Physician perspectives on the structure and function of group practice HMOs.
There is growing controversy between physicians and HMOs regarding the organizational and financial constraints placed on use of clinical resources to treat patients enrolled in capitated plans.(1,2) Physicians in several metropolitan areas are protesting the controls placed on their practices by HMOs. In some communities, these physicians are developing countervailing bargaining organizations similar to unions.(3,4) This control is getting a great deal of coverage from the news media and is generating a backlash against HMOs.
Group practice HMOs are particularly vulnerable because they are tightly coupled organizationally with medical staffs. Although physician satisfaction and support are essential for the success of group practice HMOs, little is known about the dimensions of these organizations that enhance or diminish these attitudes. Studies have shown that physicians join group practices in part to escape administrative responsibilities. Yet those in prepaid group practices appear to want substantial involvement in the administrative aspects of the plans.(5-7) Studies also indicate that physicians in group practices favor peer review as a quality assurance measure, but little is known about how they view peer review mechanisms oriented toward the conservation of resources in prepaid practices.(8) The effects of physician payment methods are similarly confusing. Group practice physicians generally favor fee for service arrangements, yet studies of those providing services for HMOs show mixed preferences for capitation versus fee-for-service payment and there is evidence suggesting that an increasing number of physicians practicing in fee-for-service group practices are being paid on a salary or modified salary basis.(7,9)
This study was designed to explore these issues from the clinicians' perspectives in three large group practice HMOs located in two large metropolitan areas. They range in size from 194 to 312 physicians and from 198,000 to 216,000 enrollees. All are more than 10 years old. HMO A, the oldest (28 years), is a staff model with salaried physicians and a financial incentive program based on productivity. HMO B, the second oldest (16 years) is a staff model with salaried physicians but no incentive program. HMO C, the third and youngest (13 years) is a group practice model with mixed productivity/risk sharing physician payment. (Recently, the older staff plan and the group model have developed network arrangements. These plans are still dominated by the original staff and group structures, however.)
The study focused on the relative emphasis planed on patient satisfaction, quality of care and cost control; methods used to control physician resource utilization; the degree of centralization of decision making; organizational support/effectiveness; and physician satisfaction/support for the HMO.
Data were collected by means of questionnaires distributed to a stratified random sample of physicians in each HMO. The physicians were stratified into generalist and specialist categories, and a random sample was drawn from each category. (The overall sampling proportion was 51 percent. Specialists were somewhat oversampled--55 percent versus 49 percent of generalists.) Of the 393 physicians surveyed, 180 questionnaires were completed, for an overall return rate of 46 percent. Individual HMO return rates ranged from 53 percent in HMO A to 42 percent in HMO B.
Because we had to provide complete anonymity for the physicians in order to obtain their cooperation, we were not able to target follow-up letters to the nonrespondents. Nor were we able to compare the characteristics of those who responsed to the general population. This raises a question regarding possible biases in our data. However, we feel that any such bias would be systematic and in the negative direction across the HMOs, because those most dissatisfied would be most likely to respond to our questionnaire. Thus, since we are studying physicians' satisfaction, it can be argued that this relatively low response provides us the acid test of physician satisfaction in HMOs based on a sample that may be biased against HMOs.
Physician satisfaction was measured by six questions dealing with overall feelings of loyalty and responsibility, satisfaction with organizations and controls and satisfaction with levels of remuneration. These statements were developed from the organizational assessment literature(10) and were then modified to reflect the group practice HMO environment by discussions with five group practice HMO medical directors not included in the study.
Measures of organizational structure, climate, and effectiveness were developed through a focused group consensus building process with the five group practice HMO medical directors noted above. This resulted in 13 statements related to the relative emphasis placed on cost control, quality of care, and patient satisfaction by the HMO; 10 statements related to method of resource use control; and 5 statements dealing with organizational support (effectiveness). After a factor analysis of the responses to these statements by the physicians in the survey, three organizational emphasis variables and one resource control variable were deleted because of low factor loadings. Centralization of decision making was measured by asking the physicians in each HMO to distribute 100 points across four decision making groups according to the influence each group has (or had) on 16 different typical decisions. The decision making groups were individual clinicians, clinicians as a group (or specialty groups), medical director, and nonphysician management. Statements were developed to represent three categories of decisions: strategic, operating, and patient care.
The data were analyzed from two perspectives. First, differences among the HMOs in the way they relate organizationally to their physicians were analyzed using one-way analysis of variance. Second, relationships between these perceived organizational attributes and physiciann satisfaction were explored using an OLS multiple regression model.
