The Maternal and Child Health Bureau (MCHB) is directing significant attention to advancing and strengthening essential public health functions, and to assisting State Title V and community MCH programs, with enhancement of their analytic capability in the assessment of needs, monitoring of efforts and evaluating outcome performance. One major issue-related to this analytic activity is the need to improve the understanding of the economics of MCH services. This need for economic analysis includes: cost effectiveness and benefits including measures and assessment of quality of care, intervention production functions, modeling provider/user behaviors, health care financing related to access and/or health outcomes, assessment of alternative disease management scenarios, value of cost and benefit information, and ethical issues resulting from explicit policies of resource allocation. Despite increased publication of literature and interest in these fields of economics, there has been little development of that literature relevant to the MCH community.

The overall mission of the grants is to increase the maternal and child health economic knowledge base, to translate MCH knowledge into action, and apply economic research methods for the development of effective and efficient interventions that can be rigorously tested and validated as program solutions to MCH problems in the community.

In order to develop and promote economic analyses as a part of health services planning and program operations, MCHB has awarded for supplemental grants to MCHB grantees in schools of public health. These supplemental grants are intended to attract MCH doctoral students to the field of health economics and to help support their scholarship and acquisition of the requisite knowledge base and skills.

Following is a list of the completed and ongoing projects. Abstracts for each study are attached.

Title XXI of the Social Security Act, the State Children's Health Insurance Program (CHIP), made federal funds available to states as of October 1997 to expand health insurance coverage for children under age 19 with family incomes up to 200% of the federal poverty level, or up to 50 percentage points above the state's current Medicaid level. This research has two main purposes: (1) to describe the socioeconomic and health status characteristics of the uninsured, CHIP-eligible population in the United States and Illinois; and (2) to simulate expected changes, along with the costs and benefits of those changes, in health care access, utilization, and selected outcomes based on enrollment in the CHIP program. For the simulation, estimates of behavioral parameters based on national data will be applied to Illinois population characteristics. Expected health outcomes will be tied to selected Healthy People 2000 objectives. The uninsured, CHIP eligible population will be compared to three other populations the Medicaid enrolled, the privately insured, and the privately insured/same income category, since each of these populations have different characteristics and may behave differently. Estimates from the three groups will used to develop ranges for what may be the expected changes for the target population. The National Health Interview Survey (NHIS) from the last two quarters of 1993 and all of 1994 will be the data source utilized for this research. This research is of considerable current policy significance. Illinois, and most states around the country, are in the process of designing and implementing their CHIP programs. While there have been many studies of the Medicaid-eligible population, there is little information as to what states can expect from the targeted population in terms of needs, enrollment and utilization. This research will provide more information on the characteristics of the CHIP-eligible population and their expected utilization patterns. In addition, by examining the expected health benefits of CHIP coverage, policy makers can justify the additional expenditures of extending health insurance coverage to children.

Health care providers regularly encounter families who are unable or unaware of how to follow through on the necessary health care, preventive and therapeutic, that their children need for optimal health. This lack of coordination within a health care clinic can overwhelm and confuse many families. Case management is a service coordination strategy and organizing technique for increasing the provision of comprehensive community-based primary care.

This project will assess the cost effectiveness of case management in increasing the compliance of a family with recommended well child exams during the first year of an infants life. The study design is a nonequivalent comparison group. A retrospective chart review will be used to obtain the data. The intervention sample will include 100 newborns who receive their well child care at an inner city hospital in Chicago and whose families received the services of a case manager during the infant's first year of life. The comparison sample will be 100 newborns who receive their well child care at a neighborhood health center in Chicago in the same community as the inner city hospital, and whose families did not receive the services of a case manager.

Los Angeles (L. A.) County provides an excellent location for examining the problems in recruiting and retaining foster families because it is a large metropolitan area whose situation mirrors the nation's. As of August 31, 1998, L. A. County had 48,183 children in foster care, almost 50% of all the children in foster care in the state of California. In addition, L. A. County is especially interesting because both the public and private sectors recruit foster families. The California State Department of Social Services and the L. A. County Department of Children and Family Services are responsible for recruiting and licensing foster families in the county. In addition, there are about 55 private, non-profit foster family agencies (FFAs) who are licensed by the state and the county to recruit and certify foster families.

The aim of the this research project is to examine how foster families are recruited and retained in L. A. County to gain insight into ways to increase the number of foster families in a cost-effective manner. The main research question is the following: Does the public sector or the private sector do a better job of recruiting and retaining foster families, why, and at what cost? The policy implication of this question is that if the private sector can do a better job (lower cost, higher quality, better retention), we may want to look for ways to expand its involvement and to shrink or eliminate public efforts. The first step in answering such a question is to have a good understanding of how the two sectors are organized and financed. Thus, this research examines how the two sectors differ in the following domains: advertising, recruiting, retention, training, payment rates, allowances, support services, administrative costs, and funding. In addition, it is crucial to know if FFAs serve the same types of children as the public agency in terms of ethnicity/race, gender, age, and emotional, mental, or physical disabilities since costs differ across types of children.

