Assessment of First-Year Use of Medicare’s Advance Care Planning Billing Codes

Abstract

This observational analysis characterizes the first year of use of the Medicare code for advance care planning and describes beneficiaries most likely to receive advance care planning.

Advance care planning (ACP) conversations occur infrequently among patients and their health care professionals, and when they do occur, the context is often a stressful clinical situation.^1,2 Advance care planning conversations that occur too late (or not at all) can result in care that is invasive, expensive, and not aligned with patients’ wishes. To encourage health care professionals to initiate ACP discussions, Medicare began reimbursing for ACP services on January 1, 2016, under a separate billing code. Our study aims to characterize the first year of ACP code use and to describe beneficiary characteristics associated with receipt of ACP services.

Methods

For this retrospective observational analysis, we used Medicare fee-for-service Parts A and B claims, the Medicare coverage and denominator files, and US Census Bureau data. All beneficiaries in New England (Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont) continuously enrolled in Parts A and B in 2016 were included in this sample. Beneficiaries who died during the study year were included until their death. The authors have access to Medicare claims data under an agreement between Healthcentric Advisors and the Centers for Medicare & Medicaid Services for use of these data to evaluate improvement in care quality. Institutional review board review is not required in these circumstances, per organizational policy.

Our primary outcome was the number of Medicare fee-for-service beneficiaries with an ACP claim in 2016. Secondary outcomes included beneficiary-level estimates of having a claim and the number of beneficiaries whose ACP service occurred during their annual wellness visit. We used χ² tests to compare beneficiary characteristics by ACP claim status. Using multivariable logistic regression, we identified variables associated with a beneficiary having any ACP claim, adjusting for age, sex, race/ethnicity, state of residence, dual-eligibility status (for Medicare and Medicaid), median income of residence, and the presence of certain chronic conditions. Findings were considered significant at P < .05 (2-sided).

Results

In 2016, 26 522 ACP claims were billed by 1996 health care professionals in New England. These claims occurred among 24 536 unique beneficiaries (Table 1), representing fewer than 1% of the total beneficiaries in New England. Most ACP claims were billed in the office setting (16 755 [63.2%]), followed by nursing home (4671 [17.6%]) and inpatient settings (2671 [10.1%]). A total of 10 622 ACP services (40.0%) occurred during an annual wellness visit.

Table 1. Use of ACP Services Among Medicare Fee-for-Service Beneficiaries in New England in 2016.

State	Total No. of Beneficiaries	Total No. of ACP Claimsa	No. of Health Care Professionals With ACP Claims	No. of Beneficiaries With ACP Claims	No. of Beneficiaries With ACP Claims per 1000 Beneficiaries
New England	2 481 762	26 522	1996	24 536	9.89
Connecticut	563 210	7512	367	7109	12.62
Maine	284 378	1118	186	1036	3.64
Massachusetts	1 094 216	14 072	1040	12 997	11.88
New Hampshire	239 738	965	163	917	3.83
Rhode Island	179 324	2526	157	2167	12.08
Vermont	120 896	329	83	310	2.56

Abbreviation: ACP, advance care planning.

^aIncludes ACP code 99497 only; ACP add-on code 99498 is not included.

Among all 2016 ACP claims, 5014 (18.9%) were billed by advanced practice clinicians and 21 508 (81.1%) by physicians. Most of the physicians were from internal medicine (14 025 [65.2%]), followed by family medicine (4764 [22.1%]) and geriatrics (1111 [5.2%]). Oncology and hematology specialists billed 92 total claims (0.3%).

Compared with beneficiaries without an ACP claim, greater proportions of those with a claim were female (59.2% with vs 56.1% without), 85 years or older (27.1% vs 12.9%), enrolled in hospice (10.5% vs 2.2%), and classified as deceased during the study year (15.1% vs 4.2%). Fewer beneficiaries with a claim had dual eligibility compared with those without (18.0% vs 24.0%).

