Cognitive stimulation to improve cognitive functioning in people with dementia

Transcription

1 Cognitive stimulation to improve cognitive functioning in people with dementia Review information Authors Bob Woods 1, Elisa Aguirre 2, Aimee E Spector 3, Martin Orrell 4 1 Dementia Services Development Centre Wales, Bangor University, Bangor, UK 2 University College London, London, UK 3 Research Department of Clinical, Educational and Health Psychology, University College, London, London, UK 4 Research Department of Mental Health Sciences, University College London, London, UK Citation example: Woods B, Aguirre E, Spector AE, Orrell M. Cognitive stimulation to improve cognitive functioning in people with dementia. Cochrane Database of Systematic Reviews 2005, Issue 4. Art. No.: CD DOI: / CD Contact person Bob Woods Professor of Clinical Psychology of Older People Dementia Services Development Centre Wales Bangor University 45 College Road Bangor Gwynedd LL57 2DG UK b.woods@bangor.ac.uk Dates Assessed as Up-to-date:6 December 2011 Date of Search: 6 December 2011 Next Stage Expected: 15 January 2013 Protocol First Published: Issue 4, 2005 Review First Published: Not specified Last Citation Issue: Issue 4, 2005 What's new Date Event Description History Date Event Description Abstract Background Cognitive stimulation is an intervention for people with dementia which offers a range of enjoyable activities providing general stimulation for thinking, concentration and memory usually in a social setting, such as a small group. Its roots can be traced back to Reality Orientation (RO), which was developed in the late 1950s as a response to confusion and disorientation in older patients in hospital units in the USA. RO emphasised the engagement of nursing assistants in a hopeful, therapeutic process but became associated with a rigid, confrontational approach to people with dementia, leading to its use becoming less and less common. Cognitive stimulation is often discussed in normal ageing as well as in dementia. This reflects a general view that lack of cognitive activity hastens cognitive decline. With people with dementia, cognitive stimulation attempts to make use of the positive aspects of RO whilst ensuring that the stimulation is implemented in a sensitive, respectful and person-centred manner. There is often little consistency in the application and availability of psychological therapies in dementia services, so a systematic review of the available evidence regarding cognitive stimulation is important in order to identify its effectiveness and to place practice recommendations on a sound evidence base. Objectives To evaluate the effectiveness and impact of cognitive stimulation interventions aimed at improving cognition for people with dementia, including any negative effects. 1 / 69

2 Search methods Cognitive stimulation to improve cognitive functioning in people with dementia The trials were identified from a search of the Cochrane Dementia and Cognitive Improvement Group Specialized Register, called ALOIS (updated 6 December 2011). The search termsused were: cognitive stimulation, reality orientation, memory therapy, memory groups, memory support, memory stimulation, global stimulation, cognitive psychostimulation. Supplementary searches were performed in a number of major healthcare databases and trial registers to ensure that the search was up to date and comprehensive. Selection criteria All randomised controlled trials (RCTs) of cognitive stimulation for dementia which incorporated a measure of cognitive change were included. Data collection and analysis Data were extracted independently by two review authors using a previously tested data extraction form. Study authors were contacted for data not provided in the papers. Two review authors conducted independent assessments of the risk of bias in included studies. Main results Fifteen RCTs were included in the review. Six of these had been included in the previous review of RO. The studies included participants from a variety of settings, interventions that were of varying duration and intensity, and were from several different countries. The quality of the studies was generally low by current standards but most had taken steps to ensure assessors were blind to treatment allocation. Data were entered in the meta-analyses for 718 participants (407 receiving cognitive stimulation, 311 in control groups). The primary analysis was on changes that were evident immediately at the end of the treatment period. A few studies provided data allowing evaluation of whether any effects were subsequently maintained. A clear, consistent benefit on cognitive function was associated with cognitive stimulation (standardised mean difference (SMD) 0.41, 95% CI 0.25 to 0.57). This remained evident at follow-up one to three months after the end of treatment. In secondary analyses with smaller total sample sizes, benefits were also noted on self-reported quality of life and well-being (standardised mean difference: 0.38 [95% CI: 0.11, 0.65]); and on staff ratings of communication and social interaction (SMD 0.44, 95% CI 0.17 to 0.71). No differences in relation to mood (self-report or staff-rated), activities of daily living, general behavioural function or problem behaviour were noted. In the few studies reporting family caregiver outcomes, no differences were noted. Importantly, there was no indication of increased strain on family caregivers in the one study where they were trained to deliver the intervention. Authors' conclusions There was consistent evidence from multiple trials that cognitive stimulation programmes benefit cognition in people with mild to moderate dementia over and above any medication effects. However, the trials were of variable quality with small sample sizes and only limited details of the randomisation method were apparent in a number of the trials. Other outcomes need more exploration but improvements in self-reported quality of life and well-being were promising. Further research should look into the potential benefits of longer term cognitive stimulation programmes and their clinical significance. Plain language summary Can cognitive stimulation benefit people with