Clinical supervision practices and determinants in health training institutions in Ghana: A multicentre mixed-methods study

Background Clinical supervision connects classroom learning with safe practice, but Ghanaian evidence identifies large student numbers, staff workload, limited supervisor preparation, and weak institutional coordination as persistent constraints. This study assessed supervision practices, student-reported effectiveness, supervisor competence, institutional structures, and determinants across Ghanaian health training settings. Methods A multicentre cross-sectional mixed-methods survey was conducted from January to July 2026. A voluntary non-probability sample of students, tutors/clinical instructors, and preceptors completed an online KoboCollect questionnaire distributed through institutional communication platforms. A 12-item CLES + T-informed measure assessed student-reported effectiveness, and a 10-item measure assessed supervisors’ self-rated competence. Psychometric assessment preceded index construction. Descriptive statistics and linear regression with HC3 robust standard errors were used. Free-text responses were analysed using dictionary-assisted qualitative content analysis and integrated with the quantitative findings. Results Of 682 submissions, 629 were eligible: 544 students, 69 tutors/clinical instructors, and 16 preceptors. Among 619 respondents with frequency data, 37.0% reported no supervision. High student numbers (72.2%), inadequate logistics/resources (63.3%), and inadequate staffing (49.0%) were the leading structured challenges. The effectiveness index showed acceptable internal consistency (α = .797; M = 3.73, SD = 0.56; n = 537), and the competence index showed excellent consistency (α = .928; M = 4.30, SD = 0.65; n = 83). In the adjusted student model (n = 484), daily (B = .208, 95% CI [.051, .366]), weekly (B = .208, 95% CI [.084, .331]), and monthly supervision (B = .324, 95% CI [.097, .551]) were associated with higher effectiveness than no supervision. Full communication of learning objectives was also positively associated (B = .263, 95% CI [.140, .385]). Compared with always completing supervision documentation, sometimes (B = − .398, 95% CI [-.526, − .270]) and rarely/never completing it (B = − .233, 95% CI [-.418, − .047]) were associated with lower effectiveness. Free-text findings explained these patterns through workload, placement congestion, resource limitations, weak feedback, and poor institutional coordination. Conclusion Supervision was inconsistently delivered in the participating settings. Regular contact, clear learning objectives, and consistent documentation were modifiable correlates of student-reported effectiveness. Improvement requires protected supervision time, manageable placement numbers, prepared and recognised preceptors, adequate resources, and accountable collaboration between training institutions and clinical facilities.