Improved pain resolution in hospitalized patients through targeting of pain mismanagement as medical error.

Current strategies to reduce excess pain among hospitalized patients remain inadequate. New, effective approaches are urgently needed. In this prospective observational study of a performance-improvement intervention, we studied the effect of computer-generated, real-time alerts used by nurses on the rate of a medical error in pain management defined as lack of reassessment within 120 minutes from the last observation of severe pain. We also studied duration of severe pain events and frequency of treatment of opioid-related adverse effects. Analyses of 51,619 consecutive observations of severe pain were performed in monthly intervals. Significant decrease in error rate (delayed pain reassessment) was observed postintervention (mean+/-standard error [SE]: 35.8%+/-0.7%) compared with preintervention rate (56.2%+/-1.4%, P<0.0001). Among 6305 unique severe pain events examined during four months pre- and postintervention, time to resolution of severe pain decreased significantly (median time preintervention [January 2006] of 195 minutes compared with median time postintervention of 117, 106, and 101 minutes [January, April, and June 2007], P<0.0001). Hospital-wide, unanticipated monthly naloxone administration decreased postintervention (mean+/-SE: 1.48+/-0.21 per month per 1000 inpatients) compared with preintervention (2.69+/-0.35, P=0.0130). Hospital-wide implementation of real-time, computer-generated alerts identifying instances of delayed pain reassessment resulted in sustained reduction of error rate and faster resolution of severe pain without oversedation.