Educational interventions for improving the skills of medical practitioners to detect, diagnose, and manage people with cognitive impairment and dementia

  • Protocol
  • Intervention



This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess and compare the effectiveness of different educational interventions that are aimed at improving the skills of primary care and secondary care clinicians to detect, diagnose, and manage people with cognitive impairment and dementia.


Description of the condition

Dementia is a chronic, progressive syndrome that manifests itself as a conglomerate of deficits in multiple domains of higher cognitive functions (Boustani 2003; NICE 2006). The typical clinical presentation of dementia is impairment of memory, accompanied by at least one of the following cognitive disturbances: aphasia, apraxia, agnosia, and deficiencies in executive functioning (DSM IV; van Hout 2007). These impairments must represent a decline from a baseline level of normal functioning and be severe enough to have a negative impact on an individual's social or occupational performance (DSM IV).

There are several subtypes of dementia, which result from different underlying pathological processes and differ in their typical manifestations and prognosis (Iliffe 2009). This, in turn, will inform the treatment approach (Iliffe 2009). In clinical practice, the three more common subtypes of dementia are Alzheimer's disease (AD), which accounts for approximately 40% of dementia cases, and vascular dementia (VD) and dementia with Lewy bodies (DLB), each accounting for approximately 25% (Boustani 2003; Iliffe 2009). Both AD and VD can coexist in what is known as mixed dementia, the prevalence of which is greater than originally thought (Alzheimer's Association). Although the exact prevalence of mixed dementia is not known, some autopsy series estimate that vascular pathology coexists in approximately 24 to 28% of cases of AD (Langa 2004). Certain types of dementia can follow an atypical pattern of neuropsychological deterioration in which memory impairment is not the most salient feature (Galton 2000).

Dementia imposes a burden on patients, their families, and society. For patients, dementia leads to difficulty in conducting a range of occupational, social, and (ultimately) personal routine activities; increased care dependency; increased likelihood of institutionalisation; and reduced life expectancy (Chengxuan 2007; Robinson 2009). The next of kin of people with dementia experience anxiety and depression (Boustani 2003). From a public health perspective, dementia contributes 11.2% of all years lived with disability; higher than the contribution from any other single clinical condition (Iliffe 2009; Koch 2010). Its socioeconomic impact is also considerable; dementia-associated costs are currently estimated to be in the order of GBP 17 billion in the UK (Koch 2010), and USD 100 billion in the US (Boustani 2003).

Dementia is among the most common conditions in elderly people, with its prevalence rising exponentially from 1.5% in people aged 60 to 69 years to more than 40% in people aged 85 years or older (Kiejna 2010). Its incidence increases with age (Chengxuan 2007; Ferri 2005). In the US, the estimated incidence of AD is 53 cases per 1000 people aged 65 to 74 years, increasing to 170 cases per 1000 people aged 75 to 84 years and 231 cases per 1000 people aged 85 years and over (Alzheimer's Association 2012). Trends in population ageing have made dementia one of the major public health priorities worldwide. Ferri 2005 estimated the total number of people living with dementia worldwide at 24 million; as the world population ages, this number could double to 42 million by 2020 and reach 81 million by 2040.

The exponential increase in the number of people with dementia raises concerns regarding their accurate and timely detection and diagnosis. Non-recognition can range from 50% to more than 75% depending on the setting (Bamford 2007; Borson 2006; Boustani 2003); and it becomes particularly evident when clinicians are confronted with atypical cases or with people in the early stages of the disease (i.e. those with mild cognitive impairments) (Galton 2000; Iliffe 2009; Koch 2010; van Hout 2007). Non-recognition and delays in diagnosis can lead to patients and their next of kin not being able to access timely and optimal pharmacological treatment and psychosocial support (Downs 2006).

General practitioners are, in most countries, the first point of contact for people with dementia (Downs 1996; Downs 2006; van Hout 2007). In the primary care setting, the evidence has shown that early recognition of dementia is limited (van Hout 2007). In the UK, diagnosis takes between 18 and 30 months (Iliffe 2009). Physicians in secondary care are usually more familiar with the early signs of cognitive deterioration (Phung 2010). In this setting, countries such as Denmark, the UK, and the US have detection rates of approximately 80% or more.

