Accelerated clinical evaluations for Alzheimer's disease: feasibility across specialty and rural care settings

Abstract

INTRODUCTION

Timely diagnosis of Alzheimer's disease (AD) is increasingly important given the availability of disease‐modifying therapies, yet current diagnostic pathways remain fragmented, resource‐intensive, and difficult to access. We developed the accelerated clinical evaluation for Alzheimer's disease (ACE‐AD), a nursing‐led workflow integrating a clinical interview, cognitive testing, blood biomarkers, and portable magnetic resonance imaging (MRI) to streamline dementia evaluation.

METHODS

ACE‐AD was evaluated in two phases. Phase 1 validated the workflow in 20 participants with standard‐of‐care (SOC) memory clinic diagnoses compared against ACE‐AD‐derived etiologic diagnosis, cognitive staging, and neuroimaging findings. Phase 2 adapted the workflow for a rural neurology practice, assessing feasibility, utility, and acceptability.

RESULTS

In Phase 1 (n = 20), primary etiologic diagnoses were universally concordant between SOC and ACE‐AD evaluations, with 90% agreement on AD as a primary or contributing etiology. Cognitive stage agreed in 13 participants (65%); ACE‐AD rated all discordant cases as less impaired. Portable MRI identified clinically relevant abnormalities consistent with SOC MRI, and patients, care partners, and clinicians rated the workflow highly acceptable. In Phase 2 (n = 7), the rural neurologist reported improved diagnostic confidence, efficiency, treatment planning, and patient communication without workflow disruption.

DISCUSSION

A structured, nursing‐led evaluation integrating blood biomarkers and portable MRI can capture sufficient clinical information to support specialist dementia diagnosis while shifting much of the diagnostic workup earlier in the care pathway. These findings support further evaluation of ACE‐AD as a scalable approach for expanding access to timely dementia diagnosis in both specialty and resource‐limited settings.

HIGHLIGHTS

• ACE‐AD integrates nursing assessment, blood biomarkers, and portable MRI, preserving specialist oversight while shifting work earlier.

• Primary ACE‐AD etiologic diagnoses were highly concordant with SOC.

• Patients, care partners, and clinicians rated ACE‐AD highly acceptable.

1. BACKGROUND

Timely and accurate diagnosis of Alzheimer's disease (AD) and related dementias remains a persistent challenge, particularly in settings where access to specialty cognitive care is limited. ¹ , ² Conventional diagnostic pathways typically require multiple providers and clinic visits, fragmented and sequential testing, and variable use of advanced biomarkers, resulting in delays and prolonged time to diagnosis. ³ , ⁴ These structural limitations pose increasingly consequential challenges to an aging population often waiting months to see a specialist. ⁵ , ⁶ The emergence of disease‐modifying therapies that require early and accurate identification for maximum benefit compounds this concern. ⁷ , ⁸

Recent advances in blood‐based biomarkers and portable, ultra‐low‐field neuroimaging provide an opportunity to redesign and scale the evaluation of cognitive impairment. Blood biomarkers such as plasma phosphorylated tau‐217 (p‐tau217) offer minimally invasive indicators of AD pathology. In particular, p‐tau217 and its derivative measures have demonstrated diagnostic performance approaching that of cerebrospinal fluid biomarkers and amyloid positron emission tomography. These findings support the emerging use of plasma markers in routine clinical evaluation, in conjunction with other markers across a range of settings. ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ Likewise, portable magnetic resonance imaging (MRI) systems enable structural brain imaging in outpatient and resource‐limited settings where access to conventional high‐field MRI may be limited. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Although image resolution is lower and current portable MRI systems are not suitable for detecting cerebral microhemorrhages or other subtle abnormalities, they can identify major structural findings and provide important diagnostic support in ruling out mass lesions, hydrocephalus, large infarcts, and substantial cerebrovascular disease as part of differential diagnosis. When combined with the conventional structured clinical interviews and brief cognitive testing, these tools offer the potential to consolidate key elements of dementia diagnosis into a triage‐based efficient workflow that could support greater access to timely care. ¹⁶

With this need in mind, we sought to develop and test a structured, single‐visit evaluation for dementia. We call this model the accelerated clinical evaluation for Alzheimer's disease (ACE‐AD). Conceptually, ACE‐AD separates testing and information gathering from clinical synthesis, enabling trained staff to efficiently collect standardized clinical, cognitive, biomarker, and imaging data and summarize the findings in a structured report that supports the clinician's evaluation and care planning. See Figure 1.

