Single Motherhood and Neonatal Survival of Twins: Materials and Methods
The data set, “The Matched Multiple Birth File,” assembled by the National Center for Health Statistics (NCHS) was used for this analysis. This data set includes U.S. live births and fetal deaths among twins and higher order multiples for the years 1995 through 1998. The completeness of this file is excellent (99%), and the procedures for quality control of the data are explained in detail elsewhere. Since correct analysis of multiple and correlated data such as that of twins is not possible without prior matching, the NCHS undertook the matching of records for deliveries involving multiple gestations for the years 1995-1998. The first stage of the matching involved the building of an algorithm consisting of variables from live birth and fetal death records that were considered the most uniquely identifying to the pregnancy. Live birth and fetal death records, which had identical values for these variables, were then identified. If the number of records with identical information equaled the reported plurality (e.g., two records reported as twins) these records were considered members of the same multiple and assigned a unique set identification number. In those instances where the number of records with identical data exceeded the reported plurality of the records, visual review was conducted and matching done as appropriate. Approximately, 93% of all records were matched in this first stage. All other records were considered unmatched and included in subsequent matching procedures that involved use of additional variables and a composite of algorithmic combinations in addition to using manual identification methods. Perfect matching was finally achieved for 98.8% of the records, and the whole process has been adequately validated and found to be very accurate.
The two groups of interest for the study were married and single mothers stratified by race. Marital status was determined through either direct questioning (90% of the states) or inferential procedures (e.g., availability of paternal acknowledgement, etc.). Race was defined as that of the mother for both blacks and whites. We compared the following selected sociodemographic characteristics between the two marital status subgroups within each racial group: maternal age, parity, maternal education, adequacy of prenatal care utilization and smoking during pregnancy. Adequacy of prenatal care was determined using the R-GINDEX (revised graduated index) algorithm, which determines whether prenatal care was adequate or not based on the trimester prenatal care began, number of visits and the gestational age of the infant at birth. Inadequate prenatal care refers to those women who had missing prenatal care information or who had prenatal care but was considered suboptimal or mothers who had no prenatal care at all. We preferred this index of prenatal care because it has been found to be more accurate than many others, especially, in describing the level of prenatal care utilization among groups that are high-risk and, therefore, may require more intense care (e.g., multiple pregnancies). The accuracy of all these aforementioned variables on the birth certificate has been validated in previous studies, and the information coded in these variables was accurately reported.
canadian cialis online
Neonatal morbidity characteristics were compared between married and single women within each racial group. These included low birthweight (LBW, <2,500 g), very low birthweight (VLBW, <1,500 g), preterm (<37 weeks) and very preterm (<33 weeks) and small-for-gestational age (<10th percentile of birthweight for gestational age). Since the growth pattern of twins differs significantly from that of singletons, we applied growth curve references constructed specifically for twins. In addition to these characteristics, gestational age and birthweight specific mortality rates were calculated for each marital status and within each racial group. For gestational age-specific mortality rates, unit intervals were used, while birthweight was divided into 500-g strata. Gestational age in completed weeks is computed from the interval between the first day of the last menstrual period (LMP) and the date of birth. Records missing the date of the LMP are imputed when there is a valid month and year. From 1989-1997, clinical estimate of gestation was used in the computation of gestational age in cases where the date of the LMP was not reported or where the LMP date was inconsistent with the birthweight. Approximately, 4-5% of the gestational ages during the period were based on clinical estimate of gestation.
For this study, the main birth outcomes of interest were: neonatal mortality (death of the newborn within the first 28 days of life), which we further subdivided into early (from day 1 to day 7 postdeliv-ery) and late (from day 8 to day 28 postdelivery) neonatal mortality.
Outcome observations made on twins represent clustered data and are highly correlated. This lack of independence of observations renders conventional statistical methods inappropriate for this type of analysis, as they tend to disregard the clustering and treat observations as though they are unrelated events. This leads to inaccurate standard errors, resulting in spurious associations. To avoid such a bias, we employed the Robust Sandwich Estimator (RSE) to adjust the estimates of the variance of the coefficients in order to account for the correlation among observations within twin sets. Adjusted and unadjusted hazard ratios were generated using the Partial Likelihood Method described by Cox after testing for the nonviolation of the proportionality assumption. We confirmed this by plotting the log-negative-log of the Kaplan-Meier estimates of the survival function versus the log of time. The resulting curves were parallel. Based on an extensive literature review and their clinical/public health significance as well as probable biologic role, the following variables were adjusted for in the Cox model: education, parity, adequacy of prenatal care and maternal smoking during pregnancy.
We also computed adjusted excess mortality among twins born to single mothers using those of married gravidas as referent using the following equation:
where RR stands for the adjusted relative risk of mortality (measured as hazard ratios) and j denotes the racial subgroup (whites or blacks). All tests of hypothesis were two-tailed, with a type-1 error rate fixed at 5%.
online pharmacy uk
This study was approved by the Institutional Review Board at the University of Alabama at Birmingham.