| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Health Policy and Practice |
1 From the Research Department, American College of Radiology, 1891 Preston White Dr, Reston, VA 20191. Received September 18, 1998; revision requested October 15; final revision received February 26, 1999; accepted May 13. Address reprint requests to J.H.S. (e-mail: jonathans@acr.org).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Questionnaires were mailed during the winter of 1994–1995 to a stratified random sample of 3,024 diagnostic radiologists, radiation oncologists, and nuclear medicine specialists; 75% responded. We weighted the responses to make them representative of all radiologists in the United States and compared the findings with those of similar previous surveys.
RESULTS: Fifty-one percent of radiologists would recommend a career in radiology to a college-age adult, which is down from 65% in 1990 but up from 42% in 1988. Forty-one percent said they liked working in radiology less than they did 5 years ago, whereas 22% said they liked it more. Radiologists' perceptions of managed care's effect on their practice had more influence on professional satisfaction than did its administrative intrusions or the actual percentage of managed care patients in their practice. Other factors associated with decreased satisfaction were increased administrative duties and government involvement.
CONCLUSION: Satisfaction in the radiology profession declined during a period of dramatic change and uncertainty. Fear about managed care rather than its actual effect was the dominant factor in the decrease. Therefore, it is unsound to predict a long-term decline in professional satisfaction.
Index terms: Radiology and radiologists • Radiology and radiologists, socioeconomic issues
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In response to these and other concerns, the American College of Radiology (ACR) periodically conducts surveys of the physician workforce in radiology. This study of professional satisfaction used data from ACR's 1995 Survey of Radiologists to build upon prior studies of satisfaction in the radiology profession (4,5). This is the fourth study from the 1995 Survey of Radiologists. The first study was an overview (6), the second investigated women in radiology (7), and the third studied characteristics of group practices (8).
After a brief introduction to the methods used in the present study, we explore the current professional satisfaction of radiologists and what they report as their level of satisfaction compared with that 5 years ago. This exploration will focus on explaining variations in professional satisfaction both within the 1995 Survey of Radiologists and between this survey and an earlier 1990 ACR study of radiologists (5). By using the 1990 study results as the baseline, we will show that there has been a decline in professional satisfaction and present evidence supporting our conclusion that this decline is directly associated with perceptions of the evolving managed care environment.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
We used weighting to account for differential sampling and the response rates of various groups within the sample. The weighted data indicate what the responses would be if all the radiologists had been surveyed and had responded. The total weighted number for the survey was 33,124; this included diagnostic radiologists, radiation oncologists, and relevant nuclear medicine specialists in the United States, as well as trainees and retired radiologists. The following example illustrates how the weighting works: On the basis of the American Medical Association Masterfile and various ACR data, we estimated that there were approximately 3,090 professionally active, posttraining radiation oncologists in the United States in 1995; we received surveys from 233 of these radiation oncologists. Therefore, a weight was calculated by dividing the number of active radiation oncologists by the number who responded to the survey (3,090 ÷ 233 = 13.26). For tests of statistical significance, the weighted number was adjusted for actual sample size so as not to inflate the probability that a finding was statistically significant.
Although the definitions of most variables and response categories are evident in the tables and text, a few variables merit explanation. For the variable we call "specialty," radiologists were asked to select the one field in radiology that best describes what they personally do in clinical practice. Those who selected radiation oncology were classified as radiation oncologists, those who selected general diagnostic radiology or general radiology (both diagnostic and therapeutic) were classified as diagnostic radiology generalists, and those who selected any diagnostic radiology subspecialty were categorized as diagnostic radiology subspecialists. For hours worked per week, the questionnaire asked about the "last normal, full work week—that is, a week without holidays, illness, professional society meetings, or other nonstandard events." The instructions were to include clinical practice, administrative duties, required hospital activities, research, teaching, etc, but not professional society meetings, continuing medical education, or on-call time.
For most of the information presented, we were interested in professionally active (not retired), posttraining (not residents or fellows) radiologists. Many of the questions about professional and practice characteristics were not asked of or relevant to residents, fellows, or retired radiologists.
Data from ACR's 1990 Survey of Radiologists also were used in this study. The 1990 survey, a stratified random sample of 2,804 diagnostic radiologists, radiation oncologists, and nuclear medicine specialists, had a 69% response rate. Many of the questions used in the 1990 survey are identical in wording to those asked in the 1995 survey. This allowed us to directly compare the results based on the 1990 data and published in an earlier study of satisfaction in the radiology profession (5) with the responses to the 1995 Survey of Radiologists.
