The Effects of Organizational Structure on Primary Care Outcomes under Managed Care

Organizational Size and Complexity and Patient Satisfaction

As organizations increase in size, work within them tends to become differentiated. Complex tasks are broken up into their constituent parts, with individual employees doing one specialized and simplified job [9-11]. Only managers have a complete view of the overall production process. However, managers typically do not actually participate in production. Their principal responsibility is one of coordination and control.

Unique problems arise in managing organizational systems that produce services such as health care rather than manufactured goods. The diffuse and indistinct nature of health care, and the difficulty in quantifying or qualifying it, can lead to a high level of ambiguity in defining job responsibilities for employees [12]. Rather than simply measuring the number and quality of goods produced per hour as is often done for manufactured items, health care managers must establish somewhat arbitrary measures of output and worker efficiency. These control mechanisms tend to impede the provision of good service to patients.

Forces such as these that develop within large organizations and interfere with internal organizational processes have been labeled transaction costs or “X-inefficiencies” [13, 14]. Newhouse [15] predicts that significant inefficiencies of this type will develop in large health care organizations. The process of providing care, which in the traditional physician's office involves only a few persons with shared responsibilities, becomes fragmented in larger organizations. Communication among and coordination of the various aspects of care become difficult. Aspects of care not involving the physician, such as registration, telephone advice, and ancillary diagnostic services, become dispersed throughout the organization. As physicians come to rely on these specialized organizational supports, they become removed from much of the actual care received by the patient and from the consequences of the interactions between the patient and many of the nonphysician employees.

These employees spend their work time providing service to patients. The quality of that interaction affects the emotional atmosphere of work and employee satisfaction [16]. Unfortunately, good service is often contrary to organizational efficiency as defined and enforced by management [17]. Emphasizing service to the patient often compromises organizational efficiency. Caught in the stress of these conflicting expectations, employees may respond by withdrawing from personal involvement with or commitment to the needs of the patient, appearing to the patient as callous or unconcerned [18].

The effect of this conflict for health care employees in large organizations is especially significant in light of the social effects of large groups on individual members of the group. When an individual needs assistance, the willingness of an observer to offer that assistance is inversely related to the size of the group of observers in contact with the person in need [19]. Translated into the context of a health care organization, the willingness of any one employee to offer assistance to a patient in need decreases as the number of employees with whom the patient has contact increases. In addition, when employees are asked to perform tasks in a large group, their individual efforts sizably decrease compared with their efforts when working alone [20].

Thus, in large health care organizations we would expect to see a tradeoff between service and efficiency in the provision of primary care, with decreased patient satisfaction a likely result. In the following section, I describe actual empirical studies comparing patient satisfaction in alternative organizational structures of health care systems. Many of these studies confirm this prediction.

Patient Satisfaction in Large Health Care Systems

The results of several studies examining patient satisfaction with outpatient care in managed care settings are summarized in Table 1. Certain types of large health maintenance organizations (HMOs) have been shown to provide significant cost savings through lower rates of hospitalization and less frequent use of expensive technologies. However, they do so at the price of patient satisfaction with outpatient care [21]. In a study comparing a prepaid group practice organizational model with a traditional fee-for-service model, Tessler and Mechanic [22] found that patient satisfaction with care was significantly lower in the larger prepaid organization. Overall satisfaction was more strongly associated with access to care than with the perceived technical quality of care [22]. Sorensen and Wersinger [23] compared satisfaction with care in a large centralized HMO and a matched sample of patients receiving traditional fee-for-service care. Patients receiving care from an HMO were more satisfied with waiting time in the physician's office but were less satisfied with appointment availability and the interpersonal aspects of care [23].

These and other early studies did not fully control for patient selection. The RAND Corporation addressed this problem in its Health Insurance Study [24], in which two randomly selected patient samples were compared. One patient group was enrolled in a large group-practice-type HMO. The other received similar insurance coverage in the fee-for-service sector, in which individual physician offices and small groups predominate. This study found that measures of primary care use were the same for the two groups, with hospital use substantially reduced for the HMO sample. Overall measures of clinical outcomes were generally the same. The two groups did not differ in patient perception of the technical quality of care. However, significant differences were observed in patient satisfaction with care. Patients receiving care from an HMO had lower overall satisfaction. The HMO group was less satisfied than the fee-for-service group with the waiting time before appointment availability, the convenience of care, access to physicians, interpersonal aspects of care, and provider continuity. Patients in the HMO group were more satisfied with office waiting time [25].

