Assessment and management pathways of older adults with mild cognitive impairment: descriptive review and critical interpretive synthesis

Article history

The research reported here is the product of an HSDR Evidence Synthesis Centre, contracted to provide rapid evidence syntheses on issues of relevance to the health service, and to inform future HSDR calls for new research around identified gaps in evidence. Other reviews by the Evidence Synthesis Centres are also available in the HSDR journal.

The research reported in this issue of the journal was funded by the HSDR programme or one of its preceding programmes as project number NIHR133541. The contractual start date was in November 2020. The final report began editorial review in July 2021 and was accepted for publication in October 2021. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HSDR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2022 Chambers et al. This work was produced by Chambers et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2022 Chambers et al.

Patient and public involvement

The Sheffield HSDR Evidence Synthesis Centre’s strategic Public Advisory Group were involved from the outset of the review. The group commented on the review scope, including whether or not the research questions were the right ones, what they thought were the most important outcomes, if there was anything else they thought should be covered and what might be the priority areas for the public, before the protocol was finalised. A topic-specific patient and public involvement group was set up with the assistance of members of the Strategic Advisory Group. This group provided further input on their experience of services for people with MCI and the advantages and disadvantages of MCI as a diagnostic label. Near the end of the review, there was a second meeting where the patient and public involvement group commented on the review findings and were involved in writing the Plain English summary. A further meeting was held to discuss dissemination of the review findings to patients and the public. Approximately 15 members of the public were involved, including people with MCI, carers and those without direct experience of the condition.

Review questions

This report addresses the following questions:

• What is the evidence base around the assessment and management pathway of older adults with MCI in acute hospital wards, community/primary care and residential settings? In particular –

  • How are older adults presenting with memory problems investigated to understand the underlying cause of impairment?

  • What are the advantages and disadvantages of a ‘diagnosis’ of MCI? (We will aim to address both patient and health/social care provider perspectives.)

  • What is known about the experience of health and care services from the perspective of people with memory problems and their support networks (e.g. family, friends and other carers)?

The report comprises two separate evidence reviews: (1) a descriptive review with narrative synthesis, focusing on diagnosis, service provision and patient experience; and (2) a critical interpretive synthesis (CIS) of evidence on the advantages and disadvantages of MCI as a diagnostic label. The first subquestion is addressed in Chapters 7 and 8, the second subquestion is introduced in Chapter 6 and is the main focus of review 2 (see Chapters 11–13) and the third subquestion is covered in Chapter 9 and review 2.

Identification of evidence

A broad literature search was developed to identify research on the assessment and management pathways of older adults with MCI. The search strategy developed by the information specialist combined thesaurus and free-text terms and relevant synonyms for the population (e.g. older adults with memory problems with or without a diagnosis of MCI), intervention (e.g. screening and assessment tools, management pathways and service modules) and uses. The search terms were then combined using Boolean operators appropriately.

The search strategy was developed on MEDLINE and then translated to the other major medical and health-related bibliographic databases. The MEDLINE search strategy is presented in Appendix 1.

The search was limited to research published in English between 2010 and 2020. The date range was chosen to reflect the introduction of the UK National Dementia Strategy in 2009. 2 Earlier publications were incorporated by including relevant literature and systematic reviews. A search filter for UK studies was applied to the search to ensure that retrieved studies were relevant to the UK context. In addition, editorials, comments and letters were excluded where database functionality allowed.

The search was conducted in January 2021 on the following databases: MEDLINE, EMBASE, PsycInfo^®, Scopus, Cumulative Index to Nursing and Allied Health Literature, The Cochrane Library (i.e. Cochrane Database of Systematic Reviews and Cochrane Central Register of Controlled Trials), Science Citation Index and Social Science Citation Index.

In addition, grey literature searches were performed to retrieve clinical guidelines, policy documents and reports related to MCI from relevant websites. A full list of the websites searched is provided in Appendix 2. Reference and citation searching of included studies and relevant existing reviews were conducted for areas where more evidence was needed.

Inclusion and exclusion criteria

Study selection

Search results were downloaded to a reference management system (EndNote X9.2, Clarivate Analytics, Philadelphia, PA, USA) and duplicates removed. Unique references were imported into EPPI-Reviewer 4 (Evidence for Policy and Practice Information and Co-ordinating Centre, University of London, London, UK) systematic review software for screening and analysis. Titles/abstracts of imported references were screened against the inclusion criteria by four members of the review team, with any queries resolved by discussion among the review team. A 10% sample of excluded references were checked by one of the reviewers to ensure consistency and guard against premature exclusion. References that appeared potentially relevant were screened as full-text documents for a final decision on inclusion or exclusion, with any uncertainties resolved by discussion among the review team.

Data extraction and quality assessment

Key data were extracted and tabulated from the included studies, including study type, area of study, population, setting, study methods, findings, conclusions and key limitations. For the CIS, data extraction included positioning in argument, cited affiliations, study methods and CIS themes. Data extraction was undertaken using the coding and reporting functions of EPPI-Reviewer 4. Data extraction was performed by the four reviewers (DC, AB, AC and KS) and a 20% sample of each other’s work was checked.

Quality (risk-of-bias) assessment was undertaken for all studies that used a recognised design for which an appropriate quality assessment tool is available. Quality assessment tools used in this review included the Joanna Briggs Institute checklist for quasi-experimental studies, the CASP (Critical Appraisal Skills Programme) tool for qualitative studies, AMSTAR (A MeaSurement Tool to Assess systematic Reviews) for systematic reviews, the Swedish Agency for Health Technology Assessment tool to assess methodological limitations of qualitative evidence synthesis, and risk of bias for cohort/cross-sectional studies and diagnostic studies from the National Heart, Lung and Blood Institute (Bethesda, MD, USA) and Cochrane Collaboration, respectively. Quality assessment was performed by the four reviewers (DC, AB, KS and AC), who also checked a 20% sample of each other’s work. Assessment of the overall strength (quality and relevance) of evidence for each research question formed part of the narrative synthesis.

Synthesis of evidence

A narrative synthesis of the evidence was undertaken based on the predefined research questions and includes textual and tabular summary and critique of the included studies. Quantitative and qualitative evidence will be synthesised using methods based on the principles of CIS. 6 Briefly, CIS is a synthesis approach designed to analyse a broad range of relevant sources and use analytical outputs to develop a conceptual framework. We planned to use a variant that mobilises the literature to construct two alternative conceptual frameworks (i.e. one that assumes a pivotal role for the establishment of a definitive diagnosis of MCI and one that progresses a management pathway in the absence of a definitive diagnosis).

We have chosen a CIS methodology given its acknowledged strengths as a form of systematic review that draws on both traditions of qualitative research inquiry and systematic review methodology. A CIS is best suited to study a phenomenon that emerges over time and that constitutes a challenge to define, as is the case for MCI. In contrast to conventional systematic reviews, in which a precise question is tightly focused, CIS methodology offers the flexibility to draw from diverse relevant sources. Furthermore, CIS is not constrained to include only prespecified designs or quality of documents. Documents are selected according to relevance and their capacity to address the research question. Starting from an initial compass question relating to the assessment and management of older adults with MCI, two alternative management pathways were created and iteratively modified and defined as the synthesis progressed. In particular, we explored the extent to which assignment of a defining diagnosis or label determined the management pathway and eventual outcome. Quantitative and qualitative empirical studies was classified by the extent to which they supported each management pathway or to which they shared a common ground between the alternative pathways. The effect of contextual factors and their influence on the likelihood that individuals will progress down one or the other pathway was also explored.

Features of included papers

There were various types of included publications from the 13 publications8–20 identified. Conceptual overview papers were identified as a result of searches. Ball et al. 8 explored underlying disorders in relation to functional cognitive disorder. Forlenza et al. 9 critically reviewed existing knowledge on the conceptual limits and clinical usefulness of the diagnosis of MCI and the neuropsychological assessment (including short- and long-term prognosis). Stewart10 presented a narrative review of literature relevant to the presentation and detection of MCI in primary care. Stephan et al. 11 conducted a systematic review of the neuropathological profile of MCI, synthesising 162 clinical studies. Blackburn et al. 12 compared MCI, SCD and functional memory disorder (FMD). Brayne and Kelly13 explored the desirability of early diagnosis. The discussion paper by Swallow et al. 20 explored the process of negotiating risk in constructing classification boundaries in the memory clinic.

Other sources contained conceptual explanations for MCI as part of their objectives. One type was an evidence briefing by the British Psychological Society (Leicester, UK). 14 Such publications distil current narratives around MCI, which influence practice. The authors note that MCI is not a simple entity and the foremost narrative is the complexity of the conceptual understanding that is required. 14 There is a continued need to provide an overview of the current evidence for MCI. Evidence is framed as the support that may be needed by those for whom problems are detected before a dementia may be diagnosed.

The remainder of the studies were empirical. Swallow et al. ’s qualitative study20 examined practitioners’ accounts of the complexity associated with constructing the boundaries around MCI, Alzheimer’s disease (AD) and age in the clinic. Pierce et al. 15 presented a qualitative discourse analysis study of seven people with MCI. A service evaluation conducted by Jenkins et al. 16 consisted of an e-mail-based questionnaire, containing a vignette of an individual presenting with subjective cognitive impairment (SCI), distributed to 112 memory clinics. Stephan et al. 17 conducted a quantitative study of the suitability of forms of diagnostic tests based on criteria for MCI to predict dementia risk in the clinic. Data were from the Medical Research Council Cognitive Function and Ageing Study. 21 The study by Guzman et al. 18 consisted of a cross-sectional study of 34 participants with MCI (47% female and 53% male, with a mean age of 76.4 years) and evaluated the relationships between cognitive impairment, illness perceptions and cognitive fusion and their effects on levels of distress and quality of life. Rodda et al. 19 conducted a survey of clinicians on the Royal College of Psychiatrist’s (London, UK) Old Age Psychiatry Register about their perspectives on MCI.

The concept of mild cognitive impairment over time

Forlenza et al. 9 described the emergence of the MCI concept in the USA from the first use of the term in New York University (New York, NY, USA) (Reisburg et al. 1988 cited in Forlenza et al. 9) to the work carried out at the Mayo Clinic (Rochester, MN, USA) in the 1990s to suggest additional criteria (Petersen 1999, Petersen 2001 and Winblad 2004 cited in Forlenza et al. 9). Following this work, the US National Institute on Aging (Bethesda, MD, USA) and the Alzheimer’s Association (Chicago, IL, USA) placed emphasis on biomarkers for early diagnosis.

The study by Stephan et al. 11 was conducted prior to standardisation of the definition of MCI and reporting of pathology and, therefore, barriers existed to the creation of an integrated picture of the clinical and neuropathological profile of MCI.

Blackburn et al. 12 explored memory loss and dementia. The authors12 identified considerable confusion about the most appropriate diagnostic labels for people with memory problems at this point in time.

Issues remain in defining MCI, most prominently the conceptualisation of MCI as a stage on a pathway to dementia lacking diagnostic specificity. 8,14 MCI is described within the context of dementia, whereby an asymptomatic phase is followed by an early symptomatic phase with subjective memory complaints or MCI. 13

Definitions of mild cognitive impairment and causes of impairment

Ball et al. 8 explored underlying disorders in relation to functional cognitive disorder. Current research, such as this, has identified that some MCI cases are due to non-neurodegenerative processes and Ball et al. 8 point out the weaknesses in conceptualising MCI as a stage on a pathway to dementia lacking diagnostic specificity. In addition, Ball et al. 8 emphasise that MCI is an aetiology-neutral description and includes patients with a wide range of underlying causes. Cognitive disorder was a way of defining MCI for some time. Forlenza et al. 9 defined the criteria for the MCI concept as a ‘[s]ubjective complaint of cognitive impairment with objective cognitive impairment adjusted for age. Normal general intellectual function. Intact basic and instrumental activities of daily living’. 9

Recommendations from the British Psychological Society evidence briefing14 also identified shifting definitions of MCI. One of the central diagnostic definitions surrounds the existence of a prodromal stage of dementia. The authors14 stressed that this does not necessarily imply that everything between normal ageing and dementia (currently labelled as MCI) should be considered prodromal dementia. A key principle that emerged is the distinct pathway of MCI because it is recognised that many people labelled with MCI do not go on to develop dementia.

The same briefing paper14 provided another example of the way that a more sophisticated understanding of MCI has developed through recognition of the flaws in accepted criteria for MCI (specifically the measurement of cognitive impairment tests and the inclusion of memory impairment, which many do not experience). The authors argued that MCI should not be defined by performance below a cut-off point, but by evidence of decline from previous performance. It was suggested that subjective memory complaints are not necessarily a precursor of MCI and may reflect other conditions, such as low mood. The final areas of concern in the criteria are the possible impact of MCI symptoms on daily living and potential difficulties in defining significant impact. 14

Mild cognitive impairment has been underpinned conceptually with reference to memory loss and dementia, leading to challenges in understanding MCI at a conceptual and diagnostic level. 12 Blackburn et al. 12 pointed out that there is considerable confusion about the most appropriate diagnostic labels for people with memory problems. Despite a defined MCI label, it is widely misunderstood. Blackburn et al. 12 focused on concepts of MCI, SCD and FMD and, therefore, locating MCI within the suite of memory complaints conditions. The findings of Blackburn et al. 12 suggest that more work is required to help clinicians differentiate between non-progressive subjective memory dysfunction and the early stages of progressive memory disorders, such as AD.

The construct of SCI has been linked to MCI, although also emerging in its own right. Jenkins et al. 16 argued that SCI may be a risk factor for both MCI and dementia. Variable responses to a questionnaire containing a vignette to describe the presentation of a patient with SCI suggested a lack of a clear concept of SCI and how to manage the condition. Jenkins et al. 16 provided an insight into the relative increased clarity for MCI:

What is also clear is that several years ago the concept of MCI was the topic of similar debate to that surrounding SCI today and that MCI is now a widely recognized clinical diagnosis.