Variables
The two variables of primary interest in this study are current satisfaction with the profession and satisfaction with working in the radiology profession relative to that 5 years ago. The current satisfaction variable was created from two questions. In one question, radiologists were asked to indicate if they would recommend a career in medicine to a college-age adult. If they would recommend a career in medicine, then in a second question they were asked if they would recommend specializing in radiology. A dichotomous satisfaction variable was created as follows: A "yes" response on both questions indicated current satisfaction with the profession, and a "no" response to either question was used as an indicator of not being satisfied.
Another survey question measured current satisfaction from the perspective of work enjoyment and was largely discussed in a previous article (6). This question appears to have the face validity necessary for a measure of current job satisfaction; however, on a five-point scale, almost all respondents selected the top two categories for enjoying their work. If we had used this five-point scale or even dichotomized the scale into "enjoy very much" or "enjoy somewhat," only one end of the continuum, satisfaction, would have been measured. The other end of this continuum, dissatisfaction, would have been missing from our analysis. Since the five-point scale was a poor measure for discriminating those who were not satisfied with the profession from those who were, we used the composite dichotomous measure of satisfaction, which also has the benefit of being directly comparable with measures in earlier studies of satisfaction (4,5) and is better able to address the broader issue of satisfaction with the profession rather than with one's job.
To produce a relative measure of professional satisfaction over time, radiologists with a minimum of 5 years experience were asked to indicate whether they enjoy radiology "much more," "somewhat more," "about the same," "somewhat less," or "much less" than they did 5 years ago. To supplement this five-point scale, respondents were asked to give reasons for their answers. Up to three reasons from each radiologist were recorded and categorized according to six common themes: work environment, personal factors, financial factors, government regulation, managed care, and legal issues.
The causes of variations in current satisfaction and the reported satisfaction relative to that 5 years ago were explored with variables encompassing the individual, practice, and managed care characteristics of the radiologist. Individual characteristics included age, sex, American Board of Radiology or American Board of Nuclear Medicine certification (yes/no), specialty, hours worked per week, experience with workplace discrimination (age-, sex-, race-based), and experience with sexual harassment. Practice characteristics included position within the organization (owner or employee), academic character, group size, time spent on practice administration duties, city size, and census region of the country. The effects of managed care were evaluated by using the reported percentage of managed care patients in the radiologist's practice; the perceived present and future effects of managed care on the radiologist's practice; and experience with managed care's use of discounted fees, practice guidelines, intrusion into quality of care, capitation payments, and preauthorizations for expensive procedures.
Statistical Techniques
In reporting bivariate results, we used a P value of less than or equal to .01 as the measure of statistical significance. This more conservative P value was used to adjust for the problem of multiple comparisons. Although it is common to conduct significance tests and report unadjusted P values for multiple comparisons (9), it is important to remember that the number of significance tests increases the probability that a statistically significant difference will be found by chance alone. One can adjust for this by dividing the P value by the number of set-wise comparisons. Since this may be an issue, we provide the exact P values, which can be adjusted as appropriate.
In analyses of the bivariate effect of each explanatory variable on current satisfaction, the 
2 statistic was used to test for statistical significance. The ordinal nature of the five-point scale for reported satisfaction relative to that 5 years ago necessitated the use of the Cochran-Mantel-Haenszel statistic to determine statistical significance. We tabulated the answers to this question so that a negative value reflected a decline in satisfaction from that 5 years ago, a positive value reflected increased satisfaction, and a zero reflected no change.
The 2 statistic was used to determine if the characteristics of radiologists who responded in the 1990 and 1995 surveys differed. Since estimates of statistical significance derived from the 2 technique can be inflated by the large sample size that resulted from combining both samples, only those differences that were statistically significant and exhibited at least a minimal level of association (.10 as measured by using Cramer's V) were reported. This measure of association is not affected by sample size. For statistical analysis of the difference between the 1990 and 1995 satisfaction levels, a difference of proportions test was used for current satisfaction and a difference of means test was used for the reported satisfaction relative to that 5 years ago.
