Medicaid Managed Care Bibliography and Resources / Health Care Reform Links

Medicaid Enrollees in HMOs: A Comparative Analysis of Perinatal Outcomes for Mothers and Newborns in a Large Chicago HMO

August 1990

Center for Health Administration Studies

The University of Chicago

Claire Kohrman, Ph.D.
James Hughes, M.A.
Ronald Andersen, Ph.D.

ABSTRACT

Attempting to control Medicaid costs, many states have experimented with prepaid, managed care programs for public aid recipients. Advocates of such programs argue that they provide higher quality care, while critics charge that prepaid managed care programs merely restrict access to care for the poor. But few studies have demonstrated any differences in outcomes among Medicaid dependents in fee-for-service and prepaid managed care. This study analyzed the perinatal outcomes of Medicaid recipients enrolled in a large Chicago HMO. These data were compared to perinatal outcomes of public aid recipients, and Chicago residents with demographic characteristics similar to the Medicaid HMO population. The groups were compared on types of delivery, incidence of low birth weight, and neonatal mortality. The differences among the groups were minimal and not statistically significant. But in the majority of comparisons the HMO recipients had relatively higher incidence of low birth weights, and lower incidence of morbidity and neonatal mortality, suggesting that, once the mothers were admitted to hospitals for delivery, HMO care was as good or better as fee-for-service. Most importantly, however, the findings suggest that in the context of urban poverty, and the complex health behaviors and problems associated with it, the difference between these two payment mechanisms has relatively little positive or negative effect on perinatal outcomes.


I. INTRODUCTION

The implementation of Medicaid and Medicare brought about dramatic improvements in health care access for the poor in the 1970s (Andersen and Aday, 1980). During the 1980s, however, the health care costs have risen dramatically. Efforts at cost containment have been more successful in Medicaid than in the private sector. From 1982 to 1987 Medicaid costs rose at 9% a year, less than the health inflation rate for the non-public aid population. Cost-containment in Medicaid was achieved largely through restricting Medicaid eligibility to less than half of those in need (Dallek, 1988; Newacheck, 1988). This restriction in access for the poor has been particularly tragic for the inner city women who have seen their rates of perinatal morbidity and mortality climb during the 80s (Rosenbaum, Hughes, and Johnson, 1988). The high infant mortality rate in inner city areas, where drug treatment, family planning, and prenatal care programs have long waiting lists, but billions are spent on expensive neo-natal intensive care units, has become a symbol of the U.S. health care crisis.

In an effort to control costs but still provide quality care, many states have begun to contract with HMOs and other forms of managed care to provide services to their public aid populations. In 1981 only 280,000 Medicaid recipients were enrolled in HMOs, but by December 1987 2.7 million Medicaid recipients were enrolled in capitated managed care plans (Hurley and Freund, 1988). Government agencies expected that capitated managed care plans, with their shared risk, would control costs through expanded preventive care efforts, as they were alleged to have done with employed populations. This was thought to be particularly likely with the accumulating evidence that money spent on prenatal care reduces later, far more expensive, neo-natal intensive care costs. Critics have charged however that HMO Medicaid contracting has worsened already faulty coordination of services to poor women, and resulted in poorer care and outcomes for the public aid population.

In Chicago, where one of the highest U.S. rates of infant mortality persists, the Illinois Department of Public Aid contracts with HMO for the care of about 10% of Medicaid recipients. It is in this context that, in 1988, the Center for Health Administration Studies at the University of Chicago worked with Chicago HMO (CHMO), the HMO with the largest public aid enrollment in Chicago, to begin to evaluate the care and outcomes for its public aid enrollees.

II. PERINATAL HEALTH INITIATIVES AND MEDICAID HMOs

The Reagan and Bush administrations have both encouraged states to experiment with Medicare and Medicaid contracts with HMOs and other managed care providers (HMO Manager's Letter, 1988). The marriage of HMOs with Medicare and Medicaid required changes in both federal regulation and in the conventional operation of HMOs. The Omnibus Budget Reconciliation Act (OBRA) of 1981 modified the long-standing program requirement that individual Medicaid recipients have the freedom to obtain services from any qualified provider of care, opening the door to state experiments in mandatory managed care for public aid recipients. Since 1981 more than half of the states have sought waivers from the freedom-of-choice clause from the Health Care Finance Administration (HCFA), and capitated and case-managed programs have expanded rapidly (see Tables One and Two). Some states have contracted with HMOs and made enrollment in HMOs mandatory; others have contracted with HMOs to provide a prepaid option as an alternative to fee-for-service (FFS) medical coverage from which the Medicaid recipients can choose.

Table One: Growth of Medicaid Managed Health Care Since 1981

	States	Plan Type	Plans	Enrollment

June 1981	18	Traditional HMOs	54	282,000

December 1987	31	Variety of plans (see below)	180+	2,691,000



Table Two: Varieties of "Medicaid HMOs," in December 1987

	States	Plan Type	Plans	Enrollment


	28	Traditional HMOs	142	947,000

	2	Capitated organizations without 	3+	1,129,000
		primary care gatekeeper

	4	Capitated organizations with 	5	145,000

		prmary care gatekeepers

	6	Partially capitated	30+	50,000

		primary care gatekeepers	

	8	Fee-for-service with	?	420,000

		primary care gatekeepers	


(Freund and Hurley, 1988)

HMOs have had a reputation for promoting prevention, as well as managing utilization, and Medicaid programs have predicted that using HMOs to control costs would have a positive effect on perinatal outcomes. Because of the strong associations of perinatal morbidity with poverty, race and education, black women on Medicaid have been the population at highest risk for adverse perinatal outcomes in the U.S. Perinatal health workers have focused on sustaining pregnancies to term and increasing the birth weights of infants by improving access to care, preventive health behaviors through prenatal care outreach. It has been expected that Medicaid HMOs would seek to control later costs of morbidity by investing in these prenatal interventions.

For instance, the National Bureau of Economic Research (Joyce et al., 1988) found that when a number of interventions (including prenatal care, teen-age family planning, supplemental food programs for mothers and children, access to abortion, community health centers and neo-natal intensive care) were compared, prenatal care was the most cost-effective expenditure in reducing infant mortality. Neonatal intensive care, on the other hand, though the most demonstrably effective method on the individual level, was the least cost-effective. Gorsky and Colby (1989) estimate, for instance, that for every dollar spent on prenatal care in New Hampshire, $2.57 could be saved. Moreover, black infants benefit even more per dollar spent on prenatal care than white infants (Murray and Bernfield, 1988).