The primary method is to survey child welfare administrators from the public and private sectors in L. A. County about their foster family recruitment and retention efforts and the about the kinds of children they serve. A questionnaire will be piloted to a select number of FFAs and then, after revisions, sent to all FFAs and L. A. County's Department of Children and Family Services. The survey will be administered during an in-person or telephone interview. In addition, the literature on foster parent recruitment, retention, quality, and economic incentives will be reviewed. Finally, data will be collected from the state and county on foster children, foster payment rates and allowances, and funding trends.

Additional funding for this research is from the John Randolph Haynes and Dora Haynes Foundation Faculty Fellowship.

The growth of managed care in the United States has increased price competition among community hospitals. Children's hospitals have specialized in the provision of inpatient pediatric services. Many of these hospitals claim to cross subsidize tertiary care by charging higher prices for less sophisticated services. It is hypothesized here that markets with greater managed care penetration will have more pediatric admissions in non-children's hospitals.

Hospital data for the study are drawn from the 1992-1994 AHCPR National Hospital Inpatient Sample, managed care penetration from MedStar Pulse Survey, and market characteristics from the Area Resource file. Analysis is limited to community hospitals located in metropolitan markets that contain a children's hospital and that are included in the AHCPR and MedStar samples. Pooled time-series cross section equations are estimated for aggregate, "routine" and "high tech" admissions. Routine and high tech were defined by DRG weight.

Controlling for other factors, 10 percentage point increase in managed care penetration increased total pediatric admissions at community hospitals by 305 admissions, routine admissions by 141 and high tech admissions by 164. These findings were all statistically significant at the 99 percent confidence level.

The findings suggest that children’s hospitals face an erosion of their speciality niche as managed care increases hospital competition. However, this erosion is not limited to "routine" pediatric admissions but affects more sophisticated services as well.

With-over 4 million births per year, the area of obstetrical care demands meaningful health indicators that measure quality of care, health status and economic outcomes. This thesis is designed to identify determinants that may effect maternity length of stay and to model variations based on patient, physician, and hospital characteristics. The study will examine whether hospital-to-hospital variability in length of stay in the state of Arizona is related to interhospital differences in patient, physician, and hospital characteristics. Describing variation should assist decision-making facing increased managed care penetration, Medicaid expansions, shrinking roles of state health departments, risk-sharing arrangements and quality of care. These unfolding phenomenon are slowly merging public and private initiatives and Maternal and Child Health state agencies must develop capacities to measure health service performance for assurance, assessment and policy development. The benefits from a population-based perspective might lead to public health interventions and policy-making initiatives directed at differences in patient outcomes, physician practices and hospital variations.

The study design includes four sources of data: 1996 Arizona birth certificate data, 1996 Arizona hospital discharge summary data, 1996 Arizona Medical Association data, and 1996 American Hospital Association data. Eligible cases include patients with diagnosis-related groups 370 through 375 (vaginal or cesarean delivery) from January 1, 1996 to December 31, 1996, reported on the 1996 Arizona hospital discharge summary data set. The dependent variable is maternity length of stay as measured by the number of days in the hospital per admission. The independent variables include maternal, physician and hospital characteristics.

The statistical methodology used for the analyses includes a hierarchical linear model (HLM). As compared to linear regression models, HLM models are preferred because they model clustering characteristics of data, they provide statistically efficient estimators, and this method allows one to estimate interhospital variation and intrahospital effects.

Systemic approaches to improve the delivery of prenatal care for low-income and Medicaid eligible women have found considerable support in the last decade. The purpose of this study is to evaluate the effect of providing ancillary prenatal care services to a public health department service population, analyzing race-specific models of WIC participation and risk of small-for-gestational-age (SGA,) at term.

There were 13,095 singleton deliveries during the period 1987-1990 to women with prenatal care in a managed maternity care program. The research design entailed comparison of the intervention group (those receiving comprehensive prenatal care in 1989-1990) with an historical, control group of women who received prenatal care in the two years (1987-88) preceding the intervention.

For each intervention group compared to their control group, the odds of nonwhite women participating in WIC were 1.7 times greater while white women were over twice as likely to participate in WIC. The impact of ancillary services on term SGA births indicated a protective odds ratio of 0.851 for nonwhite women. Results for white women were not significant.

These findings suggest that for those receiving ancillary services, WIC appears more important for white women than nonwhite, and services are associated with lower risk of term SGA births for nonwhite women but not for white women. Results add to growing evidence regarding the efficacy of ancillary services.

To 1.) identify and describe four different groups of children with special health care needs (CSHCN) using definitions which represent various groups of special needs children that are currently or who have been recommended to be carved-out of participation in managed care plans. 2.) examine how total, mean and median annual health care expenditures differ based on the definition for CSHCN used. 3.) assess the capacity of several risk adjusters including age, gender, CSHCN group to which children belong, and the Ambulatory Diagnostic Groups (ADGs) to predict annual expenditures for healthy children and for the groups of CSHCN.

