Background: The US has the most expensive healthcare system globally.1 Recent claims have raised concerns that hospitalists may be in part contributing to higher levels of high intensity billing over time through upcoding.2 However, there is little empirical evidence on this topic. As the number of hospitalists continues to grow and hospitalists disproportionately care for general medicine admissions, it is important to determine the extent to which they may be driving high-intensity billing trends.
Methods: We used a sample of Medicare Part A and B claims of Medicare fee-for-service beneficiaries from 2009 to 2018. Hospitalists were defined as physicians with primary specialties in Internal Medicine, General Practice, Family Practice, or Hospitalist where 90% of their E/M claims within a given year were for inpatient services. Our primary outcome was high severity billing codes for patients admitted by a hospitalist or non-hospitalist across initial, subsequent, and discharge encounters. We compared the characteristics of patients admitted by hospitalists vs. non-hospitalists in 2011 and in 2018. We used a multivariable linear regression model with the patient as the unit of analysis and whether high-severity billing occurred during the initial hospital encounter as the primary outcome. The primary exposure variable was the type of physician, year, and the interaction term between physician type and year. Models included covariates for age, sex, dual status, race/ethnicity, Elixhauser comorbidities, and MS-DRGs. Hospital fixed effects were included to account for correlation within the hospital over time. Identical analyses were run for high-severity billing of subsequent and discharge encounters.
Results: The number of general medicine physicians classified as hospitalists grew from 23,390 in 2009 to 41,084 in 2018, while non-hospitalists decreased from 119,102 to 116,961. There were no meaningful differences in patient’s characteristics treated by hospitalists vs. non-hospitalists. In 2009, 69.9% of initial encounters by hospitalists were billed as high severity compared to 63.0% by non-hospitalists (+6.9%, 95% CI 6.6% – 7.3%, p< 0.001). In 2018, 69.1% of initial encounters were billed as high severity by hospitalists compared to 58.4% by non-hospitalists (+10.7%, 95% CI 10.5% - 10.9%, p< 0.001). In 2009, hospitalists billed 32.9% of subsequent encounters as high severity compared to 30.1% by non-hospitalists (+2.8%, 95% CI 2.6% - 3.0%, p< 0.001). By 2018, hospitalists billed 39.9% of subsequent encounters as high severity vs 33.6% by non-hospitalists. (+6.3%, 95% CI 6.2% - 6.4%, p< 0.001). In 2009, 48.2% of discharge encounters were billed as high severity compared to 34.4% by non-hospitalists (+14.0%, 95% CI 13.7% - 14.4%, p< 0.001). By 2018, 70.6% of discharge encounters were billed as high severity compared to 45.3% for non-hospitalists. (+25.3%, 95% CI 25.1% - 25.5%, p< 0.001).
Conclusions: High-severity billing is much higher for hospitalists than non-hospitalists in 2009-2018, and these gaps are growing over time, especially for subsequent and discharge encounters. These differences do not appear to be explained by substantial differences in patient complexity of patients treated across US hospitals. The increases in hospitalists over time may be contributing to rising costs related to hospital care across the country. Future work in this area would be helpful to determine drivers of high-intensity billing, and importantly, develop strategies to mitigate potential up-coding.
