Development and Validation of EPIC-US: A Reference Model for COPD Policy Evaluation in the United States

The large and growing burden of chronic obstructive pulmonary disease (COPD) requires evidence-based policies to improve patient outcomes and reduce healthcare costs. We developed the Evaluation Platform in COPD in the U.S. (EPIC-US), a whole-disease model of COPD capable of evaluating interventions across all stages of disease progression within a single policy platform.

The objectives were to develop EPIC-US by adapting a model version developed for Canada (EPIC) and internally and externally validate EPIC-US to an independent clinical cohort.

EPIC-US is a stochastic discrete-event simulation model of COPD for U.S. individuals aged ≥40 years. It simulates COPD progression based on individualized trajectories of lung function decline and exacerbations. We matched population size to U.S. census projections for 2015-2060. All-cause mortality was taken from U.S. life tables. Historical and projected trends in tobacco smoking prevalence were aligned with 2018 and 2023 National Health Interview Survey (NHIS) results, while COPD prevalence was informed by National Health and Nutrition Examination Survey (NHANES) data. We used clinical trials, meta-analyses and U.S. administrative data to inform the rate and severity of exacerbations. We externally validated moderate exacerbations and severe exacerbations using the U.S. sub-sample of the NOVEL observational longiTudinal studY (NOVELTY).

EPIC-US simulated outputs closely matched module-specific internal validation targets. The predicted population size of ≥ 40-year-olds was within 1-2% of U.S. Census projections. Predicted mean COPD prevalence (8.1% for ages 40-59 years; 14.4% for ages 60-79 years), and smoking prevalence in 2018 (12.9%) and 2023 (10.1%) were within 0.1-0.4% of NHIS and NHANES estimates. In 2016, EPIC-US predicted 560 hospitalizations for COPD per 100,000 person-years, consistent with the observed rate of 510 hospitalizations. In external validation, severe exacerbations rates in EPIC-US (0.039; 95% CI 0.035-0.043 per patient-year) were comparable to NOVELTY (0.022 per patient-year, 95% CI 0.008-0.065). EPIC-US replicated differences in exacerbation rate across classes of inhaled therapies. However, EPIC-US overpredicted the rate of moderate exacerbations relative to NOVELTY.

EPIC-US is the first multi-purpose, whole-disease model of COPD in the U.S., providing a policy-ready platform for simulating both the clinical and economic burden of the disease.