FIGURE 1.

Comparison of conventional SOC evaluation and ACE‐AD.

We evaluated this approach in a prospective, multi‐setting feasibility and preliminary implementation study spanning two complementary cohorts representing sequential phases of development and translation. In Phase 1, the ACE‐AD workflow was performed and validated against a participant's own prior standard‐of‐care (SOC) evaluation in an academic memory clinic to refine and evaluate the approach under controlled conditions. In Phase 2, the workflow was adapted for a rural community to assess feasibility, acceptability, and implementation. By presenting both phases, we aim to capture progressive and complementary evidence on feasibility, acceptability, and informational value of the ACE‐AD workflow across distinct clinical settings. Our goal for these analyses was not hypothesis testing but to demonstrate that the ACE‐AD could capture the clinical nuance necessary for accurate diagnosis.

Both phases were approved by the University of Kansas Medical Center Institutional Review Board (IRB Nos. 161334 and 161708) and conducted in accordance with the principles of the Declaration of Helsinki. A standardized protocol was used to assess participants’ capacity for informed consent, with consent obtained from legally authorized representatives when indicated.

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed PubMed and reference lists of relevant publications addressing dementia diagnostic pathways, blood‐based biomarkers, portable MRI, nursing‐led assessment, and rural access to specialty dementia care. Prior studies support these individual approaches, but evidence integrating them within a single diagnostic workflow remains limited.

2. Interpretation: ACE‐AD produced etiologic diagnoses concordant with standard memory clinic evaluations and was acceptable to patients, care partners, and clinicians. Implementation with a rural neurology practice suggests that a structured, nursing‐led evaluation can shift diagnostic work earlier in the care pathway while preserving specialist oversight and supporting clinical decision‐making.

3. Future directions: Larger, multisite studies should evaluate diagnostic accuracy, cognitive staging, cost, implementation requirements, and effects on time to diagnosis and treatment. Further work should determine which patients and clinical settings are best suited to ACE‐AD and whether the workflow improves access and outcomes across varied specialty and community practices.

2. METHODS

2.1. Phase 1–memory clinic validation

Participants were adults recruited from an urban, tertiary care neurology clinic, with an established clinical diagnosis of mild cognitive impairment (MCI) or dementia. All participants had completed a formal SOC clinical evaluation within the previous year and a clinical brain MRI within the previous 2 years. SOC testing was determined by the treating clinician and, in some cases, a supporting neuropsychologist, well before enrollment in this study. All participants were required to have a study partner familiar with their cognitive and functional status. Individuals were excluded if they were unable to undergo MRI or had cognitive impairment severe enough to preclude meaningful participation in study procedures.

After informed consent, a registered nurse not affiliated with the memory clinic conducted the ACE‐AD in a single visit. Before conducting study visits, the nurse with more than 5 years of clinical research experience in dementia (A.M.) completed training in administration of the structured interview and cognitive and functional tests, with an experienced neurologist (J.M.B.), and operation of the portable MRI system. The ACE‐AD included a structured interview to systematically capture cognitive symptoms, functional status, neuropsychiatric features, and relevant medical history, incorporating additional input from the study partner. The Short Test of Mental Status (STMS) and the Functional Activities Questionnaire (FAQ) provided validated, objective cognitive and functional measures. ¹⁷ , ¹⁸ Blood samples were obtained for analysis of p‐tau217 according to College of American Pathologists/Clinical Laboratory Improvement Amendments standards using a validated laboratory‐developed test for plasma p‐tau217 (Lumipulse G1200, Fujirebio Diagnostics, Inc.). Participants also underwent neuroimaging using a portable MRI system (0.064T Hyperfine Swoop version 2, Guilford, CT, USA). Acquired sequences included a fluid‐attenuated inversion recovery (echo time [TE] 152.44 ms, repetition time [TR] 3000 ms, inversion time [TI] 1173.9 ms, slice thickness 5 mm), T1‐weighted imaging (TE 4.82 ms, TR 725 ms, TI 240 ms, slice thickness 5 mm), diffusion‐weighted imaging (TE 62.21 ms, TR 850 ms, slice thickness 5.88 mm), and T2*‐weighted imaging (TE 140 ms, TR 1800 ms, slice thickness 5 mm).