dementia? People with dementia and their caregivers are often advised that 'mental exercise' may be helpful in slowing down the decline in memory and thinking experienced by many people with dementia. This review examined the evidence for one form of mental exercise, described as cognitive stimulation. This involves a wide range of activities that aim to stimulate thinking and memory generally, including discussion of past and present events and topics of interest, word games, puzzles, music and practical activities such as baking or indoor gardening. Typically this is carried out by trained staff with a small group of four or five people with dementia for around 45 minutes at least twice a week. Family caregivers have also been trained to provide cognitive stimulation to their relative on a one-to-one basis. This review included 15 trials with a total of 718 participants. The findings suggested that cognitive stimulation has a beneficial effect on the memory and thinking test scores of people with dementia. Although based on a smaller number of studies, there was evidence that the people with dementia who took part reported improved quality of life. They were reported to communicate and interact better than previously. No evidence was found of improvements in the mood of participants or their ability to care for themselves or function independently, and there was no reduction in behaviour found difficult by staff or caregivers. Family caregivers, including those who were trained to deliver the intervention, did not report increased levels of strain or burden. The trials included people in the mild to moderate stages of dementia and the intervention does not appear to be appropriate for people with severe dementia. More research is needed to find out how long the effects of cognitive stimulation last and for how long it is beneficial to continue the stimulation. Involving family caregivers in the delivery of cognitive stimulation is an interesting development and merits further evaluation. Background Interventions with a cognitive focus have long been used in dementia care. They have been developed in parallel with approaches emphasising the stimulation of the senses (Woods 1977). Reality Orientation (RO) (Taulbee 1966) was developed in the late 1950s as a response to confusion and disorientation in older patients in hospital units in the USA, and 2 / 69

3 was the prototype of the cognitive stimulation approach. Classes were held for 30 minutes once or twice per day. Basic personal and current information was presented to the patient and a variety of materials used, such as individual calendars, word-letter games, building blocks and large-piece puzzles. A Reality Orientation board would be used in each session and would list the name of the unit and its location, the day, date, weather, current events etc. The approach emphasised the engagement of nursing assistants in a hopeful, therapeutic process. The first controlled evaluation of RO was reported in the UK by Brook et al (Brook 1975). They found positive benefits on ratings of intellectual and social functioning in patients attending RO group sessions for 30 minutes per day, five days per week for four months compared with a control group who visited a special RO room daily but were not given any encouragement to engage with the materials or each other. A spate of controlled studies followed (Holden 1995), with outcome measures typically including assessments of orientation, other aspects of cognitive functioning and level of independent functioning. A Cochrane review specifically examining Reality Orientation (Spector 2000a; Spector 2000b) concluded that there was some evidence that RO had benefits for people with dementia on both cognition and behaviour. However, outside of a few countries (notably Italy), RO has been little practised or researched since 1990 and has attracted some criticism (Burton 1982; Dietch 1989; Powell-Proctor 1982), especially for being applied in a mechanical, inflexible, insensitive and confrontational manner. Doubts were also raised about the clinical significance of any improvements; the person with dementia might now know what day of the week it was but would this have any meaningful impact on the person's life? Such was the concern regarding the insensitive use of RO and other cognitive approaches that one influential set of guidelines on the management of dementia (APA 1997) cautioned against their use with the possibility of a negative impact on the person's well-being that outweighed any small cognitive improvements. In addition to the 'classroom' element of RO, from the early days '24 hour RO' was also advocated. This involves staff taking every opportunity to provide current information to the person, outside of the formal setting of the RO group, as well as using environmental features such as sign-posting and orientation boards to assist orientation. There have been some positive evaluations of the effects of training and sign-posting on orientation around a care facility (for example Hanley 1981; McGilton 2003) but the effects of 24 hour RO per se have been more difficult to evaluate. This is not least of all because of the difficulty in monitoring its implementation, with informal interactions being much more difficult to document than a formal group meeting. Williams 1987 reports using a modified form of 24 hour RO where staff were trained to respond appropriately to residents' requests for information, and found that 90% of staff responses to residents' requests complied with the treatment protocol. This study reported improvements in cognition, independent functioning and orientation, compared with a control group on a separate ward, when this form of 24 hour RO was implemented. Alongside the RO literature, in recent years there has been increasing discussion of 'cognitive stimulation'. In part this reflects a general view that lack of cognitive activity hastens cognitive decline, in normal ageing as well as in dementia (Breuil 1994; Small 2002), and in part it is an attempt to make use of the positive aspects of RO whilst ensuring that it is implemented in a properly sensitive and respectful manner (Spector 2001; Woods 2002). There