Jenkins et al. 16

The neuropathological profile of MCI was presented by Stephan et al. 11 Findings from included studies indicated that MCI is neuropathologically complex and cannot be understood within a single framework. The pathological changes identified included plaque and tangle formation, vascular pathologies, neurochemical deficits, cellular injury, inflammation, oxidative stress, mitochondrial changes, changes in genomic activity, synaptic dysfunction, disturbed protein metabolism and disrupted metabolic homeostasis. Therefore, determining causation is problematic (i.e. which factors primarily drive neurodegeneration and dementia and which are secondary features of disease progression).

The application of mild cognitive impairment diagnostic concepts

Stewart10 wrote a review paper on the challenge of real-world detection and diagnosis. The paper’s central focus is the construct of MCI and controversy surrounding the way it was defined (i.e. Stewart10 echoed the point made by the British Psychological Association14 about diagnostic cut-off points for measuring cognitive decline). Stewart10 stated that MCI is both a relatively simple and an important construct. Its importance depends on the assumption that early diagnosis and intervention are likely to be beneficial. The MCI construct is controversial because it ‘involves the imposition of categorical entities on what is essentially a continuous and mostly gradual process of decline in people who do develop dementia, not to mention incorporating the heterogeneity of cognitive changes in people who do not have underlying neurodegeneration’. 10 Stewart10 discussed the diagnostic process as likely to be based on subjective memory complaints under current definitions, rather than screening at a population level (which is unlikely to be feasible or acceptable). Yet, subjective memory complaints are heterogeneous in their aetiology, poorly predict medium-term dementia risk and are unlikely to be reported to GPs. In the process of diagnosis, the review raised the issue of distinguishing MCI from dementia and the challenges of identifying declines in functional activities of patients. The paper10 highlighted the subjectivity of the judgement in aspects such as cultural expectations for functional activity levels, individuals with increased support who may have higher function as a result and individuals with existing physical conditions that can have an impact on function. Findings10 relating to identifying MCI in clinical practice stress the importance of understanding health-seeking behaviours.

Previous concepts have focused on areas such as clinical characteristics and predictors of dementia9 and identification of dementia risk through identification of MCI. 17 Forlenza et al. ’s9 critical review of the limits of the diagnostic concept for MCI argued that the diagnostic criteria are complex, requiring sophisticated assessments:

Thus, there is a need for development of assessment strategies that are cost-effective, easy to administer and that generate results that are easy to interpret, while maintaining good sensitivity and specificity to identify MCI cases.

Forlenza et al. 9

The study by Stephan et al. 17 showed that definitions are not suitable for identifying people with MCI and at high risk of dementia in the general population. However, the definitions were able to identify people at elevated risk (possibly suitable for ‘watchful waiting’), as well as those considered to be not at risk. According to Stephan et al. ,17 identification could be achieved by using a simple test, such as the Mini Mental State Examination (MMSE), and more complex MCI criteria were not required.

Narratives questioning the desirability of the diagnosis have gained increasing traction. The desirability of early diagnosis was addressed by Brayne and Kelly. 13 Brayne and Kelly13 set the concept of early diagnosis in a UK policy context and the Prime Minister’s Challenge on Dementia. 3 Although the paper by Brayne and Kelly13 focused on dementia, it has relevancy for the exploration of the concept of early diagnosis and framed MCI in the context of dementia. Brayne and Kelly13 described a process of an asymptomatic phase followed by an early symptomatic phase with subjective memory complaints or MCI. Crucially, the authors point out that this challenge called for improved dementia diagnosis rates and is based on assumptions of benefit to individuals and those who care for them. In addition, the paper13 questions what is meant by ‘early’. Dictionary definitions include ‘in good time’, ‘before the usual time’ and ‘prematurely’. ‘In good time’ suggests a time that is appropriate for an individual, ‘before the usual time’ suggests a process that may be beneficial or harmful and ‘prematurely’ suggests the possibility of harm. 13 This concept of timely diagnosis means disclosure of the diagnosis at the ‘right time for the individual with consideration of their preferences and unique circumstances’ (Watson et al. 2019 cited in Brayne and Kelly13). This interpretation surrounds help-seeking at the point of noticing symptoms and referral to specialist secondary care assessment, such as memory clinics. Therefore, a timely diagnosis is delivered at the point people are ready for and will benefit from it. However, this has to coincide with adequate services to account for benefits costs and harms.

Mild cognitive impairment concept relating to services

In a study by Rodda et al. ,19 almost all respondents (99%) were moderately or extremely familiar with MCI. One hundred and ten (24%) respondents thought that the concept was more useful for doctors, 47 (10%) felt that it was more useful for patients and 292 (65%) rated it as the same for both. Findings from the Rodda et al. study19 indicated that psychiatrists thought that it would be helpful for patients to have a name for their symptoms. However, the study demonstrated the contentiousness of the diagnosis construction and related practices at the time among this group of professionals. For example, some respondents did not consider MCI a helpful concept and a few wrote in the free-response section that they did not consider it to be a diagnosis.

The discussion paper by Swallow et al. 20 explored the process of negotiating risk in constructing classification boundaries in the memory clinic. Swallow et al. 20 framed the process of constituting the diagnostic boundaries of disease in the clinic as interpretive and relational. In addition to diagnosis as a process to sort the ‘real from the imagined’ (Jutel 2009 cited in Swallow et al. 20), it was also a space for contestation (Bowker and Star 2000, Jutel 2009 and 2015, Jutel and Nettleton 2011, and Rosenberg 2002, 2003 and 2006 cited in Swallow et al. 20). Swallow et al. 20 pointed to the complications of the diagnostic conceptualisation that categorises pathology at earlier stages and in which there is a blurring of boundaries. According to a sociology of diagnosis perspective, the discussion highlighted the way that the grey area of the diagnosis is a liminal and ‘ “uncomfortable” space for patients and practitioners, favouring more conventional categories’. 20 The ‘functionality’ of diagnosis (Jutel 2015 and Timmermans and Buchbinder 2010 cited in Swallow et al. 20) is, therefore, called into question if underpinned by uncertainty [including potential positive aspects, such as uncertainty serving to return the act of decision-making to the clinic (Latimer 2013 and Reed et al. 2016 cited in Swallow et al. 20)]. Swallow et al. 20 also incorporated wider themes, such as ageing and differing positive or negative sociocultural constructions of ageing, and loss of self and related conditions, such as AD or MCI (Beard and Neary 2013, Latimer 2018, and Swallow and Hillman 2018 cited in Swallow et al. 20). Swallow et al. 20 argued that these constructions may exist at some level within the practices of the memory clinic.

Stewart et al. 10 pointed out that most research in this area has focused on the applicability of cognitive assessments in primary care. However, in general, the detection of MCI in primary care is low and, therefore, detection may need to be expanded. The authors10 suggested the use of routine clinical data and nurse practitioner involvement.

Experience-related aspects of the mild cognitive impairment concept

Psychosocial adjustment is a conceptual term emphasised by Guzman et al. 18 and their cross-sectional study of standardised measures for cognitive assessment, illness perceptions, cognitive fusion, depression, anxiety and quality of life. Guzman et al. 18 outlined the features of a MCI diagnosis. Receiving a MCI diagnosis can evoke a broad range of emotional responses in people diagnosed with MCI, including worry, ambivalence or relief. 22,23 MCI is a vague term and makes the person confused as to whether or not they will go on to develop dementia. Some researchers argued that a MCI diagnosis merely causes undue distress for individuals and their caregivers about what may be part of a ‘normal’ ageing process. 24,25 Among their findings,24,25 the researchers showed that illness perceptions were a stronger predictor of depression and quality of life than cognitive impairment. However, limited research has focused on individual experiences of receiving this diagnosis. Recommendations24,25 include multiple treatment targets to help people adjust and accept a diagnosis, which is a factor in secondary conditions, such as depression, and in improving quality of life.

Discourses of patients with a MCI diagnosis were explored by Pierce et al. 15 A central narrative was ‘[k]nowingly not wanting to know’15 about the possibility MCI would develop into dementia. Another MCI discourse was characterised as not knowing about MCI. In addition, Pierce et al. 15 explained a third discourse: ‘[i]n the absence of a coherent discourse [of MCI], around a diagnosis given to them by experts in a memory clinic, participants turned to a more familiar discourse to help them ascertain their positioning – that of ambivalent ageing and certainty of death’. 15 Therefore, the conclusions of the study suggest that the conceptualisation of MCI by patients should be considered by clinicians in terms of how information is presented to people about MCI and, in particular, how MCI is positioned in respect to normal ageing and dementia.

Summary

The conceptual landscape of MCI has changed significantly over time, gaining its own diagnostic label. However, conceptual features of MCI remain challenging for diagnosis and patient acceptance. A conceptual understanding of MCI as purely a pre-dementia condition persists, with a reliance on categorical cut-off points to distinguish between normal ageing and MCI. In addition, conceptualisations include both those who go on to develop dementia as well as those with a wide range of cognitive problems but no underlying neurodegenerative decline. This produces a level of conceptual complexity in a single framework that locates MCI definitions in the context of dementia, leading to understandings that incorporate an absence of diagnostic boundaries with the neuropathological complexity of the condition. In practice, this means that diagnosis is subjective and there is a reliance on assessment of function and memory. Clinicians call for a more sophisticated assessment of function relative to the individual.

Help-seeking

We identified three studies30–32 of barriers to and facilitators of help-seeking for memory problems and two UK studies33,34 of interventions aimed at encouraging people with concerns to consult their GP (Table 3). A 2018 systematic review30 found that there is often a long delay between noticing symptoms and seeking help and that help-seeking is often precipitated by a ‘pivotal event’. In the absence of such an event, people may discount their symptoms as normal, reserve judgement about them or misattribute them. In a small study31 of Irish adults, facilitators of help-seeking were family, friends and peers, alongside well-informed health professionals. Barriers to seeking help were a lack of knowledge, fear, loss, stigma and inaccessible services. 31 Ethnic minority groups may encounter specific barriers to help-seeking. A qualitative study32 of South Asian people identified barriers, including not knowing what help is available, a perception of lack of time in GP consultations and a belief that ‘good families’ care for people with dementia themselves. Barriers to accessing early/timely diagnosis at the level of the health system are discussed below (see Recognition and recording in primary care).

Chan et al. 33 and Livingston et al. 34 reported on interventions in the English NHS to encourage people with memory concerns to consult their GP. Both studies33,34 involved distribution of information to patients at the level of the general practice. Chan et al. 33 found that recording and management of memory problems improved during the period of the study, but there was a greater improvement in a control area where no additional information was distributed. In the study by Livingston et al. ,34 consultations about memory problems increased in intervention practices compared with control practices. However, there was no increase in diagnoses or referrals to memory clinics. This limited evidence suggests that provision of information to patients may not be sufficient to overcome barriers to help-seeking and obtaining help for memory problems.

Recognition and recording in primary care

We included six reviews10,35–39 and 12 primary studies40–51 in this section.

As noted above, identification and investigation of memory problems generally begins when patients or family members are concerned about a person’s symptoms, leading to a consultation with the patient’s GP. The objective of the investigations is often presented as ‘early diagnosis’ of the underlying cause of the memory problems (i.e. dementia or other). 10 Brooker et al. 35 have developed the concept of a ‘timely diagnosis’ (of dementia), stating that ‘citizens should have access to accurate diagnosis at a time in the disease process when it can be of most benefit to them’. 35 The concept of timely diagnosis is discussed further in review 2 (see Chapter 11).

Two included studies40,41 identified barriers to early diagnosis in general or for disadvantaged groups. A survey of primary and secondary care physicians (n = 1365) found that barriers included patients seeing cognitive decline as a normal part of ageing and not disclosing symptoms, long waiting lists, and a lack of treatment options and definitive biomarker tests. 40 Although a relatively large sample, this survey reflected the perceptions of a self-selected group of physicians only. A mixed-methods study of homeless older people living in hostels reported that the memory assessment service (MAS) was difficult to access and not patient centred. 41

People who report memory concerns to their GP are likely to be assessed using ‘pencil and paper’ cognitive tests in the first instance (see Cognitive tests). Six included studies36,42–46 assessed recognition and recording of memory problems by GPs in the UK and the Netherlands. A meta-analysis from 201136 found that GPs have considerable difficulty identifying people with MCI or mild dementia using clinical judgement, and diagnoses are poorly recorded in medical records. 36 For MCI, GPs recognised 44.7% of cases and the diagnostic label was recorded for only 10.9%. This review was published in 2011 and so may not reflect current practice. 36

Given the limitations of clinical judgement, efforts have been made to promote the use of cognitive screening tests in primary care. In the UK, these include guidance from NICE and the Social Care Institute for Excellence and the National Dementia Strategy, published in 2006 and 2009, respectively. 2,42 A comparison of GP referrals to memory clinics before and after the launch of this guidance found that the number of referrals increased, but the rate of dementia diagnoses decreased and the use of cognitive screening tools did not change. 42 This suggests an absence of evidence for the effects of the guidance in promoting screening, at least in the short term. As an alternative or complement to screening, Olazarán et al. 43 investigated the role of data routinely collected by GPs in identifying MCI and predicting its course (i.e. progression to dementia or reversion to normal cognition). For example, old age, source of information about symptoms (informant or primary care physician), short duration and low education were associated with MCI or dementia at baseline.

