Estimating impacts on safety caused by the introduction of electronic medical records in primary care.

CONTEXT Primary care is a highly complex environment in which multiple safety problems have been identified. Each primary care practice can be viewed as a complex adaptive system with its own unique characteristics. The introduction of an electronic medical record (EMR) into such a system represents a significant perturbation that can have multiple unpredictable effects. From a safety standpoint this can mean reduction in some vulnerabilities and increase in others, as well as the introduction of new vulnerabilities that did not exist under the old system. OBJECTIVE To estimate the impacts of a new EMR on various aspects of practice function using a Failure Modes and Effects Analysis (FMEA) approach based on the concept of hazard adapted from safety engineering. SETTING/PARTICIPANTS Academic rural primary care practice with 32 staff. DESIGN At baseline, a survey instrument (Perceived Hazard Questionnaire) was used to elicit staff (physicians, nurses and administrative) perceptions of frequency and severity of multiple different primary care errors in 12 different domains in the practice. For each error, a Hazard score was calculated based on the product of frequency and severity. The Hazard scores thus derived were used to prioritise the safety problems within the practice. One year later, after partial implementation of an EMR, the survey was repeated. MAIN OUTCOME MEASURES Comparison is made between priorities identified by physicians, nursing and administrative staff before and after EMR implementation. RESULTS At baseline, a high concordance between priorities identified by physicians, nursing and administrative staff was recorded. This concordance halved after partial implementation of the EMR. The staff perceived decreased hazard in nurse-physician and physician-chart interactions but hazard increased in the already high-hazard domains of physician-patient interaction in the assessment stage and nurse-chart interactions, apart from three other domains. CONCLUSIONS This FMEA-like approach identified changes in practice hazards apparently related to EMR implementation. This in turn can help in targeting pre-existing and new vulnerabilities in primary care practices.