A number of barriers delay prompt and accurate diagnosis and management of people with dementia. In their systematic review, Koch 2010 grouped these barriers into three broad categories: (1) system factors; (2) patient or societal factors; and (3) factors related to the doctor involved. The first two groups cover issues such as the stigma associated with a diagnosis of dementia; delayed presentation of people to healthcare services; and constraints on physicians, including insufficient consultation time and inadequate financial remuneration; and the lack of support for patients, carers and physicians once the diagnosis is made (Borson 2006; Chemali 2012; Iliffe 2009; Koch 2010). In view of the rapidly increasing number of people with dementia worldwide, we need to establish strategies to overcome these obstacles and change the way in which health services are organised and delivered.

Factors related to the treating physicians' attitudes, particularly in primary care, include the uncertainties in reaching and disclosing a diagnosis of dementia; particularly in the early stages when dementia syndromes often follow an insidious and variable course. As a result, the detection of mild cognitive impairment and early dementia can be problematic (Iliffe 2009). Even when doctors are able to reach a diagnosis of dementia, they are often concerned about the risks associated with disclosing the diagnosis at "the wrong place or the wrong time" (Koch 2010; Pimlott 2009). Additional factors related to the treating physicians include lack of sufficient knowledge or experience, particularly in primary care, and the widespread therapeutic nihilistic views (Koch 2010; Koch 2011; van Hout 2007). The latter refers to the view that diagnosis is not worthwhile due to a lack of available treatments or other benefits of diagnosis. Preliminary evidence seems to suggest that educational interventions for clinicians and the introduction of evidence-based practice protocols may result in improvements in the detection, diagnosis, and management of people with cognitive impairment and dementia (Downs 2006; Iliffe 2009).

Description of the intervention

Educational interventions in this context are designed to change physicians' behaviour and attitudes in relation to cognitive impairment and dementia (Koch 2011). These interventions utilise a wide range of methods and techniques, including: practice-based workshops, use of decision-support software or clinical practice guidelines, electronic tutorials, and facilitated small group learning (Downs 2006; van Hout 2007). Interventions using blended learning methods are increasingly becoming more popular. These methods consist of a combination of traditional teaching methods with e-learning tools and other learning media (Downs 2006).

The specific aims of educational interventions have been formulated as follows (Bamford 2007; Downs 2006):

  1. to improve clinicians' knowledge acquisition and clinical performance in relation to the evaluation, diagnosis, and management of people with cognitive impairment and dementia;

  2. to modify clinicians' attitudes and perceptions about dementia and the quality of the care and support currently available to patients and their next of kin;

  3. to stimulate a more active approach to case finding in both primary and secondary care settings; and

  4. to improve clinicians' skills in the use of screening tests and neuropsychological diagnostic tools.

By promoting these changes, it is expected that educational interventions can improve physicians' performance, and, hence, health outcomes (Mann 2004; Wilcock 2002).

How the intervention might work

A key element in the development of educational interventions is the careful consideration of the value of each of the theories underpinning them (Wilcock 2002). In medical education, a theory may (1) lead to a better understanding of how clinicians learn and incorporate change; (2) inform the planning and implementation of educational programmes; and (3) facilitate and enhance learners' natural learning processes, competence, and problem-solving skills (Mann 2004; Wilcock 2002). Scientists typically group theories of adult learning depending on their underlying principles and assumptions about learning; and they may fall in any of these widely accepted groups: (1) behaviourist, (2) cognitivist, (3) social learning, (4) humanist, and (5) constructivist (Mann 2004). Behaviourist theories, for example, are based on the assumption that the environment determines an individual's overt behaviour; thus, educational interventions based on behaviourist theories may focus on feedback as a way of accelerating and increasing the speed of learning (Mann 2004). A clinical practitioner usually obtains multisource feedback, receiving information about their performance on a number of competencies from their supervisors, colleagues, and patients (Mann 2004). This feedback will help learners to achieve their current objectives and to increase their progress towards any subsequent goals they set for themselves.

