Mental Health Literacy and its correlates among Zimbabwean Health Sciences students: a cross-sectional study and exploratory psychometric evaluation
TL;DR
The MHLq total score demonstrated good internal consistency and partial support for construct validity in this setting, but the weaker domain reliabilities and incomplete replication of the original factor structure support cautious interpretation of subscale scores.
Abstract
Globally, young adults, particularly university students, experience a high burden of mental disorders and multiple barriers to care, including limited mental health literacy (MHL). This study aimed to quantify MHL and associated factors among Zimbabwean health sciences students and, secondarily, to conduct an exploratory psychometric evaluation of the Mental Health Literacy Questionnaire (MHLq). This cross-sectional survey recruited 601 health sciences students using multistage sampling. MHL and common mental disorder (CMD) symptoms were measured using the 29-item MHLq and Shona Symptoms Questionnaire (SSQ-14), respectively. The MHLq total score was analysed continuously using multivariable ANCOVA with heteroscedasticity-robust standard errors. Exploratory psychometric analyses assessed internal consistency, factorability, factor structure and hypothesis-based construct validity. The mean MHLq total score was 119.5 (SD = 10.6). The total scale showed good internal consistency (Cronbach's = 0.85; McDonald's {omega} = 0.87), although domain-level reliability was lower ( = 0.52-0.66). Sampling adequacy for factor analysis was good (KMO = 0.87; Bartlett's test p < 0.001); parallel analysis suggested five empirical factors, and the original four-domain structure was only partially reproduced. Construct-validity analyses showed higher MHLq scores among students exposed to psychology/psychiatry courses, students in later years of study, and those with higher digital literacy, while MHLq scores were only weakly related to CMD symptoms. In adjusted ANCOVA, psychology/psychiatry course exposure, digital literacy, study year and family/friend information sources were independently associated with MHLq score; drug/substance use and personal CMD history were associated with lower scores. The MHLq total score demonstrated good internal consistency and partial support for construct validity in this setting, but the weaker domain reliabilities and incomplete replication of the original factor structure support cautious interpretation of subscale scores. Analysing the total MHLq continuously is therefore preferable for the present study. Campus-wide mental health education and digital-literacy strengthening merit further evaluation in longitudinal and intervention studies.