HMOs were at first eager to enter this market of guaranteed enrollees, and confident of their ability to provide care at less than the FFS costs. However, these HMOs, which had developed their patterns of service and performance with mostly young employed populations, found it both necessary and costly to expand and tailor their services for this population with very different needs. HMOs had found their competitive edge whittled away as corporate payors enacted their own cost-containment efforts. Many HMOs have become financially unstable in recent years, and some have closed.

HMOs with public aid enrollments, in particular, have found themselves forced by their contracts with states to provide care at below FFS reimbursement levels while trying to remain profitable. Some HMOs are now pulling out of Medicaid and Medicare contracts. Maxicare in northern California, for example, cancelled its contract with the state of California for public aid recipients stating explicitly that they could not cover the costs of high risk births to drug addicted mothers. The total number of Medicaid clients in HMOs has shrunk since 1987 to 2.3 million, though the total number of plans have remained the same (Darby, 1990).

Not only have these contracts been costly, but Medicaid HMOs have also come under fire from community health advocates for discouraging utilization (Webb, 1988). For instance, Representative Henry Waxman (D-CA) opposes Medicaid HMOs on the grounds that they "herd all Medicaid beneficiaries into poor people's prepaid plans, leaving those people - and federal and state taxpayers - vulnerable to the worse profiteering and denial of care" (Darby, 1990).

Public advocates are particularly suspicious of HMOs' commitment to providing psycho-social services, which have open-ended demand among consumers but unpredictable dividends for HMOs. In a study of 30 fee-for-service health care facilities and nine HMOs in Chicago, for instance, Miller (1989) found significantly greater barriers to access to drug treatment programs among the HMOs, as well as poorer quality and continuity of care among HMOs' drug treatment plans. But broad psycho-social services, such as drug treatment, are precisely the interventions required by successful infant mortality reduction initiatives among the poor.

The Center for Public Representation of Milwaukee (Brazner and Gaylord, 1986) studied Wisconsin's experiment with capitated Medicaid service, the Preferred Enrollment Institute (PEI), and reported many anecdotal examples of discouraged utilization. PEI is one of the largest mandatory Medicaid HMO plans in the country, having enrolled more than a third of Wisconsin's Aid to Families with Dependent Children (AFDC)/Medicaid recipients into fourteen HMOs. The study describes the PEI's HMOs' avoidance of referrals to expensive specialists, and lack of encouragement of costly prenatal care and family planning services. Many enrollees were found to be ill-informed about how to use the system, especially when to use emergency facilities. PEI's HMOs were found not to have integrated their prenatal and pediatric care with local perinatal networks, the Women, Infants and Children's Supplemental Food Program (WIC) supplemental food program, or the Early Periodic Screening, Diagnosis and Treatment (EPSDT) pediatric care program. Enrollment in PEI's HMOs during pregnancy was also found to create discontinuities in prenatal care.

While there are many published critiques of Medicaid managed and capitated care, comparative research on outcomes has been scant. The Milwaukee study did not analyze differences in outcomes, in this case because the HMO's utilization and outcome data had not been computerized. In general, neither the HMO nor Medicaid information systems centrally record health outcome data. California's 20-month experiment with capitated Medicaid was terminated partly because the information system was never functional, and utilization review was ineffective (Aved, 1987). In cases where information systems are functional, and data are available, however, researchers have rarely attempted to evaluate quality of care by comparing outcomes.

Even when outcome measures are possible, the appropriateness of the measures is a subject of heated debate between government agencies (HCFA, for example) and the Medicaid community. Researchers often note the difficulties of controlling for risk and severity of illness. Vertrees et al. (1989), for instance, developed an elaborate case-mix adjusted model to test theories of discouraged utilization in a capitated Medicaid program in Philadelphia, but, again, were forced to abandon their analysis because only billing data was available in the information system.

A. National Research on Outcomes of Medicaid HMOs

The Rand Corporation, however, with the Department of HHS, did compare outcomes of the prepaid care to fee-for-service, including a small Medicaid enrollment, in a 1986 study of participants in the Group Health Cooperative (GHC) of Puget Sound, Seattle. The study concludes that for most people, and particularly for those with high incomes, GHC care saved money and may have been better for health outcomes. But for the limited group of Medicaid enrollees that the study covered, health outcomes appeared poorer than for those in fee-for-service. GHC officials acknowledged that the system would need modification and supplementation for the poor (Ware et al. 1986; Ware, 1987).

The most comprehensive comparative study which considered outcomes for Medicaid/AFDC HMOs has been conducted by the Research Triangle Institute for the Health Care Financing Administration (1988). In this research project, a quality of care study was completed in 1985 for populations in mandatory Medicaid HMO demonstration projects in Santa Barbara County, CA and Jackson County, MO. These were compared to fee-for-service Medicaid populations in adjacent communities. The project took a random sample of 2400 women on AFDC between 15 and 45 and their children under 4 years old. Over 2300 births were abstracted from the four projects, as well as other selected information such as maternal health status and outpatient services. Medical records were abstracted for this sample, and questionnaires given to doctors and clients.

The overall conclusion of the study was that the HMO management of care showed no significant effect on self-assessed health status, health habits, or use of preventive services. In particular, no significant effect was found in mean birth weight or the low birth weight rate, in the C-section rate, or in the complications of delivery. (Infant mortality was not reported as an outcome in this study because the method focused on the mothers' records at time of delivery and because the numbers were too small for statistically significant results.) The study found, however, that care in all the sites was "inadequate" and that these problems were "generic to the population served and to Medicaid programs regardless of the existence of capitation or case management" (Executive Summary, Research Triangle, p. 3, 1988)

B. Medicaid HMOs in Illinois: One Attempt at Cost Containment

The Office of Technology Assessment (OTA) has estimated that by offering Medicaid coverage to all pregnant women under the poverty level, to include those currently above the AFDC cut-off level, an additional 19% of those newly covered would initiate early prenatal care who otherwise would not have; such a program would cost only $4 million dollars. Since every averted low birth weight infant saves the system $30,000 in short- and long-term costs, this program would more than pay for itself (OTA, 1987).

This cost-benefit analysis has not, however, moved Illinois lawmakers and the public to fund Medicaid, AFDC, or other programs sufficiently to guarantee that all women have access to prenatal care. Though Illinois ranks 10th in per capita income, it is 51st (including the District of Columbia) in the percentage of charges and costs reimbursed to hospitals by Medicaid, 40th in total health care spending per Medicaid recipient, and 36th in state spending on maternal and child health care programs (IHA, 1988). Illinois hospitals are reimbursed for 67% of the costs of the Medicaid clients they treat, and the cost of completely uncompensated care to the uninsured has doubled for Illinois hospitals since 1980 (from $218 million to $550 million). Partly as a result, a dozen hospitals in poor Chicago communities have closed in the last few years.