The implications of failing to diagnose or record cases of MCI or dementia in primary care have been investigated in more recent research (three included studies44–46). A recent UK study,44 which used linked records, found that 34% of participants with known dementia had a primary care diagnosis in 2008–11 and 44% had a primary care diagnosis in 2011–13. In both periods, a further 21% of participants had a record of a concern or a referral, but no diagnosis. There was a lack of relationship between severity of non-memory symptoms and diagnosis, suggesting possible low awareness of such symptoms. In a study in the Netherlands,45 GPs showed variable awareness of the presence of cognitive impairment in their older patients (i.e. those aged ≥ 65 years) without a firm diagnosis of dementia. Lack of awareness of a problem was highest for the older age groups. Training GPs to diagnose MCI and dementia, followed by case finding, resulted in a non-significant increase in new MCI diagnoses when compared with control practices. 46

Although the common model in the UK is for people with memory problems to be identified in primary care and referred to a memory clinic for more detailed assessment, primary care-led services are available in some areas. Three included studies47–49 focused on screening and diagnosis, and the topic of different service models is discussed in more detail below (see Chapter 8). An evaluation of primary care-led services in Bristol found that, in practice, GPs rarely made independent diagnoses of dementia. 48 A need for specialist support with diagnosis was also identified in a study aimed at developing a multidisciplinary memory clinic in south-western Sydney, NSW, Australia. 49 Eye clinics have been suggested as an alternative setting for screening,47 but this approach does not appear to have been tested.

Finally, two scoping reviews37,38 have examined the potential use of telehealth technologies to support diagnosis of memory problems, especially in remote areas, as well as highlighting limitations of the available evidence. Two overlapping articles50,51 by the same author reviewed the role of mobile applications, but the review focused on features of the applications rather than evidence of their accuracy or value for diagnosis. A review39 of the international literature, included for completeness, provided few new data, but emphasised the lack of evidence from low- and middle-income countries, especially the former.

Screening in hospital settings

We identified only one study52 of screening for memory problems in a hospital setting. A Commissioning for Quality and Innovation payment led to improvements in practice, as demonstrated by audit of case notes in three wards at a large university teaching hospital in England. This study53 involved hospital inpatients aged > 75 years who were admitted as emergencies. The study was subsequently withdrawn.

Diagnostic terminology

Conceptual studies of MCI are summarised in Chapter 6 and the advantages and disadvantages of MCI as a ‘diagnostic label’ are the focus of review 2 (i.e. the CIS). We found evidence from two included studies (surveys)19,54 that indicated that clinicians regard MCI as a valuable diagnostic label, suggesting that any change to the labelling of patients may be difficult to achieve. A survey of UK psychiatrists (n = 453) found that only 4.4% of clinicians did not view MCI as a useful concept, although they showed less uniformity in their diagnostic practices. 19 The authors reported that psychiatrists saw MCI as a transitional stage between normal ageing and dementia. A more recent survey of members of the European Academy of Neurology (Vienna, Austria) and the European Alzheimer’s Disease Consortium (EADC) (Mannheim, Germany) (n = 102) reported that 92% of respondents used MCI as a diagnostic label in clinical practice. 54 A majority (68%) of respondents also used the labels prodromal AD or MCI due to AD. Over 70% of respondents thought that these labels had additional value compared with a label of MCI and influenced decisions about treatment, as well as communication with patients.

Patient/carer experience of diagnosis

This section deals specifically with experience of the process of diagnosis and labelling of memory problems. The broader topic of patient and carer experience of services is covered below (see Chapter 9), although some overlap is inevitable. This section is based on evidence from four reviews22,55–57 and six primary (mainly qualitative) studies. 1,15,24,58–60

An early scoping review55 of consent for the investigation and diagnosis of memory problems noted that early discovery of AD biomarkers in research or clinical settings is problematic because it requires patients to consent to disclosure of findings that indicate an uncertain risk of developing AD in the future. The authors of the review55 argue that expectation of future decline may lead to stigma, isolation and discrimination, but offer no solutions beyond the need for more nuanced analysis of the ethics issues involved. Closely linked to consent is the question of how patients and carers understand and describe the concept of MCI. Two studies15,58 specifically addressed this question. Roberts and Clare58 used a qualitative methodology (interpretive phenomenological analysis) to analyse transcripts of interviews with 25 people with a clinical diagnosis of MCI who had been informed of their diagnosis. Four higher-order themes were identified, but participants did not use the term MCI, suggesting that the term was not meaningful for them, although they did want a clear explanation for their memory problems. Pierce et al. 15 applied ‘discourse analysis’ to interviews with seven people diagnosed with MCI. One of the discourses identified was ‘not knowing’ about MCI. This was accompanied by ‘knowing about aging and death’ and ‘not wanting to know about dementia’. These two complementary studies set the scene for difficulties in patient experience of diagnosis and labelling.

Three included systematic reviews22,56,57 cover experience of the diagnostic pathway from diagnosis of MCI22 to investigation in memory clinics56 and diagnosis of dementia. 57 Dean and Wilcock22 noted a lack of studies on the effect of disclosing a diagnosis of MCI on patients and carers. Patients and carers receiving a diagnosis of MCI produced a range of negative and positive emotional responses, but there were no studies of experience of the diagnostic process. The review by Robinson et al. 56 included experiences of the diagnosis of both MCI and dementia, with similar findings to those of Dean and Wilcock. 22 Specifically, for MCI, a diagnosis could lead to feelings of exclusion, but could also reinforce a shared identity with others. Participants wished to be proactive in their response to a MCI diagnosis. 56 Bunn et al. 57 identified more concrete information on experience of the diagnostic process. Bunn et al. 57 reported that patients and carers perceived that, in some cases, doctors were slow to recognise symptoms or reluctant to give a diagnosis, and that even when people were referred to memory services the process could be slow, with long periods of waiting. 57

The four primary qualitative studies1,24,59,60 (Table 4) provided more detailed and richer data than those that were available from the reviews. The topic is discussed in more detail below (see Chapter 9) because the studies also include patient experience of services. Overall, the studies identified substantial difficulties with the diagnostic/labelling process from the viewpoint of patients and carers. These are well summarised by the titles of the studies, which refer to ‘making sense of nonsense’,24 ‘falling through the cracks’1 and ‘negotiating a labyrinth’. 60

Cognitive tests

Cognitive tests are the main tools available to GPs and other primary care clinicians to supplement clinical judgement in assessing people with memory problems. There is an extensive literature on the diagnostic accuracy of such tests, but this review focused on the role of cognitive tests in aiding decisions about further investigations and referral or treatment pathways.

Evidence from systematic reviews indicate that a number of comprehensive screening tests, such as the Addenbrooke’s Cognitive Examination – Revised, the Cambridge Cognitive Examination, the Montreal Cognitive Assessment (MoCA) and the Consortium to Establish a Registry for Alzheimer’s Disease, have > 80% sensitivity for distinguishing people with MCI from healthy volunteers. 61 However, the time and expertise required to administer some tests may limit access to them in primary care settings. 5 An international survey of physicians (n = 1365, 63% specialists) found that most reported using cognitive tests for early detection of MCI or dementia, with the MMSE being the most commonly used tool. 62

We included two systematic reviews63,64 and three primary studies65–67 of specific tests or types of test (Table 5). A 2018 systematic review63 of automated tests included 16 studies of 11 tools, but none was suitable for monitoring disease progression or response to treatment. A systematic review64 of the Ascertain Dementia 8 questionnaire concluded that that this was a suitable tool for use in busy primary care settings because, on average, it took < 3 minutes to administer.

Tests evaluated in single studies were ‘BrainCheck’, TYM (Test your Memory) and ‘CognoSpeak’ (a fully automated screening tool). Ehrensperger et al. 65 developed and evaluated ‘BrainCheck’, which was a brief screening instrument for use in primary care and combined different sources of data. ‘BrainCheck’ was rated highly feasible and acceptable by GPs and correctly classified 89.4% of cases (including MCI, probable AD, depression and healthy controls). The TYM test, which was designed to be self-administered under medical supervision, was evaluated in a memory clinic population. 66 TYM was considered easy to use and acceptable to patients (n = 224, 35% meeting criteria for dementia). Using a revised cut-off point, the TYM showed comparable diagnostic utility to the MMSE and the Addenbrooke’s Cognitive Examination – Revised for the differentiation of people with and without dementia. The authors66 suggested that self-administered tests, such as TYM, may be particularly useful where clinicians’ time is limited. In a pilot study involving 15 people with MCI, people with other conditions and healthy controls, the automated system achieved levels of accuracy comparable to those of current manually administered screening tools. 67 This suggests potential to improve the efficiency of screening processes in the future.

An important point about cognitive tests is that the prevalence of the condition of interest (e.g. MCI or dementia) depends on the test used and the selected cut-off point. In an early study involving people aged 70–90 years without dementia (n = 981), Kochan et al. 68 found that the prevalence of MCI ranged from 4% to 70%, depending on the criteria used. Klekociuk et al. 69 noted that existing MCI diagnostic criteria resulted in an unacceptably high rate of false-positive diagnoses of MCI. The authors69 reported that a combination of measures of complex sustained attention, semantic memory, working memory, episodic memory and selective attention correctly classified the outcome in > 80% of cases. The rate of false-positive diagnoses (5.93%) was lower than that reported in previously published MCI studies. The issue of ‘false-positive’ labelling of MCI (i.e. reversion to normal cognition) is discussed below. In practice, sophisticated combinations of tests, such as those used by Klekociuk et al. ,69 are unlikely to be available in normal clinical practice.

NHS London Clinical Networks guidance recommends the use of a validated brief screening test, such as MoCA, for the detection of MCI,70 but this appears to be in the context of a memory service assessment rather than in primary care.

Two papers by Sabbagh et al. 71,72 focus on early detection of MCI in different settings. The review,71 which was focused on primary care, identifies barriers to effective screening in primary care related to time constraints and test validation. Tests delivered in ≤ 10 minutes are unlikely to fully evaluate all dimensions of cognition. In addition, tests developed and tested on highly homogeneous and well-educated populations may perform less well in routine practice with more diverse populations. Factors such as level of education and proficiency in English may affect test performance, potentially reducing test scores and resulting in people being wrongly labelled with MCI.

A further paper72 by the same group of authors deals with early detection of MCI at home. The time constraints of primary care are less of an issue in this setting and a range of digital technologies are available as supplements or alternatives to conventional ‘pencil and paper’ tests. These technologies are less well validated than tests in clinical settings and other barriers include costs and concerns about reliability of the technology. However, the authors conclude that ‘passive’ technologies that work by monitoring user behaviour could ultimately provide a low-effort, easy-to-use solution for widespread cognitive performance monitoring. 72

We identified one economic evaluation73 of cognitive testing for MCI and dementia in primary care. The authors73 examined three cognitive tests and concluded that any of them could be considered cost-effective compared with unassisted clinical judgement by GPs. The most cost-effective option in the base case was the GP assessment of cognition. It should be noted that benefits from testing were due to early access to medication for dementia.

In summary, the time and expertise needed to administer cognitive tests is variable. Although much research has focused on diagnostic accuracy, quick tests suitable for use in primary care may effectively identify people who need further investigation for cognitive impairment.

Imaging

We included two narrative reviews and a systematic review, giving a broad overview of imaging (scanning) in the investigation of MCI and dementia, together with a study of service variation and a cost study (which also included biomarkers). Two included studies (one a systematic review) reported on the effects of amyloid imaging on diagnosis and treatment, whereas a further systematic review focused on the impact of disclosing amyloid imaging results to people without dementia. The Manchester consensus guidance5 also covers imaging and is more recent than the general reviews.

In their 2013 review, Burhan et al. 74 focused mainly on functional magnetic resonance imaging (MRI), magnetic resonance spectroscopy and diffusion tensor imaging. The authors74 concluded that, because of the limitations of current evidence, these technologies were not recommended for clinical investigation. In their 2012 review, Sullivan et al. 75 reached similar conclusions, noting that neuroimaging provided little additional benefit over cognitive testing in predicting progression from MCI to dementia. Sullivan et al. 75 did not recommend routine scanning in all patients with MCI or suspected dementia, stating that health professionals should use their clinical judgement. A 2018 systematic review76 evaluated the potential of machine learning combined with neuroimaging for the diagnosis of AD or MCI. The review76 included 111 studies and concluded that machine learning was most accurate for distinguishing people with AD from healthy controls, but poor for distinguishing MCI from normal cognition. Overall, machine learning was not considered sufficiently developed for use in routine practice.

Evidence from UK services suggests considerable variation in the use of imaging in the investigation of cognitive impairment. The 2019 London memory service audit77 reported that the percentage of patients deemed not to require a scan for dementia diagnosis varied from 6% to 46%. Of those patients who did have a scan, the percentage who received computerised tomography (CT) (rather than MRI) varied from 2% to 58%. The Manchester consensus guidance5 noted that clinicians who request neuroimaging may not have access to the original images and, instead, have to rely on written reports. CT may be used as an alternative to MRI because the latter is more expensive or not available (e.g. because of lack of access to equipment or the service is not commissioned locally).

Availability of positron emission tomography (PET) in the UK is also patchy. The Manchester consensus guidance5 noted that there were only 62 PET scanners in the country at the time of writing, with most being in university teaching hospitals or research centres. There is increasing evidence for the value of PET in the diagnosis and management of MCI and suspected dementia. In a cohort study78 of memory clinic patients [n = 507 patients, 114 (23%) with MCI], amyloid PET results were positive for 242 (48%) patients. The suspected aetiology changed for 125 (25%) patients after undergoing amyloid PET. The mean diagnostic confidence increased from 80% (standard deviation 13) to 89% (standard deviation 13). In 123 (24%) patients, there was a change in patient treatment after PET, mostly related to additional investigations and therapy. Fantoni et al. 79 performed a systematic review and combined data from seven studies. The authors79 concluded that amyloid PET contributed to diagnostic revision in almost one-third of cases and demonstrated value in increasing diagnostic confidence and refining management plans.

Disclosure of amyloid PET scanning results is controversial in view of the uncertain predictive value of the result for individuals and the lack of disease-modifying treatments for early dementia. A systematic review80 of empirical evidence and theoretical arguments for and against disclosure identified 51 different arguments. The authors80 concluded that widespread disclosure of results to people without dementia requires further research on the impact of disclosure, as well as the predictive value of the test. Communication materials and strategies to support disclosure should be developed and evaluated.

