Environmental factors (mainly 'failures in medical device') and system factors (mainly 'workload') accounted for most of AAEs in the intensive care units studied. The systems and the ergonomics approaches to error analysis can be useful in creating a comprehensive error management programme in order to minimize the gap between work demands and individual capabilities.

BackgroundAction research is a participatory research method based on active cooperation between researchers and subjects. In clinical practice, action research enables active involvement of workers in developing and implementing actions promoting patient safety. This article describes a participatory action research project that was conducted in the radiology department of a tertiary care university hospital. The main objectives were: identifying potential adverse events in the department of radiology, and offering a proactive approach to improving patient safety.MethodsPhase one of the study included observing 100 patients in three units of the department and identifying potential adverse events using an observation form. According to the data obtained from the observations, multidisciplinary research teams developed and initiated, together with front-line workers, four types of interventions: ergonomic interventions in work environment design, interventions in work procedure and task design, training and guidance, and managerial interventions. Phase two included evaluation of the interventions after six months of implementation.ResultsResults showed different weaknesses in each of the three radiology units tested, including incomplete medical information necessary for performing the radiological procedure, and discontinuity of care. Post-intervention observations showed a significant reduction in the prevalence of potential adverse events. At the Angiography unit, potential adverse events related to incomplete medical information dropped from 50% to 32%, and at the CT unit they dropped from 70% to 23%. At the MRI unit potential adverse events related to discontinuity of care dropped from 61% to 19%.ConclusionsThe current study demonstrates the value of action research in non-hospitalizing health units and the benefits of cooperation between medical teams and human factor professionals in promoting patient safety. Methods similar to those described in the current paper are applicable to medical work teams in a broad range of practices.

Inter-observer agreement is essential to medical staff members and has a major effect on communication. The goal of the study was to examine the way medical professionals evaluate the potential severity of Almost Adverse Events (AAEs) that were observed in two intensive care units (ICUs). One hundred and fourteen AAEs were observed and recorded in both units by engineering students. Each AAE was rated independently by five senior medical staff members from each ICU, chosen by the unit manager, on a three-point severity level scale. Statistical analysis (K statistic and Cohen's Kappa) yielded relatively low levels of agreement among raters in both ICUs (< 0.3), but significantly greater agreement was found among nurses than among physicians in both ICUs. Low levels of agreement are attributed to the nature of work and characteristics of each ICU. Recommendations for improving agreements including forming shared mental models are specified.