Although these studies show a relation among the organizational structure of care, the process of care, and patient satisfaction, they confound two variables: organizational size and method of physician payment. Physician groups in HMOs tend to be larger than those in the fee-for-service sector. In addition, patients who receive care from physicians who are prepaid tend to be less satisfied with their care than similar patients seeing the same physicians paid on a fee-for-service basis [26, 27]. The Medical Outcomes Study has shown that in determining patient satisfaction with care, the size of an outpatient medical care organization is a separate issue from the method of payment. In both prepaid and fee-for-service systems, patients receiving care from larger facilities are less satisfied with their care than those receiving care from smaller physician groups. Problems in patient satisfaction in the larger organizations involved structural issues such as access, waiting times, and time spent with the physician more than they involved characteristics of the physician–patient interaction [28, 29].

We should not associate all types of HMOs with organizational complexity and poor patient satisfaction. Many Independent Practice Association-model HMOs continue to use smaller physician groups that maintain fee-for-service as the primary method of payment. These smaller groups seem to avoid many of the problems common to large systems [28, 29].

Although large organizational systems inherently tend to decrease patient satisfaction with the care process, some steps can be taken to improve satisfaction. In the following section, I discuss elements of patient satisfaction that can be enhanced to offset the disadvantages of large organizations.

Elements of Patient Satisfaction

In the last 20 years, patient satisfaction has been studied extensively with various survey instruments. One such instrument developed by Ware and colleagues has been the most thoroughly validated and widely used [30-32]. It can be used in longer or shorter forms, and it was the instrument used in both the RAND Health Insurance Experiment and the Medical Outcomes Study.

Patient satisfaction with primary care encompasses access to care, continuity of care, and the interpersonal nature of care. The perceived technical quality of care seems to play a lesser role in affecting satisfaction. This point is often overlooked in the establishment of monitoring systems for primary care groups and deserves emphasis in considering the design of primary care organizations under managed care.

Access to care includes being able to make an appointment to see one's physician in a timely fashion, not having to wait a long time in the physician's office, and being able to speak to one's physician on the telephone. In two studies, each involving more than 1200 patients, access to care and provider continuity were most closely associated with patient satisfaction [33, 34].

Continuity of care should be seen as reflecting a spectrum of relationships rather than an either/or phenomenon. The greater the continuity, the greater the patient satisfaction [35]. When offered a choice, patients often choose continuity of care over the convenience of care. Careful attention to the need for continuity in the design of organizational systems of care can improve patient satisfaction [36].

A question that needs further exploration is the apparently circular cause-effect relation between continuity and satisfaction. The less frequently patients see their own physician, the less satisfied they are [37]. The less satisfied they are with care, the more they tend to miss appointments [38], change physicians [39], and leave managed care plans [35, 40]. Whether poor satisfaction or poor continuity comes first is not clear.

The nature of the personal interaction between patient and physician also strongly affects patient satisfaction [41]. Physicians' “humanistic” behavior and patient satisfaction are highly correlated [42]. This is true for pediatric patients as well as for adults [43]. The more time a physician spends discussing patients' specific questions and providing explanations and patient education, the more patients are satisfied [44, 45]. The nature of the verbal and nonverbal communication between physician and patient affects both patient satisfaction with care and physician satisfaction with work [46]. Physician care that is personal in nature and thus more satisfying to patients is less prevalent in larger organizations than in smaller practice settings [47].

In the relationship between service providers and service recipients, the behavior of each party affects the outcomes of the interaction for the other [48]. This is especially true for the interaction between health care providers and patients. “Both members of the physician–patient dyad have the same fundamental human need for connection which can be fulfilled in the clinical encounter” [49]. In two studies, patient satisfaction was strongly correlated with physician satisfaction [37, 38]. A study of nurses who provide contraceptive services in a series of clinics also showed a strong relation between nurses' job satisfaction and the satisfaction of their clients. The investigators of this study concluded that “organizations that appropriately structure professional roles and meet the needs of professional staff will promote a climate in which providers as a group deliver service to clients more effectively and in which clients will perceive providers as highly service-oriented” [50].

