies have attempted to account for the congruence or agreement between physicians and patients. Krupat and colleagues ( 2000) investigated the extent to which the individual orientations of physicians and patients and th e congruence between them are associated with patient satisfaction. Their concor dance instrument examined which roles doctors and patients believed each should play in the cour se of their interacti on. Results showed that patients were highly satisfied with their physician when the physician had a matching orientation or was using a patient-centered approach (Krupt et al. 2000). Another study explored physicianpatient congruence and its associ ation with patient sa tisfaction, adherence, and health (Jahng, Martin, Golin, & DiMatteo, 2005). Results indicated that when patients and their physicians shared similar beliefs about patient participati on, patient outcomes tended to be more positive. Research Questions The overall intent of this thesis is to e xplore physician-patient pe rception with a m easure of dyadic concordance and create an instrument that can exam ine this measure in a variety of ways. In the past, studies have developed measur ing instruments that primarily focus on only one specialty, only a few dimensions of patient-cen teredness, or only one interactants perception. Additionally, the inability to effectively meas ure communication style and its effect across different combinations of the physician-patient dyad, has led to disagreement among physicians and scholars concerning the most efficient style of the medical interview. Examination of the physicians perception, patients perception, and family or spous es perception have all been


38 studied previously. However, few studies have tried to examine the physician and patients perception at the same time. This study will add to the curr ent body of literature by examining dyadic concordance and how it can be applied in order to provide additio nal information to aid this area of research. Furthermore, this study will investigate how pe rceptions differ between patient demographics and between two medical specia lties. Upon successful implementation, the measure of dyadic concordance could be used to improve the re liability and validity of current measuring instruments. Additionally, using dyadic concor dance to describe the variability between physicians and patients could be useful as an ev aluative tool for any heal thcare professional and in any medical practice. The strength of this pe rception measure lies in its flexibility; therefore the following research questions are posed: Can a paired-survey technique can be used to explore dyadic conc ordance within the medical encounter? How do the perceptions and relational diffe rences of physician-patient dyads differ between the medical specialties of oncology and family practice?


39 Figure 2.1. Factors That Effect Satisfac tion, Medical Decisions, and Relationshi p between Physicians and Patients Patient Centered Doctor Centered Communication Style Perception Nonverbal Verbal Communication Skills Race Gender SES Health Education Attitude Gender Cultural Competency Macro Constructs Macro Constructs Micro Constructs


40 CHAPTER 3 RESEARCH DESIGN AND METHODS Physician-Patient Questionnaires In order to m easure the physicians and patien ts perceptions of th eir interaction, four individual questionnaires were cr eated: separate profile sheets for physicians and patients (see appendix A & B), and two parallel survey inst ruments (see appendix C & D). Profile sheets collected sociodemographic ch aracteristics (e.g. gender, race /ethnic background, birth date, income) from participants and in addition, the patient profile sheet included three extra questions to identify the patien ts household income, occupation, a nd experience with the corresponding physician. Both survey instruments asked physicians a nd patients about their medical encounter and documented participants perceptio ns of their medical communicat ion in order to assess dyadic concordance. This new approach to measuring per ception allows us to look at the concordance or discordance of each participants answers concerni ng their medical interaction. Six parallel items measure perceptions regarding the physicians explanation of dia gnosis, explanation of treatment, patients level of pa in, patients education level, amount of confidence and trust patients have in their physicians medical deci sions, and finally patients self-reported health. These questions can be grouped into three areas of importance regarding the medical encounter: (1) dimensions of trust and confidence, (2) physician-patient percepti ons of key clinical elements, and (3) perceptions of overall interaction. Although the strength of this instrument lies within its measurement of dyadic concordance, additional patient pe rceptions regarding their medical interaction can be examined. Two groups of questions were included to asse ss medical satisfaction a nd quality of care. The


41 first group details patients perceptions of conf idence and trust in their physicians medical decisions, and is explained by three survey items: How much confidence and trust did you have in the doctor treating you? Did the doctor treat you with a grea t deal of respect and dignity? Did the doctor understand your background and values? The second group examines patients percepti ons concerning physicians level of listening, inquiry into unanswered questions, and am ount of shared decision making. This was accomplished from the following three questions: Did the doctor listen to everything you had to say, strongly agree, agree disagree, or strongly disagree? Did you have questions about your care or treatment that you wanted to discuss, but did not? Did the doctor involve you in decisions about your care as much as you wanted, almost as much as you wanted, less than you wa nted, or a lot less than you wanted? In summation, the patient survey covers important areas of the physician-patient consultation, while also measur ing dyadic concordance between si x parallel questions. Patterns from each of the included topics were analyzed for known health disparities documented in past research studies. This provided the necessary data to test the validity of the overall instrument. Sample A total of ten consenting physicians and fift y patients were enrolled from a convenience sam ple of family medicine and oncology clinic s in the southeast. For each physician, five patients were surveyed for a total of fifty dya ds. This exploratory study utilized a convenience sample, so therefore may not be representative. Equal distributions of participants have been collected from both medical specialties. Although the physician sample was small, six male physicians and four female physicians provided an almost even divide between the tw o genders. However, the convenience sample only


42 provided three minority physicians (1 African American, 1 Hispanic, 1 Asian); without collecting more data, this sample will not be suffi cient to test any physicia n racial associations among concordance variables. Physician ages ra nged from 31 to 59, with a median age of 42, and a total of seven are categorized under the ag e of 50. Physicians reported a wide variation of medical experience (after residency), ranging from less than 5 y ears to more than 20 years of practice. Only two physicians engaged in formal communication training after medical school. Patient characteristics were evenly dist ributed among the convenience sample, and contained 29 males (58%), 21 females (42%) and 19 minorities (38%). Racial distribution among patients included 10 African Americans (20%), 6 Hispanics (12%) and 3 Asians (6%). No other racial categories were reporte d among participants. The ages va ried extensively among patients encompassing a range of 19 to 92, with a median age of 54. More than half of patient participants were over 50 years of age. Two thirds of patient participants reported a yearly household income of over $60,000, and 64% completed at least some college coursework. Interestingly, all participants in the study completed at minimum a high school diploma. One third of patients did not pursue any post-high school coursework. Recruitment and Procedures After obtaining IRB approval, physicians were contacted via em ail, phone call, and prearranged one-on-one appointment s to introduce the study and gauge their interest. Upon showing interest in the study, meetings were arranged with physicians to discuss the informed consent process and overview of the study. Once enrolled, a physician profile sheet was given to the doctor and once complete, was filed under a sequ ential physician ID number. Only one profile sheet was filled out per physician and all iden tifying information was kept confidential. Several physicians were interested in discussi ng the project, but afte r meeting decided to opt out and not participate in the study. Sixteen physicians were contacted and six physicians


43 choose not to participate. Out of the six declinin g physicians, five specialized in family medicine and only one specialized on oncology. It is poss ible that family medicine physicians were worried about receiving negative evaluations duri ng the study, even though all results would be kept confidential. Perhaps, they felt their time was too limited. Finally, some physicians may be aware of inefficient communicati on skills and not want to be ex amined, or may be generally disinterested in scientific research. Physicians were allowed to withdraw from the study at any time, however, none of the enrolled physicians elect ed to withdraw over the course of the study. Patients of enrolled physicians were appro ached during clinic or office hours during the time period in which the principal investigator was present. Consent was sought and obtained prior to each patients medical appointment (prefe rably in a private exam room while waiting to see their physician). Since the in stitutional review board strongly regulates screen ing procedures, due to the federal health inform ation protection act (HIPPA), patie nts could not be screened prior to their medical visit for any demographic info rmation. Only patients who were over the age of 18 were included in the study. Upon patient consent, each was assigned an ID number and dyad number that would relate them to their corresponding physician. Patients recei ved a profile sheet to complete prior to the medical consultation. Physician interaction surveys were inserted into patients medical charts for the physicians to fill out im mediately after the interaction. In order to protect patient confidentiality, two procedures were followed: first, physicians were never informed about enrolled patients, and second, surveys were also placed into non-enrolled patient records randomly so physicians could not track or identif y enrolled patients. Physician questionnaires were completed after every medical consultatio n and surveys regarding non-enrolled patients


44 were separated and discarded by the principal investigator after in dividual study sessions. Enrolled physicians were briefed on this procedure prior to their enrollment. Patient surveys were administered after their consultations and only after the physician left the examination room or office. The principal investigator was not in the room during the medical interaction in order to avoid any conf ounding effects concerning participant observers. However, the investigator did remain in the room while the patient filled out the questionnaire in order to answer any questions he or she may have concerning the survey, study or procedures. Family members or friends of the patient were allowed to remain in the room strictly at the patients discretion, but were not allowed to influence the patients answers (i.e., suggesting answer choices, giving their opinion on answer c hoices, or asking the pati ent to explain their answer choice). Patients with cognitive or serious communication impairments that may limit patients ability to complete the survey were not enro lled in the study. Although patients could not be formally screened regarding cognitive, physical or emotional difficulties, their informal interaction was noted by the investigator and could be withdrawn, if serious concerns arose during that interaction. Ultimately, none of the en rolled patients depicted any cognitive, physical, or emotional difficulties that would jeopardize this study and as a result, no patients were withdrawn.


45 CHAPTER 4 RESULTS This chapter provides results of analyses conducted to answer four prim ary research questions. In the first section, descriptive statis tics illustrate the sample distribution for both physicians and patients. The sample distributi on is primarily extracted from physician and patient profile sheets that accompanied the pair ed-survey instrument. Next, two sections are devoted to individual physicia n and patient findings compiled from physician and patient surveys. Individual physician re sults focus on descriptors con cerning medical training, their perception of communication training for physicians, and fina lly, prior experience with patient participants. The corresponding pa tient section examines repor ted health descriptors and variables describing patient perception in the medical encounter. Finally, the last two sections of this chapter are devoted to m easuring dyadic concordance. Dyadic concordance is examined through the use of cross-tabulati ons and significance tests which examine each parallel question for all fifty physician-patient pairs. Further tests compare conc ordance results to ordinal and dichotomous dummy variables from the physician and patient profile qu estionnaires. The last section provides results from a more in-depth analysis that illustrates how concordance methodology can explore interactional differences between the medical specialties family medicine and oncology. Physician Medical Characteristics Table 4-2 exam ines the physicians characteri stics which include clinical training and experience with enrolled patient s. The ten physicians reported a wide variation of medical experience (after residency), rang ing from less than 5 y ears to more than 20 years of practice. Only two physicians engaged in formal comm unication training after medical school. Since advanced health communication is not mandatory after medical school, it was surprising to find


46 two physicians with training with in this small sample. One physician received advanced training in the form of a seminar class and the other at tained training at a na tional conference. Even though only two physicians in the study sought additional communication training, 9 out of 10 physicians felt communication skills were extremel y important in medical practice. Since prior medical experience with the same patient could faci litate better interaction and therefore lead to higher dyadic concordance, each physician-patient dyad was asked about their experience with each other. Among the 50 physician-patient dyads, 32 (64%) confirmed that this was their first visit with each other. Patient Frequencies and Individual Results Table 4-3 reports patient health descriptors and encounter characteri stics regarding their m edical visit. Three quarters of patients reported some degree of pain with a mean score of 5.5 out of 10. Self-reported health was negatively skew ed, with 68% of patients reporting only fair or poor health. Arguably, this ma y reflect the recruitment of half of the patient sample from oncology clinics; the studys inclusion criteria did not exclude pa tients based on illness or any other health issues. Questions regarding patients perceptions of confidence and trust showed that 86% of patients reported a great deal or fair amount of confidence a nd trust in the physician treating them. Only seven patients (14%) reported little confidence or trust in their physician. However, frequencies do show that among the 32 patients w ho were seeing their physician for the first time, 93% reported having a great deal or fair amount of confidence and tr ust in their physician. Furthermore, almost all patients ( 48 out of 50) felt their physician treated them with a great deal or fair amount of respect and dignity. Patient variance did exist when asked if their physician understood their background and values. Only 62% of patients re sponded favorably (strongly agree or somewhat agree) compared


47 to 24% responding that they somewhat disa greed and 14% responding that they strongly disagreed. Among the minority patients included in the sample, 74% disagreed when asked if they thought their corresponding physic ian understood their backgr ound and values. Additionally, they were five times more likely to disagree compared to non-minorities and chi-square results confirmed a significant associa tion (p < .001). Two additional si gnificant associations existed among minority patients, over half felt their physician sp ent less time than they wanted (p =.058) and 53% reported they had unanswered questions following the medical consultation (p =.054). Furthermore, numbers within some crosstab cel ls were less than the expected count when computing the chi square analyses, suggesti ng that a more robust sample may add more statistical power and increases significant asso ciations. Additional anal yses did not reveal associations of age, gender, or household income with the patients percep tions of whether their physician understood their background and values. Most patients (78%) felt thei r physician listened to everythi ng they had to say, and 64% did not have any unanswered questions. Additiona lly, 76% reported that their physician involved them as much as they wanted during the medical visit. These results suggest that a majority of patients in this study had a positive view of their consultation with their physician. Dyadic Concordance among 50 Physician-Patient Dyads Dyadic concordance was m easured by comp aring the responses from physicians and patients on six concordant items that appeared in both physician and patient questionnaires. These parallel questions assesse d physician and patient percepti ons of (1) patient self-reported health, (2) patient confidence and trust in their physician, (3) patie nt education level, (4) patient pain rating, (5) physician explanatio n of diagnosis, and (6) physicia n explanation of treatment.