HMO Organizational Analysis
Each physician was asked to indicate his or her perception of the HMO's relative emphasis on 10 cost/quality/patient satisfaction attributes (table 1, page 44). A four-point Likert scale was used, with 4 representing emphasis "to a very great extent" in day-to-day HMO operations.
Significant differences among the HMOs are evident on 7 of these 10 attributes. The most pronounced differences center on the overall emphasis on quality and the methods used to reduce costs. Physicians in HMO B perceive their organization to place somewhat less emphasis on quality as the number one issue in practice decisions than do physicians in the other HMOs. They also perceive more emphasis on shifting patients to nurses for care and on reducing lab and xray tests to reduce costs. Moreover, they feel that their HMO devotes less emphasis to reducing administrative expenses to control cost than do physicians in the other HMOs. The linkage between perceived cost control measures in HMO B and physicians' assessment of its commitment to quality is obvious. This HMO is clearly perceived to be achieving its savings at the physicians' expense, and they feel that quality is being affected.
HMOs A and C are quite similar in their approach to these issues, although it appears that they differ in their strategies to maintain quality. Physicians in HMO A perceive it to emphasize internal controls (monitoring), while those in HMO C perceive it to emphasize retention of high-quality physicians. These alternate approaches likely flow from the structure of these organizations. HMO A is a staff model with salaried physicians, while HMO C is a multispecialty group practice with about 40 percent of the patients enrolled in fee-for-service plans.
Although all the HMOs were related quite high in terms of emphasis on patient satisfaction, HMO A received significantly higher scores than its counterparts on this dimension.
Mechanisms Used to Control Resource Use
In the next analysis (table 2, page 45), physicians were asked to score nine statements related to mechanisms used by HMOs to influence physicians use of resources. Again, a four-point Likert scale was used to determine the degree to which each HMO used these mechanisms, with 4 representing "to a very great extent."
The differences expressed on these resource control dimensions largely center on HMO C. Those physicians perceive their HMO to be less oriented toward administrative directives to influence physician practice styles than physicians in other HMOs, and they perceive their HMO to be more oriented toward physician income incentives.
Perhaps the most notable finding is that, with one exception, none of the HMOs is perceived as being very aggressive in controlling resources through these mechanisms. The scores are consistently low for virtually all of the categories. Even seemingly innocuous "educational" programs are not perceived as being emphasized by these HMOs, although HMO C does so more frequently than HMOs A and B. It appears that these HMOs are not perceived by their physicians to be aggressively working to control their practice styles, at least not as measured by these variables.
Organizational effectiveness was analyzed in two ways. First, each of the five statements was included in the analysis, and then a composite effectiveness score was constructed and evaluated. Although the HMOs differ somewhat in terms of perceived organizational support/effectiveness, the scores in general are very high (table 3, left). Except for "adequacy of information," most of the scores are 3 or above on a four-point Likert scale (4 being the highest score). However, HMO A consistently received higher scores than the other HMOs on all but one of the five support/effectiveness issues included in this assessment. Those physicians clearly perceive their organization to be responsive and well run. Except for "adequacy of information" and clarity of goals, physicians in
HMO C similarly give their organization relatively high effectiveness scores. HMO B, on the other hand, is perceived as being less effective/supportive across the board compared to the other HMOs.
When the organizational support/effectiveness statements are combined into a common dimension through factor analysis, HMO A is perceived to be much more supportive than either HMO B or C. From the analysis of the individual issues in this factor, it appears that this results primarily from the perception that information is more adequate and goals and objectives are more clearly articulated in HMO A than in either HMO B or C, and it is perceived as being more adaptive, innovative, and up-to-date technologically.
Organizational Decision Making
The fourth and final analysis of the organizational characteristics of these HMOs centered on the decision making structure. As previously noted, the physicians were asked to distribute 100 points across each of 16 typical strategic, operating, and patient care decisions to indicate the relative influence of management versus clinicians on each item. A centralization score was calculated for each HMO and decision issue area by dividing the number of points allocated to management (nonphysicians management and medical director) in that area by the total number of points in that area. For example, in the area of strategic decisions, there are six questions and 600 total points to be allocated.
As indicated in table 4, above, strategic decisions appear to be the highly centralized in all of the HMOs studied, with HMO B having the highest proportion of influence in this area allocated to management. Operating decisions are somewhat less centralized, especially in HMO C. Patient care decisions are highly decentralized to clinicians in all three HMOs. Overall, HMO C has the most decentralized decision making structure.
