D102-02 Mapping the Zone of Uncertainty in Pulmonary Function Test Interpretation

Rationale The European Respiratory Society and American Thoracic Society technical standard on pulmonary function test (PFT) interpretation posits the existence of a “zone of uncertainty” defined by the overlap of distributions of healthy and diseased lung function. As PFT parameter values that fall within this zone may represent either health or disease, such values should be interpreted with caution. However, as reference equations have been defined only for healthy lung function, the precise borders of this zone remain unspecified. To address this limitation, we used a latent class model to define distributions of healthy and diseased lung function and used these distributions to map the zone of uncertainty. Methods In this exploratory study we considered a single PFT parameter, the forced expiratory volume in 1 second to forced vital capacity (FEV1/FVC) ratio. We fit a Box-Cox Cole and Green (BCCG) distribution to spirometry from healthy adult National Health and Nutrition Examination Survey (NHANES) participants. Using this distribution of healthy lung function, we then fit a latent class model to FEV1/FVC values from adult patients from the University of Pennsylvania Health System (UPHS), with the distribution of diseased FEV1/FVC values and the prior probability that an FEV1/FVC value was healthy or diseased chosen to maximize the likelihood of the UPHS data. We mapped the zone of uncertainty by identifying the range of FEV1/FVC values in which the distributions of healthy and diseased lung function overlapped. Results Spirometry data were collected from 43,274 UPHS patients (14.6% with a chronic obstructive pulmonary disease [COPD] diagnosis). The median FEV1/FVC was 0.77 (interquartile range [IQR] 0.14). A Box-Cox Power Exponential distribution best represented diseased lung function, with a median of 0.57 (Figure). The prior probability of a healthy FEV1/FVC value was 0.75. The overlap of the distributions of healthy and diseased lung function defined the zone of uncertainty by the interval [0.64, 0.95]. The FEV1/FVC values of 35,069 (81.0%) UPHS patients fell within this zone, including 31,343 (99.2%) patients with a normal FEV1/FVC, 3,726 (31.9%) patients with an abnormal FEV1/FVC, and 2,783 (44.0%) patients with a COPD diagnosis (44.0%). Conclusions Using a latent class model to define distributions of healthy and diseased lung function, we found that the zone of uncertainty encompassed almost all of the FEV1/FVC values in our clinical cohort. Uncertainty is pervasive and further research is needed to understand how best to incorporate it into PFT interpretation.