48 A seventh question on the physician survey asked if the physicia n inquired about any unanswered questions before the end of the medical consultation, while a similar question on the patient survey asked if the patient left the medical consultation with remaining unanswered questions. Although these two questions were not ex actly parallel, both desc ribed another aspect of interaction that may affect patient satisfac tion and treatment adherence. Furthermore, one could argue that addressing patie nt questions is crucial in dya dic medical interaction and may improve patient outcomes after a medical visit. Therefore, this area will also be examined alongside the dyadic concordance results. As presented earlier, dyadic concordance among physician-patient dyads was defined by a physicians correct estimation of their patients response, on a given subject relevant to the medical consultation. Exact concordance resulte d when the physician response and the patient response were equal on two parallel questions. Arguably, a physician response that deviated by +/1 from the patient response may not have be en that different from exact concordance and since this is a measure of human interaction, so me error should be acceptable. Therefore, for the purposes of this study, physician re sponses that were either exact or deviated by +/1 from the patients response were considered concorda nt. Dichotomous concor dance variables were created following the above concordance rules and utilized to illustrate frequency distributions and examine within crosstab analyses. Dummy variables were constructed for patien t gender, age, education, income, race, and experience with their physician. Cross tabulations (2x2 tables) we re constructed to investigate the relation of independent vari ables with dyadic concordance. When applicable, chi-square analyses were run to test for significance. Tw o measures were used to determine significant associations among independent variables. First, if the statistical te st was run between two


49 dichotomous variables, than Fisc hers exact test was used. Fishe rs exact test is recommended when trying to analyze associa tions among small sample sizes. Fi nally, since Fishers test can only be utilized in 2x2 analyses, a ll other chi-square tests used Pearsons measure to determine statistical significance. The alpha level was se t to 0.10 since the study included a small sample size and the project was designed as a preliminar y study to test the effe ctiveness of the dyadic parallel survey method. Total Concordance Total concordance was m easured by assessing all potential interactions that included agreement between the patient and physician. As the survey instrument included six concordant or discordant opportunities per me dical encounter, with a sample of 50 dyads, this provided a total of 300 opportunities where the physician and patient could ag ree regarding their consultation. Arguably, numerous concordant opportunities exist in a medical encounter, and although this a valuable measure to assess, it wi ll always be restrict ed by instrument design. Among the 300 concordant opportunities, a physician or patient selected dont know, or did not include an answer 48 times. This provided a new total of 252 measurable opportunities for the physician and pati ent to agree. Listed in table 4-4 is the fr equency of total concordance by the patients characteristics. Male patients were the same (32%) compared to female patients (30%). A large difference existed between the discordance of minority a nd non-minority patients. Mi nority patients were more likely to not agree with their physicia n, which resulted in 37% discordance. The perceptions of non-minority patients matched within +/1 to th e response of their physician 76% of the time. Older patients were more conc ordant (73%) than younger patients (65%) and patients with previous experience with their phy sician produced a higher level of concordance (70%). The patients reported household in come only produced a slight difference in


50 concordance between those who earned sixty thousand a year (69%) and patients who earned less (70%), however, regardless of income, one-third of these patients produced a discordance response. Finally, patients with co llege degree were more likely to be concordant (74%), than patients without a four year degree (65%). These results provide evidence that gender, minority status, education, and age may have the largest impact on the level of agreement between the physician and patient. Although, each patient characteristic may have a different effect on the medical en counter, larger sample sizes are needed to report larger di stributions of disparities when comparing levels of dyadic concordance. Concordance and Patients Self Reported Health Two-thirds of patients scored their self reported health (S RH) as poor or only fair. On average, physicians rated their patients SRH as 2.32 (out of 5), com pared to a patients average rating of 2.10, with lower values representing poorer health. Most physicians either guessed or assumed the patients SRH in 34 consultations; however in 16 interactio ns (32%) physicians responded that they did not know their patients SRH. This leads to the question, why would some physicians mark a definitive response while others did not attempt a guess? Table 4-8 shows th e number, frequency, and chi-square results of dont know (DK) responses by a patients sociod emographic characteristics and their prior experience with the corresponding physician. Male patients caused more inconsistent responses from their physician compared to female patie nts, since 75% of DK responses occurred with a male patient (24% of total interactions). Additio nally, Fishers exact test showed patient gender to be statistically significant (p .10). No further significant associations were found when analyzing DK responses with patients race, income, or age.


51 DK responses were also compared to first visits among physicians and their patients. Eleven out of 16 (p < 0.01) of the DK responses occurred when a physic ian met with his/her patient for the first time. Table 4-10 shows th e distribution of DK responses by a physicians clinical experience. One might expect that phys icians who were more experienced would be more aware of their patients SRH. However, a majority (87%, p < 0.05) of the DK responses were from experienced clinicia ns (10+ years experience). Table 4-11 presents the crosstabulation between physician and patient responses regarding the patients SRH. Bolded numbers along the diagonal represent exac t concordance between physician and patient. DK responses were not included in the conc ordance/discordance analyses and were treated as missing data. Dyadic concor dance (see Table 4-5) occurred in 29 of the 34 interactions (85%). On this first parallel item, only five interactions we re discordant and when physicians and patients were not in agreement, the difference between both responses was quite small. Discordant responses primarily overestim ated the patients SRH, and some examples would include physicians rating the patients SRH as excellent when the patient rated SRH as only fair. Tables 4-6 and 4-7 summarize the extent of concordance by patient s sociodemographic characteristics, education, and experience between physicians and patients on the parallel items for SRH. Physicians misperceived a females SR H more often than males (4 out of 5 times), younger patients more often than older patients (4 out of 5 times), and patients with higher incomes (4 out of 5 times). Although, first time patients might be expected to have higher discordance with their physician when compared to returning patients, the opposite relationship occurred. Four out of five disc ordant interactions occurred wi th established patients. However due to the small sample number of discordant inte ractions, Fishers exact test could not estimate


52 statistically significant associations. Larger sample sizes am ong discordant responses might have allowed for stronger tests of statis tically significant associations. Concordance and Patients Confidence and Trust in their Physician Univariate statistics showed that physicia ns tended to overestim ate (mean= 3.90) how much confidence and trust their patients reported regarding medical deci sions. Physicians felt their patients had a great deal of confidence and tr ust in 70% of medical in teractions, but patients only reported a great deal of confidence and trus t 48% of the time. In fact, physicians seemed quite confident that patients would have either a great deal (70%) or fair amount (8%) of confidence and trust in their de cisions. The only other response from physicians was dont know (DK) which occurred in 11 inte ractions (22%). Patients were quite varied in their reports of confidence and trust: 38% felt they had a fair amount of c onfidence and trust, while 14% reported not too much or none at all. The answer choice, dont know only appeared on two physician survey items (patients self-reported health a nd confidence & trust) among the six pa rallel questions. Tables 4-9 and 410 examine physician DK responses by patient dem ographic characteristics and by physicians clinical experience. No signifi cant associations were found among patient variables. However, small sample sizes and cells would limit any statistical testing. Frequencies show small associations compared to the results regardi ng SRH. Male patients, non-minorities, higher incomes, and older patients were more likely to provoke DK responses from physicians. Fishers exact test did find a statically significant association between first visits (p < 0.01) and DK responses. Most (91%) of all DK responses included a first time patient. Clinical experience was strongly associated (p < 0.01) with DK responses ; all DK responses were reported by physicians with ten or more years of clinical experience.


53 Table 4-12 illustrates the cr osstabulation between physic ian and patient responses regarding confidence and trust. Bolded numbers along the diagona l represent exact concordance between physician and patient. Consistent with SRH results, DK responses were not figured into concordance/discordance values. Dyadic concordance (within +/-1) occurred in only 35 of the 50 interactions (see Table 4-5). However, af ter removing DK responses, dyadic concordance occurred in 35 out of 39 inter actions or 90%. This result is much higher compared to composition of exact item response between phy sicians and patients. Arguably, confidence and trust are critical components in patient satisfaction and adhe rence to potential treatment. Implications and strategies for gaining a patients trust have b een discussed in past studies (Klingle & Burgoon, 1995; du Pre, 2000; Eggly, 2002; Street, Gordon, & Haidet, 2007; Cegala, Mcgee, & McNeilis, 2008). Therefore, perhap s it is warranted to also examine exact concordance, and in this case, the results are quite different compared to the previous measure of dyadic concordance. Exact concordance results in only 46% of the physician-patient interactions, when compared to a more inclusive measure of dyadic concordance (90%). Further analyses illustrate an overestimation of confidence and trust by th e physician, which could cause the physician to switch from a patient-centered communication style to a more pa ternalistic style in order to speed up the medical consultation (Teutsch, 2003). Exact concordance further yields a disagreement in 21 of 50 interactions, not in cluding the 11 DK responses. DK responses could be grouped with discordant interactions since physicians reporting a DK response did not even try to wager a guess, and regardless of assumpti on, did not match their patients response. If DK responses were grouped with disc ordant interactions, then the di stribution would change to 32 of 50 discordant responses.


54 The patients gender, race, age, income or experience with their physician did not produce any patterns or significant resu lts. However, having a college de gree was significant (p < 0.10) in Fishers exact test. All discordant responses incl uded patients without a college degree, but due to the small sample size among discordant dyads (n = 4), it is unclear whether the association would still exist in a larger sample. When a cr osstabulation was run be tween exact concordance and college degrees, the association was no longe r significant and discordant responses were evenly distributed between patients with and without college degrees. Concordance and Patient Education Level Physicians in m ost of the dyads (40 out of 50) reported that their pa tient was either a high school graduate (22%) or college graduate (58%) when asked about their patients highest level of education. However, in actuality, 36% of th e patients reported high school was their highest level of education and 30% reported they were co llege graduates. Results show that physicians tended to overestimate their patients educati on level, which could inhibit communication if physicians use terminology that does not match a patients level of comprehension. Since patient education level utilizes clear and concise categori es on the survey questionnaires, and physicians should know patient education le vel from the medical history, exact concordance will be reported for this set of parallel questions. Only 29 interactions (58%) were concordant wi th regard to the patients education level (see Table 4-5 and Table 4-13) among all 50 dyads. Physicians overestim ated their patients education level 12 times (24%), and in seven di fferent medical consultations, physicians reported their patient obtained a college degree, when thei r patients highest level of education was only a high school diploma. Chi-square results revealed a significant association with a p-value less than 0.01. Patients without a college degree (p < 0.01) were more likely to be mislabeled by their


55 physician (see Table 4-7) in addition to minority patients (p < 0.10, see Table 4-6). No further significant associations or fr equencies were discovered. Concordance and Patient Pain Rating A majority of patient (n=41) reported so me degree of pain among the 50 consultations. Patients average pain rating was 5.50 (median 6.00) on a scale from 1 to 10. Physicians reported an average patient pain rating of 4.24 (median 3.00). Overall, these results suggest physicians are slightly underestimating the degree of pain their patients report. Two-thirds (33 out of 50 dyads) provided concordant responses (see Table 4-5), but only 16 out of 50 dyads (32%) matched answers exactly. Perception of pain is not an eas y observation to make, and therefore, responses that are within +/-1 from each other could be qu ite similar and therefore are classified in this study as concordant. Table 4-14 depicts the crosstabulation for physic ian and patient pain level responses, with exact concordance bolded along the diagonal. This table can be used as a tool to illustrate concordant or disc ordant responses and where they fall on the rating scale. The disparity between physician and patient responses is larger in rega rds to pain rating than any other item. Twenty percent of the dyads showed a discordance of -3 or greater from the patients response. This is a fairly large underestimation of pain by the physicia n, and in some cases, were 4 or even 5 levels away from their patients response. Physicians overestimated their patients pain only 4 times among the 17 discordant interactions. Physicians were able to correctly predict ol der patients pain rating within (+/-1) more often than younger patients (see Table 4-6). More than two-thirds (11 out of 17) of the discordant responses occurred with patients und er the age of 49 (p < 0.05). Patient age was the only variable that was statistically significant asso ciation in the chi-square tests.


