Printer Friendly

How the elderly fare in HMOs: outcomes from the medicare competition demonstrations.

Because of concern about the effects of prepaid care on outcomes for elderly enrollees in health maintenance organizations (HMOs), a prospective study of access to ca, and functional outcomes was Performed, HMOs with Medicare risk contracts in January 1985 (N = 17) were selected from ten communities and were matched for comparison with ten similar communities where no Medicare HMOs were in operation. Random samples of HMO enrollees (N = 2,098) and fee-for-service (FFS) nonenrollees (N = 1,059) were assessed at baseline and at follow-up one year later (HMO = 1,873, FFS = 916) to observe access to care and functional outcomes. At baseline, nonenrollees had more bed days and poorer functional status than HMO enrollees. While fewer HMO enrollees experienced declines in functional status between baseline and follow-up (e.g., patient's ability to function declined in one or more activities of daily living: HMOs at 5.3 percent versus FFS at 8.5 percent, p < .01), after controlling for other factors with logistic regression, enrollment status was not significantly associated with functional decline. Self-rated health, history of hospitalization, age of 80 or older and baseline functional status were predictive of decline in function. After controlling for baseline differences, HMO disenrollees also experienced similar functional declines at follow-up compared to continuously enrolled beneficiaries. These findings suggest that Medicare beneficiaries who belong to HMOs experience comparable rates of functional decline to those experienced by beneficiaries in the FFS sector with similar initial levels of function and health status. Together with results showing no significant difference in medical visits according to various symptoms, we conclude that access and quality of care delivered by HMOs is comparable to that provided in FFS settings.

Since passage of the Tax Equity and Fiscal Responsibility Act (TEFRA) of 1982, health maintenance organizations (HMOs) have been encouraged to enroll Medicare recipients and provide medical care on a capitated basis. In response to this legislation, the Health Care Financing Administration (HCFA) initiated a demonstration program, entitled the Medicare Competition Demonstration, that included HMOs that were willing to accept the financial risk of providing comprehensive care to more than 250,000 Medicare enrollees. Since these early demonstrations, the program has been expanded and there are now over one million Medicare enrollees in nearly 100 HMOs across the country.

The function of an HMO is to reduce service use when possible; this leads to concerns about the quality of care that enrollees may receive and the effect this might have on their health status. Prior experience with HMOs and empirical research on their effects on the use and cost of medical services has been well researched for nonelderly enrollees. For ambulatory care, some studies suggest HMO enrollees have higher use rates while others report lower rates of ambulatory visits (Luft 1981). For the nonelderly, most of the reduced utilization is in lowered hospitalization rates. However, virtually all studies show that HMOs significantly reduce the number of days spent in the hospital. Since HMOs affect reductions in costs primarily by reducing hospital admissions (Gaus, Cooper, and Hirschman 1976; Manning, Leibowitz, Goldberg, et al. 1984), there is concern that enrollees may experience poor health outcomes. For example, if HMOs inappropriately reduce or delay use of inpatient services, they may acutely exacerbate chronic illnesses. This could lead to functional disability, especially among the elderly (Gillick 1987).

Although the incentive structure may lead to concerns that outcomes would be worse for enrollees, there are several reasons that HMOs might also be likely to prevent dramatic functional declines among elderly enrollees. First, through administrative arrangements HMOs may provide the organizational structure necessary for implementing newer techniques in geriatric care, such as geriatric assessment programs. HMOs frequently use primary care |gatekeepers' to manage the care of enrollees and to monitor utilization of medical services. This case-management approach has been strongly advocated to provide better coordinated care for the complex health care needs of older persons. Furthermore, HMOs often provide coverage for medical expenses frequently required by the elderly (e.g., eyeglasses, hearing aids) that are not routinely covered by Medicare. In addition HMOs may be more likely than fee-for-service (FFS) settings to introduce more careful discharge planning activities that prevent costly and untoward events, such as high readmission rates (Rubenstein and Kane 1985).

