Recognising Mild Cognitive Impairment and Early Alzheimer's Disease in General and Psychiatric Practice: Clinical Clues and Appropriate Referral Authors: Kengo Shimoda Published: 2026 ABSTRACT ABSTRACTEarly cognitive decline may become evident through changes in daily activities or behaviour before it is apparent in a brief consultation. This selective narrative review supports case‐finding and appropriate referral in general and psychiatric practice. Four actions organise the review: identify decline in previously mastered instrumental activities of daily living; recognise new apathy or sustained behavioural change; use a cognitive screen sensitive to mild impairment, such as the Montreal Cognitive Assessment; and refer when clinically meaningful concern remains. Functional and behavioural findings require interpretation alongside medical, psychiatric and social factors. Mild behavioural impairment has reported associations with both amyloid and tau, varying across populations. Depression, medication exposure or a reassuring screening score should not end the assessment of otherwise unexplained decline. Collateral information strengthens the assessment but living alone and lacking a reliable informant are distinct circumstances. Municipal and care services may contribute observations without resolving every gap. Dementia disease medical centres should coordinate assessment and support with referring clinicians and community services.Keywords: activities of daily living, Alzheimer disease, apathy, cognitive dysfunction, referral and consultation Introduction Early cognitive decline is often recognised through changes in one's ordinary life: bills become difficult, medication is taken twice or a previously engaged person stops initiating familiar activities. General physicians and psychiatrists are well placed to notice these changes during consultations for physical illness, sleep problems or emotional distress. The task is to recognise a clinically meaningful change and decide what should happen next, while remaining open about its cause. Lecanemab and donanemab slowed clinical decline in trials involving mild cognitive impairment (MCI) due to Alzheimer's disease (AD) or mild AD dementia [1, 2]. Japan's optimal‐use guidance concerns these early symptomatic stages [3, 4]. Timely assessment also offers diagnostic clarification, treatment of contributing conditions and support for people who do not receive an antibody. Clinical suspicion must remain distinct from confirmation of AD pathology or assessment of treatment eligibility. Japan's Basic Act on Dementia and the national Basic Plan emphasise dignity, participation and coordinated support [5, 6]. The intended readers are general physicians and psychiatrists who encounter possible cognitive decline during routine care. The purpose is case finding: recognising meaningful changes and deciding who needs further assessment. Four practical messages organise the discussion: recognise instrumental functional decline, recognise new behavioural change, use an MCI‐sensitive cognitive screen and refer when meaningful concern remains (Table 1). Collateral information supports each action. Psychiatric symptoms are considered where they may obscure recognition of decline. Detailed specialist investigations and drug protocols are outside the scope. TABLE 1Four practical messages and their clinical applications.MessageConcrete actionInterpretationRecognise IADL declineAsk about changes in medication, finances, appointments, transport, cooking or familiar devices. Establish who performs each task and why help began.Distinguish inefficiency, cognitive dependence, and unavailable support. Consider non‐cognitive explanations.Recognise new behavioural changeEstablish the onset, persistence, and functional effects of apathy or other acquired behavioural changes.Symptoms warrant cognitive and functional assessments.Use an MCI‐sensitive screenMoCA/MoCA‐J should be considered when mild impairment is suspected. The Mini‐Cog is an optional ultra‐brief assessment.Interpret the score in context. A reassuring result does not exclude a meaningful decline.Refer when concern remainsDiscuss referral for diagnostic clarification and support. Acute presentations require prompt medical assessments.A fixed number of positive findings is not required. Note: The collateral information supports all four actions.Abbreviations: AD, Alzheimer's disease; IADL, instrumental activities of daily living; MCI, mild cognitive impairment; MoCA, Montreal Cognitive Assessment; MoCA‐J, Japanese version of MoCA. Four practical messages and their clinical applications. Note: The collateral information supports all four actions. Abbreviations: AD, Alzheimer's disease; IADL, instrumental activities of daily living; MCI, mild cognitive impairment; MoCA, Montreal Cognitive Assessment; MoCA‐J, Japanese version of MoCA. Methods This selective narrative review was organised around clinical decisions concerning recognition, interpretation and referral. The initial PubMed search covered January 2000 