The starting point.
Neo-MILK was a stepped-wedge cluster-randomized trial involving 1,627 very-low-birth-weight infants across 15 German neonatal units. Exclusive mother’s own milk at discharge increased by a model-estimated 11 percentage points (95% CI, 3–19), against a modeled control baseline of 43%. The bundled intervention cannot establish which component contributed to the result. Feeding improvement does not establish long-term benefit.
Read the linked evidence brief for further results, safety context, and limits. An 11-point model estimate should not be converted into an observed intervention rate.
Read the full paper (opens in a new tab) ↗
Read the NICU Hearts evidence brief →Questions for the table.
- How does this population and setting compare with ours?
- What does cluster randomization change about our appraisal?
- What would make a similar approach feasible or difficult here?
- How would families define supportive, pressure-free care?
- What would we measure, including unintended burdens, before drawing conclusions from a local test?
Leave with a question.
Choose one uncertainty to investigate, a source to check, and someone to bring into the next conversation. Any proposed change to care needs your local clinical, governance, and family-partnership processes.