2.5. Data abstraction and risk of bias

Selected articles were imported into Rayyan software, a semi-automated web and mobile-based tool for systematic reviews [17]. The software identified duplicates, and one of the authors verified and removed the duplicates. Two authors (N.B.T. and F.S.W.) independently screened the titles and abstracts, based on the eligibility criteria stated above to validate their selection, and screened the full text for selected articles. A third author (D.N.) resolved all conflicts after the title/abstract screening and again after full-text screening. The team developed a standardized data abstraction form coherent with the study objectives and outcome measures. This data abstraction template was pilot-tested on a subset of articles and then tested for face and construct validity.

The risk of bias across studies was assessed using the tool developed by Hoy and colleagues [18], a method that relies on the GRADE working group (Grades of Recommendation, Assessment, Development and Evaluation) and Cochrane approaches. The tool assesses external validity (items 1–4), for example “Was some form of random selection used to select the sample, OR, was a census undertaken?” and internal validity (items 5–10), for example “Was the study instrument that measured the parameter of interest shown to have reliability and validity (if necessary)?” for each study, where a score of 1 was given if the item was reported, 0 if it was not reported, and 0.5 for ‘no information’ ([18] p.4). N.B.T. and D.N. independently scored the studies, and the risk of bias was classified as low (8.5–10), moderate (5–8), or high (0–5.5).