to May 2026 (last initial search: 31 May 2026), was restricted to English‐language articles and was supplemented by reference‐list searching. The search framework combined disease stage and clinical terms: (‘mild cognitive impairment’ OR ‘early Alzheimer disease’ OR ‘prodromal Alzheimer’) AND (‘primary care’ OR ‘general practice’ OR screening OR ‘case‐finding’ OR ‘activities of daily living’ OR apathy OR ‘neuropsychiatric symptoms’ OR AD8 OR ‘Montreal Cognitive Assessment’ OR ‘Mini‐Cog’ OR ‘anti‐amyloid’). Targeted searches and source checks during revision in September 2026 addressed MBI and tau, cognitive screening, apathy and depression, living alone and Japanese service and treatment guidance. Source selection prioritised systematic reviews, relevant cohort and instrument‐validation studies and diagnostic guidance. Key MBI biomarker studies were considered separately because their populations and amyloid strata differed. Official Japanese documents support policy statements and developers' instructions support instrument scoring and conditions of use. The review did not claim exhaustive retrieval and no formal risk‐of‐bias grading or new quantitative synthesis was undertaken. Selection was guided by relevance to the four clinical actions, rather than a systematic‐review protocol. Empirical findings were distinguished from clinical interpretation and proposed service arrangements. The integrated pathway is a clinical synthesis, not an empirically derived prediction rule. Four Practical Messages for Recognition and Referral Identify Decline in Previously Mastered Instrumental Activities MCI requires concern about cognitive decline, objective cognitive impairment and relative preservation of everyday independence [7]. Preserved basic self‐care alone is insufficient. Complex instrumental activities of daily living (IADL) may become slower, less reliable or more effortful, with compensatory strategies or minimal assistance. Substantial dependence attributable to cognitive decline suggests dementia‐level functional impairment. In the AgeCoDe primary‐care cohort, IADL impairment added prognostic information to MCI [8]. The Sydney Memory and Ageing Study linked difficulty with cognitively demanding activities to MCI [9]. These findings support asking about function; they do not make any single difficulty a diagnostic sign of AD. Ask about changes from the person's own baseline in medication, finances, shopping, appointments, transport, cooking and familiar devices. Establish what happened, when it began, whether it recurs and what help is now needed. A missed payment has a different meaning from a sustained inability to manage previously routine banking. Asking whether someone can use a smartphone confounds cognition with experience; asking whether a familiar operation has become difficult is more informative. Record abilities and successful adaptations alongside errors. Apparent independence may conceal increasing assistance. Ask who now prepares medicines, checks transactions or accompanies journeys and why this arrangement changed. Conversely, receiving help does not prove cognitive incapacity: poor vision, physical disability, limited literacy, finances or household roles may explain it. A person living alone may complete tasks with difficulty because help is unavailable. The assessment should therefore distinguish task performance, underlying capacity and available support, using concrete observations where possible and recording uncertainty where these cannot be separated. Separate observation from inference in the record: a pharmacist has observed uncollected prescriptions, for example, but the reason may still be unknown. Comparing several activities and the timing of changes can help determine whether a broader pattern is emerging. Where accounts conflict, further assessment is preferable to choosing the most alarming account as automatically correct. Recognise New Apathy or Sustained Behavioural Change Behavioural change can bring cognitive decline to clinical attention. In a population‐based study of 319 participants with MCI and 1590 with normal cognition, non‐psychotic neuropsychiatric symptoms affected approximately half and one quarter, respectively; apathy was associated with prevalent MCI (odds ratio 4.53, 95% confidence interval 3.11–6.60) [10]. This cross‐sectional estimate is neither diagnostic accuracy nor future AD risk. A meta‐analysis also associated apathy with subsequent cognitive impairment, although definitions and populations varied [11]. Reduced motivation, interest or emotional engagement warrants assessment without presuming neurodegeneration. Mild behavioural impairment (MBI) describes acquired behavioural or personality change beginning at age 50 years or later, persisting at least intermittently for 6 months and affecting interpersonal or everyday functioning in the absence of established dementia [12]. Symptoms should not be better explained by a psychiatric disorder, another medical condition or substances. The five domains cover decreased motivation, affective