Hellinger FJ. The Effect of Managed Care on QualityA Review of Recent Evidence. Arch Intern Med. 1998;158(8):833-841. doi:10.1001/archinte.158.8.833
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1998
This article reviews recent evidence about the relationship between managed care and quality. With one exception, the studies reviewed represent observation periods that extend through 1990 or a more recent year. The review has led to the conclusion that managed care has not decreased the overall effectiveness of care. However, evidence suggests that managed care may adversely affect the health of some vulnerable subpopulations. Evidence also suggests that enrollees in managed care plans are less satisfied with their care and have more problems accessing specialized services. In addition, younger, wealthier, and healthier persons were more satisfied with their health plans than older, poorer, and sicker persons, even after adjusting for the type of health plan.The findings of the studies reviewed do not provide definitive results about the effect of managed care on quality. Indeed, relatively few studies include data from the 1990s, and little is known about the newer types of health maintenance organizations that invest heavily in information systems and rely on financial incentives to alter practice patterns. Furthermore, managed care is not a uniform method that is applied identically by all health plans, and research studying the different dimensions of managed care also is needed.
In their 1994 review of managed care plan performance, Miller and Luft determined that: "The HMO [health maintenance organization] and indemnity plans provided enrollees with roughly comparable quality care, according to process or outcomes measures."1 (p1516) They also determined that HMO enrollees were less satisfied with their care than enrollees in indemnity plans.
Miller and Luft1 reviewed studies published before 1994. Yet, many important studies of the effect of managed care on quality have become available since that time.2- 9 Moreover, dramatic increases in the proportion of persons enrolled in managed care plans combined with striking changes in the structure and operation of the health care system necessitate that timely data be used to evaluate the relationship between managed care and quality.
This article reviews evidence about the effect of managed care on 3 dimensions of quality (effectiveness of care, satisfaction with care, and access to care). The inclusion of satisfaction and access as separate dimensions of quality acknowledges the importance of the patient's perspective to evaluations of quality.
The Institute of Medicine defines the quality of care as: "the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge."10 (p892) In recent years, the importance of the patient's perspective in the evaluation of quality has been recognized, and the inclusion by the Institute of Medicine of "desired health outcomes" in its definition of quality reflects this recognition.
During the past decade, the percentage of persons enrolled in managed care plans has increased steadily, while the rate of increase in health care spending has decreased steadily.11,12 The annual rate of increase in health plan premiums between 1987 and 1990 exceeded 17%, while it fell to approximately 2% between 1994 and 1996.13 The percentage of employees enrolled in managed care plans rose from 25% in 1987 to 75% in 1996.12- 14
Many people have inferred a causal relationship between the rapid diffusion of managed care and the abatement in health care cost increases. Labor Secretary Robert Reich recently concluded that: "The shift, with companies bringing expenses under control as workers move into managed care, is the most dramatic change in compensation costs in recent memory."15
Because no single metric of quality exists, evaluating the relationship between managed care and quality has been more difficult than evaluating the relationship between managed care and costs. Nevertheless, there is considerable apprehension about the consequences of the growth of managed care. In part, the apprehension results from the large number of cases recounted in print media and on television describing the plight of patients enrolled in HMOs who have been denied care and who have suffered serious consequences.
Legislation frequently is a corollary of public concern, and apprehension about the effect of managed care on quality has been manifested by a surge in state legislation. More than 1000 bills affecting managed care plans were introduced in state legislatures during the first 6 months of 1996.16
To date, 25 states have passed laws that permit qualified providers who are willing to accept a health plan's reimbursement rates to join that health plan's network (so-called any-willing-provider laws).17- 19 Eight of the laws cover most providers, and the remainder apply to pharmacists. Twenty-eight states have passed legislation that mandates a minimum length of a hospital stay for deliveries, and 21 states have passed laws that ensure direct access for patients to certain types of specialists (usually obstetrician-gynecologists) without receiving a referral from their primary care physicians.20 In addition, 16 states have passed broad-based "patient protection" laws that regulate managed care plans.
