Clinician type and care setting for treatment of Medicare beneficiaries with dementia

Abstract

INTRODUCTION

Coordinating care for people living with dementia (PLWD) requires understanding which clinicians deliver care and the settings of that care.

METHODS

We used the Medicare Carrier file to characterize the settings in which clinicians deliver care to PLWD, clinician types providing care, and whether clinicians record a dementia diagnosis.

RESULTS

A total of 1,934,318 PLWD received care from 783,225 unique clinicians in 2019; PLWD saw a median of eight clinicians (interquartile range 5, 14). The most common settings were office (74.8% of PLWD), emergency room (63.9%), inpatient hospital (52.1%), and skilled nursing facility (37.1%). In addition, 87.0% of PLWD received care from a primary care physician, 62.9% from a nurse practitioner, and 33.1% from a physician assistant. Of the clinicians providing care, 2.4% are psychiatrists, 1.7% are neurologists, and 0.5% are geriatric subspecialists.

DISCUSSION

Care for PLWD must be coordinated across multiple clinicians and settings, recognizing that few PLWD receive psychiatry, neurology, or geriatric subspecialty care.

Highlights

• In 2019 the median Medicare beneficiary living with dementia saw eight different clinicians.

• Care of beneficiaries living with dementia is distributed across settings, with large percentages seen in each of the office, emergency room, inpatient hospital, and skilled nursing settings.

• Primary care physicians and nurse practitioners are the clinician types seen by the largest percentage of beneficiaries living with dementia.

• Geriatric subspecialist physicians account for less than 1% of the clinicians that provide care to beneficiaries living with dementia.

1. BACKGROUND

In 2020, over six million Americans had Alzheimer's disease and related dementias (ADRD), and the number will more than double to nearly 14 million by 2060. ¹ Given the complex medical needs of persons living with dementia (PLWD) and the change in functional status of the illness trajectory, PLWD may receive healthcare services from many different types of clinical providers and in a variety of healthcare and residential settings. Distributed across multiple providers in multiple settings, care provided to PLWD may be poorly coordinated, ² , ³ contributing to their comparatively higher rates of emergency department visits, ⁴ hospital admissions, ⁵ , ⁶ and higher treatment costs overall. ⁶ , ⁷ One additional factor that further complicates the care of someone with dementia is receiving a diagnosis of dementia ⁸ and whether, if made, that diagnosis is recognized by all the individual's treating clinicians.

While unpaid caregiving has been the focus of extensive research, much less is known about care delivery to PLWD by the paid clinical workforce, though this has now been emphasized in several recent federal initiatives. The National Institute on Aging's 2023 National Research Summit on Care, Services, and Supports for Persons Living with Dementia and Their Care Partners/Caregivers highlighted the importance of better understanding ties between the workforce and outcomes experienced by PLWD. ⁹ In the summer of 2024, the Centers for Medicare and Medicaid Services (CMS) launched the GUIDE model, which “focuses on comprehensive, coordinated care and aims to improve quality of life for people with dementia” in an effort to improve the healthcare delivered to PLWD. ¹⁰

Standardizing and improving the treatment that Medicare beneficiaries living with dementia receive requires understanding the settings where they are receiving care, the types of clinicians providing that care, and whether clinicians recognize the diagnosis of dementia in their encounters with PLWD. To investigate these questions, we used 100% Medicare fee‐for‐service claims from billing clinicians for all Medicare beneficiaries with dementia in 2019.

2. METHODS

2.1. Cohort identification

The study population included Medicare beneficiaries who had fee‐for‐service (Parts A and B but not Part C) coverage for at least 1 month in 2019. The cohort was limited to beneficiaries with an ADRD diagnosis in 2019 by applying a validated claims‐based algorithm to the Medicare Provider Analysis and Review (MedPAR), Carrier, and Outpatient files. ¹¹ To identify ADRD, beneficiaries were required to have at least one claim from the MedPAR file with an ADRD diagnosis or at least two claims from the Carrier or Outpatient files with an ADRD diagnosis at least 7 days apart.

