1 EXECUTIVE SUMMARY Pregnant Women and Substance Abuse: Testing Approaches to A Complex Problem June 1998 Embry M. Howell Craig Thornton Nancy Heiser Ian Hill Renee Schwalberg Beth Zimmerman Ira Chasnoff * *National Association for Families and Addiction Research and Education Health Systems Research, Inc.
2 PREGNANT WOMEN AND SUBSTANCE ABUSE: Testing Approaches to A Complex Problem When a pregnant woman abuses drugs or alcohol, both she and her unborn child may suffer harm. In addition, substance abuse often creates or is accompanied by an array of social problems for the abuser and those around her, including violence, child abuse and neglect, and family dysfunction. Harm to children born of substance-abusing mothers can be far-reaching, affecting both their physical and cognitive growth and development. The need to study the phenomenon of substance abuse by expectant mothers is vital in terms of the health and welfare of these women and their children. The social and economic implications for society as a whole are also important policy concerns. Outreach--specifically, identifying and recruiting pregnant substance abusers--was difficult. FINDINGS IN BRIEF The Evaluation of the Demonstrations to Improve Access to Care for Pregnant Substance Abusers, funded by the Health Care Financing Administration and conducted by Mathematica Policy Research, Inc., was designed to address the problem of substance abuse by pregnant women. The goals of this demonstration were to identify pregnant women who abuse drugs and alcohol; provide them with prenatal care, substance abuse treatment, and support services; and improve the health and well-being of both mothers and their infants. The demonstrations aimed also to reduce costs borne by the Medicaid program. Mathematica studied five diverse state demonstration programs that helped women get integrated prenatal care and substance abuse treatment by providing outreach, screening, expanded coverage for treatment, and other support services, such as case management. Significant findings associated with the demonstration programs, which ran from mid-1993 to mid-1996, include: C Outreach--specifically, identifying and recruiting pregnant substance abusers--was difficult, largely because members of this population were reluctant to be identified. They may have been reluctant because they wanted to perpetuate substance use or feared social stigma, prosecution, or losing their children to child protective services interventions. C Retention of participants was challenging. 1
3 C Costs associated with serving this population were high and about twice those for other pregnant women enrolled in Medicaid. C Program participants who received higher levels of treatment had higher birthweight infants, compared to women who received minimal treatment, suggesting that birth outcomes can be improved for some pregnant substance abusers. THE PROGRAMS States that were chosen to participate in the demonstration were selected through a competitive grant solicitation on the basis of their innovative programs and willingness to collect data for the evaluation. The demonstration programs, which supplemented existing state services, were as follows: C Maryland s Better Chance program, centered at the Johns Hopkins obstetrical clinic in Baltimore, used outreach, support groups, and case management services. Costs associated with serving this population were high... C Massachusetts Medicaid Opportunities to Help Enter Recovery Services (MOTHERS) program operated statewide. The 1 state obtained an IMD waiver, allowing the demonstration to provide residential substance abuse treatment under Medicaid. The demonstration included an in-depth study of the state s existing treatment system. C New York s demonstration operated in six sites: three in New York City and three in upstate New York. The program offered enhanced substance abuse treatment and outreach. Treatment in residential facilities was offered under an IMD waiver. C South Carolina s Transitions program operated in three primarily rural counties. The demonstration included broad-based outreach, case management, and expanded substance abuse treatment. C Washington s First Steps PLUS program in Yakima County had a broad-based outreach component, training for prenatal care providers, standard screening to identify pregnant substance abusers, parenting education, and case management. Residential substance abuse treatment was also provided under an IMD waiver. 1 An IMD waiver allows a state to cover treatment in an Institution for Mental Diseases (IMD), a category of residential substance abuse treatment that is generally excluded from Medicaid. 2
4 Outreach and Screening... all the demonstrations developed one or more approaches to outreach... Although all the demonstrations developed one or more approaches to outreach, identifying and recruiting pregnant substance abusers remained challenging. The success of some strategies, such as media campaigns and community-based outreach, was limited, largely because women were reluctant to be identified. South Carolina and Washington developed screening instruments and then trained prenatal care and social service providers to utilize them systematically and routinely to identify pregnant substance abusers. Enrollment rates were the highest in these two demonstration programs, which both used broad-based outreach and routine screening. Maryland, Washington, and South Carolina sponsored an array of other activities, generally in prenatal care settings, that helped pregnant substance abusers receive vital, needed services. Maryland provided support groups, Washington offered parenting education, and all three demonstrations provided case management. In the end, between 10 and 50 percent of pregnant substance abusers were identified. Clients from the comparison areas as well as participants in the demonstrations were older and had more children, on average, than other pregnant women on Medicaid. Treatment Services Substance abuse treatment varied from state to state, although none of the demonstrations developed entirely new programs. New York, South Carolina, and Washington modified and enhanced existing treatment programs to fit the needs of pregnant women. This process took a lot of time and did not always go smoothly, since most established programs were not designed to serve this population. Both South Carolina and Washington altered short-term residential programs that had previously targeted a non-medicaid population. New York adapted its treatment programs to include a range of support services exclusively for pregnant women. Massachusetts obtained a waiver to cover its established programs, which had already been modified to serve pregnant women. In contrast, Maryland did not offer formal substance abuse treatment as part of its demonstration. 3
