Overview of biomedical and public health reviews in Ethiopia from 1970 to 2018: trends, methodological qualities, gaps and future directions

Introduction: Globally, there has been a dramatic increment of narrative reviews, systematic reviews and overview publication rates. In Ethiopia, only small number of reviews are published and no overviews conducted in biomedical and public health disciplines. Therefore, we aimed to (1) assess the trend of narrative and systematic reviews in Ethiopia, (2) examine their methodological quality and (3) suggest future directions for improvement. Methods: PubMed, EMBASE, Web of Science, SCOPUS, CINHAL, WHO Global Index Medicus, Cochrane Library and PsycINFO electronic databases were searched and supplemented by hand searching as well. All narrative reviews and systematic reviews with or without a meta-analysis from 1970 to April 2018 were included. The International Narrative Systematic assessment (INSA) for narrative reviews and A MeaSurement Tool to Assess Systematic Reviews (AMSTAR-2) for systematic reviews with or without a meta-analysis were used for quality appraisal. Fisher exact test at the p-value threshold of 0.05 was used to compare the differences in methodological quality. Results: Of the 2,201 initially identified articles, 106 articles published from 1970 to 2018 were eligible for full-text review. Among included reviews, 50.9% were narrative reviews, 16% were systematic reviews and 33.1% were systematic reviews with meta-analyses. Twenty-nine percent were published in Ethiopia and 43.4% were published after 2015. 85.1% of narrative reviews poorly described the characteristics of included studies and 63.8% did not report a conflict of interest. In systematic reviews, 89.6%, 91.7%, and 100% did not register/publish the protocol, justifying the selection of the study designs for inclusion and report sources of funding for the primary studies respectively. Overall, 55.3% of narrative reviews and 75% of systematic reviews with or without meta-analysis had poor methodological quality. Conclusions: Although publication rate of narrative and systematic reviews have risen in Ethiopia, half of the narrative reviews and three-quarters of the systematic reviews had poor methodological quality. We recommend authors to strictly follow standardized quality assessment tools during conducting reviews. Moreover, immediate interventions such as providing methodological training and employers, editors and peer-reviewers should carefully evaluate all reviews before submission or publication.