dysregulation, impulse dyscontrol, social inappropriateness and abnormal perception or thought content [13]. MBI provides a framework for characterising change; clinicians should not wait 6 months before assessing concerning symptoms. Establish whether the behaviour is new, a recurrence of a familiar psychiatric syndrome or a response to bereavement, pain or altered circumstances. Ask what the person has stopped doing and whether prompting helps. Changes in sociability or suspiciousness should be explored with the same attention to context. Biomarker findings depend on the population studied. Lussier et al. found an association with amyloid positron emission tomography (PET) without significant tau‐PET or grey‐matter‐volume associations in 96 cognitively unimpaired participants [14]. Johansson et al. reported associations with tau‐PET in the entorhinal cortex/hippocampus in 50 amyloid‐positive cognitively unimpaired participants [15]. Among 442 dementia‐free participants with normal cognition or MCI, Naude et al. found positive MBI‐tau associations in Braak I and III regions in the amyloid‐positive stratum; in the amyloid‐negative stratum, Braak I was non‐significant and Braak III showed an inverse association [16]. Thus, the amyloid‐negative findings were not uniformly null. A 2025 systematic review included 27 studies, of which 22 were cross‐sectional and five longitudinal; its five quantitative syntheses each contained only two studies [17]. Positive pooled amyloid‐ and tau‐PET associations therefore require caution. Differences in ascertainment, tracers and selection limit comparisons. These observations do not support an amyloid‐only rule, but neither do they establish a fixed within‐person sequence from behavioural change to tau accumulation. Clinically, sustained change should prompt cognitive and functional assessment, with the aetiology left open. Use a Cognitive Screen Sensitive to Mild Impairment The Mini‐Mental State Examination (MMSE) can describe global cognitive severity but may miss mild deficits. Systematic reviews favour the Montreal Cognitive Assessment (MoCA) over the MMSE when MCI is suspected because of its greater sensitivity, although lower specificity at commonly used thresholds can increase false‐positive results [18, 19, 20]. This trade‐off varies across populations and cut‐offs. The Japanese version (MoCA‐J) is an option where feasible. Reviews include heterogeneous studies, often from selected clinical samples, so summary accuracy should not be treated as a direct comparison applicable to every outpatient service. In the original Japanese validation, a MoCA‐J cut‐off of 25/26 yielded approximately 93% sensitivity and 87% specificity for MCI [21]. These are study‐specific estimates. Interpret performance alongside education, language, hearing, vision, fatigue and mood and record factors that affected administration. Follow current developer instructions and conditions of use [22]. An abnormal score identifies a need for assessment, not its cause; the probability that it represents disease also depends on the clinical setting. A reassuring score cannot override a credible history of decline. Explain the purpose of testing and avoid framing the result as a pass or fail. If distress, communication difficulties or sensory limitations undermine validity, record the limitation and prioritise the clinical account and appropriate reassessment. Referral may still be appropriate when standard administration is not possible. Mini‐Cog is an optional ultra‐brief assessment when time is limited [23]. Its evidence does not justify using a reassuring result to exclude isolated MCI and the developer also cautions that a higher score does not exclude cognitive impairment [24]. Further assessment may involve MoCA or direct referral according to the clinical concern. Mini‐Cog, MoCA and MMSE should not become a mandatory sequence before referral. Table 2 summarises their roles alongside informant tools. An unavailable, incomplete or poorly tolerated test should be documented rather than converted into a reassuring result. TABLE 2Brief tools for case finding and their limitations.InstrumentReference thresholdRole and limitationMoCA‐J25/26 in the original Japanese validationApproximately 93% sensitivity and 87% specificity for MCI in that study [21]. Performance varies by population and threshold; consider education, language, and sensory or other testing limitations [18, 19, 20].MMSENo single threshold excludes MCIIt can describe global cognitive severity, but a high score cannot exclude mild impairment [18, 19, 20].Mini‐Cog0–2 of 5 indicates a higher likelihood of cognitive impairment [24]Optional ultra‐brief assessment. A score of 3–5 does not exclude impairment, and persistent concern warrants further assessment [23, 24].AD8At least 2 of 8 in the original study [25]Structures an informant account of the decline. The evidence is not specific to MCI due to AD. Assess the informant’s familiarity with the patient [25, 26].Short IQCODENo universal threshold for MCIThresholds and