Methodological eligibility challenges in evidence synthesis of the prevalence of mental health problems in health care workers during COVID-19

Article type
Authors
Gabaldà A1, Madrid E2, Metzendorf M3, Munar L4, Rivera F5, Roqué-Figuls M6, Serra C5, Vilahur N4
1Cochrane Iberoamerica, Institut de Recerca Hospital de Sant Pau, Barcelona, Spain
2Cochrane Iberoamerica, Institut de Recerca Hospital de Sant Pau, Barcelona, Spain; Interdisciplinary Centre for Health Studies (CIESAL), Cochrane Chile Associate Centre, Universidad de Valparaíso, Valparaíso, Chile
3Institute of General Practice, Medical Faculty of the Heinrich Heine University, Düsseldorf, Germany
4European Agency for Safety and Health at Work (EU-OSHA), Spain
5IMIM-Hospital del Mar Research Institute / University Pompeu Fabra, Barcelona, Spain
6Cochrane Iberoamerica, Institut de Recerca Hospital de Sant Pau, Barcelona, Spain; Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública (CIBERESP), Spain
Abstract
Background:
Conducting systematic reviews of prevalence is challenging due to limited available methodological guidance and the large volume and variable quality of studies. During the COVID-19 pandemic, numerous studies assessing the prevalence of mental health problems in health care workers were published, often of limited rigor and quality. Systematic reviewers may consider applying methodological criteria to select a valid and reliable body of evidence.

Objectives:
To explore the impact of chosen methodological eligibility criteria on a systematic review of the prevalence of mental health outcomes in health care professionals working during the COVID-19 pandemic across the European Union (EU).

Methods:
The review included cross-sectional studies reporting prevalence estimates of major mental health problems (anxiety, depression, stress, PTSD, burnout) in health care workers. The review protocol is available in PROSPERO (CRD42023473930).
We defined an additional set of methodological eligibility criteria by excluding studies in which (1) mental health outcomes were not assessed with validated scales or clinical interviews; (2) recruitment was not geographically representative at the municipal, regional, or country level (ie, in few centers, based on social media, by snowball, or insufficiently described); and (3) numerical prevalence data were not reported.

Results:
We included 113 studies (123 references). Over half of the full-text exclusions (55.9%) were methodological (see figure): 10.5% of studies did not use validated scales, 27.5% of studies recruited in a reduced number of centers, 14.2% of studies recruited mostly through social media and personal contacts, 22.2% of studies conducted snowball recruitment, 14.8% of studies insufficiently described recruitment, and 10.7% of studies did not report numerical prevalence data.
The exclusion ratio based on our methodological criteria varied across EU countries, with a median of 74.3%. Thus, 3 out of every 4 eligible studies were excluded by our additional methodological criteria, even though they had matched the basic eligibility criteria. The benefits clearly outweigh the potential drawbacks for all but one methodological criterion: excluding studies recruiting in few centers (see table).

Conclusions:
Our additional methodological exclusion criteria ensured internal validity and pragmatically dealt with a complex and heterogeneous body of evidence. This work has relevance to patients and policymakers by contributing to methods that result in more robust evidence syntheses.