Real‐world use of diagnostic tests for mild cognitive impairment, Alzheimer's disease, and other dementias in Medicare fee‐for‐service beneficiaries

Abstract

INTRODUCTION

This study assessed real‐world use of diagnostic tests, such as neuroimaging (e.g., magnetic resonance imaging [MRI], or positron emission tomography [PET]), and computed tomography (CT), cerebrospinal fluid (CSF) biomarker, and blood tests for mild cognitive impairment (MCI), Alzheimer's disease (AD), and other dementias in a large US elderly population.

METHODS

Medicare fee‐for‐service data (2015–2020) were used to identify patients aged ≥ 67 newly diagnosed with MCI, AD, or other dementias. Descriptive analyses were conducted to understand the test use within 1 year before disease diagnosis and trends.

RESULTS

Among 653,420 patients (9.1% MCI, 30.3% AD, 60.6% other dementias), 71.9% had blood tests, 53.9% neuroimaging (46.4% CT, 17.7% MRI, and 0.7% PET), and 2.2% CSF test. Test use slightly increased from 2015 to 2020.

DISCUSSION

Findings from this study suggest low use of diagnostic tests, especially PET and CSF.

Highlights

• Blood tests, magnetic resonance imaging, and computed tomography were predominant for diagnosing mild cognitive impairment, Alzheimer's disease, or other dementias prior to the arrival of disease‐modifying therapies.

• Cerebrospinal fluid biomarker and positron emission tomography tests were infrequently used despite their diagnostic value.

• The study indicates a modest increase in diagnostic test usage over 6 years between 2015 and 2020.

• Patients often received combined or repeated diagnostic tests.

1. INTRODUCTION

Dementia is best characterized as a syndrome marked by a decline in memory and cognitive skills severe enough to hinder daily activities. ¹ , ² , ³ Alzheimer's disease (AD) is the most common cause of dementia, comprising 60% to 80% of all cases. ⁴ AD, a progressive neurodegenerative disease, is defined by two pathological hallmarks: the accumulation of extracellular amyloid beta (Aβ) plaques and intracellular neurofibrillary tangles (NFTs). ⁵ Aβ plaque deposition begins ≈ 20 years before cognitive impairment, and NFTs emerge 10 to 15 years before symptoms. ⁶ , ⁷

Diagnostically, AD follows a continuum from preclinical AD—clinically asymptomatic individuals with AD pathology—to mild cognitive impairment (MCI) and then to AD dementia. ⁸ , ⁹ , ¹⁰ MCI is a state of cognitive decline with retained functional independence. ¹¹ Approximately 15% of individuals with MCI progress to dementia within 2 years, ¹² and 32% develop AD within 5 years. ¹³ With the aging population in the United States, the prevalence of MCI, AD, and related dementias is expected to rise dramatically. By 2060, nearly 14 million Americans aged ≥ 65 will have AD, up from 6.9 million in 2024. ⁴

The advent of disease‐modifying therapies (DMTs) emphasizes the importance of early and accurate AD diagnosis, as DMTs are more effective when administered in the early stages of the disease. Confirmation of AD is also critical for determining treatment eligibility. ¹⁴ Traditionally, AD diagnosis is made by excluding other causes of cognitive impairment through a medical history, physical exam, and neurocognitive testing in primary care settings. ¹⁵ Additional examinations (biochemistry and vitamins, and hormone levels) may also be performed to rule out systemic processes. ¹⁶ When abnormalities are found, the next step is a referral to the neurologist/memory specialist as well as to obtain tests, including laboratory tests (blood and other biologic fluids), genetic testing, and brain neuroimaging such as magnetic resonance imaging (MRI) and computed tomography (CT) scans. ¹⁴ , ¹⁷ MRI and CT scans are used to detect brain structural and functional changes. ¹⁶ However, they do not specifically confirm an AD diagnosis. ¹⁴ , ¹⁸ Additionally, the current diagnosis of AD or dementia is largely based on clinical signs and symptoms, ¹⁷ with a significant number of patients being diagnosed when their disease has already advanced. In recent years, great progress has been made in the diagnosis of AD using biomarkers. Biomarker tests, such as via amyloid positron emission tomography (PET) imaging and/or lumbar puncture with cerebrospinal fluid (CSF) analysis, can increase diagnostic accuracy and help with early detection of AD. ¹⁴

The use of diagnostic tests for MCI, AD, or dementia in real‐world settings is not well understood. Tsoy et al. studied diagnostic evaluations, including lab tests, for MCI or dementia in California but excluded neuroimaging and used outdated data (2013–2015). ¹⁹ Roth et al. recently examined AD diagnostic tests but had a small sample with limited representability. ²⁰ The Imaging Dementia–Evidence for Amyloid Scanning (IDEAS) studies ²¹ , ²² are US‐wide but focus solely on amyloid PET. Thus, a gap persists in understanding the broader use of diagnostic tools for dementia on a national scale.