Concern about the quality of care also is manifested by the efforts of numerous organizations (eg, National Committee for Quality Assurance, Joint Commission on Accreditation of Healthcare Organizations, and the Foundation for Accountability) to evaluate the quality of managed care plans. The Health Plan Employer Data and Information Set of the National Committee for Quality Assurance is the most widely used tool to evaluate managed care plans.
An ideal study design to evaluate the effect of managed care on the quality of care is randomization of a large number of patients and physicians between managed care plans and traditional indemnity plans in several locations and collection of detailed information about their treatment. With this design, researchers could accurately compare, within several locations, the effectiveness of care for managed care plan enrollees with specific diseases and conditions, for which generally accepted standards exist, with the effectiveness of care for enrollees in traditional indemnity plans. Researchers also could accurately compare the effect of managed care on access to and satisfaction with care.
Yet, no such randomized experiment has been conducted. The RAND Corporation's Health Insurance Experiment (HIE) is, perhaps, the closest to the ideal study.21 But, the HIE is dated (it was conducted between 1976 and 1981). It included patients in only 1 geographic location (Seattle, Wash) and in only 1 HMO (the Group Health Cooperative of Puget Sound), and physicians were not randomized.
Studies that have compared the effectiveness of care in managed care plans with the effectiveness of care in traditional indemnity plans often compare the experiences of patients with the same disease. However, these studies may be biased because of patient and physician selection effects.
Research has demonstrated that persons who choose to enroll in an HMO differ systematically from those who choose not to enroll in an HMO. Persons who have formed strong relationships with their physicians are less likely to join a managed care plan.22 Thus, HMOs may experience favorable patient selection. Research also indicates that physicians who choose to practice in an HMO setting often have less aggressive styles of practice.23,24 Thus, HMOs may experience favorable physician selection. These potential biases suggest that existing studies may overstate the effectiveness of care received by managed care enrollees.
With one exception, the studies reviewed in this article represent observation periods that extend through 1990 or a more recent year. The use of timely information is important because managed care has spread swiftly in recent years, and dramatic changes have occurred in our health care system.
The exception is the HIE. The HIE, a randomized study of the effect of managed care on quality, remains one of the most comprehensive and widely cited research studies of the effect of managed care on the use of services and the cost and quality of care.
Each of the studies reviewed in this article, except the HIE, was published after 1993. Thus, the results were not discussed by Miller and Luft1 in their review of the performance of managed care plans. The present review is not exhaustive; studies of the effect of managed care plans on Medicaid recipients are not included. Only studies that compared health outcomes (eg, mortality, physiologic measures such as blood pressure, measures of functional status, and measures of mental health status) are included in the review of the effectiveness of care, whereas studies that compared the use of preventive services (eg, the proportion of children who received immunizations or the proportion of women who received mammograms) and studies that compared hospital admission (or readmission) rates are not included.
Finally, the studies reviewed include all studies in the annotated bibliography of key findings on the effects of managed care health plans since 1990 published by the United HealthCare Corporation, Minneapolis, Minn25 that satisfy the aforementioned criteria. Although this publication does not include all studies of the effect of managed care plans on quality (it only includes studies of the effect of managed care plans on general medical care, not the effect of managed behavioral health care plans), it does include summaries of more than 100 studies, including all studies of the effect of managed care on quality that have been published in peer-reviewed journals.
Because managed care plans often have separate divisions for commercial and Medicare business, and because the health insurance program for aged Americans is operated by a distinct entity (the Health Care Financing Administration), the effect on quality of managed care and traditional indemnity plans is reviewed separately for Medicare and non-Medicare populations.
