Physician job satisfaction

Physician Job Satisfaction Developing a Model Using Qualitative Data Julia E. McMurray, MD, Eric Williams, PhD, Mark D. Schwartz, MD, Jeffrey Douglas, PhD, Judith Van Kirk, MS, T. Robert Konrad, PhD, Martha Gerrity, MD, PhD, Judy Ann Bigby, MD, Mark Linzer, MD, for the SGIM Career Satisfaction Study Group (CSSG) The purpose of this study was to develop a current and comprehensive model of physician job satisfaction. Information was gathered by (1) analysis of open-ended responses from a large group practice physician survey in 1988, and (2) analysis of focus group data of diverse physician subgroups from 1995. Participants were 302 physicians from large-group practices and 26 participants in six focus groups of HMO, women, minority, and inner-city physicians. Data were used to develop a comprehensive model of physician job satisfaction. The large group practice survey data supported the key importance of day–to–day practice environment and relationships with patients and physician peers. Future concerns focused on the effect of managed care on the physician–patient relationship and the ability of physicians to provide quality care. Focus groups provided contemporary data on physician job satisfaction, reinforcing the centrality of relationships as well as special issues for diverse physician subgroups of practicing physicians. New variables that relate to physician job satisfaction have emerged from economic and organizational changes in medicine and from increasing heterogeneity of physicians with respect to gender, ethnicity, and type of practice. A more comprehensive model of physician job satisfac- tion may enable individual physicians and health care organizations to better understand and improve physician work life. KEY WORDS: physician job satisfaction, qualitative analysis, women physicians, career satisfaction. J GEN INTERN MED 1997;12:711–714. M ultiple studies have detailed select aspects of physician job satisfaction,1–6 but little has been done to systematically identify variables that influence career satisfaction. In addition, important physician subgroups, such as women, minorities, and inner-city physicians, have been underrepresented in the literature on this subject.5,6 In this study, we incorporate previous research7–14 with new analyses of qualitative data to refine and further assess a multidimensional, comprehensive set of variables related to physician job satisfaction. This report details methods of model development,15,16 a revised model, and differences in job satisfaction due to ethnicity, gender, and specialty. METHODS Development of Physician Job Satisfaction Variable Set Received from the Departments of Medicine (JEM, JVK, ML) and Biostatistics (JD), University of Wisconsin School of Medicine, Madison; the Sheps Center, Division on Health Professionals (EW, TRK), University of North Carolina at Chapel Hill; Division of Primary Care, Department of Internal Medicine, New York University, New York (MDS); Department of Medicine, Oregon Health Sciences University, Portland (MG); Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Mass. (JB); and the Society of General Internal Medicine (JEM, MDS, MG, JAB, ML), Washington, DC. Supported by a grant from the Robert Wood Johnson Foundation. Dr. Williams is currently affiliated with the Program for Health Services Management, University of Missouri, Columbia. Other members of the CSSG are Donald Pathman, MD, Univ. of North Carolina, Chapel Hill; Elnora Rhodes, SGIM, Washington, DC; John Frey, MD, Dept. of Family Medicine, Univ. of Wisconsin, Madison; Kathleen G. Nelson, MD, Dept. of Pediatrics, Children’s Hospital, Birmingham, Ala.; William E. Scheckler, MD, Dept. of Family Medicine, Univ. of Wisconsin Medical School, Madison; Mary Ramsbottom-Lucier, MD, Univ. of Kentucky, Lexington; Richard Shugerman, MD, Univ. of Washington, Seattle. Address correspondence and reprint requests to Dr. McMurray: J5/210 CSC, University of Wisconsin School of Medicine, Department of Medicine, 600 Highland Ave., Madison, WI 53772. An initial set of factors important to physician job satisfaction (MDSat) was developed by physician and social scientist investigators using previous research on physician satisfaction,1–6 studies by the SGIM Career Choice Task Force,12–14 previous work of other study investigators,7–11 and a sample of open-ended responses from the 1988 Large Group Practice Physician Satisfaction Survey (Table 1) that were not used in the subsequent validation process. This item pool of variables was revised