These administrative efforts in HMOs may smooth the way for elderly enrollees who need more effective management of multiple medical and social problems. Furthermore, the barriers created to avoid hospitalizations in HMOs may actually work in favor of the elderly. The major technique that HMOs use to reduce the utilization of health care resources is substituting less expensive ambulatory care and chronic care services for more expensive, resource-intensive inpatient care. Research has shown that HMO hospital utilization rates can be as low as one-half those of traditional Medicare beneficiaries (Weil 1976; Greenlick, Lamb, Carpenter, et al. 1983). Thus, it may be that some functional declines could be prevented among enrollees by avoiding the iatrogenic events that occur during hospital stays (Steel et al. 1981).

Estimates of the effect of HMOs on functional outcomes have varied. While some studies have found functional outcomes to be no different from those in FFS settings (Yelin, Shearn, and Epstein 1986; Lubeck, Brown, and Holman 1985), at least one study has reported greater functional declines in HMOs than in FFS care among lower socioeconomic groups (Ware, Rogers, Davies, et al. 1986). Nonetheless, although the incentives and organization of prepaid care are more likely to invite unwarranted reductions in services, there have been no outcome studies of the elderly who enroll in HMOs. Since the elderly experience functional declines at a much higher rate than the nonelderly, functional status is a particularly important concern for Medicare beneficiaries who enroll in HMOs.

This study examines the issue of whether or not Medicare beneficiaries enrolled in Medicare risk plans operated by HMOs under the Medicare Competition Demonstration had functional outcomes similar to a similar group of beneficiaries who did not enroll. This study is part of a larger evaluation of the Medicare Competition Demonstrations, funded by the Health Care Financing Administration from 1983 through 1987. To evaluate the Medicare Competition Demonstrations, the Health Care Financing Administration funded an evaluation of costs, service use, and quality of care provided for elderly persons enrolled in HMOs. The purpose of the study was to determine whether or not elderly enrollees in these demonstration HMOs had functional outcomes close to those of a similar group of nonenrollees.


The evaluation of access to care and outcomes for elderly persons enrolled in HMOs involved two waves of telephone surveys. Before discussing the process and content of these surveys, however, we describe the strategy used for drawing samples for the evaluation.


The process of sample selection was designed to obtain a sample from HMOs that was approximately twice as large as the FFS sample. Samples were stratified by HMO market areas and corresponding FFS market areas to ensure proportionate representation in the two groups. The selection of the HMO sample included representation of the different HMO model types: individual practice associations (IPAs) contract with different independent physicians who care for enrollees in their office practices and often care for FFS patients as well; staff and group HMOs employ physicians or medical group practices directly, and these settings often provide care exclusively to HMO patients. The evaluation of functional status outcomes included 17 HMO plans: 5 IPA-model HMOs, 3 group models, 6 staff models, and 3 plans that were a mixture of model types. The HMOs were located in ten communities.

A random sample of HMO enrollees was drawn from those who enrolled in the 17 plans between November 1984 and January 1985. Beneficiaries less than 67 years old at enrollment were excluded to permit the measurement of service utilization prior to the survey from Medicare data files. Sample members were restricted to those with both Part A and Part B Medicare coverage. Data on eligible beneficiaries for selecting the enrollee sample were obtained from HCFA's Office of Prepaid Health Care file. A random sample of 1,000 enrollees meeting eligibility requirements was drawn from each participating HMO. This larger sample was used for other analyses intended for the demonstration project. Since the objective was to complete 123 interviews with each HMO, a random sample of 200 enrollees from each of the 17 plans was drawn for the interviews on functional status and access to care.

Because of concern that the introduction of Medicare HMO enrollment might alter the care for all Medicare recipients in the area, the FFS sample was drawn from areas that had no approved Medicare HMO activity. Each of the ten communities where an HMO was selected for evaluation was matched to a similar community where no Medicare HMOs existed. The nonenrollee sample was then selected from the comparison geographical markets that were matched to the ten demonstration market areas based on average Medicare reimbursement, hospital days per beneficiary, demographic characteristics of the population, physician supply, hospital occupancy rate, average personal income, and geographic region. These comparison geographic areas were specifically selected because there was no risk-based contracting activity at the time of the baseline survey.

