β-Amyloid Positron Emission Tomography Use Across Medicare Administrative Contractor Regions

Abstract

This cross-sectional study assesses variation in β-amyloid positron emission tomography (PET) use among Medicare beneficiaries with mild cognitive impairment or dementia by Medicare Administrative Contractor region.

β-Amyloid (Aβ) detected via positron emission tomography (PET) is the less invasive of 2 ways to confirm eligibility for monoclonal antibodies for treating Alzheimer disease (AD).¹ Historically, the Centers for Medicare & Medicaid Services (CMS) provided coverage for Aβ PET under coverage with evidence development, resulting in noncoverage beyond clinical studies.² On October 13, 2023, CMS ended this policy for Aβ PET, permitting coverage determinations to be made by Medicare Administrative Contractors (MACs). While manufacturers may expect broad coverage of Aβ PET, it is not clear how MACs will apply the new policy for coverage decisions.³ We assessed variation in Aβ PET utilization among Medicare beneficiaries with mild cognitive impairment (MCI) or dementia by MAC Part A and B regions after the policy change.

Methods

This cross-sectional study used 100% Medicare fee-for-service (FFS) data from October 1, 2023, through March 31, 2024. This database includes enrollment and claims data for beneficiaries continuously enrolled in Medicare FFS coverage for Parts A and B. This study was determined not to be human participant research by the WCG institutional review board and followed the STROBE reporting guideline.

The study period was divided into 2 quarters after the policy change to examine how variation changed over time: October 1 through December 31, 2023, and January 1 through March 31, 2024 (eMethods in Supplement 1). Current Procedural Terminology (CPT) and Healthcare Common Procedure Coding System (HCPCS) codes were used to define Aβ PET scans; 1 or more CPT code (78811, 78814) for PET of chest, head, or neck and 1 or more HCPCS codes (A9586, Q9982, or Q9983) for an Aβ tracer were required on the same claim. The denominator in each quarter was the total number of beneficiaries with at least 1 month of enrollment in Parts A and B (and absence of Part C) and with an ICD-10 medical claim including a diagnosis code for MCI or dementia in any position (eMethods in Supplement 1) in the quarter. The numerator of Aβ PET scans was assessed in months within a quarter when the enrollment criteria were met.

Sociodemographic characteristics (age, sex, and race and ethnicity [ascertained using the Research Triangle Institute race code⁴]) were reported for each quarter overall and by Aβ PET examination. Prevalence of Aβ PET among beneficiaries with MCI or dementia was estimated from separate fitted logistic regression models for each quarter using predictive margins adjusted for age, sex, and race and ethnicity within each of the 12 MAC regions (eMethods in Supplement 1).⁵ Analyses were performed with SAS, version 7.15.

Results

Beneficiaries with MCI or dementia who received Aβ PET (vs those who did not) were younger, and fewer were Black (Table 1). From October through December 2023, the adjusted prevalence of Aβ PET among beneficiaries with MCI or dementia varied from 7 (jurisdiction 8) to 176 (jurisdiction E) per 100 000 beneficiaries (Table 2). From January through March 2024, the adjusted prevalence of Aβ PET among beneficiaries with MCI or dementia varied from 31 (jurisdiction 15) to 613 (jurisdiction N) per 100 000 beneficiaries.

Table 1. Characteristics of Medicare Fee-for-Service Beneficiaries With MCI or Dementia Overall and by Receipt of Amyloid PET^a.

Characteristic	October 1 to December 31, 2023			January 1 to March 31, 2024
	Overall (N = 1 302 850)	Amyloid PET		Overall (N = 1 318 239)	Amyloid PET
		Yes (n = 908)	No (n = 1 301 942)		Yes (n = 2951)	No (n = 1 315 288)
Age, mean (SD), y	80.6 (9.5)	75.4 (6.3)	80.6 (9.5)	80.6 (9.5)	76.1 (6.1)	80.6 (9.5)
Sex
Femaleb	814 281 (62.5)	499 (55.0)	813 782 (62.5)	822 431 (62.4)	1640 (55.6)	820 791 (62.4)
Male	488 569 (37.5)	409 (45.0)	488 160 (37.5)	495 808 (37.6)	1311 (44.4)	494 497 (37.6)
Race and ethnicityc
Asian	38 758 (3.0)	32 (3.5)	38 726 (3.0)	39 202 (3.0)	79 (2.7)	39 123 (3.0)
Black	110 354 (8.5)	37 (4.1)	110 317 (8.5)	111 262 (8.4)	86 (2.9)	111 176 (8.5)
Hispanic	77 641 (6.0)	51 (5.6)	77 590 (6.0)	79 304 (6.0)	115 (3.9)	79 189 (6.0)
Otherd	33 355 (2.5)	40 (4.4)	33 315 (2.5)	35 106 (2.7)	129 (4.4)	34 977 (2.6)
White	1 042 742 (80.0)	748 (82.4)	1 041 994 (80.0)	1 053 365 (79.9)	2542 (86.1)	1 050 823 (79.9)