Information collected during the ACE‐AD evaluation was compiled into a single structured report organized around the assessment elements of Current Procedural Terminology (CPT) code 99483. The report summarized the participant's medical and social history; cognitive, functional, and neuropsychiatric concerns; medication use; safety and caregiving considerations; objective cognitive and functional test results; and blood biomarker results. The initial report format was editable to allow the research clinician to review and add a radiologic impression and care plan. The entire evaluation was completed by a single nurse, though an assistant was often helpful with positioning in the MRI scan.

The clinicians were one of two neurologists not involved in the participants’ SOC visits. The research clinician reviewed the MRI for clear clinically relevant structural findings, amended the ACE‐AD report with a research diagnosis including etiology and cognitive stage, and notified study staff to proceed to the research diagnosis meeting. Research clinicians did not review participants’ SOC diagnostic assessment or prior clinical notes. However, the ACE‐AD structured medical history presented current medications, including anti‐amyloid therapy when applicable.

Approximately 1 week after the initial ACE‐AD evaluation, participants met with the research clinician for a research diagnostic visit. Clinicians discussed the findings with participants and study partners and provided recommendations for further evaluation or management. As a secondary outcome of interest, participants were randomized to complete the research diagnostic visit either in person or via telehealth to evaluate flexibility of delivery.

Following completion of enrollment, we analyzed concordance between the SOC and ACE‐AD. The patient's most contemporaneous SOC diagnosis and cognitive stage before the ACE‐AD served as the reference standard. For each participant, a single reviewer (J.M.B.) identified the clinical note nearest in time to the ACE‐AD visit (median interval, 84 days) and abstracted the coded diagnosis together with the clinical reasoning documented, including the assessment of stage, etiology, possible contributing conditions, and testing. Etiology was taken from the diagnostic code. For participants coded as MCI and for those with suspected mixed or multiple contributing pathologies, etiology was determined from the documented clinical assessment in conjunction with the diagnostic code. Stage was classified as subjective cognitive impairment (SCI), MCI, or dementia. When the index SOC evaluation note did not fully convey the treating clinician's judgment, adjacent clinic notes were reviewed to establish the diagnostic impression, disease trajectory, and relevant biomarker or treatment information. The resulting SOC stage and etiology were then compared with the corresponding ACE‐AD determinations.

A radiologist (K.S.), blinded to the SOC MRI findings and ACE‐AD diagnostic assessment, independently reviewed each portable MRI scan. The radiologist rated overall image quality and evaluated the scans for clinically relevant structural findings, including patterns of cerebral atrophy, white matter disease, infarction, hemorrhage, and other intracranial abnormalities.

Analyses were primarily descriptive and focused on characterizing the feasibility, acceptability, and informational value of the ACE‐AD. We summarized participant demographics and visit characteristics, along with key elements captured during the structured interview and assessment, to illustrate the breadth and clinical relevance of the information obtained. Diagnostic concordance between the ACE‐AD and the SOC evaluation was assessed for cognitive stage and etiologic diagnosis using percentage agreement.

We assessed acceptability using post‐visit surveys completed by participants and by providers. Participant surveys focused on clarity, efficiency, and overall experience, while provider surveys assessed usability, workflow integration, and overall impressions. Clinician and participant responses were summarized using standard Likert scales. Overall, the goal of these analyses was not hypothesis testing but to determine whether the ACE‐AD could capture the clinical information and diagnostic nuance necessary to support dementia evaluation while providing an acceptable experience.

2.2. Phase 2–rural neurology practice methods

Following Phase 1 assessment of feasibility and acceptability in the urban memory clinic, we adapted the ACE‐AD workflow to support a rural general neurology practice. Through ongoing discussions and co‐design, we determined that an adapted ACE‐AD workflow administered by a nursing team prior to a patient visit with the partnering neurologist would best accommodate the time and support needs of the practice and the patients. The neurologist notified patients with upcoming scheduled visits of the study opportunity and provided them with study team contact information if they were interested. Participants were required to be ≥50 years old, be able to provide informed consent, either directly or via a legally authorized representative, and be scheduled for a visit with the partnering neurologist within the next 5 months. MRI contraindications were not considered exclusionary for Phase 2. To minimize burden, informed consents were obtained over the telephone with a paper copy of the consent mailed to participants for review prior to the screening and consent call.