has also been growing interest in the application of various forms of cognitive training and in teaching individual people with dementia to use memory aids and strategies to assist with their particular difficulties that have been identified with the person. A Cochrane review of cognitive training and cognitive rehabilitation in early stage dementia has been completed (Clare 2003) and it is important to ensure that clear definitions are used to avoid confusion between the various cognition-based approaches, as in the past 'training', 'stimulation' and 'rehabilitation' have been used almost interchangeably. Clare and Woods (Clare 2004) proposed the following definitions. Cognitive stimulation is engagement in a range of activities and discussions (usually in a group) aimed at general enhancement of cognitive and social functioning. Cognitive training is guided practice on a set of standard tasks designed to reflect particular cognitive functions; a range of difficulty levels may be available within the standard set of tasks to suit the individual's level of ability. It may be offered in individual or group sessions, with pencil and paper or computerised exercises. Cognitive rehabilitation is an individualised approach where personally relevant goals are identified and the therapist works with the person and his or her family to devise strategies to address these. The emphasis is on improving performance in everyday life rather than on cognitive tests, building on the person's strengths and developing ways of compensating for impairments Using these definitions, this review included studies on RO group sessions but not on 24 hour RO or direct training in spatial orientation. The primary outcomes examined will be in relation to the person's cognitive functioning. It is considered that this is the minimum expectation of a general approach with this focus. However, given the concerns discussed above of a possible negative effect, measures of quality of life, mood and well-being are highly pertinent secondary outcome measures. The effects on the person's general level of function in everyday life also need to be considered in evaluating the meaning of any changes observed for the individual and his or her supporters. The impact on family caregivers and careworkers is also important to consider as they are key partners in the process of care. Objectives To evaluate the effectiveness and impact of cognitive stimulation interventions aimed at improving cognition for people with dementia, including any negative effects. To indicate the nature and quality of the evidence available on this topic. To assist in establishing the appropriateness of offering cognitive stimulation interventions to people with dementia and identifying the factors associated with their efficacy. 3 / 69

4 Methods Cognitive stimulation to improve cognitive functioning in people with dementia Criteria for considering studies for this review Types of studies This review focused on randomised controlled trials (RCTs) for which adequate information was provided or could be obtained from the researchers. The studies included must have been published, written in English and presented in a peerreviewed journal article. Types of participants Participants with a diagnosis of dementia. The main diagnostic categories that were included were Alzheimer's disease, vascular dementia or mixed Alzheimer's and vascular dementia. These diagnostic categories were considered together. Older studies, included from the previous review of RO, used other terms for this population but were included where the review authors were satisfied that the included population would now be described as having a dementia. Participants with mild cognitive impairment, where the extent of cognitive impairment or its effects on day-to-day function were insufficient to justify a dementia diagnosis, were not included. Severity of dementia was indicated through group mean scores, range of scores, or individual scores on a standardised scale such as the Mini-Mental State Examination (MMSE) (Folstein 1975) or Clinical Dementia Rating (CDR) (Hughes 1982). All levels of severity were included. Qualifying participants received the intervention in a range of settings, including their own home, as outpatients and in day-care and residential settings. No specific restrictions regarding age were applied. Data from family caregivers were included where this was available and where the relationship between the caregiver and the person with dementia was specified, including whether they were co-resident. The number of participants receiving concurrent treatment with acetylcholinesterase inhibitors was documented, where possible. Types of interventions Studies were considered for this review if they described a cognitive stimulation intervention targeting cognitive and social functioning. These interventions may also have been described as RO groups, sessions or classes. The definition of cognitive stimulation as proposed by Clare 2004 was adopted. This meant that some studies which described their intervention as 'cognitive stimulation' were excluded. Interventions needed to offer exposure to generalised cognitive activities rather than training in a specific modality. Interventions were typically conducted in a group to enhance social functioning, or could involve family caregivers. Studies were included if a comparison was made to 'no treatment', 'standard treatment' or placebo. Standard treatment was understood to be the treatment that was normally provided to patients with dementia in the study setting and could include provision of medication, clinic consultations, contact with a community mental health team, day care, or support from voluntary organisations. Placebo conditions could consist, for example, of an equivalent number of sessions in which general support, but no structured intervention, was offered. The minimum duration of intervention for inclusion of a study was one month. There were no restrictions on the number of treatment sessions, although this was noted. Types of outcome measures Outcomes were considered in relation to the impact of the intervention on the person with dementia and on the