Further, the Illinois legislature has pursued two major initiatives in containing the costs of the Medicaid program: 1) a form of case-management by the state, called the Illinois Competitive Access and Reimbursement Equity (ICARE), and 2) the promotion of HMO enrollment for Medicaid recipients. Both options were made possible by the federal OBRA legislation of 1981.

ICARE Signed into law in 1984, the Illinois Health Finance Reform Act created the ICARE system to reduce the Illinois Department of Public Aid's annual expenditures on medical care. One method used under ICARE was the shifting of routine care from tertiary care hospitals to lower-cost community hospitals. IDPA also entered into contracts for specified rates and limited days with participating hospitals through competitive contracting. This reimburesement system has contributed to the financial crisis of inner-city hospitals (Salmon, Lieber, and Ayesse, 1988; Ernst and Whinney, 1989).

Medicaid HMOs Though Illinois Medicaid recipients have had the option of enrolling in HMOs since 1974, enrollment began to increase rapidly in the 1980s when IDPA began mailing HMO brochures to recipients. More than 100,000 recipients of AFDC in Cook County are now enrolled in seven HMOs, a little under 10% of the total. Enrollment has declined since budget shortfalls dictated a freeze on enrollment activity in 1987 (Metropolitan Chicago Health Care Council, 1987).

Chicago as a city was slow to accept HMOs and concerns about Medicaid HMOs in Illinois were raised as early as 1974, when an early Medicaid HMO, the CURE Health Plan, was accused of exaggerated sales pitches and shoddy care. IDPA froze enrollment in CURE, which claimed it was being victimized by self-interested fee-for-service physicians with Medicaid clients (Anderson et al., 1985).

In 1987, the Healthy Mothers and Babies Coalition again criticized Medicaid HMOs with their paper "Cut Rate Care" (Handler et al. 1987). The report recounted anecdotes of numerous problems of Chicago's Medicaid HMOs, including HMO salespeople who, encouraged by their commissions, used false information in marketing; HMO procedures that made disenrollment difficult; HMO perinatal transfers that ignored the regionalized perinatal system; HMO enrollees who had been ill-informed about procedures and policies; and cases of inadequate specialty referral. IDPA and IDPH were accused, in turn, of having done little to monitor HMO quality of care, which they are both mandated to do.

In Chicago, where one of the highest U.S. rates of infant mortality persists (Crown, 1987; Abraham, 1989), the Illinois Department of Public Aid contracts with HMOs for the care of about 10% of Medicaid recipients. It is in this context that, in 1988, the Center for Health Administration Studies at the University of Chicago worked with Chicago HMO (CHMO), the HMO with the largest public aid enrollment in Chicago, to begin to evaluate the care and outcomes for its public aid enrollees.

III. THE STUDY OF CHICAGO HMO

While many managed care delivery systems are new, CHMO evolved from a group which in 1976 became a prepaid capitated health plan and contracted with the state to serve public aid clients. While the city of Chicago was slow to get into the HMO market, it experienced very strong interest and growth in the mid-1980s (Anderson et al 1985). CHMO grew with the market, responding to the state of Illinois' interest in prepaid contracts and growing to be the largest single provider for Medicaid enrollees (from 2000 in 1982 to over 70,000 in 1987). At the same time, they have developed their share of the private market, and in 1987 had approximately equal numbers of public and private members distributed throughout the metropolitan area.

A. Methodology

The study gathered data on 2459 public aid births to women who delivered while enrolled in Chicago HMO between July 1, 1987 and June 30, 1988. These outcomes were compared to those of similar populations in the fee-for-service Medicaid system. It was not possible to compare the HMO population with a directly comparable Medicaid population because the Illinois Department of Public Aid does not collect individual-level outcomes information.

The HMO data were gathered at the central offices of CHMO from both the HMO information system and extensive paper records. The data included type of delivery, incidence of low birth weight, and neonatal mortality (infant death in the first 28 days) as follows:

1) Type of delivery, i.e., caesarian section (c-section), normal spontaneous vaginal delivery (NSV) and complicated vaginal delivery (NSC). Data on these delivery outcomes were complete in the information system.

2) Distribution of birth weights. Data on birth weight were not collected systematically on the information system of the HMO, so a random weighted sample of all births between July 1 and December 31, 1988 was drawn to estimate the rate of low birth weight delivery in the HMO. Individual paper records from both CHMO and the hospitals were examined.

3) Neonatal mortality rates (NNM). Data on neonatal deaths that occurred before discharge from the hospital were generally recorded in the information system. In order to find deaths that occurred outside the hospital, records in the utilization review and social service departments were fully examined.

A.1 Prenatal Care Program

The prenatal program for pregnant members is designed and supervised by the Member Services department which is responsible for all Social Services, Health Education and Discharge Planning. When we began our evaluation in early 1988 the program had a staff specifically for providing support services and programs for pregnant members which included the "Mother and Infant Home Care Program." It required (and depended on) the provider sites to notify the HMO when a member had her first prenatal visit. When notified, the HMO attempted to telephone the member to set up a prenatal home nurse visit to the member to provide information about health, pregnancy and delivery, and especially to encourage the appropriate number of prenatal visits. In addition, the program sent three mailings of relevant information to the member's home, offering to pay for Lamaze classes and "quit smoking" classes, and providing a wide selection of self help books to all pregnant members. Small incentive gifts were given to mothers when they had prenatal visits.

However, the CHMO administration and staff were disappointed in the program. First, the providers were not generally compliant with the HMO's requirement that they notify the HMO of pregnancies; the staff reports that only 60% of the doctors consistently inform the HMO of pregnancies; second, the staff reports that the patients do not return consistently for their prenatal visits and in spite of the physicians' written agreement to inform the HMOs of missed prenatal visits, they rarely complied. Third, and most irremediable, less than half of the pregnant enrollees could be reached by phone. After several tries the staff sent letters offering the benefits noted above, but the response was limited; fourth, even when reached by phone, women often refused a home visit, apparently the staff thinks, seeing it more as some manipulation by the bureaucracy and/or an invasion of privacy rather than a service that would improve their health and the health of their unborn babies.

A.2 Medicaid HMO Demography

Age The age distribution of CHMO's public aid mothers was roughly comparable to those of the non-white and public aid mothers in Chicago in general and the specific communities in which most enrollees lived.

