Intellectual Disabilities: A systematic review

Background and Aims: Screening and assessment of cognitive changes in adults with Intellectual 14 Disabilities, mainly Down Syndrome (DS), is crucial to offer appropriate services to their needs. We present 15 a systematic review of the existing instruments assessing dementia, aiming to support researchers and 16 clinicians’ best practice. Methods: Searches were carried out in the databases Web of Science; PubMed; 17 PsycINFO in March 2019 and updated in May 2020. Studies were selected and examined if they: (1) focused 18 on assessing age‐related cognitive changes in person with ID; (2) included adults and/or older adults; (3) 19 included scales and batteries for cognitive assessment. Results: Forty‐eight cross‐sectional studies and 20 twenty‐six longitudinal studies were selected representing a total sample of 5,851 participants (4,089 DS 21 and 1,801 with other ID). In those studies, we found 38 scales, questionnaires, and inventories, and 14 22 batteries for assessing cognitive and behavioural changes in adults with DS and other ID. Conclusion: The 23 most used instrument completed by an informant or carer was the Dementia Questionnaire for Learning 24 Disabilities (DLD), and its previous versions. We discuss the strengths and limitations of the instruments 25 and outline recommendations for future use. 26


Introduction
Individuals with intellectual disabilities (ID) may be at an increased risk of developing dementia when compared to the general population [1].In people with ID, the prevalence of dementia is as high as 4% in individuals under 40 years, and 40% in those 60 years or older, with an average age of onset between 51 and 56 years [2,3,4].Early detection of dementia can be challenging in individuals with ID [5], many of the instruments for assessing dementia-related cognitive changes in the general population are based on the assumption of sound premorbid cognitive functioning, which is difficult to determine in those with ID [6,7,8].Furthermore, the clinical presentation of dementia in those with ID may differ compared to the general population, with personality and behavioural changes presenting earlier [9,10].
Single domain cognitive tests are the usual approach to screen for dementia in the general population, as they indicate progressive deterioration in cognitive domains [11].However, in people with ID, these tests are not appropriate due to pre-existing conditions which makes it more difficult for them to perform meaning the results cannot be interpreted in a substantial and valid way, as there are often no norms for this population [11].This has been addressed in recent research carried out by Benejam et al. (2020) [12], who used the CAMCOG-DS in people with Down syndrome to accurately diagnose Alzheimer's disease.This shows the importance of developing reliable population norms for appropriate instruments when assessing cognitive changes in people with ID.

Down Syndrome Intellectual Disability
Among adults with ID, there is a well-established link between Down syndrome (DS) and dementia, particularly Alzheimer's disease (AD).Research indicates that 95% of people with DS will develop AD by the age of 65 [4, 13, 14].Individuals with DS also have an increased risk of developing early-onset dementia; the clinical presentation of dementia symptoms before the age of 65 [4,13, 15].The increased prevalence of AD in DS is largely due to genetic factors associated with trisomy 21, the most common form of DS.Those with trisomy 21 have a third copy of chromosome 21 [16], which is responsible for the production of β-amyloid precursor protein [17].
The increased presence of β-amyloid precursor protein leads to an accelerated build-up of senile plaque in the brain, which is a primary cause of AD [16].By age 40, most individuals with DS display neuropathological changes consistent with AD, while most individuals with DS show clinical signs of dementia by age 50 [18].Similarities of symptoms between AD and DS suggest common risk factors among AD and DS.Prasher and colleagues (2008) [19] examined Apolipoprotein (APOE) genotyping in people with DS, concluding that those with APOE E4 allele had a significantly higher risk of developing AD, had an earlier onset of AD, and a higher rate of progression to death when comparing for participants with APOE 3 allele.Screening for APOE genotype in this population may be of good clinical utility as it helps people obtain early treatment, which can reduce early mortality rates [19,20].Startin et al. (2019) [21] recently "conducted the largest cognitive study to date" (p.245) with 312 participants with DS in order to assess typical age-related and AD-related cognitive changes in this population.The authors reported memory and attention measures were most sensitive to decline, although the earliest cognitive markers of AD-related pathology were identified on most outcome measures.They also reported an age-related relationship where older age groups showed poorer performance in neuropsychological tests, except for scores on the Behaviour Rating Inventory of Executive Function -adult version; a measure of executive function.However, other research has indicated that declines in executive function may precede memory loss in those with DS and AD [22], suggesting further research is needed to determine the typical progression of AD in this population.