using the above survey data as well as focus group analysis. Large Group Practice Qualitative Data The Large Group Practice Physician Satisfaction Survey was distributed in 1988 to 8,000 physicians, 50% of whom were in primary care. Forty percent of respondents answered three open-ended survey questions regarding satisfaction, dissatisfaction, and future concerns. From a computerized randomization scheme, we analyzed a convenience sample of approximately 10% of respondents (n 5 302, 110 women and 192 men, with minorities oversampled). Using standard methods of qualitative analysis,15,16 responses were bracketed to highlight relevant phrases. Two trained coders unacquainted with the study’s hypotheses entered phrases into an Excel spread sheet and inde711 712 McMurray et al., Physician Job Satisfaction pendently coded them using the MDSat variable set (Table 1). A third coder broke ties when necessary (less than 5% of statements). A k value of 0.71 indicated good reliability of coding. Response frequencies were tabulated and analyzed according to physician gender, specialty, and ethnicity. Focus Group Analysis Physician focus groups were recruited locally in 1995 by group leaders to obtain input from managed care, female, inner-city, and minority physicians from the West Coast, Midwest, New York City, and Boston, respectively. Group selection was based on the desire to validate the JGIM MDSat variable set and expose previously unexplored issues in physician job satisfaction not available from the large group practice survey. Focus group leaders posed standardized open-ended questions regarding daily satisfaction, dissatisfaction, and future concerns. Transcripts were entered into Ethnograph, a computer software program for qualitative database analysis (Quintiles, Boston, Mass., 1991). A trained coder assigned codes using MDSat for all relevant phrases and paragraphs. A second investigator reviewed transcripts and codes for accuracy as well as major constructs. Code frequencies were tabulated within groups and comments were reviewed. Qualitative data from the large group prac- Table 1. Original Variable Set of Physician Satisfaction (MDSat) and Revised Set (MDSatR) MDSat Variable Set Variable 1. Relationships 10 Relationships with patients 11 Relationships with colleagues 12 Relationships with administrators Variable 2. Personal/family characteristics 20 Issues of aging 21 Family issues 22 Mission concordance 23 Career advancement opportunities 24 Job security 25 Keeping up/continuing medical education 26 Personal time Variable 3. Day-to-day practice issues 30 Stress and workload 31 Paperwork hassles 32 Variety of patients/intellectual stimulation 33 Ancillary staff 34 Access to specialists 35 Academics Variable 4. Administrative and organizational issues 40 Impact of HMOs 41 Ability to have input into administrative decisions 42 Being spared administrative work 43 Access or ability to communicate with leadership 44 Competency of leadership 45 Feedback from the organization 46 Size of organization Variable 5. Government issues 50 Regulations 51 Malpractice Variable 6. Autonomy 60 Control of schedule 61 Control of medical decision making 62 Control over workplace issues Variable 7. Income and prestige 70 Pay/benefits 71 Pay relative to hours worked 72 Pay relative to what others make 73 Respect/status Variable 8. Quality of care 80 Ability to provide quality care in current setting MDSatR Variable Set Variable 1. Relationships 11 Patients 12 Colleagues 13 Team members in office/hospital 14 Community 15 Administrators Variable 2. Personal and family characteristics 21 Family issues 22 Racial and ethnic and gender issues 23 Personal growth/mission concordance 24 Geography 25 Training characteristics 26 Personal time Variable 3. Day-to-day practice characteristics 31 Stress in day-to-day practice/hassle factor 32 Workload 33 Availability of office and hospital resources 34 Intellectual stimulation 35 Case mix/patient variety 36 Access to specialists 37 Academics/teaching/research 38 Malpractice worries 39 Keeping up/continuing medical education Variable 4. Administrative and organizational issues 41 Organizational characteristics (size, type) 42 Ability to have input into administrative decisions 43 Level of administrative work 44 Issues of productivity and cost containment 45 Utilization review/insurance 46 Paperwork hassles 47 Job security Variable 5. Autonomy 51 Control over workplace issues 52 Control of medical decision making Variable 6. Income and prestige 61 Income and benefits 62 Pay relative to what