Background: Direct observation of procedural skills (DOPS) is a workplace-based assessment format that combines direct observation with feedback. This systematic review examined ophthalmology-specific studies reporting learner outcomes, perceptions, and implementation characteristics associated with DOPS.
Method: We searched PubMed, Scopus, Web of Science Core Collection, and ERIC through April 2025, with backward and forward citation tracking. Eligible studies were controlled or uncontrolled quasi-experimental studies and methodological comparisons of DOPS-related activities in ophthalmology trainees. Two reviewers independently screened records, extracted data, and assessed methodological limitations. Outcomes included procedural performance, learner and assessor perceptions, and implementation characteristics. Heterogeneous findings were synthesized narratively.
Results: Six studies met the inclusion criteria: one controlled pre–post comparison with alternating allocation, three uncontrolled quasi-experimental studies, and two methodological comparisons. The controlled comparison reported higher post-test OSCE scores after a DOPS-based teaching package than after traditional instruction (ophthalmic torch examination: 6.80 vs. 5.35; fundoscopy: 6.50 vs. 4.70; both p = 0.001). The uncontrolled studies reported higher scores or more favorable perceptions after programs that included DOPS; one program also included Peyton’s four-step teaching method, supervised practice, and feedback. The methodological comparisons found score correspondence between Video Observation of Procedural Skills (VOPS) and DOPS in small single-assessor samples; these findings do not establish validity or interchangeability. Several studies described assessor preparation and standardized forms.
Conclusions: Current evidence suggests that DOPS is a promising formative assessment strategy in ophthalmology education; however, the certainty of evidence is limited by small samples, single-center designs, lack of controls, heterogeneous interventions, and insufficient psychometric evaluation.