Each of the effectiveness studies reviewed evaluated different aspects of effectiveness. In particular, the effectiveness of care was evaluated for patients with many different conditions (eg, hypertension, acute myocardial infarction, depressive disorder, arthritis, and breast cancer). Thus, it was difficult to compare findings about the effect of managed care for patients with similar conditions across studies. Instead, the findings about the effect of managed care on effectiveness were generated using information about the effect of managed care on a variety of conditions. The following section summarizes the findings from the effectiveness studies. The effectiveness studies reviewed include 2 follow-up analyses of participants in the Medical Outcomes Study (MOS) that have become available during the past 2 years.2,5 In addition, a short description of the results from the 2 most prominent comparisons of care received in managed care plans and in traditional indemnity plans (the MOS and the HIE) is provided.
This section reviews studies that compared the effectiveness of care between managed care and traditional indemnity plans (Table 1). The HIE, MOS, the study of patients with rheumatoid arthritis, and the study of breast cancer survivors excluded nonaged enrollees (individuals 65 years or older).
The largest and most widely cited study of the effect of HMOs is the RAND HIE.23,26 The HIE was conducted in many sites within 1 geographic location. However, the comparison between care provided by an HMO and traditional indemnity insurers was conducted in Seattle between 1976 and 1981. This component of the HIE included 1580 residents between 14 and 62 years of age who were randomly assigned to receive care from fee-for-service (FFS) physicians or from a group model HMO—the Group Health Cooperative of Puget Sound.
In their study of the HIE, Sloss et al27 concluded that health status and health practice measures were not affected by HMO enrollment. Data were collected on numerous measures of physiologic health status (eg, forced respiratory volume, shortness of breath scale, hay fever scale, blood pressure, cholesterol level, and grip strength) and health practices (eg, smoking, alcohol consumption, level of physical activity, and seat belt use) from questionnaires and physical examinations. Regression analysis was used to estimate the effect of HMO enrollment on these measures after controlling for enrollment values of the health measures, income, demographic characteristics of the enrollee, and length of participation in the HIE.
Ware et al28 studied data from the HIE and found that for nonpoor persons, HMO enrollment had no discernible effects on 7 general health indexes (ie, physical functioning, role functioning, days in bed, serious symptoms, mental health, social contacts, and health perceptions). However, the authors concluded that poor persons fared worse in the HMO than in the traditional indemnity plan. They stated that: "The low-income group who began the experiment with health problems appeared worse off at GHC [the Group Health Cooperative of Puget Sound] by comparison with both free and pay FFS."29(p1021)
Another widely cited study of the effect of managed care on effectiveness of care is the MOS.2,5,29 This study included patients treated by 362 physicians in Boston, Mass, Chicago, Ill, and Los Angeles, Calif. Information on 2235 patients with hypertension, type 2 diabetes mellitus, acute myocardial infarction, or congestive heart failure was collected for the years 1986 to 1990. Patients were sent a baseline health survey, and data on approximately 70% of the initial cohort was obtained from a 4-year follow-up survey.
Greenfield et al2 compared the health outcomes of patients with hypertension (eg, blood pressure, physical functioning, general health perception, and general energy level) or type 2 diabetes (eg, mean glycosylated hemoglobin level, percentage of patients with glycosylated hemoglobin levels of 0.12 or greater, and percentage of patients with foot ulcers and foot infections). A 2-year follow-up study of health outcomes, a 4-year follow-up study of health outcomes, and a 7-year follow-up of mortality were conducted for patients with type 2 diabetes or hypertension. Greenfield et al2 concluded that, of the patients with type 2 diabetes or hypertension, no differences in the health outcomes or mortality existed between enrollees in HMOs and enrollees in traditional indemnity plans.
In a follow-up study of patients with type 2 diabetes, hypertension, recent acute myocardial infarction, congestive heart failure, or depressive disorder, Ware et al concluded that: "In comparisons between HMO and FFS systems, physical and mental health outcomes did not differ for the average patient."5 (p1039) Physical health outcomes and mental health outcomes were evaluated using the Medical Outcomes Study 36-Item Short-Form Health Survey.
