Background and aims The appropriateness of intensive care for extreme prematurity continues controversial. In neonatal intensive care, an increasingly common choice is whether or not to resuscitate at 23 weeks gestational age. We sought to investigate whether such an intervention is cost effective.
Methods A decision analytic model was designed comparing resuscitation vs. non-resuscitation from a societal perspective for pre-term deliveries at 23 weeks. Estimates of death (74%) and neurodevelopmental disability (84–91%) in the setting of resuscitation were taken from the existing literature. Utilities were applied to discounted life expectancy to generate QALYs. All costs and QALYs were discounted at 3%. A cost-effectiveness threshold of $100,000 per QALY was utilized. Sensitivity analysis included univariate and bivariate comparisons and Monte Carlo simulations.
Results Non-resuscitation is the dominant strategy, as it is both less expensive ($71, 036 v. $259,358) and more effective (24.7 QALYs v. 24.4 QALYs). While resuscitation would lead to 240 live infants, in a theoretical cohort of 1,000 cases, there would be 100 severely disabled, 70 moderately disabled, and 90 non-disabled survivors. In univariate sensitivity analysis, non-resuscitation was the cost-effective strategy at all reasonable ranges of the inputs for the cost of NICU care and risk of mortality and disability.
Conclusions From a societal perspective, it does not appear cost-effective to resuscitate 23 week neonates over a wide range of assumptions. In our model, even if NICU care is free, resuscitation is not cost effective at baseline due to extreme long-term costs.