AIM: Continuous quality improvement has failed to consistently reduce morbidities in extremely low gestational age newborns 23-27 weeks. 10 Vermont Oxford Network NICUs describe a novel, sustained collaboration for progress.

METHODS: We emphasised a) commitment to inter-NICU trust with face-to-face meetings, site visits, teleconferences, scrutiny of quality improvement methodology, b) transparent process and outcomes sharing, c) evidence-based formulation of an orchestrated testing matrix to select potentially better practices, d) family integration, e) benchmarking with a composite mortality-morbidity score (Benefit Metric).

RESULTS: A total of 4709 infants, mean (SD) gestational age 25.8 (1.4) weeks, admitted to 10 NICUs 1.01.2010 to 12.31.2016. The orchestrated matrix offered 45 potentially better practices; NICUs implemented mean 29 (range 19-40). There was widespread adoption of delivery room, respiratory care and infection prevention practices, but no uniform pattern. Our Benefit Metric was significantly greater than the Vermont Oxford Network all seven years (p < 0.001). Six major morbidities decreased, two significantly (p < 0.05), mortality unchanged (14%). 34% of survivors had no morbidities, 35% just one.

CONCLUSION: Cultivating trust, transparent outcomes sharing, and tailored, potentially better practice selection is associated with encouraging improvement in 23- to 27-week survival without morbidity. Our outcomes are objective but the optimal implementation pathway to sustain progress remains murky, reflective of NICUs as complex adaptive networks.