Some CSIs are more suitable for and/or are specifically designed for use in primary care settings rather than secondary care settings, with time for administration being one of the key factors determining such suitability [78–81]. Examples include the Clock Drawing Test (see Chap. 5), GPCOG (see Chap. 10), 6CIT (see Chap. 11), short IQCODE (see Chap. 13), the Memory Impairment Screen (MIS) [41], Mini-Cog [47, 48], the Mental Alternation Test (MAT) [43, 44], Time and Change Test (T&C) [68, 69], and the cognitive disorders examination decision tree (Codex) [17, 18]. Generally these tests require little specialized test equipment beyond a pencil and paper and do not require significant training to administer.

Surveys of use of CSIs in primary care have found rather divergent results, perhaps dependent on study methodology. A much-cited postal survey suggested widespread use (ca. 80 %; [82]), whereas actual analysis of referral letters directed to cognitive clinics in secondary care presents a somewhat different picture [83]. Sequential studies in one clinic over a period of more than a decade (2004–2015) have suggested a gradual increase from around 20 % to around 40 % [84–89]. In the initial surveys, the Mini-Mental State Examination (MMSE) [90] was the test most commonly reported to be used in primary care, but this has gradually changed to 6CIT [91], perhaps in part due to enforcement of copyright restrictions on the use of the MMSE and perhaps because 6CIT is specifically recommended for use in primary care.

The CSIs described in detail in this volume can be administered in between <1 and about 20 min (Table 15.2). Test duration will determine the suitability or otherwise of certain tests for certain situations, for example ACE and its iterations ACE-R and ACE-III (see Chap. 6) will be too long for use in primary care settings, and this criticism has also been made of MMSE for primary care use, hence favoring instruments such as 6CIT and GPCOG.

Trade-off between speed and accuracy is recognized in many spheres. Examining various cognitive screening instruments and using surrogate markers of time (total test score; total number of questions), correlations were found between these and measures of test accuracy (correct classification accuracy; area under the receiver operating characteristic curve), suggesting that longer tests may improve diagnostic accuracy [92, 93].

If test duration is an issue affecting applicability, then ultra-short screening tests or “microscreening” tests, comprising just a single, or two or three, questions, may be desirable.

For example, a Chinese study reported sensitivity of 0.96 and specificity 0.45 for the diagnosis of dementia by asking a single question concerning progressive forgetfulness [94]. A single question is advocated in the United Kingdom Dementia Commissioning for Quality and Innovation (CQUIN) policy document of 2012 [95] but there is no evidence base to justify this particular question, and reasons, both theoretical [96] and empirical [97], to believe that it would identify many false positives. A systematic review of single screening questions for cognitive impairment in the elderly found only a very limited evidence base [98], so this is an area in which more work is required. Questions related to ability to manage personal finances and medications, use a telephone and public or private transport [99], or learning to use new gadgets [100] have been shown in epidemiological studies to be particularly useful for dementia diagnosis, combinations sometimes having comparable or better diagnostic utility than MMSE [100] but such simple questions have yet to be submitted to diagnostic test accuracy studies [101].

Single clinical observations may also be useful as screening tests. Verbal repetition, i.e. repeating the same question or information after only a few minutes, was observed in 100/130 (=77 %) mild-to-moderate AD patients [102]. Observation of the head turning sign (patient looks at the care-giver when asked a question) may also have screening value, although the exact operationalization of the sign has differed between reported studies [103, 104]. Attending a cognitive clinic alone despite provision of written instructions to bring a relative or friend to give collateral history (the “attended alone” sign) is a robust indicator of (i.e. is very sensitive for) the absence of dementia [105]. The same is probably also true of the presentation of a written list of symptoms (la maladie du petit papier) [106].

Some cognitive instruments may, by contrast, be too long for routine application in day-to-day clinical practice even in secondary care settings, and indeed for that reason may not be regarded as CSIs. For example, the Alzheimer’s Disease Assessment Scale-Cognitive Section (ADAS-Cog) [107] has been widely used as a reference measure, for example as an outcome measure of drug efficacy in AD clinical trials practice, and takes significantly longer to perform than the MMSE (around 30–45 min). A “calculator” to convert MMSE scores to equivalent ADAS-Cog scores is available, reflecting the strong correlation between ADAS-Cog and MMSE scores [108]. The cognitive battery proposed by the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) investigators is also time consuming, incorporating the MMSE and other subtests of memory, naming, and verbal fluency [109]. Likewise the Dementia Rating Scale (DRS) and its successor (DRS-2) [110] which comprise a number of subtests (attention, initiation, construction, conceptualization, memory) to give a global measure of dementia (score 0–144), takes about 30 min to perform.