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed the literature using traditional (e.g., PubMed) sources and meeting abstracts and presentations. While a gap persists in understanding the broader use of diagnostic tools for dementia on a national scale, there have been several recent publications describing the use of diagnostic tools for Alzheimer's disease or related dementias. These relevant references are appropriately cited.

2. Interpretation: Our study results indicate that, in the 6 years prior to the introduction of disease‐modifying therapies (DMTs), blood tests, magnetic resonance imaging and computed tomography were the most commonly used diagnostic tests, showing modest increasing trends. In contrast, cerebrospinal fluid (CSF) and positron emission tomography (PET) were rarely used. Furthermore, patients frequently underwent multiple or repeated diagnostic tests.

3. Future directions: Although there have been modest increasing trends, significant improvements are still needed in the use of confirmatory tests, particularly PET and CSF, which will be essential for the expanded use of new DMTs.

Given the lack of contemporary data on the comprehensive use of diagnostic tests for MCI, AD, or dementia in the United States, in the current study, we aimed to gain a comprehensive understanding on the real‐world use of diagnostic tests prior to the diagnosis of MCI, AD, or other dementias in a large, nationally representative sample using the latest Medicare data.

2. METHODS

2.1. Study design and data source

For this retrospective, observational study, we used the Centers for Medicare & Medicaid Services (CMS) Research Identifiable Files (RIFs) for 100% of Medicare beneficiaries enrolled in the fee‐for‐service (FFS) program from the years 2013 to 2020. At the time of our study, this was the most current database available to us. Medicare is a federal health insurance program that provided coverage for ≈ 56 million Americans aged ≥ 65 years in 2020. ²³ The Medicare 100% RIFs cover Medicare beneficiaries from all census regions and contain beneficiary‐level information on demographics, enrollment, and administrative claims for health care received by Medicare beneficiaries, including Part A (e.g., institutional claims), Part B (e.g., non‐institutional claims), and Part D (e.g., drug claims). For this study, we used both patient‐level and claims‐level data from the master beneficiary summary, inpatient, outpatient, carrier, skilled nursing facility (SNF), home health agency, durable medical equipment, and Part D files.

We obtained the study database with a data use agreement from the Research Data Assistance Center, a CMS contractor that assists researchers interested in CMS data. All study data were fully compliant with the requirements of the Privacy Act, the Health Insurance Portability and Accountability Act Privacy Rule, and CMS data release policies. The study used only de‐identified patient records and, therefore, was exempted from institutional review board approval. Informed consent from patients was not required because the use or disclosure of the requested information did not adversely affect the rights and welfare of the beneficiaries and involved no more than a minimal risk to their privacy.

2.2. Study sample selection

The study population included Medicare FFS beneficiaries ≥ 67 years old and newly diagnosed with MCI, AD, or other dementias during the study identification (ID) period between January 1, 2015 and December 31, 2020. Beneficiaries were deemed to have an MCI, AD, or other dementia diagnosis if they had at least one claim for MCI, AD, or other dementias in any of the inpatient, outpatient, and carrier files, as one of the primary or secondary diagnosis codes during the ID period. We used the previously validated Chronic Condition Data Warehouse International Classification of Disease, Clinical Modification diagnosis codes (9th or 10th editions) (ICD‐9/10 CM) to identify MCI, AD, or dementia. ²⁴

Additionally, to avoid counting “rule‐out” diagnosis or diagnosis due to coding errors, we required having a second diagnosis code within 2 years (i.e., a total of two diagnosis codes within 2 years) or death within 1 year after diagnosis (i.e., one diagnosis code + death within 1 year) to verify the diagnosis of AD, MCI, or other dementias. ²⁵ The date of the first of the two codes for the same disease of interest was considered the index date. The following three mutually exclusive patient groups were identified based on the index disease of interest and a hierarchical classification order of AD > other dementia > MCI during the study ID period: (1) beneficiaries newly diagnosed with MCI (incident MCI patient group), (2) beneficiaries newly diagnosed with AD (incident AD patient group), and (3) beneficiaries newly diagnosed with other dementias (incident other dementia patient group).