Medical educational interventions can use a wide range of techniques. Peer reflection is encouraged using practice-based workshops about real cases through case discussion in small, multidisciplinary groups. Such workshops may involve electronic tutorials; using decision support software or clinical practice guidelines, or both, applied to real time and real case learning. These interventions could potentially improve the detection rate and management of people with early or mild cognitive impairment by assisting clinical reasoning and care planning in real time (Downs 2006; Koch 2011; van Hout 2007). In order to encourage clinicians to reflect on their knowledge and past experience, particularly in relation to difficult and complex cases, electronic tutorials are used (Downs 2006). Facilitated small group sessions can give clinicians the opportunity to identify individual learning needs, together with their facilitator. Both electronic tutorials and small group sessions could potentially improve clinicians' knowledge by increasing their motivation for self directed learning (Downs 2006; Koch 2011). The effectiveness of these methods, however, can vary due to potential differences in learning styles and diagnosing methods between primary and secondary care clinicians.

Potential benefits and adverse effects of the intervention

Educational interventions can improve clinicians' skills in detecting and diagnosing cognitive impairment and dementia; this could in turn lead to earlier disclosure of the diagnosis to patients and their next of kin. The early detection of dementia will enable patients and their carers to:

  • adjust emotionally and practically to the diagnosis;

  • initiate early treatment (which can delay admission to nursing homes and time to dependency);

  • prolong the early phase of the condition;

  • discuss future care, safety, and financial planning;

  • formulate advanced directives;

  • promote awareness of relevant support organisations;

  • help relieve the psychological distress experienced by next of kin and carers; and

  • contemplate issues such as research participation, driving, and house relocation, among others (Boustani 2003; Iliffe 2009; Koch 2010; van Hout 2007).

However, some scientists argue that early disclosure may have a negative impact not only on the person receiving the diagnosis and their family and friends, but also on healthcare services (Iliffe 2009; Mattsson 2010). This is particularly evident when we consider: (1) the current paucity of validated and approved disease-modifying therapies for dementia (Mattsson 2010); and (2) the amount of support that people with dementia and their carers need in order to maintain a certain degree of functional independence and dignity (Bond 2005). Both Iliffe 2009 and Mattsson 2010 have discussed some of the potential drawbacks associated with early detection of dementia. People with dementia who are diagnosed early can experience greater restriction of their daily activities; increased preoccupation with their condition; prolonged follow-up; and stigmatisation, resulting in feelings of humiliation, hopelessness, agony, and despair. For next of kin and carers, an early diagnosis of dementia can translate into hyper-vigilance for symptoms. In addition, their worries for their own health can increase the demand for pre-symptomatic testing. False-positive diagnoses can cause unnecessary distress to both patients and their next of kin, and erode trust in the clinician's professional ability. Under-recognition of co-morbidities such as depression can occur, as several behaviours are misclassified as cognitive changes that are thought to be part of the dementia phenotype. The latter can pose additional pressure on specialist services receiving an increased number of referrals. Finally, from a legal stand point, a diagnosis of dementia can significantly affect insurance premiums, as well as rights such as that to hold a driver's license. However, if educational strategies focus on improving management of dementia, not only detection and diagnosis, some of the potential drawbacks may be minimised.

Why it is important to do this review

The need to improve the early detection and the evidence-based management of people with cognitive impairment and dementia is the focal point of a number of national strategies and clinical guidelines in countries around the world. For example, in the UK, the National Service Framework for Older People, the National Dementia Strategy (Koch 2011), and the Dementia Challenge launched by the Prime Minister in March 2012 highlighted this need.

A number of researchers have attempted to evaluate the educational interventions designed to improve the detection and management of dementia, particularly in primary care (Borson 2006; Downs 2006; Koch 2011; Valcour 2000; van Hout 2007). However, if we are to draw firm conclusions about the effectiveness of these interventions, we must evaluate this body of literature systematically. We were able to find only one systematic review that addresses this research question (Koch 2011). However, the review in question has employed a rapid appraisal approach, thus excluding many studies that were not readily accessible or that had been published in a language other than English. Moreover, the performance of secondary care clinicians in relation to the detection, diagnosis, and management of people with dementia has rarely been studied. A systematic review in this area is, therefore, warranted to address these methodological limitations and gaps and help to inform future research.


To assess and compare the effectiveness of different educational interventions that are aimed at improving the skills of primary care and secondary care clinicians to detect, diagnose, and manage people with cognitive impairment and dementia.


Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCT), including cluster RCTs. We will include reports of ongoing or unpublished work if they are associated with our outcomes of interest. In these cases, we will contact the relevant authors.