Other Intellectual Disability
There is less conclusive evidence of an increased risk of dementia in those with an intellectual disability not related to DS (herein other ID).While there may be several genetic factors, leading to increased risk of dementia in those with other ID -such as reduced baseline cognitive ability and fewer neurons and synaptic connections [1] -older adults with other ID show protective factors against developing dementia, including lower rates of smoking and greater cardiovascular health compared to the general population [23].Some research suggests the prevalence of dementia for individuals with other ID may be the same or slightly higher than the general population [24,25], although a longitudinal study by Strydom et al. (2013) [1] reported that dementia might be five times more prevalent in this population.However, epidemiological studies may underestimate true prevalence rates due to several factors.Firstly, dementia is under-diagnosed in the general population -it is likely that this is also present in those with ID [8].Secondly, those with ID generally have poorer access to health care services [26,27], which could result in lower levels of diagnosis.Finally, dementia presents differently in those with ID compared to those without, leading to difficulty in diagnosis [8]. a more robust way to screen for dementia in this population [8,11].The review will look at the strengths and limitations of instruments and aims to provide researchers and clinicians with an up to date, comprehensive list of available tools.

Materials and Methods
The methods for this review were based on the recommendations of the (3) studies including scales and batteries for cognitive assessment.Sixty-one articles were excluded based on exclusion criteria (review studies and/or intervention studies, or the age of participants not matching the criteria).In total, 140 articles were included for a thorough review (as shown in Fig. 1.).A manual search of the reference sections of the retrieved studies and review articles was conducted.However, no new articles meeting the inclusion criteria were found.
We analysed 48 cross-sectional studies and 26 longitudinal studies qualitatively, excluding 66 articles for not meeting inclusion criteria (e.g., review studies, intervention studies, and studies including children or adolescents).In total, 74 articles were included in this review.All articles were reviewed by two researchers independently.In the few cases of disagreement, discrepancies were solved by consensus.

Quality Assessment
As for critical appraisal of the studies included in this review, a standardised checklist to identify the risk of bias was used to assess the quality of included studies.The checklist was based on the Newcastle-Ottawa Scale (NOS) [30], embedded on the table A2 and A3.A total score with a maximum value of nine points provides a rating for the quality level.Quality levels of evidence were defined as high (9-7 points); medium (6-4 points), and low (3-1 point).No studies presented low-quality range.
Of the 48 cross-sectional studies, 24 included only participants with DS, while the remaining 24 included individuals with DS and other ID, described in Table 1.The tables for cross-sectional and longitudinal studies (Appendix, Tables A2 and A3) present the characteristics of the participants (age, diagnosis), intervention, comparison, outcomes, and study design [31] structured according to the eligibility criteria.The average duration of longitudinal studies was 99.9 months (sd: 217 months), with no data for one study.We found 38 scales, questionnaires, and inventories, and 14 batteries for assessing cognitive and behavioural changes in adults with ID.Among the selected studies, we identified a multitude of different instruments (single-domain cognitive tests; scales; batteries; tasks), with few replications, and a lack of descriptive data (means, standard deviations, gender ratios, specificity and sensitivity scores) in publishing material, which was not obtained from all authors upon request.Consequently, a meta-analysis could not be performed.Of the 26 longitudinal studies, the majority (n =18) focused on DS, while the remainder (n=8) included participants with DS and other ID.There was also a large degree of heterogeneity in measures used in longitudinal studies including those with both DS and other ID.Within the seven studies included, 30 measures and tasks were reported.All datasets generated for this study are included in the article or its supplementary material, including tables 4-7 list of instruments used in the studies, PRISMA checklist and SWiM checklist.