This study analyzed Alabama Medicaid claims data for children between the ages of one and 17 who were continuously enrolled in Alabama’s Medicaid program throughout calendar years 1994 and 1995.

The prevalence of CSHCN varied from 1% to 13% depending on the definition for CSHCN used. Combining the four groups yielded an estimate of 23% of the children in this study with special health care needs. Fifty-five percent of the expenditures for the children in this study were for CSHCN. However, some groups of CSCHN were found to have greater total, mean and median expenditures than the others. Generally, age and gender were able to explain very little of the variation in expenditures for each of the four groups of children without special health care needs.(.17%.-.65%) and for each of the groups of CSHCN (.04%-1.5%). Age, gender, and the 34 ADG categories were able to explain between 6.02% and 8.72% of the variation in expenditures for each of the four groups of children without special health care needs and between 8.5% and 10.9% of the variation in expenditures for each of the four groups of CSHSN. Overall, these variables could explain 3.16% of the variation in expenditures for all children without special health care needs and 8.4% of the variation in expenditures for all children with special health care needs. Therefore, when children with special health needs are combined with those who do not have special health care needs, the percentage of the variation in expenditures explained rose from 3.16% to 8.8%. Prediction estimates improved considerably when expenditures were truncated at $25,000.

Prevalence and expenditure estimates for CSHCN vary depending on the definition used. Whatever the definition, one can better risk adjust for CSHCN then for children without special health care needs using standard methods. Thus, concerns regarding the lack of predictability of expenditures for children with special health care needs that leads to carving them out of managed care plans are unfounded in this study.

While much has been made of the effect "managed care" has had on the medical care sector, little mention has been made of effects on other sectors of the economy. One sector in which managed care should have strong repercussions is the pharmaceutical industry. The pharmaceutical industry, while multinational, has a large stake in the U.S. medical care sector. The emphasis on cost savings under managed care has direct implications for the pharmaceutical industry; a major cost in the health care system is drugs and its accompanying technologies.

Economists (Weisbrod 1991) have previously described the interactive relationship between health care and the R&D field, and have hypothesized that implementation of cost controls would lead to cheaper medical technologies (including drugs). This proposal examines the hypothesized relationship. However, this proposal hypothesizes that the pharmaceutical industry response to the stressor of health care expenditure control will not be marked by the development of drugs associated with less expenses, but by an increase in its current business strategy (i.e. standard operating procedure). The pharmaceutical industry has been able to turn a profit based on monopolies through patents and by diversifying its risks through a shotgun approach to drug development. It is hypothesized that the pharmaceutical industry will attempt to increase its monopolies through relative increases in patent times and by pushing more drugs through the system in hopes of finding a multimillion dollar seller.

It is hypothesized that these are these forces have driven regulatory changes in the Food and Drug Administration, which is the governmental agency that provides oversight on new drugs. It is hypothesized that regulatory changes allowing speedier approval for new drugs will do so at the expense of more post-approval problems. These problems can be identified through the FDA's Spontaneous Reporting System. The FDA's Spontaneous Reporting System has collected reports regarding adverse drug events since 1969. This data set includes information about drugs associated with adverse drug events, the actual event itself, and the patient involved. Preliminary analysis shows that recent regulatory changes have had a significant effect on the number of adverse drug events reported to the FDA.

This proposal will examine data from the FDA's Spontaneous Reporting System and determine through econometric techniques whether significant relations exist between adverse drug event reporting and the rise of managed care. Key to this proposal is the use of children as a "sentinel' category of patients. Drug testing in children is considered inadequate; a recent FDA rule change has instituted increased testing of drugs for this age category. It is hypothesized that increases in adverse drug events are more likely in children since little is known about the effects drugs will have on children. Children are also less likely to have confounding factors, such as the multiple drug regimens of the elderly.

A final proposed chapter is the significance of the rise in adverse drug events. While direct costs of adverse drug reactions have been examined in the literature, this proposal may expand on these. In addition, this proposal will describe indirect costs: losses of public confidence and trust in the medical care systems. With changes in managed care, public trust in medical care institutions has already been decreased. However, the current literature does not comment on the additional factor of public confidence. Losses in trust compounded with losses in confidence could have severe consequences for the medical care sector; increased monitoring and regulation may accompany loss of trust but loss of confidence may result in individuals foregoing the benefits of current medical

My post-doctoral work will address economic issues and public health problems surrounding access to reproductive technologies. I plan to focus on the issue of access to abortion services among lower-income women, and the costs associated with differential access across income groups and geographic location. Specifically I propose to examine the price elasticity of abortion use among women in different socioeconomic groups; and the costs to the public sector associated with differential access to state funded abortion services.