During the initial visit, a trained registered nurse or nurse practitioner conducted the ACE‐AD structured clinical interview. Before conducting study visits, the evaluation team was trained in administration of the structured interview, STMS, FAQ with an experienced memory care nurse practitioner (S.A.F.), and operation of the portable MRI system. The STMS, FAQ, blood samples, and participant post‐visit surveys were obtained in the same manner as Phase 1. Blood was shipped in chilled containers using overnight shipping and processed in the same lab as Phase 1. Results were summarized in the same manner as in Phase 1 and provided prior to the participants’ next scheduled neurology appointment. Access to cloud‐based MRI was provided to the partnering rural neurologist and a study neurologist (J.M.B.), who reviewed the scan for any emergent clinically relevant abnormalities.

Following completion of Phase 2, the partnering rural neurologist completed a structured survey assessing the perceived utility of the ACE‐AD reports across all evaluated patients collectively. The survey included Likert‐scale and multiple‐response items evaluating impact on diagnostic confidence, clinical efficiency, and decision‐making, as well as open‐ended questions regarding strengths, limitations, and implementation preferences. Because the neurologist remained the treating physician, and regulatory and cross‐health‐system constraints made collection of final diagnoses impractical, diagnostic concordance between the ACE‐AD and the clinical evaluation was not assessed.

3. RESULTS

3.1. Phase 1–memory clinic cohort results

3.1.1. Participant characteristics

Twenty participants were enrolled (mean age 71.0 years [standard deviation (SD) 8.0]; 11 women). Participants attended the interview with either a spouse (80%) or close family member (20%) as study partner. Educational attainment was generally high, with 90% completing at least some post‐secondary education. Table 1 summarizes participant‐level comparisons between the memory clinic and ACE‐AD evaluations.

TABLE 1.

Per‐participant staging, diagnosis, cognitive, functional, and biomarker data (memory clinic) versus ACE.

	Stage of cognitive impairment		Etiologic Diagnosis		Cognitive Performance (STMS)		Functional Assessment (FAQ score unless indicated*)		Amyloid Assessment
ID	SoC	ACE‐AD	SoC	ACE‐AD	SoC	ACE‐AD	SoC	ACE‐AD	SoC	ACE‐AD
A a	Dementia	Dementia	AD	AD	+++	++++	+	+	Elevated (CSF)	Elevated (Plasma)
B	Dementia	Dementia	AD	AD	+++	+++	++	++	Elevated (PET)	Elevated (Plasma)
C	Dementia	Dementia	AD	AD	+++	+++	+*	+	None	Elevated (Plasma)
D a	Dementia	Dementia	AD	AD	+++	++	+	+	Elevated (CSF)	Elevated (Plasma)
E	Dementia	Dementia	Multi: vascular + possible AD	Non‐AD multi‐etiology	++++	++++	+*	+	None	Non‐elevated (Plasma)
F	Dementia	Dementia	FTD	FTD	+++	+++	+*	+	Non‐elevated (Plasma)	Non‐elevated (Plasma)
G	Dementia	Dementia	AD	AD	+	+	+*	+	Elevated (Plasma)	Elevated (Plasma)
H a	Dementia	MCI b	AD	AD	+++++	+++++	+	++	Non‐elevated (PET)	Elevated (Plasma)
I	Dementia	MCI b	AD	AD	+++	+++	+*	++	Elevated (CSF)	Elevated (Plasma)
J	Dementia	MCI b	AD	AD	++++	++++	+*	++	Elevated (CSF)	Elevated (Plasma)
K a	Dementia	MCI b	AD	AD	++++	++++	++	+	Elevated (PET)	Indeterminate (Plasma)
L	MCI	MCI	AD	AD	++++	++++	++*	++	Elevated (CSF)	Elevated (Plasma)
M	MCI	MCI	AD	AD	++++	++++	++*	+	Elevated (PET)	Elevated (Plasma)
N	MCI	MCI	Non‐AD indeterminate	Non‐AD indeterminate	+++++	+++++	++*	++	Elevated (CSF) Non‐elevated (PET)	Indeterminate (Plasma)
O a	MCI	MCI	AD	AD	+++++	+++++	++*	++	Elevated (PET)	Non‐elevated (Plasma)
P a	MCI	MCI	AD	AD	++++	++++	+	+	Elevated (CSF)	Elevated (Plasma)
Q	MCI	MCI	Suspected prodromal DLB	Prodromal DLB, possible AD	+++++	+++++	++*	++	None	Indeterminate (Plasma)
R	MCI	SCI b	Non‐AD indeterminate	Non‐AD indeterminate	++++	+++++	++*	++	Non‐elevated (PET)	Non‐elevated (Plasma)
S a	MCI	SCI b	AD	AD	++++	+++++	++	++	Elevated (PET)	Non‐elevated (Plasma)
T	MCI	SCI b	Non‐AD multi	Non‐AD multi	+++++	+++++	++*	++	Non‐elevated (Plasma)	Non‐elevated (Plasma)