5 THE RESULTS The outcome analysis showed that pregnant substance abusers were difficult to identify and recruit into treatment programs. Furthermore, they were difficult to retain once recruited. Their birth outcomes remained extremely poor, and costs remained about twice those of other pregnant women on Medicaid. While the outcome analysis was based on limited data and a nonexperimental design, it did not show that the demonstration programs increased the number of clients receiving prenatal care or substance abuse treatment, or that services led to higher birthweights or lower program expenditures. Those who received intensive treatment (higher levels of and greater retention in treatment) did have higher birthweight infants, compared with those who received minimal treatment. Since both groups had a need for treatment, the results suggest that birth outcomes can be improved for some pregnant substance abusers.... the results suggest that birth outcomes can be improved for some pregnant substance abusers. A client-level analysis revealed some significant differences in outcomes for demonstration clients and other pregnant substance abusers in demonstration areas. Because women chose to enroll in demonstration programs, however, and were not randomly assigned, the differences could have resulted from self-selection. In contrast, an areawide outcome analysis not subject to these selection problems showed no significant differences in key outcomes--use of prenatal care, substance abuse treatment, birthweight, and total Medicaid costs--in demonstration areas, relative to trends in comparable areas where the demonstrations were not operating. Many of the outcomes we observed cannot be attributed directly to the demonstrations because a randomized design was not used. With this caveat in mind, we tentatively conclude that intensive treatment appeared to be potentially effective for some women, although the demonstrations were not able to achieve the very ambitious goal of improving outcomes for all pregnant substance abusers. RECOMMENDATIONS FOR THE FUTURE The evaluation provides insights into the challenges of identifying, recruiting, and ultimately serving the needs of a very high-risk group within the Medicaid population. It also dramatically underscores the need for sustained efforts to address pregnant substance abusers needs. 4
6 More rigorous studies and an improved study design will contribute to the search for solutions to this vexing problem. In addition, the lessons learned from this evaluation can be used to develop a model demonstration program that would include: More rigorous studies and an improved study design will contribute to the search for solutions to this vexing problem. C Links between Medicaid, substance abuse, and health agencies at the state level, and between prenatal care and substance abuse treatment agencies at the local level C Implementation of standardized screening protocols, accompanied by uniform training in how to screen, given to traditional and nontraditional providers and agencies where substance abusers often turn for help C On-call outreach workers, trained in substance abuse counseling, available for referral and emergency response C A continuum of care, including prenatal care, detoxification, intensive substance abuse treatment (either residential or outpatient, with follow-up outpatient care for a minimum of three months), and routine provision of key support services, such as case management and child care, to increase retention in the program C An evaluation design that would involve random assignment to either intensive outpatient or residential care after detoxification, with Medicaid IMD waivers for those receiving residential services RESEARCH METHODOLOGY Mathematica conducted the national evaluation with its subcontractors, Health Systems Research, Inc., (HSR) and the National Association for Families and Addiction Research and Education, Inc. (NAFARE). HSR conducted annual site visits to the demonstrations, while NAFARE conducted focus groups with pregnant substance abusers and providers at the demonstration sites. The cross-site evaluation used qualitative and quantitative methods to study implementation, identify those served by the demonstrations, and analyze the outcomes for demonstration clients and other pregnant substance abusers in demonstration and comparison areas. Data for the evaluation came from multiple sources, including site visits, focus groups, state Medicaid claim files, state substance abuse treatment system files, birth certificates, surveys (in two states), and some limited information collected by individual demonstration programs. 5
7 ABOUT THE AUTHORS Embry Howell is a vice president at Mathematica with expertise in maternal and child health, Medicaid policy, and other health care issues. Craig Thornton is a senior fellow at Mathematica. He focuses on disability policy, evaluation design, and the impact of managed care on people with disabilities enrolled in Medicare and Medicaid. He also studies substance abuse and mental health issues. Nancy Heiser is a research analyst at Mathematica whose research focuses on maternal and child health and the impact of managed care on vulnerable populations. Ira Chasnoff is president of the National Association for Families and Addiction Research and Education. For the past 20 years, he has conducted research on the impact of substance abuse on pregnancy and the long-term outcomes of exposed children. Ian Hill is associate director at Health Systems Research, Inc., and is responsible for maternal and child health projects. Renee Schwalberg is a senior associate at Health Systems Research, Inc., focusing on managed care and vulnerable populations. Beth Zimmerman is a policy associate at Health Systems Research, Inc., focusing on maternal and child health. HOW TO GET THE FULL REPORT For a copy of the full report, please send a check for $16.25 to Jan Watterworth, Librarian, Mathematica Policy Research, Inc., P.O. Box 2393, Princeton, NJ , (609) Please ask for publication PR For more information about this study or Mathematica s other health research, please contact Embry Howell, vice president, (202) , Information is also available on our web site at 6
8 ABOUT MATHEMATICA POLICY RESEARCH Mathematica Policy Research, Inc., is a leader in policy research and analysis. The employee-owned firm, with offices in Princeton, NJ, Cambridge, MA, and Washington, DC, has conducted some of the most important evaluations of health care, education, nutrition, welfare, employment, and early childhood policies and programs in the United States. Its studies provide a sound foundation for decisions that affect the well-being of Americans. Mathematica is a registered trademark of Mathematica Policy Research, Inc. The views presented here are those of the authors and not neces- sarily those of the Health Care Financing Administration. 7