The Impact of California’s Medi-Cal Managed Care Program on
Birth Outcomes and Prenatal Care Utilization During the 1990s
DRAFT
THE IMPACT OF CALIFORNIA’S MEDI-CAL MANAGED CARE PROGRAM
ON BIRTH OUTCOMES AND PRENATAL CARE UTILIZATION
DURING THE 1990s
Tania Barham, Paul Gertler, Kristiana Raube
The University of California Berkeley
December 15, 2002
PRELIMINARY RESULTS
The authors we like to thank Claudio Ferraz, Frederico Finan, Guido Imbems, Karen Macours, Craig McIntosh, Xiangyi Meng for their helpful comments and input. All mistakes are solely those of the authors. The results are still preliminary. Please do not cite.
1
The Impact of California’s Medi-Cal Managed Care Program on
Birth Outcomes and Prenatal Care Utilization During the 1990s
Executive Summary
During the 1990s, the US turned to managed care to help curb government Medicaid expenditure and improve the access and quality of health services for the poor. State governments were encouraged to switch from fee-for-service to managed care due to private sector success with managed care in the 1980s. By 2001, all states except Alaska and Wyoming had some form of Medicaid managed care, and 57 percent of the Medicaid population nationally were part of a managed care arrangement.
California first introduced managed care for their Medicaid population in Santa Barbara in 1983. However, it was not until the early 1990s that the state mandated 19 counties to switch their Medicaid beneficiaries from fee-for-service to new commercial managed care arrangements. By January 2002, 52 percent of Medicaid beneficiaries in 22 counties in California were enrolled in managed care plans.
There are three main types of Medicaid managed care plans in California:
- Two-Plan Model (TPC)
- County Organized Health System (COHS)
- Geographic Managed Care (GMC)
Of these three models, TPC enrolls the largest number of Medicaid recipients, 42 percent, while, COHS and GMC account for 9 percent and 6 percent of the beneficiaries respectively. In this paper we examine COHS and TPC plans, as they are the models for future expansion.
In all three plans, the Calworks-linked population (state welfare recipients) must enroll in Medicaid managed care to receive their free state health care benefits. California also offers Medicaid to poor pregnant women who are not eligible for the state welfare program. Medicaid covers their prenatal care and delivery costs as well as the health care costs of their child for one year. In order to ensure continuity of care, the state does not mandate all pregnant women receiving Medicaid to enter into a managed care arrangement. Many pregnant women can apply for a medical exemption that allows them to opt out of managed care. As a result, even though 50 percent of Medicaid beneficiaries were enrolled in a managed care plan in 2000, only 29 percent of deliveries were covered by managed care, while the rest were paid for on a fee-for-service basis.
Mark Duggan has previously shown that the introduction of managed care in a county leads to a 15 to 20 percent increase in expenditures. He notes, however, that those counties that already had 35 percent or more of their Medicaid beneficiaries in managed care at the time of the mandate experienced no increase in cost. There is reason to believe that the cost increases in counties new to managed care are short term and associated with implementation. It will be important to re-examine the question of cost in a few years when more data are available to determine whether costs subsequently decreased or continued to rise.
In this paper we investigate the impact of the California managed care mandates with an opt out option for pregnant women, on four birth outcomes:
- adequacy of prenatal care index (APNCU)
- the incidence of low birth weight babies (< 2500 and <1500 grams)
- the incidence of preterm babies (born before 37 and 33 weeks of gestation)
- the cesarean section rate
To assess the impact of the Medicaid managed care program we must define the counterfactual, that is, what would have happened to birth outcomes in the absence of the mandates. Since randomization was not employed to designate Medicaid patients into the managed care and fee-for-service plans, we use two techniques from the impact evaluation literature, difference analysis and double difference analysis, to define the counterfactual. Each technique makes a different assumption about what aspect of the counterfactual group behavior is supposed to mimic the treated group had the treatment not taken place.
The counterfactual group for difference analysis is the Medicaid beneficiaries in managed care counties before the mandates took place. The first difference is achieved by comparing the means of the outcome variable between the post- and pre-mandate periods. This technique assumes that the growth rates of the outcome variables in the post-mandate period, had the mandates not taken place, is the same as the growth rate in the pre-mandate period.
To perform double difference analysis we also must define a separate comparison group. We use three different comparison groups:
- Fee-For-Service (FFS) Group: California Medicaid beneficiaries in counties not mandated to switch from fee-for-service to managed care.
- The Self-Pay Group: Mothers who were not insured so paid for their own prenatal care and delivery.
- Commercial Group: Mothers who are covered by commercial insurance plans for their prenatal care and delivery.