Although no effect of managed care was found for enrollees as a group, subgroups were significantly affected. For example, for those with low incomes, HMO enrollment was associated with greater deterioration in physical health status. Indeed, the greatest effect of HMO enrollment was found on initially ill enrollees living in poverty. The HMO enrollees in this subgroup were more than 6 times as likely to experience a deterioration in their physical health status than enrollees in traditional indemnity plans.5 (p1044)
In particular, for 1 in 3 persons enrolled in an HMO who were living in poverty and were in poor health at the beginning of the study, the physical health status deteriorated during a 4-year period. The corresponding figure for persons with similar economic and health status who were enrolled in a traditional indemnity plan was 1 in 20. However, these findings were based on relatively few persons.
A rigorous and timely study of the effect of managed care on quality was conducted by Yelin et al4 on 1025 patients in northern California with rheumatoid arthritis who were enrolled in two cycles (in 1982 to 1983 and in 1989).4 Enrollees were followed up through 1994, and the care received by 798 patients with rheumatoid arthritis who were enrolled in a traditional indemnity plan was compared with the care received by 227 patients enrolled in an HMO.
The primary source of data was an annual structured telephone interview. The chief finding was a lack of evidence that patients enrolled in HMOs experienced different health outcomes (ie, number of painful swollen joints, global assessment of pain, overall disease status, the Health Assessment Questionnaire score, and a standardized measure of physical function for people with rheumatoid arthritis) than patients enrolled in traditional indemnity plans. Yelin et al,4 however, did not study the effect of managed care on subgroups of enrollees.
Lee-Feldstein et al30 used survival data for a group of 5892 non-Hispanic white women from a population-based breast cancer registry in Orange County, California, from 1984 through 1990. The outcomes for non-Hispanic white women with breast cancer were worse in HMO hospitals than in small community hospitals or large teaching hospitals. In addition, the use of breast-conserving surgery was greater at large teaching hospitals than at small community hospitals or HMO hospitals.
A randomized study of 6476 Medicare beneficiaries enrolled in 75 HMOs from 44 market areas and 6381 Medicare beneficiaries enrolled in traditional indemnity plans from the same 44 market areas was conducted by Mathematica Policy Research Inc.31 Investigators studied medical records for 1990 for 412 HMO enrollees and for 401 enrollees in traditional indemnity plans who were hospitalized for the treatment of colon cancer. In addition, 408 medical records of HMO enrollees and 402 medical records of enrollees in traditional health plans, all of whom were hospitalized for the treatment of stroke, were studied.
The Mathematica study also compared large numbers of Medicare beneficiaries enrolled in HMOs or in traditional Medicare indemnity plans receiving outpatient care for urinary incontinence, chest pain, and joint pain. Brown et al31 found that outcomes for patients with colon cancer and stroke (ie, readmission rates and mortality rates) and outcomes for the conditions treated on an outpatient basis (ie, the likelihood of being symptom free at the time of interview) for Medicare beneficiaries with urinary incontinence, chest pain, or joint pain did not vary between HMOs and traditional indemnity plans.32
In another article based on the same sample, Clement et al3 identified 2243 Medicare beneficiaries enrolled in HMOs and 2009 Medicare beneficiaries enrolled in traditional indemnity plans, all of whom had joint pain.3 The data collected from structured telephone interviews revealed that Medicare beneficiaries with continued joint pain during the previous 12 months who were enrolled in an HMO experienced significantly less improvement than enrollees in the traditional indemnity plan. There was no difference in the likelihood of achieving complete alleviation of symptoms between HMO enrollees and enrollees in traditional indemnity plans. In addition, HMO enrollees were significantly less likely to be followed up or monitored closely than their counterparts enrolled in traditional indemnity plans.