The close link between patient and physician satisfaction has received relatively little attention but has important implications for the design of primary care systems. In the following section, I summarize the current knowledge of physician satisfaction.

Elements of Physician Satisfaction

The satisfaction of physicians and other health care providers has not received nearly the attention that patient satisfaction has. Although several investigators have constructed survey instruments to measure physician satisfaction, there is no consensus on how to measure the factors that determine satisfaction.

In general, working in large organizations tends to decrease job satisfaction for all types of workers [51]. A study focusing on primary care physicians confirmed this generalization: Physicians in larger organizational settings were less satisfied than those in smaller, traditional practices [52].

In a review of the early literature on the job satisfaction of physicians working in large, organized settings, Lichtenstein [53] found that perceived autonomy was the organizational characteristic most closely related to physician satisfaction. Two more recent studies confirmed that professional autonomy is important in determining physician satisfaction with work [54, 55].

There are two types of autonomy in medical practice: bureaucratic and clinical [56]. Bureaucratic autonomy concerns the structure of work and includes scheduling the hours to be worked, call schedules, and the number or type of patients to be seen. Clinical autonomy involves control over actual decisions in patient care and includes ordering tests and selecting treatments. The important issue of the availability of specialists for consultation or referral spans both types of autonomy.

A theoretical debate in the sociological literature on professions has centered on whether physicians working in large organizations become “deskilled” or “proletarianized” by bureaucratic constraints on their autonomy [57-59]. One study suggested that even the highly bureaucratic Soviet system did not necessarily compromise physicians' clinical autonomy [60].

Engel [56] suggests that “there is an optimal level of bureaucratic organization with respect to professional autonomy within the client-professional relationship.” A moderately bureaucratic setting may actually provide more autonomy than either a nonbureaucratic setting or a highly bureaucratic one [61]. Baker and Cantor [62] confirmed this conclusion in a study of physicians working in managed care organizations. They found that “although practicing under managed care is associated with lower levels of perceived autonomy in patient selection and time allocation, it is associated with higher levels of perceived autonomy in use of hospital care, tests, and procedures … Practicing under managed care is not uniformly associated with lower levels of satisfaction” [62].

Individual physicians seem to seek practice settings that balance the type of autonomy they value most. Some are comfortable in small, independent settings, whereas many are comfortable in settings that provide some bureaucratic structure yet allow them to maintain basic clinical autonomy [63].

In addition to professional autonomy and the nature of physician–patient interactions, other factors that have been shown to affect physician satisfaction are relationships with other professionals, physician pay, perceived professional status, and work load [64-67]. As discussed above, continuity of care is also important in determining physician satisfaction.

As primary care increasingly comes under the jurisdiction of managed care organizations, physician satisfaction with work will not necessarily have to be sacrificed [68]. Careful attention to structural relationships affecting physician autonomy, the physician–patient interaction, continuity of care, and positive relationships with other providers can provide a work environment that is supportive of both physician and patient needs. In the next section I discuss how such a system might affect the cost of care.

Organizational Size and the Cost of Physician Services: The Question of Economies of Scale

If large organizations tend to be more complex, presenting problems in satisfaction for both patients and physicians, what are their cost advantages? Do larger organizations achieve cost savings through economies of scale that offset their disadvantages?

Economies of scale in the provision of outpatient medical care was the subject of research and debate in the 1960s and 1970s; little empirical work has recently been done. In 1964, Harris concluded that “group practice contributes greatly to rising productivity. This increase in productivity can be measured in the simple manner, namely economies of scale; that is, unit costs tend to decline with rising output of the firm up to a point” [69].

Although little empirical evidence supported this concept, many found it attractive nonetheless [70]. Others were skeptical. Bailey suggested that “there are a number of erroneous and possibly disastrous ideas circulating in this country concerning the supposed economies of scale to be found in the group practice of medicine” [71]. Although he agreed that economies of scale were provided in moving from solo to group practice through the better timing and sequencing of patient flow, he contended that these economies are rapidly exhausted, resulting in constant returns to scale. At some point in the growth of the organization, increasing complexity leads to inflexibility, inefficiency, and consequent diseconomies of scale [72].

This perspective is shared by Newhouse [73], who sees the relation between the size of a medical practice group and the costs of providing outpatient care as U-shaped: Initially costs decrease with size, but at some point they level off and then begin to increase as organizations increase in size. He explains this relation on the basis of inefficiencies in patient flow and manpower use that develop as groups become larger and more complex [73].

