RESEARCH ARTICLE
The effect of exercise on behavioural and psychological symptoms of dementia: the EVIDEM-E randomised controlled clinical trial† David Lowery1,2, Arlinda Cerga-Pashoja1,2, Steve Iliffe2, Ingela Thuné-Boyle2, Mark Griffin2, James Lee5, Alex Bailey1, Rahul Bhattacharya4 and James Warner1,3 1
Central and North West London NHS Foundation Trust (CNWL), London, UK Primary Care and Population Health, University College London (UCL), London, UK 3 Imperial College London, London, UK 4 East London NHS Foundation Trust (ELFT), London, UK 5 Independent exercise instructor, Brighton, UK Correspondence to: Dr D. Lowery, E-mail: [email protected] 2
†
Trial Registration: ISRCTN01423159
Objective: The
objective of this study is to evaluate the effectiveness of a simple dyadic (person with dementia and their main carer) exercise regimen as a therapy for the behavioural and psychological symptoms of dementia. Method: A two arm, pragmatic, randomised, controlled, single-blind, parallel-group trial of a dyadic exercise regimen (individually tailored walking regimen designed to become progressively intensive and last between 20–30 min, at least five times per week). Community-dwelling individuals with ICD-10 confirmed dementia with the following: clinically significant behavioural and psychological symptoms, a carer willing and able to co-participate in the exercise regimen, and no physical conditions or symptoms that would preclude exercise participation were invited by mental health or primary care services into the study. Results: One hundred and thirty-one dyads were recruited to this study. There was no significant difference in Behavioural and Psychological Symptoms as measured by the Neuropsychiatric Inventory at week 12 between the group receiving the dyadic exercise regimen and those that did not (adjusted difference in means (intervention minus control) = 1.53, p = 0.6, 95% CI [ 7.37, 4.32]). There was a significant between-group difference in caregiver’s burden as measured by the Zarit Caregiver Burden Inventory at week 12 (OR = 0.18, p = 0.01, CI [0.05, 0.69]) favouring the exercise group. Conclusions: This study found that regular simple exercise does not appear to improve the behavioural and psychological symptoms of dementia, but did seem to attenuate caregiver burden. Further study to improve exercise uptake are needed. Copyright # 2013 John Wiley & Sons, Ltd. Key words: trial; dementia; exercise; physical activity; behavioural and psychological symptoms History: Received 12 June 2013; Accepted 17 November 2013; Published online 11 December 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/gps.4062
Introduction Dementia is associated with a cluster of non-cognitive symptoms and behaviours that are an integral part of the syndrome (Alzheimer, 1907). Commonly described as behavioural and psychological symptoms of dementia (BPSD), they include disturbances of perception, Copyright # 2013 John Wiley & Sons, Ltd.
thought content, affect or behaviour (Finkel, 2003). Over 80% of people with dementia experience BPSD at some point during the course of their illness (Aalten et al., 2003), and they are commonly associated with reduction in the quality of life for the person as well as their carers (Colerick and George, 1986); increase of caregiver stress (Finkel et al., 1996), costs of care Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
820
(Rabins et al., 1982); and premature institutionalisation (Cohen-Mansfield, 1995). Importantly, behaviours that challenge care givers—particularly agitation, psychosis and sleep disturbance, often lead to institutionalisation and increased service utilisation (Donaldson et al., 1998; Dunkin and Anderson-Hanley, 1998; Schur and Whitlatch, 2003; Miller et al., 2010). Although there is some evidence supporting the treatment of BPSD with antipsychotic drugs, there have been increasing concerns about the safety of these drugs for people with dementia. Several studies have indicated that there is a long-term risk of mortality, accelerated cognitive decline and impaired quality of life in patients with dementia who take antipsychotic medication (Kales et al., 2007; Ballard et al., 2009). The UK’s clinical regulatory body—the National Institute for Health and Clinical Excellence—recommends the use of antipsychotic medications is restricted to cases in which severe risk of harm or distress is likely (NICE, 2006). There are few other safe pharmacological interventions with a proven evidence base. Non-pharmacological alternatives are being increasingly explored, and exercise has appeared to be a promising candidate (Brodaty and Arasaratnam, 