Since one of the eligibility requirements for the sample of enrollees is that age at enrollment be at least 67 years old, a similar method had to be used to draw a nonenrollee sample comparable in age. Thus, "pseudo-enrollment dates" were randomly assigned to members of the nonenrollee sample such that their distribution in each comparison site identical to the distribution of actual enrollment dates among enrollees in the corresponding HMO demonstration site. Nonenrollee samples were randomly drawn from beneficiary lists provided through the Medicare program so that each member was at least 67 years old as of the assigned pseudo-enrollment dates. As with the enrollee sample, a subsample from each market area was then drawn for interviews. The number of nonenrollees selected for interviews varied across the comparison geographic areas in direct proportion to the number of HMOs operational in the corresponding demonstration markets. Thus, the goal was to obtain 82 completed interviews in each of the comparison sites corresponding to demonstration sites with one HMO, 117 completed interviews in each of the sites corresponding to demonstration sites with two HMOs, and 165 completed interviews in the market area corresponding to the largest demonstration site. The sample size for the baseline and follow-up surveys was designed to detect differences between groups of five percentage points or more on a binary variable, with a power of .85, where one value has a mean of .15 (e.g., whether functional status declined between baseline and follow-up according to enrollment status), for two-tailed tests at the .05 significance level (Fleiss 1981).


To study the impact of the demonstration HMOs on functional status and other outcomes, the telephone surveys for baseline and follow-up included questions on the presence of specific symptoms, access to care for symptoms, functional status, disability, and a variety of other characteristics (e.g., demographics, socIoeconomic status, living arrangements). The presence of specific symptoms during a six-month recall period was evaluated for 10 symptoms at baseline and follow-up: abdominal pain, arthralgias, bleeding (from any site), diarrhea, exertional chest pain, exertional dyspnea, involuntary weight loss, loss of vision, persistent cough, and syncope. These symptoms were specifically chosen due to their high prevalence among the elderly. Further, it was felt that it was important to determine access to care for specific symptoms to control for differences in care-seeking behaviors that might be present between enrollees and nonenrollees. At follow-up, respondents were also surveyed to find whether or not the symptoms prompted contact with the medical provider and whether or not an outpatient visit resulted.

Functional status was assessed with both activities of daily living (ADL) and instrumental activities of daily living (IADL). These scales have been used in a variety of settings and function as sensitive measures of independence and autonomy among elderly persons. Both ADL and IADL instruments have been extensively tested for interrater reliability. With Guttman scalogram analysis, the coefficients of reproducibility have ranged from .948 to .976 (Katz, Vignos, Moskowitz, et al. 1968; Katz, Heiple, Downs, et al. 1967). The IADL scale is sensitive to independence outside of the home (shopping, handling finances, taking medications, preparing meals, transportation, using the telephone); the ADL scale assesses independence in the home (dressing, eating, toileting, grooming, bathing, transferring). Each activity was addressed in a binary fashion (1 = assistance required, 0 = no assistance required). The survey was designed so that the six questions regarding basic activities of daily living were measured only if the respondent had impairment with one or more of the IADLs. Disability was measured in several ways, both at baseline and follow-up: number of bed days in the past two weeks, number of days with limited activity in the past two weeks, and number of bed days in the past year. Respondents based their answers on recall (e.g., questions on prior health status). Interviews took an average of 25 to 35 minutes to complete.

The baseline surveys were conducted by telephone between March 1 and July 31, 1985, about four to nine months after enrollees entered their HMO. The timing of the baseline interview was a concern, since the effects of HMO care could be experienced immediately after enrollment. Nonetheless, it was felt that conducting the baseline interview within nine months following enrollment would yield reasonably accurate measures of pre-enrollment health status. Nonenrollees in the FFS comparison group were interviewed over the same time period to ensure comparability between the two groups. The follow-up survey was administered approximately 12 months after the administration of the baseline survey.

