36 DATA EXTRACTION The two reviewers then performed a full-text review of the included publications and filled in a data extraction table in Excel. Extracted data regarded study characteristics, study aims, main results, and conclusions. For each study, data on associations between LHL and mediating factors were sorted in different columns in the Excel file, based on the mechanisms in the Pathway of PaascheOrlow: self-care management, patient-provider interaction, utilization of care. Our extraction file also encompassed columns to extract data on LHL and the role of the social context or competences of the health care professional, which came from IROHLA Intervention model[18]. Clinical health outcomes, such as kidney decline or blood pressure, were in a different column in the file. This structure helped to unravel the HL-mediator-health outcomes pathway. For intervention studies, we added to the table information about the chosen strategies and its effectiveness, also derived from IROHLA[18]. QUALITY ASSESSMENT MDB and EMF rated the methodological quality of the included quantitative and intervention studies with the checklist of Downs and Black[26] and three additional criteria from the Effective Public Health Practice Project (EPPHP) Quality Assessment Tool and Appraisal Tool for Cross-sectional studies (AXIS) [27,28]. Disagreements were solved in discussion with a third reviewer, AFW. The EPPHP and AXIS criteria were added to put more weight on potential participation bias, because of known lower research participation of people with LHL[29]. Qualitative studies were assessed with a checklist, derived from the Cochrane Supplemental Handbook Guidance[30]. Together, the tools provided 16 criteria for quantitative studies, 30 for intervention studies and 18 for qualitative studies within four domains: 1) reporting, 2) external validity, 3) internal validity and 4) study participation. Each criterion could be rated with 0, 1 or 2 points. The total rating for all criteria and each independent domain was expressed as a percentage of the total maximum score possible. Domains could be of low (≤50%), moderate (>50% and ≤75%) or high (>75%) quality. Both the total and domain ratings were used to determine the final study quality. A high quality study had a total score >75% and at least three domains with a high quality rating. Details on the rating system are in supplemental tables 3a-d.
RkJQdWJsaXNoZXIy MTk4NDMw