Do trials reduce uncertainty? Assessing impact through cumulative meta-analysis of neonatal RCTs