2,098 completed interviews were obtained from HMO enrollees at baseline, and 1,059 completed interviews from FFS nonenrollees (N = 3,157). Response rates for the baseline interviews varied across HMO plans, from 66.0 percent to 87.6 percent (mean = 76. 1, standard deviation = 6.0). For nonenrollees, baseline response rates varied across the comparison geographical areas from 60.9 percent to 89.2 percent (mean = 69. 1, standard deviation = 9.0). Descriptions of the demonstration cities and comparison market areas, number of HMOs in each city, and completion rates at follow-up for both HMO and FFS settings are listed in Table 1. Follow-up interviews were completed with 1,873 HMO enrollees, including 376 who had disenrolled since baseline, and 916 FFS nonenrollees (N = 2,789). The survey completion rate at follow-up, excluding deaths (N = 123), was 91.9 percent of the baseline cohort (92.4 percent for HMO, 91.1 percent for FFS). Reasons for failure to complete follow-up interviews included inability to locate the respondent (N = 54), physical or mental disability that prevented interviewing (N = 23), the physician died or retired, prompting change of medical care (N = 52), and refusal to participate (N = 94). There were no significant differences among reasons for failure to respond to the follow-up survey between HMO enrollees and FFS nonenrollees. The mean intervals between baseline and follow-up surveys were 372.6 days for HMO enrollees and 357.3 days for FFS patients, a difference that was statistically significant (p < .01), but was not felt to be clinically significant for the outcomes examined.



Analysis between groups included two-tailed student t-tests for continuous variables. For follow-up of functional status outcomes, enrollees and nonenrollees were compared on the proportion experiencing a decline in activity since baseline. A logistic regression model was also used to control for baseline differences. For the logit model, decline in at least one activity of daily living was used as a single binary (0-1) dependent variable. Maximum-likelihood estimates were computed and confidence intervals (95 percent) were constructed around each odds ratio (Haldane 1956; Kleinbaum, Kupper, and Morgenstern 1982). For the multivariate model, mean values were imputed for some demographic independent variables for missing data. Missing data were present for less than 1.5 percent of cases for all variables except income, where mean values were imputed for about 10 percent of the cases in the logit model.

The analysis of access to care and outcomes for those who disenrolled from HMOs was judged to be important, since barriers to care and functional declines may be reasons for disenrollment (Wrightson, Genuardi, and Stephens 1987). However, the inclusion of disenrollees in the analysis raises several issues that deserve special comment. The principal consideration concerns how disenrollees should be counted. Should disenrollees be included as enrollees, nonenrollees, or analyzed as a separate group? As mandated by Congress, the intervention of prepaid care was not randomized in this study. Furthermore, HMO enrollees were allowed to disenroll at any point in the demonstration program, for any reason. In fact, about half of the disenrollees in our sample left the HMO within three months after joining (Rossiter et al. 1989). Moreover, even in a randomized trial the decision about how to treat patients that drop out is unresolved (Sackett and Gent 1979). Since it involved a new system of care, the debut of prepaid care in the Medicare Demonstrations was more likely to precipitate an unusually high rate of turnover in provider choice among HMO enrollees. Thus, for the purpose of this evaluation, there was not a reasonably comparable group of individuals in the FFS setting who changed providers. For these reasons, disenrollees were first included in the analysis of the continuously enrolled, and then analyzed as a separate group.


A baseline comparison of personal characteristics of HMO enrollees and the FFS comparison group of nonenrollees is shown in Table 2. There were significant differences in the two groups in mean age, marital status, home ownership, employment, health problems that might require medical care, and hospitalizations during the previous year. None of these differences were large, however, and there were no significant differences in education, race, or gender.


As shown in Table 3, the comparison group of FFS nonenrollees had significantly poorer health status when compared to HMO enrollees at baseline. There were significant differences in both bed days (during previous two weeks and previous year) and restricted activity days. FFS nonenrollees were also consistently more likely to require assistance to complete both ADLs and IADLs compared to HMO enrollees. The prevalence of symptoms at baseline was remarkably similar between the two groups, and no significant differences were found.