In this context it is also necessary to mention the Clinical Dementia Rating (CDR) [111, 112] and the Global Deterioration Scale (GDS) [113]. These are global staging measures based on both cognitive and functional capacities, which have gained prominence through their use in the definition of mild cognitive impairment (CDR 0.5 and GDS 3 correlate, but are not necessarily synonymous, with MCI). CDR has been reported to be useful in screening for dementia [114].

Cognitive screening instruments may be classified according to whether they test general or specific cognitive functions [77, 115, 116]. One of the desiderata for CSIs as formulated by the American Neuropsychiatric Association was sampling of all the major cognitive domains, including memory, attention/concentration, executive function, visual-spatial skills, language, and orientation ([117]; see Chap. 1, at Sect. 1.​3). Many CSIs attempt this broad, multidomain, sampling to a greater or lesser extent (e.g. MMSE, ACE/ACE-R/ACE-III, MoCA; see Chaps. 3, 6, and 7 respectively). Generally, the more comprehensive the neuropsychological coverage, the longer the test takes to administer, although the Clock Drawing Test (see Chap. 5) may be an exception.

On the other hand, instruments which test a specific cognitive function may have a place in screening [116]. For example, since episodic memory impairment is typically the earliest deficit manifest in AD patients, tests for anterograde (“hippocampal”) amnesia may be particularly pertinent, such as the Memory Impairment Screen (MIS) [41], the Free and Cued Selective Reminding Test or Five Words Test [29], and the Visual Association Test [71]. Similarly, tests of visuoperceptual function such as the Poppelreuter (overlapping) figure may identify deficits in this cognitive domain which may occur early, for example in posterior cortical atrophy or the visual variant of AD [51]. Scales specifically measuring attention, executive functions, and language are also available [77], some of which may be of particular value in specific clinical situations, e.g. assessing executive and/or language function in suspected frontotemporal lobar degeneration syndromes (see below, at Sect. 15.3.4).

Cognitive screening instruments are most often administered to patients (Part II), most usually by the clinician, but are sometimes undertaken by the patient themselves, usually with medical supervision (e.g. TYM; see Chap. 9). Clinician administration of a cognitive screening instrument permits a qualitative patient-clinician interaction during testing which may inform clinical judgments over and above the raw test scores which emerge. The clinician’s gentle, persuasive technique of test administration may also ensure that liability to drop out is less likely than with patient self-administered tests.

Because of the importance of collateral history in the assessment of possible cognitive disorders, such that diagnostic guidelines for dementia have emphasized the importance of informant interview [118, 119], scales to be completed by a knowledgeable informant may also have a place in assessment (Part III). Examples include the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE; see Chap. 13), the Neuropsychiatric Inventory (NPI) [120], the Short Memory Questionnaire (SMQ) [61], and the Dementia Questionnaire (DQ) [24]. Some scales may be suitable for both patient- and informant-administration purposes (e.g. AD8; see Chap. 14). An informant component is also incorporated in the GPCOG (Chap. 10). Informant scales which help in the differential diagnosis of dementia subtype have also been reported: the Cambridge Behavioural Inventory (CBI) may assist in differentiating AD and frontotemporal lobar degenerations [121–123] (see below, at Sect. 15.3.4), and the Fluctuations Composite Scale may assist in diagnosis of DLB [124, 125] (see below, at Sect. 15.3.3).

Most CSIs produce a global score to be compared against cut-offs said to define normal/abnormal test performance (see Chap. 2, at Sect. 2.​3.​1). Test subscores may identify particular areas of weak cognitive performance. However, too much reliance should not be placed on such overall numerical values since there are many factors other than cognitive decline which may influence test performance, including patient age, educational status, culture, language, presence of primary psychiatric disorder (anxiety, depression), and presence of primary sensory deficits (see Chap. 1, at Sect. 1.​3). As previously mentioned (above, at Sect. 15.2.4), qualitative aspects of performance on administration of CSIs may also inform clinical diagnosis. Moreover, test cut-offs defined in index studies, which may utilize highly selected patient cohorts and normal control groups, may not be applicable in day-to-day clinical practice [126] wherein all patients have at least subjective memory complaint, itself not necessarily a benign condition [127]. Revision of test cut-offs to scores more appropriate for the casemix seen in a particular clinic has been reported for several cognitive screening instruments including ACE-R (see Chap. 6), MoCA (see Chap. 7; [128]), and TYM (see Chap. 9; [129]).