The above beneficiaries were required to be aged ≥ 67 years on the index date to ensure ≥ 2 years of pre‐index eligibility; have continuous enrollment in FFS Medicare, and be eligible for Medicare Parts A and B for 2 years prior to the index date. To ensure that patients were newly diagnosed with MCI, AD, or other dementias, patients were excluded if they had any claims in inpatient, outpatient, carrier, home health, or SNF claim files for each index disease of interest or any prescription fill for an AD medication (i.e., donepezil, rivastigmine, galantamine, and memantine) during the 2 year period before the index date (2 year washout period). Patients were further excluded if they had any evidence of Part C enrollment within 2 years prior to the index date to ensure complete information was retrieved. A total of 653,420 Medicare FFS beneficiaries were included in this analysis after applying these study eligibility criteria (Figure 1).

FIGURE 1.

Patient attrition diagram. The date of the first diagnosis code for a disease of interest was deemed to be the index date. AD medications included donepezil, rivastigmine, galantamine, and memantine. AD, Alzheimer's disease; FFS, fee for service; HH, home health; MCI, mild cognitive impairment; SNF, skilled nursing facility.

2.3. Study measures

The study measures were constructed using data in the 1 year period before and on the index date. Patients’ sociodemographic characteristics included age groups on the index date (i.e., 67–74, 75–84, and 85+ years old), sex (female vs. male), race/ethnicity (non‐Hispanic White, Black, Hispanic, Asian, and American Indian/Alaska native), Medicaid eligibility to determine Medicare–Medicaid dual status (yes/no), county of residence (metropolitan, micropolitan, and rural residence), usual physician types, and year of diagnosis. Race/ethnicity was identified using the Research Triangle Institute imputation algorithm, which was included by CMS in the Medicare administrative database. The imputation algorithm improved sensitivity in identifying Hispanic beneficiaries from 29.5% to 76.6% over the existing Medicare Program's enrollment database race variables, and for Asian/Pacific Islanders from 54.7% to 79.2%. ²⁶ A beneficiary's county of residence was first determined using the MBS files. The rural–urban continuum codes (RUCC) were then used to determine the rurality of a county's beneficiary: metropolitan (RUCC 1 to 3), micropolitan (RUCC 4 to 7), and rural (RUCC 8 and 9). ²⁷ The usual physician specialty was defined as the physician specialty with the largest number of outpatient or office visits during the 1 year period before and on the index date (baseline) period.

Patients clinical characteristics include adapted Charlson Comorbidity Index (CCI) by excluding dementia, ²⁸ presence of the following comorbidities relevant to MCI, AD, or other dementias: anxiety, bipolar disorder, depressive disorders, epilepsy and recurrent seizures, schizophrenia and other psychotic disorders, sleep disorders, stroke/transient ischemic attack, transient mental disorders due to conditions classified elsewhere, other conditions of brain, other degenerative diseases of nervous system, not elsewhere classified, concussion, tobacco use disorder, periodontitis, obesity, diabetes, hypertension, heart failure, cardiac dysrhythmias, acute myocardial infarction, and ischemic heart disease. ²⁹ , ³⁰ , ³¹ , ³² , ³³ The presence of these clinical characteristics was determined based upon having at least one claim with a relevant ICD‐9/10 CM diagnosis code, and Current Procedural Terminology (CPT)/Healthcare Common Procedure Coding System (HCPCS) Codes.

The diagnostic tests, or diagnostic workup of interest, included neuroimaging (i.e., MRI, PET, or CT scan) for the head or brain, CSF biomarker test, and recommended blood tests, including both genetic tests (e.g., apolipoprotein E genotyping) and rule‐out tests (e.g., vitamin B12 and thyrotropin) for patients with MCI, AD, or other dementias. We identified the list of these diagnostic tests based on existing literature ¹² , ¹⁹ , ³⁴ , ³⁵ and clinical expert inputs. The presence of these clinical diagnostic tests was determined based on having at least one claim with a relevant CPT/HCPCS code (see Table S1 in supporting information).

