Purpose: Research conducted by physicians generates knowledge and has led to important advances and changes in the health care system. Physician retention is a concern facing many health care systems, and job satisfaction and attitudes play a role in retention. This study examined whether physicians who are involved with research have greater job satisfaction and more positive job characteristics perceptions.
Method: Cross-sectional analysis examined this research question using a sample of 7,734 physicians across 135 medical centers in the Department of Veterans Affairs (VA) in 2008. Using existing survey results, ratings on job characteristics (job autonomy, skill development opportunities, work and family balance, and performance feedback) and job satisfaction were regressed on research involvement in a multilevel hierarchical generalized linear model. The model controlled for physician-level characteristics and organization-level characteristics related to research activities.
Results: Analyses revealed that physicians who spent part of their time involved with research activities were more likely to report favorable job characteristics ratings. Physicians involved with research were also more likely to be satisfied with their job. Physicians who worked in medical centers with greater levels of research funding were more likely to report favorable ratings for job characteristics and job satisfaction.
Conclusions: Involvement with research was associated with more favorable job characteristics and job satisfaction perceptions among physicians in VA. Although there is a time and opportunity cost involved with research, medical centers that provide physicians with the opportunity to conduct research may have a more satisfied workforce.
Dr. Mohr is investigator, Center for Organization, Leadership and Management Research, VA Boston Healthcare System, and research assistant professor, Department of Health Policy and Management, Boston University School of Public Health, Boston, Massachusetts.
Dr. Burgess is senior investigator, Center for Organization, Leadership and Management Research, VA Boston Healthcare System, and professor, Department of Health Policy and Management, Boston University School of Public Health, Boston, Massachusetts.
Correspondence should be addressed to Dr. Mohr, VA Boston Healthcare System (152M), 150 South Huntington Avenue, Boston, MA 02130; telephone: (857) 364-5679; e-mail: firstname.lastname@example.org.
First published online June 20, 2011
Editor's Note: A commentary on this article appears on page 923.
Research conducted in academic medical centers by active clinical physicians has led to many important advances and generated the impetus to change health systems,1 guide evidence-based practices,2 increase the ability to understand global health issues,3 and influence legislation related to health issues.4 Medical school training that emphasizes research involvement can provide opportunities for physicians to gain knowledge and skills that facilitate working in multidisciplinary teams and delivering high-quality, evidence-based care. Physicians who are involved with research during medical school training are more likely to be involved with research following graduation.5
Studies of physician satisfaction have shown that greater physician satisfaction is associated with greater patient adherence to treatment6 and with greater patient satisfaction.7,8 Zuger,9 however, highlighted several longitudinal studies reporting that the physician population has become less satisfied over time. To address this, research opportunities may be leveraged to improve physician satisfaction. Physicians may pursue research activities to gain greater autonomy.10 Job autonomy may be a strong motivating force leading to greater satisfaction.11,12 Research opportunities are a particularly salient characteristic in attracting and retaining skilled employees in academic and general hospital systems. In a multiyear study conducted in the Veterans Health Administration, a majority (61%) of employees with medical degrees indicated they would not work at the organization if research opportunities were not available.13
Little research has examined the relationships that research involvement and research-relevant organizational characteristics have with physicians' perceptions of job characteristics and physicians' job satisfaction across broad disciplinary backgrounds. The purpose of this study was to examine the extent to which organization-level factors were associated with physician research involvement and the extent to which research involvement and individual-level factors influenced ratings on job characteristics and job satisfaction.
We used a model to examine factors that may affect job characteristics and physician satisfaction from both an individual- and an organization-level perspective. The Job Characteristics Model provides a framework to relate key job features to employee satisfaction and motivation.14,15 According to this model, five core job characteristics (skill variety; task identity; task significance; autonomy; and feedback) affect three critical psychological states: experiences of meaningfulness, responsibility for outcomes, and knowledge of results. These three psychological states affect job satisfaction, growth satisfaction, and motivation.