Several existing studies support the claim that the availability of state subsidized abortions has a sizable effect on the number of poor women who have abortions. Rebecca Blank, for example, finds that 19% - 25% of the abortions among low-income women which are publicly funded do not take place after funding is eliminated. Research on the health outcomes of abortion availability show that, in addition to reducing risk to the pregnant woman by removing the risks of child-birth, abortion availability is correlated with reductions in the percentage of low-birth-weight and pre-term babies born, and reductions in the infant mortality rate. These improvements in birth-outcomes are thought to be due to the decline in the number of unwanted births when abortion is available. I plan to incorporate these findings into economic estimates of the costs to states of providing or withholding abortion services from low-income women.

In addition, I plan to work with Professor Malcolm Potts to survey existing research on alternative (low cost) abortion technologies, such as manual vacuum aspiration and non-surgical procedures. We plan to construct estimates of the costs and benefits associated with these technologies compared to traditional surgical abortion; and we will construct estimates of the extent to which increased use of these technologies could increase access to abortion in under-served regions of the United States.

Maternity care guidelines established by the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) highlight the importance of family, environmental, and social risk factor assessment during the delivery hospitalization and caution against early postpartum discharge for adolescents or women at substantial social risk. Pressures by managed care plans and other third party payors to reduce lengths-of-stay may cause physicians to overlook these criteria., resulting in early discharge before medical or social risks have been resolved. This study explores the association of individual and regional factors, and the penetration of managed care with postpartum length-ofstay in a representative sample of low-income women.

Data were obtained from a postpartum survey conducted for the evaluation of Healthy Start (HS), a national, community-based initiative to reduce infant mortality in 22 U.S. sites. Women who were < 6 months postpartum were sampled from WIC clinic attendees in each HS area. Statistical analysis relied on chi-square testsand multivariate logistic regression.

Of 2,671 respondents, 15 percent we discharged <24 hours after a vaginal delivery or <72 hours after a cesarean section. Black race, unmarried status, smoking during pregnancy, and absence of data on household income were associated with a lower probability of early discharge, while women with less than adequate prenatal care were at greater risk for having a short postpartum hospital stay. Adolescents were as likely as older women to be discharged early. A high state Medicaid managed care penetration rate lowered the probability of early discharge, though the effect was diminished with multivariate adjustment.

Appropriate safeguards are needed to ensure that disadvantaged women receive adequate postpartum health education prior to hospital discharge, and that home visits are provided for mothers and newborns with short postpartum stays.

Comprehensive health education and home visit follow-up are important components of early postpartum discharge practices. Shortened hospital stays after delivery limit the time available for offering advice on infant and self-care topics to new mothers, and routine postpartum home visits may not be provided by health plans in the absence of specific complications. This study explores the association of early postpartum discharge with the receipt of postpartum health advice and the receipt of home visits in a representative sample of low-income women.

Data were obtained from a postpartum survey conducted for the evaluation of Healthy Start (HS), a national, community-based initiative to reduce infant mortality in 22 U.S. sites. Women who were <6 months postpartum were sampled from WIC clinic attendees in each HS service area. Statistical analyses relied on chi-square tests and multivariate logistic regression.

Of 2,671 respondents, 15 percent were discharged <24 hours after a vaginal delivery or <72 hours after a cesarean section. Forty percent of women reported receiving postpartum advice on all eleven reported infant and self-care topics, and there was no difference by hospital length-of-stay following delivery. Women discharged early were significantly less likely to receive one or more postpartum home visits than mothers with longer hospital stays, and this relationship persisted after multivariate adjustment.

These findings suggest that health care providers used the early postpartum period efficiently interns of relaying essential information to new mothers in this sample. However, women discharged early did not receive additional follow-up that may be necessary to ensure a safe transition from hospital to home soon after delivery.

These findings suggest that health care providers are mindful of some of the recommended factors when considering early discharge, but not all. Individual factors and regional norms are currently more important determinants of postpartum length-of-stay than residence in a state with a high penetration of managed care. Adherence to criteria for early postpartum discharge of disadvantaged women could be improved. The increase in managed care among low-income populations does not appear to decrease compliance with professional guidelines.

Despite the widespread use of early postpartum discharge, concern remains about the safety and efficacy of this practice for socially disadvantaged women in particular. Previous research primarily has focused on the association of early postpartum discharge with maternal and neonatal rehospitalization rates, while little is known about the impact of short obstetric stays on preventive health behaviors and the utilization of preventive health services. These outcomes may be better indicators of missed opportunities for comprehensive teaching and the coordination of follow-up services caused by shortened lengths-of-stay. This study used a representative sample of low-income women to examine the association of early postpartum discharge with the initiation and duration of breast feeding, postpartum birth control use, receipt of a routine postpartum check-up, receipt of well-baby care, and Hepatitis B vaccine administration.

Data were obtained from a postpartum survey conducted for the evaluation of Healthy Start (HS), a national, community-based initiative to reduce infant mortality in 22 U.S. sites. Women who were <6 months postpartum were sampled from WIC clinic attendees in each HS area. Statistical analysis relied on chi-square tests and multivariate logistic and Cox regression.

