Design: Within-trial economic evaluation alongside a multicentre, two-arm parallel group, randomised controlled trial (SIFT).
Setting: 55 UK neonatal units from May 2013 to June 2015.
Patients: Infants born <32 weeks’ gestation or <1500g, receiving less than 30 ml/kg/day of milk at trial enrolment. Infants with a known severe congenital anomaly, no realistic chance of survival, or unlikely to be traceable for follow-up, were ineligible.
Interventions: When clinicians were ready to start advancing feed volumes, infants were randomised to receive daily increments in feed volume of 30 ml/kg (intervention) or 18ml/kg (control).
Main outcome measure: Cost per additional survivor without moderate to severe neurodevelopmental disability at 24 months of age corrected for prematurity.
Results: Average costs per infant were slightly higher for faster feeds compared to slower feeds (mean difference £267, 95% CI: -6928 to 8117). Fewer infants achieved the principal outcome of survival without moderate to severe neurodevelopmental disability at 24 months in the faster feeds arm (802/1224 vs 848/1246). The stochastic cost-effectiveness analysis showed a likelihood of worse outcomes for faster feeds compared to slower feeds.
Conclusions: The stochastic cost-effectiveness analysis shows faster feeds are broadly equivalent on cost grounds. However, in terms of outcomes at 24 months age (corrected for prematurity) faster feeds are harmful. Faster feeds should not be recommended on either cost or effectiveness grounds to achieve the primary outcome.
- milk feeding
- randomised controlled trial