In 2000, the U.S. Department of Health and Human Services, in its publication Healthy People 2010, reaffirmed its goal to improve the health of mothers and infants.1 Since such goals first were promulgated in 1990, preterm birth continues to be a major source of perinatal morbidity and mortality as well as having an important economic effect. In its recent report, the Institute of Medicine2 states that the incidence of preterm birth actually has increased during the past two decades. Specifically, preterm birth—defined as delivery before 37 weeks of gestation—accounted for 9.4% of live births in the United States in 1981; by 2004, this proportion had increased to 12.5%. According to the National Center for Health Statistics,3 much of this increase encompasses births between 32 and 36 weeks of gestation; the incidence of births before 32 weeks has been stable at about 2% for the past 20 years. Late preterm birth, recently defined by Raju et al as births between 34 and 36 weeks of gestation, now accounts for almost 75% of all preterm births.4 Importantly, these infants also have significantly increased morbidity and mortality when compared with term infants.5
One particular concern cited in the Institute of Medicine report is the long-standing and continuing disparity in the rates of preterm birth between racial and ethnic groups in the United States.2 These are most striking for African-American women, in whom the preterm birth rate in 2003 was 17.8% compared with 11.5% for white women. That same year, rates for Hispanic, Native-American, and Asian Pacific Islander women were 11.9%, 13.5%, and 10.5%, respectively.
Access to the health care system for pregnant women long has been recognized as a key determinant of birth outcomes.2 It is widely held that adverse pregnancy outcomes are related to poverty, and therefore racial/ethnic minorities would benefit from improved access to health care. To that end, policy makers have focused their efforts on improving birth outcomes by improving access to prenatal care. One example is expanded eligibility for Medicaid. In light of these efforts, particularly in minority populations, the Institute of Medicine’s findings on the increase in preterm births are particularly disappointing.
This report is from Parkland Hospital, the public hospital serving the medically indigent of Dallas County, Texas. Parkland Hospital, in conjunction with the University of Texas Southwestern Medical School, has developed a neighborhood-based, administratively and medically integrated public health care system for inner-city pregnant women with no other health care resources designed to improve access to pregnancy care. The purpose of this study was to compare the rate of preterm birth among African-American and Hispanic women at Parkland with national data.
Selected obstetrical outcomes for all women who give birth at Parkland Hospital, as well as neonatal outcomes, are entered into a computerized database. Nurses attending each delivery complete an obstetrical data sheet, and research nurses assess the data for consistency and completeness before they are stored electronically. Data on neonatal outcomes are abstracted from discharge records. Parkland Hospital is a tax-supported institution serving Dallas County that has a level III neonatal intensive care unit adjacent to the labor and delivery units. The obstetric service is staffed by house officers and faculty members of the Department of Obstetrics and Gynecology at the University of Texas Southwestern Medical School, and the neonatology service is staffed by house officers and faculty members of the Department of Pediatrics.
Between January 1, 1988, and December 31, 2006, a total of 280,961 women and adolescents delivered at our hospital. Of these, 260,197 were liveborn singleton infants weighing 500 g or more and delivered of women who had had prenatal care. A deidentified dataset was created for analysis of these mother–neonate pairs. These singleton births then were compared with the available U.S. data for the years 1995–2002 obtained from the National Center for Health Statistics.6 The U.S. data were limited to singletons weighing 500 g or more and delivered of women who had had prenatal care. Birth certificates on which birth weight or gestational age were reported as unknown or not stated were excluded from this analysis.
Preterm birth was the outcome of interest and was analyzed by subgroups to include birth at 27 weeks of gestation or less, 28–31 weeks, 32–36 weeks, and total births at less than 37 weeks of gestation. These subgroups were chosen to correspond to data groups available in the national dataset. This analysis was approved by the institutional review board of the University of Texas Southwestern Medical Center.
The obstetrical estimate of gestational age that was used to manage the care of the women and adolescents during the intrapartum period was used to assign the gestational age for the Parkland Hospital cohort. These estimates were based on the date of the last menstrual period (LMP) and the results of obstetrical ultrasonography, if any, performed during the pregnancy. The reported time of the LMP was accepted as correct if the fundal height measured between 18 and 30 weeks of gestation was correlated with the week of gestation within 2 cm. Patients with discrepancies between the two values underwent obstetrical ultrasonography. This methodology used to calculate gestational age has been validated previously.7
The primary measure used to estimate the gestational age of the newborn in the U.S. dataset was the interval between the first day of the mother’s LMP and the date of birth as recorded in the standard birth certificate. These data were edited for LMP-based gestational ages that were clearly inconsistent with the neonate’s birth weight. The U.S. Standard Certificate of Live Birth includes an item, “clinical estimate of gestation,” that was compared with length of gestation computed from the date the LMP began when the latter appeared to be inconsistent with birth weight. The clinical estimate also was used if the LMP date was not reported.