accuracy vary by setting. Requires an informant who can compare present and previous functioning [26]. Note: Interpret reference thresholds alongside clinical history. Use validated language versions and current developer instructions, including conditions of use [22, 24]. A missing informant or incomplete test should be recorded as a limitation, not as a reassuring result.Abbreviations: AD, Alzheimer's disease; AD8, Eight‐item Informant Interview to Differentiate Aging and Dementia; IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly; MCI, mild cognitive impairment; MMSE, Mini‐Mental State Examination; MoCA‐J, Japanese version of the Montreal Cognitive Assessment. Brief tools for case finding and their limitations. Note: Interpret reference thresholds alongside clinical history. Use validated language versions and current developer instructions, including conditions of use [22, 24]. A missing informant or incomplete test should be recorded as a limitation, not as a reassuring result. Abbreviations: AD, Alzheimer's disease; AD8, Eight‐item Informant Interview to Differentiate Aging and Dementia; IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly; MCI, mild cognitive impairment; MMSE, Mini‐Mental State Examination; MoCA‐J, Japanese version of the Montreal Cognitive Assessment. Refer When Clinically Meaningful Concern Remains Consider referral for credible cognitive or IADL decline, a concerning new behavioural syndrome or an abnormal assessment requiring clarification. Judge severity, trajectory, alternative explanations and safety. The four actions are complementary; all need not be positive. Explain what specialist assessment may offer, ask about the patient's priorities and provide decision‐making support when needed. Referral does not commit the person to biomarker testing or antibody treatment. State the referral question: clarification of a changing syndrome, assessment of functional dependence or advice about management and support. This helps the receiving service triage the request according to clinical need. An acute change, fluctuating attention, rapid progression, focal neurological findings or immediate safety concerns require prompt medical assessment through the appropriate acute pathway [27]. Routine referrals should not defer this assessment. For gradual unexplained decline, investigation of contributing conditions can proceed alongside referral. When immediate referral is not indicated, agree on a review plan and specify changes that should prompt earlier reassessment. Avoid a fixed waiting interval that disregards deterioration. Document who will review the patient and how the patient or a trusted contact can raise new concerns. Assessment of alternative or potentially reversible contributors is shared by all referring clinicians. Consider delirium, medication effects, sensory impairment and relevant medical causes, including thyroid disease or vitamin B12 deficiency, according to the presentation [27]. Complete appropriate assessment without treating every possible investigation as a prerequisite for referral. Record results already available and investigations pending, so that the receiving service can avoid unnecessary duplication. Obtain Collateral Information Without Creating an Access Barrier Information from someone who knows the patient can establish changes in cognition, function, mood and behaviour [27]. Record the source, frequency of contact and examples. Disagreement with the patient does not automatically demonstrate impaired insight: observation, expectations and informant distress also influence reports. The AD8 or the short Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) may structure this discussion, although their evidence is not specific to MCI due to AD [25, 26]. The patient's account and goals remain central. Living alone and lacking a reliable informant are distinct. A friend may know someone living alone well, whereas a cohabiting relative may observe little of their daily functioning. A US population‐based study documented functional difficulties and gaps in assistance among people living alone with cognitive impairment [28]. This supports attention to unmet needs without assuming the same prevalence or service arrangements in Japan. With the patient's consent and appropriate privacy safeguards, useful sources may include a trusted contact, primary‐care clinician, pharmacist, care manager, home‐visit nurse, municipal service or community general support centre. Existing records, medication discrepancies, repeated missed appointments and changes in assistance can contribute to the history. Initial intensive support teams may assist eligible people who are not adequately connected to medical or care services or whose behavioural symptoms complicate support [29]. Their remit is targeted assessment and support, not universal reconstruction of premorbid functioning. Ask which people or services the patient trusts, explain how additional observations may help and agree on what information can be shared. Seek information relevant to the