In their study of data from the MOS on 2235 patients with type 2 diabetes, hypertension, recent acute myocardial infarction, congestive heart failure, or depressive disorder, Ware et al5 found that the 476 Medicare beneficiaries enrolled in HMOs who had 1 or more of the 5 conditions fared worse than the 346 Medicare beneficiaries with 1 or more of the 5 conditions who were enrolled in the traditional indemnity plan. In particular, 257 (54%) of the 476 Medicare beneficiaries enrolled in HMOs and 97 (28%) of the 346 Medicare beneficiaries not enrolled in an HMO experienced declines in physical health status.
In their 1994 review, Miller and Luft observed that: "In seven of eight observations from five studies, fewer HMO enrollees were as satisfied with, or evaluated as highly, perceived quality of care and patient-physician interactions than did FFS enrollees."1(p1516)
The studies reviewed by Miller and Luft were published before 1994 and represent the experience of persons living in a few geographic areas or who were Medicare beneficiaries. None of the 5 studies randomly selected nonaged participants, and none that included nonaged enrollees was national in scope. More important, none compared nonaged enrollees in managed care plans and traditional indemnity plans after 1990.
Results from several large studies of randomly selected nonaged and aged enrollees in managed care and traditional indemnity plans have become available since 1993. In general, these studies support the findings of Miller and Luft about enrollee evaluations of satisfaction with care. They also reveal that enrollees in managed care plans have greater difficulty accessing specialized services and have more unmet needs.
Brief descriptions of these large studies follow (Table 2). The first 5 studied nonaged enrollees and the final 3 studied Medicare beneficiaries.
This study of 3000 insured adults aged 18 to 64 years in Boston, Los Angeles, and Miami, Fla, who were interviewed by telephone in 1994 included 500 randomly selected adults enrolled in an HMO in each city and 500 randomly selected adults enrolled in an indemnity plan in each city.6 Information about costs, satisfaction, access, choice, and various demographic variables was collected during a 20-minute interview.
Enrollees in HMOs were not as satisfied as enrollees in indemnity plans. Of the HMO enrollees, 29% rated their plan as excellent; 38% of enrollees in indemnity plans rated their plan as excellent.
Access to specialty services was more restricted in HMOs than in the indemnity plans. Of the enrollees in HMOs, 26% rated their access to specialty services as excellent, while 45% of enrollees in indemnity plans rated their access to special services as excellent.
During the 3 years before the study, almost half of the respondents had changed plans (54% of HMO enrollees and 37% of enrollees in indemnity plans). In addition, a strong relationship was revealed between income and satisfaction that persisted even after adjusting for plan effects. Persons with a higher income were more satisfied with their plan than were persons with a lower income, whether they were enrolled in an HMO or a traditional indemnity plan.
A questionnaire was mailed by the National Research Corporation, Lincoln, Neb, to 250000 households representative of the United States between May and June 1996, and responses were received from 167000 households (66.8%).8 The responses indicated that non-Medicare enrollees in HMOs are about as satisfied with their health plan as enrollees in traditional indemnity plans. However, enrollees in preferred provider organizations (PPOs) were significantly less satisfied with their plan than were HMO enrollees or enrollees in traditional indemnity plans.
In fact, 16.5% of nonaged enrollees in HMOs and 16.5% of enrollees in traditional indemnity plans were completely satisfied with their plan; only 13.1% of nonaged PPO enrollees were completely satisfied. The percentage of nonaged enrollees in HMOs who were very satisfied with their plan (41.8%) was slightly higher than the corresponding percentages of enrollees in traditional indemnity plans (39.8%) and PPOs (39.7%).
Nonaged enrollees in traditional indemnity plans rated their access to medical services significantly better than nonaged enrollees in PPOs or HMOs. In particular, 32.7% of nonaged enrollees in traditional indemnity plans indicated that they had no problems obtaining care compared with 23.4% of nonaged enrollees in PPOs and 20.1% of nonaged enrollees in HMOs. Similarly, 30.8% of nonaged enrollees in traditional indemnity plans rated their access to specialty care as excellent compared with 21.9% of nonaged enrollees in PPOs and 19.6% in HMOs.