Of 2,671 respondents, 15 percent were discharged <24 hours after a vaginal delivery or <72 hours after a cesarean section. After multivariate adjustment, infants born to women discharged early were less likely to receive a Hepatitis vaccine or a well-baby visit, though the latter was true only among respondents who received postpartum home visits. In contrast, early postpartum discharge was positively associated with the receipt of a routine postpartum checkup. No relationship between obstetric length-of-stay and birth control use or breast feeding practices was found in this sample.

These findings suggest that early postpartum discharge may decrease the utilization of infant preventive health services, namely Hepatitis B immunization and the receipt of well-baby care. The impact of shortened length-of-stay did not moderate after adjustment for postpartum teaching, the receipt of home visits, or the coordination of follow-up appointments. More intensive services may be warranted to ensure the appropriate utilization of infant preventive health care in conjunction with early discharge practices.

Marie C. McCorrnick, M.D., Sc.D.
Barbara L. Devaney, Ph.D.
Joanna E. Siegel, Sc.D.
Marcello Pagano, Ph.D.

Maternity care guidelines established by the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) highlight the importance of family, environmental, and social risk factor assessment during delivery hospitalization and caution against early postpartum discharge for adolescents or women at substantial social risk. Pressures by managed care plans and other third party payers to reduce lengths-of-stay may cause physicians to overlook these criteria, resulting in early discharge before medical or social risks have been resolved.

This study explores the association of individual and regional factors, and the penetration of managed care with postpartum length-of-stay in a representative sample of low-income women.

Data were obtained from a postpartum survey of WIC clinic attendees who had an infant <6 months of age in 15 U.S. sites. Statistical analysis relied on chi-square tests and multivariate logistic regression.

Of 2,671 respondents, 15 percent met the criteria for early postpartum discharge. Black race, unmarried status, smoking during pregnancy, and absence of data on household income were associated with a lower probability of early discharge, while women with less than adequate prenatal care were at greater risk for having a short postpartum hospital stay. Adolescents were as likely as older women to be discharged early. A high state Medicaid managed care penetration rate lowered the probability of early discharge, though the effect was diminished with multivariate adjustment.

These findings suggest that health care providers are mindful of some of the recommended factors when considering early discharge, but not all. Individual factors and regional norms are currently more important determinants of postpartum length-of-stay than residence in a state with a high penetration of managed care.

Despite the frequency of early postpartum discharge, concern remains about the safety and efficacy of this practice, particularly for socially disadvantaged women.

This study examined the association of early postpartum discharge with preventive health behaviors and the utilization of preventive health services in a representative sample of low-income women.

Data were obtained from a postpartum survey of WIC clinic attendees who were <6 months postpartum and lived in 15 U.S. communities. Statistical analysis relied on chi-square tests and multivariate logistic and Cox regression.

Outcomes examined were the initiation and duration of breast feeding, postpartum birth control use, receipt of a routine postpartum check-up, receipt of well-baby care, and Hepatitis B vaccine administration.

Of 2,671 respondents, 15 percent were discharged <24 hours after a vaginal delivery or <72 hours after a cesarean section. After multivariate adjustment, infants born to women discharged early were less likely to receive a Hepatitis vaccine or a well-baby visit, though the latter was true only among respondents who received postpartum home visits. Early postpartum discharge was positively associated with the receipt of a routine postpartum check-up. No relationship between obstetric length-of-stay and birth control use or breast feeding practices was found in this sample.

Early postpartum discharge may decrease the utilization of infant preventive health services, namely Hepatitis B immunization and the receipt of well-baby care. The impact of shortened length-of-stay did not moderate after adjustment for postpartum teaching, the receipt of home visits, or the coordination of follow-up appointments. More intensive services may be warranted to ensure the appropriate utilization of infant preventive health care in conjunction with early discharge practices.

Comprehensive health education and home visit follow-up are important components of early postpartum discharge practices. Shortened hospital stays after delivery limit the time available for offering advice on infant and self-care topics to new mothers, and routine postpartum home visits by health plans may not be provided in the absence of specific complications.

This study explores the association of early postpartum hospital discharge with the receipt of postpartum health advice and home visits in a representative sample of low-income women.

Data were obtained from a postpartum survey of WIC clinic attendees who were <6 months postpartum and lived in 15 U.S. communities. Statistical analysis relied on chi-square tests and multivariate logistic regression.

Of 2,671 respondents, 15 percent met the criteria for early postpartum discharge. Forty percent of women reported receiving postpartum advice on all 11 reported infant and self-care topics, and there was no difference by timing of hospital discharge. Women discharged early were significantly less likely to receive one or more postpartum home visits than mothers with longer hospital stays. This relationship persisted after multivariate adjustment, indicating that differences in the prevalence of home visits between the two groups could not be explained by more frequent early discharge among lower risk women who may benefit less from follow-up services.