clinical question. If it cannot be obtained, record the limitation and continue to consider what can be assessed directly. Living alone should not itself lead to assumptions about decision‐making ability. These links can improve the available account, but they cannot guarantee a reliable baseline. Staff may have only recent contact, records may describe isolated encounters and the person may have no established service connection. Asking a municipal service or support team to supply a history does not transfer the clinician's responsibility for assessment. Document what is known, what remains uncertain and how each source was interpreted. Do not infer absence of decline from absence of corroboration. When cognitive or functional decline is suspected, referral may be appropriate even without a reliable informant; incomplete information should not delay clinically indicated assessment. The dementia disease medical centre should help coordinate information gathering while clinical evaluation proceeds, drawing on its regional links [30, 31]. Repeated consultations and, where available, direct observations of everyday functioning may reduce uncertainty, but cannot always recover previous abilities. Such observations complement the history rather than automatically replacing it. If uncertainty persists, communicate it, address immediate support needs and agree on reassessment. Recognise Decline Despite Competing Explanations Depression, physical illness and medication effects can complicate the recognition of cognitive decline. The aim is to establish whether the change remains unexplained and warrants further assessment. Psychiatric expertise can help interpret behavioural symptoms and treatment response, but medication review and longitudinal observation are shared clinical responsibilities. Assess Cognition Alongside Mood and Motivation Depression can mimic, accompany or precede cognitive decline. Sadness, guilt, hopelessness and distress support a depressive syndrome, although their absence does not exclude depression and apathy may coexist [32]. Ask whether reduced activity reflects lost motivation, difficulty organising an activity, fear of failure, physical illness or lack of opportunity. The practical question is whether cognition and everyday functioning return towards the person's baseline as mood improves. Persistent decline requires assessment even when depressive symptoms respond to treatment. When the initial history is concerning, cognitive assessment can proceed alongside treatment of depression [27]. Vascular disease can further complicate interpretation, particularly with gait changes or executive slowing. When imaging is already available, white‐matter hyperintensities should be interpreted in the clinical context. Mixed vascular and AD pathology is common [33]. A vascular contribution therefore need not explain all symptoms or close the assessment of persistent decline. Table 3 links common presentations to questions that help decide whether further cognitive evaluation is needed. Treatment of contributing conditions and reassessment should proceed together. TABLE 3Clinical situations that can obscure cognitive decline.Clinical findingWhat to establishImplication for assessmentReduced initiative or interestAssess mood, motivation, ability to organise activities, physical limitations, and change from baseline [32].Assess cognition and function when the change remains unexplained.Sadness, guilt or prominent distressEstablish depressive syndrome and its timing relative to cognitive and functional changes [27, 32].Cognition and function should be reassessed if decline persists despite improvement in mood. Assessment can be performed alongside treatment.WMH with executive slowing or gait changeVascular contributors should be considered in the overall clinical and imaging context [33].Vascular contribution does not need to explain all symptoms. Reassess persistent decline; mixed pathology is common [33].Symptoms during sedative or anticholinergic treatmentReview indication, actual use, dose changes, and symptom timing; plan medication changes individually [34, 35, 36].Symptoms should be reviewed after individualised medication changes, as cognitive improvement remains uncertain.Persistent change across visitsCompare function and behaviour with the previous baseline; reconsider established diagnoses.New or progressive functional changes should prompt reconsideration. A small score change alone is insufficient. Note: These findings can coexist. The purpose is to recognise change requiring further assessment, not to assign a single cause from a symptom or imaging finding. Treat contributing conditions while investigating unexplained decline.Abbreviations: AD, Alzheimer's disease; WMH, white‐matter hyperintensities. Clinical situations that can obscure cognitive decline. Note: These findings can coexist. The purpose is to recognise change requiring further assessment, not to assign a single cause from a symptom or imaging finding. Treat contributing conditions while investigating