The importance of PET and biomarkers (see below) is expected to increase with the availability of disease-modifying therapies for dementia. Intensive investigation will be required to identify those patients who could benefit from such treatments (e.g. those with MCI) and evidence of underlying damage suggestive of early (or prodromal) dementia. Wittenberg et al. 81 attempted to estimate the potential future costs of expanding the number of PET and cerebrospinal fluid (CSF) tests in the UK. Based on a focused literature review and consultation with experts, annual costs of 100,000 extra amyloid PET scans and 100,000 extra CSF tests were estimated at £113M and £48M, respectively. These costs are likely to be higher in the first year because of the need for additional staff training. However, as noted by the authors,81 these costs are likely to be small compared with the acquisition costs of disease-modifying drugs and the impacts of inaccurate diagnoses. Ongoing economic evaluation will be needed to clarify the costs and benefits.

Biomarkers

Blood and CSF biomarkers are increasingly important as alternatives to imaging in investigating the underlying cause of cognitive impairment. Biomarkers are particularly important in investigating MCI when the underlying cause of the impairment is thought to be AD. 5

The review included four publications82–85 focused on the use of biomarkers (with the extensive clinical and diagnostic accuracy literature falling outside our scope). In line with the Manchester consensus guidance,5 a 2017 expert review82 recommends biomarker testing in people with MCI or dementia when there is diagnostic uncertainty and the result will affect diagnosis or treatment. A 2015 survey83 of EADC members revealed that the participating memory clinics were using biomarker testing for clinical as well as research purposes. Survey participants were specialist centres regarded as early adopters of new technologies and so the results may not be representative of routine practice. In addition, a further limitation is that only four of eight eligible centres in the UK participated in the study. 83

In 2020, the EADC published a survey on biomarker counselling, disclosure of diagnosis and follow-up in patients with MCI,84 and guidance on the same topic (jointly with the European Academy of Neurology). 85 The survey involved 34 centres across 20 countries. The majority of respondents had access to biomarker tests and found them useful. Arrangements for pre-test counselling and advice and referral after testing varied between centres. The study authors identified a need to improve counselling and training for clinicians to improve their communication skills. As with the previous EADC survey, this research involved specialist tertiary centres (only one from the UK) and so the findings may not be applicable to routine practice. The guidance85 emphasises the need to involve dementia specialists in counselling before biomarker testing and in disclosure of the findings. Follow-up is recommended for patients with MCI and should include advice on reducing dementia risk and, if applicable, treatment of specific underlying causes. The authors85 note that advice on advance directives may also be relevant.

Other tests

We included three studies86–88 of other tests that have been used to support investigation of the underlying cause in people with MCI. Alongside the cognitive impact of MCI, effects on gait and balance have been extensively evaluated. A 2017 systematic review86 (with 14 included studies) reported that MCI affects specific aspects of gait, particularly when patients are challenged to complete a cognitive task at the same time (i.e. dual-task conditions). The authors86 concluded that gait assessment with an additional cognitive task is useful for diagnosis and outcome analysis in the population of people being investigated for possible MCI. Interventions targeting specific components of gait could potentially slow progression to dementia.

The other two studies87,88 looked at tests intended for use in memory clinic populations. A clinical decision support tool (PredictND) was evaluated in 747 patients with MCI, FCD or dementia. 87 Compared with clinical assessment alone, the aetiological diagnosis changed in 13% of patients when the decision support tool was also used. Diagnostic accuracy (using follow-up diagnosis as the reference standard) did not change, but confidence in the diagnosis, measured on a visual analogue scale, increased by three percentage points. The authors87 concluded that clinical decision support tools could help clinicians in the differential diagnosis of dementia. Elsey et al. 88 explored the use of conversation analysis to distinguish people with FMDs from those with memory problems caused by dementia. In a small study (n = 30, including four patients with amnestic MCI), patients with functional disorders responded differently than those with dementia or MCI to questions about memory problems. The authors88 concluded that conversational profiles have the potential to assist with screening and referral in primary care and diagnosis in secondary care.

In summary, tests such as gait and conversation analysis may have the potential to aid investigation of memory problems, but the relative lack of studies suggests that they may have been supplanted by the rapid development of imaging and biomarker technologies.

Differential diagnosis

This section covers investigations to distinguish between MCI and two related conditions (i.e. SCD and FCD). SCD does not have a universally accepted definition, but can be distinguished from MCI on the basis that perceived memory problems or cognitive decline are not supported by the results of cognitive tests. 1 FCD is defined as persistent subjective cognitive difficulties that are not consistent with a recognised disease process and not supported by objective measures of cognitive functioning. 89 An additional feature is described as ‘internal inconsistency’ (i.e. an ability to perform a task well at some times but not others). 8

Two included studies8,89 focused on differential diagnosis of FCD and MCI. A recent expert review8 on FCD noted that early diagnosis of FCD can provide reassurance to patients compared with a diagnosis of MCI. A comparative study89 of 21 people diagnosed with FCD, 17 people diagnosed with MCI and 25 healthy controls found that both groups were impaired relative to controls, with the main difference being that the FCD group were younger. The authors89 concluded that clinicians need to be aware that FCD and MCI symptoms overlap and that the diagnosis can change over time. The potential for conversation analysis to identify people with FCD was noted above (see Other tests).

We included four studies90–93 on differential diagnosis of MCI and SCD. A cross-sectional survey90 of people aged ≥ 65 years in south London (n = 126) reported SCD in two-thirds of participants, and the SCD was described as significant in 31%. However, only one participant had sought help from their GP. This is consistent with studies of help-seeking reported above (see Help-seeking). A 2015 narrative review91 characterised SCD as a form of early cognitive decline that cannot be detected by current objective tests and, hence, a risk factor for progression to MCI and dementia. A small observational case series92 recruited 62 patients with SCD who were followed for a mean of 44 months. At the time of follow-up, 24% of patients had developed dementia or amnestic MCI, with the main factors predicting progression being older age at onset of symptoms and first assessment.

Some definitions of MCI include subjective memory complaints as a requirement. Yates et al. 93 investigated the presence of SCD and its relationship with mood in people with MCI. In a 2-year follow-up study, a clear association was found between SCD and mood both at a single time point and over time. Mood problems (e.g. depression or anxiety) were more closely related to the presence of SCD than that of objective cognitive impairment. The authors93 concluded that SCD may be a function of anxiety and depression rather than objectively detectable cognitive decline.

Evidence,8,89–93 varying in quantity and quality, suggests that distinguishing MCI from FCD and SCD is highly complex. FCD may be a separate entity and early diagnosis may be reassuring for the patient that they do not have dementia. SCD may, in at least some cases, be a precursor to progression to objective MCI and dementia.

Alternative causes of mild cognitive impairment

Expert reviews and clinical guidelines emphasise the importance of considering alternative (non-neurodegenerative) causes of MCI, but we found a lack of studies evaluating such investigations. An expert review26 of diagnosis and management recommended that clinicians should consider depression, polypharmacy and uncontrolled cardiovascular risk factors. Similarly, the Manchester consensus guidance5 identifies physical and psychiatric illness as alternative causes of cognitive symptoms.

Reversion to ‘normal’ status

Four studies43,69,94,95 included in the review investigated factors associated with reversion from MCI to normal cognition (Table 6). Awareness of this possibility is important for the investigation, counselling and management of people with MCI. Interpretation of the studies is complicated by different definitions of MCI and normal cognition. Overall, reversion to normal cognition was associated with younger age and a higher level of education. One study94 reported that people who reverted to normal cognition had more comorbidities (e.g. respiratory, urological and psychiatric conditions) than those who went on to develop dementia, suggesting that comorbidities may have been a factor in developing MCI and treatment of these conditions may have improved cognition.

Interestingly, for the investigation of MCI, Olazarán et al. 43 reported that data routinely collected as part of the patient history in primary care had value both for detecting MCI and predicting the outcome. Klekociuk et al. 69 took a contrasting approach, using a battery of tests to identify the best combination for distinguishing people with true MCI and reducing the rate of false-positive diagnoses.

Prognosis and progression to dementia

Progression from MCI to dementia is the reverse of return to normal cognition and the studies in the previous section may also be relevant. We included four studies17,96–98 specifically on the topic of predicting progression to dementia. A 4-year follow-up study96 of patients with amnestic MCI (n = 44) reported that 41% of patients had progressed to dementia by the end of the study. A combination of cognitive tests distinguished those who progressed from those who did not with 74% accuracy. Using a different technology, Montero-Odasso et al. 97 reported that dual-task gait analysis could be used to predict progression. The authors97 noted that gait testing is easy to administer and could be used to decide on further biomarker testing, preventative strategies and follow-up planning in patients with MCI.

An early study by Stephan et al. 17 examined different definitions of MCI and concluded that none reliably predicted risk of developing dementia over 2 years when applied to a general population, although people who were not at risk were more reliably identified. In contrast to the findings of Grande et al. 94 on comorbidities (see Reversion to ‘normal’ status), a further study by Stephan et al. 98 found no association between comorbidities and progression from MCI to dementia.

Dementia care pathways

Three narrative reviews99–101 researched dementia care pathways; further details are provided in Table 8. One review, published in 2014,100 reviewed the evidence around dementia care pathways and discussed the service pathways and the referral routes that are used by services in England and internationally. The review100 found that the term ‘dementia care pathway’ has several potentially overlapping meanings. For people with memory problems, the first encounter with health services, which is generally via a GP consultation, seems to be particularly important. The process of further assessment is not always straightforward, with barriers such as GPs reluctance to refer patients to memory services and the fact that the initial assessment appointment can be daunting and confusing for patients. The other narrative review,99 published in 2017, reviewed studies on the involvement of primary care in diagnostic care pathways for people with memory problems in England. The review found evidence of service redesign and the implementation of innovative approaches to attempt to improve dementia assessment, diagnosis and subsequent support. The development of innovative approaches, such as (1) GPs being primarily responsible for diagnosis and treatment, and only the more complex cases are referred to MASs or (2) diagnosis led by secondary care with support from older people’s mental health teams, should be encouraged and properly evaluated. In addition, attempts to improve access and signposting to postdiagnostic support should also be encouraged.

Another narrative review101 investigated how care is integrated for people with AD and other cognition disorders. This review101 found varying amounts of involvement of primary and specialist care within models of integrated community dementia assessment and management. All of these models focused on improving care co-ordination, interdisciplinary teamwork and personalised care. Although integrated care for people with dementia is highly desirable, the review101 found that efforts to improve integrated dementia care have had mixed results in terms of diagnosis rates, patient behaviour, carer burden and well-being. This limited success is partly due to the inherent complexity involved and the many different types of integration that can occur. Further steps towards a goal of integrated care would focus on improving interprofessional training, early diagnosis in primary care, technology, integrating dementia care into mainstream health and social care, and the encouragement of dementia-friendly communities.

Memory clinics/memory assessment services

Memory clinics or MASs were investigated in 16 studies16,46,48,49,70,77,102–111 (Table 9).

Four studies70,77,102,103 considered the organisation and characteristics of memory clinics throughout the UK and found considerable variability.

Chrysanthaki et al. 102 attempted to categorise MASs in the UK. The study randomly surveyed 73 MASs in 2015, finding considerable variation in staff numbers, new patients per whole-time equivalent, skills mix and nurse-to-doctor ratio, length of first appointment, follow-up time and frequency of follow-up within the first year. Following diagnosis, all MASs surveyed provided pharmacological treatment, but the availability of non-pharmacological support varied. These variations made it impossible to group characteristics within a typology of MAS, and the impact of individual structural and process characteristics should be considered in future evaluations. Another audit77 of 10 London memory services was undertaken in 2016 to compare memory services and develop a quality improvement programme. This audit, again, found considerable variation between services. An important finding was that postdiagnostic support information was provided for patients receiving a diagnosis of dementia only, meaning that patients diagnosed with MCI would generally receive no information or support. In addition, waiting times are potentially increased by people without dementia attending memory services. Services could consider developing specific pathways for these patients. In 2019, Cook et al. 103 undertook a larger audit of 85 MASs from five regions of England. This audit,103 again, revealed considerable differences across all stages of the memory service pathway, from assessment practices to choice of investigations, final diagnosis and access to support and treatment.

In 2020, the London Dementia Clinical Networks produced guidance70 for commissioners and clinicians working in memory services. The guidance70 discusses appropriate pathways for people presenting with memory problems that are not due to dementia. The London audit of memory services,77 discussed above, had found that 40% of patients were not given a dementia diagnosis, which increased to 85% for patients aged < 65 years. MCI was the most common non-dementia diagnosis. The guidance advised that people with MCI should be reviewed at least annually until a non-dementia cause is established, the condition has resolved or the person has been diagnosed with dementia. Follow-up may take place in primary or secondary care, depending on local commissioning arrangements. Similarly, the audit of memory services in London highlights the importance of follow-up for these patients, together with the likelihood that non-dementia patients can increase waiting lists and thereby delay dementia diagnosis. 103

Outside the UK, the development of memory clinics in Netherlands from 1998 to 2009 was described by Ramakers and Verhey. 107 The number of memory clinics and people accessing the services have increased significantly. In the Netherlands, memory clinics are now part of the regular care for people with memory problems and early stages of dementia.

Different ways of organising memory clinics were investigated in five studies. 16,48,49,104,108 The UK studies are discussed, followed by the studies from Australia.

A qualitative evaluation,48 published in 2014, of primary care-led dementia diagnostic services in Bristol found important themes of GPs and memory nurses working together and a lack of support following diagnosis. Although patients and carers were generally satisfied with primary care- or secondary care-led dementia diagnosis, they, together with health professionals, expressed concern about the lack of postdiagnostic support, which was a problem that was also highlighted in the 2016 audit of memory services. 77

One study108 explored the views of patients and carers on communication and information provision during memory services assessment. People with cognitive impairment and their carers were recruited from four memory clinics (two in north-east England, one in London and one in north-west England). The patients interviewed discussed the lack of information about the clinic that they were attending, the need for professionals to ensure that patients understand the process, and long waits for test results. Conversely, patients and carers found tailored information helpful, as well as signposting to local voluntary sector services that could provide information or support. The study108 concluded that the communication and information provided needs to be improved for patients being assessed for possible dementia, particularly for patients with MCI and for patients less likely to improve with medication.

