Eligibility criteria
The following criteria were needed for an article to be included in our review
  1. Time frame: study realized during the coronavirus Disease 19 (COVID-19) outbreak,
  2. Reports’ characteristics: peer-reviewed articles without language restriction,
  3. Design: Observational studies with more than 30 participants and thereby including cross-sectional, case-control and cohort studies (baseline data),
  4. Participants and outcomes: articles reporting frequencies with or without associated/risk factors of mental health issues such as depression, anxiety and stress, among African universities’ students.
We excluded unrelated or duplicated articles, as well as abstract-only papers in case of unavailability of information required for our review.
Information sources and search strategies: We applied a search strategy on PubMed/MEDLINE, Excerpta Medica Database (EMBASE) and African Journals Online (AJOL) by the July 24, 2021, respectively at 00:00, 00:15 and 00:30 (Greenwich Mean Time). To build the search strategies respectively dedicated to each of the three databases, we associated key-terms groups, with as first group the one related to mental health issues, as second the one related to “students”, and as third the one pertaining to African regions and countries. For each group, synonyms and subtypes were joined using the Boolean operator “OR”, and the three groups were linked by the Boolean operator “AND”. For instance, regarding the group related to mental issues and the PubMed/MEDLINE database, we associated terms such as “Mental Health”, “Psychological”, “Depression”, “Anxiety” and “Stress”. Our search strategies are available in table 1. After applying our strategies in each database, the respective resulting lists were exported and further uploaded in Rayyan, a web and mobile app conceived for screening purpose in systematic reviews (https://rayyan.qcri.org/) [21]. In Rayyan, all the records (from the three databases) were joined as a same review project.
Selection process: We performed a two-steps selection process. First, based on the appraisal of titles and abstracts, and considering our eligibility criteria, two authors (FTE and WADL) independently selected articles. Discrepancies were resolved by a consensus or if necessary, by a third assessor (DLTM). After this first selection step, all relevant papers were exported from Rayyan in an appropriate file which later served to create a reference list of our selected articles, by using a reference software, notably Zotero (https://www.zotero.org/) [22]. The second step of the selection process involved to search for the full texts, and explore these full-texts considering our eligibility criteria and especially our outcomes of interest. This step was independently performed by two authors (FTE and DLTM), with disagreements solved by a consensus or a third appraiser if necessary (WADL).
Data collection process: The selection process was followed by data extraction. Indeed, we reviewed and collected data of interest from each article retained after titles, abstracts and full-texts screening. For this purpose, we used a predefined and pretested extraction tool, available in table 2. The data extraction process was independently ensured by two authors (FTE and WADL), and a third one intervened if dissimilarities were not solvable through a consensus.
Data items: Throughout the data collection step, we looked at the following items
  1. Bibliometric information: family name of the first author, year and journal of publication,
  2. Study characteristics: study’s country, study design, study period (and related COVID-19 stage if mentioned), study setting (one or more Faculties, one or more universities), sampling method, minimal sample size calculation, type of data collection process (paper and/or electronic version),
  3. Participants’ characteristics: subtype of university students (medical/ pharmacy/ health science students, etc.), mean or median age, age range, total sample size, sample size for males and females respectively,
  4. Outcomes: mental health issues (depression, anxiety, stress, etc.) with the tool used to screen for each one, the cut-off scores potentially used, the reported frequency, as well as associated/risk factors if assessed.
Study risk of bias assessment: We based our risk of bias assessment on the tool developed by Hoy and colleagues [23]. This tool is made of 10 items addressing four domains of bias plus a summary risk of bias assessment. Included studies were classified according to their total score as low risk of bias (8 – 10), moderate risk of bias (6 – 7) and high risk of bias (0 – 5).
Synthesis methods: We planned to present our results as narrative, graphical and tabular summaries, with respect to the following items
  1. Information on the study selection process through the PRISMA model of study flow chart presentation,
  2. Selected articles’ individual characteristics with mental health issues assessed, displayed as narrative and tabular summary, and specifically including articles’ bibliometric profiles, study and participants characteristics, tools/ cut-off scores/ frequencies/ factors regarding each studied mental health condition,
  3. Methodological quality and risk of bias of each study, presented as narrative and tabular synopsis.