Some tests are qualitative, such as the Queen Square Screening Test for Cognitive Deficits [52]. Although the Cambridge Behavioural Inventory can be scored [122], the authors of the test suggested that the overall benefit of the instrument was in providing a structured behavioral symptom profile rather than a summated behavioral score [130].

AD is the most common dementing disorder with over 20 million cases estimated worldwide. As episodic memory impairment is the most frequent early symptom of AD, specific tests for this construct may be most appropriate when there is clinical suspicion of this diagnosis. Such tests of episodic memory include the Memory Impairment Screen (MIS) [41] and the Free and Cued Selective Reminding Test or Five Words Test [29, 151].

Of the general cognitive function tests, MMSE (Chap. 3) is thought to be rather insensitive for AD, particularly in its mild stages, but combination of MMSE with the Clock Drawing Test (“Mini-clock”) has been reported to be highly sensitive and specific in detection of mild AD [152]. Some of the MMSE variants (Chap. 4) are reported to be sensitive and specific for AD diagnosis, such as Modified Mini-Mental State Examination-Revised (3MS-R) and the Six-Item Screener (SIS). The Addenbrooke’s Cognitive Examination (ACE) and its successors, ACE-R and ACE-III, are sensitive for AD diagnosis; the VLOM subscore of these tests has good sensitivity and specificity for the diagnosis of AD (Chap. 6). The Montreal Cognitive Assessment (Chap. 7) is sensitive for mild AD, and the Test Your Memory (TYM) test (see Chap. 9) is reported to be better at identifying AD cases than the MMSE [153]. Of the commonly use informant scales, IQCODE (Chap. 13) has also been reported to show excellent screening properties for AD [154].

Other tests reported to be effective in screening for AD include the Scenery Picture Memory Test [155], the screening test for Alzheimer’s disease with proverbs [156], the Philadelphia Brief Assessment of Cognition [50], the Memory Alteration Test [39], the three-objects-three-places test [157], the traveling salesman problem (a visual problem solving task; [158]), the Short Cognitive Evaluation Battery [60], the Visual Association Test [71], and the 7-min neurocognitive screening battery [57].

The evolution of AD is characterized by asymptomatic, predementia and dementia phases, evolving over many decades, the former with or without symptoms [159], and for which criteria have been developed [160]. In the later, symptomatic, stage of the predementia phase a syndrome of prodromal AD or mild cognitive impairment (MCI) may be defined [161].

Identification of MCI is, at least theoretically, a high clinical priority since early interventions might possibly arrest or slow disease progress sufficient to prevent the development of dementia. Although probably a heterogeneous disorder at the clinical level, nevertheless tests highly sensitive for detection of MCI are desirable.

A systematic review identified a number of cognitive screening instruments capable of identifying MCI [162]. For example, MoCA (see Chap. 6) was reported to be very sensitive for diagnosis of MCI, moreso than the MMSE [163]. Both MoCA and ACE-R are highly sensitive for the diagnosis of MCI [164], and MoCA and Mini-ACE are comparable in terms of effect size (Cohen’s d) [165]. The Quick Mild Cognitive Impairment (Qmci) screen [166], derived from the ABCS135 [2, 3], also has significant promise for MCI identification (see Chap. 12). A systematic review concluded that the Clock Drawing Test was not suitable for MCI screening [167] (see Chap. 5, at Sect. 5.​6.​2, for fuller discussion). Combination of the MMSE and the Clock Drawing Test (“Mini-clock”) is reasonably accurate in separating MCI cases from healthy controls [152]. Of the informant scales, IQCODE has also been reported to show excellent screening properties for MCI [154].

“Vascular dementia” (VaD) is not a unitary construct, encompassing such entities as vascular cognitive impairment (VCI) short of dementia, poststroke dementia, multi-infarct dementia, subcortical ischemic vascular dementia (SIVD), and selective infarct dementia [168]. Such heterogeneity at clinical, etiological, and neuropathological levels poses significant problems in devising cognitive screening instruments specific for “vascular dementia”, the moreso when the frequent overlap with neurodegenerative processes such as AD is taken into account [169]. Furthermore, it is recognized that some cognitive screening instruments may be “Alzheimerized”, i.e. suitable for picking up the characteristic deficits in AD (viz. episodic memory) but not necessarily those in VaD/VCI. Although there is overlap in the profile of neuropsychological deficits, vascular cognitive syndromes may show greater impairments in attention, working memory, and executive function than encountered in AD patients [170].