These findings suggest that health care providers used the early postpartum period efficiently in terms of relaying essential information to new mothers in this sample. However, more than half of the women did not receive teaching on at least some topics, suggesting that appropriate safeguards are needed to ensure that disadvantaged women receive adequate postpartum health education prior to hospital discharge. Women discharged early also did not receive additional follow-up that may be necessary to ensure a safe transition from hospital to home soon after delivery.

On August 5, 1997 Congress passed the most significant funding increase for children's health care coverage since the original enactment of Medicaid, the State Children's Health Insurance Program (CHIP). As states decide how to use CHIP funds, there is an urgent need for new data about unmet health need, access to care, and crowd out among the population of children who are targeted under the legislation. Massachusetts is one of the few states in the nation that has had extensive prior experience providing health services to the population of children that are covered under CHIP expansions. The state's Children's Medical Security Plan (CMSP), initiated in July, 1994, has provided health insurance coverage for primary health care services to over 80,000 children from birth to age 18, 40% of whom live in families with incomes 133%-200% of the federal poverty level (FPL), the income group targeted under the expansions. Using data from CMSP administrative and claims files and information obtained from a telephone survey of CMSP participants, the proposed research aims to: 1) to determine the relationship between insurance status, unmet health need, and health service utilization among CMSP enrollees; 2) to assess the impact of enrollment in CMSP on indicators of access to health care; and 3) to determine the extent to which crowd out is occurring among CMSP enrollees. Survey findings and the linkage of survey data with claims data are expected to assist in estimating health needs and costs of children likely to enroll in CHIP programs. Survey questions related to unmet health need and access to care are modeled after those from the NHIS Access to Care Module. Questions related to family demographics and insurance status based on the 1998 random household survey of Massachusetts' residents insurance status to enable comparison to the state's population. The study population is currently enrolled children who have been continuously enrolled for a minimum of 6 months. A sample of 900 program participants stratified by income group (<133, 133-200, 200% FPL) and age and systematically selected by date of enrollment will yield estimates with sampling errors of /- 3% at a 95% CI for major research questions. Interviews will be conducted in Spanish and Portuguese among respondents more comfortable in their native language. An additional over sample of 100 Spanish speaking respondents is proposed to permit stable inferences specific to the Spanish speaking enrollees. Using the McNemar test to assess within subject changes related to unmet health need and access indicators pre and post program enrollment, the study design has >85% power to detect changes. Multiple and multinomial logistic regression models will be used to predict relationships between major outcomes (crowd out and health service utilization) and primary question variables (length of time without insurance, access indicators, unmet health need, and family demographics).

A body of research suggests that school food service and nutrition instruction can play a significant role in affecting children's diets. States, districts, and schools have adopted a variety of policies and practices in attempt to increase the healthfulness of these school nutrition programs. However, whether these policies and practices are associated with more healthful school environments has not undergone extensive evaluation, until now.

This analysis identifies the state, district, and school policies and characteristics associated with a healthful school nutrition environment. Using secondary data from the 1994 School Health Policies and Programs Study, scales rating schools based on the healthfulness of food service and nutrition instruction practices were developed. Scale scores were used as the dependent variables in a series of regression models.

Different types of variables were found to be associated with more healthful school food service practices, including policies and support from higher levels of government, and characteristics of school personnel. State and district policies that command or change incentives to induce the inclusion of nutrition in curricula positively affect school commitment to nutrition instruction exposure.

The results of this analysis are used to recommend state, district, and school policies to improve school nutrition programs. Policies suggested redress the externality and information asymmetry in the market for healthful school nutrition.

Ms. Clark is engaged in three research projects. The first project involves a cost-benefit analysis of the Violence Against Women Act of 1994. Using Justice Department data, costs were determined for direct property losses, medical and mental health care, police response, victim services, lost workdays or school days, lost housework, pain and suffering/quality of life, loss of affection/enjoyment and death. Benefits were comprised of averted costs as a result of the Act. Comparing the rates before and after the law, this analysis found that the Violence Against Women Act saved $14.2 billion in averted social costs. The second project is examining more specifically the impact of partner violence on the cost of delivery services. Approximately 3,000 women with prenatal screening about domestic violence and birth certificate data are being matched to Medicaid data. During the current pregnancy, 3% of women reported physical violence. The most frequent perpetrators of violence were the women's current husbands or current boyfriends (58%), followed by the women's ex-husbands or ex-boyfriends (28%). The cost analysis is underway.

The third project, the basis for her dissertation, is designed to determine the societal cost of partner violence during pregnancy. Part of a larger study involving primary data collection on transitions in pregnancy, this project will investigate the costs of services for women experiencing partner violence, extending beyond the cost of delivery services examined in the second project .

Ms. Navaie-Waliser has undertaken a cost-effectiveness study of care coordination and home visitation for Medicaid-eligible women and infants in North Carolina. Using linked birth and death certificates and claims data, logistic regression was used to analyze costs and infant survival in 19,066 care coordination participants, 325 coordination and home visitation participants, and 23,325 non-participants. All of the coordination/home visiting participants survived to age 1; recipients of care coordination alone were 40% less likely to survive than were eligible non-participants. The costs for each additional life saved by providing coordination and visitation was $83,043 for very low birth weight infants and $13,993 for low birth weight infants. CE ratios for coordination alone were $61,806 and $24,712 for very low birth weight and low birth weight infants, respectively.