The Robert Wood Johnson Foundation (Princeton, NJ) National Access Survey, fielded in the spring and summer of 1994, was a follow-up to the 1993 National Health Interview Survey sponsored by the National Center for Health Statistics of Hyattsville, Md.7 The National Access Survey achieved a 75% response rate and is based on a sample of 3450 persons. Information was obtained from telephone interviews augmented by in-person interviews to ensure that vulnerable populations were adequately represented.
Findings from this survey revealed that HMO enrollees, PPO enrollees, and enrollees in traditional indemnity plans have reasonably good access to services and relatively few unmet needs. Yet, a slightly higher percentage of HMO enrollees (24%) than PPO enrollees (21%) or enrollees in traditional indemnity plans (20%) had to wait more than 3 days for an appointment, and a slightly higher percentage of HMO enrollees (4.8%) than PPO enrollees (3.7%) or enrollees in traditional indemnity plans (3.0%) reported unmet medical care needs.
To improve their corporate health benefit plan strategies, 3 large employers (Xerox Corporation [Stamford, Conn], GTE Corporation [Stamford], and Digitial Equipment Corporation [Maynard, Mass]) distributed a survey during the fall of 1993 that was completed by 24306 employees.32 Each of the employers had noted a sizable migration of employees from indemnity plans to managed care plans during the early 1990s and was strongly interested in obtaining information that could be used to compare health plans across a broad range of characteristics.
Results indicated that employees generally were more satisfied with managed care plans than with indemnity plans. In particular, employees enrolled in managed care plans were more satisfied with the paperwork requirements, financial arrangements, and plan-provided information than were employees enrolled in indemnity plans. Yet, employees enrolled in managed care plans were not as satisfied with the provision of care (eg, continuity of care and physician-related choice) as were the employees enrolled in indemnity plans.
Louis Harris and Associates, on behalf of Harvard University (Cambridge, Mass) and the Robert Wood Johnson Foundation, conducted a telephone survey of 873 persons in 1994 that inquired about views on health services and health plans.33 The survey oversampled persons who reported that they were in poor health, persons with disabilities, and persons who were hospitalized during the 12 months before the study.
Enrollees in managed care plans reported more problems with their health care than did enrollees in indemnity plans. For example, 22% of enrollees in managed care plans and 13% of enrollees in indemnity plans indicated that they had problems obtaining treatment that they or their physicians believed was necessary. Furthermore, 15% of enrollees in managed care plans and 6% of enrollees in indemnity plans believed that their physicians spent inadequate time treating them.
The 1990 study by Brown et al34 of the experience of 6400 randomly selected Medicare beneficiaries from 75 HMOs and 6400 randomly selected Medicare beneficiaries enrolled in a traditional indemnity plan found that HMO enrollees were less satisfied than were enrollees in the traditional indemnity plan. This was true for almost every dimension of care (eg, access, waiting times, and perceived thoroughness of the examination) except cost. In particular, 45% of the Medicare beneficiaries enrolled in HMOs and 53% enrolled in the traditional indemnity plan rated their overall quality of care as excellent. Similarly, 43% of the Medicare beneficiaries enrolled in HMOs and 61% enrolled in the traditional indemnity plan rated their ease of access to physician care as excellent.34
In addition, data from the Medicare Current Beneficiary Survey (MCBS) conducted during the fall of 1993 support this finding.35 The MCBS is a continuous personal interview survey designed to collect information about all care received by Medicare beneficiaries. Interviews are conducted every 4 months. The MCBS conducted during the fall of 1993 includes interview data from approximately 12000 Medicare beneficiaries, of whom approximately 900 were enrolled in an HMO. Data revealed that the percentage of beneficiaries enrolled in HMOs who were satisfied with their care was less for aged and disabled beneficiaries. Approximately 87.0% of aged beneficiaries (individuals under 65 years) enrolled in an HMO were satisfied with their care compared with 91.8% of aged beneficiaries enrolled in a traditional indemnity plan. For disabled beneficiaries, the corresponding percentages were 77.3% and 88.2%.