To determine the double difference estimator, a second difference that measures the change in means of the outcome variable for the comparison group between the pre- and post- mandate period is also calculated. The impact of the program is found by subtracting the first difference from the second difference. Note, the key counterfactual assumption underlying this model is that in the absence of the mandates, the growth rates of the outcome variables in the treatment group is the same as the comparison group over the time of the analysis. We use California birth certificate data from 1993 to 2000 and perform the analysis separately for the COHS and the TPC counties.
We find that the mandates have no consistent impact on any of the birth outcomes, with the exception of the cesarean section rate. Our preliminary results show that the cesarean section rate did not rise by more than 10 percent for the TPC group and 23 percent for the COHS.
Also, the move to Medicaid managed care did not impact low birth weight, premature delivery, or use of prenatal care. Given Duggan’s findings that costs increased up to 20 percent, we conclude that California should not expand its managed care program at this time. However, it may be the case that these cost increases are temporary and associated with implementation. As a result, prenatal care access and birth outcomes, as well as costs, should be monitored as more years of data become available.
1
The Impact of California’s Medi-Cal Managed Care Program on
Birth Outcomes and Prenatal Care Utilization During the 1990s
1. 0Introduction
During the 1990s, the State of California undertook a sweeping change in Medi-Cal[1], their health care program for the poor, by by shiftinging more than half of the population from fee-for-service into managed care. The primary rationale for the switch was to improve access to quality care and control growth in expenditures. Managed care arose out of a belief that it would help curb rising government Medi-Cal expenditures which sky- rocketed, however, between 1989 and 1992. Partly due to the expansion of eligibility for pregnant mothers and infants, Medi-Cal expenditure growth rates reached 26 percent between 1991 and 1992 compared to 6 percent annually between 1980 and 1988 (Kaestner, 2002). In 2002, California Medicaid expenditures topped $17 billion, and were the second highest and are the number 2 budget item behind education.[2] Concern over lack of access to quality primary care for low-income groups also played a key role in the increased reliance on managed care, especially in California (McCall et al., 2000).
California is not alone in turning to managed care for its Medicaid population. Presently, all states except for Alaska and Wyoming have some form of Medicaid managed care (Kaiser Commission on Medicaid and the Uninsured, 1999). States embraced managed care during the 1990s following private sector success with controlling costs through the use of managed care during the 1980s (Kaestner, 2002). Between 1991 and 2001, the proportion of Medicaid recipients in managed care rose from 10 to 57 percent nationally (Table 1).
Medicaid managed care is not new to California, as Santa Barbara began a managed care program in 1983. Managed care enrollment did not accelerate, however, until the mid-to-late-1990s when the state mandated 19 counties to shift their Medi-Cal beneficiaries into new, commercial managed care arrangements. In January 2002, over 5.8 million Californians, or 17 percent of Californians, were enrolled in Medi-Cal. Approximately 52 percent of these recipients were part of California’s Medi-Cal managed care program (Table 1).
The Medi-Cal managed care mandates require that for almost all health care services, Medi-Cal beneficiaries use the county endorsed managed care Health Maintenance Organizations (HMOs). Beneficiaries are not reimbursed for care they receive from non-HMO physicians or hospitals. However, prenatal care, deliveries, and health care for infants provide an exception to this rule. Many non-welfare beneficiaries who do not normally receive Medi-Cal may qualify to receive prenatal care, delivery services, and health care for their newborns from Medi-Cal when they become pregnant. To ensure continuity of care for this expanded group, the county Medi-Cal plans allow many pregnant women to opt-out of the otherwise required managed care plan and seek prenatal care and delivery services from a provider of their choice. This provider will be reimbursed on a fee-for-service basis if they are not part of the Medi-Cal HMO system.
In this paper, we investigate the impact of the Medi-Cal managed care mandates, which includes the opt-out option for pregnant women, on access to and quality of care by examining birth outcomes. In particular, we focus on three indicators of quality of care, cesarean section rate, low birth weight and preterm babies. Access to care is measured by the Adequacy of Prenatal Care Utilization (APNCU) Index developed by Milton Kotelchuck (Kotelchuck, 1994). This index combines the month that prenatal care began and the number of prenatal care visits into one indicator of access.
We find that managed care arrangements have not changed the incidence of low birth weight or premature babies. Although we do find a positive impact on the APNCU index we argue that urban and rural differences explain this result. Lastly, there is a small increase in the cesarean section rate, though the result is not robust.
The paper is organized as follows. We provide a brief background of the main managed care models in California in Section 2. A presentation of some of the theoretical support both for and against managed care is provided in Section 3. A review of the literature is found in Section 4. The various empirical methodologies used in the analysis are explained in Section 5. Section 6 describes the construction of the data, describes the treatment and comparison groups, and the dependent variables, and lastly presents the summary statistics. The results are discussed in Sections 7 and Section 8 concludes the paper.
