Clinical utility of scores on the Blaylock Risk Assessment Screen (BRASS): An analysis of administrative data

Purpose Project was undertaken to examine the utility of the Blaylock Risk Assessment Screen (BRASS) in identifying patients who may experience discharge complications as indicated by longer hospital stays or readmission within 30-days of a discharge to home. Background Before measures can be put in place to facilitate discharge planning and to prevent unplanned readmission by recently discharged patients, those at risk of such events must be identified. Methods Project involved an analysis of 13-months of administrative data from one tertiary care hospital. Utility of the BRASS was examined in terms of its sensitivity and specificity as well as its positive and negative predictive values. Results Majority (83%) of hospital discharges were to home. Approximately 7% of patients experienced at least one readmission within 30-days of being discharged to home. Using scores of 10 or higher as an indicator of risk, BRASS exhibited a high degree of specificity suggesting it is useful for ‘ruling in’ those who have the outcomes-of-interest. However low sensitivity indicates many who experienced the outcomes were incorrectly classified by the BRASS as low risk. The low positive predictive value for 30-day readmission also suggests many who were classified by the BRASS as being ‘at risk’ were not readmitted. Conclusion The observed rate of 30-day readmission is likely conservative as the analysis involved data from only one acute care facility. One explanation for the low positive predictive value for 30-day readmission is that completion of the BRASS on admission enabled the implementation of preventive measures.