In fact, operating decisions in HMO C are more decentralized than previously found in most fee-for-service group practices.(4) Clearly, physicians in this organization dominate patient care and operating decisions while the HMO concentrates on strategic decision areas. Whether this is by design or default is unknown. These physicians initiated this HMO as an offshoot of their multispecialty group practice, and it appears that they still control the decision making process.
It is interesting to note that the most tightly coupled HMO in terms of physician-organization relationships (HMO A, the older, centralized, salaried physician model) has a relatively decentralized decision making structure. Perhaps the integration of the physicians into the organization through salaries and the fact that they concentrate most of their physicians in four clinics allows them to comfortably decentralize decision making.
Finally, physicians were asked to respond to four statements related to their satisfaction with their HMO and their degree of responsibility for and loyalty to the organization. Table 5, page 48, shows the results. Although the HMOs differed in their centralization of decision making and methods used to control resource use, they all received relatively high satisfaction and loyalty scores from their physicians. All scores were above 2.5, and most were above 3 on a 4-point Likert scale.
HMO A shows particularly high levels of physician satisfaction and loyalty, while HMO B has significantly lower scores. HMO C received high scores for the practice and standards of care categories, but its physicians are clearly less satisfied with their income levels and with methods used to control resource use. This suggests that physicians in the two staff model HMOs are more satisfied and loyal than those in the group model. In part, this may reflect the mixed HMO and fee-for-service practice of this group of physicians.
To examine the relationships between organizational attributes and physician satisfaction, we created composite variables for each set of organizational attributes and for physician satisfaction. They represent perceived relative emphasis on cost containment, patient satisfaction and quality of care in the HMO; perceived relative emphasis on educational programs, physician income incentives, and professional peer review as the means to control resource utilization; organizational support/effectiveness; degree of centralization of operating decisions, patient care decisions, and strategic decisions; and physician satisfaction.
We hypothesized that physicians' satisfaction would be higher in HMOs that relieve them of administrative responsibilities, emphasize patient satisfaction and quality, provide good organizational support, and emphasize size educational programs and peer review to control resource utilization. Previous group practice research indicates that physicians join groups in part because they do not like the administrative aspects of medical practice. (11) Therefore, centralization of administrative decisions should be expected to increase satisfaction. On the other hand, because physicians are trained to protect their patient care and professional preogatives, they may be less satisfied if the organization engages in activities that attempt to administratively influence their practice styles or that emphasize cost considerations over patient care.(12)
Four organizational attributes are higly related to physician satisfaction when only these linear effects are considered. The most pronounced is the degree to which the organization is perceived to be effective and supportive. Although somewhat less dramatic, two other variables (the use of educational programs to control resource use and the degree to which the HMO is perceived to emphasize patient satisfaction) are also positively related to physician satisfaction. A perceived emphasis on cost control in general and even the use of peer pressure to control resource utilization is negatively related to satisfaction. It appears that any type of organized (or formalized) control on resource use--whether it is administrative or professional--leads to lower levels of satisfaction. Efforts to inform and educate clinicians about resource use, on the other hand, appear to be much more acceptable and lead to higher levels of satisfaction. The decision making structures of the HMOs do not appear to influence physician satisfaction. Nor does it appear that generalists differ from specialists in terms of satisfaction.
We further tested to see if the relations between physician satisfaction and management's influence on decision making were linear. This resulted in two changes. In the linear equation, satisfaction tended to increase as strategic decisions become more centralized. When squared variables were introduced, the results were reversed, indicating that this relation is curvilinear. Physicians indicate higher levels of satisfaction as strategic decisions become more centralized. but they become increasingly dissatisfied if centralization in that area increases beyond some critical point. This may result from the nature of these organizations. Except for very large practices, traditional fee-for-service group practices make few strategic decisions that bind all of their physicians. Each of the clinicians in those practices has had wide latitude for individual decision making regarding referral patterns, acceptance of additional patients, and choice of hospital. Conversely, in an HMO, strategic decisions effect all of the clinicians and often commit them to time schedules, the use of certain hospitals, and expanded work loads. Therefore, they understandably want to be involved in those decisions. As with free-for-service group practice physicians, however, lower level strategic decisions.
The second change centered on the "emphasis on cost control" variable. Emphasis on cost control was still found to be negatively related to satisfaction, but these relationships were not statistically significant. This indicates a weak curvilinear relationship but with the effects of the variables remaining in the same direction at low and high levels of perceived cost control. In other words, relatively low levels of cost control efforts tend to cause dissatisfaction. As those efforts intensify, at some level dissatisfaction increases exponentially. Consequently, at that level, each increment of cost control effort will generate levels of dissatisfaction that are very disruptive and not cost effective. More work should be devoted to identifying those breakeven points.