For this analysis, the cohort of PLWD was further limited to beneficiaries with at least one non‐denied claim for an evaluation and management (E&M) service in the 2019 Carrier line file, which includes fee‐for‐service claims submitted by clinicians, including physicians, physician assistants, and nurse practitioners. The Carrier file contains all professional claims regardless of setting (i.e., it includes outpatient encounters as well as services provided to a patient during an inpatient stay such as consultations, surgeries, or other procedures). E&M services were identified by merging the Healthcare Common Procedure Coding System (HCPCS) code on each claim line with its associated Berenson‐Eggers Type of Service (BETOS) code based on the Restructured BETOS Classification System (RBCS). ¹²

2.2. Outcomes and statistical analyses

The primary goal of this descriptive cross‐sectional study was to identify the types of clinicians (i.e., specialty and licensure) delivering E&M services to PLWD and the settings in which that care was delivered.

First, we used the 2019 Carrier line file to identify encounters between beneficiaries with dementia and clinicians by identifying unique combinations of beneficiary, clinician (identified by National Provider Identifier [NPI]), claim, line first expense date, and place of service. We included all non‐denied Carrier line items for E&M services (denied line items were identified by an allowed charge amount of $0).

After identifying all unique encounters, we determined specific care setting using the encounter place of service variable in the Carrier line file (Appendix Table 1). We computed the overall number of beneficiaries with dementia and at least one E&M visit to a clinician performed in a setting of interest as well as how many unique clinicians (one, two, three, four, and five or more) each PLWD saw within the given setting group. We also determined the total number of beneficiary encounters in each setting of interest. The basic unit of analysis was the encounter, not a particular episode or admission (e.g., a patient hospitalized for several days with daily visits from Clinician A the first 3 days and daily visits from Clinician B the next 2 days would contribute a total of five encounters in the inpatient setting, three from Clinician A and two from Clinician B).

All clinicians with Carrier claims for E&M services are included in the analysis. To determine the specialty and/or licensure of the clinicians providing care, we (1) merged clinician NPI from the Carrier line file with taxonomy codes from the 2019 National Plan and Provider Enumeration System (NPPES) ¹³ and (2) used the Medicare Provider and Supplier Taxonomy Crosswalk to map the clinician NPPES taxonomy code to the appropriate clinician group for our analysis (Appendix Table 2), ¹⁴ including primary care physicians (including internal medicine and family medicine), psychiatry, neurology, geriatric physician subspecialists (i.e., geriatric medicine, geriatric psychiatry, and behavioral neurology and neuropsychiatry), and hospice and palliative care physicians. We included nurse practitioners overall and broken out by psychiatry and gerontology. Physician assistants were included as a single group because the NPPES taxonomy system does not provide subspecialty information.

Claims data were then collapsed to the beneficiary–clinician dyad level with an indicator flag for whether the clinician recorded an International Classification of Diseases, Tenth Revision diagnosis code for ADRD in any encounter with that beneficiary (i.e., a beneficiary may have been identified as having ADRD based on encounters with Clinician A, but that diagnosis may or may not be reflected in the beneficiary's encounter with Clinician B). For each specialty or licensure group of interest, we determined the number and percentage of PLWD seen, the number of PLWD‐serving clinicians, the number of unique clinician–PLWD dyads, and the percentage of these dyads in which the clinician indicated an ADRD diagnosis in an encounter. Finally, using setting groupings (i.e., outpatient, residential, inpatient, other) we determined the settings in which clinicians provided care to PLWD for a subset of the clinician license/specialty groups.

All data processing and analyses were conducted using SAS version 9.4 (SAS Institute, Inc., Cary, NC, USA). The study was approved by the Michigan Medicine Institutional Review Board, and informed consent was waived due to the observational study design and the use of deidentified secondary data.

3. RESULTS

The analysis included 1,934,318 Medicare beneficiaries enrolled in traditional Medicare who were living with dementia and had at least one E&M encounter in the Carrier file in 2019. These beneficiaries accounted for a total of 63,259,501 E&M encounters that were submitted by 783,225 unique clinicians.

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed PubMed and Google for high‐level analyses of the settings and types of clinicians that provide care to PLWD in traditional (i.e., fee‐for‐service) Medicare.

2. Interpretation: Using the 2019 Medicare Carrier file (i.e., where individual clinicians bill for patient care encounters) for all beneficiaries with a recorded dementia diagnosis, over 1.9 million beneficiaries living with dementia received care from 783,225 clinicians. In virtually every setting, the majority of care is delivered by primary care clinicians and nurse practitioners, while psychiatrists, neurologists, and geriatric subspecialists make up a small portion of clinicians delivering care.

