Screening for Cognitive Impairment in Primary Care: Rationale and Tools

Abstract

Cognitive impairment is common and often under diagnosed in the early stages. Patients and family caregivers benefit from early diagnosis of reversible causes and longer lead time for care planning in primary dementia diagnoses. Primary care physicians are the first and best providers for diagnosing common, serious, and progressive cognitive disorders.

Introduction

Dementia is a cluster of age-associated neurodegenerative diseases that affect increasing numbers of families and that challenge the social and financial resources of the American health system. The global prevalence of cognitive impairment in older adults may be as high as 9%, affecting 50 million people. Some predictions are that this number will triple by 2050.¹ In the advanced economies such as Japan, the US, and the EU, countries are experiencing the impact of an estimated 10 million new cases each year. China’s mandated strict population control has led to economic effects of demographic aging as well. For the first time in 2020, there were more people in the world over age 60 than children under age 5. In addition, the number of very old people increases every decade. Japan has the oldest population we have ever seen resulting in fewer working age people to financially support and care for the increasing numbers of very old people suffering from the syndromes of aging including dementia, frailty, and functional dependence.²

Primary care practitioners who provide health promotion, disease prevention, and care of acute and chronic conditions are on the leading edge of this demographic shift. Professional bodies such as the American Geriatrics Society (AGS), governmental agencies such as the United States Preventive Health Task Force (USPHTF), and others routinely review the published research literature and convene experts to assess preventive practice with age-specific guidelines for example for vaccination and periodic screening for common illnesses. Screening is a useful population strategy when conditions are common and treatable even if not curable when detected at an early, asymptomatic stage. The calculation is that early detection is cost-effective if there are enough cases that can be cured, such as early-stage breast cancer, or managed effectively to reduce long term morbidity, such as cardiovascular disease and diabetes.

The guidelines over the years have tailored screening to estimated individual risk. Better understanding of the genetics of breast, colon, and prostate cancer identifies high risk individuals who would benefit from screening at younger ages, in their 40s, compared to average risk individuals who may begin screening in their 50s. This strategy conserves individual and societal costs. Similarly, considering lag-lead time to benefit, cervical cancer screening is no longer recommended for women over age 65 who have had consistent well-woman care. But the genetics and risk factors for dementing diseases are less well understood. Effective medical and surgical treatment for cardiovascular disease and many malignancies are readily available. This is not yet true for neurodegenerative diseases resulting in loss of cognitive function. For these reasons the USPHTF has been equivocal in its recommendations for population screening for dementia.³ The British National Institute for Health and Care Excellence (NICE) guidance (NG97: June 20, 2018)⁴ also advises against routine dementia case-finding in asymptomatic patients due to the lack of demonstrably effective treatment. The recent events surrounding FDA approval of new Alzheimer disease drugs is one example.⁵

Why Should We Screen for Dementia?

The cost of caring for people with dementia is substantial. An analysis of the Health and Retirement Study (HRS) longitudinal cohort of Medicare beneficiaries aged 70 and older reported that the cost of five years of care for an Alzheimer’s disease patient was over $100,000 greater than that of a heart or cancer patient.⁶ Although Medicare paid the same amount for covered costs regardless of diagnosis, the difference is accounted for by out-of-pocket costs for paid caregiving and loss of employment income by family caregivers.⁶ About 10% of today’s generation of 70–80 year-olds already suffer with dementia. The prevalence is estimated to increase by 10% in each successive cohort so that about 20% of 80–90 year-olds and a third of the oldest ages over 90 require supervision due to cognitive decline. The family caregivers of the old are themselves older. With no imminent change in how care for “memory” patients will be financed, the bulk of the care is provided by lower income, older women caregivers.^6,7,8

