Physician Billing for Advance Care Planning Among Medicare Fee-For-Service Beneficiaries, 2016–2021

Abstract

Introduction

In 2016, the Centers for Medicare & Medicaid Services started reimbursing practitioners for their time spent providing advance care planning (ACP) with patients. This study assessed utilization of this policy by examining trends in ACP billing across medical specialties from 2016 to 2021 and differences in ACP service volume by metropolitan status.

Methods

The authors analyzed Centers for Medicare & Medicaid Services Medicare Provider Utilization and Payment files (2016–2021) from 6 specialty groups using corresponding Healthcare Common Procedure Coding System codes to derive trends in the percentage of practitioners billing any ACP visit. Wilcoxon tests were conducted to compare the average number of visits by metropolitan status.

Results

The percentage of practitioners billing ACP visits tripled from 1.76% in 2016 to 4.56% in 2021, with the highest percentage among hospice and palliative medicine practitioners (36.94%) in 2021. ACP service volume was similar by metropolitan status for hospice and palliative medicine, but it was higher in nonmetropolitan regions for cancer-related specialties, non-cancer terminal disease specialties, and primary and geriatrics care.

Discussion

This nationwide analysis showed low adoption of ACP billing by 2021, and it varied widely across specialties. This may reflect practical challenges of ACP related to comfort level with ACP discussion and documentation burden among the professional communities.

Conclusion

Despite an overall increase in the proportion of physicians billing ACP codes from 2016 to 2021, adoption remained low. Efforts are needed to address barriers to ACP and provide goal-concordant care to patients.

Introduction

Advance care planning (ACP) enables individuals to have meaningful discussions and make preparations for future medical decisions, especially for those facing serious illnesses or who may be unable to communicate their preferences. ¹ This patient-centered approach is recognized for its potential to enhance care quality and decrease the use of unnecessary aggressive treatments at the end of life. ² However, there is ongoing debate about the extent to which ACP ensures goal-concordant care and its overall value in end-of-life decision-making. ³ This misalignment between the intent of ACP and desired outcomes has been attributed to several practical challenges in clinical settings, including the quality of patient–practitioner communication, the dynamic nature of patient preferences, and the documentation and implementation burden associated with delivering goal-concordant care.

To incentivize the use of ACP, the Centers for Medicare & Medicaid Services introduced new ACP procedure codes in 2016 to reimburse practitioners for time engaged in ACP. ³ Although this policy led to an increase in ACP use among patients, evidence to date has shown that the overall utilization has remained low. For instance, a study using 20% Medicare claims showed that only 2% to 3% of Medicare beneficiaries had ACP claims from 2016 to 2017 ⁴ ; another study focusing on Medicare beneficiaries with serious illness showed only 11% had ACP toward the end of life. ¹ However, less is known about practitioners’ adoption of ACP services, who play an important role in initiating ACP conversations.

This study investigated trends in ACP billing across medical specialties from 2016 to 2021 and examined differences in ACP service volume by metropolitan and nonmetropolitan areas. By focusing on practitioner-level adoption, this research aimed to fill a critical gap in the literature and inform future design of interventions targeted toward physicians to promote equitable and widespread use of ACP services.

Methods

Data from the Centers for Medicare & Medicaid Services Medicare Provider Utilization and Payment files (2016–2021) were analyzed to identify practitioners who billed Medicare for ACP services using corresponding Healthcare Common Procedure Coding System codes (99497 and 99498). ³ Practitioners in 6 specialty groups that may care for patients needing ACP were included: hospice and palliative medicine (HPM), cancer-related specialties, non-cancer terminal disease specialties, primary care (PC), geriatrics, and advanced practice practitioners. The authors’ primary outcome was any billing for ACP services in a given year. ³ The authors also examined total ACP service volume among practitioners with any ACP billing in a year. The authors classified practitioners’ locations into metropolitan and nonmetropolitan areas based on their rural-urban commuting area (RUCA) codes. Specifically, practitioners with RUCA codes between 1 and 3 were classified as practicing in metropolitan areas, and practitioners with RUCA codes greater than 3 were classified as practicing in nonmetropolitan areas. The authors described trends in the percentage of practitioners with any ACP billing across different specialty groups. Wilcoxon tests were conducted to compare the average number of visits by metropolitan status. Statistical analyses were performed with SAS 9.4, and P values < .05 were considered as statistically significant.

This study was deemed by the Emory University Institutional Review Board as having nonhuman participants, using Emory Institutional Review Board’s self-determination worksheet, given the use of publicly available data.

Results

The analytic sample consisted of 2,688,409 practitioner-year observations from the years 2016 to 2021, including 7320 in HPM, 113,012 in cancer-related specialties, 333,771 in non-cancer terminal disease specialties, 1,061,877 in PC, 12,105 in geriatrics, and 1,160,324 advanced practice practitioners. The percentage of practitioners with any ACP billing tripled from 1.76% in 2016 to 4.56% in 2021, with the highest percentage among HPM practitioners (36.94%) in 2021, followed by geriatrics (14.40%), PC (7.66%), advanced practice practitioners (3.27%), cancer-related specialties (1.08%), and non-cancer terminal disease specialties (0.53%) (Figure 1).

Figure 1:

Percentage of Medicare Fee-For-Service practitioners billing any advance care planning by specialty groups, 2016–2021. Cancer-related specialties included hematology–oncology, medical oncology, radiation oncology, hematopoietic cell transplantation and cellular therapy, gynecological oncology, and surgical oncology. Non-cancer terminal disease specialties included cardiology and advanced heart failure specialties (including interventional and transplant cardiology), gastroenterology (with a focus on conditions like cirrhosis), nephrology (focusing on kidney failure), and pulmonary disease (especially terminal chronic obstructive pulmonary disease). Primary care included family practice, general practice, and internal medicine. Geriatrics care included geriatric medicine and geriatric psychiatry. Advanced practice practitioners included nurse practitioners and physician assistants. HPM = hospice and palliative medicine.