In Table 4, follow-up comparisons of HMO enrollees and FFS nonenrollees are shown for symptom prevalence, care-seeking behavior, and access to care. Again, as with the baseline comparison, the prevalence of symptoms at follow-up was similar between the two groups. Further, the proportion of patients in both groups who sought

care for the symptoms they reported was also similar. Finally, among those with symptoms who sought care, the vast majority in both groups actually saw a medical provider. Only differences in the proportion of patients who saw providers for arthralgias (i.e., joint pain) were significant, with HMO enrollees being more likely to see medical providers. Overall, 97 percent of enrollees and 95 percent of nonenrollees sought and received care for a specific symptom. For the small proportion who sought care but were not seen by a provider, there were no significant differences between groups for reasons given (e.g., could not get an appointment, illness self-limited) for any of the reported symptoms.


In Table 5, HMO enrollees and FFS nonenrollees are compared for declines in ADLs and IADLs. There were significant differences in the proportions of enrollees and nonenrollees who experienced a decline in activities of daily living and an increase in bed days over the follow-up interval. Nonenrollees were more likely to report a decline in ADLs and more likely to describe increased restricted-activity days, increased annual bed days, and increased bed days over the previous two weeks. The differences in IADLs between the two groups were not as marked, although nonenrollees were significantly more likely to experience a decline in at least one IADL.


Since the observed difference between HMO and FFS beneficiaries in functional decline could be due to the baseline differences noted between the two groups, logistic regression analysis was performed. As shown in Table 6, after controlling for baseline differences between the groups for age, baseline functional status, self-reported health, and other variables that could reasonably affect functional status (e.g., history of hospitalization), the results indicated that nonenrollees were not significantly more likely to experience functional declines. Functional decline was more likely among those age 80 years or older, with a history of a previous hospital stay during the 12 months prior to enrollment, with poor or fair self-rated health status, and with at least one ADL requiring assistance at baseline. However, the presence of a health problem at baseline that might require medical care, marital status, gender, and annual income did not appear to affect functional decline at follow-up.


Evaluations of access to care and functional status at follow-up for disenrollees were also performed as separate analyses. Disenrollment from HMOs was high -- 18 percent overall -- though it varied greatly among plans. Reasons for voluntary disenrollment have previously been reported (Rossiter et al. 1989) and largely have to do with misunderstanding the HMO policy at enrollment or dissatisfaction with the plan. High disenrollment rates in one HMO were due to plan closing. Because there is no comparable group of disenrollees in the FFS sector, comparisons were made with those enrollees who continued in HMO plans. As shown in Table 7, there were no differences at follow-up between HMO disenrollees and enrollees for symptom prevalence. However, there were significant differences in care-seeking, with disenrollees much less likely to seek care. Among those who did seek care, disenrollees were seen with similar frequency. There were no significant differences in the number of disenrollees who experienced an increase in restricted activity days or bed days at follow-up. A trend for greater declines in at least one ADL among disenrollees was only marginally statistically signficant, though declines in at least one IADL were highly significant. Comparisons of the continuously enrolled members with disenrollees through logistic regression revealed that, after controlling for baseline differences, functional declines between the two groups were not statistically significant at follow-up.



In the past two decades there has been a rapid increase in the growth rate of the older population. Together with advances in medical technology, this expansion of the elderly community has greatly escalated costs in the Medicare program. Since this rise in health care expenditures is likely to continue, several federal programs have been enacted to contain utilization of medical resources. Prepaid care has been one of these initiatives.

This study used survey data to examine access to care and functional outcomes in selected samples of Medicare HMO enrollees and comparable FFS Medicare nonenrollees. Since the mean age of both groups was approximately 75 years old, it is not surprising that both enrollees and nonenrollees experienced substantial declines in functional status, a consequence of an aging cohort, even during a relatively short follow-up period. While HMO enrollees and FFS nonenrollees reported symptoms, sought care, and were seen in outpatient settings at consistently similar rates, there were significant differences in functional status. At baseline, FFS nonenrollees reported more disability days and were less likely than HMO enrollees to be able to perform various common activities without assistance. The findings also indicate that nonenrollees were more likely to experience declines in functional status and increases in bed days between baseline and follow-up when compared with enrollees. However, these differences in rates of functional decline between enrollees and nonenrollees were shown by logistic regression to be due to differences between the groups that were present at baseline, such as functional status and prior hospital use.