Recent studies have illustrated both the health benefits as well as the cost savings associated with the use of maternity care services,. While use of these services is beneficial and provides system-wide cost savings, maternity care in the U.S. continues to comprise an increasing proportion of annual health expenditures. Recent estimates of national maternity care costs range from $23 to $34 billion annually, comprising approximately 5% of total annual U.S. health expenditures. In an era of increasing health expenditures, there is a need to create a more efficient use of health care resources.

One proposed solution to address the inefficient use, as well as the inequitable distribution, of maternity care resources has been the efficient utilization of advanced practice nurses (APNs). While certified nurse-midwives (CNMS) and nurse practitioners (NPs) have been documented to increase access to the provision of high-quality, "cost-effective" maternity care, their inefficient use have proved costly. A method, based in welfare economic theory, has been developed to estimate the costs associated with the inefficient use of APNS. Nichols' model is the first and only neoclassical economic model applied to the health services market and underutilization of APNS. While this model has notable advantages, including the estimation of state-level economic loss, estimations were developed base on 1987 data for the primary care services market, rather than on recent data for maternity care services. There has not been an application of this model specific to maternity care services.

We propose a study with the following specific aims: First, to estimate the parameters of the social welfare loss model by: (a) obtaining state-level estimates of the two primary model parameters, the average price paid per maternity care visit (P) and the quantity of maternity care services delivered (Q), based on small area estimation techniques applied to national and county level data; and (b) defining specific estimates of other parameters from the literature. Second, to compute social welfare loss estimates associated with the inefficient use of CNMs/NPs for each U.S. state. Third, to perform sensitivity analyses of selected parameters for each U.S. state. Fourth, to assess state-level social welfare loss as it relates to legislative change.

An economic model described by Nichols (1992) will be modified to compute the social welfare loss society is presently bearing for the inefficient utilization of advanced practice nurses due to restrictive regulatory and reimbursement environments.

The conceptual model presents the relationships among parameters within the framework adapted from Nichols' economic model. In this model, the final outcome is the estimated economic premium, or social welfare loss, society pays for restrictive state practice environments for CNMs/NPs. The model will incorporate -estimates of clinical underutilization, as well as economic parameters, to compute these state-level economic premiums.

The state-level social cost estimates that ultimately emerge from the proposed model will be based on actual health services market data and derived values with an analytical foundation in economic theory.

Several researchers have argued that APNs have been underutilized in the current health services delivery system. Research defining the costs associated with this underutilization has been limited. To redress this concern, the research will provide state-level estimates of the economic loss society is currently bearing for the underutilization of CNMs and NPs in maternity care.

This dissertation will examine how one genre of providers, family planning clinics, is adapting to the growth of managed care. Family planning clinics are important providers of basic health care for uninsured, underinsured and adolescent women. Increasing competition for privately and publicly insured clients and declines in public funding threaten the financial viability of these and other "safety-net" providers. Managed care contracting is an important mechanism for recapturing sources of revenue traditionally used to cross-subsidize uninsured patients, yet safety-net providers have been relatively slow to adapt to managed care. The literature suggests family planning clinics are adopting various strategies to attract managed care contracts, but little is known about the factors related to the strategies or the types of contracts being negotiated.

This study will operationalize factors identified in previous studies (primarily case studies) as influencing the ability to obtain managed care contracts among family planning clinics, and test the external validity of these factors. Specific aims of the study are to identify organizational and environmental attributes that predict: 1) whether clinics enter into managed care contracts; and 2) the magnitude and types of contracts being negotiated (by payment method, payer type and covered services).

National-level secondary data from: 1) the 1994 National Survey of Women's Health Centers;2) the 1996 Planned Parenthood Federation of America Managed Care Contracting Survey and Annual Affiliate Services Census; 3) InterStudy (on managed care markets in 1994 and 1996); and 4) the Health Care Financing Administration (on Medicaid managed care statistics in 1994 and 1996) will be utilized to explore associations between the predictor and outcome variables using multivariate methods.

Study findings will provide policy relevant information concerning: the financial viability or providers of basic health care for millions of underserved women; trends in health care delivery for women, and the ways in which safety-net providers as a whole are responding to managed care.

This study will examine utilization patterns of selected medical and health-related services by chronically ill children with functional limitations. Previous research has focused mostly on the utilization patterns of medical (e.g., physician, hospital) services among chronically ill children. There is, however, little empirical evidence on the utilization patterns of other services and items use by this population and their families, such as therapeutic services, assistive devices, and respite care. Given the limitations of the existing delivery system, as well as, the changing environment in which these children receive their care, it is important to develop a better understanding of the broad range of services used by this population.