2012). As well as the physical benefit, exercise has been demonstrated to have a positive effect on cognitive symptoms and mood in other populations (Berger and Motl, 2000) and could provide other benefits such as a sense of empowerment. There have been few robust clinical trials of exercise, and conflicting metaanalyses (Heyn et al., 2004; Forbes et al., 2008). Heyn et al., reported a positive effect of exercise on BPSD (ES = 0.54; 95% CI, 0.36–0.72) (Heyn et al., 2004); however, this does not provide a full picture of the effectiveness of exercise for BPSD (Thune-Boyle et al., 2012). There is considerable heterogeneity among trialled interventions, and exercise is often combined with other behavioural interventions. Some exercise regimens are complex and required a degree of physical fitness that would preclude many older adults with complex physical problems and moderate or significant dementia from performing them; finally, many trials are relatively underpowered. Method
Trial design
We conducted a pragmatic, randomised, controlled, single-blind, parallel-group trial of a dyadic exercise regimen (tailored walking) for community-dwelling individuals with dementia and their carers. The trial was designed following the UK Medical Research Council’s guidance on the development and testing of complex interventions (Craig et al., 2008). Individuals with a diagnosis of dementia or suspected dementia who had a carer willing and able to be a co-participant in the exercise regimen were eligible for recruitment. We registered and reported the trial design prior to recruitment (Trial Registration: ISRCTN01423159) (Cerga-Pashoja et al., 2010). The study opened for recruitment in January 2010 and final follow-up was in December 2012. After screening and consent, participants were randomly allocated (ratio 1 : 1) to one of two arms: the treatment arm, which received an individually tailored regimen of walking in addition to TAU; or the control arm, which received TAU only. A simple randomisation protocol was performed by an independent randomisation officer using a computer algorithm. Allocation was communicated to participants, carers, the participant’s General Practitioner and the therapist; other study personnel were kept blind to randomisation status until completion of the analysis. Blinding of group allocation from participants was impossible, but we adopted several strategies to minimise de-blinding of research staff, and these are discussed in the published trial protocol (CergaPashoja et al., 2010). Funding was provided by the UK’s National Institute for Health Research (RP-PG-0606-1005), sponsorship by CNWL NHS Foundation Trust, and ethical approval by Outer North East London Research Ethics Committee (ref 09/H0701/67). The study was overseen by an independent steering group including a data and safety monitoring subcommittee. Informed consent was sought from all participants. When an individual lacked capacity to consent, assent was sought and an appropriate consultee was asked for an opinion on the individual’s involvement and assent was periodically reassessed (Department of Constitutional Affairs, 2007).
Objectives Study population
The primary objective was to determine the effectiveness of a dyadic exercise regimen delivered through a programme of incremental walking for treating BPSD compared with treatment as usual (TAU). Copyright # 2013 John Wiley & Sons, Ltd.
This study was conducted in several inner city, urban and semi-rural locations in and around London. Participants were recruited either directly from the Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia
North Thames Dementias and Neurodegenerative Diseases Research Network’s dementia research register (Iliffe et al., 2011), by self-referral or via primary clinical services or specialist mental health services (e.g., memory assessment and community mental health). Participants with a clinical diagnosis of dementia (confirmed using ICD-10 Diagnostic Criteria for Research (1992)) with at least one significant BPSD symptom defined by the Neuropsychiatric Inventory (NPI) (NPI minimum score of frequency = 2, and severity = 2) (Cummings et al., 1994) were eligible for the trial. For the purpose of the trial, we excluded the NPI domains of delusions and hallucinations from screening and outcome as we did not anticipate exercise would affect these symptoms. A risk assessment protocol to check individuals’ suitability for the exercise programme included the falls risk assessment tool and timed unsupported standing tool (Studenski et al., 1994; Nandy et al., 2004) and primary care physician case note review.