56 Concordance and Physicians Explanation of Diagnosis A total of 37 physician-patient dyads reported discussing a diagnosis during the medical consultation. Physicians were asked how well they had explained their diagnosis to the patient and reported a mean score of 9.57 on a scale of 1 to 10, showing that physicians were quite confident in their explanations (a rating of 10 would indicate a pe rfect explanation). Only three interactions produced a physician ra ting of 7 or 8, whereas every ot her score was either 9 or 10. When patients were asked how well the physician had explained the diagnosis, the mean rating of 7.95 was lower than the physicians however only half responded that they understood everything their physician had explained. Fewer than one fourth (2 2%) of patients reported that they only understood some or only a little of what their physician had said. These results suggest that physicians were sometimes mi sperceiving confus ed patients. Table 4-5 indicates that 51% ( 19 out of 37) of responses we re concordant and table 4-15 shows the full crosstab of physician and patient responses regarding the physicians explanation of the diagnosis. On only one occasion, did a pati ent rate the physicians explanation higher than their doctor. Every discordant response resulted from the physic ian rating the explanation much higher than the patient, and some disparities we re quite high. 6 out of 18 of the discordant response differed by 4, 5 or 6 response levels suggesting a misunderstanding among patients that was not perceived by their physicians. Tables 4-6 and 4-7 report the frequencies a nd chi-square results of dyadic concordance regarding patient sociodemographi c characteristics. One variable stands out among the others: minority patient dyads are more likely to be discordant (71%, p < 0.05) than non minorities. Furthermore, only four minority patient dyads were concordant out of all of the interactions. Medical experience did not increase concorda nce among physicians when examining this parallel item.


57 Concordance and Physicians Explanation of Treatment A total of 42 physician-patient dyads reported discussing a treatm ent during the medical consultation. Physicians were asked how well they had explained their treatment to the patient and reported a mean score of 9.77 on a scale of 1 to 10. Again, physicians were quite confident in their explanations since a rating of 10 would i ndicate a perfect explanati on, therefore nearly all (98%) of the physicians rated their treatment expl anation as 9 or 10. Pati ents reported a higher mean (8.00) regarding explanation of treatment than for their assessment of the diagnosis. However, only 50% of patients rated their physic ians explanation as a 9 or a 10, and 33% of patient responses rated their physic ians explanation as a 7 or lowe r. This resulted in a dyadic discordance of 48% (see Table 4-5) Six patients provided a rating of 5 or less (see Table 4-16) and, perhaps more perplexing; th eir corresponding physician rated th e same explanation as a 9 or 10. This may be the largest disconnect among all c oncordant items. Further analysis showed that these same patients also left the medical vis it with unanswered questions and low levels of satisfaction. Male patients were more likely to disagree with their physicians rating than female patients concerning the physicians explanation of treatment (see Table 4-6), along with minority patients. Patients without a college degree were more likely to disagree with their physicians rating (see Table 4-7) and additiona l analyses showed that these pa tients were also less likely to understand everything their physician had said. Although frequencies of each independent variable showed small frequency associations, none of the variables were statistically significant in chi-square analyses. This is most likely due to small sample sizes and low cell counts. It is possible that gender, race, and education could be significant with a larger sample size.


58 Results Concerning Unanswered Questions Table 4-4 also depicts physician and patient responses with regard to unanswered questions. P hysicians were asked if they had asked the patient about any unanswered questions and 72% reported that they did. Patients were asked if they ha d questions they had wanted to discuss but did not, and only 36% re ported that they had left the medical visit with unanswered questions. This suggests that a majority of physic ians were trying to identify and address any misconceptions or confusion before th e end of the medical consultation. Dyadic Concordance within the Medical Specialties Family Medicine and Oncology As a final testing step for the paired survey m ethod, descriptive statistics were run to examine if relationships existed between th e specialties family medicine and oncology. Crosstabulations, frequencies, a nd t-tests were run separately with the 25 dyad pairs from each area of medicine. The first anal ysis compared the mean respon se difference between physician and patient parallel answers. Larger differences (whether negative or positive) indicated greater discordance. Additionally, a t-test was run to determine sta tistical significance among the means. Table 4-17 displays the results for both specia lties. Since both specialties only contained 25 dyads, and some concordant items further lowe red the sample size depending on DK responses, little association was found between concordance and independent variables. Therefore, these results are largely descriptive and primarily compare concordance va lues using univariate statistics. Results examining dyadic concordance of a pa tients SRH (see Table 4-18) indicate that family medicine interactions were overall more discordant than dyads from oncology. Among the 29 concordant dyads, nearly half (47%; p < 0.05) included oncology dyads, while 38% (p < 0.05) included family medicine. However, all five discordant interactio ns regarding SRH came from Family medicine interactions and each discordant response was due to an underestimation


59 by the physician. Oncologists were slightly more likely to report that they did not know their patients SRH (56%) than family medicine physi cians (44%). T-test results did not indicate a significant difference when comparing the means of both specialties. Frequencies examining the patients confid ence and trust in thei r physician were not statistically significant in either family medi cine or oncology. Family medicine dyads were slightly more concordant (54%) than oncology dyads (46%), which was due to oncology physicians responding with slightly more DK answers (see Table 419). Patients within the study averaged less confidence and tr ust (3.16) in their physicians co mpared to patients in oncology (3.44) and further testing showed this association to be strongly significant with a p-value of 0.001. As results indicated from th e overall sample, physicians in both family medicine and oncology assumed their patients were either high school or college graduates. It is unclear whether physicians guessed their patients educat ion level or mistakenly wrote an incorrect response. Family medicine intera ctions were more discordant (64%), oncology patients had less variation in their highest leve l of education, and oncologists only deviated from their two primary responses five times, and three of those occasions consisted of previous patients (whom the physician personally knew). Similarly, family medicine physicians devi ated only four times. Unfortunately, results do not e xplain why physicians were only choosing between two potential categories or why those categories did not matc h up more frequently with patient responses. Surprisingly, among the 9 discordant responses in family medicine, 6 out of 9 differed by a maximum of 3 categories. This means that fam ily medicine physicians were assuming their patient received a college edu cation when the patient reporte d only a high school degree. The largest mistake by oncologists was only a difference of two categories.


60 Concordance results for a patients reported pain level showed oncologists produced higher concordance (58%) compared to family medicine physicians (42%). Again, family medicine physicians were more discordant (65%) and tended to underestimate their patients pain by a higher margin (-1.28, p < 0.001) than oncologists (-0.12, p < 0.001). Additionally, among the discordant responses in Family medicine, physicia ns perceived their patien ts pain rating 3, 4 or 5 units away from their patients rating 9 out of 11 times. By comparison, the oncologists largest margin of error was a difference of 3 units and only occurred in 2 interactions. On average, patients rated their oncologists explanation of diagnosis higher (8.67) than patients from family medicine dyads (7.11). Earlier results repor ted that physicians in both specialties rated themselves quite high (mean= 9.57), which led to a rather large disparity in concordance data. Family medicine dyads accounte d for 64% of the discordance when compared to oncology dyads (36%). Further analysis (see Table 416) shows a significan t association (p < 0.001) between the average margin of discordance in family medicine dyads (2.28) and oncology dyads (1.21). Similar results were found when comparing physician and patient responses to the explanation of treatment. Again, family medicine dyads were more disc ordant (55%, p < 0.05) and among the discordant responses illustrated a larger margin of difference (2.17, p < 0.001) compared to oncology dyads (1.46, p < 0.001). Overall, these results show that measuri ng dyadic concordance provides new information and sheds light on how physician and patient perceptions may differ during the medical consultation. The instrument includes some flaw s, which are easily corrected, but excels in providing informative input that indirectly relates to health communication. Dyadic concordance may reveal something intriguing and different and furthermore, could be useful in examining how physicians communication style impacts patie nt perception and satisfaction. In summation,


61 the findings show male patients, minorities, and patients under the age of 50 were more likely to differ from their physicians perception during th e medical consultation. Discordance was also found among patients with higher salaries, no college degree and with patients visiting their physician for the first time. Family medicine cons ultations produced a higher rate of discordance when compared to oncology consultations. Am ong the six components, family medicine consultations produced higher discordance levels when measuring the pa tients self reported health, patient education level, patient pain rati ng, explanation of diagnosis, and explanation of treatment. Patients further reported that more family medicine physicians did not ask them if they had any questions compar ed to patient reports in onc ology. These results and their implications will be discussed in chapter five.


62 Table 4-1. Frequency of characteristics for physicians (N= 10) and patients (N= 50) Characteristic No. (%) Patient Physician Demographics Male 29 (58%)6 (60%) Female 21 (42%)4 (40%) Minority 19 (38%)3 (30%) Race White 31 (62%)7 (70%) Black 10 (20%)1 (10%) Hispanic 6 (12%) 1 (10%) Asian 3 (6%) 1 (10%) Age Median age, years (range) 54 (19-92)42 (31-59) Under 50 22 (44%)7 (70%) Over 50 28 (56%)3 (30%) Household income, yearly (US$) Under 60k 17 (34%)Over 60k 33 (66%)Education High school graduate or le ss 18 (36%)Some college/college graduate22 (44%)Post-college education 10 (20%)Notes: Percentages in table are ro unded to nearest whole percent.


63 Table 4-2. Frequency of physician characteristics concerning medica l training and experience with patients Characteristic No. (%) Physician Characteristics (N=10) Years in practice <5 2 (20%) 5-10 2 (20%) 10-20 4 (40%) 20+ 2 (20%) Formal communication Training Yes 2 (20%) No 8 (80%) How important are communication skills in medical practice? Not important 0 (0%) Moderately important 1 (10%) Extremely important 9 (90%) Physician-Patient Dyad (N=50) 1s t visit with patient? Yes 32 (64%) No 18 (36%) Notes: Percentages in table are ro unded to nearest whole percent.


64 Table 4-3. Patient health descript ors and reported health encounter characteristics from patient surveys (N= 50) Characteristic N/Mean (%/range) Pain rating Responding yes 38 (76%) Mean rating of pain (low=1 ; high=10) 5.5 Diagnosis Interactions in which diagnosis was discussed 39 (78%) Mean perception of physician explanation (low=1; high=10) (range) 7.95 (3-10) Treatment Interactions in which treatmen t was discussed 42 (84%) Mean perception of physician explanation (low=1; high=10) (range) 8.00 (2-10) Self rated health status Excellent or very good 8 (16%) Good 8 (16%) Only fair or poor 34 (68%) Physician listened to ev erything I had to say Strongly agree or agree 39 (78%) Strong disagree or disagree 11 (22%) Patients with unanswered questions after medical interaction Minority 9 (50%) Total 18 (36%) Patients who report physician did no t spend enough time with them Minority 10 (56%) Total 18 (36%) Confidence and trust in physician Great deal or fair amount 43 (86%) None or not too much 7 (14%) Physician understands your background and values Minority (disagree or strongly disagree) 14 (74%) Physician treat you with respect and dignity (yes) 48 (96%) Notes: Percentages in table are ro unded to nearest whole percent.


65 Table 4-4. Total concordance out of 300 physician-pa tient opportunities by pa tient characteristics Patient Characteristic N Total Concordance Concordant Discordant Gender Male 14168% 32% Female 11170% 30% Minority Yes 98 63% 37% No 17476% 24% Age 50 or older 14273% 27% Younger than 50 11065% 35% Household income 60k 16369% 31% 59k 89 70% 30% College degree (4-year or higher) Yes 12574% 26% No 12765% 35% First visit Yes 81 68% 32% No 17170% 30% Notes: Percentages in table are rounded to nearest whole percent. Concordance was achieved if physician response was +/-1 from patient response among parallel question. N=48 DK/NA responses were excluded from the frequency calculations.


66 Table 4-5. Frequency of concordant/discordan t responses between physician-patient dyads on parallel questions Item n (%) N DK/NA Patient self reported health 34 16/0 Concordant 29 (85%) Discordant 5 (15%) Confidence & trust in physician 39 11/0 Concordant 35 (90%) Discordant 4 (10%) Patient education level 50 Concordant 29 (58%) Discordant 21 (42%) Patient pain rating 50 Concordant 33 (66%) Discordant 17 (34%) Explanation of diagnosis 37 0/13 Concordant 19 (51%) Discordant 18 (49%) Explanation of treatment 42 0/8 Concordant 22 (52%) Discordant 20 (48%) Physician asked patient about unanswered questions 36 (72%) 50 Patients who report having unanswered questions 18 (36%) 50 Notes: Percentages in table are rounded to nearest whole percent. Concordance was achieved if physician response was +/-1 from patient response among parallel question.