primary family caregiver. Studies could present data in both these categories. Short term (immediately after the intervention) and medium term (follow-up one month to one year after the intervention finished) outcomes were considered. Outcomes for the person with dementia and the caregiver were considered where these were assessed using scores on standardised tests, rating scales and questionnaires. Rates of attrition and reasons for participants dropping out from the study were noted. Outcomes for the person with dementia Outcome measures for the person with dementia sought to identify whether changes were observed following the intervention. The following variables were considered as outcome measures for the person with dementia. Performance on at least one test of cognitive functioning (including tests of memory and orientation). Self-reported, clinically-rated or carer-reported measures for mood of the person with dementia. Self-reported or carer-reported quality of life or well-being measures for the person with dementia. Observer or carer ratings of everyday functioning (activities of daily living) of the person with dementia. Carer ratings of the participant's behaviour. Clinician or carer ratings of neuropsychiatric symptoms or behaviour problems of the person with dementia. Clinician or carer ratings of the social engagement of the person with dementia. 'Carer' in this context included care staff as well as family caregivers. Outcomes for the family caregiver The outcomes for the family caregiver that were considered included any of the following. Self-reported well-being, depression and anxiety. 4 / 69

5 Self-reported burden, strain and coping. Satisfaction with the intervention. Search methods for identification of studies We searched ALOIS ( the Cochrane Dementia and Cognitive Improvement Group Specialized Register, on 6 December The search terms used were: cognitive stimulation, reality orientation, memory therapy, memory groups, memory support, memory stimulation, global stimulation, cognitive psychostimulation. ALOIS is maintained by the Trials Search Co-ordinator of the Cochrane Dementia and Cognitive Improvement Group and contains studies in the areas of dementia prevention, dementia treatment and cognitive enhancement in healthy populations. The studies are 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: meta Register of Controlled Trials; Umin Japan Trial Register; WHO portal (which covers ClinicalTrials.gov; 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) (The Cochrane Library); 4. six-monthly searches of a number of grey literature sources: ISI Web of Knowledge Conference Proceedings; Index to Theses; Australasian Digital Theses. To view a list of all sources searched for ALOIS see About ALOIS on the ALOIS website. Details of the search strategies used for the retrieval of reports of trials from the healthcare databases, CENTRAL and conference proceedings can be viewed in the methods used in reviews section within the editorial information about the Dementia and Cognitive Improvement Group (CDCIG). Additional searches in each of the sources listed above, to cover the timeframe from the last searches performed for the Specialized Register to December 2011, were run to ensure that the search for the review was as up to date as possible. The search strategies used can be seen in Appendix 1. A total of 804 references were retrieved from the December 2011 search. After de-duplication and a first assessment, authors were left with 41 references to further assess for either inclusion, exclusion or discarding. Data collection and analysis Searches were conducted as detailed above to identify all relevant published studies. The date and time of each search, together with details of the version of the database used, were recorded. Additional information was sought, as outlined above, and hard copies of articles were obtained. Quality assessment RCTs were identified and the two review authors (BW and EA) worked independently to determine which studies met the criteria for inclusion. Trials that did not meet the criteria were excluded, and reasons for exclusion were noted in the table 'Characteristics of excluded studies'. Review authors' selections of trials were compared and the final list of included studies was reached by consensus. The selected RCTs were described in tabular form, permitting an evaluation of their methodological quality. Studies were assessed against a checklist of quality requirements using the Cochrane approach (see risk of bias tables). Grade A, '': adequate concealment (randomisation; concealed allocation). Grade B, 'Unclear risk': "randomised", but methods uncertain. Grade C, 'High risk': inadequate concealment of allocation or no randomisation, or both. Only trials with a grade A or B ranking were included in the review. Again, the review authors worked independently to ascertain which studies met the quality criteria, and consensus was reached through discussion. Attempts were made to obtain additional information from the study authors when further data were needed. Data extraction Data from the RCTs selected for inclusion were extracted, recorded and entered into RevMan. The summary statistics required for each trial and each outcome for continuous data were the mean change from baseline, the standard error of the mean change, and the number of patients for each treatment group at each assessment point. Where changes from baseline were not reported, the review authors extracted the mean, standard deviation and the number of patients for each treatment group at each time point, if available. The review authors calculated the required summary statistics from the baseline and post-treatment group means and standard deviations, assuming in this case a zero correlation between the measurements at the baseline and follow-up time points. This method overestimates the standard deviation of the change from baseline but this conservative approach was chosen as it is preferable in a meta-analysis. For binary data, the review authors sought the numbers in each treatment group and the numbers experiencing the outcome of interest. The baseline assessment was defined as the latest available assessment prior to randomisation, but no longer than two months prior. For each outcome measure, data were sought on every patient randomised. To allow an intention-to-treat analysis, the data were sought irrespective of compliance and whether or not the patient was subsequently deemed ineligible or otherwise excluded from treatment or follow-up. If intention-to-treat data were not available in the publications, 'on-treatment' data or the data of those who completed the trial were sought and were indicated as such. In studies where a cross-over design was used, only data from the first treatment phase after randomisation were eligible for inclusion. 5 / 69