3. Future directions: Improving care delivery to PLWD within the Medicare program must account for the multiple clinicians who deliver care to each PLWD and the multiple settings across which care is delivered.

Among unique PLWD, the most common encounter settings were office (Table 1; 74.8% of PLWD), emergency room (63.9%), inpatient hospital (52.1%), and skilled nursing facility (37.1%). Inpatient hospitals accounted for the largest percentage of encounters overall (35.7%), followed by office (20.8%), skilled nursing facility (14.1%), and nursing facility (11.4%). In the outpatient, residential, and inpatient settings, the plurality of PLWD saw five or more clinicians (Figure 1): 41.5%, 31.5%, and 53.1%, respectively. Among PLWD overall, 75.2% saw five or more Medicare clinicians in 2019, seeing a median of eight clinicians (interquartile range 5, 14).

TABLE 1.

Setting of clinician encounters for persons living with dementia in traditional Medicare.

	Unique patients with ≥1 visit by setting (N = 1,934,318)		Total patient encounters by setting (N = 63,259,501) a
	N	Percent b	N	Percent
Outpatient c
Office	1,446,271	74.8	13,149,419	20.8
Outpatient hospital	629,040	32.5	2,575,640	4.1
Urgent care facility	91,243	4.7	137,697	0.2
FQHC, independent clinic, public health clinic, or rural health clinic	4697	0.2	16,438	0.0
Other outpatient	1405	0.1	1989	0.0
Residential
Skilled nursing facility	717,485	37.1	8,938,255	14.1
Nursing facility	646,141	33.4	7,190,605	11.4
Assisted living	269,508	13.9	2,649,663	4.2
Custodial care facility	70,088	3.6	354,673	0.6
Other residential	13,937	0.7	61,778	0.1
Inpatient
Inpatient hospital	1,008,300	52.1	22,605,662	35.7
Comprehensive inpatient rehabilitation	46,735	2.4	567,637	0.9
Other settings
Emergency room – hospital	1,235,065	63.9	3,291,972	5.2
Home	181,480	9.4	1,039,202	1.6
Specialty psychiatric care	34,168	1.8	485,181	0.8
Telehealth provided other than in patient's home	31,697	1.6	158,563	0.3
Community mental health center	4536	0.2	22,474	0.0
Other	5314	0.3	12,652	0.0

Abbreviations: FQHC, Federally Qualified Health Center.

^a A given episode of care (e.g., inpatient admission) may contribute multiple encounters.

^b Column does not sum to 100% because a given patient could have encounters in multiple settings.

^c Appendix Table 1 outlines how distinct place of service codes are grouped into the settings used in analysis.

FIGURE 1.

Number of clinicians seen by persons living with dementia in Medicare, overall and by encounter setting. The figure presents the percentage of Medicare beneficiaries living with dementia who saw one, two, three, four, or five or more unique billing clinicians overall and within setting of interest in 2019. Beneficiaries who saw zero clinicians within a given setting are not presented (e.g., 20% [N = 387,585] of beneficiaries with dementia had no visits in the outpatient setting).

Eighty‐seven percent of PLWD had an encounter with a primary care physician (Table 2), while 62.9% saw a nurse practitioner, and 33.1% saw a physician assistant. Further, 16.3% and 27.3% of PLWD had an encounter with a general psychiatrist and neurologist, respectively. A total of 13.6% of PLWD saw a geriatric physician subspecialist: 12.0%, 1.7%, and 0.3% with a geriatrician, geriatric psychiatrist, and behavioral neurology or neuropsychiatry, respectively. In addition, 6.4% saw a gerontology nurse practitioner. Among clinician–PLWD dyads, specialists in both geriatric psychiatry and behavioral neurology or neuropsychiatry recorded an ADRD diagnosis in 69.1% of their encounters with a PLWD. Primary care physicians recorded an ADRD diagnosis in 35.4% of dyads, nurse practitioners in 35.3%.

TABLE 2.

Clinician types providing care for PLWD.