Data from the 1994 National Health Interview Survey, including supplemental questionnaires on childhood disability, will be used to investigate utilization patterns of four medical services (physician, hospital, home health, assistive devices and supplies) and five health-related services (respite care, therapeutic, social work, case management, transportation) among a sample of children between the ages of 5 and 17 with chronic conditions that result in functional limitations. Specifically, this study aims to: 1) determine the mix and volume of selected medical and health-related services used by a nationally representative sample of chronically ill children with functional limitations; 2) compare the volume of selected medical and health-related services used by selected subgroups of chronically ill children with functional limitations; and 3) determine if similar factors explain variations in the mix and volume of selected medical and health-related services used by chronically ill children, both within and across subgroups.

After performing descriptive analyses of the population, multiple logistic and multiple linear regression will be used to examine the relationship between "predisposing, enabling, and need" factors, and a) the likelihood of service use and b) the volume of utilization among users. Separate regression equations will be constructed for each service/item category. Additional analyses will determine if the association between the use of medical and health-related services for chronically ill children with functional limitations.

The results of this study will be useful for policymakers and program planners interested in developing comprehensive delivery systems for children with chronic and disabling conditions. The results of this study may also as a preliminary tool for monitoring access and utilization of this population in a changing health care system, by establishing baseline levels of resource use.

Frederick Connell (Health Services),
Ichiro Kawachi (Harvard),
Bruce P. Kennedy (Harvard),
and John W. Lynch (Michigan)

Data were obtained from the National Center for Health Statistics and the U.S. Census. Income inequality was calculated for the 414 counties with 1990 population >100,000. Linear regression was used to assess the associations between county income, income inequality, and teen birth rates.

Higher income inequality and lower median household income (MHI) were significantly associated with higher teen birth rates, with the association stronger among younger teens. Higher income inequality was associated with higher teen birth rates at every level of community income - an effect beyond that of income alone.

Although teen births are often regarded as a problem of individual behavior, the data indicate that contextual, community-level factors also contribute significantly to this undesirable health outcome.

The Office of the Insurance Commissioner assessed benefit coverage for contraceptive, sexually transmitted disease, HIV/AIDS, gynecological care (including sexual health counseling), infertility, sterilization, pregnancy termination, and reproductive cancer screening services.

A representative sample of health insurance carriers (n=12) was surveyed in July 1998. Data were colleced on the 91 top-selling plans, representing a total enrollment of 1,399,650 (44% of all those with health insurance in Washington State). Of these plans, 8% were indemnity plans, 16% were point-of-service plans, 22% preferred provider organization plans, and 54% were HMO-type plans. Policies for the coverage of "core" reproductive services based on US Preventive Services Task Force recommendations were assessed for each plan.

There are wide variations in coverage of "core" services by plan type. Coverage for core gynecologic (92%), maternity (93%), reproductive cancer screening (99%), STD (99%), and HIV/AIDS (100%) services was high. Half the plans covered no contraceptives; 30% covered the five FDA-approved reversible methods; and only two plans covered over-the-counter contraception. Three of 4 plans covered core sterilization (76%) and pregnancy termination (77%), and few plans covered infertility (25%) services. Benefit coverage is less than national figures, based on one available study. Most carriers do not have specific policies for health information related to provision of sexual health services to minors and victims of domestic abuse.

In this large sample a significant proportion of insured women and men did not have basic coverage for key reproductive health services.

U.S. reproductive health policy has concentrated largely on women, but researchers and policy makers are beginning to understand that encouraging men to be physically, emotionally, socially and financially responsible for reproductive health outcomes is a crucial element in the minimization of STD's. New advances in chlamydia testing make this possible. Urine based methods include leukocyte esterase (LE) and newer methods that amplify DNA make detection of Chlamydia in urine accurate and easy. However, these tests, called, polymerase chain reaction (PCR) and ligase chain reaction (LCR), can be significantly expensive. Cost-effectiveness studies enable us to analyze the trade-off between universal screening with an inexpensive test (which is not highly accurate), versus selective screening with a more expensive test (which is highly accurate), versus no screening. To date, there are no published comprehensive cost-effectiveness analysis, assessing the new urine based tests to screen asymptomatic men.

To compare the cost effectiveness of three screening programs for Chlamydia trachomatis in a population of asymptomatic males. The three interventions include (1) universal screening (test all males with urine LCR), (2) selective screening (test all males with urine LE; test all LE+ males with urine LCR), and (3) no screening.

The research design is a decision analysis conducted using TreeAge software to assess the cost-effectiveness of the three screening strategies using of a detailed decision tree to reflect the male infection, male transmission to female partners, and resulting complications such as PID and other sequelae. The analysis incorporates probabilities based on primary clinical data on 5,400 asymptomatatic males (mean age 16.2 years) and through a meta-analysis of current research, studies, and expert opinion. Direct medical costs will be estimated with Medicaid, Medicare, and/or Washington State Uniform Health Plan reimbursement schedules.