What is the purpose of cognitive screening? This issue may be addressed by considering the classic criteria for disease screening published under the auspices of the World Health Organization (WHO; see Box 1.1) [36, 37], and also published guidelines and criteria for developing screening programs [38] such as those from the UK National Screening Committee (www.​nsc.​nhs.​uk).

Many of these conditions are fulfilled for dementia as a syndrome, and for specific subtypes of dementia, most importantly Alzheimer’s disease (AD). For example, the public health implications of dementia and its huge economic costs are unequivocally established [12–17]. It is also evident that the natural history of most forms of dementia encompasses a presymptomatic phase, with disease evolution occurring over many years before clinical presentation. Longitudinal epidemiological studies suggest almost 10 years of cognitive decline in AD preceding dementia [39]. Biomarker studies indicate that the neurobiological changes which underpin Alzheimer’s disease commence many years, indeed decades, before the emergence of clinical symptomatology [40–42]. This long presymptomatic phase presents a potential window of opportunity for disease identification, and intervention should disease modifying drugs become available.

Equally, many of these screening criteria are yet to be fulfilled for dementia. For example, it has yet to be established that any of the available pharmacotherapies for AD are more beneficial when applied at the presymptomatic stage compared to the later symptomatic stage. Application of pharmacotherapies in presymptomatic AD has, to my knowledge, yet to be reported but there is no evidence that cholinesterase inhibitors, a symptomatic treatment for AD, prevent conversion of prodromal AD (mild cognitive impairment) to AD in the long term [43–45]. It is not clear that healthcare systems have the capacity and policies to test for dementia and deal with the consequences, nor that the cost of case finding, including diagnosis and treatment, would be economically balanced in relation to possible expenditure on medical care as a whole.

Putting aside these issues, which may possibly be resolved by ongoing research, the key screening criterion considered in this book is whether there are suitable tests or examinations available to detect dementia and its subtypes with reasonable sensitivity and specificity, and which are acceptable to the population. The population in question needs careful definition in this context, since prevalence rates of dementia may differ greatly in different populations. Hence, a cognitive screening instrument to be applied at the whole population level might be very different to one applied to at-risk groups (e.g. older persons) or to the highly selected population attending cognitive/memory clinics. The latter, pretty much without exception, have at minimum subjective memory complaints. It is to the constituency of those presenting to clinical attention with memory complaints that the current volume is addressed.

As with all medical activities, such as investigation and treatment, a screening process may be associated with both clinical benefits and risks, which should be recognized at the outset. Screening for dementia is not equivalent to diagnosis, which remains at least in part a clinical judgment made by those experienced in the diagnosis of these conditions, a process which needs to take into account the marked clinical and etiological heterogeneity of the dementia syndrome [34, 46–51] and the inadvisability of accepting “one size fits all” approaches [52, 53]. Screening can therefore never replace the clinical interview.

Because screening tests for dementia can never have perfect sensitivity and specificity (i.e. = 1), there will always be a risk of false positive and false negative diagnoses (see Chap. 2). Highly sensitive tests, which are generally thought desirable for screening purposes, will ensure that early cases are not missed but at the risk of making false positive diagnoses (with all the attendant, and ultimately unnecessary, anxiety, treatment risks, etc., that false positive diagnosis may entail). Highly specific tests minimize incorrect diagnoses but may miss early cases (false negatives). Screening tests that disclose abnormalities only when a disease is clinically obvious are of limited applicability, indeed measures of test performance may be inflated by using patients with established diagnoses.

What features would be desirable for the optimal cognitive screening instrument?

A number of criteria for such an instrument were enunciated nearly 20 years ago by the Research Committee of the American Neuropsychiatric Association [54]:

To these criteria one may add:

Other issues may also require consideration when selecting a cognitive screening instrument, for example the location in which testing is undertaken (primary or secondary care) and the suspected dementia diagnosis being screened for (see Chap. 15, at Sects. 15.​2.​1 and 15.​3 respectively). In primary care settings, briefer tests may be optimal [8, 61, 62]. If the suspected diagnosis being screened for is AD then tests which focus on the examination of episodic memory, to the relative exclusion of other cognitive domains, may be preferred.

Cognitive screening instruments are “noisy”, which is to say that a variety of factors may influence patient performance to obscure any signal of cognitive impairment due to brain disease (i.e. factors unrelated to the construct the tests have been designed to assess). These include patient age, educational status, culture, language, the presence of primary psychiatric disorder (anxiety, depression), and presence of primary sensory deficits (visual or hearing impairment). For example, one study found that poor performance on the MMSE [63] due to causes other than dementia was recorded in around 10 % of an elderly population, increasing with age (>40 % in those ≥85 years), most commonly due to poor vision and hearing, deficient schooling, and the consequences of stroke [64].

It is well-recognized that test performance may vary with factors such as the environment in which testing is undertaken (e.g. the alien surroundings of an impersonal clinic room vs. the familiar location of the patient’s home) and tester (e.g. perceived to be sympathetic and encouraging vs. brusque and impatient). All these factors may need to be taken into account when using cognitive screening instruments, rather than relying solely on raw test scores. Corrections to test scores or revision of cut-offs may be applicable to allow for patient age and education [65–67].

Educational and cultural biases are evident in many typical screening test items [68]. For example, tests which rely heavily on literacy will be challenging for individuals with limited education or from cultures using a different language. Screening tests may thus need adaptation for these factors. Tests which may be characterized as tests of performance have a long history [69] and continue to be developed [70]. Similar considerations apply to patient ethnicity. Cultural modifications have been reported for a variety of cognitive screening instruments, including the MMSE, the Short Portable Mental Status Questionnaire, and the Short Orientation-Memory-Concentration Test [68]. Cultural factors may also affect willingness to be screened for cognitive impairment [71]. Ideally culture-free cognitive screening tests should be developed: claims for such status have been made for the Mini-Cog [72] and the Time and Change Test [73]. Patient assessment by means of informant reports (see Part III of this book) may be relatively culture-free, as may also be the case for functional assessments.

Cognitive screening instruments are not equivalent to a neuropsychological assessment administered by a clinical neuropsychologist, which remains the “gold” or reference standard for cognitive assessment. The tests used in neuropsychological assessment are potentially many [10, 74–76] and tend to focus on function within individual cognitive domains or give a global measure of intelligence (verbal, performance, and full-scale IQ). Requirement for a trained neuropsychologist to administer such tests means that access is not universal. The test battery administered is often time-consuming (much greater than the 15 min suggested by the Research Committee of the American Neuropsychiatric Association [54]), fatiguing for patients, and may sometimes require multiple outpatient visits. Hence neuropsychological assessment is not a plausible means for screening cognitive function, although it may be necessary to clarify diagnosis in those identified as cognitively impaired by screening instruments.

In an age in which dementia biomarkers, based on the findings of sophisticated neuroimaging and biochemical testing, are beginning to be used to define disease entities even before the onset of dementia per se [77–79], it may be questioned what role there may be for cognitive screening instruments in dementia diagnosis. The interrelationships of cognitive screening instruments and biomarkers are only beginning to be investigated [80].