Predictive factors of preterm delivery in French Guiana for singleton pregnancies: definition and validation of a predictive score

Objective: Given the high rate of premature birth in French Guiana (13.5%), and its stability in time, the aim of the present study was to define a predictive score for preterm birth in women with a unique pregnancy in order to help prioritize health resources in the local context.Methods: A retrospective study was conducted on all deliveries of unique pregnancies in French Guiana collected between 1 January 2013 and 31 December 2014 in the Registre d'Issue de Grossesse Informatisé (RIGI), a registry that collects data on live births over 22 weeks of amenorrhea on the territory. Statistically significant predictors (p < .05) of preterm delivery were included in a logistic regression model. The selected variables were chosen to be available during the first trimester. Coefficients were used to establish a score which was categorized and prospectively validated using data from 2015.Results: Seven explanatory variables, all measurable during the first trimester of pregnancy, were significantly associated with preterm birth. The predictive score divided in deciles allowed to establish sensitivity and specificity thresholds. Overall, depending on the chosen threshold the score sensitivity was low and the specificity was high. Lowering the threshold identified half of women as "at risk" for preterm birth.Conclusion: This first trimester score was insufficiently sensitive to identify individual women at risk for preterm delivery.