In discussing economies of scale, the issue addressed is the cost of providing a given amount of service, such as a physician office visit. Although the aggregate cost of care under different types of plans is important, it is not the same as economies of scale. Economies of scale are usually achieved through the efficient use of capital. Unfortunately, in complex health care organizations, any savings in the efficient use of capital are often offset by the added costs of developing new management systems and by “the surprising inability of … administrators to appreciate the complexity of the decision-making process within their own institutions” [74]. These potential problems inherent in large facilities led one hospital administrator to caution that “economies of scale are often difficult to verify, and belief that they should be present is not a rational a priori justification for (larger organizations)” [75].

The few existing data suggest that economies of scale affect small practice groups but lead to decreasing returns to scale as groups grow beyond a certain as-yet undetermined size. The relation among organizational size, the cost of physician care, and patient satisfaction is shown in Figure 1.

Investigators of a study of pediatric practices concluded that larger settings, especially multispecialty groups, compare unfavorably with smaller groups in physician productivity [76]. In a study of several hundred practice groups of different sizes and specialty mixes, small group practices had increasing returns to scale and large practice groups had decreasing returns to scale. The investigators hypothesized that the optimal group size is approximately seven physicians [77]. Similar conclusions were reached in studies of outpatient mental health services [78] and dental visits [79]. A pattern of initial economies of scale with subsequent diseconomies also exists in the design and implementation of primary health care systems in developing countries, with higher costs and more complexity coming with larger organizations and systems [80].

In the last 20 years, the forces of cost-containment and competition have been associated with a large-scale aggregation of health care institutions and physicians into more and larger practice groups, often on the assumption that “the more organized and systematized, the more cost-effective the grouping will become” [81]. Although these economic benefits may exist for some types of hospital and other specialized care, it is not at all clear that the same benefits are realized for primary care [82]. Few data currently indicate that large groupings of physicians achieve cost-economies in the provision of primary care and other outpatient services. This conclusion deserves added emphasis and takes on increased importance when combined with the adverse effects of increasing organizational size on patient and physician satisfaction with the process of primary care.

Summary and the Need for Future Research

In summary, the following can be concluded from available data:

1. Larger health care organizations tend to be more complex, more fragmented, and more impersonal.

2. Patients and physicians tend to be less satisfied with larger practice settings.

3. Factors that promote a high-quality primary care encounter include accessibility, continuity, and open physician–patient communication.

4. Autonomy is a principal determinant of physician satisfaction; autonomy need not be sacrificed in a managed care environment.

5. The relation between the size of a primary care group and the cost of providing care is probably U-shaped: Costs decrease as physicians move from solo practice into small groups, level off fairly rapidly, and at some point begin to increase as organizational size increases because of organizational complexity.

It is also possible to summarize what is not known:

1. The size at which a primary practice group achieves optimal cost-effectiveness in providing patient care.

2. The practice group size at which patient satisfaction begins to become a problem.

3. The size or structure of primary care groups that best promotes physician and other employee satisfaction.

From the standpoint of classic economic theory, these issues should not be of great concern. Proponents of managed competition base their proposals on classic concepts of market efficiency and predict that under a competitive market for health care, “the efficient will prosper and the inefficient must improve or fail” [83]. From this perspective, the problems of organizational complexity and inefficiency in primary care delivery will be solved by market forces.

However, markets have not always done a good job of weeding out inefficient organizations or systems. In his Nobel Prize-winning argument, North [84] points out that historically, market forces “are always a mixed bag of those that increase and those that decrease productivity” [84]. Even in a more competitive market for health care, a dynamic tradeoff will occur between efficiency and satisfaction (of both patients and providers) within large primary care organizations. Given the paucity of data on the cost of providing physician services under alternative organizational models, no firm basis seems to yet exist for identifying or implementing the optimal primary care system under managed care.

If, as the investigators of the Medical Outcomes Study conclude, we are shifting to a new paradigm for primary care that emphasizes providing care in a structured organizational environment, we must be sure that such a shift is “a purposeful one, delineated to ensure high standards of care and acceptability to patients and clinicians” [24]. If primary care is to be cost-effective, satisfying to patients and physicians, and still part of large managed care systems, we must find the answers to these and other questions as rapidly as possible through carefully designed research.