In addition to bivariate analyses, multivariate models were used to study both the current satisfaction and the reported satisfaction relative to that 5 years ago. In reporting results for the multivariate models, we used a P value of less than or equal to .05 as the measure of statistical significance. The dichotomous nature of the current satisfaction measure (satisfied or not satisfied) required the application of logistic regression analysis to estimate the simultaneous effect of multiple explanatory variables on the probability that a radiologist will be satisfied with his or her profession. In contrast, the measure of reported satisfaction relative to that 5 years ago produced a scale more amenable to using ordinary least squares regression techniques. Ordinary least squares regression analysis is appropriate for modeling ordinal dependent variables with continuous independent variables (10); however, there is the potential for the assumption of equal distance between response categories to be violated. To evaluate whether our results were sensitive to the type of modeling used, we also conducted a polynomial logistic regression analysis (each response category is compared against all others). The results of the significance tests were consistent with those of the ordinary least squares model.
Perceptions concerning the current effect of managed care on the radiologist's practice were correlated with the percentage of managed care patients in the practice (r = 0.42). This required the use of an interaction term in the regression models to ensure that the evidence of a relationship between perceived effect and satisfaction was independent of the effects of managed care patient loads.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Multivariate analyses.—Bivariate comparisons cannot take into account the simultaneous effects that multiple variables may have on current satisfaction and reported satisfaction relative to that 5 years ago. For example, the association between managed care and satisfaction may disappear or differ from our initial conclusions after we control for variations in other factors such as individual and practice characteristics.
To better clarify which variables have an independent effect on satisfaction when the effect of other factors are considered, two multivariate models were developed (Table 2), one for current satisfaction and the other for reported satisfaction relative to that 5 years ago. For current satisfaction, five variables were statistically significant. With other individual, practice, and managed care variables held constant, being board certified and working in an academic practice had a positive association with current professional satisfaction. In contrast, experience with workplace discrimination, working excessive hours (46 or more hours per week), and negative perceptions of the future effect of managed care had a negative association with current satisfaction.
|
As expected, when individual, practice, and managed care variables were held constant, having obtained a first permanent posttraining position in the past 5 years had a statistically significant and positive association with reported satisfaction compared with that 5 years ago. In addition, practice owners reported less satisfaction relative to that 5 years ago than did employees. Somewhat similar to the results found in the analysis of current professional satisfaction, radiologists' perceptions of the present and future effects of managed care on their practices had a negative association with reported satisfaction compared with that 5 years ago.
Explanations for satisfaction relative to that 5 years ago.—To further study the factors affecting reported satisfaction relative to that 5 years ago, the respondents' written explanations were analyzed. (Recall that survey participants were given the opportunity to provide open-ended rationales for their ratings of reported satisfaction relative to that 5 years ago.) The explanations were grouped into six general factors: work environment, personal, financial, government, managed care, and legal. Specific explanations within each general factor are displayed in Table 3 according to the relevant satisfaction rating.
|
Of the reasons cited to explain why radiologists were either somewhat or much more satisfied than they were 5 years ago, an improved work environment was cited in 53% of cases. This primarily involved a general change in work environment, moving to a different group or practice, and/or reduced administrative duties. Another 38% cited personal reasons, which were attributable to their enjoyment of using new technology but also to their feelings of increased competency and intellectual and personal rewards. Financial reasons accounted for 5% of all the reasons for improved satisfaction. Managed care issues were not mentioned as a reason for increased satisfaction.
Of the reasons cited to explain why radiologists were either somewhat or much less satisfied than they were 5 years ago, at least one work environment issue was cited in 37% of cases. As an example, 15% of the reasons cited for being either somewhat or much less satisfied involved increased administrative duties. Another 5% of the explanations involved control by other physicians and increased bureaucracy. The second largest category of explanations for ratings of lower satisfaction (30%) was related to managed care issues. This primarily involved interference by managed care, but it also included similar explanations, such as loss of control or independence from outsiders, the changing health environment, and fee control by insurance companies.
In addition to managed care and work environment issues, 14% of the explanations involved government regulation, control, and "red tape." Overall, for both managed care and government factors, there was a clear trend: As satisfaction relative to that 5 years ago decreased, the percentage of explanations involving problems with managed care and government regulation increased.
Trends in Satisfaction
The compatibility of the 1995 and 1990 ACR Surveys of Radiologists allowed us to make meaningful comparisons between professional satisfaction in 1990 and professional satisfaction in 1995. The questions related to current satisfaction and reported satisfaction relative to that 5 years ago were identical in both surveys. In addition, many of the explanatory variables available from the 1995 survey were also available in the 1990 data.
