Trends in medically-indicated versus spontaneous preterm birth, 2004-2013

BACKGROUND: Despite decades of research aimed at prevention, preterm birth remains an enormous leading cause of infant mortality in the United States and worldwide. Of concern, racial disparities in preterm birth remain an intractable public health issue. In an effort to reduce preterm birth, organizations such as the American Congress of Obstetricians and Gynecologists (ACOG) released policy statements in 2009 aimed at reducing early elective deliveries. Subsequently, the incidence of preterm birth in the United States has decreased, but whether this decrease is due to a reduction in iatrogenic or "medically-indicated" preterm birth is unknown. Further, the effect of the reduction in early elective deliveries on racial disparities is unknown. Our hypotheses were that 1) after 2009, preterm births would be less likely to be medically-indicated than due to spontaneous causes and 2) black-white differences in preterm births would be unchanged. OBJECTIVES: 1) Determine the proportion of preterm deliveries at Beth Israel Deaconess Medical Center (BIDMC) from 2004-2013 that were medically-indicated versus spontaneous. 2) Due to persistent disparities, determine if shifts in type of preterm delivery varied by race/ethnicity. METHODS: We reviewed the first 87 deliveries in 2013 and randomly selected 15% of the records for each year from 2004-2013. Additionally, we reviewed 69 charts to oversample black women's deliveries. We manually abstracted data from BIDMC's online medical record and designated each delivery as either medically-indicated (preeclampsia, poor fetal growth, hypertension, or other fetal/maternal condition) or spontaneous (preterm labor, preterm premature rupture of membranes or cervical incompetence). Two reviewers independently reviewed 18 records for concordance of medically-indicated versus spontaneous preterm birth typing. If the first reviewer could not phenotype the delivery, then a neonatologist and obstetrician were consulted. We reviewed 971 out of the 5,566 preterm deliveries and included 930 that were confirmed preterm and had a clear medically-indicated or spontaneous phenotype. We dichotomized the time period into early (2004-2009) and late (2010-2013). Statistical methods included comparisons of early versus late using Chi-Square tests, logistic regression models to adjust for potential confounding variables, and stratified analyses (singletons and black versus white). RESULTS: There were 46,981 deliveries at our institution during the study period, 5,566 of which were preterm. Among the 930 preterm deliveries sampled from the 10-year period, 45.6% were medically-indicated with a non-significant, subtle difference between the early (48.3%) and late (41.9%) (P=0.05) time periods. The odds ratios of medically-indicated versus spontaneous preterm birth in late versus early were 0.77 (P=0.05) and 0.73 (P=0.03) for all participants, unadjusted and adjusted, respectively. While not statistically significant, a higher proportion of preterm deliveries among black women were medically-indicated in the early (50.4%) versus late (40.6%) periods (P=0.19). There was a similar trend among white women between the early (50.0%) and late (46.9%) periods (P=0.48). The odds ratios of medically-indicated versus spontaneous preterm birth from late versus early were 0.67 (P=0.19) and 0.63 (P=0.14) for black participants, unadjusted and adjusted, respectively. For white participants, the odds ratios were 0.88 (P=0.48) for unadjusted and 0.80 for adjusted (P=0.20). Overall at BIDMC, the preterm delivery rate was significantly higher in the early period (12.3%) compared to the later period (11.2%) (P=0.0003). While we observed a reduction of preterm birth among all women, black women experienced a 20.8% decrease (from 16.2% in the early period to 12.8% in the late) in preterm birth, while white women experienced just a 4.9% decrease (from 12.4% to 11.7%), resulting in a narrowing of the racial disparity of preterm birth in our institution. CONCLUSION: At a Massachusetts birth hospital we found a reduction in the incidence of preterm deliveries over a 10-year period that coincided with policy efforts to reduce early elective deliveries. There was a reduction in the proportion of preterm births that were medically-indicated from 48.3% to 41.9%. The reduction in medically-indicated preterm birth was most evident among black women at BIDMC with concurrent decrease in the overall preterm birth rate among black women resulting in a near elimination of the racial disparity in preterm birth at BIDMC. Future work includes statistical analysis to account for the oversampling of deliveries in 2013 as well as oversampling of black women's deliveries using inverse probability weighting. We also plan to analyze which underlying conditions (preeclampsia, intrauterine growth restriction, fetal distress, etc.) were responsible for the reduction of the medically-indicated deliveries