Discussion and Implications
This study aimed to systematically review scales and batteries for screening for cognitive changes in adults with ID and provide a guide for practitioners and researchers to choose valid, reliable instruments.This review found a multitude of materials used with adults with ID, with much of the research focusing on those with DS.We focused on batteries and scales as the best approach to evaluate cognitive changes and age-related changes in individuals with ID [8,11].
The current evidence encourages the focus on two measures (DLD and CAMCOG-DS) which should be further explored psychometric, clinic and longitudinally based on the available literature.
Identified instruments can be divided into two categories: informant-based measures (answered by a carer) and self-report measures (answered by the individual).Across the literature, the diagnosis of dementia in this population is a major concern and subject to a disagreement regarding which instrument to use; there is also considerable disagreement surrounding which instruments better discriminate mild neurocognitive disorder and preclinical dementia [8].Studies are discussed according to the study design and clinical groups.

Longitudinal Studies in Participants with Down Syndrome
The present review identified a multitude of measures used to assess cognitive change in those with DS -35 separate measures and tasks were used across the 17 studies.The Dementia Questionnaire for Learning Difficulties (DLD -previously referred to as the Dementia Questionnaire for Persons with Mental Retardation, or DMR) [32,33,34] was the most frequently used measure, appearing in six studies [4,35,36,37,38,39].The frequent use of the DLD may reflect its recommendation by the National Institute for Health and Clinical Excellence -Social Care Institute for Excellence in the UK [40].The DLD, an informant-based measure, was developed by Evenhuis (1990) [32] for use with Dutch speakers but has since been translated and used in several countries, allowing cross-cultural comparisons [36,39,41].The DLD consists of 50 items and eight subscales and provides scores for cognitive and social domains.Previous research has noted that the DLD is widely used due to high levels of agreement between its scores and clinician's diagnosis [42] as well as its good sensitivity and specificity [43].
In the included studies, the DLD was effective in identifying deterioration in cognitive and social skills in adults with DS over time [39] , although Nelson et al. (2007) [38] noted that while DLD total scores showed good overall test-retest reliability after one year (r=.77), there was low test-retest reliability for the social scale (r=.45).In another study, [36] using the cognitive element of the DLD as a secondary measure to examine the impact of seizures on cognitive impairment in adults with DS, Lott et al. (2012) [36] found that the cognitive scale of the DLD identified increased deterioration in adults with DS and AD with seizures compared to those without seizures.Similarly, a 14-year longitudinal study by McCarron et al. (2014) [37] found that epilepsy was identified as a significant predictor of dementia in adults with DS and noted the DLD was the most sensitive instrument for tracking cognitive changes over time.However, another study [39] reported that the DLD showed poor sensitivity in distinguishing between dementia-related cognitive decline and depression, which is likely due to the inclusion of the social skills element of the questionnaire.Furthermore, Evenhuis et al. (2009) [33] suggested that this measure may not have adequate sensitivity when used with people with severe and/or profound ID due to a floor effect; similarly, it may also be problematic with those with mild ID due to a ceiling effect on cognitive function.
The Severe Impairment Battery (SIB) [44] is another measure of cognitive functioning which has been used longitudinally.The SIB is a self-report measure assessing cognitive function across nine domains: attention, language, orientation, memory, praxis, visuospatial perception, construction, social skills, and orientating head to name [45].The SIB was used in 3 longitudinal studies exclusively examining those with DS [35,36,37].Like the DLD, [36] the SIB was effective at tracking the cognitive decline in adults with DS and seizures; it was used as a secondary measure and provides a limited description of its effectiveness [35,37].

Longitudinal Studies Including Participants with DS and other ID
There was no overlap between measures used across studies, with no measure included in more than one study.This is illustrative of the lack of standardised measures for assessing cognitive decline in those with other ID and highlights the need for an accepted, recommended measure to allow synthesis across different studies.
It is interesting to note that the DLD was only used in a single study including participants with other ID [41].The study found that the DLD showed good test-retest reliability within their sample and reported that DLD scores showed agreement with other measures of cognitive change used in their study.