unexplained decline. Abbreviations: AD, Alzheimer's disease; WMH, white‐matter hyperintensities. Review Medicines That May Contribute Sedative‐hypnotics and medicines with anticholinergic properties can affect attention, memory and everyday function. Observational associations between cumulative anticholinergic exposure and dementia do not establish causation [34]. Confounding by indication and treatment of early, unrecognised symptoms complicates interpretation. Review prescription and non‐prescription medicines, recent changes, actual use and timing relative to symptoms. Balance current benefits against adverse effects and withdrawal risk; a medication list alone cannot establish drug‐induced cognitive impairment. Structured interventions can reduce benzodiazepine receptor agonist use in older adults [35], but this does not demonstrate prevention of dementia. Evidence that anticholinergic deprescribing improves cognition remains of very low certainty [36]. Agree an individualised plan, including gradual tapering when appropriate and monitor sleep, mood and function. Medication adjustment can clarify potentially reversible contributions while investigation of progressive decline continues. Persistent deterioration should not be attributed indefinitely to psychotropic exposure, nor should patients be promised reversal of neurodegeneration through withdrawal. Use Follow‐Up to Clarify Persistent Change Repeated consultations offer a view of a person's conversational style, initiative, appointment behaviour and ability to organise treatment. New reliance on others, repetitive questioning or declining function despite improved mood may become clearer over time. Record concrete changes and, when useful, repeat assessments under comparable conditions. Small score changes alone may reflect measurement variability or practice effects. Continuity also creates a risk of attributing a new syndrome to a longstanding psychiatric diagnosis. When the pattern changes, reconsider the formulation and the need for specialist assessment. Comparison with earlier episodes can be particularly useful: has recovery restored the person's usual organisation and initiative or has a new functional deficit persisted? Patients may also describe compensatory effort that was not apparent to others. These observations can refine the referral question while avoiding the assumption that either longstanding psychiatric illness or late onset necessarily determines the aetiology. Follow‐up should clarify uncertainty without postponing referral when concerning decline is already evident. Coordinate Specialist Assessment and Continuing Care Japanese dementia disease medical centres have a regional coordination role linking differential diagnosis, medical consultation and support [30, 31]. The proposed pathway builds on this role: the centre should act as a hub connecting primary care, municipal services, community general support centres, initial intensive support teams and long‐term‐care providers, according to individual needs. This means agreeing responsibilities and helping people reach assessment and support. This does not imply that every centre can provide every service or resolve all informational gaps. Identify the clinical question, the gaps that materially affect interpretation and the service best placed to help. Bring available observations together and return a usable plan to the patient and referring clinician. Repeated requests for missing information should not leave responsibility for assessment unclear. A useful referral describes the trajectory, concrete functional changes, cognitive assessment, psychiatric and neurological features, medication and relevant medical history. Include the patient's priorities, existing support, available collateral sources and unresolved gaps. The receiving service should communicate its working diagnosis, remaining uncertainty and proposed care plan, with responsibilities agreed between services. Continuing care includes comorbidity, medication, mood, daily function and access to practical support, irrespective of antibody treatment. Where appropriate, the referring clinician retains these familiar elements of care, while the centre coordinates specialist input. The patient should know which service to contact for deterioration or a support problem. A named contact and explicit follow‐up responsibilities are proposed operational safeguards, not demonstrated outcomes of this review. Figure 1 shows clinical assessment and information gathering proceeding alongside coordinated support. Proposed pathway for case‐finding, referral, and continuing support. Functional and behavioural change, cognitive assessment, and available collateral information inform the decision to refer. The elements are complementary and do not form a mandatory test sequence. Acute presentations require prompt medical assessment. Dementia disease medical centres coordinate with referring clinicians and municipal and long‐term care services. Clinical assessment and further information gathering proceed in