Evaluating asthma control: a comparison of measures using an item response theory approach

Self-reported symptoms, FEV(1), and clinician judgment are all used to evaluate asthma control. The relative utility of each measure of control cannot be easily assessed. Item response theory (IRT) approaches allow for the direct comparison of the utility of different types of measures used to assess control. The objective of this study was to evaluate the validity and reliability of evaluating asthma control using symptom, clinical, and physiologic measures by applying an IRT approach. Subjects receiving care at an asthma clinic were evaluated on measures of asthma control. Based on 114 evaluations, IRT parameters were estimated to evaluate whether measures assessed a single underlying construct, the hierarchical relationship between the measures and the level of control each measure assessed, whether measures targeted all levels of asthma control, and whether the scoring categories distinguished between different levels of control. Infit statistics (0.74-1.5) for individual items showed that all items fit the underlying concept of asthma control. The reproducibility of the hierarchal scale was high (0.9). The results also demonstrated that items differentiated two strata (high, low) of control. The gaps in the hierarchal scale showed that for many subjects (37%) there were no items at their level of asthma control. The IRT approach identified gaps in current measurement that need to be addressed to provide more precise evaluations of control required to accurately monitor changes in patient status.