821
Interventions
Physical exercise was delivered as an individually tailored regimen of walking designed to become progressively more intensive and last between 20–30 min. This was facilitated by one registered exercise professional qualified in instructing physical activity and exercise (National Vocational Qualification Level 3) and delivered to participants in the treatment arm of the trial in and around their own home. To help distinguish between a therapist effect operating through social contact and the effect of physical activity, the intervention was divided into two 6-week epochs. The exercise therapist progressively withdrew support at the end of the first epoch, with the expectation that the dyad would perform the exercise regimen regularly and independently at least five times per week (Table 1). Outcomes were measured at the end of both epochs. All participants were asked to record their daily activities throughout the 12 weeks of participation using a diary designed for this study. The intervention group
Table 1 Administration of outcome measures Day
Administrator
0
Research worker
0 1–2 3–4 5–8 6–8 13–17 17–28 40–46
Randomisation Exercise therapist Exercise therapist Exercise therapist Exercise therapist Exercise therapist Exercise therapist Independent researcher
40–46
Research worker
40–46 80–88
Exercise therapist Independent researcher
80–88
Research worker
80–88 90–99 182–196
Exercise therapist Research worker Research worker
Intervention/assessment ICD-10 DemQol MMSE GHQ NPI ZBI Demographics Medication CSRI Vital signs Allocation and diaries RPE timed walk, diarya RPE timed walka Telephone contact—advice and guidancea RPE timed walka Telephone contact—advice and guidancea Telephone contact as neededa NPI Diary reminder DemQol GHQ ZBI Medication Vital Signs RPE timed walka NPI Diary reminder DemQol GHQ ZBI Medication Vital Signs RPE timed walka Collection of diaries Telephone contact Check for change of status (residence and mortality)
ICD, International Classification of Diseases; MMSE, Mini Mental State Examination; NPI, Neuropsychiatric Inventory; CSRI, Client Service Receipt Inventory; RPE, Rating of Perceived Exertion; GHQ, General Health Questionnaire; ZBI, Zarit Caregiver Burden Inventory. a Treatment group only.
Copyright # 2013 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
822
diary contained an additional visual analogue scale, the rating of perceived exertion (RPE) (Borg, 1982). Perceived exertion is based on the physical sensations a person experiences during exercise, including increased heart rate, respiration or breathing rate, sweating, and muscle fatigue. Although this is a subjective measure, an exertion rating may provide a reasonable estimate of the actual heart rate during exercise (Borg, 1982). The therapist trained participant dyads in its use and validated their ratings. Participants were encouraged to extend (or reduce in some circumstance) the level of intensity to between 12 and 14 on the RPE scale. At this level, individuals should be able to respond verbally to questions, but may have to pause to take another breath more frequently than they would normally. Beyond 14, individuals would struggle to respond verbally and maintain pace. Study endpoints
Both the treatment and control arms were assessed for all outcomes at 6 and 12 weeks after randomisation. Further telephone contact took place at 26 weeks to assess additional adverse events, mortality, change in domiciliary status and adherence to exercise regimen (Table 1). The primary endpoint was a between-group difference of proportions of people with a reduction of three or more points on the composite NPI score (Cummings et al., 1994) at 12 weeks follow-up. This was chosen as a clinically meaningful change in BPSD symptoms. Secondary outcome measures were: mean differences in NPI score, the General Health Questionnaire (GHQ) (Goldberg, 1972), DemQOL-Proxy (DemQol) (Smith et al., 2005), and Zarit Caregiver Burden Inventory (ZBI) (Bedard et al., 2001). Participants’ level of exercise and compliance with the intervention were measured by carers using daily diaries and the RPE scale. Analytic methods
Assuming 80% of patients in the control group will have BPSD symptoms measured by the NPI at the 12 week follow-up and based on an anticipated between-group (control minus intervention) absolute risk difference of 30% in the proportion of people with BPSD, we calculated that a sample size of n = 116 participants would provide 90% power to detect this difference with a 5% (two-sided) significance level. To allow for anticipated attrition, the target sample was set at n = 146. Copyright # 2013 John Wiley & Sons, Ltd.