parallel. Links represent proposed coordination according to need and local capacity, with consent and privacy safeguards. IADL, instrumental activities of daily living; MoCA, Montreal Cognitive Assessment. Treatment Context for Referral As of September 2026, Japanese optimal‐use guidance limits anti‐amyloid treatment to MCI due to AD or mild AD dementia and requires amyloid confirmation by PET or cerebrospinal fluid testing before treatment [3, 4]. Referring clinicians do not need to complete these tests or determine eligibility. Specialists assess the individual benefit–risk balance and preferences; diagnostic clarification and support remain valuable when treatment is not chosen. Apolipoprotein E (APOE) ε4 status informs the assessment of amyloid‐related imaging abnormalities (ARIA) risk, counselling and specialist safety monitoring, rather than serving as an independent Japanese eligibility criterion [3, 4]. Treatment‐related genotyping is recommended after clinical and amyloid confirmation of AD [37]. Since 1 July 2026, it is reimbursed under specified patient and facility conditions [38]. Predictive testing of asymptomatic people is a separate context, is not covered by this provision and is not recommended routinely [37]. APOE testing is neither a general screening nor a prerequisite for referral. The clinician arranging APOE testing should explain its implications for the patient and biological relatives before consent and when disclosing results, with access to genetic counselling, including external referral where needed [37]. Limitations This selective narrative review is subject to selection bias and does not provide a formally graded evidence base. Diagnostic‐accuracy studies often use selected populations. The principal MBI‐PET comparisons were cross‐sectional studies of North American or European research cohorts [14, 15, 16] and the broader literature includes longitudinal studies with uncertain individual prognostic value [17]. Applicability to routine Japanese care requires evaluation, including differences in referral populations, language and available collateral information. The proposed referral pathway has not been prospectively validated. Japanese service policy supports the centre's coordinating role; however, the effects of the proposed arrangements on access, diagnostic accuracy, workload and patient experience remain unknown. Regional capacity and information‐sharing arrangements will constrain implementation and a reliable baseline may remain unavailable. Evaluation should include people living alone and those without informants, particularly whether clinically indicated assessment proceeds without avoidable delay. Conclusion Recognise change in previously mastered activities, new apathy or sustained behavioural change, assess cognition with an appropriate brief tool and refer when meaningful concern remains. Dementia disease medical centres should link diagnostic assessment with regional support. Collateral history strengthens this work, but its absence should not delay a clinically indicated assessment. Coordinated care must remain attentive to the person's priorities, continuing needs and unresolved uncertainty. Funding The author has nothing to report. Ethics Statement This narrative review used published literature and involved no new human‐participant data; ethics committee approval was not required. Conflicts of Interest Kengo Shimoda has received lecture and/or consulting fees from Eisai Co. Ltd. and EP Pharma. Acknowledgements The author thanks the reviewers for their constructive comments. The author is also grateful to the staff of the Dementia Disease Medical Center, Nippon Medical School Chiba Hokusoh Hospital and in particular to the medical liaison office, which collated questions about referral received from partner clinics and from the Elderly Welfare Division of Inzai City; to Dr. Mineo Yamazaki, former Director of the Dementia Disease Medical Center, for information on the insurance coverage of apolipoprotein E ε4 testing; and to Professor Amane Tateno, Department of Psychiatry, Nippon Medical School, for advice on identifying original studies relevant to this review. Eisai Co. Ltd. and EP Pharma had no role in the design, content, writing or revision of this review or in the decision to submit it for publication. All individuals named have given permission to be acknowledged. ChatGPT (OpenAI) assisted with drafting in British English and the preparation of the tables and figures. Literature searches and source selection were performed by the author without the use of artificial intelligence tools. The author determined the structure, selected and critically reviewed the material, revised and approved the final manuscript and takes responsibility for its content and the accuracy of its references. Data Availability Statement Data sharing is not applicable because no new research data were created or analysed. References Associated Data Data Availability Statement Data sharing is not applicable because no new research data were created or analysed.