Background: Occurrence variance reporting (OVR) systems are essential tools for identifying, mitigating, and preventing risks in healthcare, yet they are widely under-used because of inadequate training, fear of blame, and unclear definitions of what constitutes a reportable event. In Saudi Arabia, Vision 2030 reforms place incident reporting at the centre of quality and patient-safety improvement. Objective: To examine whether a hospital-wide staff training program on the OVR system was associated with changes in the number of incident reports, the distribution of risk levels, and the distribution of incident categories at Al-Noor Specialist Hospital, Makkah. Methods: A quantitative, retrospective, single-group pre–post quasi-experimental design without a control group was applied to secondary data extracted from the hospital’s electronic OVR system. The intervention was a 40-minute face-to-face session (presentation, case study, and system-demonstration video) delivered from mid-May 2024 to 430 documented attendees (≈18.7% of 2,300 employees). A total census of 4,981 reports was analysed across a pre-intervention period (1 July 2023–30 June 2024; n = 2,018) and a post-intervention period (1 July 2024–30 June 2025; n = 2,963). Monthly volumes were compared with the Mann–Whitney U test after Shapiro–Wilk testing; risk-level and category distributions were compared with chi-square tests of independence. Results: Total reports increased by 46.8%, and monthly volume rose significantly (median 154 vs. 241; U = 11.00, z = −3.52, p < .001, rank-biserial r = 0.847). The reporting rate rose from 15.90 to 24.99 per 1,000 patient-days (+57.2%) despite a 6.6% fall in patient-days. Risk-level distribution changed significantly (χ²(3, N = 4,981) = 262.44, p < .001, Cramér’s V = 0.230), shifting toward low- and moderate-risk reports (82.0% to 95.9%). Incident-category distribution also changed significantly (χ²(5, N = 4,981) = 243.42, p < .001, Cramér’s V = 0.221), with notable increases in the Incident (0.8% to 10.5%) and Adverse Event (29.7% to 34.8%) categories. Conclusion: Staff training on the OVR system was associated with a sustained increase in incident reporting and with changes in the risk-level and category profile of reports. Because the design lacked a control group and classifications were not independently verified, the changes cannot be attributed to training alone or interpreted as improved classification accuracy. Training appears to be a promising, low-cost strategy against under-reporting when combined with leadership support, feedback to reporters, and evaluation in controlled studies.