Women in the U.S. dataset who had had prenatal care was based on the 1989 version of the U.S. Standard Certificate of Live Birth. This birth certificate specifies prenatal care based on the “month prenatal care began.” The sources for this information are not required to be specified in the 1989 version of the standard birth certificate. We excluded women from this analysis when prenatal care was not stated or was listed as unknown on the birth certificate. Prenatal care in the Parkland Hospital dataset was defined as at least one visit to a hospital prenatal care clinic before presentation for delivery.
Statistical analyses using SAS 9.1 (SAS Institute, Inc., Cary, NC) included χ2 and Mantel-Haenszel test for trend. Race-specific comparisons for grouped preterm births were analyzed as multiple comparisons, and the significance levels are adjusted using the method of Bonferroni with a cutoff of P<.01. P values for all other comparisons were considered significant at less than .05. Of note, the Parkland births are unavoidably included in the U.S. dataset and could not be removed for analysis because the datasets are deidentified.
Maternal demographic characteristics for the U.S. cohort (n=29,366,816 women) are compared with those for the Parkland Hospital cohort (n=260,197) in Table 1. The Parkland Hospital cohort included predominantly minority women, with Hispanic women representing the greatest proportion (70%) and African-American women the second largest group (20%). A great majority (89%) of the Hispanic women delivered in Texas are from Mexico.3
Data are limited to singletons with birth weights of 500 g or greater and prenatal care. Shown in Figure 1 is the overall annual percentage of births at less than 37 weeks of gestation in the United States compared with Parkland Hospital. The comparisons according to maternal race/ethnicity are also shown in Figure 1. The overall rate of births at less than 37 weeks of gestation was significantly lower (P<.001) at Parkland, as were the rates for African-American and Hispanic women. Trend analysis for the Parkland data also showed a significant (P<.001) decline over the study time period. The comparison for Hispanic women throughout the remainder of this analysis was limited to Mexican women in both the U.S. and Parkland datasets. The decline in preterm births in African-American and Mexican women was independent of maternal age and parity. Subgrouped preterm births rates for 27 weeks of gestation or less, 28–31 weeks, and 32–36 weeks according to maternal race/ethnicity are shown in Table 2. Except for births to white women, subgrouped preterm birth rates were significantly lower at Parkland compared with the United States.
The percentage differences in births at less than 37 weeks of gestation, comparing rates among African-American and Mexican women with rates for white women in the United States and Parkland Hospital cohorts, are shown in Figure 2. The disparity vis-à-vis white women decreased for both minority populations at Parkland Hospital.
The rates of preterm birth at Parkland Hospital according to maternal race/ethnicity divided into those related to spontaneous rupture of the fetal membranes, idiopathic spontaneous labor, and indicated preterm births are shown in Figure 3. The trend in rates of ruptured membranes and spontaneous preterm labor significantly decreased (P<.001) during the study period in all three race/ethnicity groups, and indicated preterm births significantly increased in each of these groups.
Overall, 99% of the U.S. cohort received prenatal care compared with 94% at Parkland Hospital. The rate of prenatal care at Parkland increased significantly from 88% in 1988 to 98% in 2006 (P<.001). Shown in Figure 4 is the rate of births at less than 37 weeks of gestation synchronized with the proportion of women delivered at Parkland Hospital after receiving prenatal care in the hospital system. The decline in preterm births at Parkland coincided with an annual increase in prenatal care rates beginning about 1992. Analysis of births at Parkland to women who did not have prenatal care (n=16,031) did not show a similar decrease in preterm birth rates during the study time period. Specifically, the rate of preterm birth in women who did not have prenatal care remained above 15% and did not change significantly between 1988 and 2006 (P=.65).
This study was prompted by the 2006 Institute of Medicine report “Preterm Birth: Causes, Consequences, and Prevention.”2 Our purpose was to compare the incidence of preterm birth in the United States with similar outcomes in indigent pregnant women cared for within the tax-supported Dallas County hospital system. At least two significant findings emerged from this comparison. First, unlike that for the United States, the preterm birth rate at Parkland Hospital actually decreased consistently during the past 15 years. Second, the magnitude of the disparity in preterm birth rates between white women and racial ethnic/minorities was decreased in the Parkland cohort relative to the U.S. cohort.
The differences in preterm birth rates between our inner-city minority populations and the United States lead us inevitably to question why preterm birth rates were lower at our hospital when compared with the national data. One potential explanation recently was suggested by Joseph and colleagues,8 who noticed that preterm birth rates were much higher in the United States compared with Canada. This difference, however, was reduced when the U.S. rate was based on the “clinical estimate of gestational age” rather than the LMP, the latter being the basis for previously reported U.S. data. For example, the preterm birth rate in the United States in 2002 was 12.3% if computed using the LMP compared with 10.1% when the clinical estimate of gestation was used. In Canada, as well as in many industrialized countries with preterm birth rates lower than those in the United States, the rate is based on clinical estimates of gestational age similar to the Parkland data; this could account for at least some of the lower rates observed at our hospital. However, this explanation cannot account for all of the differences we observed. For example, the Parkland preterm birth rate was 5.5% in 2002, which still compares favorably with the corrected U.S. rate of 10.1% for that year.