Cognitive and functional measures: STMS scores (range, 0–38) are presented as binned performance ranges to enhance privacy: +++++, ≥35; ++++, 30–34; +++, 20–29; ++, 10–19; and +, <10. FAQ scores (range, 0–30) are similarly grouped: ++, 0–9 (no or uncertain functional impairment); +, 10–30 (possible functional impairment). FAST stages range from 1–7, with higher stages indicating greater functional impairment, and are grouped as ++, stages 1–3; +, stages 4–7. SoC evaluations documented either FAST or FAQ, whereas ACE‐AD evaluations included FAQ for all participants.

Amyloid assessment: Results are presented as Elevated, Indeterminate, or Non‐elevated based on the documented clinician interpretation or contemporaneous guidance and therefore may differ from classifications using current diagnostic thresholds. Test modality is shown in parentheses: CSF, phosphorylated tau‐217/amyloid‐β42 ratio; PET, florbetapir F‐18 positron emission tomography; Plasma, phosphorylated tau‐217 concentration.

Abbreviations: SoC, standard of care; SCI, subjective cognitive impairment; MCI, mild cognitive impairment; ACE‐AD, Accelerated Clinical Evaluation; STMS, Short Test of Mental Status; FAQ, Functional Activities Questionnaire; FAST, Functional Assessment Staging Tool. identified with a *; CSF, cerebrospinal fluid; PET, florbetapir F‐18 positron emission tomography; Plasma, plasma phosphorylated tau‐217 concentration.

^a Participant received anti‐amyloid therapy.

^b ACE‐AD stage differed from the memory clinic stage; all seven differences represented a less impaired classification by ACE‐AD.

3.1.2. Etiology

Based on the SOC evaluation, AD was the primary etiology in 14 participants; two participants had mixed etiologies, one had frontotemporal degeneration, one had suspected prodromal dementia with Lewy bodies, and two had cognitive impairment of uncertain etiology. ACE‐AD identified the same primary etiologic diagnosis as the SOC evaluation in all 20 participants (100%). Agreement regarding AD as either a primary or contributing etiology was 90%. The two discordant cases reflected differing assessments of possible secondary Alzheimer's pathology in the absence of confirmatory amyloid biomarkers. Seven participants were receiving anti‐amyloid therapy at the time of the ACE‐AD, limiting direct comparison of contemporaneous plasma biomarkers with previously established amyloid status.

3.1.3. Diagnostic staging

Based on the SOC evaluation, 11 participants were classified as having dementia and nine as having MCI. ACE‐AD assigned the same cognitive stage in 13 of 20 participants (65%). Among the seven discordant cases, ACE‐AD consistently assigned a less impaired stage: four participants with SOC dementia were classified as having MCI in the ACE‐AD, and three participants with SOC MCI were classified as having SCI in the ACE‐AD. No participant was assigned a more advanced stage by ACE‐AD. In five of the seven discordant cases, the less impaired ACE‐AD classification was supported by relatively preserved cognitive and functional performance (STMS 33 to 37/38; FAQ 0 to 7). The remaining two participants had initiated anti‐amyloid therapy between the SOC and ACE‐AD evaluations. Thus, these differences may reflect changes in clinical status or the information available at each evaluation rather than misclassification.