To frame our analyses, we also relied on self-determination theory, which focuses on the locus of causality to explain motivation. According to this theory, the perceived locus of causality includes external, internal, and impersonal factors that explain success or failure.16 People who perceive a greater locus of internal control are more likely to experience inherent satisfaction and enjoyment.17 The locus of causality also influences perceptions of job characteristics, such as autonomy. Because physicians who reported more autonomy also reported lower work overload, burnout, and quality-of-care ratings,18 the autonomy component may be a particularly relevant aspect in research to explain satisfaction.
Medical center characteristics may also influence job characteristics and job satisfaction. Physicians working within an academically affiliated hospital may have greater opportunities for involvement in collaborating on scholarly work, such as grants or manuscripts, and more supportive peer feedback for improvement. The activities may occur through the medical center or through the academic affiliate's research groups and committees. The distance between the university and the hospital, however, may influence opportunities for research involvement. Commuting to work can play a role in stress and job satisfaction.19 Universities located far away from the affiliated hospital may present an opportunity cost disincentive, because the physical distance could disrupt the physicians' commutes between home and the research and clinical setting.
Figure 1 displays a conceptual layout of the variables we tested in the study. At the provider level, research may meet goals identified as important to physicians: autonomy, flexibility, and skill development opportunities. Accordingly, we hypothesized that ratings on these dimensions should be higher for physicians involved with research. Research involvement alone may be directly associated with higher job satisfaction, but research involvement may additionally moderate the relationship between job characteristics and job satisfaction; a positive relationship between job characteristics and job satisfaction increases when physicians are involved with research.
At the organization level, we anticipated that medical center research activities and academic affiliate structures influence physician involvement in research. Physicians who work at a medical center with an academic affiliation and within a relatively short traveling distance may have more opportunities to become involved with research using the established transorganizational infrastructure. Such organizational factors should lead to higher ratings on job characteristics and job satisfaction. We offer an alternative possibility that, although physicians can benefit from working at an academic affiliate, there may be significant challenges and difficulties to doing so that may minimize benefits or present more conflicts and demands. Specifically, the medical center and academic affiliate may differ on several important priorities and policies. The discrepancies could strain the physician to balance multiple competing and conflicting demands. This could lead to role ambiguity or conflict if work methods, scheduling, and performance criteria are unclear.20 This may also lead to lower job satisfaction, commitment, and performance.21
Data sources and sample
The current study focuses on the Department of Veterans Affairs (VA) health care system, an organization with a strong and mission-guiding research focus, though the implementation of this mission varies considerably across medical centers. The four-part mission of VA is to provide patient care, conduct research, provide teaching opportunities, and serve as a contingency backup for the Department of Defense.22 VA is the largest provider of medical and health education in the country.23 VA devotes a significant budget to research. Including funds from outside sources, VA had a total research budget of approximately $1.8 billion in 2008.24
The institutional review board at the VA Boston Healthcare System provided ethical approval of the study.
To model VA physician characteristics and job attitudes, we used the results of the 2008 All Employee Survey (AES), which VA has administered annually since 2004. VA administered the AES to the census of full-time employees during April and May 2008. For the first time, a question about involvement in research was included on the questionnaire. The questionnaire contained 8 demographic items and 58 items using a five-point Likert response scale. The questionnaire contained three sections to assess employees' satisfaction with key job features, perceptions of work group conditions, and perceptions of organizational culture. To raise awareness of the survey, VA routinely engages in large-scale marketing efforts, including the presence of survey coordinators at every medical center, regular e-mail updates from local medical center leadership, publicly visible notices, and group incentives for high participation rates. Survey coordinators requested that only full-time VA-based employees complete the survey, and from this group, we examined only physicians' responses. Employees responded to the survey using one of three modes: Internet, paper-and-pencil survey forms, or interactive voice response. Most employees completed the survey using the Internet mode of administration. Over 226,000 VA employees were eligible to participate in 2008; of these, 164,502 completed the survey for a response rate of 72.8% overall. In 2008, the national full-time employee equivalent (i.e., full-time, part-time, or other) number of physicians was 15,175; 12,560 physicians had full-time employment status. The AES received responses from 8,215 self-identified physicians, and we were able to fully match 7,734 respondents, for a 65% full-time physician response rate. Our response rate compares favorably to non-VA physician survey response rates among surveys using different modes of administration.25–27
To model VA institutions' number of academic affiliations and distance between the medical centers and affiliated institutions, we used the Inventory of Organization Characteristics (IVOC). This instrument was administered in the fall of 2005 and completed by all VA medical centers (n = 136) and networks (n = 21) to collect information on organizational structures (e.g., service lines), leadership activities (e.g., decision making), and performance and quality activities (e.g., physician champions) that could affect clinical practices and patient care.28 We obtained information about research dollars received by investigators at each medical center for VA- and non-VA-funded projects from the Veterans Equitable Resource Allocation database.
