A Model for Early Identification of Chronic Obstructive Pulmonary Disease

Chronic obstructive pulmonary disease (COPD) is highly underdiagnosed. Early diagnosis is, therefore, essential to reduce costs and exacerbations and prevent disease progression. This calls for the development of a COPD case-finding tool. The present study aimed to develop a model for early identification of COPD with an eye to optimizing COPD case finding. The study was based on data from the US National Health and Nutrition Examination Survey (NHANES) dataset (2007-2012). For the analysis, 772 participants with spirometry defined COPD were included. Potential predictors for COPD (n=42) were extracted. The model was based on logistic regression, and the predictors were included using a stepwise forward selection. A five-fold cross-validation was used to train and validate the model. The predictors included age, gender, and pack-years of smoking. The model obtained an AUC of 0.71. In conclusion, such a model can be useful for identifying individuals that should perform post-bronchodilator spirometry to aid early identification of COPD.