Where available, we will include process evaluations and cost evaluations that are associated with a primary trial. This will allow us to describe in more detail the development and evaluation of the components of these educational interventions.

We will exclude any other type of study design as this systematic review is concerned with the effectiveness of educational interventions.

Types of participants

We will include participants who are generalist clinicians working in either primary care or secondary care settings, and specialist consultants in any related medical field and who are working in any clinical setting. We will also include studies specifically targeting specialist consultants in the fields of neurology, old age psychiatry, and geriatric medicine. Since people with cognitive impairment and dementia form an important part of these clinicians' daily workload (Mitchell 2010), we would expect their clinical skills and detection rates to be extremely good. However, the complexity and variability of the clinical presentation of people with dementia, in particular the atypical cases, can challenge even the most experienced of clinicians.

We will exclude healthcare professionals other than doctors (e.g. nurses, community care workers, counsellors, psychologists, etc.). We will also exclude people with cognitive impairment or dementia, their next of kin, and their carers. We will not make exclusions based on age, gender or any other sociodemographic characteristics of the participants.

Types of interventions

We will include educational interventions whose primary objective is to:

  1. improve clinicians' skills in evaluating, diagnosing, managing (or a combination of these) people with cognitive impairment or dementia;

  2. improve clinicians' skills in the use of screening tools or neuropsychological diagnostic tests, or both;

  3. increase clinicians' knowledge and awareness in relation to cognitive impairment and dementia; and

  4. encourage clinicians to follow the recommendations of evidence-based clinical guidelines or decision support algorithms.

We will include interventions that use any delivery method, such as didactic teaching, traditional continuing medical education (CME), mailings, clinical practice guidelines, electronic tutorials, and facilitated small group learning.

We will consider studies for inclusion if they compare different educational interventions against each other or against no intervention. We will exclude studies in which the primary intervention is not an educational programme but which focus, instead, on assessing the effects of pharmacological treatments; population screening; validating or evaluating a particular cognitive function test; or estimating the prevalence of cognitive impairment or dementia in a particular patient population or setting, or both.

Types of outcome measures

Primary outcomes

We will include studies reporting on the following primary outcomes:

  • clinicians' accuracy of diagnosis, assessed against a confirmed diagnosis of dementia made by a cognitive disorders specialist (e.g. psychiatrist or neurologist);

  • referral rates of people with cognitive impairment or dementia from primary care to secondary care or specialist centres; and

  • clinicians' concordance with evidence-based clinical guidelines or decision support algorithms related to the diagnosis or management of dementia.

Secondary outcomes
  • Clinicians' knowledge of cognitive impairment or dementia, or both, using any validated tool (e.g. Clinical Knowledge Summaries of the National Institute for Health and Clinical Excellence).

  • Changes in clinicians' attitudes or perceptions of dementia, or both, using any validated tool (e.g. the Dementia Attitudes Scale).

  • Changes in clinicians' confidence in using screening tools or specific cognitive function tests, or both (e.g. Mini-Mental Status Examination).

  • Costs of educational interventions.

Search methods for identification of studies

Electronic searches

We will search ALOIS ( - the Cochrane Dementia and Cognitive Improvement Group’s Specialized Register. The search terms used will be: education, educational, training, teach, teaching, instructing, instructions, lesson, workshop, computer, tutorial.

The Trials Search Co-ordinator maintains the ALOIS register, which contains dementia and cognitive improvement studies identified from:  

  1. monthly searches of a number of major healthcare databases: MEDLINE, EMBASE, CINAHL, PsycINFO, and LILACS;

  2. monthly searches of a number of trial registers: metaRegister of Controlled Trials, Umin Japan Trial Register, and World Health Organization portal (which covers, International Standard Randomised Controlled Trial Number (ISRCTN), Chinese Clinical trials Register, German Clinical trials register, Iranian Registry of Clinical trials, and the Netherlands National Trials Register, plus others);

  3. quarterly search of the Cochrane Central Register of Controlled Trials (CENTRAL);

  4. monthly searches of a number of grey literature sources: ISI Web of knowledge Conference Proceedings, Index to Theses, and Australasian Digital Theses.

To view a list of all sources searched for ALOIS see About ALOIS on the ALOIS website.

We will run additional separate searches in many of the above sources to ensure that we retrieve the most up-to-date results. Appendix 1 shows the search strategy that we will use for the retrieval of reports of trials from MEDLINE (via the Ovid SP platform).