Using individual data from the AES, we modeled the following biological markers: age, race, and gender. We assessed organizational status indicators using categorical variables for tenure and supervisory authority and physician specialty occupation from a list of six classes. We modeled work setting as either Community Based Outpatient Clinic (CBOC) or hospital.
To assess participants' research involvement, we used the question: “In what type of setting do you spend at least 20% of your time?” Physicians selected up to five options: research, inpatient, outpatient, administrative, or extended care. The use of a 20% cut point to classify individuals into a research category is similar to a 15% cut point used in a recent study.29 We coded the responses dichotomously (0 if involvement was not reported; 1 if it was).
We computed the number of major medical school affiliations with VA medical centers and networks using IVOC data. Respondents also indicated whether there was a medical school located “on the same campus,” “within walking distance,” or “not within walking distance” for each of the affiliated schools. Additionally, we dichotomously coded whether the named academic affiliate was listed as a top 25 medical school (0 if it was not; 1 if it was).30 We used the total amount of research dollars that each organization received to account for the medical center's research connection using four ranges for funding categories: none; less than $1 million; $1 million to $5 million; and more than $5 million. Urban or rural setting was also included in the model.
On the basis of the theoretical model, we used four items from the AES to assess job characteristics: new skill development opportunities, feedback from supervisors, job autonomy, and work and family balance. Physicians rated these items using a five-point Likert scale (1 = strongly disagree; 5 = strongly agree). Similar to other research studies on physician satisfaction,7,31,32 we dichotomously coded items to indicate positive perceptions of the attribute in question. We coded “agree” or “strongly agree” responses with a value of 1, and we coded “disagree,” “strongly disagree,” and “neither agree nor disagree” responses with a value of 0.
Overall job satisfaction.
We assessed overall job satisfaction using responses to the question, “Compared to what you think it should be, what is your overall level of job satisfaction?” Response options ranged from 1 (“not at all satisfied”) to 5 (“very satisfied”). We coded a response of “satisfied” or “very satisfied” with a value of 1, and all other nonmissing responses were coded with a value of 0.
The question asks respondents to compare their current level of satisfaction with their ideal level of satisfaction. This item applies discrepancy theory, in which dissatisfaction occurs when someone receives less than what was expected.33 A single question framed this way is a reliable and valid approach compared with asking a series of questions on specific job satisfaction facets.34 In previous research using VA primary care team teams, we found a positive association between this item and process and intermediate outcome quality-of-care indicators, suggesting it is a potentially important marker for care effectiveness.35
First, we examined descriptive statistics for employee and organizational characteristics. We computed inferential statistics comparing physician research involvement with organizational characteristics and job characteristics, and comparing overall job satisfaction ratings with research involvement. To build the multilevel model, we calculated Spearman correlations to identify and delete variables with values greater than 0.50.
To examine the independent associations of each employee and organizational characteristic, we constructed a cross-sectional hierarchical generalized linear model (HGLM) using HLM 6.06 (Scientific Software International, Inc., Lincolnwood, Illinois). We used robust standard errors. We estimated HGLMs where each of the four job characteristics was a dependent variable with employee characteristics entered as level 1 variables of interest and organizational characteristics entered as level 2 variables of interest.