Race The race of members is not recorded by the HMO but it was possible to estimate race from the neighborhoods and zip codes in which members lived. We inferred that the public aid mothers in CHMO were roughly comparable to the general Medicaid/AFDC population. AFDC enrollees in Chicago are 75% black and more than 75% of the public aid women in the HMO lived in the 25 Chicago community areas which are almost entirely black, and have a high rate of public aid enrollment.

Health status of the HMO population was again considered to be similar to that of the neighborhoods in which the HMO members lived. While there are theories of differential selection into HMOs, most do not apply to public aid enrollees, and those that do present conflicting viewpoints on whether HMOs would differentially select for healthier or sicker public aid enrollees. We found no reason to expect that the enrollees would have different prior health statuses than the general public aid population.

A.3 Comparison Populations

Data from three comparison groups were used:

1) Data from two perinatal networks which were chosen for comparison because they had both good data collection systems and included a significant percentage of the CHMO population. The hospitals in the two networks with suitable data covered about 50% of CHMO's public aid deliveries. The hospitals do not gather information on infant mortality after the infant has been released (the rate of neonatal mortality before discharge is recorded below as "in-hospital NNM").

2) Community area perinatal health data that are gathered by the Illinois Department of Public Health, particularly the data gathered on the ten community areas composed of 25 poor inner- city neighborhoods in which the majority of CHMO's public aid mothers live, and which have been targeted for infant mortality reduction programs. On the basis of the community area data, combined with information on the racial composition and AFDC enrollment rate of the communities, a regression model was also constructed to estimate the low birth weight rate for Chicago's public aid deliveries.

3) Data from the Nationwide Evaluation of Medicaid Competition Demonstrations, a Health Care Financing Administration study of outcomes of Medicaid managed care demonstrations projects.

B. Findings

The outcomes of the 2459 births to Medicaid mothers in CHMO in the twelve month period studied are displayed in Table Three. As displayed in the table, data were gathered for two six month periods and aggregated for a one year rate. While the significance of the outcomes can be best understood in comparison to the same outcome measures for similar groups, not all measures are available for all groups, and a patchwork of comparisons must be made to understand the larger picture. Most important, because of the very small numbers of deaths, rates vary sharply in different time periods (See six month rates.)

Table 3: CHMO Medicaid Perinatal Outcomes July '87 - June '88


		Deliveries	       	Low	

		normal	complicated	Cesarian	Birth	Neo-Natal Mortality**

	Births	vaginal	vaginal	sections	Weight	In-Hosp	Total

Full Year	2459	75%	11%	14%	16.0%*	10.6	11.4


July-Dec. '87	1309	73%	14%	13%	16.0%	12.2	13.6

Jan.-June '88	1150	78%	7%	15%	-----	8.7	8.7

* Inferred from 6-month sample. Birthweight averages do not fluctuate the way that mortality rates do, and thus are used to estimate for the entire year.

**Per thousand births.
Table Four presents data on hospital births to Medicaid enrollees in the two Chicago perinatal networks in which over half of the CHMO public aid enrollees are delivered: Mt.Sinai and the University of Chicago networks. (These data are important in part because they are the only outcome data available associated with method of insurance payment.)

When CHMO perinatal outcomes are compared to fee-for-service Medicaid for all races at the University of Chicago (a tertiary care hospital where there is a concentration of high risk births), CHMO births have a considerably lower C-section rate, slightly lower LBW rate, and similar in-hospital NNM rate. When CHMO births are compared with the all black Medicaid population, the CHMO C-section and LBW rates are again lower and the in-hospital NNM rate is also lower for CHMO. However, for Mt. Sinai hospital (an inner-city hospital with both a tertiary neonatal facility for high risk births, and a large primary care service for low risk pregnancies), the C-section, LBW and NNM rates are lower than for CHMO. Data from 1988 show that Mt. Sinai had the lowest C-section rate in Illinois (12%), followed closely by St. Francis Cabrini (15%), University of Chicago (16%), and Cook County (16%), which together making up a significant proportion of the CHMO Medicaid delivery population (Griffin, 1990).

Table 4: 1987 Chicago Perinatal Network Data 


                        Births	NSV	NSC	C-Secs* LBW	In-Hosp NNM	

CHMO	2459	75%	11%	14%	16.0%	10.6


Medicaid: All Races	

U. of C. Network	7211	75%	7%	18%	18.9%	10.5


Medicaid: Black

Mt. Sinai Network	2347	83%	6%	11%	13.9%	8.6

U. of C. Network	5788	77%	6%	17%	20.8%	11.7

Table Five compares outcomes of the CHMO Medicaid population --birth weight and NNM-- with births from a broader population of all residents in the poorest Chicago community areas where the largest part of the CHMO Medicaid population lives. (Data from these twenty-five neighborhoods have been collected because they have been targeted by the public health department for infant mortality reduction programs.) When compared with birth outcomes in their own communities, HMO enrollees have higher incidence of low birth weight (16%), but lower neonatal mortality rate. Their NNM rate of 11.4 is among the lowest in the ten community areas, where, over a three year period, infant mortality ranged from 11.4 to 17.6 and averaged 13.6.


Table 5: Chicago's High Public Aid Recipient Communities 


                       Births         LBW       NNM

CHMO	2459        16.0%      11.4


Infant Mortality Reduction Initiative data

All Races, 19 poorest areas - 1987	20385	14.3%	13.5


Illinois Health Data

Black Only, 25 poorest areas - 1986	20087	14.8%	14.3



Table 6: Extrapolated Chicago Medicaid Population Low Birth Weight Rate*


75% Black, 100% public aid		14.6%		


100% Black, 100% public aid		18.3%	


*Projection from regression analysis of public aid and community perinatal data

Table Seven presents the only suitable comparison available of national data from the Nationwide Evaluation of Medicaid Competition Demonstrations--a national effort to evaluate care under Medicaid HMO contracts. The numbers of births were limited and although the study reflects a higher C-section rate than Chicago HMO clients and a lower LBW, the neonatal mortality numbers were too small to report.


Table Five: Nationwide Medicaid Competition Evaluation


	          Births    C-Secs    LBW

CHMO	2459	14%	16.0%	


Research Triangle Demonstrations

Medicaid Fee for Service	575	16%	11.9%

Medicaid Managed Care	599	16%	11.2%



IV. DISCUSSION

As just noted in our study and with other studies of perinatal outcomes, it is difficult to demonstrate statistically significant differences using usual levels of confidence when studying the relatively small numbers of low birth weights, and particularly of infant deaths. For example, the Nationwide Evaluation of Medicaid Competition Demonstrations (Research Triangle Institute 1988) does not report infant deaths because the numbers are too small for appropriate analysis. In this study, the differences in mortality and morbidity rates are generally not statistically significant.