To detect cognitive impairment related to cerebrovascular disease, derivations from existing tests may be used, or adaptations of existing tests, such as the CAMCOG (R-CAMCOG) [171] or ADAS-Cog (VADAS-Cog) [172]. Although the MMSE apparently remains the most widely used instrument to screen for VaD/VCI, a systematic review found it to have insufficient criterion validity, and favored other instruments such as MoCA (e.g. [173, 174]) (see Chap. 7, at Sect. 7.​7), Cognistat [13], and the Functional Independence Measure-cognition as having good predictive values [175], although the latter compared unfavorably to R-CAMCOG in one study [176]. Screening for vascular cognitive impairment using the Diagnostic Checklist for Vascular Dementia but using the MMSE rather than the detailed neuropsychological part of the checklist has been reported [177] and a subscore of the MMSE has also been reported to identify VaD [178] (see Chap. 4, at Sect. 4.​3.​1).

The Hachinski Ischemic Score is a brief clinically based scale (Table 15.4) used to differentiate AD and multi-infarct dementia [179], in which context it performs well, although there are problems with the diagnosis of mixed dementia [180]. The scale score has been used in many AD drug trials as an exclusion criterion for possible cases of vascular dementia.

The Brief Memory and Executive Test (BMET) was specifically designed as a quick bedside screening test for VCI due to cerebral small vessel disease and is reported to have high sensitivity and specificity for differentiating such patients from those with AD, in which it outperformed the MMSE [10].

It must be remembered that motor impairments following stroke may affect performance on cognitive screening instruments. How these omissions are handled may have implications for how tests are rated, and this requires to be made explicit [181].

Compared to AD, visual and executive cognitive functions are recognized to be more frequently impaired in cognitive syndromes (dementia, MCI: PDD, PD-MCI) associated with Parkinson’s disease (PD) and in dementia with Lewy bodies (DLB), with relative preservation of orientation in time and place (e.g. [182, 183]). A number of tests which seek to exploit these differences and thereby facilitate diagnosis of cognitive impairment in PD and DLB have been developed (Table 15.5), in addition to the more standard screening instruments.

The Mini-Mental Parkinson (MMP) [145], a derivative of the MMSE, has already been discussed (see Chap. 4, at Sect. 4.​2.​8). The Parkinson neuropsychiatric dementia assessment (PANDA) instrument comprises five cognitive tasks and a depression questionnaire and was reported to have sensitivity of 0.90 and specificity of 0.91 for PDD [184]. The Parkinson’s Disease – Cognitive Rating Scale (PD-CRS) was designed to cover the full spectrum of cognitive deficits found in PD, and was found to diagnose PDD accurately [185] A shorter version, the PDD-Short Screen (PDD-SS) [186], takes about 5–7 min to administer. The Scales for Outcomes in Parkinson’s Disease – Cognition (SCOPA-COG) instrument consists of ten items based on the most common cognitive deficits in PD (maximum score 43) and which proved sensitive and specific [187]. SCOPA-COG may be more discriminative than MMP [188]. To these disease specific scales may be added the Fluctuations Composite Scale (FCS), derived from the Mayo Fluctuations Questionnaire of Ferman et al. [124], which has been reported in a pragmatic study to identify synucleinopathies (PDD, PD-MCI, DLB) when these conditions have entered the initial differential diagnosis of cognitively impaired patients [125].

Usage of the commonly used cognitive screening scales to detect cognitive deficits in PD and DLB has been reported. A subscore of the MMSE defined by Ala et al. [189] was reported to facilitate detection of DLB versus AD (see Chap. 4, at Sect. 4.​3.​2). Similar weighted subscores can be derived from the ACE [190] and MoCA [191]. ACE-R has been reported a valid tool for dementia evaluation in PD [192], and useful as one component of a three-step procedure to identify dementia in PD, as have MoCA and the Frontal Assessment Battery [193]. A number of other studies (e.g. [194–196]) have shown utility of MoCA in detecting cognitive impairment in PD (see Chap. 7, at Sect. 7.​8).

ACE may be used to detect cognitive impairment in the “atypical” parkinsonian syndromes such as progressive supranuclear palsy, corticobasal degeneration, and multiple system atrophy [150, 197].

The heterogeneous group of frontotemporal lobar degenerations (FTLD) may present with either behavioral or linguistic impairments [198]. Delayed diagnosis of these conditions, particularly behavioral variant frontotemporal dementia (bvFTD), is a frequent observation, despite informant report of behavioral change, with the syndrome often being labeled psychiatric and treated as such [199]. Hence, instruments sensitive to frontal lobe dysfunction which might facilitate diagnosis of bvFTD have been described (Table 15.6).