Clinician type	Unique PLWD seen by clinician, n	Percentage of PLWD who have seen this clinician type (n = 1,934,318)	Unique clinicians who have seen PLWD, n	Distinct PLWD–clinician dyads, n	Percentage of unique dyads where clinician recorded ADRD diagnosis
Primary care physician	1,683,668	87.0	159,111	4,853,843	35.4
Internal medicine	1,289,748	66.7	81,443	3,211,650	32.8
Family medicine	1,013,011	52.4	77,668	1,642,193	40.5
Psychiatry	315,295	16.3	18,992	460,023	53.7
Neurology	527,700	27.3	13,496	707,997	45.5
Geriatric subspecialist physician	263,548	13.6	3873	308,506	59.2
Geriatric medicine	232,131	12.0	3447	267,977	57.7
Geriatric psychiatry	32,939	1.7	360	34,443	69.1
Behavioral neurology & neuropsychiatry	6054	0.3	66	6,086	69.1
Hospice and palliative care	51,516	2.7	1571	55,847	36.0
Nurse practitioner (NP)	1,217,057	62.9	119,539	2,724,547	37.8
Psychiatry NP	141,420	7.3	5830	168,991	62.7
Gerontology NP	124,379	6.4	2,610	134,646	51.4
Physician assistant	640,589	33.1	70,580	1,017,589	19.5
Other physician
Emergency medicine	1,131,408	58.5	43,533	2,475,709	9.9
Internal medicine, cardiovascular	541,043	28.0	20,295	857,739	8.1
Internal medicine/family medicine, other subspecialty	478,106	24.7	36,924	748,617	8.6
Hospitalist	432,993	22.4	14,522	852,130	24.8
Ophthalmology	412,318	21.3	16,461	498,558	0.6
Surgery (non‐orthopedic)	315,272	16.3	31,016	469,582	7.7
Internal medicine, pulmonary disease	269,413	13.9	9932	444,819	9.1
Orthopedic surgery	259,690	13.4	21,306	310,317	2.8
Internal medicine, nephrology	226,726	11.7	8836	413,616	7.1
Urology	223,980	11.6	8425	280,018	2.5
Internal medicine, gastroenterology	222,138	11.5	12,689	308,503	5.5
Physical medicine and rehabilitation	203,235	10.5	7386	286,613	15.2
Internal medicine, infectious disease	188,594	9.7	5992	298,622	9.4
Dermatology	188,029	9.7	10,862	206,054	0.2
Other clinician
Podiatry	560,594	29.0	14,899	663,029	1.9
Optometry	453,887	23.5	27,752	491,941	0.3
Clinical psychology	173,000	8.9	10,869	199,129	38.1
Clinical neuropsychology	43,435	2.2	1724	44,321	57.1
Clinical nurse specialist	48,756	2.5	2013	52,582	42.7

Abbreviations: ADRD, Alzheimer's disease and related dementias; PLWD, persons living with dementia.

Among other clinician specialties, the largest percentage of PLWD were seen by emergency medicine (58.5%), podiatry (29.0%), cardiology (28.0%), optometry (23.5%), and hospitalist physicians (22.4%). Among these other specialties or disciplines, neuropsychology, clinical nurse specialists, clinical psychology, and hospitalists were the most likely to record an ADRD diagnosis at 57.1%, 42.7%, 38.1%, and 24.8%, respectively.

Among the 783,225 unique clinicians that billed traditional Medicare for encounters with PLWD in 2019, 20.3% were primary care physicians, 15.3% were nurse practitioners, 9.0% were physician assistants, 2.4% were psychiatrists, and 1.7% were neurologists; just 0.5% were geriatric subspecialist physicians (Table 3). Within each clinician type, the largest percentage of clinicians provided care in the office setting, with the exception of neurologists, where a slightly higher percentage provided inpatient hospital care (67.8% vs 66.2%). A larger percentage of geriatric subspecialty physicians provided their care in residential settings, with 41.5% providing care in a skilled nursing facility, 27.6% in a nursing facility, and 21.1% in assisted living.

TABLE 3.

Encounter setting of PLWD‐serving clinicians by clinician type.