A preliminary service evaluation16 in the UK used vignettes to consider knowledge of SCI and how it is managed in specialist clinical practice. The evaluation included 23 memory clinics located throughout the UK and found that most clinicians would discharge patients with SCI and only a small number of clinicians would arrange follow-up or diagnose SCI. The study16 concluded that there is a need for a coherent and consistent framework for the management of people with SCI.

Two further studies14,104 described service initiatives in Australia that could be introduced in the UK. A descriptive study104 aimed to identify differences between a nurse-led memory clinic with an open referral policy and other memory clinics in Australia. The characteristics of profiles to be identified were demographics, assessment scores and diagnostic profiles of people attending. The open referral policy led to a large proportion of patients being self-referred and nearly one-third of these patients were diagnosed with MCI or dementia. Open referral policies and nurse-led services may overcome current barriers to early diagnosis. An evidence briefing from the British Psychological Society14 noted that increasing access to memory clinics could lead to identification of more people with MCI.

A qualitative study49 elicited the views of GPs, older people, community/government local representation and community health-care workers to co-create a multidisciplinary memory clinic in Australia. Community forum participants felt that they had a good knowledge of available dementia resources and services, but that services were highly fragmented and needed to be easier to navigate for the patient/carer. A ‘one-stop shop’ and the provision of a dementia key worker were potential recommendations. Participants also recommended that the memory clinic should offer diagnostic services, rapid referrals, case management, education, legal services, culturally sensitive and appropriate services, allied health, research participation opportunities and clear communication with GPs. The study also highlighted that it is important to work with stakeholders to co-design models of care for people with dementia that take into account the needs of the community.

Patient outcomes were investigated in four studies. 46,105,106,109

A cohort study105 investigated changes in patients’ health-related quality of life 6 months after referral to a MAS in England. Patients’ mean health-related quality of life improved in the first 6 months irrespective of diagnosis. Another cohort study106 found that self-reported health-related quality of life improved in all study participants over the first year after their initial appointment at a MAS. People with a diagnosis of dementia saw greater improvement in health-related quality of life than people with MCI or no diagnosis. The presence of advisory and support staff at MASs was the only characteristic that was associated with large increases in health-related quality of life.

A qualitative study investigated the health care experiences of people with MCI and their caregivers,109 and this is discussed in greater detail in Chapter 9. Patients were recruited from six mental health trusts in the south of England. The study found that the specific needs of patients with MCI should be considered in the design of memory clinics and other related services.

A cluster-randomised controlled trial46 assessed the effect of a two-stage intervention of GP education case finding on diagnostic yield and the mental health of patients and carers in the Netherlands (i.e. a potential service innovation for the UK). Among patients and relatives who consented to stage 2 of the trial (n = 145, 25%), there were no differences in mental health between the intervention and control groups. There was a non-significant increase in the number of new MCI diagnoses in the intervention arm. 46

The costs of memory clinics were investigated in two studies110,111 and audits of memory clinics have demonstrated significant differences in the provision of services. A cohort study,111 published in 2015, compared the costs and quality of memory clinics with a traditional Community Mental Health Teams (both services were based in eastern areas of England). The mean cost for the memory clinic-based service was £742, which was cheaper than the mean cost for Community Mental Health Team service, which was £807. However, the difference was not statistically significant. The memory clinic-based service provided more comprehensive and multidisciplinary care, including counselling for patients before and after diagnosis, more screening of blood tests for comorbidities or reversible causes of dementia, increased use of structured assessment tools for cognitive screening of patients and carers, more regular copying of letters to patients/carers and signposting to third-sector services. Another cohort study110 assessed the cost of diagnosis and early support for patients with cognitive decline of patients attending randomly selected MASs in England. The study found considerable variation in costs across the MASs. Differences appeared to be due to differences in the workload of different staff. Further research to investigate the impact of this variation on patient outcomes would be helpful.

Role of other services in early or increasing dementia diagnosis

Four studies33,47,112,113 researched the role of other services in early or increasing diagnosis of people with MCI. These studies included primary care services, such as cancer clinics and an ophthalmology clinic. Further details of these studies are provided in Table 10.

Two studies47,112 investigated how services (ophthalmology clinics47 and cancer clinics112) could help to identify people with memory problems or cognition disorders when providing services. A narrative review47 explored the role of ophthalmology clinics in screening for early AD. Ophthalmology services and eye clinics could have a role in the early diagnosis of AD by communicating abnormal eye test findings to primary care physicians to follow-up and assess as required. However, the review47 did not include any studies evaluating this approach and was more a suggestion for further research. In addition, a outpatient cancer clinic in Wales investigated112 how it could identify people with memory problems to improve treatment experience and outcomes. The mixed-methods case study112 found five actions that could potentially improve the experience for people with dementia or MCI: (1) communicating common problems that people with memory problems could experience to increase awareness, (2) providing an environment that would help people to feel that they could discuss their memory problems, (3) training for staff on memory issues in patients, (4) providing tools that could help people to self-manage their memory problems and (5) tackling the support needs of carers. For people with dementia or MCI, embedding the clinical treatment of cancer within an environment that is dementia friendly could improve the cancer treatment experience for people with dementia or MCI.

Chan et al. 33 undertook a locality-based controlled study to evaluate whether or not a leaflet campaign to raise awareness of memory problems led to more people visiting their general practice to discuss concerns about their memory. The leaflet campaign increased the recording and management of memory problems, but the improvement was greater in control than intervention practice. This study,33 which was set in the UK, from 2010, may have questionable relevance to the current situation because of increased ease of accessing information using the internet, etc.

Certain ethnic groups can be reticent about reporting memory problems to their GP. A qualitative study113 investigated perceptions of associated symptoms, causes, consequences and preferred treatments of dementia for South Asian adults with memory problems in Manchester, UK. Perceptions varied among South Asians in different circumstances and in different age cohorts, and understanding these variations in perceptions of dementia can enable targeted information and interventions and lead to identification of those older adults at greater risk of distress.

Treatment/follow-up

Treatment or follow-up for MCI and dementia were investigated in six studies29,114–118 and further details of the studies are provided in Table 11.

A systematic review115 reviewed self-management interventions for people with dementia and MCI. The review115 found that group-based interventions for people with dementia or MCI included components of self-management, information, communication, social support and skills training. Preliminary evidence reported that self-management interventions can potentially help people with MCI or dementia and are feasible in this population. However, included studies had important methodological issues, with many not discussing the intervention’s theoretical basis and many not reporting measurable outcomes. To find the effectiveness of these interventions, more methodologically strong research is required.

A systematic review,116 published in 2013, reviewed the evidence on the impact of psychological intervention studies on adjustment depression and anxiety in MCI and early dementia. Several studies treating depression in people with early dementia using problem-solving and modified cognitive–behavioural therapy approaches had promising findings. Modified cognitive–behavioural therapy showed promise for improving adjustment and quality of life in patients with MCI and early dementia. Strong methodological studies are needed to properly test these interventions and enable the development of clinical recommendations. The British Psychological Society published an evidence briefing117 in 2016 on a similar topic, reviewing the evidence on psychological therapies for people with dementia. The briefing117 found that the evidence base is developing and, although methodological quality remains variable, conclusions can start to be drawn. Psychological therapies should be performed by trained professionals. For people with a dementia, diagnosis can either be formal or informal. These different formats consider the context of the therapy and the patient’s specific needs. The briefing identified the following three actions for moving forward. First, psychological therapies should be generally available in dementia care throughout the journey, from pre-assessment counselling to support following diagnosis, and at times when people with dementia become anxious or distressed. Second, psychological therapies, where necessary, should focus on carers’ and family needs and help a patient’s wider family to understand the diagnosis of dementia and enable planning for the future. Third, routine care should integrate psychotherapy skills to help improve interactions between people with dementia and their carers.

A qualitative study114 interviewed patients and their relatives from MASs and falls prevention services in the UK. The study114 reported that participants would consider taking exercise specifically to prevent falls and reported that participants would join a group if invited. Relatives wanted to provide practical support, but without undermining the patient’s independence.

The Royal College of Psychiatrists developed a consensus guideline118 that considers the role of clinicians in advising patients with MCI about driving. Clinicians managing patients with a MCI diagnosis should consider if patients are able to drive safely. Concerned clinicians should advise patients to contact the Driver and Vehicle Licensing Agency. Clinical reassessment of an individual’s ability to drive should be part of the ongoing management of a patient with a MCI diagnosis. The Manchester consensus5 felt that annual follow-up would be beneficial, although expensive to implement, and this could include assessment of driving ability.

A systematic evidence review,29 undertaken to inform the US Preventive Services Task Force, assessed the evidence on screening and treatment for cognitive impairment in older adults. The review29 included a large number of studies, but found that it remains unclear whether or not interventions for patients or caregivers provide clinically important benefits for older adults with earlier detected cognitive impairment or their caregivers.

Specific populations

Two studies41,119 investigated services and pathways for specific populations of people with dementia (i.e. prisoners119 and older homeless people41). A mixed-methods study119 investigated the prevalence of dementia and cognitive impairment in prisoners aged ≥ 50 years in England and Wales. In terms of services, the study found that local initiatives to improve the experience of prisoners had been developed but were difficult to sustain, given that they were not commissioned services. Multidisciplinary working in prisons is hindered by agencies continuing to work separately and the limited or inadequate communication between different professionals. Another mixed-methods study41 investigated service provision for homeless people experiencing memory problems. The study41 found that the links between different hostels and local primary care and mental health services varied significantly and some staff had substantially more sector experience than others, but high levels of staff turnover were reported. Training levels and availability varied, and differences existed in the extent to which hostel staff were permitted to access local NHS and local authority courses. The study41 concluded that, given the declining number of hostels in England, the limits of NHS engagement with this sector and growing homelessness, homeless people with memory problems are under-recognised and excluded from other sources of support. Services and support for prisoners and homeless people with MCI are insufficient and need to be developed.

The role of technology

Four studies38,120–122 investigated the use of technology in diagnosis and interventions for people with MCI and dementia (Table 12). For details of the study by Sabbagh et al. ,122 see Table 16.

A US expert panel considering the detection of MCI at home72 found that, despite barriers yet to be addressed, electronic point-of-contact testing holds great promise. It offers a critical method to support large-scale cognitive screening for the early detection of MCI. Supplementing in-clinic evaluation with at-home assessment may help to identify individuals with MCI, allowing physicians to intervene and ultimately to monitor progression, potentially without requiring the individual to present to the physician’s office.

A literature review38 of telemedicine in AD and MCI considered the diagnostic and interventional implications. Included studies reporting the use of telemedicine to support patients, their carers or both during the disease stages generally had positive results, although the review38 did not perform quality appraisal of the included studies.

A small focus group study120 researched how technology could help or hinder the daily lives of people with mild cognitive deficits. Participants in the study were from Sweden, and had cognitive deficits, lived with someone with cognitive deficits or were members of volunteer health organisations or health professionals at memory investigation clinics. The study found that technology could help, cause stress or bring about feelings of control. In addition, the study found that technology was used for orientation and managing finances. A small feasibility-usability study121 investigated a tablet application (app) that was adapted for people with cognitive impairment. The app was tested in home and clinical settings in Sweden and Spain. The app was found to be feasible and had an acceptable interface for people with MCI. The testing of the app resulted in further development and improved procedures.

COVID-19

Two studies123,124 covered the challenges arising from the COVID-19 pandemic for diagnosis and memory services and how services could develop (Table 13).

One policy/guidance,123 produced by NHS England and NHS Improvement, described changes to memory services in England due to COVID-19 and potential new ways of working, discussing general principles for configuring and managing memory services. A blended model, incorporating telephone, video and face to face, is needed. Personalised care remains key to MAS delivery and needs to evolve to adapt safely to COVID-19 restrictions while maintaining patient choice wherever possible. COVID-19 may become endemic and adaptation to a new way of working is required to maintain high-quality care. In addition, a narrative review124 considered the international evidence on remote clinics during and after the COVID-19 pandemic and potential solutions for after the pandemic. Memory services have experienced considerable challenges due to COVID-19; however, this has provided the opportunity for services to consider new technologies and approaches for cognition assessment. Solutions are available that could enable clinical assessment in people who are old, frail and limited by social distance measures. Moving forward, this approach, both now and in the future, could enable people who are living remotely or who would find attending a memory clinic stressful to be assessed in their own homes. This is linked to the study in technology section on home assessment (see The role of technology). 122

Overall pathway summary

In the UK, access to specialist medical care is through referral from primary care. When a person or their family or carer starts to identify memory problems and/or suspects dementia, they often first go to a primary care service. 60 Specific triggers for help-seeking for subjective cognitive concern include being prompted by family and knowing a relative with dementia. 125 Patients may have concerns about impaired memory, mood and personality changes or psychosis that could be caused by dementia. The GP will conduct an initial assessment to rule out any underlying physical or mental health causes. If further investigation is required, an onward referral is made to a MAS. The stages of the pathway are shown in Box 1 (based on NHS London Clinical Networks70) and are then explored narratively in the following sections. Table 15 summarises coverage of the pathway in the studies included in this section.

Patient/carer suspects dementia

Carers commonly precipitate the search for help from a doctor. 57 Patients may differ in their willingness and readiness to receive a diagnosis. 57 Persistent barriers to early diagnosis include stigma, the normalisation of symptoms and a lack of awareness about the signs and symptoms of dementia. 132 It often takes a trigger event or tipping point, such as a hospitalisation or bereavement, before people seek help. 57,100 Family members often recognise something is wrong before the person with dementia does and are frequently instrumental in obtaining a diagnosis. 56,57 Carers commonly report changes in behaviour, effects on the patient’s ability to socialise and exacerbation of pre-existing problems, such as marital difficulties and medical conditions. 22,56,133 Carers attribute changes to multiple causes, including the effect of ‘normal’ ageing, incipient dementia, longstanding personality traits and physical health problems. 22 The conceptual studies, reviewed earlier (see Chapter 6), reveal an emerging emphasis that the diagnosis pathway should make more consideration of the relative function of each individual.

