Mini‐Mental State Examination (MMSE) for the detection of Alzheimer’s dementia and other dementias in asymptomatic and previously clinically unevaluated people aged over 65 years in community and primary care populations
Abstract
This is a protocol for a Cochrane Review (Diagnostic test accuracy). The objectives are as follows:
To determine the diagnostic accuracy of Mini‐Mental State Examination (MMSE) at various thresholds for Alzheimers disease dementia (ADD) and other dementias in asymptomatic and previously clinically unevaluated people aged 65 years and over in community and primary care populations.
Background
This protocol is based on ‘Neuropsychological tests for the diagnosis of Alzheimer’s disease dementia and other dementias: a generic protocol for cross‐sectional and delayed‐verification studies’ (Davis 2013). The review forms part of a suite of reviews that will address the utility of different neuropsychological tests for the cross‐sectional and delayed‐verification diagnosis of dementia in a range of populations, for example The Montreal Cognitive Assessment for the diagnosis of Alzheimer’s disease and other dementia disorders (Davis 2013b); IQCODE for the diagnosis of Alzheimer’s disease dementia and other dementias within a general practice (primary care) setting (Quinn 2013); Mini‐Cog for the diagnosis of Alzheimer's disease dementia and other dementias within a community setting (Fage 2013) and Mini‐Mental State Examination (MMSE) for the detection of Alzheimer's disease and other dementias in people with mild cognitive impairment (MCI) (Arevalo‐Rodriguez 2013).
This review addresses the use of the Mini‐Mental State Examination (MMSE) for the cross‐sectional (current) diagnosis of Alzheimer’s dementia and other dementias when used in the community and primary care, which are populations with a relatively low prevalence of dementia compared to memory clinics and secondary or inpatient care. We include studies that examine the utility of the MMSE in previously unevaluated symptomatic people or asymptomatic people (akin to screening) because we seek to address the utility of the MMSE when applied to the clinically relevant question of patients and clinicians 'does this person have dementia now?'. The utility of the MMSE for delayed verification of dementia diagnosis at some future point (utility of MMSE for addressing the question 'is the current level of cognition sufficiently poor that this person has a pre‐dementia syndrome') is addressed by a separate review.
Target condition being diagnosed
Dementia is a progressive syndrome of global cognitive impairment that affects 6.5% of the UK population aged over 65 years (Matthews 2013). There is a significant global disease burden (36 million patients worldwide) that is predicted to increase to over 115 million by 2050, particularly in developing regions (Ferri 2005; Wimo 2010). Dementia encompasses a group of neurodegenerative disorders that are characterised by progressive loss of cognitive function and ability to perform activities of daily living that can be accompanied by neuropsychiatric symptoms and challenging behaviours of varying type and severity. The underlying pathology is usually degenerative and subtypes of dementia include Alzheimer’s disease dementia (ADD), vascular dementia, dementia with Lewy bodies (pathological clusters of alpha‐synuclein protein (McKeith 2005)), and frontotemporal dementia. There is considerable overlap in the clinical and pathological presentations, for example Alzheimer’s disease pathology can be found in people who have a clinical phenotype of vascular or Lewy body dementia, and vascular changes and Lewy bodies are commonly found in the post‐mortem examination of brains of people with an ADD phenotype (CFAS 2001; Matthews 2009; Savva 2009). Some commentators have advised against the use of neuropathological criteria as the 'gold standard' for the diagnosis of dementia, including subtypes (Scheltens 2011).
The target condition in this review will be dementia or its subtypes (as defined by the reference standards described below) identified at the same time as the index test is administered.
Index test(s)
The Folstein Mini‐Mental State Examination (MMSE) (Folstein 1975) is a 30 item assessment of cognitive function that assesses attention and orientation, memory, registration, recall, calculation, language and ability to draw a complex polygon. It takes around seven minutes to administer to a person with dementia and five minutes to a person with normal cognition (Borson 2000). The MMSE has recently been subject to copyright restrictions (de Silva 2010). The MMSE is scored out of 30 with a conventional cut‐off of 24, with scores less than this indicating cognitive impairment (Mitchell 2009); although other cut‐off points have been suggested (Crum 1993; Kukull 1994). There is a wide spectrum in the severity of disease that people with dementia have and this will affect the diagnostic properties of a diagnostic test such as the MMSE.