Abbreviations: MCI, mild cognitive impairment; PET, positron emission tomography.

^a Data are presented as number (percentage) of beneficiaries with MCI or dementia unless otherwise indicated.

^b Imputation to the modal sex value (ie, female) was done for fewer than 11 beneficiaries with unknown sex from October 1 through December 31, 2023. The exact number of imputed values was redacted due to the Centers for Medicare & Medicaid Services patient confidentiality policy.

^c Race and ethnicity were ascertained using the Research Triangle Institute race code⁴ and were included in the analysis to describe beneficiaries who received amyloid PET.

^d American Indian or Alaska Native, other, and unknown.

Table 2. Amyloid PET Utilization Among Beneficiaries With MCI or Dementia by MAC Region.

MAC regiona	Estimated prevalence of amyloid PET, per 100 000 beneficiariesb
	October 1 to December 31, 2023	January 1 to March 31, 2024
Jurisdiction 5	42	138
Jurisdiction 6	27	129
Jurisdiction 8	7	92
Jurisdiction 15	9	31
Jurisdiction E	176	430
Jurisdiction F	31	148
Jurisdiction H	48	165
Jurisdiction J	33	117
Jurisdiction K	56	222
Jurisdiction L	100	190
Jurisdiction M	105	243
Jurisdiction N	129	613

Abbreviations: MAC, Medicare Administrative Contractor; MCI, mild cognitive impairment; PET, positron emission tomography.

^a Jurisdiction 5: Iowa, Kansas, Missouri, and Nebraska. Jurisdiction 6: Illinois, Minnesota, and Wisconsin. Jurisdiction 8: Indiana and Michigan. Jurisdiction 15: Kentucky and Ohio. Jurisdiction E: American Samoa, California, Guam, Hawaii, Nevada, and Northern Mariana Islands. Jurisdiction F: Alaska, Arizona, Idaho, Montana, North Dakota, Oregon, South Dakota, Utah, Washington, and Wyoming. Jurisdiction H: Arkansas, Colorado, Louisiana, Mississippi, New Mexico, Oklahoma, and Texas. Jurisdiction J: Alabama, Georgia, and Tennessee. Jurisdiction K: Connecticut, Maine, Massachusetts, New Hampshire, New York, Rhode Island, and Vermont. Jurisdiction L: Delaware, District of Colombia, Maryland, New Jersey, and Pennsylvania. Jurisdiction M: North Carolina, South Carolina, Virginia, and West Virginia. Jurisdiction N: Florida, Puerto Rico, and US Virgin Islands.

^b Adjusted for (but not conditional on) age, sex, and race and ethnicity within each MAC region.

Discussion

Despite the goal of CMS that MAC regions would provide consistent coverage for Aβ PET, this study found that sizable variation existed across regions in coverage determinations following the policy change. Possible reasons for variation include differences in proximity to facilities equipped to conduct Aβ PET, patient and practitioner preferences, and coverage decision-making across MACs. Access to Aβ PET has important implications for guiding treatment of monoclonal antibodies for AD both to detect Aβ presence to initiate treatment and to determine whether treatment may be stopped or modified following the clearance of Aβ to a centiloid level indicating removal of all detectable plaque.⁶ Uneven access to imaging has the potential to widen existing disparities to treatment and outcomes for AD.³ A study limitation is that presence of a diagnosis code for dementia is one factor, among others, used to make coverage determinations.

Supplement 1.

eMethods.

Supplement 2.

Data Sharing Statement