To examine the effect of individual and organizational characteristics on job satisfaction, we used a multistep model process. In the first step, we entered individual and organizational characteristics into the regression to determine their effect on overall job satisfaction. In the second step, we added the four job characteristics variables to the regression. In the third step, we added the interaction term between research involvement and each of the four job characteristics variables. We also ran models using the full five-point range of response options.
In Table 1, we present the descriptive characteristics of the sample. Approximately 12% (n = 901) of respondents reported research involvement. Because respondents could indicate working in only one or multiple settings, we tested whether the job characteristics and job satisfaction of physicians working only in a research setting (11%, n = 103/901) differed from those of physicians working in research and at least one other setting (89%, n = 798/901) using chi-square tests. Because none of the comparisons reached a significance level less than .10, we did not further specify the research involvement variable in regression analysis.
In Table 2, we present data on respondents who provided favorable responses to the job characteristics items and the overall job satisfaction item. In general, physicians involved with research activity provided higher ratings on all dimensions. The largest differences occurred in the dimensions of job autonomy and skill development opportunities. Of those with research activity, 62.1% (577/929) responded favorably about their level of autonomy compared with 45.3% (3,233/7,132) of those with no research activity, and 75.2% (700/931) of those with research activity responded favorably about their opportunity to develop skills compared with 63.8% (4,545/7,121) of those with no research activity. For overall job satisfaction, 78.6% (741/942) of physicians involved with research reported a favorable rating compared with 71.9% (5,198/7,234) of physicians not involved with research.
As seen in Table 3, physicians involved with research were more likely to provide a favorable rating on job characteristics variables; odds ratios (ORs) ranged from 1.29 to 1.33. Physicians located in medical centers with academic affiliates reported less favorable ratings for skill development opportunities and work and family balance. When the academic affiliate was located within walking distance, we found significant effects on performance feedback, skill development opportunities, and work and family balance; ORs ranged from 1.16 to 1.18. The organizational research funding level was significantly related to higher ratings for all job characteristics; as the level of funding increased, the estimates for favorable responses also increased.
In Table 4, we present the three additive models for overall job satisfaction. In the first step of the model, research involvement was significantly associated with a greater likelihood of favorable overall job satisfaction (OR = 1.23). In the second step of the model, involving overall job satisfaction and job characteristics, we found that all job characteristics variables were significantly associated with a greater likelihood of favorable overall job satisfaction (ORs ranged from 2.29 to 4.76). While accounting for job characteristics, research involvement was not significant. In the third step of the model testing interactions between research involvement and job characteristics, all job characteristics variables were also significant (ORs ranged from 2.28 to 4.84), but none of the interactions were significant. Additionally, we examined these models using the full five-point range of response options from the AES. Our findings, however, did not differ in a meaningful way or influence the conclusions.
We examined whether physician research involvement and research-oriented organizational characteristics were associated with higher ratings on job characteristics and overall job satisfaction in VA. In multilevel logistic regression models, we found that research involvement had a positive association with job characteristics and overall job satisfaction. Additionally, we found that favorable job characteristics ratings were associated with greater overall job satisfaction, with the strongest effect for skill development opportunities. Our results are exploratory in a field with limited data and literature; however, findings provide initial support for potential benefits of physician research involvement. As an application of our findings, we suggest that health care organizations place an emphasis on research activities and allow protected physician time for research. Although this may come at a cost in physicians' ability to see patients, in the long run the time may lead to more organizational commitment, job satisfaction, and tenure.
Although VA physicians involved with research reported higher overall job satisfaction, it remains unclear how this difference translates into specific outcomes over time. For example, an analysis testing whether physicians involved with research have lower turnover rates, different career trajectories, leadership roles, or provide more evidence-based care may add value. Further longitudinal research and analysis can foster a better understanding of long-term associations and implications.
We note some limitations and future research directions. The study examined variables commonly used to explain job satisfaction. The connection between individual and organizational factors with job satisfaction is complex. We note that variables not measured, such as coworkers, skill variety, professional identification, climate, and culture, may influence physician satisfaction. Using census survey data was also a limitation because it prevented flexibility and precision in asking relevant questions, which would be a next step for future studies.