	Percentage of PLWD‐serving clinicians within a given clinician type by encounter setting a
	Overall	Primary care physician	Psychiatry	Neurology	Geriatric subspecialty physician	Nurse practitioner	Physician Assistant
Unique clinicians, n	783,225 b	159,111	18,992	13,496	3,873	119,539	70,580
Percentage of PLWD‐serving clinicians	100 b	20.3	2.4	1.7	0.5	15.3	9.0
Outpatient
Office	60.2	62.9	52.1	66.2	56.3	50.9	48.1
Outpatient hospital	27.8	34.5	19.6	51.6	35.2	22.8	21.2
Urgent care facility	2.5	3.2	s c	s	0.7	4.4	8.0
FQHC, independent clinic, public health clinic, or rural health clinic	0.1	0.2	0.4	0.1	0.4	0.2	0.1
Other outpatient	0.1	0.0	s	s	s	0.4	0.2
Residential
Skilled nursing facility	6.3	11.7	4.6	2.0	41.5	10.6	2.9
Nursing facility	4.4	7.1	3.6	0.7	27.6	7.8	1.8
Assisted living	2.7	3.2	2.3	0.3	21.1	6.2	1.3
Custodial care facility	0.6	1.0	0.8	0.1	5.5	1.1	0.2
Other residential	0.3	0.5	0.6	0.2	1.8	0.5	0.1
Inpatient
Inpatient hospital	38.6	38.5	41.2	67.8	41.1	25.1	26.6
Comprehensive inpatient rehabilitation	0.9	1.2	0.5	1.7	1.6	0.6	0.4
Other settings
Emergency room – hospital	12.3	5.8	9.3	13.7	4.6	7.4	20.6
Home	3.1	5.4	0.9	0.6	19.4	5.3	1.0
Specialty psychiatric care	0.9	0.8	16.1	1.1	2.7	1.2	0.4
Telehealth provided other than in patient's home	0.5	0.2	5.3	4.3	0.8	0.7	0.2
Community mental health center	0.3	0.0	3.9	s	s	0.5	0.1
Other	0.2	0.1	0.3	s	s	0.2	0.0

Abbreviations: CMS, Centers for Medicare and Medicaid Services; FQHC, Federally Qualified Health Center; PLWD, persons living with dementia.

^a Because a clinician may work in multiple settings, the sum across settings within each clinician type (i.e., column totals) will exceed 100%.

^b The overall values for “Unique clinicians” and “Percentage of PLWD‐serving clinicians” reflect all clinicians that provide care to PLWD, including licensures or specialties not separately reported here (e.g., cardiology). Because the other columns (e.g., primary care physician) reflect a subset of all clinician types, the sums across the top two rows (e.g., Percentage of PLWD‐serving clinicians) are less than the overall totals presented.

^c Values for these cells are suppressed because CMS does not allow reporting of results based on fewer than 11 beneficiaries.

Within the majority of settings, the most common clinician type was primary care physicians who, for example, accounted for 21.2% of clinicians who provided office‐based outpatient care to PLWD and 37.9% of clinicians who provided care to PLWD in skilled nursing facilities (Table 4). However, nurse practitioners accounted for the most clinicians in assisted living (35.2%). Geriatric subspecialist physicians make up just 0.5% of office‐based PLWD‐serving clinicians and 3.3% of those in skilled nursing facilities.

TABLE 4.

Clinician type of PLWD‐serving clinicians by encounter setting.

