An evaluation of formal risk scoring for preterm birth.

Formal risk scoring systems for the prediction of preterm birth lack many of the characteristics that are required from effective screening tests. They show poor positive predictive values, poor reproducibility, and large differences in performance among different populations. In clinical practice the implementation of these scoring systems has not shown to confer more benefit than harm to the women and infants involved. There is no conclusive evidence for reduction of the incidence of preterm birth, especially among women considered to be at high risk on the basis of such scores. The use of these scores has led to a profusion of interventions being applied to women without demonstrable benefit for them from these interventions. The introduction of formal risk scoring for preterm birth thus profoundly alters the type of care available to pregnant women. The main effect thus far appears to be that a potential but relatively imprecise risk of preterm birth tends to be replaced by the certain risk of dubious treatments, whose merits are undocumented and whose hazards are unknown. There is great need for controlled studies to establish whether and for whom potential benefits of formal risk scoring can outweigh its hazards.