67 Table 4-6. Frequency of concordant/discordant responses of parallel questions by patient sociodemographic variables Gender Minority Age M F No Yes 49 50 Patient self reported health (N=34) Concordant 16 (94%) 13 (76%) 16 (84%) 13 (87%) 12 (75%) 17 (94%) Discordant 1 (6%) 4 (24%) 3 (16%) 2 (13%) 4 (25%) 1 (6%) Confidence & trust (N=39) Concordant 18 (86%) 17 (94%) 20 (91%) 15 (88%) 17 (94%) 18 (86%) Discordant 3 (14%) 1 (6%) 2 (9%) 2 (12%) 1 (6%) 3 (14%) Patient education level (N=50) Concordant 20 (69%) 16 (76%) 25* (81%) 11* (58%) 15 (68%) 21 (75%) Discordant 9 (31%) 5 (24%) 6* (19%) 8* (42%) 7 (32%) 7 (25%) Patient pain rating (N=50) Concordant 20 (69%) 13 (62%) 20 (65%) 13 (68%) 11** (50%) 22** (79%) Discordant 9 (31%) 8 (38%) 11 (35%) 6 (32%) 11** (50%) 6** (21%) Explanation of diagnosis (N=37) Concordant 11 (52%) 8 (50%) 15** (65% ) 4** (29%) 8 (50%) 11 (52%) Discordant 10 (48%) 8 (50%) 8** (35%) 10** (71%) 8 (50%) 10 (48%) Explanation of treatment (N=42) Concordant 11 (46%) 11 (61%) 16 (57%) 6 (43%) 8 (50%) 14 (54%) Discordant 13 (54%) 7 (39%) 12 (43%) 8 (57%) 8 (50%) 12 (46%) Notes: *.05 < p .10, **.01 < p .05, ***p .01 Percentages in table are round ed to nearest whole percent. ChiSquare results are taken from Fishers exact test due to small sample size


68 Table 4-7. Frequency of concordant/discordan t responses on parallel questions by patient income, college degree and prior experience with physician Income College degree 1s t visit 59k 60k No Yes No Yes Patient self reported health (N=34) Concordant 11 (92%) 18 (82%) 16 (89%) 13 (81%) 23 (85%) 6 (86%) Discordant 1 (8%) 4 (18%) 2 (11%) 3 (19%) 4 (15%) 1 (14%) Confidence & trust (N=39) Concordant 11 (79%) 24 (96%) 16* (80%) 19* (100%) 29 (94%) 6 (75%) Discordant 3 (21%) 1 (4%) 4* (20%) 0* (0%) 2 (6%) 2 (25%) Patient education level (N=50) Concordant 11 (65%) 25 (76%) 14*** (56%) 22*** (88%) 23 (72%) 13 (72%) Discordant 6 (35%) 8 (24%) 11*** (44%) 3*** (12%) 9 (28%) 5 (28%) Patient pain rating (N=50) Concordant 12 (71%) 21 (64%) 17 (68%) 16 (64%) 21 (66%) 12 (67%) Discordant 5 (29%) 12 (36%) 8 (32%) 9 (36%) 11 (34%) 6 (33%) Explanation of diagnosis (N=37) Concordant 9 (64%) 10 (43%) 10 (56%) 9 (47%) 9 (41%) 10 (67%) Discordant 5 (36%) 13 (57%) 8 (44%) 10 (53%) 13 (59%) 5 (33%) Explanation of treatment (N=42) Concordant 8 (53%) 14 (52%) 9 (43%) 13 (62%) 14 (52%) 8 (53%) Discordant 7 (47%) 13 (48%) 12 (57%) 8 (38%) 13 (48%) 7 (47%) Notes: *.05 < p .10, **.01 < p .05, ***p .01 Percentages in table are rounded to nearest whole percent. ChiSquare results are taken from Fishers exact test due to small sample sizes


69 Table 4-8. Frequency of physicians dont know responses regarding th eir patients self reported health and characteristics N (%) Patient gender Female 4 (25%)* Male 12 (75%)* Patient race Minority 4 (25%) Non-minority 12 (75%) Patient income 59k 5 (31%) 60k 11 (69%) Patient age 49 6 (38%) 50 10 (62%) First visit with patient Yes 11 (69%)*** No 5 (31%)*** Notes: *.05 < p .10, **.01 < p .05, ***p .01 Percentages in table are ro unded to nearest whole percent. Percentages computed from overall sample of N= 50. Chi-Square results are taken from Fishers exact test due to small sample sizes


70 Table 4-9. Frequency of physicians dont know re sponses regarding thei r patients confidence and trust in their medical decisions by patient characteristics N (%) Patient gender Female 3 (27%) Male 8 (63%) Patient race Minority 2 (18%) Non-minority 9 (82%) Patient income 59k 3 (27%) 60k 8 (63%) Patient age 49 4 (36%) 50 7 (64%) First visit with patient Yes 10 (91%)*** No 1 (9%)*** Notes: *.05 < p .10, **.01 < p .05, ***p .01 Percentages in table are rounded to nearest whole percent. Percentages computed from overall sample of N= 50. Chi-Square results are taken from Fishers exact test due to small sample sizes Table 4-10. Frequency of physicians dont know responses by physicians medical experience (N=50) Clinical experience after residency < 5 years 5-10 years 10-20 years > 20 years Patients self reported health 0 (0%)* 2 (13%)**9 (56%)** 5 (31%)** Number of interactions (out of 50)10 10 20 10 Patients confidence and trust 0 (0%) ***0 (0%)***6 (55%)*** 5 (45%)*** Number of interactions (out of 50)10 10 20 10 Number of Physicians 2 2 4 2 Notes: *.05 < p .10, **.01 < p .05, ***p .01 Percentages in table are rounded to nearest whole percent. Significance is reported from Pearsons 2-sided chi-square


71 Table 4-11. Illustration of dyadi c concordance by comparing physician and patient response when asking about the patie nts self reported health Patient response 1 Poor 2 Only fair 3 Good 4 Very good 5 Excellent Total Physician response 1 Poor 11 4 1 0 0 16 2 Only fair 2 2 0 0 0 4 3 Good 2 0 1 1 0 4 4 Very good 0 1 2 4 0 7 5 Excellent 0 1 0 1 1 3 9 Dont know 5 6 4 1 0 16 Total 20 14 8 7 1 50 Notes: Bold numbers represent exact concordance where physician response = patient response. This study defines concordance as +/-1 from exact concordance. Physicians responding dont know is not factored into dyadic concordance. Table 4-12. Illustration of dyadi c concordance by comparing physician and patient response when asking about the patients c onfidence and trust in their physician Patient response 1 None at all 2 Not too much 3 A fair amount 4 Great deal Total Physician response 1 None at all 0 0 0 0 0 2 Not too much 0 0 0 0 0 3 A fair amount 0 1 1 2 4 4 Great deal 1 3 14 17 35 9 Dont know 1 1 4 5 11 Total 2 5 19 24 50 Notes: Bold numbers represent exact concordance where physician response = patient response. This study defines concordance as +/-1 from exact concordance. Physicians responding dont know is not factored into dyadic concordance.


72 Table 4-13. Illustration of dyadi c concordance by comparing physician and patient response when asking about the patients highest level of education Patient response 1 HS grad 2 Business, technical, vocational school 3 Some college, no four-year degree 4 College grad 5 Post grad training Total Physician response 1 HS grad 10 0 0 1 0 11 2 Business, technical or vocational school 0 0 0 2 0 2 3 Some college, no four-year degree 1 0 2 0 0 3 4 College grad 7 3 2 12 5 29 5 Post grad training 0 0 0 0 5 5 Total 18 3 4 15 10 50 Notes: Bold numbers represent exact concordance where physician response = patient response. This study defines concordance as +/-1 from exact concordance. Table 4-14. Illustration of dyadi c concordance by comparing physician and patient response when asking about the patients pain rating Patient response 0 No pain 1 Low 2 3 4 5 6 7 8 9 Total Physician response 0 No pain 8 0 0 1 0 0 0 0 0 0 9 1 Low 2 3 1 0 0 0 0 0 0 0 6 2 2 0 1 1 1 1 2 0 0 0 8 3 0 0 0 2 1 0 1 2 1 0 7 4 0 0 0 0 0 1 1 0 0 0 2 5 0 0 0 0 0 1 0 1 2 0 4 6 0 0 0 0 1 0 0 2 0 0 3 7 0 0 0 0 0 0 2 2 2 0 6 8 0 0 0 0 0 0 0 1 0 1 2 9 0 0 0 0 0 0 1 0 2 0 3 Total 12 3 2 4 3 3 7 8 7 1 50 Notes: Bold numbers represent exact concordance where physician response = patient response. This study defines concordance as +/-1 from exact concordance.


73 Table 4-15. Illustration of dyadi c concordance by comparing physician and patient response when asking about how well the phy sician explained the diagnosis Patient response 1 Not well 2 3 4 5 6 7 8 9 10 Very well Total Physician response 1 Not well 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0 1 0 0 0 0 0 0 1 8 0 0 0 0 0 0 1 0 0 1 2 9 0 0 1 1 2 3 1 1 0 0 9 10 Very well 0 0 0 2 0 0 0 6 7 10 25 Total 0 0 1 4 2 3 2 7 7 11 37 Notes: Bold numbers represent exact concordance where physician response = patient response. This study defines concordance as +/-1 from exact concordance. Table 4-16. Illustration of dyadi c concordance by comparing physician and patient response when asking about how well the phy sician explained the treatment Patient response 1 Not well 2 3 4 5 6 7 8 9 10 Very well Total Physician response 1 Not well 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0 0 1 0 0 0 0 0 1 8 0 0 0 0 0 0 0 0 0 0 0 9 0 0 1 1 0 1 2 1 1 0 7 10 Very well 0 1 0 1 1 2 3 6 5 15 34 Total 0 1 1 2 2 3 5 7 6 15 42 Notes: Bold numbers represent exact concordance where physician response = patient response. This study defines concordance as +/-1 from exact concordance.


74 Table 4-17. Mean difference betwee n physician and patient parallel responses by the specialties family medicine (n=25) and oncology (n=25) Family medicine Oncology Patient self reported health 0.44 b -0.06a Confidence & trust in physician 0.52*** d 0.61*** b Patient education level 0.40 f 0.36 f Patient pain rating -1.28*** f -0.12 f Explanation of diagnosis 2.28*** b 1.21***cExplanation of treatment 2.17*** b 1.46***e Notes: *.05 < p .10, **.01 < p .05, ***p .01 Significance was tested using a t-test of the mean differences a Mean computed from N= 16. b Mean computed from N= 18. c Mean computed from N= 19 d Mean computed from N=21. e Mean computed from N= 24. f Mean computed from N= 25. Table 4-18. Frequency of concordant /discordant responses of parall el questions by the specialties family medicine and oncology Family medicine Oncology N DK/ NA Patient self reported health 34 16/0 Concordant 13 (72%)** 16 (100%)** Discordant 5 (28%)** 0 (0%)** Confidence & trust in physician 39 11/0 Concordant 19 (91%) 16 (89%) Discordant 2 (9%) 2 (11%) Patient education level 50 Concordant 16 (64%) 20 (80%) Discordant 9 (36%) 5 (20%) Patient pain rating 50 Concordant 14 (56%) 19 (76%) Discordant 11 (44%) 6 (24%) Explanation of diagnosis 37 0/13 Concordant 7 (39%) 12 (63%) Discordant 11 (61%) 7 (37%) Explanation of treatment 42 0/8 Concordant 7 (39%) 15 (63%) Discordant 11 (61%) 9 (37%) Physician asked patient about unanswered questions 14 (39%)** 22 (61%)** 50 Patients who report having unanswered questions 13 (72%)** 5 (28%)** 50 Notes: *.05 < p .10, **.01 < p .05, ***p .01 Percentages in table are rounded to nearest whole percent. ChiSquare results are taken from Fishers exact test due to small sample sizes


75 Table 4-19. Frequency of physicians dont know responses by the specialties family medicine and oncology (N=50) Family medicine Oncology Patients self reported hea lth 7 (64%) 9 (56%) Patients confidence and trust 4 (36%) 7 (44%) Notes: Percentages in table are ro unded to nearest whole percent.


76 CHAPTER 5 DISCUSSION In the opening chapter, the significance for studying physician-pati ent interaction was outlin ed and there is a need for additional research if we wish to keep improving the physicianpatient relationship. Due to large amounts of va riability between human interactions, it is important to examine multiple dimensions of the physician-patient encounter and suggest potential solutions with reporte d findings. Arguably, research ha s created more questions than answers when trying to decide how to meet th e needs or goals of physicians and patients. Literature has documented that research has not provided a comprehensiv e solution to meet the expectations of health care professionals (Talen, Grampp, Tu cker, & Schultz, 2008), patients (Brenner & Bartholomew, 2005; Swenson et al. 2006), or researchers (Cardarelli & Chiapa, 2007). Additionally, the Institute of Medicine (IOM) issued two reports (2000, 2001) weighing in on Americas health care qua lity and documenting several ineffi cient dimensions of physicianpatient interaction. The IOM furt her argued that a new level of partnership needs to occur between physician-patient dyads a nd a potential solution may be fo r health care professionals to adopt some aspect of patient-centered care. Arguably, crucial elements of medical communi cation and dyadic inte raction (i.e., human perception, relational theory, cultural competency ) have not been exhaustively documented in past literature. This is par tly due to various methodological problems that have plagued the measurement of patient satisfaction (Aragon & Gesell, 2003), patient centeredness (Mead & Bower, 2000) and shared-decision making (Elwyn et al. 2001). Only a few st udies have explored how simultaneous physician and patient percep tions of each other influence the medical consultation (Krupt et al. 2000; Jahng et al. 2005; Ahlena et al. 2007; Zebiene et al 2008).


