Economic evaluation alongside the Premature Infants in Need of Transfusion randomised controlled trial
Journal Articles
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
BACKGROUND: The Premature Infants in Need of Transfusion (PINT) Outcome Study showed no significant difference in the primary outcome of death or neurodevelopmental impairment (NDI) in extremely low birthweight (ELBW) infants. However, a post-hoc analysis expanding the definition of NDI to include borderline intellectual functioning (Mental Development Index (MDI) <85) found an improvement in outcomes in the group maintained at higher haemoglobin levels. OBJECTIVE: To determine the cost effectiveness of more frequent red blood cell transfusions (high-Hb threshold) compared with less frequent transfusions (low-Hb threshold) in ELBW infants. DESIGN/METHODS: The authors performed an economic evaluation using patient-level data collected during the PINT randomised trial. The authors measured comprehensive costs from a third-party payer's perspective over a time horizon from birth through 18-21 months corrected age. RESULTS: The average total cost in the high-Hb threshold group was CAN$149 767 compared with CAN$150 227 in the low-Hb threshold group (difference of CAN$460, p=0.96). Cost-effectiveness analysis estimated savings of CAN$6879 for every additional infant surviving without severe NDI. There was a 48% chance that the high-Hb threshold reduced costs while improving outcome and a 90% chance that it would be cost effective at a willingness-to-pay threshold of CAN$250 000 per additional survivor without severe NDI. Post-hoc analysis defining cognitive delay as MDI score <85, instead of <70, revealed savings in the high-Hb threshold group of CAN$4457 per additional survivor without NDI. Results were robust to deterministic sensitivity analyses. CONCLUSION: A high-Hb threshold for transfusion, as measured in ELBW PINT study infants through 18 months corrected gestational age, may be an economically appealing intervention. The estimates were associated with moderate statistical uncertainty that should be targeted in larger, future studies.