Before evaluating professional satisfaction, an analysis was conducted to illuminate the potential differences between the 1990 and 1995 population parameters that were related to the individual, practice, and managed care characteristics of the radiologists responding to the two surveys. Substantive differences were found between the years 1990 and 1995 in age, academic character, and percentage of managed care patients. Specifically, a higher percentage of the 1995 sample was under the age of 35 years (17% in 1995 vs 11% in 1990) and over the age of 55 years (30% in 1995 vs 23% in 1990) compared with those in the 1990 survey. The percentage of radiologists from academic practices was lower in 1995: About 20% of the 1995 respondents were in academic practices compared with 28% in 1990. There was an increase in the percentage of managed care patients, from a mean of 13% in 1990 to a mean of 31% in 1995.
The comparison between 1990 and 1995 estimates of current satisfaction is displayed in Table 4. Overall, the percentage of radiologists who were currently satisfied with the profession in 1995 was significantly lower (51%) than that in 1990 (65%). Although the data are not displayed in Table 4, it is worth noting that the relative satisfaction compared with that 5 years ago also had a similar but less dramatic decline; in 1990, 27% were somewhat or very much more satisfied than they were 5 years ago compared with 22% in 1995.
|
Multivariate analysis of current satisfaction revealed similar results. With simultaneous controlling for the explanatory variables that were available for both the 1990 and 1995 surveys, the satisfaction in the radiology profession remained significantly lower in 1995 than that reported in 1990. The significant difference between 1990 and 1995 remained when we controlled for the effect of the percentage of managed care patients. This indicates that other factors that decreased satisfaction but were not included in the model changed over the 5-year period. For example, questions pertaining to perceptions toward managed care were not included in the 1990 survey, and, as a result, had to be excluded from the multivariate analysis.
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Models of professional satisfaction risk oversimplifying a very complex concept by excluding other relevant variables such as family environment, pay, benefits, and relationships with co-workers. In addition, our measures of satisfaction were not based on a survey instrument validated for that sole purpose. The measures did, however, behave (vary) similarly to those used recently by other researchers (11). To compensate for these weaknesses, information from other sources, such as prior research findings and open-ended responses from the Survey of Radiologists, were used to bolster some of the conclusions. It is important to consider that in our interpretation of the open-ended responses, the reported percentages are "of those reasons cited" and cannot be considered a direct reflection of the radiology profession (ie, some did not answer, whereas those who provided more explanations had a greater influence on the percentages than did those who provided fewer explanations).
Perceptions, as compared with facts, can be very time sensitive and would likely differ, even if we surveyed today the same radiologists who participated in the 1995 study. This is a common problem in surveying attitudes, but it should not dilute the insight attitudes offered into the radiology profession, and specific to this study, the perceptions about the effects of managed care on radiology practices. Although the data provide a window into the satisfaction of radiologists during a period of dramatic change, more contemporary information on professional satisfaction is not known. The results of this study, however, provide a valuable reference point for future studies on professional satisfaction.
Our results that suggest a decrease in current satisfaction from 1990 may sound more dramatic than they are. With a similar but not identical measure of satisfaction from a 1988 study (4), it was found that only 42% of radiologists were satisfied with the profession then, a figure lower than the one we found for 1995. The transient quality of perceptions coupled with the observed cycles in attitudes toward the profession suggests that satisfaction may be a fairly volatile quantity that, although an important gauge of the health of the profession, is highly variable in response to recent perceptions and events.
Satisfaction, Perceptions, and Managed Care
As with other physicians, there are many factors associated with the professional satisfaction of radiologists (12–14). With our measures of professional satisfaction, we found that board certification was associated with increased current satisfaction, whereas heavy workloads and experience with workplace discrimination were associated with decreased current satisfaction with the profession. In turn, radiologists who owned their practices reported having less satisfaction relative to that 5 years ago than nonowners. Of the reasons cited to explain why radiologists were somewhat or much more satisfied than they were 5 years ago, 56% involved an improved work environment. This primarily involved moving to a different group or practice, reduced administrative duties, and/or a general change in work environment and is consistent with research on sources of satisfaction for physicians in general (13,15).
