Characterizing and predicting rates of delirium across general hospital settings.

Abstract

To better understand variation in reported rates of delirium, this study characterized delirium occurrence rate by department of service and primary admitting diagnosis. Nine consecutive years (2005-2013) of general hospital admissions (N=831,348) were identified across two academic medical centers using electronic health records. The primary admitting diagnosis and the treating clinical department were used to calculate occurrence rates of a previously published delirium definition composed of billing codes and natural language processing of discharge summaries. Delirium rates varied significantly across both admitting diagnosis group (X210=12786, p<0.001) and department of care (X26=12106, p<0.001). In both cases obstetrical admissions showed the lowest incidences of delirium (86/109764; 0.08%) and neurological admissions the greatest (2851/25450; 11.2%). Although the rate of delirium varied across the two hospitals the relative rates within departments (r=0.96, p<0.001) and diagnostic categories (r=0.98, p<0.001) were consistent across the two institutions. The frequency of delirium varies significantly across admitting diagnosis and hospital department. Both admitting diagnosis and department of care are even stronger predictors of risk than age; as such, simple risk stratification may offer avenues for targeted prevention and treatment efforts.