Introduction of a prehospital critical incident monitoring system--pilot project results

Hospital medical incident monitoring improves preventable morbidity and mortality rates. Error management systems have been adopted widely in this setting. Data relating to incident monitoring in the prehospital setting is limited.

Implementation of an incident monitoring process in a prehospital setting.

This is a prospective, descriptive study of the pilot phase of the implementation of an incident monitoring process in a regional prehospital setting, with a focus on trauma care. Paramedics and emergency department staff submitted anonymous incident reports, and a chart review was performed on patients who met major trauma criteria. Selected trauma cases were analyzed by a structured interview/debriefing process to elucidate undocumented incidents. A project committee coded and logged all incidents and developed recommendations.

Of 4,429 ambulance responses, 41 cases were analyzed. Twenty-four (58.5%; 95% CI = 49.7-67.4%) were reported anonymously, and the rest were major trauma patients. A total of 77 incidents were identified (mean per case = 1.8; CI = 1.03-2.57). Anonymous cases revealed 26 incidents (mean = 1.1; CI = 0.98-1.22); eight trauma debriefings revealed 38 incidents (mean = 4.8; CI = 0.91-8.69) and nine trauma chart reviews revealed 13 incidents (mean = 1.6; CI = 1.04-2.16). A total of 56 of 77 (72.7%; CI = 65.5-80.0%) incidents related to system inadequacies, and 15 (57.7%; CI = 46.7-68.6%) anonymously reported incidents related to resource problems. A total of 35 of 77 (45.5%; CI = 40.4-50.5%) incidents had minimal or no impact on the patients' outcomes. Thirty-four of 77 (44.2%; CI = 39.3-49.1%) incidents were considered mitigated by circumstance. Incident monitoring led to generalized feedback in most cases (65 of 77; 84.4%; CI = 77.6-91.3%); in three cases (3.9%; CI = 3.7-4.1%), specific education occurred; two cases were reported to an external body (2.6%; CI = 2.5-2.7%); three cases resulted in remedial action (3.9%; CI = 3.7-4.1%); four for trend/further observation and analysis responses (5.2%; CI = 4.9-5.5%).

The pilot project demonstrates successful implementation of an incident monitoring system within a regional, prehospital environment. The combination of incident detecting techniques has a high yield with potential to capture different error types. The large proportion of incidents in the "near miss" category allows analysis of incidents without patient harm. The majority of incidents were system related and many were mitigated by circumstance. The model used is appropriate for ongoing incident monitoring in this setting.