A statistical analysis plan was written and locked before data analysis began. Descriptive statistics are presented as numbers and percentages for categorical variables and means with standard deviations for continuous data. Where assumptions for normality and linearity were met, analysis of covariance adjusting for baseline scores was used for between-group comparisons of continuous data (intervention-control). Analysis was conducted on an intention-to-treat basis with available data. At endpoint, participants were dichotomised according to a clinically significant reduction (Baseline—Week 12) of BPSD (Composite NPI score change ≥ 3 points); and validated thresholds for ‘caseness’ relating to psychological wellbeing (GHQ) (0–24; optimal cut-off ≥23) (Goldberg et al., 1997) and for care giver burden (ZBI) (0–28; optimal cut-off ≥24) (Schreiner et al., 2006); change in domiciliary status and mortality. Binary logistic regression was used to analyse binary outcomes. All analysis was performed blind by the researcher and overseen by the study statistician using the statistics package SPSS (v20) (IBM, Armonk, New York, USA). Results Description of the sample
One hundred and thirty-one participant dyads were randomised to either receive the exercise regimen in addition to TAU (treatment), or to receive TAU only (control). Eighty nine percent of dyads (n = 116) completed the trial, with seven and eight dyads lost from the control group and the intervention groups, respectively (Figure 1). Descriptive statistics for the baseline demographic and the outcome data are presented for each group (Table 2). The groups were relatively similar in terms of age, type of dementia and other characteristics (Table 2). Behavioural and psychological symptoms as measured by the NPI were similar in prevalence to those reported in the original validation paper (Cummings et al., 1994). At baseline, 36 carers were at or above the validated threshold for ‘caseness’ for psychological morbidity on the GHQ and 25 for care giver burden on the ZBI. Analysis of primary outcome
Although the intervention group were 41% more likely to reach a clinically significant reduction of three or more points of NPI composite score by week 12, there was no significant difference between the groups (OR = 1.41, p = 0.36, 95% CI [0.67, 3.01]). Using Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia
823 Screened n = 474*
Excluded n = 82 No BPSD (n=39) Falls Risk (n=29) No carer (n=8) GP Excluded (n=4) No dementia (n=2) Declined n=252
Randomised n = 131
Allocated to Treatment as Usual N = 64
Allocated to Treatment as Usual N = 67
Lost to follow up Withdrew
Included in Primary Analysis N = 57
n=3 n=4
Lost to follow up Withdrew Died
n=1 n=5 n=2
Included in Primary Analysis N = 59
* Confirmation of ability to form a dyad was part of screening process. Therefore, numbers screened are presented as individuals with dementia only.
Figure 1 Participant Dyad Flow. Table 2 Socio-demographic statistics (baseline) Control (n = 64) n Person with Dementia Age Gender: female Ethnicity: white Marital status: married FRAT Score: ≥2 Living: private residence Years Education MMSE Alzheimers Years since diagnosis: ≤2 years Carer Age Gender: female FRAT score: =0 Relationship: ‘partner or spouse’ Carer distress (NPI)
Mean ± SD (range) or frequency (percentage)
Intervention (n = 67) n
Mean ± SD (range) or frequency (percentage)
64 64 64 63 64 64 60 64 59 61
78 ± 7.4 (58–99) 39 (60.9%) 50 (78.1%) 45 (71.4%) 36 (56.3%) 57 (89.1%) 11.92 ± 5.9 (0–36) 14.9 ± 8.7 (0–29) 38 (64.4%) 38 (64.4%)
67 67 67 67 67 67 63 67 65 59
79 ± 6.8 (64–97) 35 (52.2%) 56 (83.6%) 46 (68.6%) 27 (40.3%) 59 (88.1%) 12.1 ± 4.1 (6–23) 16.3 ± 7.4 (0–30) 44 (67.7%) 35 (54.7%)
59 64 64 64 64
60.9 ± 17 (22–88) 39 (60.9%) 46 (71.9%) 35 (54.7%) 11.9 ± 8.1 (0–39)
53 67 67 67 67
65.4 ± 14.9 (27–89) 50 (74.6%) 40 (59.7%) 42 (62.7%) 11.8 ± 8.9 (0–38)
FRAT, Falls Risk Assessment Tool; MMSE, Mini Mental State Examination; NPI, Neuropsychiatric Inventory.
