The review will be conducted and reported in accordance to the preferred reporting items for systematic reviews and meta-analysis (PRISMA) statement (11, 12). This systematic review protocol title will be registered with the International Prospective Register of Systematic Reviews (PROSPERO).
We will include articles that have documented the psychosocial interventions to address depression in elderly patients in low and middle income countries.
Psychosocial interventions are defined as any interventions that emphasize psychological or social factors rather than biological factors (13). This definition allows for the inclusion of psychological therapies and health education as well as interventions with a focus on social aspects, such as social support and networking. Interventions with a physiological component in addition to a psychosocial component (e.g., exercise groups for older people) will also be considered.
The psychosocial interventions could appear in any format, for example, in groups or individually, as long as they are described in the study and would allow for replication. Interventions with organization of care as the main focus will not be considered for this review. All trials of secondary and tertiary preventive interventions, relapse prevention, and pharmacological interventions will be excluded.
The review will include articles from studies done following cross-sectional, longitudinal, cohort, case control and randomized control study designs. Sub-group analysis will be performed on the included studies to cater for the variations in the study designs.
Unpublished but accessible studies will be included in the review. Studies on non-suicidal deliberate self-harm will be excluded. Articles with only abstracts available (with no full articles) will be excluded as they may be difficult to compare with full articles.
The selection criteria of the articles for inclusion will be strictly applied and no double reporting of the same outcome will be included in the review. We will report the study as a single one, but in the event that there could be separate outcomes, we will report them as such. Each publication will have the name of the author, year, title and publisher.
The trials to be considered for this review will have depressive symptoms or depression as a measured outcome. Furthermore, to be eligible, trials will have to encompass a control condition: either care as usual, waiting list, or no intervention.
We will conduct a systematic review of observational studies (cross-sectional, cohort, longitudinal and case control studies) that documented psychosocial interventions to address depression in elderly patients in low and middle income countries. We will also describe the most common psychosocial interventions for depression, identify factors that promote psychosocial interventions for depression and assess the effectiveness of psychosocial interventions for depression among elderly patients.
The systematic review will only include studies conducted in low and middle income countries all over the World.
The participants in this systematic review will be elderly persons aged 50 years and above.
The comparison groups will be elderly provided with psychosocial interventions and those without (biological care).
The primary outcome of the review will be the occurrence of depression and depressive symptoms, as measured by Depression Rating Scales. Secondary outcomes will be functional level and quality of life. The outcome measures will be recorded immediately after the intervention and at end of follow-up.
The search strategy will be carried out by the research team using the following electronic databases and search engines, from inception, using the same search strategy (Appendix A) with alterations as appropriate for each database: the Cochrane Library, PsychINFO, PubMed, EMBASE, Africa wide-information and global health. We will hand search the references of the included studies. We will use a combination of mesh terms, text words and combine them with appropriate Boolen operators in order to identify as many studies as possible. Grey literature will also be included. We will identify the relevant grey literature using web searching, web-based catalogues as well as using bibliographic databases.
All publications retrieved from the databases will be screened for inclusion by the primary reviewer and an independent researcher separately. Available data will be extracted and coded independently by the primary reviewer and a second data extractor.
The methodological quality of the included intervention studies will be assessed and rated according to the Cochrane Collaboration Handbook (14). Quality of the study will be assessed through risk of bias assessment tool which will be done using a checklist developed from the strobe statement .
The data collected will also include the authors’ names, the title of the article and the year the study was conducted. This data will be extracted by two content experts on the research team. For articles with missing information, the primary authors will be contacted by the principle investigator to provide the missing or additional data. For any discrepancies, the principle investigator will make the final decision. Data extraction will be done in two stages by four people of different specialties that include a Psychiatrist, Psychologist Statistician and an expert in reviewing systematic reviews. We will then download the full text articles for further screening by independent reviewers who will be identified. In the event that there is disagreement about which study to include or exclude, the PI will be the arbitrator.
All publications retrieved from the databases will be screened for inclusion by the primary reviewer. Available data will be extracted and coded independently by the primary reviewer and the systematic review expert.
The methodological quality of the included intervention studies will be assessed and rated according to the Cochrane Collaboration Handbook (14). Specifically, the Cochrane risk of bias (RoB) tool will be used to assess the quality of research papers. The studies will be rated by taking five individual domains into consideration (sequence generation; allocation concealment; personnel, and outcome assessors; incomplete outcome data; selective outcome reporting; and other sources of bias) and giving them a quality rating of “low risk of bias,” “unclear,” or “high risk of bias.” We will use GRADE criteria to assess the quality of evidence. Specifically, we will consider risk of bias, imprecision, indirectness, inconsistency and possible publication bias.
The primary outcome of the review will be the occurrence of depression and depressive symptoms, as measured by depression rating scales (such as the Geriatric Depression Scale). Secondary outcomes will be functional level and quality of life. The outcome measures will be recorded either immediately after the intervention or at end of follow-up.
Data will be entered into Review Manager 5.0 software by the principal reviewer and systematic review expert separately. For binary efficacy outcomes (e.g., depression), the Mantel-Haenszel random effects model for calculating odds ratio (OR) will be applied. Where intention-to treat (ITT) data will not be available, end-point continuous data for trial completers will be used. For continuously distributed outcomes, the Weighted Mean Difference (WMD) or Standardized Mean Difference (SMD) will be calculated as appropriate using a random effects model. SMD will be calculated when outcomes will be measured using different scoring systems.
If measures of variance of outcomes are found through publications, through calculations, or by contacting the authors, the outcome will be excluded from the meta-analysis. Substantially, skewed data (where the standard deviation is more than twice the mean value) will not be entered in the meta-analysis. The impact of statistical heterogeneity on the meta-analysis will be assessed by quantifying inconsistency among the studies with the I2 Index test . Sensitivity analyses will be conducted by looking at randomized trials only and at studies with low risk of bias only (i.e., at least two out of five domains will be rated as low risk of bias for the study).