77 Measuring and describing agreement between physicians and patients gained momentum in the mid-nineties when Liaw, Young, and Farish (1996) developed a conceptual framework to study doctor-patient discordance. In their study, patients were assessed before and after the consultation in regards to their main health pr oblem, treatments, investigations ordered, their satisfaction with the consultation, and whether thei r expectations were met. Additionally, patient information was compared to physician notes and the corresponding medical record in order to assess concordance. This study provided useful in formation regarding the patient perspective and issues in medical charting, but did not truly assess the physicia n perception of the interaction. Since the mid-nineties, few studies revisited physi cian-patient concordance. Some studies have analyzed concordance in different aspects of care (Ahlena et al. 2007; Zebiene et al. 2008), but have missed a critical component of perception and did not descri be the level of agreement or how it fluctuates between both physic ians and patients in relation to sociodemographic variables. Past communication approaches have been st rictly one dimensional, and therefore the focus has usually strayed to either the physician or patient. Therefor e, there may be a need to reconceptualize our notion of physic ian-patient relationships to more accurately measure a coconstructed relationship. Theref ore, a measure is needed that can account for both the physician and patients perception at the same time and assess those perceptions for effective medical dialogue. Furthermore, additional research is needed to understand how medical perceptions change and why they may be different, de pendent on the composition of participant characteristics or medical specialties. In order to accomplish this task, the idea of concordance was re-c onceptualized into a useful interactional measure. C oncordance is simply defined as a state of agreement or harmony; however, this definition can also be applied to medical perception. Utilizing the basic definition


78 of the word concordance, the concept of dyadic concordance will describe the level of agreement between both physician and patient perceptions after their medica l encounter and furthermore, provide a numerical measure for comparison. This dimension of medical interaction may allow us to further identify communication barriers or disparities between physician-patient dyads. Therefore, the main objective of this study was to measure communication indirectly by focusing on the perception of each medical participant. Two research goals were examined: (1) how does dyadic concordance vary when compared to patients gender, age, race, income, education, and experience with physician and (2) what informa tion might be valuable when using a pairedsurvey method to measure dyadic concordance within family medicine and oncology. In addition to measuring concordance, the patient survey instrument included two additional consultation dimensions that explor ed a patients confidence and trust in their physicians medical decisions and their percep tion of time management during the medical consultation. Both additions could lead to re levant information since arguably, a patients perception of time management during the medical consultation may be quite different from physicians. Patients want to spend as much tim e as possible during the medical encounter in order to address all of their pertinent questions and compre hend medical dialogue. However, physicians may want to make the medical consulta tion as efficient as possi ble, while adhering to any time constraints in their schedule or dictated by medical policies. This could cause physicians to provide shorter clin ical explanations or not addre ss all of a patients concerns. Therefore, if one goal of the medical consultatio n is to provide a satisfactory medical encounter for both the patient and physician, then recogniz ing time constraints and finding an efficient solution for both participants is crucial.


79 Patient responses could help id entify their particular perceptio ns and needs in relation to the physicians ability to explain medical dia gnoses or treatments, and to follow up on any unanswered questions. Additionally, patients were asked to report their self-reported health, which could be compared to clin ical findings for the assessment of perceived health. Finally, one item of cultural competency was measured by asking the patients if they felt their physician understood their background and values. This meas ure is useful when combined with other survey items because it allows for the assessmen t of potential communicati on barriers resulting from a lack of cultural competency. By no means was this an exhaustive examination of cultural competency, and better models exist if tryi ng to accurately measure culturally competent interactions (Teal & Street, 2008). Adding extra assessments to coincide with dyadi c concordance benefited this pilot study in the following ways. First, since the medical cons ultation was not video ta ped or audio recorded, the additional questions aided the instrument and dyadic concordance measure by providing a more in-depth examination of the physicia n-patient encounter. Second, assessing dyadic concordance is not the only measure of importance and adding additional questions helps test the feasibility of a more complex instrument for future adaptations. This iteration of the pairedsurvey instrument was primarily a test of feasib ility, and not necessarily intended to complete a comprehensive investigation of dyadic interac tion. However, the instrument did excel at exploring several critic al perceptions from the overall patient sample in addition to dyadic concordance. A total of 50 patients were enrolled in the study, and although this sample was not representative, it did provide quality data concerning local area patients and how their perceptions of medical encounters. Findings revealed that a major ity of patients believed their


80 physician treated them with respec t and dignity. Furthermore, 78% of patients felt their physician listened to everything they had to say and 64% responded that not only did their physician spend enough time with them, but also an swered all their questions. Take n together, we can infer that physicians included in this st udy did not utilize a paternalis tic model of communication but employed a more relationship-cente red interaction. If the interactions were primarily doctorcentered or strictly paternalistic, then patient perc eptions of their clinicia n would be expected to be lower, especially when patients were asked if they were allotted time to voice concerns (Back & Hutchinson, 2006). Additionally, we would have expected lower scores when patients were asked if physicians listened to those concerns (T eutsch, 2003), or treated them with respect and dignity (Grol et al. 1990). Communication styles th at reflect cultural competency include understanding a patients background and values (Beach, Saha, & Cooper, 2006), medical beliefs (Flores, 2000), and how stereotypes impact medical cons ultations (Malat & Hamilton, 2006) To assess one element of cultural competency, patients we re asked to assess their physic ians understanding of their own background and values. This is an important per ception to assess because health disparities are quite prevalent today (Gorman & Read, 2006; Card arelli & Chiapa, 2007) and disparities related to patient characteristics may emerge during th e medical consultation (Van Ryn & Burke, 2000; Ngo-Metzger et al. 2006). This follows on recen t reports recommending the use of culturally competent communication because it may decrease racial and ethnic disparities (Teal & Street, 2008). Findings from this study suggested some al arming patterns. Among the minority patients included in the sample, 74% disagreed when as ked if they believed their physician understood their background and values. They were five times more likely than white patients to disagree,


81 and chi-square results confirme d a highly significant association (p < 0.001). Furthermore, two additional significant associations (p < 0.06) exis ted among minority patients: over half felt their physician spent less time than they wanted, and 53% reported they had unanswered questions after the medical visit. Only one quarter of white patients felt thei r physician did not spend enough time with them and consequently onl y 25% reported any unanswered questions. It is unclear what led to these outcome s and perhaps physicians were utilizing a paternalistic communication approa ch (i.e., physicians make the decisions) during these medical encounters, which arguably may have led to nega tive responses (Charles et al. 1997). Patients with past negative medical expe riences or with extensive medi cal experience may have been primed or sensitive to perceived problems and therefore, more likely to report those issues on questionnaires. Furthermore, it is difficult to determine whether these discrepancies may be the result of the physician misinterpreting similar behaviors from different patients or if any stereotyping or prejudice is o ccurred. Unfortunately, the measur ing instrument does not provide enough depth with regard to cultural competency or information on physicians perception of racial attitudes to explain this finding. Howeve r, these results do suggest that some disparity exists between minority patients perceptions and their physicians. This data may reflect similar patterns discusse d in past literature. For example, patients who visited physicians of a different race were gene rally less satisfied and ra ted their interactions with those physicians as less pa rticipatory than racially concor dant dyads (Cooper-Patrick et al. 1999; Saha et al. 1999). Many ra cial and ethnic minorities also reported poor communication with their physician, which could lead to more problems during the medical interaction (Flores et al. 2002; Cardarelli & Chiapa, 2007). Finally, physicians may fail to process individual patient


82 behavior, therefore be ing swayed by a set of beliefs or attitu des that lead to the misidentification of minority concerns (McKinlay, 1997). Three-quarters of patients re ported some measure of pain and responded negatively when asked about self-reported health, which compares to samples obtained from more representative oncology populations (Malin et al ., 2006). If the measuring instrume nt evaluated a larger sample from a variety of medical specia lties, then patient health de scriptors would most likely be different and more evenly distribu ted across answer choices. Dyadic Concordance, Perception, and Implications for Practice Physician-patent communication has been scruti nized for decades am id reports of patient dissatisfaction, medical errors, heal th disparities, increased malp ractice litigation, and lack of interpersonal care. Numerous models for impr ovement have been suggested; however, many of these negative outcomes have not dramatically d ecreased and still occur t oday. One aspect of the physician-patient relationship, dya dic concordance, has rarely been discussed, and perhaps could provide further insight into h ealth communication and possible avenues for decr easing negative medical outcomes. Examining Total Concordance Since 50 physician-patient interactions provided six opportunities per dyad to m easure concordant perceptions, a tota l of 300 concordant opportunities were assessed. Among those opportunities, 252 included paralle l responses which could be examined for concordance. Twothirds of measured opportunities were classifi ed as concordant, which depicts room for improvement since one-third of th e opportunities missed their mark. In particular, male patients, minorities, and patients under the age of 50 were more likely to differ from their physicians perception during the me dical consultation.


83 Since gender has been identified as one s ource of systematic variation in medical encounters (Street, 2002), it was not surprisi ng to find variability among male and female patients. Literature has documented distinct di fferences in clinicians perception of male and female patients. Elderkin-Thomps on and Waitzkin (1999) believe a gender bias may exist in the medical encounter, where physicians misperce ive the seriousness and nature of womens symptoms. Additionally, physicians may misattri bute womens health problems to emotional issues or stress (Hall & Roter, 1995) and potentia lly misperceive their ed ucation level along with their need to communicate (Todd, 1989; West, 1993; Beck, 2001). However, female patients in this study were more concordant with their physicians than male patient s, therefore providing a counter argument to the findings of these earlier studies. It is possible that physicians were able to find out more about female patients during the medical encount er, making them more likely to match their patients response and perception. Past literature has confirmed that female patients are more likely to openly express feelings of concern and ask ques tions (Roter et al. 1991; Hall et al. 1994; Bernzweig, Takayama, Phibbs, Le wis, & Pantell, 1997; Kiss, 2004). Discordance in this study betw een minority patients and thei r physicians may be tied to a multitude of medical disparities involving race and ethnicity. Stepanikova et al. (2006) have reported that minorities may provide their physicians with lower scor es on indirect measures of trust. Minority patients may also be subject to the social distance hypothesis and provide lower interactional scores for their physician (Malat, 2 001). Finally, it is possible that the race of the physician may impact minority patient ratings. Dyad ic racial discordance could result in lower satisfaction scores, trust, and decreased participation in medical decision-making (Saha et al. 1999; Cooper et al. 2003; Stepanikova et al. 2006). Taken together past associations regarding


84 minority patient perceptions seem to support the findings from this instrument and lend credibility to the methological design. Discordance was also found am ong patients with higher salari es, no college degree and with patients visiting their physician for the fi rst time. Patient education level may be an important dimension to examine for the medical consultation. If physicians know or ask about patients terminal education, then they can ad just their communication style to appropriately reflect potential comprehension levels. It wa s surprising that this instrument found more discordance in patients with high er salaries, because literature suggests lower salaries may create more discordance due to the physician underestim ating patient comprehension. This association is likely comparative to findings concerning patient educati on and socioeconomic status (SES) since there is evidence that an individuals SES may have an impact on physician behavior, diagnosis, treatment explanations, and patient perception during the medical encounter (van Ryn & Burke, 2000; Epstein, 2005). Research ha s identified that i ndividuals from lower socioeconomic backgrounds tend to receive less health-re lated information from their physicians compared with individuals from higher soci oeconomic backgrounds (N go-Metzger et al. 2006). Utilizing this instrument to measure the total dyadic concordance out of all concordant opportunities is a useful appro ach to evaluating the physicianpatient encounter. It has the potential to catch a snapshot of communicati on barriers and sociodemographic disparities for individual physicians, or a team of physicians. Perhaps the strength of the instrument lies in its adaptability to measure interaction in a multit ude of ways. The following section illustrates a more specific utilization for th e concordance instrument, where six components of the medical encounter were examined for each physician-patient dyad.