Clinical pathway
Dementia develops over several years, from a presumed initial asymptomatic period where pathological changes accumulate in the absence of clinical manifestations, through subtle impairments of recent memory or personality or behavioural changes, until the disease has become more apparent with multiple cognitive domains involved and difficulty in planning complex tasks.
Standard diagnostic practice
Standard diagnostic assessment relates to the evaluation of people where there is concern about possible dementia, particularly to exclude alternative diagnostic hypotheses, and includes history, clinical examination (including neurological, mental state and cognitive examination) and an interview with a relative or other informant. Before diagnosing dementia, other physical and mental disorders, for example hypothyroidism or depression, that might be contributing to cognitive impairment should be identified and if possible treated. A neuroradiological examination (computed tomography (CT) or magnetic resonance imaging (MRI) scan) of the brain is also recommended in most recent guidelines to exclude structural causes for the clinical phenotype, for example a subdural haematoma (McKhann 2011; NICE 2006.), but sometimes the diagnosis is made on the history and presentation alone. Dementia diagnosis is defined by a deficit in more than two cognitive domains of sufficient degree to impair functional activities. These symptoms are usually progressive over a period of at least several months and should not be attributable to any other brain disorder. The International Classification of Diseases (ICD)‐10 (World Health Organisation 1992) and Diagnostic and Statistical Manual of Mental Disorders (DSM) (American Psychological Association 1994) general diagnostic criteria for dementia are detailed in Appendix 1.
Screening
Screening is the identification of unrecognised or asymptomatic disease by the administration of tests that can be applied quickly, and is not intended to be diagnostic (Porta 2008). Recent UK health policy has encouraged opportunistic testing of older primary care attenders who have presented for reasons other than a memory complaint (Brunet 2012; Le Couteur 2013; Rasmussen 2013). General practitioners in the UK are incentivised to actively find people with dementia through routine annual questions in a Direct Enhanced Service (DOH recommendations; NICE 2013). In some cases, after further evaluation, the general practitioner may then make a diagnosis of dementia with or without a subtype (Ahmad 2010). Dementia screening is not recommended by the United States Preventative Services Task Force (US Preventative Services) but the Patient Protection and Affordable Care Act requires an annual assessment of cognition for people who are enrolled in Medicare (Cordell 2013). The UK government has also incentivised case finding for dementia on acute admission to secondary care services (Dementia CQUIN).
Because people with dementia may not experience subjective memory problems, and a diagnosis of dementia often requires diagnostic evaluation by an experienced clinician, triage tests such as the MMSE are used in clinical practice to help rapidly identify people with a high likelihood of having normal cognition who do not require onward referral and investigation. Some investigations have blurred the distinction between the use of tests as a screening instrument in individuals without manifest disease and their use as clinical triage tools (Kamenski 2009).
Clinical pathway
In the UK, people with memory problems usually initially present to their primary care practitioner who may administer the MMSE to 'rule in', or confirm, the possibility of dementia and will then potentially refer them to a specialist hospital memory clinic. Some people with dementia do not present until much later in the disorder or will follow a different pathway to diagnosis, for example being identified during an admission to general hospital for a physical illness. Diagnostic assessment pathways may vary in other countries and diagnoses may be made by a variety of clinicians including neurologists, psychiatrists and geriatricians.
How might the index test improve diagnoses, treatments and patient outcomes?
Many countries in Europe and worldwide have, or are developing, dementia strategies which emphasise the importance of accurate diagnosis to access appropriate health and social care services. Despite copyright restrictions, current experience is that the MMSE is still used extensively in clinical practice, including in the primary care setting, where it may be used as either a screening test for dementia or as part of a more detailed evaluation of a person for whom there is concern about dementia. A systematic evaluation of the diagnostic test accuracy of the instrument is needed to determine what confidence patients and clinicians can have in the clinical diagnosis of dementia based on the MMSE. A confirmed diagnosis of dementia is believed to offer opportunities for interventions, both social and medical, which may reduce the associated behavioural and psychiatric symptoms of dementia (Birks 2006; Clare 2003; McShane 2006) and help people with dementia and their families and potential carers to plan and avoid admissions to hospital or institutional care (Bourne 2007).
Prior tests
It is likely that no prior tests will have been performed as this review investigates the diagnostic utility of the test in people who have not been evaluated clinically for dementia. In some settings a two‐stage screening and assessment process takes place. Screening of people with suspected dementia usually requires a brief test of cognitive function or informant questionnaires, or both, and a low score indicates those people who require more in‐depth assessment (Boustani 2003). Some studies that have been carried out in the community and in primary care may have administered a very brief high sensitivity instrument before application of the MMSE, and then investigated all of those who screened positive and a subsample of those who screened negative. In this case we would include the study and consider the prior test as a potential source of heterogeneity.