In the absence of clearly effective therapeutics, the USPHTF did not change the 2014 guideline in 2020.³ However, professional and advocacy organizations including the Gerontological Society of America Workgroup on Cognitive Impairment and Earlier Diagnosis advocate for early diagnosis of dementia to allow patients and families time for financial planning and mobilization of support systems.⁹ In 2018 the American Diabetes Association recommended annual screening for adults over age 65 at initial and annual clinic visits for early detection of MCI (mild cognitive impairment not meeting DSM 5 criteria for a diagnosis of dementia), and dementia to improve diabetes outcomes.¹⁰ A survey of patients and caregivers sampled their opinions about dementia screening. There was agreement that early diagnosis in addition to planning time, would allow patients to participate in planning for their own futures while they still had capacity. Family members also believed that having planning time would help them learn about the disease and make housing decisions.^7,11 Patients and family members report being more comfortable discussing cognitive impairment with a familiar PCP.^12,13

A delayed diagnosis of dementia encourages the thought that memory loss is normal aging until it results in crises that challenge relationships, mistaken attributions of mental illness, wandering, traffic accidents, and police encounters. Financial exploitation of cognitively-impaired elderly is a real risk. Often there were signs that might have been evident to an expert examiner. Several screening tools we will discuss are available to primary care providers and have been calibrated to detect early-stage cognitive impairment MCI.^14,15 Similarly to a slightly elevated HgA1c indicating glucose intolerance not meeting criteria for diabetes, MCI informs a surveillance strategy. About 15% of MCI diagnosed patients will progress to dementia in a year. The finding of MCI should prompt patients and clinicians to engage in evidence-based activities to preserve cognitive performance. These include physical exercise, cognitively engaging activities such as music and chess, social engagement, and cardiovascular risk modification.¹⁶

Primary care providers have a distinct advantage over episodic providers in that they have a relationship with the patients. Visiting out-of-town family members may be shocked at changes that were not evident on the weekly phone call. At a routine visit a family member may raise concerns about getting lost while driving, repetitive questions or stories and “apathy.” Patients themselves may complain about word-finding problems, forgetting names, and losing their keys. Sometimes there is an “unmasking” in which an older patient becomes agitated or confused in response to an acute illness.¹⁴

What Are the Barriers to Cognitive Screening in Primary Care?

The main barriers to dementia screening in primary care are the insidious onset of cognitive impairment, the tendency for us to politely “normalize” it in old patients and the lack of a clear mandate to screen. This is coupled with the rapid pace of primary care visits that are already crowded with routine monitoring and documentation demands. A third barrier is lack of training and familiarity with screening tools and referral resources.^12,13,14

In 2011 the Centers for Medicare and Medicaid (CMS) recommended physicians use the Medicare Initial and Annual Wellness Visits (MAWV) to assess cognitive function by direct observation, history gathering including, with patient consent, from family members if present, chart review for concerns expressed by other clinicians and with consent, a brief, validated assessment tool.¹⁷ The Welcome to Medicare Initial Visit, Medicare Annual Wellness Visit, or Medicare Annual Exam can be conducted face-to-face or by telehealth as convenient, fully reimbursed and contextually easily understood platforms for cognitive screening.^17,18 For residents of long-term care and skilled nursing facilities, CMS mandates quarterly assessments using the Minimum Data Set 3.0 which includes cognitive screening of verbal patients. The Veterans’ Affairs Health System¹⁶ guideline advised screening only in the presence of warning signs such as inattention, failure to follow up on medication instructions, inability to provide information, family reports of confusion, wandering, deficits in self-care, nutrition, and safety.¹⁷ The AGS, however, recommends routine screening at all Medicare mandated visits to establish a documented baseline for other clinicians. Insurance companies with Part B and Part C plans have Annual Wellness home visits during which they perform a routine screen.

The utility of a screening test depends on its diagnostic accuracy as measured by sensitivity, which has few false negatives, and specificity, few false positives. Greater sensitivity is useful for surveillance and longitudinal follow up of subjective symptoms and MCI. Sensitivity is influenced by the setting, for example delirium during acute illness, environmental distractions in an ED, sensory deficits of hearing and vision and importantly, the lack of a common language between the patient and the examiner. All these may result in false positive cognitive screening. When embedded in a more comprehensive functional assessment such as those contained in MAWV templates, the finding can guide further decisions for evaluation. There are more than 30 cognitive screening tools in use in the US and internationally. We do not intend to present a comprehensive review, but we will compare several widely used English language tools that are available to primary care providers, and which can be easily uploaded to EHR pull down menus or dot phrases. Value-based EHRs have already incorporated such tools especially in the Medicare Annual Visit templates.