Searching other resources

For all included studies, we will conduct an author and citation search in Science Citation Index. We will check the reference lists of all potentially included studies and review articles for additional references. We will contact authors of relevant trials and ask them to identify additional published, ongoing, and unpublished studies. We will revise and update the key original database search strategy according to any additional information.

Data collection and analysis

Selection of studies

Two of the review authors (JMB and KIT) will implement the search strategy, and import all the references identified to EndNote. They will remove duplicate records of the same report. JMB and KIT will independently screen titles and abstracts to exclude those obviously irrelevant studies. JMB will obtain full-text copies of potentially relevant studies. JMB and KIT will independently screen the full text of these studies to assess for compliance with our inclusion and exclusion criteria. We will resolve any disagreements through discussion and, if we cannot agree, JC or LM will act as arbiters.

Data extraction and management

JMB and KIT will independently extract data from included studies using a structured data extraction form. They will then compare their completed forms and follow up any discrepancies, with reference to the original article. They will contact authors of studies containing missing or incomplete data in an effort to obtain any missing information.  

We will attempt to extract the following information from each included study:

  • general information about the study;

  • study protocols including aims of the study, study design, methods of and setting for recruitment, inclusion and exclusion criteria, details of the control and comparison groups, and incentives for participation;

  • description and number of participants, and setting and place where the intervention was delivered;

  • details of the intervention, including the underlying educational theory; and

  • specific objectives of the intervention and skills it aims to promote.

In view of the anticipated complexity of the interventions under consideration in this systematic review, we will follow the framework proposed by Möhler 2012 and attempt to collect data on the developmental and evaluating processes of the included studies (see Table 1). If available, we will collect data on a number of process components such as recruitment, reach, fidelity, satisfaction, and implementation components (dose delivered and dose received) (Driessen 2010).

Table 1. Criteria for reporting the development and evaluation of complex interventions (CReDECI) checklist
  1. Source: Möhler 2012.

First stage - development
1Description of the intervention's underlying theoretical considerations
2Description of all components of the intervention
3Rationale for the aim/essential functions of the intervention's components
4Illustration of any intended interactions between different components
5Rationale for the aim/essential functions of the intervention's components, including the evidence whether the components are appropriate for achieving this goal
6Consideration of contextual factors and determinants of the setting in the modelling of the intervention
Second stage - feasibility and piloting
7Information on pilot-testing
8In case of pilot-testing: presentation of all relevant results and their impact on the modelling of the final intervention
Third stage - introduction of the intervention and evaluation
9Description of the control intervention (comparator)
10If the study was conducted in different clusters or centres: description of a standardised implementation strategy throughout the centres
11Description of all materials or tools used for the implementation of the intervention to allow a replication of the study
12Description of an evaluation of the implementation process
13Description of any deviation from the study protocol during the implementation process
14Description of facilitators or barriers revealed by the process evaluation that have influenced the intervention's implementation
15Description of unexpected interactions between components of the intervention and the environment in which the intervention was implemented
16Description of costs or required resources for the intervention's implementation

This information will be summarised in a 'Characteristics of included studies' table. For each outcome measure, we will attempt to extract sufficient data to allow for an intention-to-treat (ITT) analysis.

Assessment of risk of bias in included studies

Two review authors (JMB and KIT) will assess the risk of bias for all included studies, using The Cochrane Collaboration's tool for assessing the risk of bias in RCTs (Higgins 2011). We will assess the risk of bias across the following domains:

  1. random sequence generation;

  2. allocation concealment;

  3. blinding of participants and personnel;

  4. blinding of outcome assessment;

  5. incomplete outcome data;

  6. selective outcome reporting; and

  7. other bias.

Other sources of bias will include the comparability of intervention and control group characteristics at baseline; validation of outcome assessment tools; reliability of outcome measures; and protection against contamination. 'Protection against contamination' refers to those circumstances in which the control group is likely to have received the intervention.

For all studies, we will attempt to locate the original study protocol or study registration record to compare reported methods and outcomes against those originally planned.

