Original scoring system of safety reports for ensuring medical security-trajectory for 18 years at a single center.

Abstract

Objectives: Safety reports are core components of secure medical systems, but their significance have remained obscure, mostly owing to limited quantitative approaches other than the numbers of reports. In 2005, we developed "management levels" and their equally converted points (p) of scores, which indicated the grades of required correspondences of medical systems. Moreover, products of those scores multiplied by risk levels (equally converted scores), "total scores" were also proposed for weighting harmful events from biphasic aspects; severity of patient damage and the required levels of correspondence as medical systems. In this study, we assessed this scoring system using extensive longitudinal experience.

Methods: Risk, management and total scores were retrospectively surveyed between 2012 and 2022 when a consistent reporting system had worked throughout the study period.

Results: The aforementioned three scores inconsistently decreased along with a decrease in number of safety reports from 2,445 to 1,194 cases, or 2,069 to 1,052 cases/10⁵ admitted patients (c/pap) in eleven years. Of the scores, the most prominent decrease was found in the management scores from 2,164 to 1,070 points/pap (p/pap). The risk and total scores had decreased even more modestly: from 1,879 to 1,484 p/pap, and 5,470 to 4,007 p/pap, respectively. When compared with the proportion (%) of risk and management levels in 2012, risk level 0 decreased, and risk level 1 and 3a increased in 2022. However, the proportion of management levels did not change during this period.

Conclusion: The proposed scoring system showed that the cumulative severity of patient damage did not worsen, or decrease in some years accompanied by decreasing correspondence levels of the medical system, although the numbers of reports also decreased. These findings show the appropriateness of the on-going safety activities of this hospital even when restricted to reported events. However, their effectiveness including unreported events, requires further study.