3.1.4. Structural neuroimaging

Portable ultra‐low‐field (portable) MRI and conventional high‐field MRI were concordant for clinically relevant structural findings, including mass lesions, large‐territory infarcts, hydrocephalus, and severe cerebrovascular disease. Both modalities also uniformly identified prominent patterns of cerebral atrophy.

From a clinical perspective, ACE‐AD clinicians considered portable MRI sufficient for the initial diagnostic evaluation in all participants, although conventional high‐field MRI was recommended in six of 20 participants, including five pursuing anti‐amyloid therapy. As expected, radiologists noted lower image quality on portable MRI than on conventional MRI in 75% of portable MRI images (15 of 20). The portable MRI identified a small medial temporal T2 hypointensity not reported on the conventional MRI, whereas the conventional MRI identified an incidental parotid gland mass not visualized on the portable MRI.

3.1.5. Clinical characterization

During the ACE‐AD interview participants endorsed a range of cognitive, functional, and neuropsychiatric concerns, with memory complaints nearly universal and substantial representation of executive, visuospatial, and neuropsychiatric symptoms. Functional impacts were most evident in instrumental activities of daily living and aligned with cognitive stage. Cognitive and functional performance showed substantial variability (mean STMS 30.3 [SD 8.2]; mean FAQ 10.8 [SD 8.7]). A summary of symptom endorsement and participant‐reported perceptions of the evaluation is provided in Table S1.

3.1.6. Participant and care partner experience

Participants and care partners rated both the nurse‐led ACE‐AD evaluation and the follow‐up diagnostic visit uniformly high: All respondents found the visits easy to complete and the information understandable, and at least 95% expressed confidence in the diagnosis and plan and endorsed the single‐visit approach. The lowest‐rated item was in‐scanner comfort during portable MRI (75% agreement).

3.1.7. Virtual versus in‐person research diagnostic visit

Responses on post‐visit surveys were highly favorable regardless of visit format, with no clear pattern, suggesting that virtual diagnostic visits were less understandable or acceptable than in‐clinic visits. Two virtual respondents noted logistical or communication challenges specific to the telehealth platform.

3.1.8. Clinician experience

ACE‐AD clinicians unanimously rated the workflow as feasible (100%) and acceptable (100%) across participants. Satisfaction with workflow guidance was also high, with one early implementation issue related to information security delaying MRI acquisition and prompting refinement of written guidance for subsequent visits.

3.2. Phase 2–rural neurology practice cohort results

3.2.1. Participant characteristics

Twelve participants contacted the study team, and seven were ultimately enrolled. Participants had a mean age of 75.8 years (SD = 6.6); six were male. Most attended visits with a spouse (n = 5) or close family member (n = 1) who contributed collateral history. Educational attainment was high, with six of seven participants (86%) completing at least some post‐secondary education. Only four participants were eligible for MRI.

3.2.2. Clinical characterization

As in Phase 1, participants endorsed a range of cognitive, functional, and neuropsychiatric concerns, with memory complaints most common and additional representation of executive, visuospatial, and behavioral symptoms. Basic activities of daily living were largely preserved, while instrumental activities demonstrated greater variability and reflected cognitive stage. Cognitive and functional performance showed substantial variability (mean STMS 31.0 [SD 6.1]; mean FAQ 7.6 [SD 9.4]). A summary of symptom endorsement and participant‐reported perceptions in this phase is provided in Table S2.

3.2.3. Participant and care partner experience

Due to a data capture issue, only four participants and care partner dyads completed a visit experience survey following the ACE‐AD evaluation. Participants and care partners rated the ACE‐AD evaluation uniformly, agreeing or strongly agreeing with the following statements: The visits were easy to complete, the information understandable, and the MRI was comfortable. Surveys following the diagnostic visit were not administered in Phase 2 because, by design, the research team was not notified when those visits occurred.

3.2.4. Clinical utility

The ACE‐AD reports did not alter the neurologist's final diagnoses. Rather, the summary reports were perceived to “support earlier diagnosis,” “inform treatment decisions,” and “improve patient communication,” as well as “greatly increasing diagnostic confidence,” with clinic visits described as “much more efficient” when supported by the ACE‐AD report. The reports were seen as “very helpful” in informing clinical decision‐making across the evaluated cases.

