Predicting future risk of asthma exacerbations using individual conditional probabilities.

Abstract

Determination of future risk of exacerbations is a key issue in the management of asthma. We previously developed a method to calculate conditional probabilities (π) of future decreases in lung function by using the daily fluctuations in peak expiratory flow (PEF). We aimed to extend calculation of π values to individual patients, validated by using electronically recorded data from 2 past clinical trials. Twice-daily PEF data were analyzed from 78 patients with severe (study A) and 61 patients with poorly controlled (study B) asthma. For each patient, the π value was calculated from 5000 PEF data points simulated based on the correlation and distribution properties of observed PEF. Given an initial PEF, the π value was defined as the probability of a decrease in PEF to less than 80% of predicted value on 2 consecutive days within a month. These probabilities were then compared with actual occurrences of such events and clinically defined exacerbations within the following month. π Values were related to actual occurrences of decreases in PEF (adjusted R(2) > 0.800 for both studies). Every increase of 10% in π value was associated with an odds ratio of having a future exacerbation of 1.24 (95% CI, 1.07-1.43) for study A and 1.13 (95% CI, 1.02-1.26) for study B, with better sensitivity and specificity than clinic-measured FEV(1). These results from 2 independent datasets with differing asthmatic populations and differing exacerbation criteria provide support that clinically relevant quantification of individual future risk of exacerbations is possible.