Evaluation of intervention systematic reviews on chronic non-communicable diseases and lifestyle risk factors in low-middle income countries: meta-research.

Abstract

Background: Systematic Reviews (SRs) rigorously synthesize findings on a theme, but some articles with this design are redundant due to errors and conflicts. Meta-research aims to rigorously analyze research, assessing SRs' methodological quality and result reliability. This study evaluates SRs' overall quality in low- and middle-income countries (LMICs) on chronic non-communicable Diseases (NCDs) and key modifiable risk factors, using assessment tools.

Methods: A search strategy was conducted in the following databases: MEDLINE (via PubMed), Embase, (via Elsevier), Cochrane Library, and Grey Literature for published studies from January 1, 2014 - April 5, 2024. SRs addressing the association between at least one of the four most important modifiable behavioral risk factors (tobacco use, inadequate diet, alcohol consumption, and physical inactivity) and chronic NCDs in populations classified as LMICs according to the 'World Bank list of countries' were included. The selected studies were imported into the EndNote 20 software and analyzed using a form for the extraction of their main data and four tools were chosen to assess each of the most important domains of scientific evidence: Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) for article writing; Template for Intervention Description and Replication (TIDieR) for intervention description; A Measurement Tool for Evaluating Systematic Reviews (AMSTAR-2) for methodological assessment; and Grading of Recommendations, Assessment, Development and Evaluation (GRADE) for certainty of evidence.

Results: Nine studies were included in this analysis. The average Overall Score on the PRISMA 2020 checklist was 13.5 for articles published before 2020 and 25.67 for those published after 2020. TIDieR analysis revealed complete correspondence (100%) for item 'Brief Name', while other items, like 'Why' (89%), and 'What', 'Who Provided', and 'How' (78%), were partially met but significantly so. Regarding AMSTAR-2 criteria, only one study fulfilled all critical items, meeting item 7 by providing a detailed list of excluded studies and justifying each exclusion motive. Additionally, among critical items applicable to multiple articles, only item 11 was consistently fulfilled by all studies. In the final classification, one article achieved a moderate quality rating, three were critically low quality, and five had low quality among the nine evaluated articles. In the GRADE tool evaluation, limitations resulted in estimations for only 19 outcomes and 8 intervention-exposure sets.

Conclusion: The results demonstrated that the writing of recent scientific articles meets most of the PRISMA 2020 criteria, with a checklist being the most used tool. Interventions and exposure were also very well reported, with the TIDieR checklist not being cited in any study as a guiding tool. AMSTAR-2 revealed a methodological approach of varied quality, mainly low and critically low. The GRADE approach classified the certainty of the evidence as generally very low. Therefore, it is necessary to encourage adherence to these approaches to improve the methodological quality in SR studies on chronic NCDs and behavioral factors in LMICs.