	Percentage of PLWD‐serving clinicians within given encounter setting by clinician type
	Overall	Primary care physician	Psychiatry	Neurology	Geriatric subspecialty physician	Nurse practitioner	Physician assistant
Unique clinicians, n	783,225 a	159,111	18,992	13,496	3,873	119,539	70,580
Percentage of PLWD‐serving clinicians	100*	20.3	2.4	1.7	0.5	15.3	9.0
Outpatient
Office	471,456	21.2	2.1	1.9	0.5	12.9	7.2
Outpatient hospital	217,585	25.2	1.7	3.2	0.6	12.5	6.9
Urgent care facility	19,495	26.3	s b	s	0.1	27.1	29.0
FQHC, independent clinic, public health clinic or rural health clinic	1164	27.9	7.3	1.0	1.2	20.3	6.1
Other outpatient	693	3.2	s	s	s	72.0	16.2
Residential
Skilled nursing facility	48,970	37.9	1.8	0.5	3.3	25.9	4.2
Nursing facility	34,109	32.9	2.0	0.3	3.1	27.3	3.7
Assisted living	21,102	23.9	2.0	0.2	3.9	35.2	4.3
Custodial care facility	4868	33.1	3.0	0.2	4.4	26.0	2.8
Other residential	2370	31.8	5.1	1.1	3.0	26.4	3.6
Inpatient
Inpatient hospital	302,147	20.3	2.6	3.0	0.5	9.9	6.2
Comprehensive inpatient rehabilitation	6707	28.8	1.3	3.4	0.9	10.7	3.8
Other settings
Emergency room – hospital	96,652	9.6	1.8	1.9	0.2	9.1	15.0
Home	24,183	35.8	0.7	0.3	3.1	26.4	3.1
Specialty psychiatric care	7372	16.2	41.5	2.0	1.4	19.6	4.1
Telehealth provided other than in patient's home	4168	8.1	23.9	14.0	0.7	21.3	3.0
Community mental health center	2340	0.7	31.4	s	s	27.7	1.7
Other	1501	9.1	3.9	s	s	15.1	2.1

Abbreviations: CMS, Centers for Medicare and Medicaid Services; FQHC, Federally Qualified Health Center; PLWD, persons living with dementia.

^a The overall values for “Unique clinicians” and “Percentage of PLWD‐serving clinicians” reflect all clinicians that provide care to PLWD, including licensures or specialties not separately reported here (e.g., cardiology). Because the other columns (e.g., primary care physician) reflect a subset of all clinician types, the sums across the top two rows (e.g., Percentage of PLWD‐serving clinicians) are less than the overall totals presented.

^b Values for these cells are suppressed because CMS does not allow reporting of results based on fewer than 11 beneficiaries.

4. DISCUSSION

In this study of Medicare fee‐for‐service beneficiaries who live with dementia, we found that the care PLWD receive is widely distributed across clinician types and care settings, with the vast majority delivered by clinicians who are not geriatric subspecialists or neurologists. These data illustrate that policies and practice strategies to improve the care of PLWD need to encompass multiple clinicians and settings and must focus on non‐specialists, with whom PLWD have the vast majority of their healthcare encounters.

It is important to note the critical role played by primary care physicians, nurse practitioners, and physician assistants, who account for 20.3%, 15.3%, and 9.0%, respectively, of the clinicians who cared for PLWD in 2019. In contrast, psychiatrists and neurologists made up just 2.4% and 1.7% of PLWD‐serving clinicians, while geriatric subspecialist physicians accounted for just 0.5% of this workforce. The importance of advanced practice providers (APPs) – particularly nurse practitioners and physician assistants – is consistent with prior work examining APP care for Medicare beneficiaries overall, with dementia among the specific conditions for which APPs provided a larger proportion of care. ¹⁵ Our estimates likely underestimate the role of APPs because of the practice of “incident to” billing, where services delivered by an APP are billed by a supervising physician and thus paid at a higher reimbursement level. ¹⁶ However, our estimate that APPs accounted for 24.3% of clinicians that cared for PLWD is only slightly lower than the Patel et al. finding that 27.1% of visits for beneficiaries with dementia were provided by APPs. ¹⁵

More than 30% of PLWD encounters are in residential settings such as skilled nursing or assisted living. These residential settings face workforce challenges because both the supply and training of the clinician workforce offering services in these settings may be inadequate. ¹⁷ , ¹⁸ Given a residential patient population with multiple complex conditions, including ADRD, there have been calls for education of “nursing home specialists,” with recent research documenting a positive relationship between nursing home adoption of “SNFists” and higher‐quality end‐of‐life care. ¹⁹ Our findings demonstrate that residential care is an important practice setting for PLWD‐serving geriatric subspecialty physicians, 41.5% of whom practiced in skilled nursing facilities and 21.1% in assisted living. However, 37.9% of clinicians that care for PLWD in skilled nursing facilities are primary care clinicians and 25.9% are nurse practitioners. Therefore, initiatives to meaningfully improve dementia care must focus on supporting primary care generalists who are the de facto dementia care workforce.

