A Cox Regression Analysis of Covariates for Asthma Hospital Readmissions

Abstract

Background. Asthma hospital admissions and readmissions are unacceptably high, thus, a method to identify those at greatest risk could be helpful. Methods. An observational retrospective study using a Cox regression to determine the relationship between the time interval between admissions and possible covariates of a readmission. The covariates were age, sex, ethnicity, smoking habit, history of allergy or eczema/hay fever, age of onset, Townsend index (TI), Jarman score (JS), and drugs on discharge. Those with p < 0.2, together with interacting covariates, from the preliminary analysis were eligible for the multivariate Cox regression analysis. Results. Of the 523 patients admitted between 1994 and 1998 because of their asthma, complete data were available for 440. Of these, 112 were readmitted. Eligible covariates for the multivariate Cox regression analysis were sex, allergy status, history of eczema/hay fever, the JS and TI together with interactions between JS and TI, JS and allergy, and allergy with eczema/hay fever. There were 278 subjects (71 with a readmission) with complete data for these eligible covariates. The multivariate analysis revealed that female sex (odds ratio [OR] = 2.65, 95% confidence interval [CI] 1.42, 4.92), high JS (OR = 2.03, 95% CI 1.13-3.65), and history of allergy (OR= 1.88, 95% CI 1.06-3.32) formed the final model as significant predictors of readmission. Conclusion. Females with a history of allergy that were registered at a practice with a high workload (JS) had a higher risk of readmission. The analysis method used highlights how those at risk of readmission can be identified so that they can be targeted post discharge.