Subsequently, family members may continue to find decision-making being deferred to them, particularly in relation to advance care planning. 128 Patients and relatives associate dementia with ‘loss of the self’. 56 Whether or not patients and relatives will want to be informed of the diagnosis depends on how they perceive the consequences of dementia or how they react to previous family experiences. 56 Studies describe how patients may display the phenomenon of ‘knowingly not wanting to know’. 15 Patients fear a loss of control (particularly with regard to issues such as continence) and express a loss of well-being. Patients are concerned that their family will be upset and they fear having to move to a care home. At this particular point in the pathway, patients do not hold specific expectations of support services, but they express a desire for advice. 56

Minority ethnic populations often perceive greater stigma for dementia symptoms. 32,57 These populations are less likely to recognise symptoms of dementia as an illness than white individuals, and more likely to ascribe these symptoms to the ageing process. 32,57 In addition, symptoms of dementia are sometimes given cultural or religious explanations. 32,57 One study reported that some people who have not had a consultation consider memory problems to be given by God32 and, consequently, see acceptance of fate as an alternative response. 32 Another study found that some people do not identify medical support as appropriate. 130

The evidence is inconclusive regarding whether or not level of education or socioeconomic status has an impact on awareness of dementia and help-seeking behaviour. 57

Presentation to primary care

Some people with memory problems approach the GP themselves, whereas others make a joint approach with family members, although carers often act proactively. In a few instances, the carer approaches the GP without informing the person who is experiencing difficulties. Those with memory problems or other symptoms may seek a diagnosis in the belief that catching the disorder early is crucial to receiving treatment for it. The media may be influential in this decision:

We’re still waiting for something to happen. Because they reckon early diagnosis don’t they and then they can do something and try and hold it if that is the case. Give you something to slow it up a bit. But we’re still waiting.

Carer59

A thematic analysis identified the point of deciding to seek help as one of three ‘transition points’ at which trust can either be developed or undermined. 125 Patients who perceive health-care practitioners’ behaviour as dismissive are less likely to trust the outcome of the health-care encounter.

Initial general practitioner assessment

Most patients report either positive or neutral experiences of consulting their GP about their memory problems. 109 In contrast, most patient advocates report at least some negative aspects to their encounter with the GP. 109 Patients who are referred to the memory service tend to recall little, if any, diagnostic input from their GP. However, these patients consistently acknowledge the GP’s importance as the person who starts the ball rolling at the beginning of the diagnostic process. 48 Many patients recognise the important gatekeeping role of the GP. Carers spoke of GPs not always taking action when concerns are first raised, but instead telling them to ‘keep an eye on it and in a couple months if it is still the same we’ll refer her for tests’. 48 Carers may feel that they are not being listened to by the GP or that they are being ‘fobbed off’ and that memory problems are not being taken seriously by GPs. 48 Patient advocates also report issues relating to ‘reporting’ on the person with MCI in their presence. 109

Studies suggest that sometimes doctors are slow to recognise symptoms or reluctant to give a diagnosis. 48,57 In terms of communication with professionals, people with memory problems and their carers identify two areas as most important to them: (1) being kept informed during the assessment process and (2) being told the outcomes of the assessment, including the results of tests and the diagnosis. 108 The health-care practitioners’ response to help-seeking is the second ‘transition point’ at which trust can be developed or undermined. 125 Although rates of detection of MCI in primary care have increased from the low rates of the early 2010s,10 a need to expand coverage by nurse practitioners and routine clinical data remains.

Treat or review

Experiences of referral are mixed, with some patients reporting a delay in referral to memory services and others reporting confidentiality obstacles, with doctors reluctant to talk to carers about their family member with dementia. 57 South Asians who have had a consultation are more likely to identify forgetfulness and loss of social meaning as symptoms of dementia. 130

Referral to a memory assessment service

Even when people have been referred to memory services, the process may be slow, with long periods of waiting. 57,59,60 Some experience significant delay between appointments. While waiting, patients report having little expectation of what they were waiting for. Consequently, intervals between appointments for results and explanations may be seen as burdensome. 59

Delays in the assessment process sometimes lead patients to believe that a diagnosis could not be confirmed. This can lead to disappointment with the health system. 59 The passage of time does not always resolve uncertainty about the diagnosis and how to respond to it and, in fact, it can exacerbate patient concerns. 59

Referral accepted

Seeking medical advice may offer practical benefits. Changing a set of signs and symptoms into a ‘known’ condition may offer the prospect of clinical management and the relief of suffering. However, in the specific case of MCI (i.e. with no recommended treatments being available) this otherwise promising pathway may become a dead end, leading to potential frustration. The discussion of timely diagnosis in the conceptual papers and the CIS reveal the importance of disclosure of the diagnosis at the ‘right time for the individual with consideration of their preferences and unique circumstances’. 13 This requires that help-seeking can be met by adequate services to optimise the benefits and minimise the costs and harms, as itemised below. An appropriate service response would include multiple treatment targets to help people to accept and adjust to their diagnosis. This would then realise benefits in terms of reduction of secondary conditions, such as depression, and improved quality of life. 18

Common sources of support for carers are friends, social groups and family members, including, in some cases, the person with MCI themselves. 109

Person offered dementia assessment

The process and outcome of assessment is a third and final ‘transition point’ at which trust can be developed or undermined. 125 Misunderstandings and absence of trust in assessment processes may lead to patients failing to fully agree with the outcomes of the assessment. 125

Dementia assessment

Patients express variable experience of a memory clinic. 109 Many patients experience the assessment process as confusing, with one study describing it as ‘labyrinth like’. 60 Patients are able to recall in detail what happened, the impact it had on their lives and the meanings they attributed to the process as they began to accept their condition. 60 Many patients experience rising anxiety as they approach the point of diagnosis. 60 Patients receive few explanations about the various medical tests, examinations and scans they underwent, including ‘X-rays of the head’60 and ‘general scans, magnetic resonance imaging scans, electrocardiography and blood and cholesterol tests’. 60 Even when these procedures are discussed, patients may remain confused about details and what the tests would reveal. 60

Neuropsychological testing can be particularly daunting when patients try to interpret their scores. 60 Patients respond to testing with words such as ‘worry’, ‘concern’ and ‘anxious’, starting from being visited by a specialist doctor to being concerned about what the tests are for, what scores mean and whether or not they have underperformed. 60 Without a clear explanation of the meaning of these scores, test results appear misunderstood or meaningless, making patients even more anxious. 60 Interpreting scan results can be frightening for patients and, for some, it is the first time that the possibility of AD is raised. 60 Patients express relief at the outcome of the assessment58 and relay positive comments on how the service is run. 109 However, most patients receive little or no information about their diagnosis. 58,109

Information reported by informants is important to confirm that the patient remains able to perform activities of daily living. An informant might report, for example, that the patient needs more time to complete tasks, but is still able to do them independently. 70

Clinicians are recommended to use a brief screening tool validated for the detection of people with MCI, such as the MoCA. Neuroimaging to identify underlying brain pathology may be useful, but is not currently considered mandatory. Neuropsychological assessment is potentially helpful in cases where screening test results are inconsistent with the results of clinical assessment.

Patients with early-stage dementia may struggle to complete the early cognitive decline assessment without some help. 126 Completion of the assessment using a computer may provoke frustration in patients. 126 Many may experience ‘dexterity limitations associated with ageing despite having no formal diagnosis of conditions affecting motor control e.g. lack of familiarity with touch screen technology, manifested in frequent accidental screen presses’. 126

Patients generally want professionals to provide understandable information and manage expectations during the assessment process. In addition, patients want memory clinic staff to provide specific information about local services. 108 Information needs to be provided in an appropriate form for people with memory problems and their carers. Abley et al. 108 found that few patients could recall positive examples of direct communication between memory services and themselves, or their carers, such as letters that could be easily understood. Overall, the memory services appeared to have few mechanisms to enable patients to remain informed. In some cases, families were informed of the outcomes of the investigations before the patient. 108

Mild cognitive impairment diagnosis

Mild cognitive impairment is characterised by the subjective experience of a decline from a previous level of cognitive functioning. Diagnosis of MCI requires objective evidence of impairment in performance on one or more cognitive domains relative to that expected given the individual’s age and general level of intellectual functioning. The cognitive impairment should not be sufficiently severe to significantly interfere with independence in the person’s performance of activities of daily living,9 but is not entirely attributable to normal ageing. Patients with MCI occupy a difficult middle ground where they often require assistance to maintain optimum function, but do not require enough help for it to be clearly classified as ‘caregiving’. 22 As discussed in Chapter 6, the use of cut-off points to distinguish between MCI and normal ageing may be problematic. For this reason, some authors argue that MCI should not be defined by performance below a cut-off point, but by evidence of decline from previous performance.

Studies of the communication of the diagnosis129,134 reveal that doctors commonly present a diagnosis of MCI as ‘not dementia’. Over three-quarters of doctors specifically name MCI, with almost a half of these doctors explaining MCI as a stage between ‘normal’ ageing and dementia. Where diagnosis is not named, doctors’ explanations focus on patient symptoms, often attributing these to vascular disease. 129

Emotional responses to a diagnosis of MCI range from the positive (e.g. relief that the diagnosis was not dementia) to negative (e.g. fear about the future). 22 Patients in all diagnostic groups face challenges in dealing with cognitive changes and with their changing self-identity. 127 Patients specifically with a MCI diagnosis face the challenge of causal attribution of cognitive decline, as well as anxiety and concern related to perceived decline. 127 Even where their cognitive state does not decline, patients still face anxiety, fearing that decline will occur.

People respond very emotionally, and often differently, to diagnosis, with responses that include sadness, anger, loss, rationalisation and humour. 131 The two most common responses to the diagnosis are worry and relief. 131 Many people who expect to receive a diagnosis of dementia or AD describe relief when this is not the case. One participant who had previously been told she had AD based on the result of her CT describes being ‘devastated, absolutely wiped out’ on hearing this and when she was later given the MCI diagnosis she was ‘over the moon’. 131

Following disclosure of a diagnosis, people with dementia develop diverse coping mechanisms from positive responses, such as feeling empowered, to difficulty in accepting the diagnosis and active denial. Some people respond to their diagnosis by seeing it as an opportunity to ‘reconsider their priorities and make positive changes, such as spending more time with family, focusing on positive tasks which bring enjoyment, undertaking future care planning and sharing the diagnosis with the outside world’. 56 People with MCI display varied coping strategies, being initially aware of their difficulties, but then, if progressing to dementia, losing this insight. The ability to cope and how to do so relate to how people perceive normal ageing and how they have previously experienced people with dementia. People find the label of AD more difficult to cope with than dementia or memory loss.

This sense of relief contrasts with previous widespread fear about the possibility of developing dementia or, specifically, AD. Other concerns may relate to brain cancer or other illnesses affecting cognition, after which being diagnosed with ‘a touch of Alzheimer’s’ may still be met with relief. 131 Nevertheless, many patients and carers continue to worry that MCI could develop into dementia. The MCI diagnostic label may fail to assuage anxiety given the similarities between MCI and dementia. However, other patients are reassured at the diagnosis. The relationship of MCI to dementia and AD may be interpreted in diverse ways. Some patients may see a diagnosis of MCI as implying no further danger from AD. Other patients may miss the mention of the term MCI all together, perhaps latching on to familiar lay terms, such as memory loss. 131 People with MCI and their families report struggling to deal with ongoing changes in relationships, with both parties fearing that the person with MCI could become dependent. 131

Patients struggle to make sense of the MCI label. 131 The term ‘cognitive’ is particularly problematic. Patients do not usually encounter the term outside their contact with medical and research professionals. However, it is not only the ‘cognitive’ part of the label that patients have a problem with, but describing their impairments as ‘mild’ fails to grasp the depth of their difficulties. 131 This phenomenon has been described as ‘not knowing’ about MCI. 15

Many patients question the MCI diagnosis, rating their own experience over ‘the abstract discourse of medicine’. 131 The ability to trust the doctor, or to rely on the ability of the doctor, is critical, with many patients believing that more is going on than the doctors uncover. Ruling out an AD diagnosis may be seen as a clinician’s strategy to mask something more serious or sinister, or to stop the patient from contemplating serious actions. The converse is that some patients believe that MCI is a form of overmedicalisation whereby a diagnostic label is inappropriately attached to the normal ageing process. However, not all patients react against the MCI label, as others trust the diagnosis because of their respect for clinical judgement. Such patients may also express a willingness to leave clinical issues in the hands of the clinicians so that they can get on with living their own lives. 58

A tension regarding medication sees some people expressing a desire for medication, whereas others are cautious of overmedicating. Medical advice may be solicited, not as a way to medically ‘treat’ memory loss, but to secure a sense that support is available to be called on as and when needed. People with MCI come to view clinical services not only as a source of appropriate treatments, but also as a way of receiving reassuring support as and when they need it. It is not unusual, in the absence of a definitive medicine to slow down or prevent cognitive decline, for people with MCI to draw on ‘self-management strategies to maintain cognition – crosswords, puzzles, “brain training” games, and diary keeping’. 23 Clinical services are, therefore, viewed very much as a back-up service, as persons with MCI take the lead on managing the changes they experienced. Patients may be provided with information, such as the NHS Health Check Dementia Leaflet,135 which is available in different languages and as a video.

Those with MCI are frequently not told or given written information about the clinic or service they are attending. Professionals may ask people with memory problems if they would like copies of the clinic letter sent to their GP, but this is not always actioned. 108 Some patients are told that they have MCI. 108 Although seemingly accurate from a professional perspective, on its own, without additional information or practical help, this may seem meaningless and not particularly helpful. 108 Patients may react by questioning whether or not their condition is a disease. 24 On diagnosis, patients struggle with the social implications of a potential future diagnosis of AD. 24 The resulting ambiguity may precipitate both social and psychological tensions. 24 Patients particularly resist attribution of their symptoms to AD, explained by the considerable stigma associated with that condition24 and this illustrates that a significant psychosocial impact is associated with how MCI is framed and stigmatised. 24 Subjective memory impairment can be more concerning than angina, asthma, hypertension or a previous heart attack.