While HMOs have been shown to be effective in treating acute problems of the nonelderly (Luft 1981), the quality of managed care for the chronic illnesses that frequently affect the elderly has been challenged (Iglehart 1987). In addition to apprehensions regarding the conflict of interest involved in placing the provider at financial risk, there have been concerns regarding how elderly persons might react to methods that HMOs use to channel patients. In particular, there have been misgivings over mechanisms used by HMOs to control use of health care services and the effects that these techniques for limiting utilization might have on access to ambulatory care (Luft 1982; Bates and Brown 1988). Nonetheless, the results from this study do not indicate that HMO enrollees face more barriers to care compared to beneficiaries in the FFS sector. Enrollees and nonenrollees reported a similar prevalence of symptoms and care-seeking behaviors, and both groups were seen with high regularity by medical providers.

The data suggesting that enrollees were significantly healthier than nonenrollees at baseline warrant special consideration. There are at least two possible reasons for this finding. First, beneficiaries' decisions regarding enrollment could be a reflection of their health status levels. Beneficiaries with serious health problems are more likely to have strong ties to a particular physician and are less likely to be willing to switch to an HMO-affiliated physician than are beneficiaries with few health problems. For example, beneficiaries without a primary care physician are more likely to join an HMO (Brown, Langwell, Berman, et al. 1986). Since the absence of visits to a physician prior to enrollment is an indicator of fewer health problems, these beneficiaries are more likely to be healthier than the nonenrollees with primary care physicians. Second, the HMO plans themselves tend to target groups of beneficiaries for enrollment through special presentations to active senior citizens groups or by providing special benefits (e. g., preventive care) that appeal to the well elderly (Brown 1988).

While disenrollees reported less care-seeking behavior, they were equally as likely as continuously enrolled beneficiaries to see a provider when care was sought. Although disenrollees were more likely to experience functional declines (in one or more ADLs) at follow-up, these declines were of marginal statistical significance and mostly due to baseline differences. Thus, whereas there appeared to be no important functional consequences of disenrollment, the findings emphasize the dissimilarity of enrollees and disenrollees. Since about half of the disenrollment from HMOs is due to dissatisfaction with the Medicare HMO program (Rossiter et al. 1989), further monitoring of disenrollment patterns may be needed.

Due to several limitations in the design of this study, caution must be exercised in the interpretation of the results. First, due to a congressional mandate, this was not a randomized trial. Thus, the conclusions from this study may be biased if those who voluntarily enrolled in HMOs were substantially different from nonenrollees in ways that were not captured by the control variables in the multivariate model. Clearly, nonenrollees were less functional at baseline. However, for the evaluation of functional declines, cohorts were observed for changes in functional status between two time periods and baseline function was included in a logistic model. Moreover, even if randomization would have been done, the effects of HMOs on FFS medical practices in the same market areas make contamination an issue. To avoid the competitive effects of managed care on non-HMO practices, the comparison group of FFS nonenrollees was purposefully selected from communities without any Medicare HMO activity. In fact, because of an increasing market penetration of Medicare HMOs in most urban areas, these data may represent a unique opportunity for comparing communities with and without prepaid care options.

Another limitation of the study was the multiple comparisons performed. Because many variables were compared between the two reimbursement models, some of the findings may have been due to chance alone. Nonetheless, the findings consistently revealed several patterns, including baseline differences indicating selection of functionally healthier HMO enrollees. Further, a multivariate model was used to control for baseline differences at follow-up, and included variables conventionally regarded as important predictors of functional decline as well as differences between those who choose and those who do not choose HMO care. Since the results of the multivariate analysis of functional status at follow-up revealed no differences between groups, the statistical power of the study could also be questioned. However, since modest differences in functional status between enrollees and nonenrollees were detected (i.e., [greater than or equal to] 5 percent) with adequate power in the sample size at follow-up, these negative findings should be viewed with confidence. Finally, the results from this study are consistent with other studies of the quality of medical care in Medicare HMOs that has been shown to be no worse (Retchin and Brown 1990a, 1991), and in some areas better (Retchin and Brown 1990b), when compared to FFS care.

