Utility of advance care planning billing codes: a cross-sectional analysis of U.S. hospitalists' documentation.

objectiveTo establish the validity of Advance Care Planning (ACP) billing codes 99497 and 99498 to measure ACP occurrence in the hospital.

resultsAs part of a clinical trial testing the effect of a physician-directed intervention on ACP for hospitalized older adults, we conducted a retrospective chart review of physician notes to assess validity of ACP billing codes on Medicare claims. Among 434 inpatient stays managed by 39 hospitalists at 35 hospitals, 85 (19.6%) had chart documentation of ACP, and 349 (80.4%) did not. Of the charts with documented ACP, 49 had a bill for ACP (57.6% sensitivity). Of the charts without evidence of ACP, there were no ACP (100% specificity). Qualitative analysis of the documentation found that 39 (46%) focused on guiding treatment during the current hospitalization and not on establishing preferences for future care. ACP bills capture clinician performance specifically but not sensitively, which means using claims data will underestimate clinician ACP practice. Further, billed inpatient ACP conversations were often related to guiding treatment during the current hospitalization, not on establishing preferences for future care.