6 As the outcomes measured in clinical trials of dementia and cognitive impairment often arise from ordinal rating scales, where the rating scales had a reasonably large number of categories (more than 10) the data were treated as continuous outcomes arising from a normal distribution. Data analysis The meta-analyses included the combination of data from trials that may not use the same rating scale to assess an outcome. Therefore, the measure of the treatment difference for any outcome used was the weighted mean difference when the pooled trials used the same rating scale or test and the standardised mean difference (the absolute mean difference divided by the standard deviation) when they used different rating scales or tests. The duration of the trials may vary considerably. If the review authors considered the range too great to combine all trials into one meta-analysis, they divided it into smaller time periods and conducted a separate meta-analysis for each period. Some trials may have contributed data to more than one time period if multiple assessments were made. The meta-analyses presented overall estimates of the treatment difference from a fixed-effect model and a test for heterogeneity was performed using a standard Chi 2 statistic. Where there was evidence of heterogeneity of the treatment effect between trials then a random-effects model was utilised (which results in broader confidence intervals than for those of a fixed-effect model). The review authors discussed and reached consensus on the interpretation of the statistical analyses, seeking specialist statistical advice from CDCIG as required. The review authors discussed and reached consensus on the presentation of the findings in the background to the review. Results Description of studies From the initial set of references identified by the updated systematic searches (since our last review, Spector 2000a), 53 potentially relevant studies were identified. Of these, seven studies met the inclusion criteria (Baldelli 2002; Bottino 2005; Chapman 2004; Onder 2005; Requena 2006; Spector 2001; Spector 2003) and were included in this review. Our previous review (Spector 2000a) included eight studies in the meta-analysis, six of which were included in this replacement review ( Baines 1987; Baldelli 1993a; Breuil 1994; Ferrario 1991; Wallis 1983; Woods 1979). For the two studies excluded at this stage, the data needed for the current analyses were not available (Gerber 1991; Hanley 1981). Full details of included studies and reasons for exclusion of excluded studies are presented in the tables 'Characteristics of included studies' and 'Characteristics of excluded studies'. The pre-publication search in December 2011 identified 41 further studies for consideration. Two further studies (Buschert 2011; Coen 2011) met the inclusion criteria and have been added to the review, with three further studies awaiting classification (Buettner 2011; Fernandez-Calvo 2010; Niu 2010). Overall, 718 participants, 407 in the treatment groups and 311 in the control groups, were included in the analyses of the 15 included studies. The included studies varied in many aspects: (1) participant characteristics; (2) number and duration of cognitive stimulation sessions; (3) activities which defined cognitive stimulation; (4) the activity of the control group; and (5) outcome measures. These factors will be considered in turn. 1) Participant characteristics Diagnosis: eight of the nine new studies specified the diagnostic criteria used. Coen 2011 simply described their participants as having mild to moderate dementia; Spector 2001 and Spector 2003 used DSM-IV criteria, but did not break the participants down by dementia subtype; Baldelli 2002 included similar numbers of participants with Degenerative senile dementia of the Alzheimer s type (SDAT) (N = 46) and vascular multi-infarct dementia (N = 41), although all had experienced at least one cerebrovascular accident resulting in motor deficits; Bottino 2005, Buschert 2011, Chapman 2004, Onder 2005 and Requena 2006 all specified a diagnosis of probable Alzheimer's disease (AD) according to NINCDS- ADRDA criteria linked with either ICD-10 or DSM-IIIR criteria. In these studies, participants were on a stable dose of an acetylcholinesterase inhibitor (ACHEI) (rivastigmine in the Bottino 2005 study; donepezil in the remaining studies apart from Buschert 2011 where a variety of medications including memantine were being taken). Amongst the earlier studies, Breuil 1994 specified DSM-III criteria for dementia and Baldelli 1993a stated that their participants were diagnosed with "Alzheimer s (SDAT) ". The four studies from 1991 and earlier specified more general criteria using cognitive measures that indicated that dementia diagnoses were justifiable (Baines 1987; Ferrario 1991; Wallis 1983; Woods 1979). Ten of the 11 most recent studies provided mean baseline scores on the Mini-Mental State Examination (MMSE), and the 11th (Ferrario 1991) provided the MMSE score range for participants. Coen 2011, Spector 2001 and Spector 2003 were the only studies with a mean MMSE score in the moderate range (10 to 20), perhaps reflecting the upper limit of 24 for participants to be included in these studies. The mean scores (16.9, 13.1 and 14.4 respectively) were several points lower than those in the eleven studies reporting mean scores (mean average 19.7), and outside the range of 18 to 25 specified by Ferrario However, those studies where the mean score was in the mild range (> 20) may well have included participants in the moderate or even the severe range, for example the lowest score in the Breuil 1994 study was reported to be 9. In general, however, it can be said that the studies included in this review had targeted participants in the mild to moderate range of cognitive impairment. The average age of participants was over 70 years in all studies (except Wallis 1983 where it was 69.8 years); in 6 studies it was over 80 years. The average mean age across the 15 studies was 78.8 years, with the range of ages that were reported from 38 to 97 years. Over half the studies reported inclusion of participant(s) aged 90 years and above. Participants were resident in care homes, nursing homes or hospitals, apart from six studies (Bottino 2005; Breuil 1994; 6 / 69