As recommended by Puffer 2003, we will assess the risk of bias for cluster RCTs across the following domains:

  1. secure cluster allocation (evidence that cluster randomisation took place securely);

  2. cluster attrition (loss of entire clusters after randomisation);

  3. cluster imbalance (evidence of imbalances at the cluster level);

  4. differential individual recruitment or consent (evidence that different proportions of participants were recruited into the different arms of the trial); and

  5. differential individual exclusion or inclusion (evidence that eligibility criteria were applied differentially after randomisation).

JMB and KIT will independently assign each domain of each individual study to one of three categories: low, high, or unclear risk of bias. They will resolve any differences by discussion; if they cannot reach consensus, JC or LM will act as arbiters. For each study, we will present a 'Risk of bias' table summarising our judgements. We will incorporate the results of our assessment of risk of bias into the review through a narrative description of each of the quality items for each included study.

Measures of treatment effect

We will evaluate the characteristics of included studies in order to determine whether they are suitable for pooling in a meta-analysis. We will include all the studies that meet our inclusion criteria and report the outcomes of interest in a meta-analysis.

For dichotomous outcomes, we will calculate the odds ratio (OR) and 95% confidence intervals (CI). For continuous outcomes, we will calculate the mean difference (MD) and 95% CI. For studies using different measurement scales, we will calculate the standardised mean difference (SMD). In the case of missing data, we will conduct a modified ITT analysis. We will perform all statistical analyses using Review Manager 5 (RevMan 2011).

Unit of analysis issues

In the case of cluster RCTs, we will attempt to obtain data at the patient level. If these data are not available from the study report, we will request them directly from the contact author(s). We will meta-analyse patient-level data using a generic inverse-variance method in Review Manager 5 (RevMan 2011), which accounts for the clustering of data. If access to patient-level data is not possible, we will extract a summary effect measurement for each cluster. The sample size will be considered as the number of clusters and the analysis will proceed as if the trial was individually randomised. The drawback of this technique, however, is that it will reduce the statistical power of our analysis. In those cases in which the statistical analysis of a cluster RCT has already been adjusted for the clustering of data, we will extract the reported effect estimates and use them directly for our analysis.

Dealing with missing data

Where necessary, we will contact study authors to request additional information on missing data. We will report any missing data if we cannot gather any additional information. For the purpose of the analysis, we will conduct an ITT analysis.

Assessment of heterogeneity

We will assess studies for diversity across the following domains: intervention, quantitative outcome, and participants. Studies deemed sufficiently similar in these domains will be meta-analysed and evaluated for heterogeneity by visual inspection of forest plots and by using the I2 statistic, accompanied by 95% CIs. If substantial heterogeneity is found (as indicated by an I2 value greater than 0.5), we will not assess publication bias or perform a meta-analysis. In this case, we will attempt to determine the potential reasons for the observed heterogeneity.

Assessment of reporting biases

Where possible, we will include studies published in any language to minimise language bias. We will also attempt to minimise the risk of publication bias by conducting a comprehensive search of multiple databases, including those of ongoing and unpublished trials. If at least 10 studies are identified, we will assess reporting biases using a funnel plot regression weighed by the inverse of the pooled variance. We will interpret a regression slope of zero as absence of publication bias.

Data synthesis

If we find appropriate numerical data and conducting a meta-analysis is indicated, we will synthesise data according to their type. For continuous outcomes, we will calculate the MD or the SMD depending on whether the study authors reported outcomes on the same scale or not. For dichotomous variables, we will calculate the OR and 95% CI. Where continuous outcomes have been categorised by the study authors, we will treat them as categorical outcomes and calculate an OR and 95% CI between pairs of categories if the outcomes are not already dichotomised. If appropriate numerical data are not available or if meta-analysis is not appropriate, or both, we will perform a narrative synthesis of the evidence.

We will use the GRADE system to assess the quality of the evidence, the magnitude of effect of the interventions examined, and the sum of available data on the main outcomes, to produce a 'Summary of findings' table for each primary outcome (Higgins 2011).

We will perform one or more meta-analyses (using a random-effects model) for any of the primary outcome measures if at least two studies remain for each outcome with the I2 statistic less than 0.5. We will not conduct a meta-analysis in the case of significant heterogeneity. If we consider the duration of the different studies too heterogeneous, we will divide the data into smaller time periods and conduct a separate meta-analysis for each time period.