Among individual report components, blood‐based biomarkers, MRI findings, and functional assessments (including FAQ and functional history) were rated as “extremely useful,” while objective cognitive testing (STMS) was rated as “very useful.” The clinical history summary was considered “moderately useful,” reflecting that this information is often independently obtained during the clinical encounter.

In open‐ended responses, the neurologist identified blood‐based biomarkers as the most valuable component of the evaluation. The overall workflow was viewed favorably, with a preference for external support to conduct future evaluations rather than relying on internal clinic staff. The neurologist emphasized the importance of targeting rural areas with limited imaging access and suggested that further expansion for more rural and frontier populations would enhance the impact of the ACE‐AD model.

4. DISCUSSION

We conducted a multi‐phase feasibility and acceptability study of a streamlined, single‐visit, nursing‐led dementia evaluation integrating clinical history, cognitive and functional assessments, blood biomarkers, and portable MRI (ACE‐AD). Findings were synthesized into a structured report, organized around the assessment elements of CPT 99483, for use by the diagnosing clinician. Phase 1 evaluated the workflow in an academic memory clinic, while Phase 2 assessed its implementation in a rural neurology practice with fewer specialty resources. Overall, patients and clinicians found the workflow feasible, acceptable, and easy to complete. In Phase 1, ACE‐AD findings showed complete concordance with SOC evaluations for primary etiology, although agreement on cognitive stage was more modest. In Phase 2, the report was considered useful for diagnostic decision‐making, treatment planning, and patient communication.

Our approach allows diagnostic information typically obtained across fragmented and asynchronous visits to be acquired within a single visit, substantially reducing delays inherent in traditional sequential models. The success of the ACE‐AD demonstrates that a structured, single‐visit diagnostic evaluation can be delivered earlier in the care pathway using a team‐based approach. By separating standard clinical history taking and testing from diagnosis and care planning, the model has the potential to reduce neurologist time devoted to front‐end evaluation while preserving the specialist's central role in clinical interpretation and decision‐making. This improved efficiency is needed, since dementia diagnosis is often a lengthy, fragmented process, with some timelines approaching 2 years from first reported concern to diagnosis. ² , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ In areas with limited specialist resources or high travel burden, such as many rural communities, this often leads to underdiagnosis and treatment delays. ²⁴ , ²⁵ , ²⁶

Collectively, these findings suggest that condensed models such as ACE‐AD could facilitate the reorganization of initial dementia evaluation process, enabling much of the diagnostic workup to occur before a specialty cognitive visit. For example, we envision ACE‐AD, triggered by a cognitive concern identified during a Medicare Annual Wellness Visit, being implemented by clinical staff (in primary care or memory clinic) prior to a diagnostic visit with the treating clinician. This could occur at the clinic, through a mobile health team external to the clinic setting, or even in community environments such as community centers or churches. This approach also has the potential to improve patient triage, allowing primary care and general neurology to diagnose and manage straightforward cases while reserving memory clinics for patients requiring complex diagnostic clarification or advanced treatment decisions. By aligning clinical expertise with patient complexity, the ACE‐AD workflow could shorten the time from evaluation to diagnosis and treatment while expanding access to specialty dementia care. The next step is to develop and evaluate implementation strategies tailored to memory clinics, general neurology practices, and primary care.

Our concordance analysis provides insight into the potential value and challenges of the ACE‐AD model. Agreement for primary etiologic diagnosis was encouraging, suggesting that the structured interview, cognitive assessment, and portable MRI frequently captured the key information necessary to identify the underlying cause of cognitive impairment. Agreement for cognitive stage was more modest, although all disagreements occurred between adjacent stages and ACE‐AD consistently classified participants as less impaired. This pattern may also reflect differences in clinical staging and documentation. In routine practice, clinicians may preferentially classify borderline presentations as MCI rather than SCI because of implications for further diagnostic evaluation or access to services. Consequently, some apparent disagreements may reflect differences in clinical classification rather than meaningful differences in diagnostic interpretation. However, it is also possible that discordance was related to more in‐depth neuropsychological testing, treatment, or even natural fluctuation. Importantly, no participant was classified as more impaired by the ACE‐AD than by the SOC evaluation, suggesting that the workflow preserved diagnostic caution while accurately identifying the underlying disease process. Although distinguishing MCI from early dementia remains clinically important, timely identification of patients in the earliest stages of cognitive impairment may be more consequential for ensuring prompt diagnostic evaluation and access to disease‐modifying therapies.