others make 63 Respect and status Variable 7. Quality of care 71 Ability to provide quality care Variable 8. Expectations 81 Discrepancy between job expectations and experience 713 Volume 12, November 1997 JGIM tice survey were triangulated with focus group responses to further assess validity and to clarify the experiences of each subgroup. Modification of Physician Job Satisfaction Variable Set Patterns, similarities, and differences in qualitative responses from the 1988 survey and 1995 focus groups were analyzed in monthly conference calls and in a modified Delphi technique during the investigators’ meeting in November 1995, resulting in a revised variable set (MDSatR). Two coders then used MDSatR to code an independent second set of questionnaires from the 1988 study (n 5 144). Response frequencies were found to be similar to those in the initial questionnaire analysis. A k value of 0.61 was considered acceptable for this second analysis, although somewhat less than desired. This may have represented the coders’ lack of familiarity with the revised coding scheme. Table 2. Percentage of Physicians Making Comments About Satisfaction, Dissatisfaction, and Future Concerns in 1988 Large Group Practice Physician Survey (N 5 302 Physicians) Variable Relationships Personal/family Day-to-day practice Administrative issues Government issues Autonomy Income Quality of care Satisfaction, % Dissatisfaction, % Future Concerns, % 36 10 22 5 20 15 47 43 17 15 34 30 1 11 10 8 1 12 16 2 5 13 14 18 Focus Groups RESULTS Large Group Practice Survey As sources of satisfaction, physicians cited day-today practice issues, relationships with patients and colleagues, and positive aspects of administrative issues such as “concentration on patient care with management done by professionals.” Physicians were dissatisfied with stress-related aspects of day-to-day practice, such as workload and patient volume. Future concerns emphasized the anticipated effects of managed care on physician relationships with colleagues and patients and the negative effects that an intensified focus on cost containment and productivity would have on the quality of care. Table 2 shows the rank ordering of these variables. Managed care group participants discussed paperwork hassles and noted problems with continuity of care when patients switched plans. Women physicians emphasized the satisfaction from providing “total care” to their patients, but had concerns about workload, case mix, balance or role conflict, and delayed professional advancement. Minority physicians sought more like-minded colleagues and discussed the pressures of being a role model and of being “all things to all people.” Inner-city physicians had a sense of “returning to one’s roots” and expressed a strong sense of mission; major concerns included isolation and the “burden of caring.” Table 3 is a tabulation of the comments made by physicians in the separate focus groups and shows the rank ordering of the variables. Table 3. Components of Physician Job Satisfaction Identified by Physician Focus Groups Variable Relationships Patients Colleagues Administration Personal/family Balance of work/family Mission concordance Day-to-day practice Stress Paperwork Administrative issues Government issues Autonomy Income/prestige Quality of care * n 5 number of comments. HMO, % (n 5 186)* Women, % (n 5 163) Minority, % (n 5 95) Inner-City, % (n 5 52) 45 23 18 4 10 4 2 14 3 5 25 1 3 1 1 42 20 18 4 14 12 1 20 9 4 6 1 6 7 4 53 23 26 4 16 3 11 6 0 2 17 0 0 4 3 42 23 2 17 15 4 12 10 2 0 19 4 2 6 2 714 McMurray et al., Physician Job Satisfaction MDSat was modified to MDSatR by expanding the variables on relationships, personal and family characteristics, and administrative and organizational issues. Government regulation was dropped, and a new variable (“expectations”) was added to assess discrepancies between job expectations and experiences. DISCUSSION Previous research on physician satisfaction has suffered from using nonrepresentative physician populations,6 or has studied only limited facets of satisfaction.4–6 Our study provides a multidimensional taxonomy of satisfaction applicable to a variety of health care environments and relevant to the careers of special physician populations. It incorporates variables that reflect crucial contemporary concerns with productivity, cost containment, and the impact of variation in case mix. Our findings suggest the following: (1) relationships and day-to-day practice issues are key components of physician satisfaction, and (2) different components of