2.4. Statistical analysis

Descriptive statistics were reported for the overall study population and the three patient groups. Specifically, continuous variables were summarized by the number of available observations, means, standard deviations (SD), quartiles, minimum, and maximum. Categorical variables were summarized by frequencies and percentages. Chi‐squared tests were used to evaluate differences in proportions of categorical variables across the study patient groups.

To understand the test use at a granular level, we reported the test rates for each test (e.g., MRI, CT, PET, and CSF) or test categories (e.g., any imaging test, any blood test) alone and in combinations, with and without a chronological order, within 1 year before and on the index date. To understand the test use over time, we reported the test rates by each index year. Additionally, among patients who received a test, we reported the mean (SD) number of tests used per patient. A sensitivity analysis using the 3 months prior to the index date was also performed. All data transformations and statistical analyses were performed using SAS Enterprise Guide Version 7.15 (SAS Institute Inc.).

3. RESULTS

3.1. Patient demographic and clinical characteristics

Of the 653,420 Medicare FFS beneficiaries included in the study, 395,936 (60.6%) were incident other dementia patients, 198,045 (30.3%) were incident AD patients, and 59,439 (9.1%) were incident MCI patients. The majority of the study population were female (61.9%), White (88.3%), Medicare‐only (i.e., non‐dual, 76.1%) beneficiaries, and lived in a metropolitan area (72.6%; Figure 2).

FIGURE 2.

Patient sociodemographic characteristics and physician specialty. PCP, primary care physician/geriatrician.

The study population had a high comorbid disease burden (Table S2 in supporting information). The majority of the study population had comorbid illnesses such as hypertension (82.2.1%) or hyperlipidemia (64.7%). There were statistically significant differences in comorbid illness across the three study groups. The mean CCI score was observed to be higher in patients in the other dementia group than those in the AD or MCI patient groups (P < 0.001). A higher percentage of patients in the other‐dementia patient group had a comorbid illness such as cardiac dysrhythmias, diabetes, ischemic heart disease, or heart failure (P < 0.001).

3.2. Use of diagnostic test within 1 year before disease diagnosis

Overall, during the 1 year before the index date or disease diagnosis, 71.9% of the study population received at least one blood test, 53.9% had any neuroimaging test (i.e., MRI, CT, or PET), 46.4% had CT, 17.7% had MRI, 0.7% had PET, and 2.2% had CSF. The use of diagnostic tests differed across the three patient groups (P < 0.001). Patients in the other dementia group had the highest use of any neuroimaging test (56.2% vs. 51.4% in AD vs. 46.9% in MCI). Patients in the MCI group had the highest use of CSF (3.5% vs. 2.2% in other dementia vs. 1.8% in the AD group; Figure 3).

FIGURE 3.

Use of diagnostic tests within 1 year before or on the disease diagnosis date. AD, Alzheimer's disease; CSF, cerebrospinal fluid; CT, computed tomography; MCI, mild cognitive impairment; MRI, magnetic resonance imaging; PET, positron emission tomography.

Some patients received multiple tests before their MCI, AD, or other dementia diagnosis. Specifically, for the overall study population, when examining their test combinations in a non‐chronological order, 39.9% received blood + imaging tests, 1.5% had blood + CSF + imaging tests, 0.41% had blood + CSF tests, and 0.23% had CSF + imaging tests. When examined in a chronological order, the most common combination of tests is a blood test followed by an imaging test (15.3%; Figure 4A,B). Excluding blood tests, CT + MRI is the most common combination without chronological order (9.5%). CT followed by MRI is the most common combination in chronological order (3.2%; Figure 5A,B). Among patients who received any test, on average, patients received 2.1 (1.5) blood tests, 1.8 (1.4) any imaging tests (1.7 [1.3] for CT, 1.2 [0.6] for MRI, 1 [0.1] for PET), and 1.3 (1.1) CSF tests (Table S2). Similar results were found for each patient group (Table S3 in supporting information).

FIGURE 4.

Use of diagnostic test combinations within 1 year before or on disease diagnosis date. CSF, cerebrospinal fluid.

FIGURE 5.

Top seven diagnostic test combinations (excluding blood tests) within 1 year before or on disease diagnosis date. CSF, cerebrospinal fluid; CT, computed tomography; MRI, magnetic resonance imaging; PET, positron emission tomography.