Rationale
Policy for dementia diagnosis is developing rapidly and has changed since the publication of the generic protocol for neuropsychological tests (Davis 2013). There is a great need for a systematic appraisal of the diagnostic utility of neuropsychological tests, including the MMSE, in unselected (clinically unevaluated) populations.
Objectives
To determine the diagnostic accuracy of Mini‐Mental State Examination (MMSE) at various thresholds for Alzheimers disease dementia (ADD) and other dementias in asymptomatic and previously clinically unevaluated people aged 65 years and over in community and primary care populations.
Secondary objectives
To investigate the heterogeneity of test accuracy in the included studies.
Methods
Criteria for considering studies for this review
Types of studies
The inclusion criteria for studies in this review are based on the generic protocol for neuropsychological tests in dementia (Davis 2013). This review will evaluate the diagnostic accuracy of the MMSE when used in people aged over 65 years in non‐specialist settings, including community settings (population‐based screening) and primary care settings (where people may be screened opportunistically or may present to the primary care practitioner with memory problems). We will include studies that examine the diagnostic utility of the MMSE in people considered to have a memory problem (by patient, informant or clinician), as well as studies that examine the diagnostic utility in people regardless of a memory complaint (screening studies). These will be analysed separately as described in Investigations of heterogeneity.
We will include a diagnosis of dementia at any stage of disease (as long as the dementia has not been previously identified by a specialist) as this pragmatic approach is likely to be most useful in informing current health policy and clinical practice. For the purposes of this review we will not examine the utility of MMSE for the diagnosis of pre‐clinical dementia (Sperling 2011) as this will be the subject of a separate review.
We will include studies that are cross‐sectional in design, administering the index test and the reference standard(s) within a short time span (less than six months). These will include cohort studies and nested case‐control studies (nested within cohort studies) but will exclude case‐control studies as they are likely to provide a biased estimate in the meta‐analyses (Whiting 2013). We will not include delayed verification studies as these will be examined in a separate review, as described in the Background.
Figure 1 outlines the process for including articles in the review; further details are given in the selection of studies section.
Inclusion of studies
Examples of applicable studies for this review
-
Participants selected regardless of suspicion of cognitive disorder. This would be an unselected cross‐sectional survey that administers the MMSE to all participants and then evaluates all participants (or a random sample), regardless of MMSE result, with a full assessment for the presence or absence of the target disorder. These studies are analogous to screening and could be conducted in:
-
-
the community;
-
people attending primary care, as defined in the Participants section, though we consider these studies will be uncommon as they would involve screening people for cognitive disorder in a doctor's office waiting room and the ethics of this are not established.
-
-
Participants selected as suspected of having a cognitive disorder. These studies could be conducted in:
-
the community, though we anticipate these studies will be uncommon as many people who are suspected of having a cognitive disorder will seek evaluation in a healthcare setting;
-
primary care.
-
Therefore, we anticipate that we may find studies that use the MMSE in two different ways (evaluating people with and without suspicion of cognitive disorder) in two different clinical settings (community, before seeking diagnostic evaluation; and primary care, at the point of seeking diagnostic evaluation). We expect the diagnostic accuracy of the MMSE to differ between studies based in community and in primary care populations (and particularly in primary care participants selected on the basis of memory symptoms) and will conduct separate analyses in these four potential study populations.
Participants
We will include all participants who meet the criteria for inclusion in primary care or community‐based studies described above. We define primary care as non‐specialist, office‐based care with a first‐contact healthcare provider.
Studies where the MMSE is administered in a secondary care population, for example an emergency department, neurology ward or memory clinic, will be excluded.
We will exclude studies of participants with current, or a history of, alcohol or drug abuse, central nervous system (CNS) trauma (for example subdural haematoma), tumour or infection. Similarly, studies that only recruit participants on the basis of disease state (for example Parkinson disease, multiple sclerosis, brain injury, motor neurone disease) or residence (for example residential home, nursing home, prison) will be excluded as they are not applicable to the general population. Studies that recruit participants conditional on these criteria will have a different prevalence of dementia to the general population, and therefore the test accuracy is likely to be not applicable. We will include studies that recruit participants from retirement communities (defined as non‐residential non‐nursing elderly person independent facilitated living communities) as residents in these settings are likely to be similar to the population regarding cognition and comorbidities.
