Upcoding in Medicare Population; Evolving frequencies of comorbidities among Medicare beneficiaries hospitalized for Chronic Obstructive Pulmonary Disease

Introduction: The economic impact of upcoding on health care system has been profound. Current data of how chronic obstructive pulmonary disease (COPD) patients’ comorbidities have evolved over time are lacking. Objective: To describe the trends in COPD patients’ comorbidities over time and whether certain hospitals tend to report them more frequently than others.
Design. We used patient and hospital characteristics to build a two-level logistic regression model to predict mortality as a function of patients’ comorbidities. Settings: 5% Medicare database. Participants: 336,856 COPD hospitalizations over six-year period, 07/01/2009-06/31/2015. Methods: We divided comorbidities into three main categories based on their odds (ORs) of predicting mortality in the year prior to the study period. These were high-risk (OR>= 1.15), intermediate (0.85 =<OR<= 1.15) and low risk (OR<=0.85). We examined the yearly and interval changes (07/2013-06/2015 vs. 07/2009-06/2010) in comorbidities reporting. Then, we examined the interval changes in the average total number of comorbidities per patient. Finally, we examined variation in comorbidities reporting based on hospital characteristics. Results: While the patient mortality adjusted OR was not statistically significant (0.965; 95% CI 0.927-1.005), the hospital and patient mortality adjusted OR was (0.961; 95% CI 0.922-0.999). Thirty out of forty-one comorbidities have increased over the study period and the rest have declined. The average total number of comorbidities increased amongst all three categories (low, intermediate and high-risk). While the absolute change was highest for intermediate-risk one (0.51) and similar for low and high-risk ones (0.18 and 0.17), the percent change for high (8.42%) and intermediate-risk (8.93%) comorbidities were almost twice the percent change in low-risk category (3.88%). Findings were similar when data were re-analyzed based on hospital characteristics. Conclusion: The high-risk and intermediate-risk comorbidities had more than 8% increase and the low-risk comorbidities had almost 4% increase over the study period. Evaluation of other diagnoses and different time periods may shed light on whether patients are getting merely sick or there is upcoding in which hospitals and/or providers report high and intermediate-risk comorbidities more frequently.