COMPUTER-BASED MODELS TO IDENTIFY HIGH-RISK CHILDREN WITH ASTHMA

Effective management of populations with asthma requires methods for i dentifying patients at high risk for adverse outcomes. The aim of this study was to develop and validate prediction models that used compute rized utilization data from a large health-maintenance organization (H MO) to predict asthma-related hospitalization and emergency department (ED) visits. In this retrospective cohort design with split-sample va lidation, variables from the baseline year were used to predict asthma -related adverse outcomes during the follow-up year for 16,520 childre n with asthma-related utilization. In proportional-hazard models, havi ng filled an oral steroid prescription (relative risk [RR]: 1.9; 95% c onfidence interval [CI]: 1.3 to 2.8) or having been hospitalized (RR: 1.7; 95% CI: 1.1 to 2.7) during the prior 6 mo, and not having a perso nal physician listed on the computer (RR: 1.6; 95% CI: 1.1 to 2.3) wer e associated with increased risk of future hospitalization. Classifica tion trees identified previous hospitalization and ED visits, six or m ore beta-agonist inhalers (units) during the prior 6 mo, and three or more physicians prescribing asthma medications during the prior 6 mo a s predictors. The classification trees performed similarly to proporti onal-hazards models, and identified patients who had a threefold great er risk of hospitalization and a twofold greater risk of ED visits tha n the average patient. We conclude that computer-based prediction mode ls can identify children at high risk for adverse asthma outcomes, and may be useful in population-based efforts to improve asthma managemen t.