Agreement between Cochrane Neonatal reviews and clinical practice guidelines for newborns in Denmark: a cross-sectional study

Objective: To assess agreement between Cochrane Neonatal Group reviews and clinical practice guidelines in Denmark. Design: Retrospective analysis of clinical guidelines for newborn infants. Materials: All Cochrane neonatal reviews and Danish clinical guidelines for newborn infants. Main outcome measures: The recommendations from the Cochrane reviews and local clinical guidelines were compared and classified as being in agreement, in partial agreement or in disagreement. Authors of guidelines were asked whether Cochrane reviews had been considered during guideline development and reasons for any disagreements. Heterogeneity among departments was assessed. Results: 173 interventions evaluated in Cochrane neonatal reviews were included. All 17 Danish neonatal departments agreed to participate, but only 14 (82%) delivered data. Agreement between reviews and guidelines was observed for a median of 132 interventions (76%) (range 129-134), partial agreement was observed for 31 interventions (18%) (range 29-33), and disagreement was observed for 10 interventions (6%) (range 813) (kappa = 0.56, range 0.53-0.59). Most of the latter 10 interventions were not recommended in the reviews but were recommended in the guidelines. There were numerous reasons for disagreement, the most common being usage of evidence with higher bias risks than randomised trials in guidelines development. Overall, Cochrane reviews were rarely (10%) used during guideline development. For nine guideline topics (5%) there was diversity among the Danish departments' recommendations. Conclusions: There is good agreement between Cochrane reviews and neonatal guidelines in Denmark. However, Cochrane reviews were rarely used for guideline development. Heterogeneity among guidelines produced by the various neonatal departments seems moderate.