Still, certain descriptive comparisons and the directions of the findings are revealing. When the CHMO population of mothers and infants is directly compared to aggregate populations of mothers and infants with similar demographic characteristics (race, age, socioeconomic status), and, in fact, from the same communities in 75% of the cases, the CHMO rate of low birth weight is consistently in the high range of variation while the NNM rate is consistently in the low range of variation. Thus infants delivered in CHMO seem to have a lower birth-weight-specific neonatal mortality rate than comparable populations. While the HMO infants, as a group, are defined as higher risk because of their weight, the probability of their dying in the first 28 days is relatively low.

In addition to these general results, the research has highlighted certain issues of continuing sociological interest that we feel need further attention. Four issues with important consequences are:

1) The effects of the public aid and social service system on continuity of care for Medicaid clients, i.e., issues of eligibility, enrollment and disenrollment;

2) The structural and attitudinal barriers to HMOs' successful preventative, including the conditions of poverty, patient and provider atitudes;

3) The consequences of the poverty of information on the poor;

4) The effects on outcomes of the increased use of neonatal technology, at the expense of family planning and prenatal care.

1. Enrollment, Disenrollment and Continuity of Care

Lack of continuous care is said to be a significant problem in the prenatal care of mothers from high risk groups. We noted earlier the continuous turnover of HMO public aid members -- both those who are pregnant and those who are not. Staff in CHMO Consumer Services report that both enrollment and disenrollment are often caused by a change in Medicaid eligibility. 1 Research and agencies focused on infant mortality emphasize the importance of sustained benefits during pregnancy and, in fact, the Medicare Catastrophic Coverage Act of 1988 mandates that by July 1990, all pregnant women and infants to one year who have incomes below 100% of the federal poverty level, and meet AFDC resource standards, will be covered.

The volatile nature of eligibility, and the delayed processing of public aid records affect those enrolled in the HMO as they do those using the Medicaid green card. In the current competitive environment turnover is made more likely as HMOs change their premiums and benefit offerings, and employers change plans as they seek to contain their health care costs. Wintringham and Bice (1985) note that two years of sustained coverage in an HMO are usually necessary for public aid clients before utilization patterns stabilize. Welch (1988a) also found that Medicaid clients were more likely to disenroll from HMOs, after an initial period of high utilization, than privately insured HMO members. This suggests that HMOs are more likely to participate in Medicaid contracting if their enrollees are locked in for at least a year or two, as indicated by a DHHS survey of HMOs (La Jolla Management Corporation, 1985). HMO lock-ins would also ensure greater continuity of care for the clients, and encourage HMOs to control costs through long-term investments in preventive services rather than short-term discouraged utilization (Welch, 1988b; Omenn, 1987). A recent study in Michigan (Carpenter, 1989) compared outcomes for Medicaid populations who voluntarily enrolled in HMOs to those who were mandatorily assigned, and found no differences in health outcomes or evidence of more discouraged utilization by the mandatory assignees.

2. Structural Conditions and Preventative Health Services

All models of HMOs are expected to emphasize preventative care, and, in fact, all HMO administrators may recognize the opportunity to control utilization with active preventive programs. Staff and group models, however, because they have more centralized organization and peer interactions, are better able to create and sustain programs than are IPAs and network HMOs which contract with autonomous physicians. An effective preventive program requires the HMO administration to support programs with centralized information systems and physician incentives; the providers must see the value for both their patients and themselves; and the client/patients must use the programs offered.

Summarizing HMO studies, Luft (1988) concludes that the quality of HMOs' care is generally the same as fee-for-service. But he notes that most of these studies are of staff-model HMOs and not of the more decentralized Independent Practice Associations. We observed that the fragmentation of CHMO's decentralized IPA, combined with the enrollee behavioral patterns, contributed to the problems of CHMO's prenatal care program. First, there were many different departments, and professional cultures (social service vs. utilization) that required coordination. Secondly, physicians who contracted with CHMO were neither rewarded when they reported pregnancies or prenatal care visits to the central office, nor penalized if they did not report. Thirdly, even when the pregnancies were reported to the central office, they were not recorded on the central information system, but rather in a card file, losing the opportunity for overall coordination and monitoring of prenatal care.

CHMO also has no control over the supply or quality of physicians in poor communities, with which they are required to contract. Summarize Fosset/Perloff work

For the privately insured, smaller numbers of obsterician-gynecologists per person in an area has been found to be correlated not only with higher patient loads but with more frequent travel to adjacent communities for delivery, which is in turn correlated with more complicated deliveries, rates of prematurity and neonatal care costs (Nesbitt, 1990). Clearly, an IPA-type HMO has little leverage in changing the distribution and quality of its contracted physicians; this is a federal and state responsibility.

We found the obstacles to motivating the public aid enrollees to utilize preventative care were also daunting. The prenatal hazards faced by this community are more complex and severe, and probably less responsive to preventative advice than those of employed populations with more stable family support for compliance (Bryant, 1982; Ward, 1987) and less severe prenatal hazards. Many of the clients had no phones, or did not answer. Many did not understand the concept of managed care and assigned primary care physicians, and continued to seek familiar alternative sources of care.

[Inner city residents often have a much broader definition of prenatal care and where they may get it than do providers and public agencies. For instance, many rely to an unknown, been significant extent, on the free City and County prenatal services, which further fragments the provision of care (Fossett, Peterson and Ring, 1989).]

Consistent with reports in the literature, CHMO staff reported that enrollees often missed scheduled physician appointments, sometimes due to lack of money for transportation, lack of child care, or other unexplained reasons (Dutton, 1978, 1986). When they were reached by the program, they often refused a prenatal care nurse visit.

Medicaid recipients in the HMO have many services officially available to them which they do not use. Staff at the HMO who administer these services --like their professional counterparts in other institutions and agencies outside of the HMO-- are not able to fully explain the general ineffectiveness of their prenatal programs. While there is documented success with many of the programs and clients, the staff seems disproportionately aware of their failed or frustrated efforts. For example, the social service office of CHMO reports widespread resistance among public aid clients to visits by nurses, social workers, and home helpers.

Providers often described successful programs and many good outcomes in interviews, but emphasized their failures with their clients/patients when we asked them specifically to evaluate the success of their efforts. When clients do not respond to them or inexplicably fail to use the services they offer, they are angry and seem particularly frustrated. They do not know why, when, or which clients will behave in this "self-destructive manner." In spite of providers' many successes, they seemed to have developed a self- protective cynicism since they could not predict with which clients they would fail. This pessimism and "burnout" is undoubtedly another barrier to effective care.