Cross-Sectional Studies in Participants with Down Syndrome
As was the case with longitudinal studies, the DLD [32] was the most frequently used instrument, appearing in eight studies [42,48,49,50,51,52,53].While the DLD was generally reported as a good marker of cognitive decline and dementia in those with DS, [18], one study found no association between scores on the DLD and the presence of beta-amyloid precursor protein, a biological marker of senile plaques and neurofibrillary tangles present in AD.While this may indicate that the DLD lacks sensitivity in identifying early cognitive changes associated with AD in those with DS, the authors suggest that small sample size and lack of statistical power may have influenced their findings.
The SIB [44] was also frequently used, appearing in four cross-sectional studies [16,49,54,55] Witts and Elder (1994) [55] carried out a preliminary study on the use of the SIB with adults with DS and concluded that the measure was suitable to assess cognitive function in this population.Furthermore, they noted that no floor or ceiling effects were observed in scores on the SIB -this is advantageous as it indicates that the measure can be used to assess cognitive function in a wide range of individuals with ID.A later study [49] reported that the SIB showed good concurrent validity with the DLD.However, unlike Witts and Elder (1994) [55], the authors reported evidence of ceiling effects, which has implications for the clinical usefulness of the measure [49].They also identified the need for more longitudinal research to determine the effectiveness of the measure over time.A further potential limitation of the measure is reported by Head et al. (2011) [18], who noted that, like the DLD, there was no association between scores on the SIB and the presence of beta-amyloid precursor protein, which may indicate that the measure lacks sensitivity.

Cross-Sectional Studies Including Participants with DS and other ID
The DLD [32,33,34] revealed good psychometric properties in studies with participants with both DS as other ID.Eight studies used the DLD [43,56,57,58,59,60,61,62]. Shultz et al. (2004) [43] reported the sensitivity of the DLD as 0.65 and specificity 0.93.The instrument was found to be a good marker of the cognitive and affective symptoms observed in the early signs of dementia [60], and displays good inter-test validity with other instruments like the SIB [49] and the Alzheimer's Functional Assessment Tool (AFAST) [58].The DLD has shown adequate inter-rater reliability for all subscales, except behaviour and disturbance, with correlations of 0.68 or higher [33].
Due to problems with floor and ceiling effects in the assessment of people with ID, researchers have attempted to address this issue.Startin et al. (2016) [52] created a comprehensive neuropsychological assessment to evaluate people with DS and avoid ceiling and floor effects.
The LonDownS Consortium identified a set of tests for the evaluation in people with DS with minimum floor and ceiling effects.The authors suggest that the battery is suitable for most adults with DS, although half the participants with both dementia and DS were unable to undertake any of the cognitive tasks in the battery, indicating that it may be useful for screening before the development of dementia [52].This instrument has also been found to have "limited diagnostic value as a single assessment" because it is not possible to estimate the extent of the decline in cognitive functioning based on scores -the instrument is also limited at determining whether cognitive decline is due to ID, dementia, or other reasons [63].
There is evidence that the Test for Severe Impairment (TSI) is reliable for monitoring the progression of dementia in people with severe ID [66].The TSI was used in three of the crosssectional studies including participants with DS and other ID [56,67,68].This instrument was developed to assess cognition in people with severe cognitive impairment, and most individuals with moderate/severe ID score on this test and only those with advanced dementia fail to score.
In addition to its use in cross-sectional studies, the TSI is reliable and valid in longitudinal studies as it monitors rates of changes and indicates a decline in cognitive function over time that can indicate dementia.In one of the earliest studies using the TSI, [66], the authors assessed the reliability and validity of the instruments in a sample of 60 adults with DS.They found that the convergent validity of the TSI for all samples was good (r=.94), with satisfactory interrater reliability (r=.97) and test-retest reliability (r=.98) over a two-year period.The instrument also showed good internal consistency, with a Cronbach's alpha of 0.89.