Copyright # 2013 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
824
analysis of covariance adjusting for baseline covariates, no significant difference was found between the groups of mean NPI at week 12 (intervention-control) (β = 1.53, p = 0.6, 95% CI [ 7.37, 4.32])(Table 3).
week) was achieved by 30.77% of the treatment group, and prescribed intensity (RPE, 12–14) was achieved in 53.25% of the walks. Safety and adverse events
Analysis of secondary outcomes
Caregiver’s burden as measured by the ZBI doubled from 16% at baseline to 32% by week 12 for the control group participants but decreased from 23 to 17% for those in the intervention group; (OR = 0.18, p = 0.01, CI [0.05, 0.69]). There were no statistically significant differences between the groups at week 12 in carers’ mental health (measured using the GHQ), carers’ distress as measured by the NPI and quality of life of participants with dementia (measured using the Demqol-proxy), as shown in Table 3. There were no significant differences between groups at week 26 in change of domiciliary status (OR = 0.39, p = 0.42, CI [0.38, 3.85]), or mortality (OR = 1.76, p = 0.55, CI [0.28, 11.08]). Treatment
There were no significant differences between the groups of self-reported walking time (minutes) at baseline (β = 5.2, p = 0.09, 95% CI [ 11.33, 0.93]), week 6 (β = 0.55, p = 0.82, 95% CI [ 5.32, 4.21]) and week 12 (β = 1.18, p = 0.697, 95% CI [ 4.81, 7.17]). Prescribed frequency of walks (five times per
There were six reported falls (two in the control arm and four from the intervention), but none required hospitalisation. There were also five reported deaths, three in the intervention arm and two in the control. Two deaths were reported during the intervention period and the other three at long-term follow-up. All adverse events were reviewed by a senior clinician and were deemed not to be related to trial participation. Discussion Previous studies have reported that exercise is beneficial in reducing some BPSD, especially depressed mood and agitation and may also improve sleep and reduce ‘wandering’. However, research evaluating the efficacy of exercise as a therapeutic tool for other important symptoms such as anxiety, apathy and repetitive behaviours is limited (Thune-Boyle et al., 2012). This study shows that physical exercise tailored to participant-carer dyads, designed to become progressively intensive and lasting between 20–30 min per day made no significant difference to NPI score after 12 weeks. This contradicts the metaanalysis reported by Heyn et al., (Heyn et al., 2004).
Table 3 Between group comparison at weeks 6 and 12 Outcome NPI GHQ ZBI
NPI Demqol NPI carer distress
Week
Control frequency (percentage)
Intervention frequency (percentage)
Baseline 6 12 Baseline 6 12 Baseline 6 12
64/64 (100%) 32/56 (57.1%) 33/57 (57.9%) 46/63 (73%) 24/57 (42.1%) 24/56 (43%) 52/62 (83.9%) 14/55 (25.5%) 18/56 (32%)
67/67 (100%) 39/62 (62.9%) 39/59 (66.1%) 48/67(71.6%) 16/62 (25.8%) 17/55 (31%) 50/65 (76.9%) 14/59 (23.7%) 10/57 (17.5%)
6 12 6 12 6 12
Mean ± SD 26.6 ± 17.5 25.6 ± 16.6 101.1 ± 14.9 101 ± 13.5 11.07 ± 7.2 9.98 ± 5.9
Mean ± SD 25.7 ± 20.5 23.9 ± 20.6 103.6 ± 11.9 104 ± 10 11.5 ± 8.5 10.9 ± 9.3
OR†
95% CI
1.27a 1.41a
0.61 : 2.66 0.67 : 3.01
0.52 0.36
0.42 b 0.59 b
0.18 : 1.00 0.24 : 1.43
0.05 0.19
0.48 b 0.18 b
0.14 : 1.67 0.05 : 0.69
0.25 0.01
β* 0.81 1.53 1.27 2.62 0.06 1.14
6.08 : 4.45 7.37 : 4.32 2.33 : 4.86 0.78 : 6.02 2.25 : 2.14 1.31 : 3.58
0.76 0.60 0.49 0.09 0.96 0.76
p
†
OR of aImprovement (reduction) of composite NPI score of 3 or more points between baseline and weeks 6 and 12 bScoring at or above the validated threshold for ‘caseness’ at weeks 6 and 12. *Analysis of covariance of the adjusted difference in means (intervention-control) at weeks 6 and 12. The denominator of week 6 and week 12 data varies as result of participant availability.