In our fully adjusted model (Table 2), older age (OR for ≥85 years, 0.86; 95% CI, 0.83-0.90), lower income (OR, 1.36; 95% CI, 1.30-1.42), and the presence of certain cancers (OR, 1.52; 95% CI, 1.47-1.58), heart failure (OR, 1.69; 95% CI, 1.64-1.75), stroke (OR, 1.37; 95% CI, 1.31-1.42), chronic kidney disease (OR, 1.91; 95% CI, 1.85-1.97), and dementia (OR, 2.16; 95% CI, 2.09-2.24) were associated with higher odds of having a claim. Being male (OR, 0.63; 95% CI, 0.58-0.68), Asian (OR, 0.84; 95% CI, 0.73-0.96), black (OR, 0.63; 95% CI, 0.58-0.68), and Hispanic (OR, 0.40; 95% CI, 0.34-0.48) and having dual eligibility (OR, 0.73; 95% CI, 0.70-0.75) were associated with lower odds of having a claim.

Table 2. Unadjusted and Adjusted OR Estimates of an ACP Claim Among New England Medicare Fee-for-Service Beneficiaries^a.

Characteristic	Unadjusted OR (95% CI)	P Value	Adjusted OR (95% CI)	P Value
Age, y
<65	0.21 (0.20-0.22)	<.001	0.48 (0.45-0.51)	<.001
65-74	0.37 (0.36-0.38)	<.001	0.71 (0.68-0.73)	<.001
75-84	0.60 (0.58-0.62)	<.001	0.86 (0.83-0.90)	<.001
≥85	1 [Reference]		1 [Reference]
Sex
Male	0.88 (0.86-0.91)	<.001	0.90 (0.88-0.93)	<.001
Female	1 [Reference]		1 [Reference]
Race
Asian	0.75 (0.65-0.86)	<.001	0.84 (0.73-0.96)	.01
Black	0.74 (0.69-0.79)	<.001	0.63 (0.58-0.68)	<.001
Hispanic	0.34 (0.29-0.41)	<.001	0.40 (0.34-0.48)	<.001
White	1 [Reference]		1 [Reference]
Other		<.001	0.839 (0.771-0.914)	<.001
Dual eligibility for Medicare and Medicaid	0.70 (0.68-0.72)	<.001	0.73 (0.70-0.75)	<.001
Median income of residence, $
<47 700	0.79 (0.76-0.82)	<.001	1.36 (1.30-1.42)	<.001
47 700-71 550	0.91 (0.88-0.93)	<.001	1.22 (1.19-1.26)	<.001
>71 550	1 [Reference]		1 [Reference]
State of residence
Connecticut	1.06 (1.03-1.09)	<.001	1.10 (1.07-1.14)	<.001
Maine	0.284(0.26-0.30)	<.001	0.28 (0.26-0.30)	<.001
Massachusetts	1 [Reference]		1 [Reference]
New Hampshire	0.28 (0.26-0.3060)	<.001	0.26 (0.25-0.28)	<.001
Rhode Island	0.91 (0.87-0.96)	<.001	0.97 (0.92-1.02)	.22
Vermont	0.17 (0.15-0.20)	<.001	0.17 (0.15-0.19)	<.001
Presence of chronic condition
Cancerb	2.16 (2.09-2.23)	<.001	1.52 (1.47-1.58)	<.001
Stroke or TIA	2.75 (2.65-2.85)	<.001	1.37 (1.31-1.42)	<.001
Chronic kidney disease	3.04 (2.96-3.12)	<.001	1.91 (1.85-1.97)	<.001
Heart failure	3.54 (3.44-3.65)	<.001	1.69 (1.64-1.75)	<.001
Dementia	4.08 (3.96-4.21)	<.001	2.16 (2.09-2.24)	<.001

Abbreviations: ACP, advance care planning; OR, odds ratio; TIA, transient ischemic attack.

^aIncludes 2 481 762 beneficiaries.

^bIncludes breast, colorectal, prostate, lung, and endometrial cancers.

Discussion

Results of this study found limited use of ACP services in 2016 and differences in use among various subgroups, including by race/ethnicity and by dual eligibility status, even after adjustment. Health care professionals may be having ACP conversations and missing out on reimbursement or not discussing these issues with patients at all,³ even those beneficiaries who would likely benefit the most. Our findings can inform interventions to increase awareness and use of ACP services and can encourage strategies for incorporating these codes into clinician workflows. We suggest focusing on the opportunities for increased uptake among physicians caring for dually eligible and nonwhite beneficiaries, among hematologists and oncologists, and during annual wellness visits, when beneficiaries would not be responsible for co-payments.