Our intention is that the findings of this review will have relevance to clinicians in primary care and community health, and be applicable to people with 'usual dementia' (Brayne 2012) and, therefore, we will exclude studies that investigate specific clinical groups. For example participants with a family history of Alzheimer’s dementia may be more readily diagnosed with dementia, perhaps leading to verification bias. On this basis, we will exclude studies which exclusively investigate those with a known genetic predisposition. We will also exclude studies specifically investigating early‐onset dementia.
Index tests
The index test is the 30‐item (original) version of the MMSE (Folstein 1975). We recognise that there are other versions of the test (including Grace 1995; Harrell 2000; Haubois 2012; Kabir 2000; Molloy 1991; Tschanz 2002) but these are best investigated in separate studies because we anticipate substantial heterogeneity in the diagnostic test performance of these instruments, and their inclusion would create an unfeasible workload for this review. However, we will not exclude studies on the basis of language, that is studies that report, for example, the diagnostic test accuracy of the Korean version of the 30‐item MMSE.
Target conditions
The target condition is all‐cause dementia and any dementia subtype. We expect to find studies that focus on all‐cause dementia, ADD, vascular dementia, Lewy body dementia and frontotemporal dementia. We will appraise findings separately if studies examining dementias of differing aetiologies or differing stages have been extracted and are included in the review.
Reference standards
In this review the target condition will be dementia or its subtypes as defined by the clinical reference standards described in the generic protocol (Davis 2013) and outlined below. We will exclude from this review studies that use neuropathological criteria as the only reference standard as this review seeks to determine a cross‐sectional diagnosis of dementia.
Clinical diagnosis will include all‐cause (unspecified) dementia, as defined by the DSM, Fourth Edition (American Psychological Association 1994); ICD, 10th edition (World Health Organisation 1992) (see Appendix 1); and the Clinical Dementia Rating.
In studies in which a reference standard refers to different criteria for dementia (for example McKhann 1984: unlikely, possible, probable, definite), we will consider people as having the disease if they are classified as having either probable or definite Alzheimer's dementia.
Alzheimer's dementia
The National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association (NINCDS‐ADRDA) (McKhann 1984) have put forth the best antemortem clinical consensus ‘reference standard’ for ADD, defining three ante‐mortem groups: probable, possible, and unlikely Alzheimer’s dementia. Newer criteria for ADD introduced in 2011 include the use of biomarkers (such as brain imaging and cerebrospinal fluid analysis) to contribute to the diagnostic categories (McKhann 2011). However, we will present in a separate category any studies that use these criteria and will test the findings in a sensitivity analysis.
Lewy body dementia
The reference standard for Lewy body dementia is the revised or earlier McKeith criteria (McKeith 1996; McKeith 2005).
Frontotemporal dementia
The reference standard for frontotemporal dementia is the Lund criteria (Lund 1994).
Vascular dementia
The reference standard for vascular dementia is the National Institute of Neurological and Communicative Disorders and Stroke and the Association Internationale pour la Recherché et l'Enseignement en Neurosciences (NINCDS‐AIREN) criteria (Román 1993).
We recognise that different iterations of reference standards over time may not be directly comparable (for example DSM‐III‐R versus DSM‐IV, ICD‐9 versus ICD‐10) and that the validity of diagnoses may vary with the degree or manner in which the criteria have been operationalised (for example individual clinician versus algorithm versus consensus determination). Data on the method and application of the reference standard will be collected and, if considered to be a source of bias, will be examined as a source of heterogeneity. Although it is unlikely that a specific reference standard might favour particular index tests there is the more general issue of incorporation bias, in which the reference standard is applied with knowledge of the index test, because neuropsychological deficits are integral to the definition of dementia. This is less problematic in cross‐sectional studies because the index test and the reference standard may be administered completely independently.
Search methods for identification of studies
Electronic searches
We will search MEDLINE (OvidSP), EMBASE (OvidSP), BIOSIS previews (Web of Knowledge), Science Citation Index (ISI Web of Knowledge), PsycINFO (OvidSP) and LILACS (BIREME). See Appendix 4 for a proposed draft strategy to be run in MEDLINE (OvidSP). We will design similarly structured search strategies using search terms appropriate for each database. Where appropriate, we will use controlled vocabulary such as MeSH terms and EMTREE. There will be no attempt to restrict studies based on the sampling frame or setting in the searches developed. We will not use search filters (collections of terms aimed at reducing the number that need to be screened) as an overall limiter because those published have not proved sensitive enough (Whiting 2011). We will not apply any language restriction to the electronic searches; we will use translation services as necessary.