More information about the perspectives of Medicare beneficiaries on managed care plans was obtained from a telephone survey, conducted by Mathematica Policy Research, of 3080 beneficiaries who had been enrolled in an HMO during the 12 months ending February 29, 1996.9 The sample was randomly selected from the group health plan file maintained by the Health Care Financing Administration; the survey collected information about access, satisfaction, health status, and demographics.
Although the survey by Mathematica Policy Research did not collect information from Medicare beneficiaries enrolled in a traditional indemnity plan, responses about access to care were compared with responses from the MCBS conducted during the fall of 1994.9(pp95-96) The HMO enrollees were more than 3 times as likely to report problems accessing care than were their counterparts enrolled in a traditional indemnity plan (13% vs 4%). Moreover, older, sicker, and poorer enrollees had more problems accessing care whether they were enrolled in HMOs or a traditional indemnity plan.
Because the phrasing of the questions related to satisfaction varied considerably between the 2 surveys, satisfaction with care was not compared between beneficiaries enrolled in HMOs and beneficiaries enrolled in traditional indemnity plans. However, investigators determined that sicker, older, and poorer beneficiaries were less satisfied with care than were healthier, younger, and richer beneficiaries whether they were enrolled in HMOs or traditional indemnity plans.
To isolate the effect of managed care on the quality of care, the care received by enrollees in managed care plans must be compared with the care received by enrollees in indemnity plans (a cross-sectional study) or, for persons who change enrollment between managed care and indemnity plans, the care received must be compared for the 2 types of plans (a before-and-after study). All studies that compared care in managed care plans with care in indemnity plans have been cross-sectional studies (ie, they did not observe changes in care associated with a change in health plan enrollment).
Several reasons explain the lack of before-and-after studies. First, obtaining information about a sufficient sample of patients who have changed enrollment between managed care and indemnity plans is difficult. In any year, relatively few persons may make such a change. Second, tracking the care received by individuals through time is difficult, especially if they change health plans. To compare the effectiveness of care received by enrollees who change enrollment between managed care and indemnity plans (recently, most changes have been from an indemnity plan to a managed care plan), personal identifiers must be obtained from each plan. Yet, plans are hesitant to provide detailed data about the treatment of specific persons and to provide information that may be used to identify the persons (eg, name, date of birth, place of birth, and social security number) because of confidentiality issues and marketing concerns.
Overall, the evidence presented in this article does not support the premise that managed care has lowered the effectiveness of care. However, it does suggest that managed care may adversely affect some subgroups of enrollees. In particular, evidence suggests that sick enrollees who are poor or elderly fare worse in HMOs.
The best designed and most comprehensive studies of the effect of managed care on the effectiveness of care (the HIE and the MOS) revealed no significant differences between the overall health outcomes of HMO enrollees and enrollees in traditional indemnity plans. However, both studies revealed important differences for subgroups of enrollees.
The HIE found that low income HMO enrollees in poor health experienced more serious symptoms than did low income enrollees in poor health who used traditional indemnity plans.28 The MOS found that elderly or poor patients with 1 of 5 conditions who were enrolled in an HMO experienced greater declines in physical health status than did their counterparts enrolled in traditional indemnity plans.5 Furthermore, enrollees in the MOS with 1 of the 5 conditions who were elderly and poor experienced the steepest declines in physical health status.
