How Should We Test For Pre-Term Labour? An Economic Model And Cost-Effectiveness Analysis

Abstract

Objectives: Develop a prognostic model including quantitative fetal Fibronectin for predicting pre-term labour and assess its cost-effectiveness. Methods: Based on data derived from four European studies comprising 1,835 women and 171 events of preterm delivery in the QUIDS IPD meta-analysis, a decision analytic model assessed the cost-effectiveness of alternative prognostic strategies. Costs and probability of correct prognosis were calculated for each strategy, and expressed as net monetary benefit (NMB). Strategies included (i) clinical risk factors alone (CR) (ii) CR plus fetal-fibronectin (fFN), (iii) CR plus quantitative fetal-fibronectin (qfFN), (iv) CR plus transvaginal ultrasound of cervical length (TUCL), (v) CR plus fFN plus TUCL, and (vi) CR plus qfFN plus TUCL. The cost-effectiveness of each strategy is modelled over three gestation periods: extremely premature (<28 weeks), very premature (28-32 weeks), and premature (32-37 weeks). Results: qfFN testing dominated fFN in all three gestational periods with a lower mean cost per patient and greater probability of correct diagnosis at 7 days. qfFN had the greatest NMB for the Extremely Premature model (£17,301 (95% CI: £16,818 - £17750) per correct diagnosis), while in the Very Premature model TUCL had the greatest NMB (£16,375 (95% CI: £15,670 - £17,028) per correct diagnosis), while qfFN + TUCL had the highest NMB in the Premature model (£17,199 (95% Ci: £16,236 - £17,965) per correct diagnosis). Conclusions: Current NICE clinical guidelines support the use of TUCL for predicting pre-term labour in women ≥30 weeks pregnant. Our findings support this recommendation in Very Premature women, however, the optimal choice between qfFN and TUCL varies across the three gestation periods.