Abstract
We hypothesized that certainty of the available evidence is relatively low in neonatology. Thus, we designed a meta-epidemiological review to examine what is the certainty of evidence in the latest Cochrane neonatal reviews and investigate if the number of trials and enrolled patients is associated with the certainty of evidence. We searched Cochrane neonatal reviews published between January 2022 and May 2024. We included all reviews on interventions concerning neonates that had at least one meta-analysis performed with GRADE-rated evidence certainty. From those reviews, we extracted the presented certainty of evidence and analyzed its association with the number of trials and participants by ANOVA. We screened 55 Cochrane reviews and included 49 of them. In these 49 reviews, there were 443 reported outcomes with graded certainty of evidence. The certainty was reported to be high in 8 (1.8%), moderate in 89 (20.2%), low in 195 (44.0%), and very low in 151 (34%) of the outcomes. Reviews reporting outcomes with higher certainty of evidence had significantly more trials and patients (approximately 3 and 1.5 times more, respectively) than those with only low certainty of evidence. CONCLUSION: In the past 2 years, Cochrane neonatal reviews have generally had low or very low certainty of evidence for most outcomes. Only 2% of the reviewed outcomes had high certainty. The number of included patients and trials significantly affected the certainty. These findings highlight the continuous need for better quality and larger trials. WHAT IS KNOWN: • Neonatology is among the largest specialities and the evidence certainties of interventions have been varying. • Neonatal patients and studies need to consider the uniqueness of the patients and the acute situations in the study designs. WHAT IS NEW: • The included 49 reviews consisted of 443 outcomes and of these only 1.8% were classified as high certainty of evidence. • Higher evidence certainties were associated with higher number of included trials and participants.