Given these structural and attitudinal barriers, public aid HMOs trying to stay solvent may not be motivated to develop and support preventative programs

[A recent program in New York City paid outreach workers a commission for every low-income woman they located and got into prenatal care, and found that the high costs of such a program were nonetheless lower than the costs of additional low-birth weight babies (Brooks-Gunn, 1989).]

3. Social Class, Race and the Collection of Data

It was a major goal in this study to devise a suitable surrogate public aid group from the community to compare to the HMO. We sought data on race and socioeconomic status. Race is widely used as a risk factor in perinatal health, but in controlled studies race is revealed to be also a proxy for other variables. While low birth weight is more likely among blacks (Lee et al., 1988), when socioeconomic conditions are controlled, race is nearly insignificant as a predictor of infant morbidity and mortality (Institute of Medicine, 1985; Lieberman, 1987). These socioeconomic conditions are themselves proxies for a host of health-related conditions, such as poor nutrition, stress, cigarette smoking, exposure to toxins, drug abuse, and lack of preventive medical care.

Yet health data bases often code by race but rarely by economic condition. There is almost no coding to designate if payor is "Medicaid," or if employment status is "public aid." As Paul Starr et al. (1989) have described in their recent The Politics of Numbers, the power of social groups is reflected in the quantity of information collected on them and made available about them. The presence of racial information (black, Asian, Hispanic) in federal and state health records contrasts with the absence of income, employment, health insurance and public aid information on health records and suggests that race is a more politically powerful division or a classification more readily acknowledged than class. Using most governmental and organizational data sources it is virtually impossible to control for the intercorrelation of race and class in their effects on health.

The lack of information is not only a research problem but contributes to the spotty and inadequate care that the public aid population often receives (Alexander et al., 1989; Buescher, 1989; Davis, 1986). If there is to be any continuity in the care that the public aid population receives, from emergency room to primary physician, greater centralization of health records between hospitals, primary care givers, and public agencies will be necessary (Fox and Stuart, 1989).

4. Reliance on High Technology to Improve Outcomes

The U.S. Office of Technology Assessment (1987:4) notes that the overall decline in infant mortality in the last 20 years comes through "two routes: 1. Changes in birth weight-specific infant mortality rates (through neonatal intensive care); and 2. Changes in the distribution of birth weights toward heavier babies. But 91% of the improvement has been due to advances in costly high technology solutions of the first route. In fact, birth weight distributions actually deteriorated." Had the technological support for low birth weight babies not improved, there would have been an overall increase in U.S. infant mortality rates.

Wise et al. (1988) similarly show that, as access to regionalized neonatal intensive care has stabilized the mortality rates of low birth weight and premature infants, the infant mortality rate has become even more sensitive to the influences of socioeconomic conditions and prenatal care. Wise and his co-authors found that there had been a rapid increase in access to tertiary neonatal intensive care units in Boston, such that more than 95% of Boston's low birth weight infants are delivered in hospitals with such units. Similarly Mayfield (1990) found that the mortality rate of low-birth weight newborns born in Level III facilities was half that of infants born in facilities without advanced intensive care, controlling for maternal risk, birthweight and obstetrical volume.

This increased access to highly sophisticated technological care, and consequent increased survival rate for low birth weight infants, has depressed the infant mortality rate, which is increasingly composed of a rising rate of mortality in normal birth weight and post-neonatal infants (Ekwo and Gosselink, 1989). This change in the composition of infant mortalities makes the rate even more sensitive to access to perinatal and pediatric care, and to socioeconomic and political factors, such as restricted Medicaid eligibility (Wise et al. 1985; Starfield 1985; Khoury et al. 1984).

In addition to an over-reliance on intensive care, and an under-utilization of prenatal care, maternal and infant health among the poor has been strongly effected by restricted access to subsuduzed family planning and abortion services. Legge (1985) finds a strong correlation between liberal abortion policies and infant and maternal health, and Winikoff (1987) estimates that increased use of family planning could reduce infant mortality by 20%. Joyce and Grossman (1989) found that black Americans' birthweights have been especially sensitive to fluctuations in availability and use of abortion, and predict that restrictions in abortion access will increase low-birth weight outcomes.

Infant mortality was not correlated with enrollment time in CHMO, and thus with the HMO's opportunity to provide prenatal care. Rather, any effects of CHMO membership seemed to manifest in the delivery and postpartum period. While the concern of critics is that HMOs with Medicaid contracts will try to cut costs with Medicaid clients, this study suggests that CHMO's services parallel those available to the poor in the public and private sector. That is, the HMO pays a high price to provide high technology intensive care to high risk premature and low birth weight infants after birth, rather than successfully targeting family planning and prenatal care, the more cost effective ways to lower infant mortality and morbidity (U.S. OTA, 1987; Joyce et al., 1988).



V. SUMMARY AND CONCLUSIONS

In summary, this focused evaluation of the delivery of care to Medicaid clients in an HMO, in addition to documenting the efforts and outcomes of CHMO, brings further attention to 1) the serious problems of the inner city poor as they are manifested in poor health outcomes-- particularly for mothers and infants, 2) the continuing inability of public or private delivery systems--in spite of apparent efforts--to consistently connect poor inner city women with standard prescribed, and demonstrably medically effective and cost effective prenatal care, and 3) the documented increase in sustenance of premature and low birth weight infants with costly high technology intensive care through the neonatal period. Further study at the community level is crucial to better understand why childbearing women in poverty, both within and outside of the HMO, are not well connected with prenatal care systems which are proven to best protect their pregnancies and enhance their health and the growth and health of their infants in the most cost effective way.

Until that issue is understood, however, this study has documented, Chicago HMO continues to compensate --as does the larger system of which it is a part-- with costly high technology care, and thus accomplishes relatively low neonatal mortality among the many vulnerable infants born in the HMO.

Thanks

We would like to offer our thanks to the many public and private officials, researchers, managers and health care providers across the nation who facilitated this study. While it is not possible to name them all we especially want to mention the following for their data and substantive discussion: Leatrice Berman, Maria Corpuz, Noah Epstein, Richard Ferguson, Dr. Lawrence Gartner, Arden Handler, Dr. Arthur Kohrman, Phyllis Jones, Dr. Kwang-sun Lee, Sara Loevy, James Masterson, Ross Mulner, Dr. Steven Myers, Janet Perloff, Randy Pletcher, Iris Shannon, Rodney Taylor, Mark Testa, and Dr. Ernestine Willis.