A single researcher with extensive experience of systematic reviews will perform the initial searches.
Searching other resources
We will check the reference lists of all relevant papers for additional studies. We will also search the:
-
Meta‐analyses van Diagnostisch Onderzoek (MEDION database) (www.york.ac.uk/inst/crd/crddatabases.html),
-
Database of Abstracts of Reviews of Effects (DARE) (www.york.ac.uk/inst/crd/crddatabases.html),
-
Health Technology Assessments Database (HTA Database) in The Cochrane Library,
-
Aggressive Research Intelligence Facility (ARIF database) (www.arif.bham.ac.uk).
Relevant studies will be used in PubMed to search for additional studies by applying the related articles feature. Key studies will be examined in citation databases such as the Science Citation Index and Scopus to ascertain any further relevant studies. Some grey literature will be identified through Web of Science Conference Proceedings. We will aim to access theses or Doctor of Philosophy (PhD) abstracts from institutions known to be involved in prospective dementia studies. We will also attempt to contact researchers involved in studies with possibly relevant but unpublished data. We will not perform handsearching as there is little published evidence of the benefits of handsearching for reports of diagnostic test accuracy (DTA) studies at present (Glanville 2010).
Data collection and analysis
Selection of studies
Figure 1 shows a flowchart for inclusion of studies in the review. Preliminary searches indicate a substantial number of citations for the MMSE. Because of the need for timeliness in conducting this Cochrane review we propose the following inclusion and exclusion criteria.
The inclusion criteria are:
-
population is either community or primary care (see 'Types of studies');
-
reference standards as described above;
-
MMSE is used as an index test (may be as one of many tests) and is administered to all study participants;
-
study design is cohort or nested case‐control.
The exclusion criteria are:
-
classic case‐control study design (subject to spectrum bias);
-
index test administered to only cases or controls, rather than to both groups (cannot calculate diagnostic utility).
We will select studies from title and abstract screening undertaken by a team of trained assessors. Two assessors will independently review all citations retrieved by the search strategy and will classify them as either definitely relevant (study designed as a diagnostic test accuracy study), possibly relevant (study might report data that could be used to calculate diagnostic test accuracy), or irrelevant. Pairs of authors will then assess full text papers and we will resolve any disagreements by discussion with a senior author. We will detail the study selection process in a PRISMA flow diagram.
Data extraction and management
We will extract data on study characteristics to a study specific pro forma, independently by two authors, and will include data for the assessment of quality and for the investigation of heterogeneity. The pro forma will have been piloted against 10 primary diagnostic studies.
We will dichotomise the results if necessary and cross‐tabulate them in two by two tables of index test result (positive or negative) against target disorder (positive or negative). We will also extract the results directly into RevMan tables.
Assessment of methodological quality
We will asess the methodological quality of each study using the QUADAS 2 tool (Whiting 2011) as recommended by The Cochrane Collaboration (see Appendix 4 for QUADAS 2 statements and Appendix 5 for anchoring statements). The assessment of quality table will help the reader to evaluate the strength of evidence to support the diagnostic accuracy of the MMSE.
Statistical analysis and data synthesis
We expect the diagnostic accuracy of the MMSE to differ between studies based in community and in primary care settings (and particularly in primary care participants selected on the basis of memory symptoms). We will conduct separate analyses in the five potential study populations (see 'Types of studies' above).
The target condition comprises two categories: (1) dementia (not otherwise specified) and (2) dementia subtypes (Alzheimer’s, vascular, Lewy body, frontotemporal). Studies may detail one or more outcomes, however we expect most studies to focus on dementia (not otherwise specified) or ADD, or both. Each of the target conditions will merit a separate meta‐analysis in each of the target populations if there are sufficient data. We intend to conduct a primary analysis in which we propose combining all subgroups of dementia: the two by two table disease positives will be any dementia, disease negatives will be absence of any dementia, and test positives will be persons scoring below the traditional cut‐point on the original MMSE, as defined in the Index tests section. The rationale for this is that in primary care it is often of most relevance to the patient and clinician to know whether the person has any dementia syndrome, rather than a particular dementia sub‐type. For subgroups, if two or more studies report the diagnostic accuracy of the MMSE for that subgroup at the same cut‐point we will construct two by two tables with disease positives being for that subgroup only, disease negatives being all other persons, and test positives being people scoring below the specified cut‐point.