Since HMOs reduce health care expenditures mainly by avoiding the need for hospitalization, it is understandable that there are concerns about the elderly being susceptible to the harmful consequences of limitations on hospital services. This study of the effects of prepaid care on Medicare beneficiaries indicated that, after correcting for baseline differences between groups, declines in functional status appeared to be similar when compared with traditional FFS care. Federal efforts to shift care in the Medicare program to HMO environments appear to be successful in maintaining or improving the health status of beneficiaries who choose to enroll in prepaid plans.


The authors acknowledge the guidance of many individuals who helped in the design, data collection, and analysis phases for the national Medicare Competition Demonstration. Among these, particular gratitude is extended to Kathy Langwell, Tamara Faulknier, and James Hadley. The authors also gratefully appreciate the valuable assistance and support provided by Mrs. Anne Brooks and Mrs. Beverly DeShazo.


Bates, E. W., and B. S. Brown. "Geriatric Care Needs and HMO Technology: A Theoretical Analysis and Initial Findings from the National Medicare Competition Evaluation." Medical Care 26, no. 5 (May 1988): 488. Brown, R. Biased Selection in the Medicare Competition Demonstrations. Princeton, NJ: Mathematica Policy Research, Inc., 1988. Brown, R., K. Langwell, K. Berman, A. Ciemnecki, L. Nelson, A. Schrier, and A. Tucker. Enrollment and Disenrollment in Medicare Competition Demonstration Plans: A Descriptive Analysis. Princeton, NJ: Mathematica Policy Research, Inc., 1986. Fleiss, J. L. Statistical Methods for Rates and Proportions. New York: John Wiley & Sons, Inc., 1981. Gaus, C. R., B. S. Cooper, and C. G. Hirschman. "Contrasts in HMO and Fee-for-Service Performance." Social Security Bulletin 39, no. 5 (May 1976):3-14. Gillick, M. R. "The Impact of Health Maintenance Organizations on Geriatric Care." Annals of Internal Medicine 106, no. 1 (January 1987): 139-43. Greenlick, M. R., S. J. Lamb, T. M. Carpenter, T. S. Fischer, S. D. Marks, and W. J. Cooper. "Kaiser Permanente's Medicare Plus Project: A Successful Medicare Prospective Payment Demonstration." Health Care Financing Review 4, no. 4 (Summer 1983): 84-98. Haldane, J. B. S. "The Estimation and Significance of the Logarithm of a Ratio of Frequencies." Annals of Human Genetics 20, no. 4 (April 1956): 309-11. Iglehart, J. K. "Health Policy Report: Second Thoughts about Health Maintenance Organizations for Medicare Patients." New England Journal of Medicine 316, no. 23 (June 1987): 1487-92. Katz, S., K. G. Heiple, T. D. Downs, A. B. Ford, and C. P. Scott. "Long-Term Course of 147 Patients with Fracture of the Hip." Surgical, Gynecology and Obstetrics 124, no. 6 (June 1967): 1219-30. Katz, S., P. J. Vignos, R. W. Moskowitz, H. M. Thompson, and K. H. Svec. "Comprehensive Outpatient Care in Rheumatoid Arthritis, a Controlled Study." Journal of the American Medical Association 206, no. 6 (November 1968): 1249-54. Kleinbaum, D. G., L. L. Kupper, and H. Morgenstern. Epidemiologic Research: Principles and Quantitative Methods. Belmont, CA: Lifetime Learning Publications, 1982. Lubeck, D. P., B. W. Brown, and H. R. Holman. "Chronic Disease and Health System Performance: Care of Osteoarthritis across Three Health Services." Medical Care 23, no. 3 (March 1985): 266-77. Luft, H. S. Health Maintenance Organizations: Dimensions of Performance. New York: John Wiley & Sons, Inc., 1981. --. Health Maintenance Organizations and the Rationing of Medical Care." Milbank Memorial Fund Quarterly 60, no. 2 (Spring 1982): 268-306. Manning, W. G., A. Leibowitz, G. A. Goldberg, W. H. Rogers, and J. P. Newhouse. "A Controlled Trial of the Effect of a Prepaid Group Practice on Use of Services." New England Journal of Medicine 310, no. 23 (June 1984): 1505-10. Retchin, S. M., and B. Brown. "The Effects of Prepaid Care on the Management of Congestive Heart Failure." American Journal of Medicine 90, no. 6 (February 1991): 236-42. --. "Management of Colorectal Cancer in Medicare Health Maintenance Organizations." Journal of General Internal Medicine 5, no. 1 (January/ February 1990a): 110-14. --. Quality of Ambulatory Care in Medicare Health Maintenance Organizations." American Journal of Public Health 80, no. 4 (April 1990b): 411-15. Rossiter, L. F., K. Langwell, T. T. H. Wan, and M. Rivnyak. "Patient Satisfaction among Elderly Enrollees and Disenrollees in Medicare Health Maintenance Organizations." Journal of the American Medical Association 262, no. 1 (July 1989): 57-63. Rubenstein, L. Z., and k. L. Kane. "Geriatric Assessment Programs: Their Time Has Come." Journal of the American Geriatrics Society 33, no. 9 (September 1985): 646-47. Sackett, D. L., and M. Gent. "Controversy in Counting and Attributing Events in Clinical Trails." New England Journal of Medicine 301, no. 26 (December 1979): 1410-12. Steel, K., P. M. Gertman, C. Crescenzi, and J. Anderson. "Iatrogenic Illness on a General Medical Service at a University Hospital." New England Journal of Medicine 304, no. 11 (March 1981): 638-42. Ware, J. E., W. H. Rogers, A. R. Davies, G. A. Goldberg, R. H. Brook, E. B. Keeler, C. D. Sherbourne, P. Camp, and J. P. Newhouse. "Comparison of Health Outcomes at a Health Maintenance Organisation with Those of Fee-for-Service Care." Lancet 5, no. 1 (May 1986): 1017-22. Weil, P. "Comparative Costs to the Medicare Program of Seven Prepaid Group Practices and Controls." Milbank Memorial Fund Quarterly 54, no. 1 (Winter 1976): 339-65. Wrightson, C. W., J. Genuardi, and S. Stephens. "Demographic and Utilization Characteristics of HMO Disenrollees." Group Health Association Of America Journal 8, no. 1 (January 1988): 23-24. Yelin, E. H., M. A. Shearn, and W. V. Epstein. "Health Outcomes for a Chronic Disease in Prepaid Group Practice and Fee-for-Service Settings: The Case of Rheumatoid Arthritis." Medical Care 24, no. 3 (March 1986):236-47.