Subgroup analysis and investigation of heterogeneity

We will analyse the effect of different groups of educational theories on outcomes of interest separately. We will group educational interventions according to their underpinning theories and distribute them into subgroups. We will then compare the results between subgroups.

We will explore the effect of potential differences in learning styles and diagnosing methods between primary and secondary care clinicians (see How the intervention might work), by conducting subgroup analyses according to the different clinical specialities being targeted by the included studies.

Sensitivity analysis

A sensitivity analysis will be considered if one or more studies are dominant due to their size; or if one or more studies have results that differ from those observed in other studies; or if one or more studies have quality issues that may affect their interpretation judged using The Cochrane Collaboration's tool for assessing risk of bias (Higgins 2011).


Appendix 1. MEDLINE search strategy

1. exp Dementia/

2. Delirium/

3. Wernicke Encephalopathy/

4. Delirium, Dementia, Amnestic, Cognitive Disorders/

5. dement*.mp.

6. alzheimer*.mp.

7. (lewy* adj2 bod*).mp.

8. deliri*.mp.

9. (chronic adj2 cerebrovascular).mp.

10. ("organic brain disease" or "organic brain syndrome").mp.

11. ("normal pressure hydrocephalus" and "shunt*").mp.

12. "benign senescent forgetfulness".mp.

13. (cerebr* adj2 deteriorat*).mp.

14. (cerebral* adj2 insufficient*).mp.

15. (pick* adj2 disease).mp.

16. (creutzfeldt or jcd or cjd).mp.

17. huntington*.mp.

18. binswanger*.mp.

19. korsako*.mp.

20. or/1-19

21. "cognit* impair*".mp.

22. exp *Cognition Disorders/

23. MCI.ti,ab.

24. ACMI.ti,ab.

25. ARCD.ti,ab.

26. SMC.ti,ab.

27. CIND.ti,ab.

28. BSF.ti,ab.

29. AAMI.ti,ab.

30. MD.ti,ab.

31. LCD.ti,ab.

32. QD.ti,ab.

33. AACD.ti,ab.

34. MNCD.ti,ab.

35. MCD.ti,ab.

36. ("N-MCI" or "A-MCI" or "M-MCI").ti,ab.

37. ((cognit* or memory or cerebr* or mental*) adj3 (declin* or impair* or los* or deteriorat* or degenerat* or complain* or disturb* or disorder*)).ti,ab.

38. "preclinical AD".mp.

39. "pre-clinical AD".mp.

40. ("preclinical alzheimer*" or "pre-clinical alzheimer*").mp.

41. (aMCI or MCIa).ti,ab.

42. ("CDR 0.5" or "clinical dementia rating scale 0.5").ti,ab.

43. ("GDS 3" or "stage 3 GDS").ti,ab.

44. ("global deterioration scale" and "stage 3").mp.

45. "Benign senescent forgetfulness".ti,ab.

46. "mild neurocognit* disorder*".ti,ab.

47. (prodrom* adj2 dement*).ti,ab.

48. (episodic* adj2 memory).mp.

49. ("preclinical dementia" or "pre-clinical dementia").mp.

50. or/21-49

51. 20 or 50

52. ((educat* or train* or teach*) adj3 (program* or intervention* or course* or workshop* or seminar*)).ti,ab.

53. Education, Medical/

54. Education, Medical, Continuing/

55. Teaching/

56. workshop*.ti,ab.

57. ("e-tutorial*" or "electronic tutorial*" or "web-based tutorial*").ti,ab.

58. (clinical adj3 guideline*).ti,ab.

59. ("decision-support software" or (decision adj3 (software or computer or web*))).ti,ab.

60. "group learning".ti,ab.

61. Computer-Assisted Instruction/mt, st [Methods, Standards]

62. Decision Support Techniques/

63. "self directed learning".ti,ab.

64. or/52-63

65. 51 and 64

66. randomized controlled

67. controlled clinical

68. randomi?ed.ab.

69. placebo.ab.

70. randomly.ab.

71. trial.ab.

72. groups.ab.

73. cluster*.ti,ab.

74. RCT.ti,ab.

75. or/66-74

76. (animals not (humans and animals)).sh.

77. 75 not 76

78. 65 and 77

Contributions of authors

JMB and KIT wrote the protocol. JMB developed the search strategy. JC conceived the idea for the systematic review and reviewed the protocol.

Declarations of interest

None known.