Objective: To systematically review meta-analyses (MAs) and randomised controlled trials (RCTs) of interventions for infants at risk of cerebral palsy (CP), to determine if consensus exists in study end-points.
Methods: MAs within the "Neonatal" and "Pregnancy and Childbirth" Review Groups in Cochrane Database of Systematic Reviews (to June 2011) were included if they contained risk factors for CP as a study end-point, and were either published in 2010 or 2011 or cited >20 times in Sciverse Scopus. Up to 20 RCTs from each MA were included. Outcome measures, definitions and cut-points for ordinal groupings were extracted from MAs and RCTs and frequencies calculated.
Results: Twenty-two MAs and 165 RCTs were appraised. High consistency existed in types of outcome domains listed as important in MAs. For 10/16 most frequently cited outcome domains, <50% of RCTs contributed data for meta-analyses. Low consistency in outcome definitions, measures, cut-points in RCTs and long-term follow-up prohibited data aggregation.
Conclusions: Variation in outcome measurement and long-term follow up has hampered the ability of RCTs to contribute data on important outcomes for CP, resulting in lost opportunities to measure the impact of maternal and neonatal interventions. There is an urgent need for and long-term follow up of these interventions and an agreed set of standardised and clinically relevant common data elements for study end-points.
Keywords: Cerebral palsy; clinical trials; meta-analysis; outcome measures; research design.