Overall, the use of diagnostic tests increased slightly from 2015 to 2020 (Figure 6). Similar trends were found for each patient group, except for CT use in the MCI patient group (Figures S1–S4 in supporting information).

FIGURE 6.

Trends in real‐world use of diagnostic tests within 1 year before or on the disease diagnosis date. CSF, cerebrospinal fluid; CT, computed tomography; MRI, magnetic resonance imaging; PET, positron emission tomography.

3.3. Use of diagnostic test within 3 months before disease diagnosis

The diagnostic test use patterns and trends were similar during the 3 months prior to an MCI, AD, or other dementia diagnosis (Tables S4–S6 in supporting information). However, the test use rates and the average number of tests received during this time period were lower compared to those found during the 1 year prior to disease diagnosis. For example, the percentages of the study population receiving at least one blood test or an imaging test were reduced to 44.2% and 38.8%, respectively. Less than 1% of the Medicare FFS beneficiaries received a CSF or PET test during the 3 months before disease diagnosis (Table S6).

4. DISCUSSION

Early diagnosis of AD and other dementias is a core public health strategy. ³⁶ To diagnose AD and other dementias early, diagnostic tests play a crucial role as they can help with identifying cognitive impairment, differentiating AD from other types of dementia, tracking the course of the disease, and determining patients’ eligibility for treatments such as DMTs. ¹⁴ , ³⁷ Our study fills the existing data gap by quantifying the use of diagnostic tests, both traditional ones such as MRI and CT, and novel ones such as CSF, for MCI, AD, or other dementia types in a large, nationally representative sample of older adults in the United States. Our study results suggest that, during the 6 years preceding the arrival of DMTs, despite the modest increasing trends, blood tests, MRI, and CT were the predominant diagnostic tests, while the use of CSF and PET was very infrequent. Additionally, patients often received combined or repeated diagnostic tests.

The overall low use of CSF biomarker and PET tests, including the Aβ PET observed in our study, is consistent with what has been reported in prior studies. ²⁰ , ³⁸ Even though our study did not explore which factors may have influenced these low rates, the invasiveness of lumbar puncture, a necessary procedure for getting CSF samples, and its link to more adverse events than imaging or blood‐based testing may be the main barrier to using CSF biomarker tests. ²⁰ , ³⁹ , ⁴⁰ Access and availability constraints may be other significant barriers to CSF testing, as most CSF analyses are carried out in specialized laboratories ³⁷ and there is a need for skilled clinicians to perform lumbar puncture. ⁴¹ Additionally, the low reimbursement rate for lumbar puncture may limit the use of CSF testing. ⁴¹ The national average Medicare reimbursement for an outpatient diagnostic lumbar puncture in 2023 was ≈ $135. ⁴¹ Compared to the use of CSF biomarker tests, the use of amyloid PET scans is even lower. The “out‐of‐pocket expense to patients” or lack of reimbursement was likely to be the main reason for their low use. Without reimbursement, patients would typically need to pay ≈ $5000 to $7000 for amyloid PET scans, compared to < $1000 for CSF biomarkers or as little as $437 for a non‐contrast brain MRI. ⁴² , ⁴³ Prior to October 2023, CMS had limited reimbursements of Aβ PET scans to patients enrolled in coverage with evidence development (CED) programs. Specifically, patients had to be part of an authorized clinical trial, preferably a randomized controlled study, to examine “evidentiary gaps” in the usage of such scans, particularly with respect to how they potentially improve health outcomes. In addition, Medicare beneficiaries were limited to one reimbursed one Aβ PET scan per lifetime. In October 2023, CMS lifted the CED requirement and one‐scan limit and expanded access to amyloid PET scans for Medicare beneficiaries. ⁴⁴ Despite the invasiveness of CSF testing and the high costs of PET, with the advent of DMTs, the use of CSF biomarker and PET tests is expected to increase due to the confirmation requirement of the presence of Aβ pathology prior to DMT initiation.

