Introduction:

Reporting systems (RSs) are the first step to improve patient safety in health institutions, consequently determining their performance is relevant.

Objective:

To determine the performance in terms of positive predictive value (PPV), the concordance, and the coincidence of passive and active RSs in the detection of true clinical incidents and reportable events with unwanted effects (REUWEs), in a teaching hospital in Bogotá, Colombia.

Methods:

Cross-sectional study, assembled in a retrospective cohort, consisting of hospitalized patients (>12hours) in the surgery service, between May and July 2017. The PPV was calculated for the detection of patients with clinical incidents or REUWE in both RS. Concordance and coincidence between RS were determined.

Results:

The incidence of REUWE from the passive RS was lower than the incidence from the active RS (2% vs 11.8%), the PPV for the identification of patients with clinical incidents and REUWE was similar (PPV patients with clinical incidents: passive 95% confidence interval [CI] 34.6-66.2 vs active 95% CI 45.1-71.7; and PPV patients with REUWE: passive 95% CI 36.8-65.4 vs active 95% CI 29.3-54.9). Concordance was acceptable (Kappa 0.38) and the coincidence of patients and their REUWEs was 15.38%.

Conclusion:

In the active and passive RSs, the detection of true clinical incidents and REUWE (PPV) was similar and the concordance in the detection of subjects with REUWE was acceptable. However, the coincidence between the REUWEs in the patients detected by each RS was substantially different and should be considered when analyzing the information coming from one or another RS.

        · resumo em Espanhol     · texto em Inglês | Espanhol     · Inglês ( pdf ) | Espanhol ( pdf )