Another explanation for the decreased rate of preterm births at Parkland Hospital might be related to the demographic shift to a predominantly Hispanic population during the past 20 years. During this time, the proportion of Hispanic women in our prenatal population has increased from 40% in 1988 to 85% in 2005. However, this explanation also seems unlikely because the decline in the preterm birth rate seen in Hispanic women also occurred in African-American women. Moreover, the disparity in preterm birth rates between white women compared with Hispanic and African-American women narrowed during the years studied. Another finding that speaks against the decline in preterm birth rates being attributable solely to demographic factors is the unchanged rate of preterm birth at Parkland in women who did not have prenatal care during the same time period.
In an effort to examine our findings from a different perspective, we compared birth weight distributions of neonates born at Parkland Hospital with the more than 60 other hospitals that voluntarily report their outcome data to the National Perinatal Information Center. This is a nonprofit organization that, among other services, performs analyses on the cost, management, and outcomes of perinatal services and provides comparative results to member hospitals.9 In a comparative analysis of birth weight distributions in 250-g to 500-g increments between 500 g and 2,499 g during 2007, the proportion of preterm births based on these birth weight groups was significantly lower at Parkland Hospital compared with the other hospitals in the dataset (Table 3). We are of the view that this result from an independent dataset serves to buttress the findings of our analysis.
There were a number of factors during this time period that could be contributory to the lowered preterm birth rates in our patient population. In the early 1990s, we began a concerted effort to improve access to and use of prenatal care with the intention of developing a program of seamless obstetrical care beginning with enrollment during the prenatal period and extending through delivery into the puerperium. Prenatal clinics were placed strategically throughout the county to provide convenient access for indigent women. In addition, these clinics were colocated with comprehensive medical and pediatric clinics that enhanced patient use. Because the entire clinic system as well as the hospital was operated by Parkland, administrative and medical oversight was seamless. For example, prenatal protocols were used by nurse practitioners at all clinic sites to guarantee homogeneous care, which included standardized referrals to the hospital, with its central clinic system, for women with high-risk pregnancy complications. These high-risk pregnancy clinics operate each weekday, with specialty clinics for women with prior preterm birth, gestational diabetes, infectious diseases, twins, and hypertensive disorders. Each clinic is staffed by maternal–fetal medicine faculty with special interests in such complications. Importantly, Parkland has a closed medical staff, and all attending physicians are employed by the University of Texas Southwestern Department of Obstetrics and Gynecology, whose members adhere to agreed-on practice guidelines using an evidenced-based outcomes approach. The same is true for inpatient management to include pregnancy complications. Thus, prenatal care is considered one component of a comprehensive and orchestrated public health care system that is community-based. Putting this all together, we hypothesize that the decrease in preterm births experienced at our inner-city hospital was attributable to a geographically based public health care program specifically targeting minority populations of pregnant women.
We must admit that the results now reported were unexpected. Why should preterm birth at our inner-city hospital be so strikingly different than in the U.S. dataset? Could it be so simple that access to an organized prenatal care system resulted in a decline in the rate of preterm birth? Although we cannot definitively answer this question, it is noteworthy that women in our healthcare system who did not have prenatal care demonstrated no such decline in preterm births. Even though we cannot provide a definitive explanation, we are compelled to report our results and suggest that public health care can prove useful in ameliorating preterm births in the most vulnerable women giving birth in the United States.
1.U.S. Department of Health and Human Services. With understanding and improving health and objectives for improving health. In: Healthy People 2010. 2nd ed. Washington (DC): U.S. Government Printing Office; 2000.
2.Committee on Understanding Premature Birth and Assuring Healthy Outcomes Board on Health Sciences Policy, Behrman RE, Butler AS, editors. Preterm birth: causes, consequences, and prevention. Washington (DC): National Academies Press; 2006.
3.Martin JA, Hamilton BE, Sutton PD. Births: final data for 2004. Natl Vital Stat Rep 2006;55:1–101.
4.Raju TN, HigginS RD, Stark AR, Leveno KJ. Optimizing care and outcome for late-preterm (near term) infants: a summary of the workshop sponsored by the National Institute of Child Health and Human Development. Pediatrics 2006;118:1207–14.
5.McIntire DD, Leveno KJ. Neonatal mortality and morbidity rate in late preterm births compared with births at term. Obstet Gynecol 2008;111:35–41.
6.United States Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics, Division of Vital Statistics. Natality 1995–2002. Natality data summary. Available at: http://wonder.cdc.gov/natality.html
. Retrieved August 4, 2008.
7.McIntire DD, Bloom SL, Casey BM, Leveno KJ. Birth weight in relation to morbidity and mortality among newborn infants. N Engl J Med 1999;340:1234–8.
8.Joseph KS, Huang L, Liu S, Ananth CV, Allen AC, Sauve R, Kramer MS, et al. Reconciling the high rates of preterm and postterm birth in the United States. Obstet Gynecol 2007;109:813–22.
9.National Perinatal Information Center. Available at http://www.npic.org/AboutNPIC/Aboutnpicindex.php
. Retrieved March 1, 2007.