Development and Validation of Personalized Prediction to Estimate Future Risk of Severe Exacerbations and Uncontrolled Asthma in Patients with Asthma, Using Clinical Parameters and Early Treatment Response

Abstract

Current level of asthma control can be easily assessed by validated instruments, but it is currently difficult to assess individuals' level of future risk. Develop, and validate, a risk prediction score for level of future risk, including patient characteristics and information on early treatment response. We used data of 304 adult patients with asthma from a 12-month primary care randomized controlled trial with 3-monthly assessments. With logistic regression we modeled the association between the level of future risk and patient characteristics including early treatment response. Future risk was defined as Asthma Control Questionnaire (ACQ) score of 1.5 or more at 12 months or the experience of at least 1 exacerbation during the final 6 months. We developed a risk prediction score on the basis of regression coefficients. Performance of the risk prediction score improved, taking into account data on early treatment response (area under receiver-operating curve [AUROC]= 0.84) compared with a model containing only baseline characteristics (AUROC= 0.78). The score includes 6 easy-to-obtain predictors: sex, ACQ score and exacerbations in the previous year at baseline and at first follow-up, and smoking status and exacerbations in the previous 3 months (indicating early treatment response). External validation yielded an AUROC of 0.77. The risk prediction score classified patients into 3 risk groups: low (absolute risk, 11.7%), intermediate (47.0%), and high (72.7%). We developed and externally validated a risk prediction score, quantifying both level of current asthma control and the guideline-defined future risk. Patients' individual risk can now be estimated in an easy way, as proposed but not specified, by asthma management guidelines.