For all included studies the data in the two by two tables (showing the binary test results cross‐classified with the binary reference standard) will be used to calculate the sensitivities and specificities, with their 95% confidence intervals. We will present individual study results graphically by plotting estimates of sensitivities and specificities in both a forest plot and in the receiver operating characteristic (ROC) space. We will use RevMan software for descriptive analyses and to produce summary ROC curves. If more than one threshold is reported in an individual study, then we will present the graphical findings for all thresholds reported.
We will perform meta‐analyses on pairs of sensitivity and specificity if it is appropriate to pool the data. If the majority of the studies report results with consistent thresholds we will use the bivariate random‐effects model approach based on pairs of sensitivity and specificity (Cochrane DTA handbook chapter 10; Reitsma 2005). If different thresholds are reported, we will use hierarchical summary ROC models (Cochrane DTA handbook chapter 10). We will avoid data from a single study being included in the calculation of a summary statistic of sensitivity and specificity on more than one occasion (in the same population) by using only the threshold which is considered to be 'standard practice' for the target population in question. If there is no agreed standard practice for the target population in question then we will use the optimal threshold (the threshold nearest to the upper left corner of the ROC curve) in the calculation of the summary ROC curve in RevMan and for any subsequent meta‐analysis; we recognise that this may lead to an overestimate of diagnostic accuracy (Leeflang 2008). We will assess the model fit by using likelihood ratio tests.
We will use Stata software, version 13 (StataCorp, Texas) to carry out the additional analyses using the bivariate approach; for the hierarchical summary receiver operating characteristic (HSROC) approaches we will use either R, WinBUGS or Statistical Analysis Software (SAS).
Investigations of heterogeneity
There are many potential sources of heterogeneity in this review, which are fully detailed in Davis 2013. In this review we consider that the most important sources of heterogeneity are the characteristics of the study populations, how the MMSE was used, and the reference standard employed.
We will investigate heterogeneity in the first instance (informally) through visual examination of forest plots of sensitivities and specificities and through visual examination of the ROC plot of the raw data. Where there is evidence of heterogeneity we will attempt to adjust for these variables in the models through subgroup analyses as we anticipate that study characteristics will be inadequately operationalised to allow inclusion in the hierarchical regression model. We anticipate that sources of heterogeneity that are sufficiently well operationalised for us to investigate are likely to be investigated in categories only (for example we would classify age as: average age category (65 to 69 years; 70 to 74 years; 75 to 79 years; 80 to 84 years; over 85 years). We recognise that it is likely there will be insufficient power to formally investigate all possible sources of heterogeneity.
Potential sources of heterogeneity include the following.
Target population
In addition to the four potential study populations (for which all analyses will be performed separately), differences may occur in studies performed in different countries where the background levels of education and prevalence of disease may affect the performance of the MMSE. We will also evaluate the effects of differences in age and sex distribution in the study populations.
Index test
We will examine differences in MMSE thresholds used in the included studies. We will also analyse use of different language versions as a source of variation. The characteristics of the examiner may result in differences in test application and outcome, such as whether there was training prior to the test, and the professional background of the examiner. Other issues, such as the scoring (for example serial 7's or WORLD spelt backwards) may be important aspects of heterogeneity but may be poorly reported. Though likely to be important, training required for test administration may not be well reported or easily operationalised.
Reference standard
If data are available we will conduct separate analyses for studies examining the different subtypes of dementia. We expect most studies to examine either all‐cause (non‐specific) dementia and ADD as the outcome; if these studies use different reference standards for the target disorder we will examine this as a potential source of heterogeneity.
Sensitivity analyses
For the main analysis we will include all studies regardless of the risk of bias as assessed by QUADAS‐2, but we will perform sensitivity analysis to determine the effect of excluding studies that are deemed to be at high risk of bias. We will explicitly consider incorporation bias in studies where examiners used the MMSE as part of their diagnostic assessment of study participants (which is possible if the study was not designed as a DTA study, for example in a dementia prevalence study).
Assessment of reporting bias
Quantitative methods for exploring reporting bias are not well established for studies of DTA (Cochrane 2013)