Copyright # 2013 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia
Although overall BPSD were lower at week 12 in the group that received the exercise intervention in comparison with the control group, this difference was not statistically significant. This may be related to the poor adherence to the full exercise regimen, and/or the duration of the programme; longer follow-up periods and methods for increasing adherence might result in a significant effect, and this warrants further investigation. A large proportion of the TAU group experienced an improvement in BPSD over the course of the study, which may be a Hawthorne effect attributable to the engagement of research staff with participants (McCarney et al., 2007), and/or the natural fluctuation of BPSD. An alternative explanation is confounding due to changes of exercise behaviour in the control group. More objective measurement of control participants’ exercise behaviour may have allowed us to test for this, but would have brought its own methodological challenges (see succeeding text). Caregiver’s burden as measured by the ZBI decreased significantly in the intervention arm but increased in the control arm. This finding is potentially important because caregiver burden mediates decisions to relocate the person with dementia to a care home (Dunkin and Anderson-Hanley, 1998) and is associated with greater use of services (Miller et al., 2010); consequently, it may provide a useful proxy measure of both a longer-term outcome and future service use. It seems somewhat counter-intuitive that declining caregiver burden was attenuated in the treatment group given the absence of effect for BPSD. Although, this may have been as a consequence of the beneficial effects of exercise direct to the carer, rather than because of any change in the person with dementia. Feelings of mastery (or lack of in the control group), because of an anticipated positive effect of the intervention is one possible explanation; this could have profound ethical implications for the design of future research and should be explored along with other potential mechanisms and long-term effects. This will also be important for optimising this intervention and the development of other interventions. Strengths and limitations of this study include strengths, the relatively large sample and low attrition rates providing sufficient power to detect an effect (positive or negative) of clinically significant magnitude; the recruitment of a clinically relevant sample in terms of diagnosis, severity of disease and mobility allowing greater generalisability; use of well validated and blinded outcome measures; the use of a pragmatic exercise programme, which does not require specialist Copyright # 2013 John Wiley & Sons, Ltd.
825
training or equipment; and efforts to control for the impact of therapist contact. This study has shown that an acceptable exercise intervention for people with dementia can be developed and tested. The limitations of the study include the single-blind design and the weakness of self-report in measuring exercise. The single-blind design had the inherent and unavoidable risk of affecting control participants’ exercise behaviour. Additionally, there was also a risk that participants would divulge information about the group allocation to the researcher. We attempted to reduce this risk through the inclusion of a second, independent, researcher who collected primary outcome data (NPI) at 6 and 12 weeks. Assessing physical activity with minimum intrusion to study participants is challenging and we relied on participants’ self reports through daily diaries. Self-report is prone to information bias, especially when participants are not blinded to their group allocation (Bauhoff, 2011). We gave considerable thought to the issue of measuring physical activity, in particular, GPS receivers, actigraphs and pedometers. We decided not to use such instruments to avoid excessive intrusion and prevent use of such instruments affecting exercise behaviour in their own right. Phasing out the effect of psychosocial contact (from the exercise therapist) from the exercise intervention appeared to allow a relatively successful control for contact between participants and the exercise therapist—results were consistent across both time periods. In addition, this was a complex intervention in which the social contact between the person with dementia and their carer formed an important component. Dyadic walking may have encouraged psychosocial support from the carer, and this may partly explain the changes we observed for caregiver burden. Conclusions A regular simple exercise such as walking does not appear to be effective for reducing BPSD but does appear to help reduce carer burden. However, we cannot be sure, particularly as increasing physical activity in people with dementia and their carers is difficult. The intervention group were 41% more likely to experience a clinically important reduction of NPI score, but this was not a significant difference. Given that less than half the intervention group achieved the prescribed exercise frequency and intensity targets, further studies attempting to increase uptake are warranted, as are longer studies to measure the effects of reduced carer burden. Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
826
Conflict of interest None to declare. The study sponsor had no role in the study design, in the collection, analysis and interpretation of data, in the writing of the report and in the decision to submit this paper for publication. Key points
• •
This study does not provide any evidence that exercise is an effective therapeutic intervention for the behavioural and psychological symptoms of dementia. Dyadic exercise supported by up to 3 hours therapist input significantly attenuates worsening caregiver burden.