85 Interpreting Discordance The first component of the instrum ent expl ored dyadic concordan ce in patients self reported health (SRH). Every patient reported thei r perception of SRH, however, physicians only responded in 34 of the 50 interactions. Physicians selected dont know (DK) as their answer choice for 16 interactions. At first glance, it c ould be assumed one or two individual physicians may have marked DK in all of their interactions therefore showing a meas ure of consistency in respect to their own perception. However, this was not the case; the DK responses were spread out among 6 out of 10 physicians. Furthermore, th e six physicians were quite varied in their responses, selecting dont know for a few patients and selecting a definitive answer for the rest of their patients. Simila r patterns were also found with regard s to patients confidence and trust in their physician. Among the 50 dyads, physicians selected DK 11 times. Results mirrored those from the SRH component, where male patients elic ited a majority of the DK responses from their physicians. Why do some medical interactions elicit DK responses from physicians, while other interactions involving the same physician elicit a direct re sponse? An analysis of sociodemographic variability shows male patie nts accounted for 75% (p < 0.10) of physicians DK responses when asked about their patients SR H and perception of confidence and trust. No other patient sociodemographic vari able was statistically significan t. Physicians may have had a harder time reading male patients and their a ssociated verbal and non verbal cues. Female patients tend to be more talkative (Gorman & Read, 2006) and empathic (Kiss, 2004), therefore, allowing the physician more opportunities to eval uate and read female patients perceptions during the medical encounter. When comparing the two components assessing the patients SRH and confidence and trust, physicians also had difficulty reading the perceptions of first time patients. This may be


86 expected since the physician had not developed an in-depth relationship yet with the new patient. However, this could present an opportunity to en gage the new patient in a more patient-centered communication style, which would allow the physic ian to get a better read on their new patient. Additionally, incorporating a mo re patient-centered approach to communication could improve the way physicians relate to and communicate with patients leading to more positive patient perceptions for matters such as trust, respec t and general satisfacti on within the medical encounter (Cegala, Mcgee, & McNeilis, 2008). It is unclear what communication style each physician utilized or whether they changed communication styles during the medical consultation because the instrument lacked components to distinguish between communication styles. The third component measured th e physicians perception of th eir patients education level and the highest level of educati on achieved by the patient. Almost all physician responses varied between two answer choices: high school graduate or college graduate. Rarely did physicians select any other answer choice. It is unclear w hy physicians only picked 2 out of 6 choices 80% of the time and unfortunately, this created 42% discordance among all the dyads. Further analyses identified higher discordance among ma le patients, minority patients, and patients without a college degree. This may imply that physicians were not reviewing medical and personal history prior to every interaction. One last possibility could result from physicians stereotyping some patients, however, more information would need to be collected with an instrument designed to measure stereotypes and prejudice before this possibility could be validated. The last three components measuring dyadic concordance produced the largest and potentially most problematic discordance levels in the st udy. These items measured the


87 concordance between clinical eval uations and explanations. Physicia ns were asked to assess their patients pain level and assess how well they expl ained a diagnosis or treatment. Patients were also asked parallel questions. Overall, physicians had a lot of confiden ce in their ability to explain diagnoses and treatment options, howev er this was not equally perceived by their patients. In fact, almost one-half of all patien ts responded differently than their physician and their ratings differed from physicians quite dram atically. Although this pa ired survey technique succeeded at identifying this potential disconnect between physicians and patients, it does not provide enough detail to determine why only half of the patients reported that they did not comprehensively understand clinical explanations. Sociodemographic differences may help provid e more information, but also add another layer of complexity to the results. Minority pati ents were more discorda nt when assessing their physicians explanation of the diag nosis, but not treatment. There was more discordance between younger patients and physicians when assessing the patients degree of pa in. Additionally, older patients produced higher discordance levels when asked about physicians explanation of treatment and diagnosis. Furthermore, patients wi th lower education produced more discordant responses regarding treatment e xplanation, but not the diagnosis. Theses disparities may stem from physician difficulty in reading patients verbal and non verbal behavior. Without video taping or tape recording the interaction, however, it is difficult to assess if the patient was confused at any time during the interaction or if the physician detected patient confusion. Since many patients reported leve ls of lower self reported hea lth and indicated a specific level of pain, it is possible that the physical or emotional demands of their illness could interfere with decision making (Harvey, Kazis & Lee, 1999) and affect concordance compared to physicians perceptions. This could also lead phy sicians to misperceive many elements of the


88 medical encounter. The best way to correct these discrepancies is not always clear, but it is important to take steps to maximize effectiv e communication in orde r to ensure positive consultation outcomes. How we interpret concordance and discordanc e should be carefully evaluated in future studies. Potential conceptual or evaluative issu es may reside when distinguishing physicianpatient agreement versus concordance level. In other words, are all manifestations of concordance good? If a scenario exists where a patient negatively rates a physicians explanation of diagnosis and the corresponding physician also rates their explanati on negatively, then the potential for concordance is high. However, that type of concordance represents the acknowledgement of inefficiency. Therefore, it is important to explain what that concordance means and how it would be measured in a studys conceptual design. Discordance should also be carefully examined. If a physicia n under/over estimates patient pa in, resulting in discordance, what does that mean? Perhaps, some physicians are good diagnosticians and recognize more non verbal behavior in reference to pain than the patient. On the ot her hand, patients could inflate or exaggerate their level of pain in many situ ations. This measure could be subjected to misinterpretation and researcher s should discuss the level of dyadic discordance/concordance and to what extent does that level matter. Comparing Medical Specialties Exam ining dyadic concordance between several medical specialties ma y provide a useful way to indirectly assess interaction or comm unication and compare t hose differences across various medical fields. The feasibi lity test related to this phase of the study explored the same six components in order to maintain consistency. However, it is possible to utilize different interactional components relevant for individual medical specialties, in order to provide a more efficient and relevant concordance instrument.


89 After measuring dyadic concordance between th e six components of interaction, analyses reported a higher rate of discorda nce within family medicine c onsultations compared to oncology consultations. Among the six components, family medicine consultations produced higher discordance levels when measuring the patients self reported health, patient education level, patient pain rating, explanation of diagnosis, and explanation of treatment. Patients further reported that more family medicine physicians did not ask them if they had any questions compared to patient reports in oncology. Both specialties followed similar response patterns, where they were equally prone to label their patients educational degree as a high school or college diplom a, believe their patients had a high degree of confidence and trust in their me dical decisions, and rate their diagnosis and treatment explanations at a very high level. A strength of the concordance instrument was the ability to distinguish the degrees of discordance between specia lties regarding each of the six components. For example, when physicians in Family medicine were asked to assess their patients highest level of education, their discor dant response differed by three categories above what the patient reported 67% of the time. This means that the Family medicine physicians enrolled in this study were assuming their patients had completed a college education when the patient only completed high school. This may have an effect on the type of communication style chosen and how a physician chose to explain medi cal information. Furthermore, if this led to incorrectly assuming the comprehension level of th e patient, it may have resulted in inefficient decision making or misconceptions on behalf of the patient. Physicians perception of pain may be a more important skill set in some specialties compared to others. Arguably, physicians needing this skill in their prac tice could utilize dyadic concordance in order to evaluate their performance or the perfor mance of a resident or student.


90 Even if not used as an evaluative tool, it could still prove useful if trying to maximize the effectiveness of medical care. The instrument wa s able to extract two significant pieces of information: the level of agreement between physician and patient and whether the physician underestimated or overestimated the patients perception of pain. Combining this information with diagnostic criteria coul d aid the physician in making informed choices that satisfy both parties. Family medicine physicians and Oncologists di ffered in their percepti ons of patient pain, diagnosis explanation, and treatmen t explanation. Family physicians were more discordant than Oncologists on all three measures and among the disc ordant interactions, were also more likely to have larger errors on all three measures Some interactions did produce fairly large misconceptions, where the physician response differe d from the patient response by five or more units, but this primarily resulted in Family medicine consultations only. Due to sampling restrictions and conceptual design, findings can not conclude that Oncologists had better communication practices or outcomes than Fam ily physicians, but this pattern should be explored further in future studies. Oncology patients often look towards their health care providers to meet several key information and support needs (Rose, 1990). Therefor e, efficient and empathic communication is emphasized between patients and Oncologists due to the fear of a cancer diagnosis, complexity of medical information, and uncertainty regardin g the course of their disease or treatment benefits (Siminoff, Ravdin, Colabianchi, & Saud ers-Sturm, 2000). It is quite likely that two different communication styles we re used, and that oncology trai ning is more likely to include relationship-centered communi cation training (Arora, 2003; Back & Hutchinson, 2006).


91 Implications for Practice The concordance instrum ent is a successful approach to measure dyadic concordance and could benefit medical interaction in a variety of ways. First, the in strument is highly flexible and additional components could be a dded in order to assess different interactions than the ones tested in this study. With further testing and more comprehensive items, the instrument could be used to evaluate medical students or experien ced physicians after a consultation. Physicians could utilize the concep t of dyadic concordance in order to create a more satisfactory medical encounter for their patients, especially if they run a private clinic or practice. Additionally, the parallel items that conceptually measure dyadic concordance could be extracted and incorporated into existing or validated communication inst ruments. This may provide an exhaustive examination of physician-patien t communication and allow the re searcher to explain more variability than ever before. Suggestions for Inst rument Improvement Overall, the instrum ent was able to succe ssfully measure multiple aspects of physicianpatient interaction and produced quality data w ithin a feasibility study. However, it was not designed to provide an exhaustiv e analysis of physician-patient communication or distinguish between physician communication st yles. It may be difficult to craft an instrument which can provide that level of detail usi ng a survey technique. To accomplish this feat, a mixed methods approach combining paired surveys and video taping/audio recordi ng could provide the necessary design. However, adding questions or scales that can distingu ish between physician communication styles could be added to the current instrument. This would provide more data to accompany dyadic concordance, and allow a re searcher to report changes in physician communication style during the medical consultation.


92 When asking the patient to assess physician understanding of their own background and values, questionnaires should be sp ecific in order to decrease the chance of interpretation error. The corresponding question on the c oncordance instrument for this study was a bit vague and should have included more detail. When patien ts are asked about their background and values they could interpret the statement in several ways They may wonder if the survey is referring to their heritage, ethnic values, family background, or perhaps background and values related to medical decisions. Breaking this question into several items would help describe cultural perceptions more distinctively and allow for a more accurate analysis along side concordance data. The sampling needs to be more representative to allow for generalizations of the findings. Increasing the sample size from this study would al so add greater reliability to the results, and allow for more powerful statistical tests and eval uations. The instrument was designed to analyze dyadic interaction and therefore not originally meant for population inferences. In my opinion, the instrument should be used as an evaluative tool in medical practice, with the end goal of increasing positive medical outcomes for both physicians and patients. Improving Physician-Patient Interactio n: Sug gestions for Future Research Literature evaluating physician-patient enc ounters has slowed due to a decrease in new ideas or reliable measures. Since health care di sparities have arguably in creased, more research within the field is warranted. Within the last te n years, scholars have generated more questions than answers. For example, what is the most ef fective communication style? Is there a universal definition of patient-centeredness? How do you teach patient-centered communication to physicians? Do patients need medical communication training too? How about cultural competency training? Each question has a multit ude of answers, but few solutions. Perhaps a critical component is missing from the current body of research and th is study argues that the


93 perception of both physicians and patients n eeds to be analyzed, alongside communication measures, in order to provide a more compre hensive assessment of medical interaction. If the ultimate goal for physician-patient enc ounters is to create sa tisfactory relationships, increase patient compliance, improve illness ou tcomes, and reduce health disparities, then a communication approach is needed that can a ccount for each goal. Paternalistic communication styles create the worst disp arities (Roter, 2000; Back & Hu tchinson, 2006) and the suggested solution was adopting a patient-centered approach (Balint, 1 969; Byrne & Long, 1976; Cockerham, 2007). However, patient-centered comm unication has struggled to establish a viable conceptual model and evaluative approach in order to achieve desired goals. Currently, culturally-competent communicatio n is being emphasized as a solution to accompany patientcentered communication, w ith the hope of eliminating consistent problems in the physicianpatient relationship. It remains uncertain if this will solve a majority of communication barriers or disparities, but it should he lp. Adding dyadic concordance into the mix of research ideas will account for another missing aspect of interaction by providing va luable data on both patients and physicians perceptions of the medical encounter. It is hard to predict how phys ician-patient interaction will ch ange over the next decade, but this study adds a new piece of information that should prove to be a valuable asset when measuring the perceptions of physicians and patie nts. In conclusion, eval uating those perceptions as a measure of dyadic concordance, can add an important dimension to studying healthcare interaction.