Evidence from a large randomized study of Medicare beneficiaries revealed no significant differences in the quality of care (ie, mortality, readmission rates, and complication rates) received by hospitalized beneficiaries with colon cancer or stroke or in the quality of care (ie, likelihood of receiving medical attention, estimated likelihood of being symptom free at the time of interview) for beneficiaries with urinary incontinence, chest pain, or joint pain.32 However, evidence from the same study found that Medicare beneficiaries enrolled in HMOs who had continued joint pain experienced poorer outcomes than similar beneficiaries enrolled in a traditional health plan.3
The findings of almost all studies of satisfaction with care and all studies of access to care revealed that HMO enrollees were less satisfied and had greater difficulties accessing care than enrollees in traditional health plans. The exception was found in a survey with 167000 responding households in 1996 conducted by the National Research Corporation.8 Nonaged enrollees in HMOs were about as satisfied as nonaged enrollees in traditional indemnity plans.
In all studies of Medicare beneficiaries, HMO enrollees were less satisfied and had more problems accessing specialized services than did enrollees in traditional indemnity plans. Beneficiaries enrolled in HMOs who were sick, poor, or both, were the most dissatisfied and experienced the most problems accessing specialized service.
The findings reviewed do not provide definitive results about the effect of managed care on quality. Indeed, relatively few studies compare the effectiveness of care in managed care plans with that in traditional indemnity plans after 1990. Owing to the rapid pace of change in our health care system, the relevance of findings reported in earlier review articles is disputable.
Furthermore, few data are available on newer models of HMOs, such as individual practice associations, point-of-service plans, networks, and hybrid plans. The follow-up study of the MOS by Ware et al5 included only 98 persons not enrolled in group or staff model HMOs, and in the study by Yelin et al,4 only 31 HMO enrollees were not enrolled in the Northern California Kaiser Permanente HMO enrollees. Thus, few data are available about the effect on the effectiveness of care of newer types of managed care plans that invest heavily in information systems and rely on financial incentives to alter practice patterns.
Research studying the different dimensions of managed care also is needed.36 Managed care is not a uniform method that is applied identically by all health plans. Managed care plans diverge importantly in configurations, financial inducements, and the characteristics of participating health care providers and enrollees. Moreover, managed care plans use many means (eg, financial incentives for physicians, utilization review, educational activities, peer pressure, and clinical guidelines) to improve their performance.
Research on the relationship between the dimensions of managed care and quality requires detailed information about the characteristics of health plans, providers, and enrollees. Thus, claims data alone are inadequate. For each claim, researchers must have information about how each provider is compensated, the medical condition of the patient, and the structure of the relevant health plan.
Acquisition of this information and estimation of the relationship between the dimensions of managed care and quality are time consuming, costly, and demanding. The alternative is continued ignorance about the characteristics of managed care plans that can control use of health care services and maintain, or improve, the quality of care.
In particular, there is a need for research that studies the relationship between methods of reimbursing physicians and the quality of care. A recent review concluded that the financial incentives for physicians are a key reason that managed care plans have successfully reduced the use of health care services.37 Yet, there is no research about how financial incentives affect quality.
Managed care plans use fee-for-service, capitation, "withholds," bonuses, and a variety of other methods to reimburse physicians. Many managed care plans establish referral accounts to cover the cost of specialists, and, in some instances, primary care physicians receive a portion of the surplus in this account or are liable for some portion of the deficit. Physicians also may face financial incentives related to their equity interests. For example, Columbia/HCA (Nashville, Tenn) enables some physicians to invest in Columbia/HCA regional health care systems.
Knowledge about how different types of financial incentives affect cost and quality is a fundamental component in the development of an efficient managed care plan. In the absence of such knowledge, managed care plans likely will continue to use financial incentives to control costs without a full understanding of the effect of the incentives on the quality of care.
Accepted for publication July 31, 1997.
The views expressed in this article are those of the author and do not necessarily reflect the views of the Agency for Health Care Policy and Research.
Reprints: Fred J. Hellinger, PhD, Agency for Health Care Policy and Research, Suite 500, 2101 E Jefferson St, Rockville, MD 20852.