Consistent with the findings of another study (6) that used the same data but a slightly different analytical approach, our findings indicate that uncertainty has a greater influence on attitudes toward managed care than do actual experiences. Thus, one factor that consistently explained the variations in our measures of satisfaction was perceptions about managed care, and this had a more persistent effect than did the realities of managed care, such as the percentage of managed care patients. Most revealing were the results of the multivariate models. Contrary to conventional belief, managed care patient ratios did not have a significant association with satisfaction when we controlled for perceptions of managed care. In contrast, negative perceptions concerning the future effect of managed care had a direct and deteriorating association with the professional satisfaction of radiologists.
Placing our results in context with those of other studies, prior research on the relationship between satisfaction and percentage of managed care patients has either found low (6) or no correlation (16) or promoted the conventional view that satisfaction decreases as the percentage of managed care patients increases (17). Why have we not seen consistent results? Apart from the dominant role of perceptions of managed care, one possible explanation may be that the relationship between satisfaction and percentage of managed care patients is not uniformly linear. Our data show that there is a general but somewhat unstable decrease in satisfaction as the percentage of patients with managed care plans increases, but this decline reverses as the percentage increases beyond 60% (Table 1).
We tested this curvilinear relationship by excluding the variables that measure perceptions of managed care from our multivariate models. When the overpowering effects of perceptions were removed from the models, we found that the measure of managed care ratios behaved in the hypothesized curvilinear fashion. This evidence of a rebounding effect must be considered tentative, however, since it was based on a relatively small number of radiologists with high managed care patient populations. In addition, we have no evidence that the satisfaction of radiologists at the higher levels of managed care utilization will return to the premanaged care levels reported in the 1990 data.
Measuring the concept of satisfaction is inherently difficult, as is establishing a firm causal relationship between the professional satisfaction of radiologists and the changing managed care environment. However, on the basis of our two quantitative measures of satisfaction, the open-ended comments of respondents, and logical inference and experience, we believe that there is sufficient cumulative evidence of a causal link between managed care and satisfaction. More specifically, we conclude that uncertainty about the effects of managed care on the radiology profession had a significant and deteriorative influence on the professional satisfaction of radiologists during the middle 1990s. However, assuming that other important factors such as working environment and government policy will remain relatively stable, it is not unreasonable to hypothesize that the decline of professional satisfaction reported in this study may reverse as radiologists become more accustomed to a managed care environment.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, P.E.C.; study concepts and design, P.E.C.; definition of intellectual content, P.E.C., J.H.S.; literature research, P.E.C.; data acquisition, J.H.S.; data analysis, P.E.C.; statistical analysis, P.E.C.; manuscript preparation, P.E.C.; manuscript editing and review, P.E.C., J.H.S.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. M. Zafar, R. S. Lewis, and J. H. Sunshine Satisfaction of Radiologists in the United States: A Comparison between 2003 and 1995 Radiology, July 1, 2007; 244(1): 223 - 231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. S. Lewis, J. H. Sunshine,, and M. Bhargavan A portrait of breast imaging specialists and of the interpretation of mammography in the United States. Am. J. Roentgenol., November 1, 2006; 187(5): W456 - W468. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Merewitz and J. H. Sunshine A Portrait of Pediatric Radiologists in the United States Am. J. Roentgenol., January 1, 2006; 186(1): 12 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Sunshine, R. S. Lewis, and M. Bhargavan A Portrait of Interventional Radiologists in the United States Am. J. Roentgenol., November 1, 2005; 185(5): 1103 - 1112. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. H. Deitch and J. H. Sunshine The Relationship of Managed Care To Business, Professional, and Organizational Aspects of Radiology Practices Am. J. Roentgenol., January 1, 2004; 182(1): 29 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Sunshine, Y. S. Cypel, and B. Schepps Diagnostic Radiologists in 2000: Basic Characteristics, Practices, and Issues Related to the Radiologist Shortage Am. J. Roentgenol., February 1, 2002; 178(2): 291 - 301. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Sunshine, C. Simon, G. R. Bushee, and B. Schepps 1997 Graduates Speak: Initial Employment Experiences of Residency and Fellowship Graduates Am. J. Roentgenol., November 1, 2000; 175(5): 1225 - 1232. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. E. Crewson and J. H. Sunshine Diagnostic Radiologists' Subspecialization and Fields of Practice Am. J. Roentgenol., May 1, 2000; 174(5): 1203 - 1209. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
