OBJECTIVE: To evaluate the cost effectiveness of three different approaches to the care of neonates born at 22 weeks of gestation: universal resuscitation, selective resuscitation, or no resuscitation. METHODS: We constructed a decision-analytic model using TreeAge to compare the outcomes of death and survival with and without neurodevelopmental impairment in a theoretical cohort of 5,176 neonates (an estimate of the annual number of deliveries that occur in the 22nd week of gestation in the United States). We took a societal perspective using a lifetime horizon, and all costs were expressed in 2017 U.S. dollars. Effectiveness was based on combined maternal and neonatal quality-adjusted life years (QALYs). The incremental cost-effectiveness ratio was determined (cost/QALY) for each additional survivor. The willingness to pay threshold was set at $100,000/QALY. All model inputs were derived from the literature. Deterministic and probabilistic sensitivity analyses were performed to interrogate model assumptions. RESULTS: Universal resuscitation would result in 373 survivors, 123 of whom would have severe disability. Selective resuscitation would produce 78 survivors with 26 affected by severe impairments. No resuscitation would result in only eight survivors and three neonates with severe sequelae. Selective resuscitation was eliminated by extended dominance because this strategy had a higher incremental cost-effectiveness ratio than universal resuscitation, which was a more effective intervention. The incremental cost-effectiveness ratio of universal resuscitation compared with no resuscitation was not cost effective at $106,691/QALY. Monte Carlo simulations demonstrated that universal resuscitation is more effective but also more expensive compared with no resuscitation, with only 35% of simulations below the willingness to pay threshold. CONCLUSION: In our model, neither selective nor universal resuscitation of 22-week neonates is a cost-effective strategy compared with no resuscitation.