Cognition and thinking involves several anatomically distinct yet integrated neural pathways. Neurogenerative diseases may initially affect specific circuits. A screening tool must be able to sample different cognitive functions.¹⁹ Although specific diagnosis is desirable, it is sometimes more important in primary care to establish a baseline for referral or surveillance and establish what the patient can or cannot safely do. The tools we discuss probe neural pathways involved in several cognitive domains because we believe that reliance on only one short-term memory can miss important functional deficits.

Antegrade memory loss (STM) is rapid forgetting of new information and emerges early in Alzheimer’s disease. Visual spatial dysfunction may begin as a new difficulty with procedural tasks such as cooking, filling out forms or way finding in familiar areas emerge later in Alzheimer’s disease. Visual spatial difficulty appears early in diseases such as posterior cortical atrophy and dementia with Lewy bodies (LBD). Visual hallucinations emerge early in LBD but are rare in mild-moderate Alzheimer’s disease. Apathy and social withdrawal may represent depression which should be treated before a diagnosis of Alzheimer’s disease is made. Personality changes and social or emotional disinhibition emerge early in frontotemporal dementia (FTD) before there is evidence of STM loss. Executive dysfunction, impaired judgement, difficulty with multitasking or complex tasks that were easily done in the past occurs in multiple diagnostic groups. Vascular dementias, including post-stroke, multi-infarct, chronic small vessel disease or microangiopathy have variable presentation depending on the brain region affected. Classically, multi-infarct dementia develops as stepwise, abrupt changes compared to insidious decline seen in Alzheimer’s disease. Cardiovascular risk factors are present.

Less common syndromes affect isolated abilities. Prosopagnosia affects facial recognition; primary progressive aphasia affects language function without a history of alcohol abuse. Normal pressure hydrocephalus, Parkinson’s disease, multiple sclerosis, and supranuclear palsy (SNP) have characteristic early onset physical and radiological findings. Prion diseases, e.g., Creutzfeldt Jakob (CJD), follow a rapid course. A screening test might detect cognitive impairment but for these uncommon diseases the specific diagnosis should be delegated to a neurological specialist or research center.

Screening Tools Compared

Table 1 compares several commonly used brief (≤5 items) or short (≤15 items) screening tools. Table 1 is not exhaustive, but it presents comparative characteristics of the screening tools most used in primary care. The most well-known tool is the Folstein Mini-Mental Status Examination (MMSE).²⁰ It is an 11-item 30-point questionnaire introduced in 1975. It takes 8–10 minutes to administer to an alert, cooperative patient with intact hearing, visual, and motor functions. It is widely used in clinical practice and research. However, the MMSE is now copyrighted so clinicians and researchers have had to seek alternative tools.^21,22

Table 1.

Summary Characteristics of Three Dementia Screening Tools and Short Forms

Tool (date of publication)	MMSE20 (1976)	MoCA25 (2005}	SLUMS22 (2006)	Mini Cog46 (2003)	MiniMoCA30 (2015)	RCS (2015)26
Minutes to administer	8–10	10–12	8–10	3	5	3
Public domain	No	No (current)	Y	Y	partially	Y
Sensitivity to MCI%	18	90	92	NA	NA	NA
Sensitivity to Dementia%	90	100	100	76–99	NA	89
Specificity for dementia%	100	87	81	89–93	NA	94 (70 MCI)
Adapted for visually impaired?	Y	Alternate form available	Y	Y	Y	Y
Adaptation for hearing impairment?	Y	Alternate form available	Y	Y	Y	Y
Influenced by education	Y	Y	Y	N	NA	NA
Scoring adjusted for education	N	N	Y	N	N	N
Score ranges for severity (WNL MCI, mild, moderate, severe)	26–30 normal 19–25 mild 10–20 mod. ≤ 9 advanced	26–30 normal 19–25 MCI 11–21 mild, ≤ 10	<HS 25–30 nl >HS 27–30 nl 24–26 MCI ≤ 23 mild+	5 nl 3–4 possible 0–1 likely	24 normal Not scored for MCI	10 normal 5–7 mild <5 dementia
Adaptation for telehealth?	N	Y	N	N	N	N