2.0The Major Managed Care Plans in California
Managed Care Expansion
Managed care has been legally available for Medi-Cal beneficiaries in California since the Knox-Keene Health Care Service Plan Act in 1975 that authorized the state to license HMOs or prepaid health plans. Until the 1990s, enrollment into a managed care plan in most counties was voluntary. Due to the mandates during the 1990s, presently, 22 counties enroll some or all of their Medi-Cal recipients in managed care programs (Table 2).
By January 2002, 57 percent of all Medi-Cal patients, or 2.97 million recipients, were enrolled in one of the three major types of managed care programs (Table 2). They are the:
- Two-Plan Model (TPC)
- County Organized Health System (COHS)
- Geographic Managed Care (GMC)
Of these three models, TPC enrolls the largest number of Medi-Cal recipients, 42 percent, while, COHS and GMC account for 9 percent and 6 percent of the beneficiaries respectively (Table 2).
COHS was the first of the three main managed care models to be enacted in California. In 1982, state legislation gave permission for three counties to create a COHS. Santa Barbara led the way in 1983, and became the first county to implement a major managed care plan for their Medi-Cal beneficiaries. They were soon followed by San Mateo in 1987, and Monterey a decade later in 1999. There are now seven COHSs covering eight counties in California (Table 2).
Under a COHS, there is one HMO that is responsible for providing services to all the Medi-Cal beneficiaries under its jurisdiction. This HMO is operated by the county but must be an independent public entity that meets the Knox-Keene Health Care Plan Act requirements for pre-paid health plans. However, they do not need a Knox-Keene license. Counties contract with the California Medi-Cal Assistance Commission and receive a prepaid capitated rate each month for each Medi-Cal recipient. Enrollment in a COHS is mandatory for the entire Medi-Cal population, and it is performed concurrently with enrollment into the Medi-Cal program.
GMC first began in Sacramento in 1994. To date, San Diego is the only other county using a GMC model. The GMC model permits six or seven commercial plans to operate within a designated region. Enrollment is mandatory for the CalWORKS-linked[3] (Welfare) population and is composed primarily of women and children. Enrollment into the managed care plan is not concurrent with Medi-Cal enrollment. Instead, Maximus, an outside organization, serves as the enrollment broker (McCall et al. 2001).
In 1995, 12 counties were designated to participate in the state’s expansion of TPC managed care. Under TPC, the state’s Department of Health Services contracts with a commercial managed care plan and a county-developed plan called the local initiative. The local initiatives are intended to preserve the network of traditional safety net providers. Like GMC, enrollment is mandatory for the CalWORKS-linked population, and Maximus serves as the enrollment broker (McCall et al. 2001) .
Setting of Capitation Rates
Capitation rates for the Two-Plan model are set by the state and are public information. Capitation rates for the other two models are negotiated by the California Medi-Cal Assistance Commission and are confidential. However, the Department of Health Services for California provides the California Medi-Cal Assistance Commission with maximum allowable capitation rates that can be set during negotiations (McCall et al. 2001).
Coverage of Services and Opt-Out Option
COHS generally covers a broader range of general health services than either GMC or TPC, and provides most of the services that were offered under the fee-for-service program. The Sacramento GMC plan is an exception. Dental coverage and mental health services, which are not included in any of the other models or plans, are covered by the Sacramento GMC, though mental health services are only provided by two of the seven commercial HMOs (McCall et al. 2001).
Furthermore, the TPC and GMC plans allow all pregnant women to obtain a medical exemption so they can choose to opt-out of the managed care plan and seek prenatal care and delivery services from a provider of their choice. Under the COHS plan, only non-welfare beneficiaries who are only covered for pregnancy, delivery and infant health care have the opt-out option.
3.0 The Theoretical Argument for the Impact of HMO Enrollment on Medicaid Expenditures, Quality and Access
There are a number of arguments both for and against the ability of managed care to control costs, increase access, and improve quality of care. In theory, managed care changes the incentives concerning the types of services provided. Payment based on fee-for-service encourages spending resources on acute care services, since a doctor is able to bill these services to the insurance company. Under a managed care system, the organization delivering the care, the HMO, receives a prepaid rate for each member enrolled in the plan and maintains some level of risk for providing all necessary services for enrolled members within that prepaid rate. The HMO is encouraged to keep patients healthy, and hence costs low, by managing a patient’s care and providing services that a fee-for-service insurance does not typically cover, such as offering preventive care, encouraging a healthier lifestyle, or beginning a smoking cessation program (Feldstien, 1999).