Our study was not designed to understand the reasons why patients received combined or repeated diagnostic tests. However, as dementia has many different types, in practice, clinicians often need to evaluate a range of potential diagnoses, simultaneously using combined data from multiple tests and biomarkers. Previous studies have suggested that combining MRI and CSF measures gave better accuracy for discriminating between AD and vascular dementia, between AD and dementia with Lewy bodies, and between frontotemporal dementia and dementia with Lewy bodies. ⁴⁵ Additionally, AD is characterized by gradual cognitive decline and progressive cerebral atrophy. Clinicians may repeat MRI or CT scans to measure patients’ brain atrophy over time. ⁴⁶ , ⁴⁷ With DMTs becoming available, the use of repeated MRI imaging testing will also be expected to rise. Monoclonal antibodies directed against aggregated forms of Aβ can cause amyloid related imaging abnormalities (ARIAs). Frequent brain MRI tests are needed to monitor ARIAs. For example, for the administration of kisunla (donanemab‐azbt), patients are required to obtain a brain MRI prior to initiating treatment as well as prior to the second, third, fourth, and seventh infusions. ⁴⁸

Our study had limitations. First, we rely on CPT codes to identify the diagnostic tests of interest for our study. Because of the non‐specific nature of CPT codes, we were not able to include common cognitive assessment tests, such as the Mini‐Mental State Examination (MMSE). It is possible that some of the rule‐out tests, such as PET scans, may not be exclusively related to dementia diagnosis. As a result, their actual use for dementia‐related evaluations in real‐world practice may be overestimated. Additionally, there may be some misclassification of the frequency of diagnostic tests included in the study. Second, claims data are extremely valuable for addressing the study objectives. However, they are derived from reimbursement information for the payment of bills. The use of claims data to identify MCI, AD, or dementia cases is a far less precise approach than has been undertaken in existing prospective cohort studies like The Aging, Demographics, and Memory Study (ADAMS), the Framingham Heart Study, or the Health and Retirement Study in which participants receive comprehensive screenings and must meet specific diagnostic criteria to be considered a case. ²⁷ Third, AD or dementia in general tends to be underdiagnosed. ⁴⁹ , ⁵⁰ Undiagnosed patients are not captured in this study. Fourth, Medicare advantage (MA) enrollees are excluded from the study because of the lack of data. Thus, the study findings could not be generalized to this population. Given that MA is a rapidly growing segment of the Medicare market, the study needs to be replicated in the MA population in the future. Fourth, our analysis of amyloid PET use was limited to the period prior to October 2023, when CMS lifted the CED requirement. Therefore, the amyloid PET scans captured in the study may reflect research‐related use rather than routine clinical use in real‐world practice. Fifth, our study population did not fully represent the IDEAS cohort. The IDEAS study assessed the impact of amyloid PET scans on health outcomes for Medicare beneficiaries aged ≥ 65, already diagnosed with MCI or dementia. Our study, however, examined the use of diagnostic tests—both to rule out and confirm—in Medicare patients before they received a diagnosis of MCI, AD, or other dementias, focusing on those newly diagnosed. Finally, we cannot measure many variables that may be associated with the use of the selected diagnostic tests, such as physician knowledge and health‐care access.

In conclusion, this is the first real‐world study to provide a comprehensive overview on understanding the use of diagnostic tests prior to the diagnosis of MCI, AD, or other dementias in a large, nationally representative sample. The study suggests that in the 6 years prior to the availability of DMTs, blood tests, MRI, and CT were the predominant diagnostic tests, while the use of CSF and PET was very infrequent. Despite the modest increasing trends, substantial improvements are needed in the use of confirmatory tests, especially PET and CSF, which will be necessary for increased use of new DMTs. Additionally, patients often received combined or repeated diagnostic tests. Future research is needed to understand the reasons for patients undergoing multiple or repeated testing to optimize the timing of a definitive diagnosis.

CONFLICT OF INTEREST STATEMENT

J.Y., T.M., V.R., V.S., and S.R. are employees of and shareholders in F. Hoffmann‐La Roche Ltd. A.D. is an employee of Genesis Research Group, which receives a consultation fee from Roche Diagnostics International. M.N.S. is from the Barrow Neurological Institute and received consulting fees from Roche Diagnostics International. All author disclosures are available in the supporting information.

CONSENT STATEMENT

The Medicare claims data used for this study were created in accordance with the principles of the Health Insurance Portability and Accountability Act (HIPAA). Consent from human subjects was unnecessary.

Supporting information

Supporting Information

Supporting Information

Yan JT, Dillon A, Meng T, et al. Real‐world use of diagnostic tests for mild cognitive impairment, Alzheimer's disease, and other dementias in Medicare fee‐for‐service beneficiaries. Alzheimer's Dement. 2025;17:e70156. 10.1002/dad2.70156