94 APPENDIX SURVEY INSTRUMENTS Physician P rofile Please check or write in your res ponse to the questions below. 1. Gender [ ] Male [ ] Female 2. What is your racial or ethnic background? [ ] White/European [ ] Black/African-American [ ] Hispanic/Latino [ ] Asian/Pacific Islander [ ] American Indian or Alaskan Native [ ] Other: ______________ 3. How old are you? _________________ 4. Where did you receive your medical education? _________________________________________________ 5. Number of years in practice: [ ] less than 5 [ ] 5-10 [ ] 10-20 [ ] 20 or more 6. Have you had any formal training in communi cation skills with your patients outside of medical school (i.e. seminars, lect ures, conferences, group training)? [ ] yes [ ] no If yes, please specify ________________________________________________ 7. How important are physicians co mmunication skills with patients? [ ] Not important [ ] Moderately important [ ] Extremely important


95 Patient Profile Please check or write in your res ponse to the questions below. 1. Gender [ ] Male [ ] Female 2. What is your racial or ethnic background? [ ] White/European [ ] Black/African-American [ ] Hispanic/Latino [ ] Asian/Pacific Islander [ ] American Indian or Alaskan Native [ ] Other: ______________ 3. How old are you? _________________ 4. How much is your YEARLY total household income before taxes (including income from all sources and from all members who live in your household)? [ ] under $9,999 [ ] $10,000 $19,999 [ ] $20,000 $29,999 [ ] $30,000 $39,999 [ ] $40,000 $49,999 [ ] $50,000 $59,999 [ ] $60,000 $69,999 [ ] $70,000 or more 5. What is your occupation? __________________________________________ 6. How long have you seen this physician? __________________________________________


96 Physicians Interaction Survey Answer the following questions by checking or circling the answer that relates to you. 1) In general, how would your patient describe their self-reported health? [ ] Excellent [ ] Very good [ ] Good [ ] Only fair [ ] Poor [ ] Dont know 2) How much confidence and trust do you feel your patient has in your medical decisions regarding diagnosis or treatment? [ ] Great deal [ ] A fair amount [ ] Not too much [ ] None at all [ ] Dont know 3) What is your patien ts educational level? [ ] High school incomplete (grades 9-11) [ ] High school graduate (gra de 12 or GED certificate) [ ] Business, technical or vo cational school after high school [ ] Some college, no four-year degree [ ] College graduate (BS, BA or other four-year degree) [ ] Post-graduate training or professional schooling after co llege (e.g., toward a masters degree or PhD, law or medical school) [ ] Dont know 4) If your patient is in pain, rate the amount of pain your patient is feeling? 1 2 3 4 5 6 7 8 9 10 (Low) (High) 5) If a diagnosis was made, how well did you explain the diagnosis to your patient? 1 2 3 4 5 6 7 8 9 10 (Not well) (Very well) 6) If treatment options were discussed, how well did you explain the course of treatment for your patient? 1 2 3 4 5 6 7 8 9 10 (Not well) (Very well) 7) Did you ask your patient if they had any unanswered questions regarding their care or treatment that they wanted to discuss before the end of the medical consultation? [ ] Yes [ ] No


97 Patient Interaction Survey Answer the following questions by checking or writing in the best answer that relates to your situation. 1) What is your reason for todays doct or visit (Please be specific)? ______________________________________________________ ______________________________________________________ 2) What is your educational level? [ ] High school incomplete (grades 9-11) [ ] High school graduate (gra de 12 or GED certificate) [ ] Business, technical or vo cational school after high school [ ] Some college, no four-year degree [ ] College graduate (BS, BA or other four-year degree) [ ] Post-graduate training or professional schooling after co llege (e.g., toward a masters degree or PhD, law or medical school) 3) Are you in any pain today? [ ] Yes [ ] No (if no, skip to question 5) 4) On a scale of 1-10, how mu ch pain are you currently in? 1 2 3 4 5 6 7 8 9 10 (Low) (High) 5) Did you receive a diagnosis today of any ailment, injury or illness? [ ] Yes [ ] No (if no, skip to question 7) 6) How well did your doctor explain the diagnosis of your ailment, injury or illness? 1 2 3 4 5 6 7 8 9 10 (Not well) (Very well) 7) Did you receive a course of treatment to day for any ailment, injury or illness? [ ] Yes [ ] No (if no, skip to question 9) 8) How well did your doctor explain the course of treatment for your ailment, injury or illness? 1 2 3 4 5 6 7 8 9 10 (Not well) (Very well)


98 9) In general, how would you describe your own health? Would you say it is excellent, very good, good, only fair or poor? [ ] Excellent [ ] Very good [ ] Good [ ] Only fair [ ] Poor 10) Thinking about todays doctor visit, how would you answer the following question: The doctor listened to everything I had to say. [ ] strongly agree [ ] agree [ ] disagree [ ] strongly disagree 11) Did you have questions about your care or treatment that you wanted to discuss, but did not? [ ] Yes [ ] No 12) Did the doctor spend as much time with you as you wanted, almost as much as you wanted, less than you wanted, or a lot less than you wanted? [ ] As much as wanted [ ] Almost as much [ ] Less than wanted [ ] A lot less than wanted 13) Did the doctor involve you in decisions ab out your care as much as you wanted, almost as much as you wanted, less than you wanted, or a lot less than you wanted? [ ] As much as wanted [ ] Almost as much [ ] Less than wanted [ ] A lot less than wanted 14) Regarding todays doctor vi sit, did you understand everything the doctor said, most of what the doctor said, some, or only a little of what the doctor said? [ ] Everything [ ] Most [ ] Some [ ] Only a little 15) How much confidence and trust did you have in the doctor treating youa great deal, a fair amount, not too much, or none at all? [ ] Great deal [ ] A fair amount [ ] Not too much [ ] None at all


99 16) Did the doctor treat you with a great deal of respect and dignity, a fair amount, not too much, or none at all? [ ] Great deal [ ] A fair amount [ ] Not too much [ ] None at all 17) I feel that my doctor understands my background and values. [ ] Strongly agree [ ] Somewhat agree [ ] Somewhat disagree [ ] Strongly disagree 18) If you could choose, would you prefer to be treated by a doctor who is male or female or do you have no preference? [ ] Prefer male doctor [ ] Prefer female doctor [ ] No preference [ ] Depends on type of doctor 19) Have you had any of the following medi cal conditions? (check all that apply) [ ] Heart Disease [ ] Diabetes [ ] Stroke [ ] Cancer specify ___________________________________ [ ] Major Surgeriesspecify ____________________________ [ ] Otherspecify _____________________________________ [ ] None of the above

PAGE 100

100 LIST OF REFERENCES Ada ms, A., Buckingham, C.D., Lindenmeyer, A. McKinlay, J.B., Link, C., Marceau, L. et al. (2007). The influence of patient and doctor ge nder on diagnosing coronary heart disease. Sociology of Health and Illness 30(1), 1-18. Ahlena, G.C., Mattsson, B., & Gunnarsson, R. K. (2007). Physicianpatient questionnaire to assess physicianpatient agreement at the consultation. Family Practice 24, 498. Armstrong, D. (1979). The emancipation of biographical medicine. Social Science and Medicine 13A, 1-8. Arnold, C.L. & Coran, J.J. (2008). The develo pment of effective physician-patient communication: The dyad-centered approach. Paper presented at the National Communication Associations Confer ence, San Diego, CA November, 2008. Arora, N.K. (2003). Interacting with cancer patients: the signifi cance of physicians communication behavior. Social Science & Medicine 57, 791-806. Back, A., & Hutchinson, F. (2006). Patient-phys ician communication in oncology: What does the evidence show? Oncology 20 (1), 67-74. Bakalar, N. Doctors gender may be factor in heart diagnoses. New York Times October 2, 2007. Bales, R. (1950). Interaction Process Analysis: A Method for the Study of Small Groups Reading, MA: Addison-Wesley. Balint, E. (1969). The possibilitie s of patient-centered medicine. Journal of the Royal College of General Practioners 17, 269-276. Balint, E., Courtenay, M., Elder, A., Hull, S., & Julian, P. (1993). The Doctor, the Patient and the Group: Balint Revisited. London: Routledge. Beach, M.C., Saha, S., & Cooper, L.A. (2006). The role and relationship of cultural competency and patient-centeredness in health care quality. [Rep. No. 960]. The Common Wealth Fund. Beck, C.S. (2001). Communicating for Better Health: A Guide through Medical Mazes Boston: Allyn and Bacon. Bernzweig, J., Takayama, J.I., Phibbs, C., Lewis, C., & Pantell, R.H. (1997). Gender differences in physician-patient communication: evidence from pediatric visits. Achieves of Pediatric Adolescent Medicine 151, 586-591. Betancourt, J.R., Green, A.R., Carillo, J.E., & Pa rk, E.R. (2005). Cultural competence and health care disparities: key perspectives and trends. Health Affairs, 24, 499-505.

PAGE 101

101 Brenner, R. J., & Bartholomew, L. (2005). Communi cation errors in radiol ogy: A liability cost analysis. Journal of American College of Radiology, 2(5), 428-431. Byrne, P., & Long, B. (1976) Doctors talking to patients: a study of the verbal behaviors of doctors in the consultation London: HMSO. Cardarelli, R., & Chiapa, A.L. (2007). Educating pr imary care clinicians about health disparities. Osteopathic Medicine and Primary Care 1(5). Cassell, E.J. (1985). Talking with Patients vol 2. Cambridge, Mass: MIT Press. Cassileth, B.R., Zupkis, R. V., Sutton-Smith, K., & March, V. (1980). Information and participation preferences among cancer patients. Annuals of Internal Medicine 92(6), 832-836. Cegala, D.J., McGee, D.S., & McNeilis, K.S. (2 008). Components of pa tients and doctors perceptions of communication competence duri ng a primary care medical interview. In L.C. Lederman (Ed.), Beyond these walls: Readings in health communication (p. 74-91) New York: Oxford University Press. Charles, C., Gafni, A., & Whelan, T. (1997). Shared decision-making in the medical encounter: what does it mean? (or it ta kes at least two to tango). Social Science & Medicine 44, 681. Cegala, D.J., McGee, D.S., & McNeilis, K.S. (2 008). Components of pa tients and doctors perceptions of communication competence duri ng a primary care medical interview. In L.C. Lederman (Ed.), Beyond these Walls: Readings in Health Communication New York, NY: Oxford University Press. Chavez, L., Hubbell, F., McMullin, J., Martin ez, R., & Mishra, S. (1995). Understanding knowledge and attitudes about brea st cancer: a cultural analysis. Archives of Family Medicine, 4, 145-152. Cockerham, W.C. (2007). Medical Sociology 10th Edition Upper Saddle River, NJ: Prentice Hall. Commonwealth Foundation. (2002). Diverse comm unities, common concerns: Assessing health care quality for minority Americans. 2001 Health Care Quality Survey [computer file]. United States: Princeton Survey Research Associates. Cooper-Patrick, L., Gallo, J.J., Gonzales, J.J ., Vu, H.T., Powe, N.R., Nelson, C. et al. (1999). Race, gender, and partnership in the patient-phys ician relationship. Journal of American Medical Association 282, 583-589.

PAGE 102

102 Cooper, L.A. Roter, D.L., Johnson, R.L. (2003) Patient-centered communication, ratings, of care, and concordance of patient and physician race. Annals of Inte rnal Medicine 139, 907-915. Coran, J. J., Arnold., C. L, & Arnold., J. C. ( 2007) Physician-patient communication: This time from the physicians perspective. Paper presented at the Florida Communication Associations Conference, Daytona Beach, Florida, October, 2007. Desharnais, S., Carter, R.E., Hennessy, Kurent, J. E., & Carter, C. (2007). Lack of concordance between physician and patient: Repo rts on end-of-life care discussions. Journal of Palliative Medicine 10(3), 728-740. Dowsett, S.M., Saul, J.L., Butow, P.N., D unn, S.M., Boyer, M.J., Findlow, R. et al. (2000). Communication styles in the cancer consultation: Preferences for a patientcentered approach. Psychooncology 9(2), 147-156. Doyal, L. (2000). Gender equity in health: debates and dilemmas. Social Science and Medicine 51, 931-939. du Pre, A. (2005). Communicating about health: Curre nt issues and perspectives (2nd ed.). New York: McGraw-Hill. Eggly, S. (2002). Physician-patient co-construction of illness narratives in the medical interview. Health Communication 14, 339-360. Elderkin-Thompson, V. & Waitzkin, H. (1999). Differences in clinical communication by gender. Journal of General Internal Medicine 14, 112-121. Elwyn, G., Edwards, A., Mowle, S., Wensing, M ., Wilkensin, C., Kinnersle y, P. et al. (2001). Measuring the involvement of patients in sh ared-decision making: a systematic review of instruments. Patient Education and Counseling 43, 5-22. Epstein, R.M. (2005). Measuring patient -centered communication in patientphysician consultations: theore tical and practical issues. Social Science & Medicine 61(7), 1516-1528. Flores, G. (2000). Culture and the patient-physician relationship: achieving cultural competency in healthcare. Journal of Pediatrics 136, 14-23. Flores, G., Fuentes-Afflick, E., Barbot, O., Carter-Pokras, O., Claudi o, L., Lara, M. et al. (2002). The health of Latino children: urgent prio rities, unanswered questions, and a research agenda. Journal of American Medical Association 288, 82-90. Foucault, M. (1975). The Birth of the Clinic: An Ar cheology of Medical Perception New York: Random House.