In response to that, the Saint Louis University Mental Status Exam (SLUMS) was developed in 2006 and validated by Divisions of Geriatric Medicine at Saint Louis University and the John Cochrane Veterans Administration Medical Center for use with male veterans. ^23,24 It has since been validated in women, and multiracial community samples. It has been translated into French, Spanish, Korean, and Mandarin Chinese. The SLUMS is calibrated to the MMSE as an 11-item 30-point scale. It differs from the MMSE in that it assesses episodic and rote memory. Episodic memory tests remembering embedded details of a story, a necessary ability for remembering medication instructions. The SLUMS assesses visual spatial executive function, important for driving, with the clock draw test (CDT) which is more sensitive than the overlapping pentangles figure drawing in the MMSE. The SLUMS queries verbal fluency as well as verbal recall. The SLUMS also has published calibration of scoring for educational level and MCI.^25,26 The SLUMS is available online for free with accompanying training video on the Gateway Geriatric Education Center. ²⁷

The most widely used and validated dementia screening tool is the Montreal Cognitive Assessment (MoCA.)²⁸ The MoCA is also a 30-point scale. Standardization for discriminating between MCI and dementia are not established,²⁹ however, a short form and a telephone form are available.³⁰ The MoCA is available in 65 languages. Paper versions 7.1 and 8.1 are free on-line for clinicians, teachers, and researchers on the MoCA institute website. To protect reliability of the instrument, since 2020 the MoCA Institute requires paid on-line training and certification for non-licensed professional use.²⁸ The Institute training site is on youtube.com (January 28, 2018) The MoCA also uses the CDT to assess executive function. MOCA tests visual naming recognition rather than free verbal recall. The items vary²² and the MoCA takes a few minutes longer to administer than the MMSE and the SLUMS.²⁴

There are several published head-to-head comparisons of the three screening forms. The SLUMS is not as widely used as the MMSE and the MoCA so there are fewer psychometric data for comparing populations.^24,29 There are limited data on test-retest variability in practice for the SLUMS and the MMSE,^28,23 however, the SLUMS was crafted to increase sensitivity to detect MCI.²² The Alzheimer’s Disease Assessment Scale cognitive subscale (ADA-Scog) is often used in clinical trials. However, it is intended for research use by clinically-trained specialists and is not appropriate for rapid primary care screening.³¹ Two systematic reviews concluded that the MOCA and the Mini-Cog had the best sensitivity among available tools in common use to detect MCI.^21,29,32,33

Each of the 30-point scales can take up to 15 minutes of office time. There are several widely used rapid screening tools that can trigger one of the more sensitive tools. The Mini-Cog asks the subject to remember three words, perform the CDT, and then recall the three words.³¹ It cannot be used on the telephone or reliably with patients with hearing or visual impairment. The Rapid Cognitive Screen (RCS) uses three SLUMS items: the five-word delayed recall, the CDT and one item of embedded memory from the story.³² The mini-MoCa is a five-minute version of the MoCA that correlated r=0.87 with the full MoCA and has published test-retest reliability.³³

The Short Blessed Test has been used since 1983 but it requires calculating weighted scores from six items. The scores corelate very well with MMSE but calculations make it more difficult to use.³⁴ If a patient is not able to perform a screening test due to illness, hearing loss or other cause, the Alzheimer’s Disease eight-item caregiver checklist (AD8) is widely used internationally. ³⁵ It queries whether the family member has noticed specific problems in memory, self-care, and task performance. For residents of long-term care and skilled nursing facilities, CMS mandates quarterly assessments using the Minimum Data Set 3.0 which includes cognitive screening of all verbal patients. The MDS 3.0 uses a tool called BIMS, the Brief Interview for Mental Status.³⁶ A social worker usually administers the 15-point BIMS. It has been validated as a predictor of need for ADL assistance in LTC, but it does not perform well outside that setting.³⁷