Other measures
Across most studies, the findings suggest that people with ID performed more poorly in verbal tasks, with significant declines with age [56,69,70,71].Phonological tasks are more likely to be sensitive to the detection of cognitive decline among individuals with DS compared to those with other ID, based on significant declines in these tasks [69,72].This is an important finding when considering which assessment should be used for those with DS and those with other ID.
Another important aspect of the screening instruments for dementia in ID is their ability to assess the behaviour changes commonly seen during the onset of dementia Another limitation is the lack of psychometric data for some of the instruments used.
Although we aimed to create a review to help clinicians and researchers to find the most suitable instrument, many studies did not provide psychometric properties based on their samples, and we considered it inappropriate to use secondary sources such tests and batteries handbooks.
As with any diagnostic assessment, we recommend following practical medical guidelines with multiple diagnostic approaches assessing cognitive, behavioural, and independent functioning.The use of informant and self-report instruments alongside medical examinations, neuroimaging techniques, and genetic and biological measures of various types of dementia is also recommended [76].

Conclusions
In conclusion, there is a multitude of instruments being used to screen for cognitive changes associated with dementia in those with ID.This review highlights the variation between measures used across studies and illustrates the need for unified, standardised measures to allow for the synthesis of results in research and greater consistency of diagnosis in clinical practice.Reliability scored 0.93. [

Figure 1 .
Figure 1.Preferred reporting items for systematic review and meta-analysis (PRISMA) flow chart concerning study retrieval and selection.
Contrasting cross-sectional and longitudinal studies, we recommend the use of specifically designed instruments such as the DLD [8] and the CAMCOG-DS [62] to assess cognitive functioning and behaviour changes related to ID and dementia.The use of measures designed for the general population should be avoided due to their lack of sensitivity in differentiating between those with and without dementia.
adults did not change over time, but that of younger adults with DS and adults without DS improved; Adults with DS showed significant and unique declines only in test of verbal fluency; diagnostic validity between the CAMDEX-DS and the CAMCOG-DS; with initial lower scores had lower levels of adaptive behaviour, were rated as more depressed and had a higher frequency of problem behaviours.had lower total scores that participants without DAT; Poor performance on the adaptation of CRT was associated with early-stage DAT.Male participants with ID no DS performed better than female participants with ID NDS; Females with DS performed better than males with ID in object assembly and block design; groups showed significant poorer performance in all tests, especially in verbal and working memory; MCI-DS showed poorer performance than control DS in the CAMCOG and DVM.

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 August 2020 doi:10.20944/preprints202008.0115.v1
Due to the prevalence of dementia in those with ID, particularly DS, it is important that researchers and clinicians have validated, reliable measures for diagnosis.Standardised measures

Table 1 .
Demographics of included individuals in the eligible studies

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 August 2020 doi:10.20944/preprints202008.0115.v1
One potentially promising new measure for assessing cognitive decline in those with other ID is the Wolfenbütteler Dementia Test for Individuals with Intellectual Disabilities (WDTIM).The WDTIM was used in a 2-year longitudinal study carried out byKuske et al. (2017) [46]and was effective at detecting cognitive changes over time.The authors noted that the WDTIM was more effective when used in conjunction with the Dementia Screening Questionnaire for

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 August 2020 doi:10.20944/preprints202008.0115.v1
Another measure was The Cambridge Cognitive Examination (CAMCOG).This was originally designed for use with the general population but was later adapted for the assessment of dementia in those with DS (CAMCOG-DS)[63].Cross-sectional studies have shown that this instrument can reliably differentiate between older and younger participants, is useful when [12].However, it has been noted that this measure may not be suitable for those with severe learning disabilities, severe sensory impairments, or advanced dementia due to floor effects[64].

5 August 2020 doi:10.20944/preprints202008.0115.v1 adaptation
of the AADS (AADS-I) that displays good psychometric properties and satisfactory interrater reliability for the six subscales (coefficients from 0.67 to 0.79).A further limitation is the lack of studies found in grey literature and open science databases -while only including papers from peer-reviewed journals helps to ensure the quality of included studies is high, it also limits a large amount of research which may provide additional insights.
behaviour changes in those withID and dementia [6].The ABDQ was used in two cross-sectional studies[74,75].LimitationsThere are some limitations to this review.There is a lack of findings from studies published in other languages.For instance, De Vreese et al. (2011) [57] carried out an Italian Preprints (www.preprints.org)| NOT PEER-REVIEWED | Posted:

Table A3 .
Longitudinal Studies assessing cognitive changes in mixed groups of ID participants