Although the HMOs differ in geographic location, size, age, and physician payment system, they appear to function in a relatively similar manner. Most of the differences center on HMO B, which appears to be more aggressive in managing physician practice styles. HMO B, for example, appears to place more cost control emphasis on the use of nurses to replace physicians and on the reduction of lab and x-ray use than do the other HMOs. Perhaps as a result, physicians in this HMO perceive quality to receive less emphasis and cost reduction in administrative areas to be less than do physicians in the other HMOs. These physicians also perceive less emphasis on educational programs as a means of controlling resource use in their HMO than do the other HMO physicians, although these differences are slight. None of the physicians rated their HMOs high on the use of educational or peer review methods to control resource utilization. Similarly, except for HMO C, the physicians do not perceive their income to be closely linked to resource use or efficiency measures. Because all of these HMOs have relatively low utilization rates of hospital days per 1,000 enrollees (HMO A=329 days, B=364 days, and C=330 days), they must be achieving these practice styles through mechanisms other than those identified in this study.
The decision making structures of these HMOs provide some insights into the practice style issue. While direct patient care decisions are highly decentralized to the clinician level in all three organizations, operating and strategic decisoons (many of which influence practice styles) are more centralized. It appears that, through this decision making strategy, the HMOs develop structures that conserve resources by controlling program coverage, enrollment, referral patterns, and access to technology without interfering with the day-to-day office practice of their physicians. Moreover, it appears that, up to a point, the physicians support this approach. Physician satisfaction and loyalty scores were quite high in all three organizations.
While physicians support this approach at certain levels of decision making, they become dissatisfied when strategic decisions are highly centralized. Clearly the most satisfied physicians in this study are those in the HMO that pays them on a salary basis and operates with a moderately decentralized decision making structure (HMO A). It appears that this HMO balances its patient care/resource utilization goals through the development of a conservative corporate (or practice) culture, starting with the recruitment of physicians who are willing to work for a salary. This HMO uses some economic incentives, but its physicians do not perceive their income to be linked to resource use in any direct way. Because this HMO has been very successful financially and has a hospital utilization rate of 329 days per 1,000 enrollees, it appears that it has successfully inculcated a conservative practice style culture in its physicians while maintaining high levels of satisfaction. Moreover, it appears that this culture enables it to operate successfully with a relatively decentralized, participative decision making structure.
1. "HMO Industry in Trouble" and "The National Medical Marketplace: The Reemergence of the Power of Physicians." Reese Report 2:13, Oct.-Nov. 1987.
2. "Physicians' Not-So-Civil War." Wall Street Journal, Feb. 9, 1988.
3. "Doctors' Dilemma: Unionizing." New York Times, July 13, 1987.
4. Kralewski, J., and others. "The Changing Climate of Medical Practice: The Physician Rebellion." New England Journal of Medicine 316(6): 339-42, Feb. 5, 1987.
5. Wolinsky, F. "Why Physicians Choose Different Types of Practice Settings." Health Services Research 17(4): 339-419, Winter 1982.
6. Mick, S., and others. "Physician Turnover in Eight New England Prepaid Group Practices: An Analysis." Medical Care 21(3): 323-337, March 1983.
7. Goldberg, J., and Martin, H. "Physician Attitudes Toward Provider Relations, Reimbursement, and Control in HMOs." Group Health Association of American Journal 10(1): 55-67, May 1989.
8. Mechanic, D. "The Organization of Medical Practice and Practice Orientations Among Physicians in Prepaid and Nonprepaid Primary Care Settings." Medical Care 13(3): 189-204, March 1975.
9. Cotter, P. "An Analysis of the Changing Patterns in Physician Employment Status." In Socioeconomic Characteristics of Medical Practice 1986, edited by Gonzales, M., and Emmons, D. Chicago, Ill.: American Medical Association, 1986.
10. Van de Ven, A., and Ferry, D. Measuring and Assessing Organizations. New York, N.Y.: John Wiley & Sons, 1980.
11. Goodman, L., and Swartwout, J. "Comparative Aspects of Medical Practice: Organizational Setting and Financial Arrangements in Four Delivery Systems." Medical Care 22(3): 255-67, March 1984.
12. Ottensmeyer, D., and Smith, H. "Patterns of Medical Practice in an Era of Change." Frontiers of Health Services Management 3(1): 3-39, Aug. 1986.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Managed Care; health maintenance organizations|
|Date:||May 1, 1992|
|Previous Article:||The limits of medical quality assurance systems.|
|Next Article:||A process for objective review of physician performance.|