Additionally, we were unable to assess the number of VA physicians who wanted to be involved with research but were unable to do so. Restricted opportunities may lead to disappointment or lower job satisfaction. For example, physicians spending at least 20% of their time in the aspect of work (i.e., patient care, education, research) they find most meaningful reported lower burnout rates.36 There are also differences in the types of research activities that physicians may prefer. For example, some physicians may enjoy interacting with patients on clinical trials, whereas others may prefer grant writing. The difference between the actual activities and preferred activities for research and other work domains could lead to important differences among groups.
VA physicians who worked at medical centers with one or more academic affiliates were less likely to provide favorable ratings for job characteristics and overall job satisfaction. Competing policies, priorities, and demands could possibly be involved in determining this result. Exploring discrepancies and tensions between medical centers and academic policies and their influence on physician attitudes and behaviors also would be important to investigate. Physicians were more likely to provide favorable job characteristics and overall job satisfaction if they worked at a medical center with an academic affiliate within walking distance. The proximity of the medical center may facilitate travel and research collaboration with other physicians. This study, however, did not directly assess ratings for convenience of conducting research, or collaborating partner location; thus, further exploration is warranted.
There also are selection and endogeneity problems that arise from the general knowledge that VA is research-friendly. Physicians choosing VA may be systematically different from other physicians. Ideally, more settings with nonresearch environments and controls for employee selection would also sharpen conclusions from this line of research.
Although the ability to examine a large number of medical centers and employees was a strength, we did not investigate differences in research norms and practices among clinical departments. For example, wide variation can exist on research practices across departments within a university.37 Thus, although the surgery department may be very active in research, it does not necessarily mean that the rehabilitation department will be as active or have the same research practices and norms. Qualitative observational research can provide additional context for interpretation and assessment of interrelationships among research behaviors and values, job attitudes, and organizational characteristics among physicians and hospital management.
We found that research involvement was associated with more favorable job characteristics and overall job satisfaction ratings among a sample of VA physicians in 2008. At the organization level, medical centers with academic affiliations were associated with less favorable ratings on some of the job characteristics and overall job satisfaction. When the academic affiliate was near the medical center, physicians generally reported more favorable job characteristics ratings. Our findings are generally supportive of the VA research mission being associated with a positive impact on the working conditions and overall job satisfaction of physicians. Further research to understand the complete picture and the impact of research on the quality and cost of patient care will be valuable. More generally, many of the readers of this journal (both physicians and nonphysicians) have elements of these multiple objective job characteristics. We hope this study motivates future research on these working conditions within the research context and implications of those job design issues.
The authors wish to thank colleagues at the VA Center for Organization, Leadership and Management Research and the VA National Center for Organization Development.
This material is based on work supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development for IIR 05-221.
The institutional review board at the VA Boston Healthcare System provided ethical approval of the study.
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
1 Commission on Health Research for Development. Health Research: Essential Link to Equity in Development. New York, NY: Oxford University Press: 1990.
2 Oxman AD, Fretheim A, Schünemann HJ. Improving the use of research evidence in guideline development: Introduction. Health Res Policy Syst. 2006;4:14.
3 Pittman MA, Svensson P. Increasing the global exchange of evidence-based research. Health Serv Res. 2006;41:303–309.
4 Oliver TR, Singer RF. Health services research as a source of legislative analysis and input: The role of the California Health Benefits Review program. Health Serv Res. 2006;43(3 pt 2):1124–1158.
6 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.
7 Haas JS, Cook EF, Puopolo AL, et al. Is the professional satisfaction of general internists associated with patient satisfaction? J Gen Intern Med. 2000;15:122–128.
8 DeVoe J, Fryer GE Jr, Straub A, McCann J, Fairbrother G. Congruent satisfaction: Is there geographic correlation between patient and physician satisfaction? Med Care. 2007;45:88–94.
