The 4AT’s role in screening for cognitive impairment and dementia

Why cognitive impairment and delirium belong in the same first-pass assessment

Delirium and dementia are different conditions, but they often overlap in acute hospital care. Delirium is an acute and fluctuating disturbance, while dementia is usually a chronic progressive disorder. A person may have either condition or both at the same time.

This creates a practical problem. A brief cognitive error may reflect longstanding impairment, an acute delirium, the effects of illness or medicines, a communication difficulty, or more than one of these. A separate “delirium workflow” and “dementia workflow” can duplicate questions without resolving that uncertainty.

The 4AT offers a useful starting point because it combines:

• alertness, looking for altered level of arousal;
• AMT4, a short assessment of orientation;
• Months of the Year Backwards, a brief test of attention; and
• acute change or fluctuation, which brings in information about change from the person’s usual state.

The cognitive items detect current difficulty. Together with the alertness and acute-change items, they make the 4AT a practical way to detect delirium while also flagging possible cognitive impairment. A dementia diagnosis, however, requires a wider assessment of persistent change, function and other possible causes.

The assessment usually takes around 2 minutes, requires no special formal training, and is free for clinical use. This makes it realistic for high-volume hospital pathways. Implementation still benefits from clear local guidance, suitable electronic recording and audit.

What the 4AT assesses

Is the 4AT a cognitive test?

Yes, in part. AMT4 and Months of the Year Backwards are brief cognitive tests of orientation and attention. However, the 4AT is more than a cognitive test because it also assesses alertness and acute change or fluctuation. Its main validated purpose is rapid detection of delirium.

The cognitive information is nevertheless clinically useful. An abnormal result can flag possible impairment that was not previously documented and prompt collateral history, fuller assessment and appropriate adjustments to communication.

Can the 4AT distinguish delirium from dementia?

Not by score alone. Acute onset, fluctuation, altered alertness and inattention make delirium more likely, but people with dementia may also score on the cognitive items, and delirium commonly occurs on top of dementia.

The distinction depends on the whole clinical picture, including:

• the person’s usual cognition and function;
• the time course of the change;
• information from family, carers or records;
• illness severity, medicines and other possible causes;
• repeated clinical observation where needed; and
• appropriate follow-up after the acute episode.

What the score means for cognition

On smaller screens, swipe sideways to see all three columns.

See the official 4AT user guide for administration, scoring and interpretation.

What the dementia evidence shows

Penfold et al.: a large routine-care study

The largest published study to date examining the association between routine 4AT scores and recorded dementia included 75,221 emergency medical admissions of people aged 65 or over across three hospitals in NHS Lothian. An admission 4AT was available for 62,188 admissions.

Among 11,145 admissions with dementia recorded in linked health records, 86.8% had a 4AT score of at least 1. At this threshold, sensitivity for a recorded dementia diagnosis was 0.87 and specificity was 0.71. The positive predictive value was 0.39 and the negative predictive value was 0.96.

These figures support a careful conclusion: a 4AT score of at least 1 was associated with recorded dementia in this population and may help identify people who need assessment for possible unrecognised dementia, but it is not a dementia diagnosis. The positive predictive value of 0.39 means that most admissions with 4AT ≥1 did not have dementia recorded in the linked data. This may reflect delirium, acute illness or other causes of cognitive difficulty and, in some cases, unrecorded dementia.

The reference standard was dementia recorded in primary care, hospital or prescribing data, rather than prospective specialist assessment of every patient. Some dementia will have been unrecorded, 17% of admissions had no 4AT, and the cohort came from one predominantly White Scottish region. Predictive values will also vary between settings.

The paper was published in 2025. A 2026 correction changed the study dates to 1 April 2016 to 1 April 2020; it did not change the results.

Read Penfold et al. in Age and Ageing · Read the published correction

Tan et al.: a staged US hospital pathway

Tan and colleagues reported implementation across eight units in a 990-bed academic hospital in Los Angeles, covering 11,180 hospitalisations. The pathway first used electronic health record information to identify an existing dementia diagnosis or dementia medication. Patients without that flag were eligible for nurse-administered 4AT, followed by the AD8 when the 4AT score was above 0. The AD8 was completed with the patient or a family member; it was not completed when the patient was considered unreliable and no reliable collateral informant was available.