7 Buschert 2011; Chapman 2004; Onder 2005; Requena 2006) where all the participants were outpatients living in the community. The participants included in the Spector 2001 and Spector 2003 studies were recruited from both residential care homes and day centres, with the former being in the majority. 2) Length, number and duration of sessions The length of the intervention varied from four weeks (the minimum for inclusion in the review) to 24 months, with the stated length of sessions varying from 30 minutes to 90 minutes. In general, the sessions of longer duration were associated with the lowest frequency (once a week). The median session length across the studies was 45 minutes, and the median frequency was three times a week, ranging from once to five times a week. The total possible exposure to the intervention varied dramatically, from 10 to 12 hours (Baines 1987; Breuil 1994; Chapman 2004; Coen 2011; Spector 2001; Spector 2003 ) to 375 hours in the two-year (Requena 2006) study. Across the 15 studies, the median exposure time was 30 hours. Requena 2006 presented data from both the 12 month and 24 months time point in their study. As there was less attrition at the 12 month time point, and this was more comparable (although still longer) in duration to the other studies, the 12 month data were used in combination with other studies in the meta-analyses, with the 24 month data reported separately. 3) Activities during cognitive stimulation The level of detail provided in the published papers regarding the activities undertaken varied greatly. All studies used small group sessions, typically with groups of five to seven participants, with the exception of Onder 2005 where family caregivers were taught to carry out cognitive stimulation with the person with dementia on an individual basis. These individual sessions, led by family caregivers, included current information, topics of general interest, historical events and famous people, attention, memory and visuo-spatial exercises and the use of clocks, calendars and notes. Early studies described the use of an RO board and discussion of current orientating information through newspapers, photographs, calendars and clocks etc., with materials selected to stimulate all five senses (for example Baines 1987; Wallis 1983; Woods 1979). Breuil 1994 introduced a number of more specific cognitive activities including drawing, associating words, object naming and categorising. Spector 2006 provided a detailed session by session treatment manual for the approach used in their studies. Activities in their sessions were designed with four themes: (1) the senses, (2) remembering the past, (3) people and objects, and (4) everyday practical issues. Activities included naming objects and people, association of words, remembering the past, discussion of hobbies, activities and current affairs, using money, knowing the way around and orientation topics. This treatment manual was also used by Coen Chapman 2004 reported topics including current events, discussion of hobbies and activities, education regarding Alzheimer s disease, life story work, and links with daily life with groups of six to seven participants. Bottino 2005 described temporal and spatial orientation, discussion of interesting themes, reminiscence activities, naming people, planning of daily activities and use of calendars and clocks and other external memory aids. Requena 2006 described, for groups of five people with dementia, visual images being shown on a TV screen from a computer and that reflected seven themes: orientation, bodily awareness, family and society, caring for oneself, reminiscing, household activities, animals, people and objects. These were accompanied by questions for discussion. None of the included trials adopted 24 hour RO in addition to group sessions, although Bottino 2005 described involving family caregivers in encouraging the use of external memory aids at home, Buschert 2011 described the use of exercises and tasks to be carried out at home between sessions and Onder 2005 encouraged family caregivers to informally engage in reality-based communication with the person with dementia two or three times a day. 4) Control group(s) activities In the earlier studies, alternative group activities were offered that were of a social (Woods 1979) or diversional (Wallis 1983) nature. Baines 1987 offered an alternative treatment, reminiscence groups, but for the purposes of this review it was the notreatment group that was included in the analyses. 'Treatment as usual' or no treatment was the control condition in a number of studies (Baldelli 1993a; Breuil 1994; Coen 2011; Ferrario 1991; Spector 2001; Spector 2003). In those studies where participants were also taking ACHEIs, the control group were typically monitored in relation to the medication ( Chapman 2004; Bottino 2005; Onder 2005; Requena 2006). Requena 2006 reported that their control participants watched TV whilst the cognitive stimulation groups were in session. Baldelli 2002 engaged both the control and cognitive stimulation participants in a physical therapy programme. Buschert 2011 asked control participants to complete pencil and paper tasks at home, encouraged by monthly group meetings. 