North carolina birth certificate

LOW BIRTH WEIGHT is a serious problem in the United States, especially in the southeastern region. Much of the reduction in neonatal mortality in the last 15 years has been attributed to improving the survival of low weight babies, while efforts at preventing low weight births have been much less successful [1-4].

Strategies for reducing low birth weight infants of women in poverty are of special interest, since their rates have been particularly high and resistant to improvement. The high rate of low birth weight in the United States compared with other developed countries is due in part to high rates among indigent women in this country [5]. Preterm birth prevention methods have been shown to be effective in reducing low birth weight in some settings [6-8], though some evidence indicates that such programs may be less successful among low-income women [5]. Methods that work among well-educated, high-income women may not be as applicable for poor women.

For women in poverty, just increasing the quantity of prenatal care (earlier entry and more visits may not be sufficient to improve birth weights substantially. Several studies have suggested that provision of nutrition, education, social work, and other services in addition to obstetrical medical care is also important [9-12]. These studies also suggest that such comprehensive prenatal care is more likely to occur in the public health setting. Our study uses several years of birth data for the States of North Carolina and Kentucky to examine the hypothesis that prenatal care provided in public health departments effectively reduces low birth weight among infants of women in poverty.

In the State maternal and child health programs in both North Carolina and Kentucky, contracts with local public health departments make funding contingent on the provision of a standard package of services. In addition to a broad array of prenatal medical services, a variety of ancillary services are required, including nutritional counseling, formal linkages to the Special Supplemental Food Program for Women, Infants, and Children (WIC), and other health education components. Thus, State-level funding and quality assurance mechanisms promote a consistent level of comprehensive prenatal care among the local public health departments in each State.

Methods

In this study, health program data files were linked to certificates of live birth to evaluate the effectiveness of prenatal care provided in public health departments. In the absence of a measure of income on the birth certificate, Medicaid-paid claims were matched to the birth records to identify a group of births to women in poverty. Records for women receiving prenatal care in health departments were then matched to the birth records. Finally, records for women who were in WIC in the prenatal period were matched to the birth certificates.

The primary comparison of the incidence of low birth weight is between women in Medicaid receiving prenatal care at public health departments and women in Medicaid receiving prenatal care from other providers (classified as such if the Medicaid birth certificates are not matched to a health department program record). Participation in WIC for these two groups is then compared. All live births to North Carolina residents for 1986 through 1988 and Kentucky births for 1985 and 1986 for which the birth certificate was matched to a Medicaid claim were included in our analyses.

In North Carolina, Medicaid claims paid for newborn hospital stays were matched to the birth records in several computerized iterations using the baby's name, hospital of birth, date of birth, sex, and county of residence. Around 98 percent of the newborn Medicaid claims for 1986 through 1988 were successfully matched to a birth record. In Kentucky, Medicaid claims paid for delivery were matched to births using the mother's social security number and then name and other variables for records not matching in the first step. A matching rate of 98 percent was achieved for the years 1985 and 1986. During these periods, the income level for Medicaid eligibility was about 40 percent of poverty in North Carolina and 27 percent of poverty in Kentucky.

In both States there is a statewide computerized client information system for public health departments that includes records for maternity care. In North Carolina, pregnancy closure records with a live birth outcome were linked to 1986 and 1987 births using mother's name, date of birth of the baby, and county of residence in two iterations. Eighty-nine percent of the health department records for the 2 years were successfully matched using these criteria. In Kentucky, delivery records from the health department system were matched to births by social security number and then name and demographic variables, with a matching rate of about 95 percent.

Matching outcome or delivery records has the potential problem that women who begin prenatal care at the health department and then are referred to other providers for high-risk conditions could be counted in the not-health-department group because of nonmatching. In both States, however, women who are referred to other providers for high risk medical care but continue to receive WIC or other support services from the health department or who come to the health department for a post-partum visit for family planning will usually have a delivery good in the health department system, especially if referral is to a tertiary hospital clinic rather than to a private physician. These women were counted in this study as health department patients.

With the use of the new birth certificate in North Carolina in 1988, records of prenatal visits to the health department could be matched directly to the birth certificates using mother's name and her date of birth (month, day, year). This arrangement precludes having to rely on the baby's date of birth on potentially incomplete records of pregnancy closure to identify health department births. If a woman has just one prenatal visit in a health department, she is counted in the health department category. This procedure minimizes provider referral bias and may in fact lead to a conservative estimate of a positive effect of health department care on low birth weight. Data from the 1988 North Carolina birth cohort are compared to the earlier data for North Carolina (1986-87) and Kentucky (1985-86) as a means of validation. Beginning in October 1987, the Medicaid eligibility level in North Carolina for pregnant women was raised to 100 percent of the Federal poverty level, so some of the women in Medicaid in 1988 will have had higher incomes that those during the 1986-87 period.

In North Carolina, 90 percent of the records for women in WIC during the prenatal period were matched to a 1986-1987 birth certificate using name, date of delivery, and county of residence. Some women in WIC do not return for a post-partum visit, and they were not eligible to be matched since no date of delivery was recorded. Therefore we expect that the number of births to women in North Carolina who were in WIC shown in this study is an underestimate. In Kentucky, the social security number was available for matching WIC records to births, followed by name and demographic variables for those not matching in the first step, so that 95 percent of prenatal WIC records were matched to a birth certificate.

The incidence of low birth weight (less than 2,500 grams), very low birth weight (less than 1,500 grams), and other birth characteristics are compared for two source-of-prenatal-care groups: health department, Medicaid; and Medicaid, not health department. Further comparison of the two Medicaid groups is done using logistic regression to assess the effect of health department prenatal care on low birth weight while controlling for differences between the two groups on other risk factors for low birth weight (under age 18, unmarried, less than 12 years of education, previous fetal death or live-born infant who died, and less than adequate quantity of prenatal medical care). For the 1988 North Carolina analysis, two control variables available on the new birth certificate are added: maternal smoking and presence of medical risk factors during the pregnancy (from check boxes).