Supported by contract no. 500-83-0047 from the Health Care Financing Administration, Department of Health and Human Services.

Sheldon M. Retchin, M. D., M. S. P. H. is Associate Professor and Chairman, Division of Geriatric Medicine; Dolores Gurnick Clement, Dr. PH. is Assistant Professor of Health Economics, Department of Health Administration; Louis F. Rossiter, Ph.D, is Professor of Health Economics, Department of Health Administration; and Barbara Brown, Ph.D. is Assistant Professor, Department of Health Administration, all at Virginia Commonwealth University, Richmond, VA. Dr. Retchin is also with the Project HOPE Center for Health Affairs, Chevy Chase, MD; Randall Brown, Ph.D. is Senior Economist; and Lyle Nelson, Ph.D. is Senior Economist, both at Mathematica Policy Research, Princeton, NJ. Address correspondence and requests for reprints to Sheldon M. Retchin, M.D., M.S.P.H., Associate Professor, Division of Geriatric Medicine, Box 287 MCV Station, Richmond, Virginia 23298. This article, submitted to Health Services Research on April 6, 1991, was revised and accepted for publication on January 29, 1992.
COPYRIGHT 1992 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:health maintenance organizations
Author:Retchin, Sheldon M.; Clement, Dolores Gurnick; Rossiter, Louis F.; Brown, Barbara; Brown, Randall; N
Publication:Health Services Research
Date:Dec 1, 1992
Previous Article:Choosing quality of care measures based on the expected impact of improved care on health.
Next Article:The effect of the Illness Episode Approach on Medicare beneficiaries' health insurance decisions.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters