Modeling asthma exacerbations through lung function in children

Abstract

Background: Formal economic evaluation using a model-based approach is playing an increasingly important role in health care decision making.

Objective: To develop a model by using an objective measure of lung function-- prebronchodilator FEV(1) as a percent of predicted (FEV(1)% predicted)--as the primary independent factor to predict the frequency of adverse events related to the exacerbation of asthma on a population level.

Methods: We developed a Markov simulation model of childhood asthma by using data from the Childhood Asthma Management Program. The primary outcomes were the result of asthma exacerbations defined as hospitalizations, emergency department (ED) visits, and the need for oral corticosteroid therapy. Predicted monthly frequencies for each acute event were based on negative binomial regression equations estimated from the placebo arm of the Childhood Asthma Management Program with covariates of age, prebronchodilator FEV(1)% predicted, time in study, prior hospitalizations, and prior nocturnal awakenings.

Results: Simulated versus observed mean number of acute events were similar within the placebo and treatment groups. While the trial demonstrated treatment effects of 48% reduction in hospitalizations, 46% reduction in ED visits, and 44% reduction in the need for oral corticosteroid therapy at 48 months, the model simulated similar reductions of 49% in hospitalizations, 41% in ED visits, and 46% in the need for oral corticosteroid therapy.

Conclusions: Our findings suggest that longitudinal intervention effects may be modeled through FEV(1)% predicted to estimate hospitalizations, ED visits, and need for oral corticosteroid therapy in childhood asthma for planning and evaluation purposes.