Of 9,168 eligible hospitalisations, 7,637, or 83.3%, received a 4AT. Among the 1,718 with a score above 0, an AD8 was completed in 1,488. Of these, 480 admissions met the study’s screen-defined “potential dementia” category and 1,008 met its “newly detected CIND” category (4AT >0 and AD8 <2). Neither category confirmed the presence or absence of dementia. Around 55% of both groups were also in the possible-delirium range, underlining how often acute and chronic cognitive concerns overlap.

This single-centre study demonstrated implementation of a staged pathway across eight units in one US academic hospital. It does not establish diagnostic accuracy, improved outcomes, or that the 4AT should be used alone. In this pathway, the 4AT was used to select patients for an additional AD8; the study did not determine who ultimately required diagnostic follow-up.

Read Tan et al. in the Journal of the American Geriatrics Society

Supporting evidence

• A prospective emergency-department study of 419 older people found that a 4AT threshold of 2 or more had sensitivity 0.74 and specificity 0.87 against expert dementia assessment. This supports further evaluation of cognitive case-finding, not diagnosis from the 4AT alone. O’Sullivan et al., 2018
• In a national hip-fracture cohort, postoperative 4AT scores of 1–3 were associated with lower odds of returning home and regaining outdoor mobility. This suggests that even sub-delirium-threshold cognitive findings can identify clinical vulnerability, although the observational study does not show that the score caused the outcomes. Hawley et al., 2023
• Months of the Year Backwards has shown useful accuracy as a brief bedside attention test for delirium. Failure is not disease-specific and may occur for several reasons. O’Regan et al., 2014
• A 2026 analysis of two prospective cohorts found broadly consistent diagnostic accuracy for delirium detection for an operationalised 4AT across different levels of baseline cognition. The 4AT items were reconstructed from study assessments rather than administered as the standard clinical 4AT, which limits the conclusion. Hogan et al., 2026

4Ms Mentation and the CMS measure

IHI Age-Friendly Health Systems

The Institute for Healthcare Improvement describes the 4Ms as What Matters, Medication, Mentation and Mobility. Mentation includes preventing, identifying, treating and managing dementia, depression and delirium.

The 4AT can contribute to the delirium and cognitive-impairment parts of Mentation. It is not a depression screen and is not a complete Mentation pathway. IHI’s 2026 assessment resource includes the 4AT among delirium tools and describes its combined delirium and cognitive-impairment assessment role.

Read IHI’s introduction to Age-Friendly Health Systems · Read IHI’s assessment tools guide

CMS Age-Friendly Hospital measure

The CMS Age-Friendly Hospital measure is a pay-for-reporting structural measure in the Hospital Inpatient Quality Reporting Program. It does not mandate or endorse a named cognitive instrument.

Domain 3 is called Frailty Assessment and Intervention. It asks hospitals to use validated instruments to screen for risks concerning mentation, mobility and malnutrition, and it also covers management plans, discharge communication, outcome data and emergency-department protocols.

The 4AT can therefore be:

• a well-validated, scalable tool for detecting delirium that also provides a signal of possible cognitive impairment;
• a way to create a common starting point for nurses, doctors and the wider team;
• a trigger for fuller assessment, management and discharge communication; and
• a structured result that can support implementation and audit.

It does not, on its own:

• establish a dementia diagnosis or distinguish all causes of cognitive impairment;
• cover depression, mobility or malnutrition assessment;
• replace management, communication or follow-up;
• satisfy Domain 3 on its own; or
• be described as “CMS approved” or required by CMS.

As with any clinical assessment tool, the 4AT informs clinical decision-making and is interpreted alongside the person’s history, examination and other relevant information.

A precise policy description: The 4AT can form part of the validated mentation-assessment element of a wider Domain 3 pathway. A single initial cognitive workflow may reduce duplication, but the hospital still needs the other assessment, action, data and protocol elements.

Policy status reviewed 28 July 2026. The current performance period runs from 1 January to 31 December 2026, with submission scheduled for 1 April to 17 May 2027. Hospitals should check current CMS Hospital IQR guidance before making reporting decisions.

Read the CMS final rule · Read the current FY 2028 Hospital IQR Program Guide · Read current dates and deadlines

A practical hospital pathway

The 4AT is most useful when the response to each result is agreed in advance.