The following institutions provided data for Illinois and Chicago: Chicago Department of Health; Illinois Department of Health; Illinois Department of Public Aid; University of Chicago's Center for Urban Research and Policy Studies, Department of Pediatrics, Perinatal Network, Woodlawn Maternal and Child Health Center; Mount Sinai Hospital's Department of Pediatrics and Department of Obstetrics and Gynecology; and the National Opinion Research Center.

Finally we would like to thanks David Citron and Dr. Charles Kelly, and all the departments at Chicago HMO, without whose encouragement and support the study would never have been conducted.



BIBLIOGRAPHY on MEDICAID MANAGED CARE

Anderson, Maren and Peter Fox. 1987. "Lessons Learned from Medicaid Managed Care Approaches." Health Affairs 71-86.

Andrews, Roxanne et al. 1989. "The Effects of Method of Presenting Health Plan Information on HMO Enrollment by Medicaid Beneficiaries." Health Services Research 24:311-328.

Ansberry, Clare. 1987. "After Five Years, Experimental Health Project for Medicaid Recipients Shows Mixed Results." Wall Street Journal, (, Part June 1):

Aved, Barbara M. 1987. "The Monterey County Health Initiative: A Post-Mortem Analysis of a California Medicaid Demonstration Project." Medical Care 25:35-45.

Bartlett, Lawrence. 1986. "The Management of Medicaid Inpatient Hospital Expenditures." Affording Access to Quality Care: Strategies for State Medicaid Cost Management, eds. Richard Curtis and Ian Hill. Wash. D.C.: National Governor's Association.

Bice, Thomas et al. Dept. of Epid. and Public Health, Yale U. 1988. "HMO Compared to Fee-for-Service among Medicaid Beneficiaries: Use, Cost and Quality." Center for Health Administration Studies Workshop,

Blackwell, Barry, Mary Gutmann and Lily Gutmann. 1988. "Case review and quantity of outpatient care." American Journal of Psychiatry 145:1003-1006.

Brazner, K., C. Gaylord. 1986. Medicaid HMO's and Maternal and Child Health: An Assessment of Wisconsin's Program for AFDC Families. Milwaukee: Center for Public Representation.

Camille, Mark. 1986. "State-local perspectives: Medicaid and HMOs: Beyond Fee-for-service." New England Journal of Human Services 6:32-33.

Cantrell, Daniel. 1988. "The HMO/Medicaid Perspective." IAHMO Forum May:8-9.

Carey, Timothy S., Kathi Weis and Charles Homer. 1991. "Prepaid versus Traditional Medicaid Plans: Lack of Effect on Pregnancy Outcomes and Prenatal Care." Health Services Research 26:165-181.

Carpenter, Eugenia. President Gini Associates, Ann Arbor 48106. 1989. "Evaluation of Medicaid Mandatory Enrollment Project." APHA 117th Annual Meeting,

Center for Public Representation. 1986. Medicaid HMOs and Maternal and Child Health: An Assessment of Wisconsin's Mandatory HMO Enrollment Program for AFDC Families. Milwaukee, Wisconsin: Center for Public Representation.

Center for Policy Studies. 1985. Prepaid and Managed Care under Medicaid: Characteristics of Current Initiatives. Washington: National Governor's Conference.

Center for Health Research. "Volume 5, Quality of Care Study from the Nationwide Evaluation of Medicaid Demostrations." Nationwide Evaluation of Medicaid Competition Demonstrations . In Research Triangle Institute, 1988

Chicago Hospital Council. Medicaid Full Service Capitation RFP . 1982, October

Cohen, Robert. 1987. "HMO Quality of Care Program." Journal of the ACURP 2:111-125.

Coltin, K. et al. 1981. "Effects of Removing the Waiting Period for Maternity Benefits in an HMO." Group Health J 2:.

Curtis, R. E. et al. 1904. "States Use Managed Care, Subsidies to Address Working Poor Problem." Business and Health 53.

Darby, Mary. 1990. "Administration pushes managed care for Medicare, Medicaid." Managed HealthCare, (, Part March 12):

DesHarnais, S. L. 1985. "Enrollment in and Disenrollment from Health Maintenance Organizations by Medicaid Recipients." Health Care Financing Review 6:39-50.

DHHS. Competitive Medical Plan . In U.S. Department of Health & Human Services, 1985, June

Douglass, R. L. and R. E. Torres. 1994. "Evaluation of a managed care program for the non-Medicaid urban poor." Journal of Health Care for the Poor & Underserved 5:83-98.

Federal Register. 1986. "Medicare and Medicaid Programs; Health Care Financing Research and Demonstration Cooperative Agreements and Grants; Amendents." The Federal Register

Freund, D. A. and R. E. Hurley. 1987. "Managed Care in Medicaid: Selected Issues in Program Origins, Design, and Research." Annual Review of Public Health 8:137-63.

______. Research Triangle Institute. 1988. "Medicaid Managed Health Care Projects." GHAA Annual Conference, Chicago: GHAA.

Glablum, Trudi and Sidney Trieger. 1982. "Demonstrations of alternative delivery systems under Medicare and Medicaid." HCFR 3:1-11.

Goldberg, Victor. 1975. "Some Emerging Problems of Prepaid Health Plans in the Medi-Cal System." Policy Analysis 1:55-68.

Handler, Arden et al. 1987. Cut-Rate Care: An Evaluation of HMO Service for Cook County's Medicaid Mothers and Children. Chicago: Healthy Mothers and Babies Coalition.

Handler, Arden, Roy Petty, Melody Bookenfeld and Janet Evans. Cut-Rate Care: An Evaluation of HMO Service for Cook County's Medicaid Mothers and Children . In Healthy Mothers and Babies Coalition, 1987

Hester, J. and E. Sussman. 1974. "Medicaid Prepayment: Concept and Implementation." Milbank Quarterly 415-444.

Hillman, Diane and Jon Christianson. 1984. "Competitive bidding as a cost-containment strategy for indigent medical care: The implementation experience in Arizona." Journal of Health Policy, Politics and Law 9:427-451.

HMO Manager's Letter. 1988. "Bush Will Encourage Medicaid HMOs, Aide Says." HMO Manager's Letter 9.

1989. "HMO Medicaid Payments Vary Nationwide." HMO Manager's Letter 6.

1988. "HMO'S Role in Medicaid Expands." HMO Manager's Letter 3.

Hohlen, Mina, Larry Manheim and Gretchen Fleming. 1990. "Access to Office-based Physicians Under Capitation Reimbursement and Medicaid Case Management." Medical Care 28:59-68.

Hurley, Robert, Deborah Freund and Donald Taylor. 1989. "Emergency Room Use and Primary Care Case Management: Evidence from Four Medicaid Demonstration Programs." American Journal of Public Health 79:843-846.