Relatively few clinician encounters for PLWD overall (1.6%) occurred in patients’ homes, though 9.4% of PLWD did have a home visit. Visits to assisted living communities – which, for resident beneficiaries, is their home – were more common, accounting for 4.2% of encounters and 13.9% of PLWD. A national consensus panel of leading experts in dementia care recommended the expansion of home‐based dementia care to improve the quality and patient‐centeredness of care. ²⁰ Among clinicians serving PLWD, geriatric subspecialist physicians were most likely to have any E&M encounters in the home setting (19.4%) and in assisted living (21.1%). However, because there are relatively few of these specialist clinicians, these subspecialists still accounted for relatively small proportions of the visits within those settings (3.1% and 3.9%, respectively). Although PLWD are significantly more likely to receive home‐based clinical care than those without dementia, ²¹ many home‐based primary care practices report that they do not have adequate knowledge to meet the needs of PLWD. ²² , ²³ As with other settings of care, programs that seek to expand capacity for home‐based services to PLWD – including to assisted living – will require supporting primary care clinicians.

While this analysis does not allow us to comment on the quality of care delivered by these clinicians, it is worth noting that, among PLWD observed in Medicare claims, primary care physicians, nurse practitioners, and physician assistants documented an ADRD diagnosis among just 35.4%, 37.8%, and 19.5% of dyads with PLWD, respectively. For primary care providers, the care they provide to PLWD may not focus on their dementia diagnosis; however, given the cognitive and functional impacts of dementia, its presence would potentially impact the management of virtually any healthcare condition. If the lack of the dementia diagnosis in a clinical encounter suggests it is not being considered in the course of clinical care, this is a significant concern. Unfortunately, because nearly one‐quarter of probable dementia is not recognized in claims data while another 18% of individuals have a delay in recognition of nearly 3 years between the onset of cognitive impairment and an encounter diagnosis appearing, ²⁴ the interactions of many PLWD and their clinicians are not captured at all in this analysis.

In contrast, dyads of PLWD with geriatricians, geriatric psychiatrists, and gerontology nurse practitioners recorded a diagnosis in 57.7%, 69.1%, and 51.4%, respectively. These higher rates of a recorded dementia diagnosis may suggest that dementia is a focus of the clinical care delivered by these geriatric subspecialists, and they are more attuned to recording it as a diagnosis. Yet it is also surprising that these subspecialists did not diagnose dementia in a notable proportion of these dyads. In a supplemental analysis, we examined the top non‐ADRD diagnoses among the specialties of geriatric psychiatry and behavioral neurology and neuropsychiatry, finding diagnoses including “mild cognitive impairment of uncertain or unknown etiology” (G318.4), “unspecified symptoms and signs involving cognitive functions and awareness” (R41.9), and “other amnesia” (R41.3). These findings suggest that these clinicians were in fact seeing individuals with impaired cognition, although the specific encounter diagnoses they were using are not part of the validated Bynum standard algorithm. ¹¹ , ²⁵

Our analysis has several limitations. First, while we used a validated approach to identify dementia in Medicare claims, the condition is still underdiagnosed in administrative data. ²⁵ The analysis reflects billed services, which, given the use of “incident to” billing, may not actually reflect who provided the care to the beneficiary. ¹⁶ The taxonomy codes available in NPPES likely reflect specialty at the time of NPI registration and may be outdated because there is no requirement that clinicians review or update their taxonomy. Finally, our analysis only includes beneficiaries in traditional Medicare, which may not reflect the experience of those in Medicare Advantage.

This comprehensive analysis of clinician services provided to traditional Medicare beneficiaries with dementia in 2019 reveals the large and diffuse clinical workforce that provides care to these individuals. In addition, the findings underscore the limited number of psychiatry, neurology, and geriatric subspecialty clinicians relative to primary care physicians, nurse practitioners, and physician assistants. Improving the healthcare delivered to PLWD will only be achieved through services that support non‐specialty clinicians across the multiple settings where PLWD receive care.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the Supporting Information.

CONSENT STATEMENT

For this analysis of observational data from the Centers for Medicare & Medicaid Services, the Michigan Medicine institutional review board does not require individual informed consent.

Supporting information

Supporting Information

Supporting Information

Maust DT, Davis RC, Muench U, Marcus SC, Spetz J. Clinician type and care setting for treatment of Medicare beneficiaries with dementia. Alzheimer's Dement. 2025;21:e70102. 10.1002/alz.70102