9 Zuger A. Dissatisfaction with medical practice. N Engl J Med. 2004;350:69–75.
10 Tight M, Kayrooz C, Åkerlind GS. Autonomy in Social Science Research: The View From United Kingdom and Australian Universities. Oxford, UK: Emerald Group Publishing; 2007.
11 Linzer M, Konrad T, Douglas J, et al. Managed care, time pressure and physician job satisfaction: Results from the physician worklife study. J Gen Intern Med. 2000;15:441–450.
12 Pratt WR. Physician career satisfaction: Examining perspectives of the working environment. Hosp Topics. 2010;88:43–52.
13 Valley D, Meterko M, VanDeusen Lukas C, Nealon Seibert M, Charns M. National Survey of VA Researchers. Boston, Mass: Management Decision and Research Center; 2002.
14 Hackman JR, Oldham GR. Development of the job diagnostic survey. J Appl Psychol. 1975;60:159–170.
15 Hackman JR, Oldham GR. Motivation through the design of work. Organ Behav Human Perf. 1976;16:250–279.
16 Deci EL, Ryan RM. Intrinsic Motivation and Self-Determination in Human Behavior. New York, NY: Plenum; 1985.
17 Ryan RM, Deci EL. Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. Am Psychol. 2000;55:68–78.
18 Halbesleben JRB, Rathert C. Linking physician burnout and patient outcomes: Exploring the dyadic relationship between physicians and patients. Health Care Manage Rev. 2008;33:29–39.
19 Koslowsky M, Krausz M. On the relationship between commuting, stress symptoms, and attitudinal measures: A LISREL application. J Appl Behav Sci. 1993;29:485–492.
20 Breaugh JA, Colihan JP. Measuring facets of job ambiguity: Construct validity evidence. J Appl Psychol. 1994;79:191–202.
21 Singh J. Striking a balance in boundary-spanning positions: An investigation of some unconventional influences of role stressors and job characteristics on job outcomes of salespeople. J Market. 1998;62:69–86.
22 Kizer KW. The “new VA”: A national laboratory for health care quality management. Am J Med Qual. 1999;14:3–20.
25 Meterko M, Young GJ, White B, et al. Provider attitudes toward pay-for-performance programs: Development and validation of a measurement instrument. Health Serv Res. 2006;41:1959–1978.
26 Sibbald B, Addington-Hall JM, Brenneman D, Freeling P. Telephone versus postal surveys of general practitioners: Methodological considerations. Br J Gen Pract. 1994;44:297–300.
27 Cummings SM, Savitz LA, Konrad TR. Reported response rates to mailed physician questionnaires. Health Serv Res. 2001;35:1348–1355.
29 Chokshi NK, Simeone DM, Chari RS, Dorey F, Guner YS, Upperman JS. A survey of academic surgeons: Work, stess and research. Surgery. 2009;146:462–468.
31 DeVoe J, Fryer GE Jr, Hargraves L, Phillips RL, Green LA. Does career dissatisfaction affect the ability of family physicians to deliver high-quality patient care? J Fam Pract. 2002;51:223–228.
32 Katerndahl D, Parchman M, Wood R. Perceived complexity of care, perceived autonomy, and career satisfaction among primary care physicians. J Am Board Fam Med. 2009;22:24–33.
33 Berry LM. Psychology at Work. San Francisco, Calif: McGraw Hill Companies, Inc.; 1997.
34 Nagy MS. Using a single-item approach to measure facet job satisfaction. J Occup Organ Psychol. 2002;75:77–86.
35 Mohr DC, Young GJ, Meterko M, Stolzmann KL, White B. Job satisfaction of primary care team members and quality of care. Am J Med Q. 2011;26:18–25.
36 Shanafelt TD, West CP, Sloan JA, et al. Career fit and burnout among academic faculty. Arch Intern Med. 2009;169:990–995.
© 2011 Association of American Medical Colleges
37 Fairweather JS, Beach AL. Variations in faculty work at research universities: Implications for state and institutional policy. Rev Higher Educ. 2002;26:97–115.