Timing of antenatal corticosteroids for optimal neonatal outcomes: a Markov decision analysis model

Abstract

Objectives Antenatal corticosteroids (ACS) are administered to pregnant women at high risk of preterm delivery and are important for reducing neonatal morbidity and mortality. They have a limited timeframe of effectiveness and timing can be difficult due to the uncertainty surrounding a patient's clinical course and risk of preterm delivery. The objective of the current study was to design a decision analysis model to optimize timing of ACS administration and identify important variables which impact timing. Methods We created a Markov decision analysis model with a base case consisting of a patient at 24 weeks gestation with an antepartum hemorrhage. The decision strategies included immediate ACS administration compared to delayed or no administration. Outcomes are from the perspective of the neonate and consist of lifetime quality adjusted life years (QALYs). Model assumptions and data for model inputs were derived from current literature and clinical recommendations. Results Our base-case analysis revealed a preferred strategy of delaying ACS for two weeks, which resulted in an expected value of 39.176 lifetime discounted QALYs. This was associated with reduced neonatal morbidity, but also resulted in 0.1% more neonatal deaths compared to immediate ACS. Sensitivity analyses identified a baseline probability of delivery of 6.19% above which immediate steroids were preferred. Other important sensitive variables include gestational age and the relative risk reduction of ACS. Conclusions Clinicians should carefully consider these factors prior to ACS administration, with a low threshold for immediate administration if the probability of delivery in the next week is estimated to be greater than 6.19%. Antenatal corticosteroids (ACS) are administered to pregnant women at high risk of preterm delivery and are important for reducing neonatal morbidity and mortality. They have a limited timeframe of effectiveness and timing can be difficult due to the uncertainty surrounding a patient's clinical course and risk of preterm delivery. The objective of the current study was to design a decision analysis model to optimize timing of ACS administration and identify important variables which impact timing. We created a Markov decision analysis model with a base case consisting of a patient at 24 weeks gestation with an antepartum hemorrhage. The decision strategies included immediate ACS administration compared to delayed or no administration. Outcomes are from the perspective of the neonate and consist of lifetime quality adjusted life years (QALYs). Model assumptions and data for model inputs were derived from current literature and clinical recommendations. Our base-case analysis revealed a preferred strategy of delaying ACS for two weeks, which resulted in an expected value of 39.176 lifetime discounted QALYs. This was associated with reduced neonatal morbidity, but also resulted in 0.1% more neonatal deaths compared to immediate ACS. Sensitivity analyses identified a baseline probability of delivery of 6.19% above which immediate steroids were preferred. Other important sensitive variables include gestational age and the relative risk reduction of ACS. Clinicians should carefully consider these factors prior to ACS administration, with a low threshold for immediate administration if the probability of delivery in the next week is estimated to be greater than 6.19%.