5) Outcome measures As a condition of inclusion, cognitive tests were used in all the studies. Eleven studies used the MMSE (Folstein 1975) and eight of the more recent studies also used the Alzheimer's Disease Assessment Scale - Cognitive (ADAS-Cog) (Rosen 1984 ). Unfortunately, only the 10 month follow-up data on these and other measures could be utilised from Chapman 2004 as it has not proved possible to obtain their data at earlier time points in an extractable form. Four studies used a self-report quality of life measure with participants with dementia (Buschert 2011; Chapman 2004; Coen 2011; Spector 2003), but again the Chapman 2004 data was not in a useable form. Five studies used a self-report depression measure (four making use of a version of the Geriatric Depression Scale: Yesavage 1983), and four studies used a depression or anxiety scale completed from carer reports as well as from interviews with the participants (such as the Cornell Scale for Depression in Dementia: Alexopoulos 1988). A variety of scales have been used to evaluate behaviour, with activities of daily living (ADL) scales used in four studies, general behaviour ratings in seven studies and problem behaviour scales used in three studies. Family caregiver outcome measures were used in three studies (Bottino 2005; Onder 2005; Spector 2001). A full list of the outcome measures used in the included studies can be found in the table Characteristics of included studies'. 7 / 69

8 Risk of bias in included studies Cognitive stimulation to improve cognitive functioning in people with dementia Details for each study are provided in the 'Characteristics of included studies' table. Allocation (selection bias) For a study to be included in this review, the review authors had to be satisfied that random allocation to treatment conditions had been used. To ascertain this, in several cases it was necessary to seek further information from the study authors (for example Baldelli 1993a; Baldelli 2002; Ferrario 1991; Requena 2006). Remote or computerised randomisation was only used in four of the most recent studies (Bottino 2005; Chapman 2004; Buschert 2011). Earlier studies described drawing names from a hat or a sealed container, where it was possible to obtain details of the randomisation procedure. Blinding (performance bias and detection bias) Performance bias With psychological interventions, unlike drug trials, it is impossible to totally blind participants and staff to treatment. Participants will often be aware that they are being treated preferentially and the staff involved may have different expectations of treatment groups. There may also be 'contamination' between groups in terms of group sessions not being held in separate rooms and staff bringing ideas from one group to another, so that control participants receive elements of cognitive stimulation. The latter effect would be reduced with clear therapeutic protocols, the existence of which was not clear in most study reports. In relation to contamination, Wallis 1983 and Baines 1987 both stated that the staff were unaware of the allocation of participants to groups, as they were removed from the ward setting for treatment, and several other studies described the groups being run in a separate or specific room (for example Ferrario 1991; Spector 2001; Spector 2003; Woods 1979). Detection bias Most studies took steps to ensure that at least part of the assessment of outcomes was carried out by assessors blind to treatment allocation. Only three studies (Baldelli 1993a; Baldelli 2002; Ferrario 1991) did not report blinding of assessors. Of course, even independent assessors may be given clues from participants during the assessments, but this was not reported as an issue in the studies reviewed here. Using independent assessors works well for evaluating changes in cognition or self-reported mood, well-being and quality of life. Ratings of day-to-day behaviour and function are typically carried out by care staff, who may be more difficult to keep blind to group allocation, unless the group sessions were carried out in a separate location to which all participants were taken. Incomplete outcome data (attrition bias) Only two studies described following an intention-to-treat analysis plan (Chapman 2004; Spector 2003). In contrast, Breuil 1994 stated that "All those who for any reason did not attend all evaluation and training sessions were eliminated", with five participants excluded on this basis (three from the cognitive stimulation group). All studies reported data on attrition. Given the nature of the condition and the age of the participants, attrition in several studies was remarkably small, with zero attrition recorded in six studies (Baines 1987; Baldelli 1993a; Baldelli 2002; Bottino 2005; Buschert 2011; Coen 2011), out of 180 participants. The largest attrition rate was reported by Wallis 1983 where there was 39% attrition in the group of participants with dementia. In this study patients who attended less than 20% of the group sessions were eliminated from the study. Requena 2006 reported 32% attrition but this was over a two year period. The two largest studies had rates of 19% (Onder 2005) and 17% (Spector 2003) over periods of six months and two months respectively. Other potential sources of bias The absence of detailed treatment protocols raised queries regarding the extent to which the cognitive stimulation was delivered as intended (having noted that there may have been differences in emphases between studies in any case). Several studies noted that staff received training or supervision, or both, in running the groups and, from an early study, Woods 1979 stated in a personal communication that "A sample of sessions were tape-recorded and rated to ensure compliance with the therapeutic protocol". More recently, Chapman 2004 described weekly meetings to ensure their treatment programme was implemented as designed. Subgroups were led by a licensed speech-language pathologist and three master's level speech-language pathology students; all underwent two hour training before the groups started and weekly meetings were held to ensure that the programme was implemented as designed. Onder 2005 described how family caregivers were trained by a multi-disciplinary team and given a manual and specific schedules for each session. No records were made, however, of how often caregivers did deliver the sessions, or how closely the manual was followed. Effects of interventions For meta-analyses we used RevMan 5.1. Cognition (See Figure 1) For the overall evaluation of the effects of cognitive stimulation on cognitive function, all 14 RCTs which included useable data immediately post-treatment were included, including a total of 658 people with dementia of whom 377 received cognitive stimulation and 281 received no treatment or a placebo treatment. As most studies included more than one measure of cognitive function, this analysis was conducted on the most extensive assessment included. For seven studies this was the ADAS-Cog, and for two each it was the MMSE and CAPE Information/Orientation scales. The overall effect size, the 8 / 69