Care plan

In MCI, the difficulties patients experience with their mental abilities, such as memory or thinking, are worse than would normally be expected for a healthy person of their age. However, the symptoms are not severe enough to interfere significantly with daily life. Many people with MCI remain stable or improve in cognition, but this condition does involve an increased risk of dementia.

Care plans should include strategies to help those with MCI to cope with practical problems, for example no longer being able to deal with finances and withdrawing their involvement in hobbies and other activities, as well as with difficulties in planning and spatial orientation. Those with MCI need assistance with activities of daily living. They not only need to be able to recognise their neuropsychiatric symptoms, but also need to develop coping strategies that help them to come to terms with their memory problems.

Memory retraining classes and memory strategy groups run by memory clinics are reported to have positive outcomes for patients and carers. 108 Where patients attend weekly sessions at memory sessions, they find information integrated within these sessions beneficial. 108 Memory clinics also offer opportunities to meet others ‘in the same boat’ in addition to providing practical strategies for dealing with memory problems. Written handouts provided at the clinics may be considered accessible and useful. 108 The source of the information can be important. Carers may contact voluntary organisations, particularly when services do not provide timely or sufficient information. Such information is found to be helpful, including ‘buddying schemes, carers’ leaflets and regular newsletters’. 108

It is considered good clinical practice to assess people with MCI on a yearly basis, provided that they consent to this. This should take place until:

• a non-dementia cause is established

• it has resolved, such as treated sleep apnoea (using screening tools such as the Epworth Sleepiness Scale or STOP-Bang)

• the patient has been diagnosed with dementia.

Follow-up can occur in primary or secondary care, according to local commissioning. Some general practices keep their own MCI registers so that they can recall patients for their annual review and, if there are concerns, then a referral is made to the memory service. Efforts to embed review within each general practice may be facilitated by commissioners adding this to their QIPP (quality, innovation, productivity and prevention) workstream.

Exit pathway

Patients will continue to experience changing levels of insight and also to maintain differing perceptions of well-being in ageing. 127 Diagnostic labels can take multiple sources of complex information and reduce them to a label that explains the whole. 131 As Verhaeghe136 puts it:

. . . a name always carries with it [ . . . ] the illusion of control and mastery. Nothing is worse than not being able to name something; once a name has been found, it seems manageable.

Verhaeghe136

Relief at receiving a serious diagnosis has even been noted in studies of dementia. 56

Diagnosis of dementia

Making the diagnosis of dementia and conveying it to patients and carers is challenging. As discussed earlier (see Chapter 6), controversy surrounds whether or not everything between normal ageing and dementia (currently labelled as MCI) should be considered as prodromal dementia. 137 Negative connotations associated with dementia, inconsistent symptoms and not knowing enough about the signs and symptoms are commonly reported barriers to early dementia diagnosis. Being told one has dementia has a large impact on a patient’s identity and often causes feelings of loss, anger, fear and frustration. 22,57,108 Spouses have to adjust to increasingly unequal relationships and the transition to a role as carer. 22,108 Carers may be vulnerable to the strain associated with these relationships, often leading to their own health problems. 57 On the other hand, studies examining the experience of couples often report how they find ways to continue working together as a team.

Adjusting to a dementia diagnosis is a complex process. 57 Initially, patients and carers experience conflicts, such as a conflict between autonomy and safety, between recognising the need for help and reluctance to accept it, or between living in the present and dealing with anxiety about and preparing for the future. 57 Clinicians need to be able to identify the levels of emotional distress and difficulty in coping with the illness in people with dementia and their family carers in a timely manner, offering early referral for specialist psychological support. 56

Regardless of culture and context, individuals often share common experiences of becoming a person with dementia. Dementia has an impact on identity,57 leading to feelings of loss, anger, uncertainty and frustration. 22,57 People with dementia struggle to preserve aspects of their former self. Family carers may support them by focusing on the person’s abilities, rather than drawing attention to mistakes. 57 A desire to preserve a pre-dementia identity sometimes leads to people being reluctant to disclose their diagnosis,57 which could lead to social isolation. 57 Despite this, studies suggest that eventually individuals with dementia and their carers come to accept their situation. 57

Dementia has a significant impact on both the individuals with dementia and their families. Spouses have to recalibrate to increasingly unequal relationships57 and communication between the couple is often affected. However, studies looking at the experiences of couples often found an emphasis on working together as a team, with a high degree of mutuality. 57 Significant strain on carers may, subsequently, have an adverse impact on their own health. 57

Some individuals respond to receiving a diagnosis as the beginning of adjustment, whereas other individuals who have been experiencing symptoms for some time have already made considerable adjustment. Although diagnosis is commonly traumatic,57 the validation of suspicions can come as a relief. 57 Some individuals with dementia and their carers may continue to consider memory loss insignificant even after diagnosis. 57

Tensions and conflicts tend to resolve over time as the disease progresses. Patients and carers become more balanced and accepting of the dementia. Many patients and their families adopt strategies to cope with the impact of dementia on their lives to manage the disease and maintain a ‘normal life’. Practical strategies include use of reminders22 and social strategies may require relying on family support. Carers seek to maintain patients’ previous level of function and activity through encouragement and by planning activities together. 22

Some patients utilise emotional strategies, such as using humour. 57 At some point, patients and carers report that they are able to adopt positive mindsets and incorporate dementia in their lives.

Patients with a dementia diagnosis have an urgent need for support from outside the family, both immediately after diagnosis and subsequently. 57 GPs play an important role in helping patients and carers to gain access to information, social and psychological support, and community care. This ongoing need for information requires information from diverse sources and in varied formats. 108 Several studies57,108 attest to the need for patients and carers to access information on financial aids and entitlements early on. The studies also highlight a need for continued access to supportive professionals and specialists.

Families with a member who has experience of dealing with dementia as a health professional or with a relative with dementia are more likely to completely acknowledge a diagnosis than those with no previous exposure to dementia. 57 Adjusting to a diagnosis is complex. Several conflicts and tensions need to be resolved to accommodate a diagnosis. For example, studies identify the conflict that may arise as people strive to preserve identity and autonomy in the face of increasing symptoms. This sometimes leads to apparent unawareness of or resistance to acknowledging a diagnosis. 57 Although this might be interpreted as denial, it may be seen ‘as a self-maintaining strategy or a deliberate choice to be seen as an agent rather than an object’. 57 Although some people with dementia and their carers actively seek information,57 others reject new knowledge. However, understanding of and attitudes towards dementia are not fixed and can evolve throughout the disease trajectory.

The GP is generally the first point of contact for people with dementia and their carers. 57 GPs have an important role to play in facilitating service access. 57 Attending memory clinics could be shocking or frightening57 and receiving a diagnosis could lead to increased tension as someone negotiates a new identity as a person with dementia. 57 Barriers to memory clinic use could include transport access, funding, awareness and costs. 49

Patients express a clear need for greater support after diagnosis, including advice, social and psychological support, access to community care and respite. 57 Valuable support is provided by voluntary organisations, such as the Alzheimer’s Society. 57,108 Signposting to voluntary organisations needs to be improved. Information provision is seen as key in many studies, but better knowledge-sharing at the point of diagnosis is not always the solution. 57 The information needs of patients vary over time and information provision needs to be ongoing, offering flexibility in timing and format. 57,108

Many people find peer support valuable. 57 However, for others, there could be negative consequences, as the inclusion of people at different stages in the dementia trajectory could make people aware of what the future held for them. 57 The timing of referral to community-based support groups may be key. 57 Appropriate referral to such groups is more likely to occur within a continuous therapeutic relationship between the person with dementia and practitioners involved in their care.

Patients consistently express negative perceptions attached to a diagnosis of dementia. 127

In terms of receiving a positive diagnosis of dementia, some patients and carers appreciated honest and clear communication and the opportunity for follow-up discussion. Sometimes, however, the amount and timing of follow-up was thought to be inadequate, leading to distress. 108 When opportunities are provided, patients do not always know what questions to ask, especially immediately following a diagnosis. 108

Implications for pathways

Patients who have been diagnosed are less likely to be anxious and upset if they have had a previous opportunity to discuss the possibility of dementia as a diagnosis with health professionals. 108 In England, patients can access memory clinic services via a GP referral only. Although the pathway presents referral from GP as an interim stage on the way to assessment, it is clear that formal discussion between GPs and their patients with memory problems prior to referral should be seen as key in communication. 108 Information leaflets that outline what a patient and their family can expect from their local memory clinic are also helpful. Changes in the commissioning of services in England may enhance GPs’ roles in this area. 108 Commissioning guidance for diagnostic and early assessment dementia services has already emerged. 59 Although national guidance is useful in terms of setting evidence-based standards, local services will need to construct flexible diagnostic disclosure pathways. 108

Protocol

The review was registered on the PROSPERO database (reference CRD42021232535). PRISMA guidelines have been followed in the reporting of this synthesis. In addition, recent observations on reporting from a subset of CISs have been factored into the methods and the documentation of the report. 141

Open research question

Critical interpretive synthesis typically starts with an open research question. For this study, we used ‘how are people with a cognitive impairment investigated to understand the underlying cause of impairment?’. The question was then refined into meaningful subquestions during progress towards a finalised question, which became apparent within the line of argument constructed by the end of the review.

Literature search

See Chapter 2.

Inclusion criteria

In addition to the inclusion criteria for review 1, studies included in the CIS had to meet the following criteria.

Study selection and data extraction

See Chapter 2.

Quality appraisal

For CIS, quality appraisal explores the likely relevance and contribution to the theory that is being developed. A form of triage took place to ensure that the included studies were specifically robust as an example of the type of study under consideration (e.g. a systematic review or a qualitative study). However, the subsequent synthesis attached emphasis to the likely relevance of items, but not to a consideration of relative rigour. This decision is compatible with interpretive forms of synthesis in which the contribution of an article to a line of argument is considered important, in comparison with the risk of gaps in the overall argumentation.

Formulation of a synthesising argument

A CIS was performed to identify unifying or explanatory themes emerging from the literature. Essentially, the CIS requires analysis and familiarisation with the studies, iterative exploration of the extracted data and consideration of the identified themes. Critical examination of the ways that the literature had constructed two competing narratives helped in constructing the research questions. The first narrative relates to the nature and adverse consequences of diagnostic uncertainty and the desirability of attaching a label for the patient, clinician and health service. The second narrative relates to the ethics challenges of offering a potential diagnosis when diagnosis is unclear, prognosis uncertain and disease-modifying therapies unavailable. Communication between clinician and patient and their carer was isolated as a central issue within the care pathway. Several studies of patient–clinician communication were prioritised in the analysis. 129,142–144 CIS offers a versatile tool for handling qualitative, quantitative and mixed research without excluding methodologically weaker studies of clear relevance. Within the framework of the two competing overall narratives, we utilised an inductive approach to identify recurring themes and develop a synthesising line of argument.

Characteristics of included literature

Table 16 shows details of articles included in the synthesis together with their methodologies. Forty publications were identified, including 10 qualitative studies, with a further three qualitative studies that used either conversation analysis or discourse analysis, two mixed methods studies, four surveys and one audit study. Only two quantitative studies were identified to inform the debate (a cross sectional study and a cluster-randomised controlled trial). We examined six expert consensus statements and two guidelines. In addition, we selected seven literature reviews, comprising three systematic reviews, one qualitative meta-synthesis, two scoping reviews and one unspecified literature review. The remaining two items represented editorial opinion contributions. All items exclusively related to high-income countries. The majority of papers (n = 19) originated from the UK, including three papers from Wales and one each from Northern Ireland and Scotland. Other studies represented the Netherlands (n = 4), the USA (n = 3), Canada (n = 2) and Germany (n = 1). Four studies were from Europe more generally and two were ‘international’. One of the European studies also included survey respondents from the USA and Canada and so was coded multiple times. The remaining seven items were literature reviews for which their coverage, not their geographic origin, is salient to this discourse (see Table 16).

What is known about the communication of diagnostic uncertainty in mild cognitive impairment?

Summary

This review reveals that MCI has fulfilled its original pragmatic purpose as a mid-way ‘holding stack’ between normal cognitive functioning and dementia. However, the absence of unambiguous and operational definition has resulted in persistent uncertainty as to whether it refers to cognitive difficulties from any cause or constitutes a prodrome of AD. In the presence of such uncertainty, commentators advocate selective avoidance of the label in diagnostic practice until more is known about its core features and predictive value. Since Ritchie and Ritchie’s previous editorial,161 the major shift in the categorisation of MCI has been the introduction of biomarkers to inform the identification – if not yet the prognosis – of AD. 180 However, the introduction of biomarkers has led to a reconceptualisation of AD (seen in the popularisation of the newly proposed concepts of ‘preclinical’ and ‘prodromal’ AD for those who test positive for AD biomarkers but have no dementia). 151

Synthesising argument for diagnosis of mild cognitive impairment

Our synthesising argument (i.e. the output generated by CIS) reads as follows.

The core concept along the pathway of diagnosis and treatment for dementia is ‘timely diagnosis’. 145,167 This concept acknowledges that an attempt to pinpoint a precise point for intervention that is common to all patients who display MCI is likely to meet only limited success. Historically, ‘timely diagnosis’ can be considered intervention at an individual patient-defined point that allows the patient and their carer to undertake mitigating changes in lifestyle and/or to make preparations for future activities of living and financial arrangements162,168 in anticipation or in advance of a worsening of their symptoms. 145 Planning for information provision and social support become important in this context. 162 Alternatively, patients and carers may make an active decision to handle their situation by living in the present and taking one day at a time. Against this patient-oriented conception of timely diagnosis has been ranged a health service imperative for early intervention, driven by a prevalent assumption that earliest possible intervention is necessarily a good thing and supported by indicative evidence that medical costs decrease with an earlier, more accurate diagnosis. 162,181 Finally, the clinical perspective sees the previously limited ambition of intervention being potentially transformed by biomarkers, both in diagnosis and in treatment with disease-modifying agents. These clinical developments have served to bring the clinical horizon more sharply in view for individual patients.