How to Perform Office-Based Screening

Medicare Annual Wellness and annual physical exams are a convenient way to introduce screening to establish a baseline. It should be part of the flow of the visit, in a quiet private space. If the patient does not speak English fluently, an on-line interpreter or family interpreter is not optimal because instrument reliability depends on standardized delivery and minimizing distractions and coaching. If a native language translation and administration is not possible, in this setting, AD8 may be more useful.²⁸ For English-speaking patients, the family should be instructed to avoid talking, coaching, or gesticulating. It is important to keep to the recommended time limits as speed is a component of the scoring. Using interpreters has not been factored into reliability and it is not advisable without extensive pre-visit training. It is important to note whether the patient makes a good effort to respond or gives up easily, which may indicate depression or delirium. Visually-impaired patients can skip the visual tasks in the SLUMS and calculate the score as a percentage.³⁷ Patients with hearing loss may perform better if examiners take off their masks to normalize frequencies and permit lip-reading. It is also useful to find out if the patient has “a better ear.” Hearing loss is increasingly being identified as a risk factor for cognitive impairment. ³⁸

In the flow of a routine follow up or acute visit, complaints of memory loss should be directly addressed with screening, even if it only occurs in the review of systems and not in the chief complaint. It can be revisited at another time if there is not time immediately. Having “trouble remembering things lately,” having to think harder about things that used to be easy are clues as are respectful queries about family complaints about missed bill payments or repetitive questions or stories. For new patients, a brief screen will indicate if one of the longer screening tests will be useful. Knowing the patient’s educational and occupational attainment are important indicators of health literacy.

Screening for cognitive impairment is required to complete a MAWV. EHR templates for AWV, Welcome to Medicare and Medicare Annual Physical Visits include brief screening tools. Screening, evaluation, and counselling about cognitive screening is covered by CMS under the AWV G codes. Use the CPT 99483 if screening is performed in the setting of a newly detected cognitive impairment during the AWV or during a routine office visit. It is reimbursed at $266 in most locations. More detail on coding is available at the CMS website.¹⁸

What To Do with Screening Results

The Alzheimer’s Association video demonstrates administering screening tests and, more importantly, video models of explaining results to family and patients.³⁹ Figure 1 walks through a branching clinical pathway for following up on cognitive screening results. As with other tests, what the clinician does with the results is important. A “normal” test is a baseline, however even small “mistakes” within the tests’ range of normal represent a benchmark for follow up at another time or as an advice to another clinician.

Figure 1.

Decision Tree for Cognitive Screening in Primary Care

Reversible causes of cognitive impairment should be addressed. A thorough medication review is mandatory at the first sign of cognitive impairment. This should include “natural” supplements, over-the-counter remedies, alcohol, and cannabis products. All prescription drugs must be reviewed for dose and timing with respect to onset of symptoms, indication, dose, and correct use. In our experience anticholinergic medications, sedatives, hypnotics and opiates, overly aggressive blood pressure, and glycemic control, especially in combination, are very commonly at the root of reversible cognitive impairment. Even if used safely for many years, pharmacodynamics and kinetics do change with aging and that may be the cause of new adverse effects of old drugs. The Beers criteria is a guide to drugs poorly tolerated by older adults which is available on-line at the American Geriatrics Society website.⁴⁰ We also use validated tools for “deprescribing.”⁴¹

A laboratory evaluation should identify electrolyte imbalances, renal and liver dysfunctions, endocrine abnormalities including thyroid dysfunction (both hypo- and hyper-), or evidence of poorly controlled or overly tightly controlled glycemia. Unrecognized anemia can be primarily responsible for or due to a common underlying condition. We always check vitamin B12 levels. Though it is a rare cause of dementia, it is easily treatable. Sleep disorders are common in older adults. Persistent complaints about insomnia may lead patients and physicians to use over the counter and prescription remedies with adverse effects on cognition.

Sleep disorders should be evaluated in any case. There is an established association between sleep apnea and mild cognitive impairment.⁴² Several studies provide evidence of improved cognitive performance with appropriate treatment of obstructive sleep apnea.⁴³

Abnormal results, especially if they are unexpected, require discussion of the meaning and next steps. If the patient is willing and the resources are available, referral to a PhD neuropsychologist or other certified psychometrician for a diagnostic assessment is especially useful and normally covered by insurance. A standardized two-hour battery of paper and pencil tests assess the “cognitive circuits.” The value of formal testing is that the pattern of deficits may indicate a specific diagnosis which is useful for prognosis and treatment. It can suggest whether treatment for co-morbid mental illness or depression may be useful. The battery also develops information about the patients’ relative strengths and deficits to guide them and their caregivers about how best to preserve their independence and dignity.