9 standardised mean difference (SMD) was 0.41 (95% CI 0.25 to 0.57). This was a statistically significant finding (Z = 5.04, P < ). For the seven studies, including 434 participants, using the ADAS-Cog as an outcome measure (Figure 2), the mean difference between the cognitive stimulation and control groups was 2.27 points (95% CI 0.99 to 3.55), a statistically significant difference (Z = 3.48, P < ). In total, 10 studies involving 600 participants used the MMSE (Figure 3). The overall mean difference was 1.74 points (95% CI 1.13 to 2.36). Again, this was a statistically significant difference (Z = 5.57, P < ). These analyses were strongly influenced by two of the more recent studies, Spector 2003 and Onder 2005, which were relatively large and had smaller confidence intervals around their reported mean differences. Five of the older studies involving 81 participants used other measures of information or orientation (Figure 4); here the SMD was no smaller (SMD 0.45, 95% CI to 0.90) and was just statistically significant despite the much smaller numbers involved (Z = 1.93, P = 0.05). The largest effect sizes were seen at the 12 month time point in the Requena 2006 study (SMD 0.70 on ADAS-Cog) and the Baldelli 1993a study (SMD 0.99 on MMSE), both of which offered above average durations of exposure to cognitive stimulation. However, Breuil 1994, which offered only 10 hours exposure, also had an above average effect size (0.63 on global cognitive score) and other studies with longer exposure (for example Ferrario 1991) had below average effect sizes. The 24 month data from Requena 2006 indicated that effect sizes appeared to be maintained through continued exposure (ADAS-Cog SMD 0.66: MD points, 95% CI to 24.85; MMSE SMD 0.56: MD 5.99 points, 95% CI to 13.56). However, these effects require replication as the confidence intervals were broad and crossed zero. Communication and social interaction (See Figure 5) Four studies, involving 223 participants, included staff ratings of the person's communication and social interaction (outside of the cognitive stimulation group), three using the Holden Communication Scale. The overall effect size (SMD) was 0.44 (95% CI 0.17 to 0.71) with participants in the cognitive stimulation groups showing a significant improvement in this area (Z = 3.15, P = 0.002). Spector 2003 was the most influential study in this analysis, although the effect was not reported as significant in the primary study report. Well-being and quality of life (See Figure 6) Four studies, involving 219 participants, included relevant self-report measures. Baines 1987 used the Life Satisfaction Index and Spector 2003, Buschert 2011 and Coen 2011 used the QoL-AD. The meta-analysis indicated that cognitive stimulation was associated with a significant benefit to well-being and quality of life compared with no treatment (SMD 0.38, 95% CI 0.11 to 0.65) (Z = 2.76, P = 0.006). The Spector 2003 findings were again a major influence. Mood (See Figure 7; Figure 8) Five studies, involving 201 participants, used a self-report measure of mood (the Geriatric Depression Scale or the MADRS). Cognitive stimulation was not associated with a clear improvement in mood across these studies. The SMD was 0.22 (95% CI to 0.53) (Z = 1.42, P = 0.16). Staff ratings of mood and anxiety similarly did not show any benefit from cognitive stimulation. Four studies, involving 239 participants, contributed to this analysis, two using the Cornell Scale for Depression in Dementia, a third using a subscale of the MOSES scale and the fourth the Rating of Anxiety in dementia. The SMD was close to zero in this domain (SMD 0.05, 95% CI to 0.31). Behaviour (See Figure 9; Figure 10; Figure 11) Three separate meta-analyses were conducted in this domain. One focused on activities of daily living (ADL) and basic selfcare skills; a second focused on behaviours seen as a problem, such as irritability, being demanding and difficult. The third included 'general' behaviour rating scales, which may include some of the previous two aspects, together with some higher level daily living skills. Four studies, involving 160 participants, used ADL scales. There was no benefit identified with cognitive stimulation (SMD 0.21, 95% CI to 0.47) (Z = 1.56, P = 0.12). Three studies, including 166 participants, used scales evaluating behaviour problems. Again there was no difference related to cognitive stimulation (SMD -0.14, 95% CI to 0.17) (Z = 0.86, P = 0.39). General behaviour rating scales showed a similar picture, with no difference emerging. Eight studies, including 416 participants, reported data on relevant scales (SMD 0.13, 95% CI to 0.32) (Z = 1.30, P = 0.20). Caregiver outcomes (See Figure 12) Three studies reported on outcomes for family caregivers. The largest of these (Onder 2005) taught family caregivers to deliver the cognitive stimulation, so the effects on caregivers were especially pertinent for that study. The effect sizes for anxiety, depression and caregiver burden were all close to zero (SMDs 0.11, 0.04, respectively), with confidence intervals crossing zero indicating no differences between the caregivers in the cognitive stimulation condition and those in the control conditions. Only one study (Baines 1987) reported outcomes for care staff, so no meta-analysis was possible. The results of this study indicated a significant increase in staff knowledge about residents participating in the cognitive stimulation intervention 9 / 69