The neuroimaging findings similarly illustrate both the promise and current limitations of integrating portable MRI into dementia evaluation. Portable MRI had sufficient quality and resolution to identify the major structural abnormalities necessary to support the differential diagnosis of cognitive impairment, ²⁷ while conventional MRI remained necessary for the evaluation of cerebral microhemorrhages requiring higher‐resolution imaging. Although each modality identified isolated incidental findings not reported by the other, these did not alter the clinical assessment of cognitive impairment. Radiologists predictably viewed image quality less favorably than clinicians, reflecting the current limitations of portable imaging, whereas both the ACE‐AD research clinicians and the rural neurologist found the studies clinically useful for their intended purpose of excluding major structural pathology and informing diagnostic decision‐making. Future implementation will require clearer operational guidance regarding MRI safety screening and image acquisition responsibilities, particularly in settings where advanced imaging expertise is limited. File S3 provides a structured operating framework for deploying portable MRI we developed following this study, although broader clinical and regulatory guidance will be necessary if portable MRI is to fulfill its potential for expanding access.

Several limitations should be considered. The sample size was modest, particularly in the rural implementation of Phase 2, and this study was not designed to establish diagnostic accuracy or clinical effectiveness. Diagnostic concordance was assessed against clinician‐abstracted SOC documentation rather than an independently adjudicated reference standard, and comparison of MRI findings relied on the abstraction of conventional MRI reports rather than blinded rereview of the original images. The structured ACE‐AD report used by the research clinician included medications such as anti‐amyloid therapies that may have influenced determinations and our analysis of etiologic agreement. In addition, participants in Phase 1 were recruited from specialty memory clinics, which may limit generalizability to broader primary care populations, and the ACE‐AD workflow was implemented by investigators experienced with the model. Additional research on scaling and implementation is critical to establish ACE‐AD or similar models as suitable for use. Further, scope‐of‐practice and safety guidance considerations need to evolve to allow more expansive deployment of portable MRI in clinical practice. Finally, we did not assess whether patient and care‐partner acceptability ratings may have been influenced by social desirability bias, as both phases were implemented in conjunction with real‐world care. Nonetheless, these findings support the continued evolution and scaling of the model.

The growing need for dementia evaluation requires the rapid development of care models that extend diagnostic capacity beyond specialty memory clinics while preserving the quality and rigor of the diagnostic process.

Models such as ACE‐AD are unlikely to fully address the limited availability of cognitive specialty care in rural communities. Broader strategies may include incentives for rural neurology training and practice, as well as expanded telehealth access to urban‐based neurologists and appropriately trained advanced practice providers. ⁴ Appropriate training of clinicians will be critical, especially for under‐resourced communities, regardless of approach. ²⁸ However, by standardizing information collection before the diagnostic consultation with trained staff, these accelerated approaches can complement the spectrum of workforce strategies and help clinicians evaluate both urban and rural patients more efficiently.

The ACE‐AD model demonstrates the feasibility of reorganizing much of the initial dementia evaluation into a structured, standardized workflow that can support clinicians across diverse practice settings. Although additional refinement, validation, and implementation work is needed, this approach has the potential to improve access to timely diagnosis and treatment, better align specialty resources with patient complexity, and provide a scalable framework for expanding high‐quality dementia care in both metropolitan and rural communities.

CONFLICT OF INTEREST STATEMENT

RAT reports institutional research support for clinical trials from Eisai, Ono, Novartis, and Eli Lilly. JLW serves on advisory boards for Eli Lilly and Novo Nordisk. JMB reports institutional research support from Eli Lilly, Amylyx, Biogen, Eisai, AbbVie, Roche, Ionis, AstraZeneca, National Institutes of Health, and the Alzheimer's Association; consulting fees from multiple pharmaceutical and biotechnology companies; honoraria and travel support from Eli Lilly; and advisory board service for Intra‐Cellular Therapies. The remaining authors declare no competing interests. Full disclosures are provided in the submitted ICMJE forms. Author disclosures are available in the Supporting Information.

CONSENT STATEMENT

All participants provided written informed consent prior to participation. The study was approved by the University of Kansas Institutional Review Board and conducted in accordance with the Declaration of Helsinki.

Supporting information

Supporting Information

Supporting Information