When viewed against these three competing drivers, the decision on diagnosis of MCI transforms from a debate between early and late diagnosis to, instead, an attempt to optimise and synchronise the three competing time cycles of patient, service and clinician. Adoption of the preferred objective ‘timely’ suppresses the relative value of both ‘early’ and ‘late’ and seeks to synchronise timelines informed by patient values, health service policy and evidence-based clinical care.

Furthermore, the subsidiary debate about the value of attribution of the label of MCI is deflected from its questionable value as an intermediate ‘holding’ stage where the patient may deny or conceal symptoms,40 which may or may not lead to subsequent dementia162 and in which a patient is typically ‘parked’ in a memory clinic. Instead, the label can be interpreted as a timely detection point at which the patient and carer are able to prepare and the clinician is able to meaningfully intervene, using cost-effective drug (e.g. acetylcholinesterase inhibitors) or non-drug interventions (e.g. cognitive stimulation therapy) to delay cognitive deterioration and improve quality of life. 162 Although subverting the early diagnosis targets arbitrarily specified for health services planning and delivery, appropriate, timely and acceptable intervention represents an effective and, largely, cost-effective strategy for health systems. 73

Implications of ‘timely diagnosis’ for the clinical encounter

Based on the synthesis and the overall synthesising argument, we were able to identify the implications of operationalising ‘timely diagnosis’. Communication is key so that the clinician can identify what exactly the patient wants to know and when, and any information requirements that will require supplementary information. A focus on activities of daily life and how memory problems affect the activities in which an individual patient is likely to engage is to be privileged over a standard approach based on some ‘typical’ disease trajectory. The US Alzheimer’s Association (Chicago, IL, USA) has developed best practice clinical guidelines for the evaluation of neurodegenerative cognitive behavioural syndromes, AD and dementias. 179

Critical interpretive synthesis: strengths and limitations

This was a comprehensive review of the literature on issues surrounding the label of MCI. However, the studies identified were heterogeneous, with primarily the qualitative studies facilitating richer analysis. All of the studies were conducted in high-income countries. By including published papers in English only, we may have missed research more relevant to other cultural settings. In addition, we included only studies published between 2010 to 2021. However, several articles25,145,152 adopted a historical approach to the development of the MCI concept, offering a longer time perspective.

Limitations to our CIS include the fact that reliance on the label MCI may have excluded some studies that document a pre-dementia stage in less specific terms. However, we compensated for this weakness by following up systematically through citation searching and use of a citations-in-context web tool (scite).

As characteristic of interpretivist research, our values and research experiences have informed how we interpreted the data and, therefore, presented the advantages, disadvantages and considerations for each clinical pathway. Therefore, other researchers may interpret the analysis differently because of different preferred paradigms and experiences. In addition, as we sought neither to ‘vote count’ nor to balance equal numbers of papers supporting or negating the claims of early diagnosis, it is possible that we have quantitatively given more attention to one of the two counterarguments. Nevertheless, the overall resting position that we have adopted of a need for ‘timely diagnosis’ places value on both competing positions and, indeed, seeks to reconcile their imperatives. Finally, the authors of this study identify with a patient, rather than a clinician, perspective. This may privilege social and psychological concerns over clinical and technological considerations and may endorse the rights of the individual over the concerns of the health service. Nevertheless, our perspectives and experience as health service researchers do acknowledge the need to make decisions that seek to optimise the effectiveness, efficiency and appropriateness of service delivery.

Beyond the specific context of MCI, the issues explored in this study provide a valuable exemplar of how clinical uncertainty, which subsequently translates to patient and carer uncertainty, acts out in a face-to-face context. The questionable ethics and clinical value of a diagnosis that is not tied securely in with a specific prognosis, and of screening against the acknowledged principles of mass screening, namely in the absence of a treatment, are confirmed by the clinical and patient perspectives that were engaged by this synthesis. This potential sticking point has been eroded by the development of biomarkers and more so by promising treatments. However, escape from this particular impasse should not mask the fact that effective communication, which involves acknowledging and learning to deal with uncertainty, is key not only to this situation, but also to other clinical areas of uncertain diagnosis.

Main findings

The two reviews presented in this report represent complementary approaches to the synthesis of a substantial body of evidence on the pathway from awareness of memory problems to a potential diagnosis of dementia. The CIS (review 2) includes additional evidence from commentary and opinion to understand the pros and cons of MCI as a diagnostic label from the clinician and patient/carer perspectives. Although our brief was to examine the investigation of cognitive impairment and not the diagnosis of dementia, some degree of overlap was inevitable.

The descriptive review highlights multiple barriers to efficient diagnosis of memory problems, starting with patient reluctance to seek help. Interventions to encourage people with concerns about their memory to see their GP have been evaluated, but without clear evidence of effectiveness. 33,34 GPs have a variety of cognitive tests available, but recent evidence suggests that substantial numbers of patients meeting criteria for dementia do not have a diagnosis recorded. 44 Options for management for people with memory problems include follow-up in the community (with individual advice about reducing dementia risk) and referral to a memory clinic/memory service for further assessment. Memory clinics are mainly intended to identify and support people with dementia and those not meeting diagnostic criteria may find themselves being discharged back to the care of their GP.

The review identified considerable variation in the way memory clinics in the UK are organised and their approach to investigating the underlying cause of memory problems. 102,103 Recent expert consensus guidance suggests that the availability of imaging and biomarker tests is limited outside specialist tertiary centres. 5 During the COVID-19 pandemic, memory clinics have changed their way of working to include virtual assessments. 123 Another recent development that is likely to have major implications for service providers and patients with MCI is the development of aducanumab (Aduhelm^®, Biogen Inc., Cambridge, MA, USA) [the first disease-modifying therapy for dementia (specifically AD)], which was recently approved by regulators in the USA (see Implications for research and Implications for service delivery).

The descriptive review found clear evidence from qualitative studies that patients with MCI and their carers find the process of investigation and diagnosis difficult and frustrating to negotiate. 1,60 Receiving a diagnostic label of MCI involves living with uncertainty23 and the terminology itself may be problematic for patients.

The CIS (review 2) investigated the advantages and disadvantages of MCI as a diagnostic label from various standpoints. The key finding from the synthesis is that the need for a ‘timely’ diagnosis outweighs the ongoing debate about the value, or otherwise, of early investigation and labelling of memory problems. Determining what is a timely diagnosis involves balancing the perspectives of the patient, the health system and the clinician. For the patient, a timely diagnosis is one that enables them to make changes and provision for the future while avoiding unnecessary anxiety and stress. The health service in the UK has to balance a policy that favours early diagnosis and treatment of memory problems against the limitation of available resources and interventions. From the clinical perspective, scanning and biomarker tests offer the possibility of a diagnosis of early dementia prior to the appearance of symptoms, with associated ethics issues. 85 The concept of what is a timely diagnosis may be transformed in the relatively near future if disease-modifying treatments for dementia prove to be effective and enter routine clinical practice.

Strengths and limitations

Implications for service delivery

We identified the following implications for service delivery:

• Our results suggest the need for a more formalised discussion between GPs and their patients with memory problems prior to memory clinic referral, covering the implications of dementia as a possible diagnosis.

• Services should consider the potential value of efforts to improve the recording of diagnoses of dementia in primary care (e.g. by provision of training and support for GPs to perform assessments using validated tools). 44

• Quality improvement work at a local and national level is expected to produce benefits in terms of improving and standardising services provided in memory clinics. 103

• Changes to the operation of memory clinics necessitated by the COVID-19 pandemic will require evaluation to ensure that services are delivered efficiently and effectively in the aftermath of the pandemic.

• The Manchester consensus guidance5 identified a need for NICE guidance on diagnosis and management of MCI and such guidance could reduce variation in service delivery. Although national guidance is useful in terms of setting evidence-based standards, local services will need to construct flexible diagnostic disclosure pathways. 70

• There is likely to be a need to plan and resource services to optimise the delivery of disease-modifying therapies if such therapies are approved for use in the NHS. This is a priority for both service delivery and research.

Implications for research

We identified the following priorities for research:

• There is a need to strengthen the evidence base for primary care-led investigation (including appropriate tools and thresholds for referral) and management of memory problems compared with service delivery through hospital-based memory clinics.

• The descriptive review identified limited research on screening for memory problems outside general practice. 47 Research to evaluate models of service for other settings, such as emergency departments, acute hospital wards and care homes, would be of value.

• In view of the move towards remote delivery of health care forced by the COVID-19 pandemic,123,124 research is needed to evaluate remote methods of memory assessment as part of mainstream services, as well as for remote locations.

• Further qualitative research is needed to ensure that services for people with memory problems are patient centred and provide people with a timely diagnosis that is expressed in terms that they can understand and act on. This research should include people with diverse memory problems (i.e. SCD and FCD, as well as MCI) and different underlying causes (for those diagnosed with MCI).

• As noted above, research is needed to optimise the introduction of disease-modifying treatments for early dementia on approval. This research could build on modelling work already carried out to estimate the costs of increased use of scanning and biomarkers, including equipment and training costs. 81

• Research should continue to develop and evaluate evidence-based programmes to reduce dementia risk for people with MCI, which can be implemented at scale, taking into the needs and preferences of people with MCI. 1 Research should consider the needs of minority groups and potential impact on health inequalities.

Conclusions

The concept of MCI as a state between normal ageing and dementia has been in use for many years. Conceptual studies indicate that MCI is understood by many clinicians as a pre-dementia (prodromal) condition related to AD or other forms of dementia. In practice, the diagnostic label of MCI is applied to people with a variety of underlying conditions whose cognitive status may decline, remain stable or improve over time. The pros and cons of the diagnostic label from the patient and clinician viewpoints are considered in review 2 and the findings suggest that the label is valued by clinicians, but is more problematic for patients and their carers.

Investigation of memory problems normally starts when people seek help from their GP. Delays in seeking help after noticing symptoms are common and members of ethnic minority groups may face specific barriers to help-seeking.

Evidence suggests that GPs may have difficulty recognising and recording memory problems using clinical judgement alone. Screening with cognitive tests may be a good use of resources73 and is likely to be of increasing importance with the development of disease-modifying treatments that may benefit people in the early stages of dementia.

Further investigation involves tests (i.e. MRI or PET and analysis of biomarkers in blood and CSF) that are generally available at specialist centres only. The process is complicated by potential overlap with at least two other recognised conditions (i.e. SCD and FCD).

Pathways for people with memory problems may involve follow-up in primary care or referral to a memory clinic/service. We found wide variation in organisation and access to memory clinics in the UK and internationally. A key finding is that memory clinics are primarily commissioned to identify and support people with dementia, suggesting that different service models may be needed for people with MCI. The lack of an evidence-based population-level dementia prevention programme may be a barrier to developing such services.

People with MCI interviewed for qualitative studies frequently portrayed their experiences prior to diagnosis in negative terms. The findings suggest a need for research and practice to make the investigation and management of MCI more patient centred.

The key finding from the CIS (review 2) is that the need for a ‘timely’ diagnosis outweighs the ongoing debate about the value, or otherwise, of early investigation and labelling of memory problems. Determining what is a timely diagnosis involves balancing the perspectives of the patient, the health system and the clinician.

The two reviews reported here have applied different ‘lenses’ to the same body of evidence (supplemented by some additional studies for the CIS). Taken together, the reviews identify the concept of a timely diagnosis for memory problems and identify barriers to obtaining such a diagnosis, from reluctance to seek help through to patchy availability of advanced diagnostic tests. We have also identified priorities for service delivery and research. The report is particularly timely in the light of the recent US approval of a disease-modifying treatment for AD. 183

Contributions of authors

Duncan Chambers (https://orcid.org/0000-0002-0154-0469) (Research Fellow in Public Health) commented on drafts of the protocol and drafts of the report, and was responsible for project co-ordination, study selection, data extraction, quality assessment and report writing.

Anna Cantrell (https://orcid.org/0000-0003-0040-9853) (Research Fellow in Health Economics and Decision Science) commented on drafts of the protocol and drafts of the report, and was responsible for information retrieval, study selection, data extraction, quality assessment and report writing.

Katie Sworn (Research Associate in Health Economics and Decision Science) commented on drafts of the report, and was responsible for study selection, data extraction, quality assessment and report writing.

Andrew Booth (https://orcid.org/0000-0003-4808-3880) (Reader in Evidence-Based Information Practice) commented on drafts of the protocol and drafts of the report, and was responsible for information retrieval, study selection, data extraction, quality assessment and report writing.

Data-sharing statement

Any additional data not included in this report and its appendices are available on request. All queries should be submitted to the corresponding author.

Disclaimers

This report presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the HSDR programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the HSDR programme or the Department of Health and Social Care.

MEDLINE search strategy

Database: Ovid MEDLINE(R) and Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Daily.

Date of search: January 2021.

Dates searched: 1946–8 January 2021.

Mild cognitive impairment grey search

Grey literature searches were performed in January 2021 to retrieve clinical guidelines, policy documents and reports related to MCI from relevant websites:

• Age UK (URL: www.ageuk.org.uk)

• American Academy of Neurology (URL: www.aan.com/)

• Alzheimer’s Research UK (URL: www.alzheimersresearchuk.org)

• Alzheimer’s Society (URL: www.alzheimers.org.uk)

• British Geriatrics Society (URL: www.bgs.org.uk)

• The British Psychological Society (URL: www.bps.org.uk)

• GOV.UK (URL: www.gov.uk).

In addition, searches were conducted on the Google search engine™ (Google Inc., Mountain View, CA, USA; URL: www.google.com/) for relevant literature (MCI and guideline, MCI and policy, etc.).