Further diagnostic evaluation includes brain imaging. The recommended first step is a noncontrast CT to rule out obvious pathology such as strokes, hemorrhages, mass lesions, or infiltrating lesions. For example, an early sign of central nervous system lymphoma can be cognitive loss. In the presence of known vascular disease, focal neurological deficits, gait abnormalities, or tremors and sensory abnormalities, referral to a neurologist is useful. Neurologists often prefer MRI for the finer structural detail but as a first step it is not necessary. If physical and radiological findings suggest normal pressure hydrocephalus (NPH) referral to neurosurgery for a diagnostic large volume spinal tap and shunting can be curative. Prior to referral and advanced imaging, routine blood tests including CMP, CBC, drug levels, thyroid functions, and vitamin B12 should be performed to rule out both acute problems and reversible subacute causes of cognitive dysfunction.

There are now many imaging options for specific diagnosis of cognitive impairment and dementia. Functional PET and MRI studies examine patterns of uptake of radiolabeled glucose, beta-amyloid, and dopamine to map regional brain metabolism and locate pathologic deposits of amyloid (Alzheimer’s disease) or deficiency in dopamine (Parkinson’s disease). There have been recent breakthroughs in identifying useful markers for diagnosis and disease monitoring using blood and CSF markers of beta amyloid and phosphorylated tau for tracking Alzheimer’s patients in research and drug trials. They are not widely available for clinical use outside of research centers.

Many primary care providers feel comfortable prescribing oral medications including rivastigmine, galantamine and donepezil, the central acetylcholinesterase inhibitors (CAI). Memantine is a central NMDA receptor antagonist for clinical cognitive impairment. The indications for starting CAIs are not entirely clear. Early treatment for MCI or mild dementia may provide the greatest benefit. Except for adverse drug reactions, the guidelines for stopping therapy are not clear. Memantine is added later for moderately advanced to severe dementia though the benefits also require discussion. The cost-effectiveness of starting or continuing these drugs in moderately advanced and advanced disease is unclear.⁴⁴

There are presently two FDA approved antibody infusion therapies directed at beta amyloid.⁴⁵ Lecanemab blocks the formation of beta-amyloid plaque in neuron cell bodies. Aducanumab removes beta-amyloid from the brain. They are expensive and not widely available. At present they should be used by specialists at specialty centers with the capacity for beta-amyloid scanning as proof of diagnosis and treatment response. As always, it is essential to stop any drugs that may be impairing cognition before starting another one to treat it.

The most important intervention for a patient and family with a new diagnosis of a progressive cognitive illness is the assurance of support from their physicians and other health and social service providers. Early referral to senior and caregiver support services in the community and on-line is important to help patients address the need for direct care, caregiver support and education and other needs that emerge. If the patient experiences behavioral symptoms, early referral to psychiatry is useful. In sum, therapeutic alliance among patients, family caregivers, and primary care practitioners is the bedrock of care for patients with dementia.

Conclusion

Dementia is a cluster of neurogenerative diseases that affect cognition. They are most common among older adults. As the population ages, the prevalence of dementia is increasing. There are no cures for these diseases despite considerable progress in research on Alzheimer’s disease. Because of the lack of effective treatment, the question of screening for early detection of dementia has been debated. However, the current consensus is that it is useful for patients and families because it provides a framework for adaptive strategies, time to learn about the disease , and time to plan for future care needs. Screening, care planning, and patient education are billable and easily incorporated into annual exams and particularly the MAWV.

The initial work up for cognitive impairment includes a rigorous medication review and reconciliation, basic imaging, and laboratory investigation for reversible causes. Referral to neuropsychological testing, specialized neurological investigation, and psychiatric assessment should be individualized based on the clinical presentation and family and patient preferences. Effective pharmacological management is emerging. It is always necessary to connect patients and caregivers with community support and educational resources.