overall job satisfaction may be more or less relevant to specific physician subgroups. For example, balance of work and family commitments was an issue for women physicians, a sense of mission was important to minority and innercity physicians, and administrative issues were relevant for those in managed care. The strengths of our analysis include the use of national survey data, the ability to triangulate data from the quantitative aspects of the survey with open-ended responses from both survey and focus groups, and the use of homogeneous focus groups as a way of uncovering current satisfaction determinants of groups not usually represented. Our study is limited by the use of 1988 survey data obtained only from large group practice physicians, whose values and satisfaction issues may be different from those of currently practicing physicians in other practice types. Also, the model may not apply to all physicians (e.g., those in rural-based or solo practices). Better understanding of physician satisfaction may improve retention and performance in clinical practice. Use of this model may allow managers of health care to better understand physician practice styles, to maximize quality of care, and to maintain a stable workforce. It is critical that the job experiences and values of currently practicing physicians be understood to maintain what is vital to those practitioners and to safeguard the profession and the health of the public. JGIM The authors thank Peter Allen, Akiko Yagi, and Steven Pattrick for their technical assistance and expertise and Janet Merlo and Jan Holmes for secretarial assistance. REFERENCES 1. Lewis CE, Prout DM, Leake B. How satisfying is the practice of internal medicine? Ann Intern Med. 1991;114:1–5. 2. Cohen AB, Cantor JC, Barker DC, Schuster AC, Reynolds RC. Young physicians and the future of the medical profession. Health Affairs. 1990;9:138–48. 3. DiMatteo MR, Sherbourne CD, Hays RD, et al. Physicians’ characteristics influence patients’ adherence to medical treatment: results from the Medical Outcomes Study. Health Psychol. 1993;12: 93–102. 4. Melville A. Job satisfaction in general practice: implications for prescribing. Soc Sci Med. 1980;14:495–9. 5. Stamps PL, Cruz NTB. Issues in Physician Satisfaction. Ann Arbor, Mich: New Perspectives in Health Administration Press; 1994. 6. Lichtenstein R. Measuring the job satisfaction of physicians in organized settings. Med Care. 1978;16:850. 7. Madison DL, Konrad TR. Large medical group-practice organizations and employed physicians: a relationship in transition. Milbank Q. 1988;66(2):240–82. 8. Konrad TR, Madison DL, Zavala E, Defriese GH. Managed care systems and transitions in large medical practices. In: Proceedings Group Health Institute, 1988. Chicago, IL: Group Health Association of America; 1989. 9. DeFriese GH, Konrad TR. The impact of financial and organizational changes on primary care. In: Primary Care Research: An Agenda for the 90’s. Washington, DC: U.S. Dept. Health and Human Services, Public Health Service, Agency for Health Care Policy and Research; 1990:97–106. 10. Konrad TR, Kory WP, Madison DL, et al. The salaried physician: medical practice in transition. Final report developed under grant HS05557 for the Agency for Health Care Policy and Research. Chapel Hill, NC: Cecil Sheps Center for Health Sciences Research, University of North Carolina at Chapel Hill; 1989. 11. Pathman DE, Konrad TR, Ricketts TC. The National Health Service Corps experience for rural physicians in the late 1980’s. JAMA. 1994;272:1341–8. 12. Schwartz MD, Linzer M, Babbott D, Divine GW, Broadhead E, and the SGIM Interest Group on Career Choice in Internal Medicine. Medical student interest in internal medicine. Initial report of the SGIM Interest Group Survey on factors influencing career choice in internal medicine. Ann Intern Med. 1991;114:6–15. 13. McMurray JE, Schwartz MD, Genero NP, Linzer M, for the SGIM Task Force on Career Choice in Internal Medicine. The attractiveness of internal medicine: a qualitative analysis of the experiences of female and male medical students. Ann Intern Med. 1993;119:812–8. 14. Schwartz MD, Linzer M, Babbott D, Divine GW, Broadhead WG. The impact of an ambulatory rotation on medical student interest in internal medicine. J Gen Intern Med. 1995;10:542–9. 15. Miles MB, Huberman AM. Qualitative Data Analysis. Thousand Oaks, Calif: Sage Publications; 1994. 16. Crabtree BF, Miller WL, eds. Doing Qualitative Research. Research Methods of Primary Care. Vol. 3. Thousand Oaks, Calif: Sage Publications; 1992.