PAGE 103

103 Friedson, E. (1970). Professional Dominance Chicago, IL: Aldine Press. Freidson, E. (1988). Profession of Medicine: A Study of the Sociology of Applied Knowledge 2nd ed. Chicago: University of Chicago Press. Freund, P.E., McGuire, M.B ., & Podhurst, L.S. (2003). Health, Illness, and the Social Body Upper Saddle River, NJ: Prentice Hall. Gorman, B.K., & Read, J.G. (2006). Gender disparitie s in adult health: an examination of three measures of morbidity. Journal of Health and Social Behavior 47, 95-110. Greenberg, C.C., Regenbogen, S.E ., Studdert, D.M., Lipsitz, S.R ., Rogers, S.O., & Zinner, M. J. et al. (2007). Patterns of communication breakdowns resu lting in injury to surgical patients. Journal of the American College of Surgeons, 204(4), 533-540. Grol, R., de Maeseneer, J., Whitfield, M., & Mokkink, H. (1990). Disease-centered versus patient-centered attitudes: comparison of gene ral practitioners in Belgium, Britain and the Netherlands. Family Practice 7(2), 100-104. Hall, J.A., Irish, J.T., Roter, D.L., Ehrlich, C.M., & Miller, L.H. (1994). Gender in medical encounters: an analysis of physician and pa tient communication in a primary care setting. Health Psychology, 13, 384-392. Hall, J.A. & Roter, D.L. (1995). Patient gender and communication with phys icians: results of a community-based study. Womens Health 1, 77-95. Halpern, J. (2007). Empathy and patient-physician conflicts. Journal of General Internal Medicine, 22(5), 696-700. Hamilton, E. & Cairns, H. (1961). Plato: The Collected Dialogues Princeton, NJ: Princeton University Press. Haskard, K.B., Williams, S.L., DiMatteo, R.M., Rosenthal, R., White, M.K., & Goldstein, M.G. (2008). Physician and patient communicat ion training in primary care: Effects on participation and satisfaction. Health Psychology 27(5), 513-522. Harvey, R.M., Kazis, L., & Lee, A.F. (1999). Decision-making preferences and opportunity in VA ambulatory care patients: associ ation with patient satisfaction. Research in Nursing & Health 22, 39-48. Haug, M.R. & Ory, M.L. (1987). Issues in elderly patient-provi der relationships. Research & Aging, 9, 3-44. Henbest, R., & Stewart, M. (1989). Patient-cen tredness in the consultation: a method for measurement. Family Practice, 6(4), 249-253.

PAGE 104

104 Heritage, J., & Maynard, D. (2006). Problems a nd prospects in the study of physician-patient interaction: 30 years of research. Annual Review of Sociology 32, 351-374. Institute of Medicine (2000). Committee on th e National Quality Report on Health Care Delivery. Board of Health Care Services. Envisioning the national health care quality report Washington, DC: National Academy Press. Institute of Medicine (2001). Comm ittee on Health Care in America. Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academy Press. Irish, J. (1997). Deciphering the phys ician-older patient interaction. International Journal of Psychiatric Medicine 27, 251-267. Irish, J.T., & Hall, J.A. (1995). Interruptive patterns in medical visits: the effects of role, status and gender. Social Science and Medicine 41(6), 873-881. Jahng, K.H., Martin, L.R., Golin, C.E., & DiMa tteo, R.M. (2005). Preferences for medical collaboration: patient-physician congruence and patient outcomes. Patient Education and Counseling 57, 308-314. Johnson, R.L., Saha, S., Arbelaez, J.J., (2004). Racial and Ethnic Differences in Patient Perceptions of Bias and Cultural Competence in Health Care. Journal of General Internal Medicine, 111. Klingle, R.S., & Burgoon, M. (1995). Patient co mpliance with and satisfaction physician influence attempts: Reinforcement expectan cy approach to compliance-gaining over time. Communication Research 22, 148-187. Korsch, B., & Negrete, V. (1972) Doctor-patient communication. Scientific America 227, 66-74. Krupt, E., Rosenkranz, S.L., Carter, M., Yeager, M.A., Barnard, K., Putnum, S.M. et al. (2000). The practice orientations of physicians and patients: the effect of doctor-patient congruence on satisfaction. Patient Education and Counseling 39, 49-59. Kundhal, K., & Kundhal, P. (2003). Cultural Dive rsity: an evolving ch allenge to physicianpatient communication. JAMA 289-294. Langewitz, W., Phillipp, E., Kiss, A., & Wossme r, B. (1998). Improving communication skills: a randomized controlled behavior ally-oriented intervention study for residents in internal medicine. Psychosomatic Medicine 60, 268-276. Levenstein, J.H., McCracken, E.C., McWhinney, I.R., Stewart, M.A., & Brown, J.B. (1986). The patient-centered clin ical method: A model for th e doctor-patient interaction in family medicine. Family Practice 3(1), pg. 24-30.

PAGE 105

105 Liaw, S.T., Young, D., & Farish, S. (1996). Improving patient-doctor concordance: an intervention study in general practice. Family Practice 13(5), 427-431. Link, B., & Phelan, J. (1995). Social conditi ons as fundamental causes of diseases. Journal of Health and Social Behavior 36, 80-94. Malin, J.L., Ko, C., Ayanian, J.Z., Harrington, D ., Nerenz, D.R., Kahn, K.L. et al. (2006). Understanding cancer patients experience and outcomes: development and pilot study of the Cancer Care Outcomes Research and Surveillance patient survey. Support Care Cancer 14, 837 Mast, M.S. (2007). On the importance of nonverb al communication in the physician-patient interaction. Patient Education and Counseling, 67, 315-318. Mast, M.S., Hall, J.A., & Roter, D.L. (2007) Disentangling physician sex and physician communication style: their effects on patient satisfaction in a virtual medical visit. Patient Education and Counseling 68, 16-22. May, C., & Mead, N. (1999). Patient-centeredne ss: a history. In C. Dowrick, & L. Frith, General practice and ethics: uncertainty and responsibility London: Routledge. McKinlay, J. (1997). Some contribut ions from the social system to gender inequalities in heart disease. Journal of Health and Social Behavior 37(1), 1-26. McWhinney, I. (1985). Patient-centered and doctorcentered models of clin ical decision making. In M. Sheldon, J. Brook, & A. Rector Decision making in general practice London: Stockton. Mead, N., & Bower, P. (2000). Patient-centredness: a conceptual framework and review of the empirical literature. Social Science and Medicine 51, 1087-1110. Morrell, D. (1972). Symptom interp retation in general practice. Journal of the Royal College of General Practitioners 22, 297-309. National Center for Health Statistics (2004) Health, United States, 2004. Washington, D.C.: U.S. Government Printing Office. Ngo-Metzger, Q., Telfair, J., Sorkin, D.H. (2006). Cultural Competency and Quality of Care: Obtaining the Patient's Perspective The Commonwealth Fund; October 2006. lications_show.htm?doc_id=414116. Accessed November 3, 2007. Parsons, T. (1951). The Social System. New York: free Press.

PAGE 106

106 Rogers, C. (1967). On becoming a person: a therapists view of psychotherapy London: Constable. Rogers, R., Hummer, R. & Nam, C. (2000). Living and dying in the USA: Behavioral, health, and social differentials of adult mortality New York: Academic Press. Roland, M. (1999). Quality and efficiency: enemies or partners? British Journal of General Practice 49, 140-143. Rose, J.H. (1990). Social support and cancer: Adu lt patients desire for support from family, friends, and health professionals. American Journal of Community Psychology 18, 439464. Roter, D. (1977). Patient participation in patient-provider interactions: th e effects of patient question asking on the quality of inte raction, satisfaction, and compliance. Health Education Monographs 5, 281-315. Roter, D. (2000). The enduring and evolving nature of the patientphysician relationship. Patient Education and Counseling 39, 5-15. Roter, D., Hall, J., & Katz, N. (1987). Relatio ns between physicians be haviors and analogue patients satisfaction, recall and impressions. Medical Care 25, 437-451. Roter, D., Lipkin, M., & Korsgaard, A. (1991). Sex differences in patients and physicians communication during primary care medical visits. Medical Care 29, 1083-1093. Roter, D.L. & Hall, J.A. (1992). Doctors talking to patients/patients talking to doctors: improving communication in medical visits Westport (CT): Auburn House. Roter, D.L., Stewar, M., Putnam, S.M., Lipki n, M.J., Stiles, W., & Inui, T.S. (1997). Communication patterns of primary care physicians. Journal of the American Medical Association 277(4), 350-356. Saha, S., Komaromy, M., Koepsell, T.D., & Bi ndman, A.B. (1999). Patin et-physician racial concordance and the perceived quality and use of health care. Archives of internal Medicine, 159, 997-1004. Schoenbaum, D.K., Collins, S.C., Tenney, K.S., Hughes, D.L., & Audet, A.J. (2002, April). Room for Improvement: Patients Report on the Quality of Their Health Care. The Commonwealth Fund, from /publications_show.htm?doc_id=221270. Shorter, E. (1985). Bedside Manners New York, NY: Simon and Schuster.

PAGE 107

107 Siminoff, L.A., Ravdin, P., Colabianchi, N., & Sauders-Sturm, C.M. (2000). Doctor-patient communication patterns in breast can cer adjuvant therapy discussions. Health Expectations, 3, 26-36. Stepanikova, I., Mollborn, S., Cook, K.S., Thom, D. H., & Kramer, R.M. (2006). Patients race, ethnicity, language and trust in a physician. Journal of Health and Social Behavior 47, 390-405. Stiles, W. (1978). Verbal response modes and dime nsions of interpersonal roles: a method of discourse analysis. Journal of Personality and Social Psychology 36, 693-703. Street, R.L (1992). Analyzing communication in medical consultations : do behavioral measures correspond to patients perceptions? Medical Care 30, 976-988. Street, R.L. (2002). Gender differences in health care provider-patient co mmunication: are they due to style, stereotypes, or accommodation? Patient Education and Counseling 48, 201206. Street, R.L., Gordon, H., & Haidet, P. (2007). Physicians communication and perceptions of patients: Is it how they look, how th ey talk, or is it just the doctor? Social Science and Medicine, 65, 586-598. Swenson, S.L., Zettler, P., & Lo, B. (2006). She gave it her best shot right away: patient experiences of biomedical and patient-centered communication. Patient Education and Counseling 61, 200-211. Szasz, P.S., & Hollender, M.H. (1956). A contribu tion to the philosophy of medicine: the basic model of the doctor-p atient relationship. Archives of Internal Medicine 97, 585-592. Talen, M.R., Grampp, K., Tucker, A.,& Schultz, J. (2008). What physicians want from their patients: Identifying what makes good patient communication. Family Systems & Health 26(1), 58-66. Teal, C.R. & Street, R.L. (2008). Critical elements of culturally competent communication in the medical encounter: A review and model. Social Science & Medicine in press, 1-11. Teutsch, C. (2003). Patie nt-doctor communication. The Medical Clinics of North America 87, 1115-1145. Todd, A.D. (1989). Intimate adversaries: cultural conf lict between doctors and women patients Philadelphia, PA: University of Pennsylvania Press. Tuckett, D., Boulton, M., Olsen, C., & Williams, A. (1985). Meetings between experts: an approach to sharing ideas in medical consultations London: Tavistock.

PAGE 108

108 U.S. Census Bureau. (2006). Alachua County, Florida Quick Facts Retrieved January 20, 2009, from Van Ryn, M. & Burke, J. (2000). The effect of patient race and socio-economic status on physicians perceptions of patients. Social Science & Medicine 50, 813-828. Walker, K., Arnold, C., Miller-Day, M., & Webb, L. (2002). Investigating the physician-patient relationship : examining emerging themes. Health Communication 14, 45-68. West, C. (1993). Reconceptualizing gende r in physician-patient relationships. Social Science and Medicine, 36, 57-66. Winefield, H., Murrell, T., Clifford, J., & Farmer, E. (1996). The search for reliable and valid measures of patient-centeredness. Psychology and Health 11, 811-824. Zandbelt, L., Smets, E., Oort, F., & de Haes, H. (2005). Coding patientcentered behavior in the medical encounter. Social Science and Medicine 61, 661-671. Zebiene, E., Svab, I., Sapoka, V., Kairys, J., Dots enko, M., Radic, S. et al. (2008). Agreement in patient-physician communication in primary care: A study from Central and Eastern Europe. Patient Education and Counseling 73, 246-250.

PAGE 109

109 BIOGRAPHICAL SKETCH Justin Joseph Coran earned his Bachelor of Arts degree from Universi ty of Florida in 2006 and received his Master of Arts degree from the University of Florida in the spring of 2009. He majored in sociology and supplemented his major with a strong medi cal science background. While finishing his doctoral degree in the Departme nt of Sociology at the Un iversity of Florida, he plans to concurrently comple te his Master of Public Hea lth degree through the College of Public Health and Health Professions at the Un iversity of Florida. Co rans primary research examines physician-patient inte raction with a particular focus on communication and the intersection of race, gender, and socioeconomic h ealth disparities. He has written manuscripts on a wide range of topics including new health co mmunication approaches, perceptions of physician competency and thoroughness, physicians percepti ons of communication, a nd key predictors of patient satisfaction with health care. Coran is currently conducting rese arch on instituting and examining a new medical communication approach for physicians and patients. He was recently elected to the National Communication Associati ons Health Communicati on Board as Graduate Student Representative. His current professional memberships include the American Sociological Association, National Communication Association, American Public Health Association, Southern Sociol ogical Association, and Flor ida Communication Association.