Among practitioners who billed any ACP service, the average ACP service volume was similar by metropolitan status among HPM (75.0 vs 66.7, P = .650), and it was higher in metropolitan areas among advanced practice practitioners (70.2 vs 54.8, P < .001). However, average ACP service volume was higher in nonmetropolitan areas among cancer-related specialties (53.6 vs 142.9, P < .001), non-cancer terminal disease specialties (82.8 vs 107.0, P = .200), PC (91.4 vs 99.3, P < .001), and geriatrics care (73.8 vs 87.0, P = .40). (Figure 2) Such differences by metropolitan status were persistent in earlier (2016–2018) and later (2019–2021) years of the study period (Figure 3A and B).

Figure 2:

Average service volume per practitioner and 95% confidence interval for advance care planning visits by specialty groups, 2016–2021. Cancer-related specialties included hematology–oncology, medical oncology, radiation oncology, hematopoietic cell transplantation and cellular therapy, gynecological oncology, and surgical oncology. Non-cancer terminal disease specialties included cardiology and advanced heart failure specialties (including interventional and transplant cardiology), gastroenterology (with a focus on conditions like cirrhosis), nephrology (focusing on kidney failure), and pulmonary disease (especially terminal chronic obstructive pulmonary disease). Primary care included family practice, general practice, and internal medicine. Geriatrics care included geriatric medicine and geriatric psychiatry. Advanced practice practitioners included nurse practitioners and physician assistants. Error bars represent standard deviation.

Figure 3:

Average service volume per practitioner and 95% confidence interval for advance care planning visits by specialty groups, 2016–2018 and 2019–2021. (A) 2016–2018. (B) 2019–2021. Cancer-related specialties included hematology–oncology, medical oncology, radiation oncology, hematopoietic cell transplantation and cellular therapy, gynecological oncology, and surgical oncology. Non-cancer terminal disease specialties included cardiology and advanced heart failure specialties (including interventional and transplant cardiology), gastroenterology (with a focus on conditions like cirrhosis), nephrology (focusing on kidney failure), and pulmonary disease (especially terminal chronic obstructive pulmonary disease). Primary care included family practice, general practice, and internal medicine. Geriatrics care included geriatric medicine and geriatric psychiatry. Advanced practice practitioners included nurse practitioners and physician assistants. Error bars represent standard deviation.

Discussion

In this nationwide analysis of Medicare Fee-For-Service practitioners from 2016 to 2021, the authors observed an increase in the proportion of physicians billing ACP codes, but adoption remained low in 2021 (~4% overall). This overall low utilization among practitioners may reflect practical challenges of ACP and recent debates among the professional communities. ⁵ First, patients must overcome emotional challenges and be willing to initiate discussions. Second, both patients and clinicians can effectively communicate about the goals of care. Third, busy clinicians must be aware of and willing to bill the dedicated ACP discussion code and its detailed documentation requirement. Finally, health systems must provide adequate support, including ensuring sufficient visit time and physician-level, quality-based incentives for ACP and goal-concordant care.

The highest ACP code adoption rate among HPM physicians may indicate a greater comfort level discussing ACP acquired by training, ⁶ or it may reflect greater knowledge of the existence of the code. Despite this, a substantial majority of HPM physicians (63%) did not bill for any ACP in 2021, indicating that the existing setup for ACP billing codes might not be very effective. The authors also observed a decline in average ACP service volume among cancer-related specialties from 2016–2018 to 2019–2021. This was consistent with the authors’ recent finding of decreasing oncologist involvement in palliative care billing among patients with advanced cancers. ⁷

In nonmetropolitan areas, the higher average ACP service volumes among non-HPM specialists may reflect a shortage of HPM specialists in these regions. Therefore, addressing barriers to ACP and realizing the full potential of goal-concordant care requires a comprehensive approach. This might include enhancing awareness among patients, providing more resources to support clinicians, providing more training for goal of care discussion, reducing documentation burden for clinicians across clinical specialties, and designing creative care models in nonmetropolitan areas. Future research also should account for differential clinical effectiveness of ACP delivery models led by HPM, other specialists, or nurses, across different care settings.

Limitations

This study had several limitations. First, the analysis relied on practitioner-level billing data, which did not include patient-level information that may have affected their use of ACP, such as diagnosis of severe illness and comorbidities. The authors tackled the potential confounding of patient mix by classifying practitioners into specialty groups with varying levels of likelihood of seeing patients with serious illness, including HPM specialists, cancer-related specialties, and non-cancer terminal disease specialties. Nevertheless, future studies using patient-level data are needed to understand how patient mix affects physicians’ billing of ACP. Second, due to limited sample sizes of practitioners who billed ACP codes in nonmetropolitan areas, the authors’ examination of ACP service volume was conducted for all study years, and it was stratified by 2 time periods (2016–2018 and 2019–2021). This may have masked potential year-to-year variations in ACP billing patterns. Future studies with larger sample sizes could explore more granular temporal trends in ACP service volume by metropolitan status. Finally, due to data limitation, the authors could not further distinguish between advanced practice practitioners with varying medical specializations that may have affected their adoption of ACP. Future research using more granular data on their practice settings would help to create a better understanding of how advanced practice practitioners contribute to ACP services across different specialty contexts.

Conclusion

Despite an overall increase in the proportion of physicians billing ACP codes from 2016 to 2021, overall adoption has remained low. Among the different specialties assessed, HPM physicians had the highest ACP code adoption rate. Future efforts are needed to address barriers to ACP adoption and improve the delivery of goal-concordant care to patients.