The aim of this study was to analyze the strength of safety measures described in incident reports in outpatient care.Methods: An incident reporting project in German outpatient care included 184 medical practices with differing fields of specialization. The practices were invited to submit anonymous incident reports to the project team 3 times for 17 months. Using a 14-item coding scheme based on international recommendations, we deductively coded the incident reports and safety measures. Safety measures were classified as "strong" (likely to be effective and sustainable), "intermediate" (possibly effective and sustainable), or "weak" (less likely to be effective and sustainable). Results:The practices submitted 245 incident reports. In 160 of them, 243 preventive measures were described, or an average of 1.5 per report. The number of documented measures varied from 1 in 67% to 4 in 5% of them. Four preventive measures (2%) were classified as strong, 37 (15%) as intermediate, and 202 (83%) as weak. The most frequently mentioned measures were "new procedure/policy" (n = 121) and "information/notification/warning" (n = 45). Conclusions:The study provides examples of critical incidents in medical practices and for the first time examines the strength of ensuing measures introduced in outpatient care. Overall, the proportion of weak measures is (too) high, indicating that practices need more support in identifying strong measures.

The aim of the study was to support the development of future critical incident reporting systems (CIRS) in primary care by collecting information on existing systems. Our focus was on processes used to report and analyze incidents, as well as strategies used to overcome difficulties.Methods: Based on literature from throughout the world, we identified existing CIRS in primary care. We developed a questionnaire and sent it to operators of a purposeful sample of 17 CIRS in primary care. We used cross-case analysis to compare the answers and pinpoint important similarities and differences in the CIRS in our sample.Results: Ten CIRS operators filled out the questionnaire, and 9 systems met our inclusion criteria. The sample of CIRS came from 8 different countries and was rather heterogeneous. The reporting systems invited a broad range of professions to report, with some also including reports by patients. In most cases, reporting was voluntary and conducted via an online reporting form. Reports were analyzed locally, centrally, or both. The various CIRS used interesting ideas to deal with barriers. Some, for example, used confidential reporting modes as a compromise between anonymity and the need for follow-up investigations, whereas others used smartphone applications and call centers to speed up the reporting process. Conclusions:We found multiple CIRS that have operated in primary care for many years and have received a high number of reports. They were largely developed in accordance with recommendations found in literature. Developers of future systems may find this overview useful.