Preventing critical failure. Can routinely collected data be repurposed to predict avoidable patient harm? A quantitative descriptive study.
After a cluster of perinatal deaths was identified in 2015, the authors assessed 15-years of routinely collected observational data from 7 different sources (administrative, patient complaint and legal data) preceding the cluster to determine whether the incidents could have been predicted and prevented. The extent of clinical activity along with direct-to-service patient complaints were found to be the more promising for purposes of potential predictive signals. The authors suggest that use of some routinely collected data of these types show promise; however, further work needs to be done on specificity and sensitivity of the data and to gain access to comparator data is needed.