Hurley, Robert E. and Deborah Freund. 1988. "A Typology of Medicaid Managed Care." Medical Care 26:.

Iglehart, John. 1983. "Health Policy Report - Medicaid turns to prepaid managed care." New England Journal of Medicine 308:976-980.

La Jolla Management Corporation. Contracting Activities of HMOs and State Medicaid Agencies . In DHHS, 1985

Langa, Kenneth and Elliott J. Sussman. 1993. "The effect of cost-containment policies on rates of coronary revascularization in California." New England Journal of Medicine 329:1784-9.

Liang, Bryan. 1990. "Continuing Mandatory Medicaid HMO Enrollment Programs: An Analysis of State, HMO and Delivery Site Characteristics." Doctoral Dissertation. Univ. of Chicago.

Lipson, Debra. Intergov. Health Policy Project, George Washington Univ. 1988. "Summary of Data: Have Managed Care Programs Met Their Goals." National Conf. on Restricted Choice: Costs and Benefits,

Medicaid Payments to Health Maintenance Organizations and Competitive Medical Plans. Social Security Act of 1976, Sec. 114 (a) . In HHS, 1976

Morris, Steven. 1989. "HMO drops Bay Area Public Aid: Maxicare cites high costs in cocaine-affected births." Tribune, (, Part Jan. 5):

Mullooly, John and Donald Freeborn. 1979. "The Effect of Length of Membership Upon the Utilization of Ambulatory Care Services: A Comparison of Disadvantaged and General Membership Populations in a Prepaid Group Practice." Medical Care 17:922.

Nassirpour, Mehdi. HMO Disenrollment Patterns Among Medicaid Beneficiaries in Illinois (a Life-Scale Analysis) . In IDPA, 1988

Neuschler, E. 1988. "HMOs and Medicaid; How can states sustain market interest?" Business and Health 5:50-1.

Oberg, Charles and Cynthia Longseth. 1986. "Medicaid - entering the third decade. Enrollment in HMOs and alternative health systems." HMOs and Medicaid: New intiatives and challenges, GHAA.

Omenn, Gilbert S. 1987. "Lessons from a Fourteen-State Study of Medicaid." Health Affairs 118-122.

Paul, J. E. and T. A. Sherrill. Research Triangle Institute. 1989. "Survey of Medicaid Beneficiaries in Minnesota, 1986 and 1988." APHA 117th Annual Meeting,

Perloff, Janet. 1987. "Safeguards are necessary for Medicaid HMOs to work." Chicago Sun-Times, (, Part Nov. 10):35.

Perloff, Janet, Phillip Kletke and Kathryn Neckerman. 1987. "Physicians' Decisions to Limit Medicaid Participation and Policy Implications." Journal of Health Politics, Policy and Law 12:221-235.

Prottas, Jeffrey and Eugenia Handler. 1987. "The Complexities of Managed Care: Operating a Voluntary System." Journal of Health Policy, Politics and Law 12:253-269.

Randall, V. R. 1905. "Impact of managed care organizations on ethnic Americans and underserved populations." Journal of Health Care for the Poor & Underserved 3:.

Research Triangle Institute. 1988. Medicaid Managed Health Care Projects, Group Health Association of America.

______. 1988. Nationwide Evaluation of Medicaid Competition Demonstrations. Baltimore: HCFA.

Rosenbaum, Sara, Dana Hughes, Elizabeth Butler and Deborah Howard. 1988. "Incantations in the Dark: Medicaid, Managed Care, and Maternity Care." Milbank Quarterly 66:661-693.

Rowland, Diane and Barbara Lyons. 1987. "Mandatory HMO Care for Milwaukee's Poor." Health Affairs 87-100.

Salmon, J. Warren, H. Stephen Lieber and Mary C. Ayesse. 1988. "Reducing Inpatient Hospital Costs: An Attempt at Medicaid Reform in Illinois." Journal of Health Policy, Politics and Law 13:103-127.

Savich, Robert. The Impact of HMO Contracting on the Integrity and Development of the Regional Perinatal Networks . In IDPH, 1985

Schnedier, A. and J. Stern. 1988. "HMOs and the Poor." Northwestern University Law Review 70:90-138.

Spitz, B. 1987. "A National Survey of Medicaid Case-Management Programs." Health Affairs Spring:61-70.

Spitz, Bruce. 1987. "A National Survey of Medicaid Case-Management Programs." Health Affairs Spring:61-70.

______. 1979. "When a Solution is Not a Solution: Medicaid and HMOs." Journal of Health Politics, Policy and Law 3:497-518.

Stebbins, J. and M. Newman. 1987, October. "Health Care Hustlers." Chicago Sun Times,

A Summary of Continuing Mandatory Medicaid HMO Enrollment Programs: A Description of State, HMO and Delivery Site Characteristics .

Temkin-Greener, H. 1986. "Medicaid Families under Managed Care: Anticipated Behavior." Medical Care 24:721-732.

Timberlake, Robert W. 1986. "Recipient satisfaction under a mandatory system." HMOs and Medicaid: New Intiatives and challenges, GHAA.

Trieger, S., T. W. Galblum and G. Riley. HMOs: Issues and Alternatives for Medicare and Medicaid . In DHHS/HCFA-Office of Research, Demonstration and Statistics,

Vertrees, James C., Kenneth G. Manton and Keith C. Mitchell. 1989. "Case-mix Adjusted Analyses of Service utilization for a Medicaid Health Insuring Organization in Phildelphia." Medical Care 27:397-411.

Ware, John E. et al. 1986. "Comparison of Health Outcomes at a Health Maintenance Organization With Those of Fee-For-Service Care." Lancet 1017-22.

Warren, Barbara. 1989. "Preliminary Evaluation of SOBRA: Relative Perinatal Risks and Costs in a Managed Care System." APHA 117th Annual Meeting,

Webb, Gary. 1988. "Poor Health: For patients on welfare, there are no rules against lousy care." Progressive May:17-19.

Welch, W. P. 1988. "HMO Enrollment and Medicaid: Survival Analysis with a Weibull Function." Medical Care 26:45-52.

Welch, W. P. and Mark Miller. 1988. "Mandatory HMO Enrollment in Medicaid: The Issue of Freedom of Choice." Milbank Quarterly 66:618-639.

Wintringham, Karen and TW Bice. 1985. "Effects of Turnover on Use of Services by Medicaid Beneficiaries in a Health Maintenance Organization." Group Health J Spring:12-18.

Wolfsen, J., P. J. Levin and W. J. Brock. 1987. "Linking health care for the poor to health care for profit." Health Affairs 6:129-135.