1. Define who should be assessed and when. Acute medical, emergency, perioperative and other non-ICU pathways may use different triggers. An ICU-specific delirium tool may be more suitable in critical care.
2. Make the assessment workable. Provide glasses and hearing aids, use suitable language support, explain the purpose, and record why an item or the whole assessment could not be completed.
3. Record enough context. Keep the total and item scores, date and time, assessor, clinical setting, relevant communication or sensory factors, and evidence about acute change.
4. Respond promptly to possible delirium. Assess for causes and start the local delirium management pathway. Do not wait for dementia assessment before treating an acute problem.
5. Follow up the cognitive signal. Establish usual cognition and function, seek collateral history, review medicines and reversible contributors, and decide whether reassessment after recovery or referral is needed.
6. Communicate across transitions. Include relevant findings, actions and unresolved concerns in discharge information and communication with primary care or the next care setting.
7. Audit the pathway, not just completion. Measure who receives the assessment, whether positive results lead to action, and whether follow-up and communication occur. Review differences between patient groups and reasons for non-completion.

One assessment may reduce duplication, but it is not a shortcut

The practical benefit is not that the 4AT replaces every cognitive or dementia assessment. It is that the same brief assessment can alert the team to possible delirium and possible cognitive impairment at the start of care.

That can make the next decision clearer:

• possible acute disturbance: assess and manage delirium now;
• possible cognitive impairment: establish baseline and plan follow-up;
• both signals present: address delirium while recognising likely underlying vulnerability; or
• continuing concern despite a low score: use clinical judgement and assess further.

The value comes from what the team does next, not from recording a score alone.

Communication, language and sensory impairment

Cognitive assessment is affected by hearing, vision, language, speech, literacy, education, culture, fatigue and acute illness. These factors should be considered before interpreting any brief result.

Use the person’s usual aids and an appropriate interpreter where needed. Record the circumstances of testing and avoid treating non-completion as a dementia diagnosis. Services should audit completion and follow-up across demographic and social groups rather than assuming that one standard process works equally well for everyone.

Clinical commentary

The cognitive contribution of the 4AT has often been under-described. Penfold et al. show that abnormal admission scores are associated with recorded dementia and may help identify people who warrant assessment for possible unrecognised dementia, while Tan et al. show how the result can sit within a staged US hospital pathway.

The practical conclusion is also the useful one. The 4AT gives the team a rapid, structured and reliable way to detect delirium while flagging cognitive concerns at a point when both are easily missed. Clinical interpretation, collateral history and follow-up remain important where indicated, as they do for every assessment process. For hospitals trying to make age-friendly care workable, the 4AT is a strong first step in a wider pathway.

Frequently asked questions

Is the 4AT a cognitive assessment?

Yes, in part. Two items assess orientation and attention. The full 4AT also assesses alertness and acute change or fluctuation, and its primary validated purpose is rapid detection of delirium.

Does the 4AT diagnose dementia?

No. An abnormal score can flag possible cognitive impairment and the need for fuller assessment. Dementia diagnosis requires evidence of a persistent decline from previous ability, effects on everyday function, clinical assessment, and consideration of delirium and other causes.

Does CMS require hospitals to use the 4AT?

No. In Domain 3, hospitals attest whether they use validated instruments for mentation, mobility and malnutrition. CMS does not specify the 4AT or any other named cognitive tool.

Can the 4AT alone satisfy Domain 3?

No. It can support the cognitive impairment and delirium part of mentation assessment. Domain 3 also covers mobility, malnutrition, management and discharge plans, outcome data and emergency-department protocols.

What should happen after a positive cognitive result?

The team should check for delirium and other acute causes, establish the person’s usual cognition and function, obtain collateral information, address communication and sensory factors, and arrange reassessment, referral or follow-up where appropriate.

Can the 4AT be used when a person already has dementia?

Yes. Delirium commonly occurs in people with dementia. In this situation, change from the person’s usual state, altered alertness, fluctuation and collateral history are particularly important. The result still requires clinical interpretation.

Does a score of 0 rule out dementia or delirium?

No. It makes delirium or severe cognitive impairment less likely at that time, but it does not exclude mild impairment, dementia, an evolving delirium or another clinically important problem.

Related 4AT resources

• How to use the 4AT
• The 4AT Manual
• The 4AT evidence base
• Evidence for the AMT4 component
• Evidence for Months of the Year Backwards
• Download the 4AT

References

---

This page explains evidence and implementation options. It is not a substitute for local clinical policy, professional judgement or current CMS reporting guidance.