Ethics This study adhered to and obtained the requisite approvals necessary to conduct research with patients in the NHS. Ethical approval was given by Outer North East London Research Ethics Committee (ref 09/H0701/67). Permissions (CSP20305) were obtained to recruit participants from the following Sites: Central & North West London NHS Foundation Trust, Surrey and Borders Partnership NHS Foundation Trust, West London Mental Health Trust, East London NHS Foundation Trust; and Participant Identification Centres: Camden & Islington NHS Foundation Trust and Barnet, Enfield and Haringey Mental Health NHS Trust. Acknowledgements We gratefully acknowledge the contribution of all study participants and their carers. The study was supported by a proactive and valuable steering group (Chaired by Dr Craig Ritchie); a wide number of recruitment hubs (North Thames DeNDRoN, Surrey and Borders Partnership NHS Foundation Trust—Dr Nilforooshan, and West London Mental Health NHS Trust—Dr Regan); and several volunteers who kindly provided their time in exchange for experience of working on a research trial (Mr Aeron-Smith, Miss Dimitrova, Miss Theobald, Miss Pulford, Ms Mystris, Ms Sula, Dr Jordanova, Dr Rajenthran and Dr Holzer). The study was supported by Central and North West London NHS Foundation Trust. This report presents independent research funded by the National Institute for Health Research under Copyright # 2013 John Wiley & Sons, Ltd.
its Programme Grants for Applied Research scheme (RP-PG-0606-1005). The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. Author contributions J. W. conceived of the idea of the study and together with S. I. obtained funding for it. J. W., along with D. L. and A. P., had primary oversight of the design and day to day management of the study. A. P. was responsible for data collection, A. P., D. L. and M. G. led the analysis of data. D. L. wrote the first draft of this paper. All other authors, along with the steering group, made essential contributions to the design of the study in line with their individual expertise. All authors contributed to the drafting and/or critical revision of this manuscript for important intellectual content, approval of the final version, and take public responsibility for its content. References Aalten P, DeVught ME, Lusberg R, et al. 2003. Behavioural symptoms in dementia: a factor analysis of the neuropsychiatric inventory. Dement Geriatr Cogn Disord 15: 99–105. Alzheimer A. 1907. Über eine eigenartige Erkrankung der Himrinde. Allgemeine Zeitschrift fur Psychiatrie und Psychisch-gerichtliche Medizin. Ballard C, Hanney ML, Theodoulou M, et al. 2009. The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. Lancet Neurol 8(2): 151–157. Bauhoff S. 2011. Systematic self-report bias in health data: impact on estimating cross-sectional and treatment effects. Health Serv Outcomes Res Method 11(1–2): 44–53. Bedard M, Molloy DW, Squire L, et al. 2001. The Zarit Burden Interview: a new short version and screening version. Gerontologist 41(5): 652–657. Berger BG, Motl RW. 2000. Exercise and mood: a selective review and synthesis of research employing the profile of mood states. J Appl Sport Psychol 12(1): 69–92. Borg G. 1982. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 14(5): 377–381. Brodaty H, Arasaratnam C. 2012. Meta-analysis of nonpharmacological interventions for neuropsychiatric symptoms of dementia. Am J Psychiatry 169(9): 946–953. Cerga-Pashoja A, Lowery D, Bhattacharya R, et al. 2010. Evaluation of exercise on individuals with dementia and their carers: a randomised controlled trial. Trials 11: 53. Cohen-Mansfield J. 1995. Assessment of disruptive behavior/agitation in the elderly: function, methods, and difficulties. J Geriatr Psychiatry Neurol 8(1): 52–60. Colerick EJ, George LK. 1986. Predictors of institutionalization among caregivers of patients with Alzheimer’s disease. J Am Geriatric Soc 34: 492–498. Craig P, Dieppe P, et al. 2008. Developing and Evaluating Complex Interventions: new Guidance. Medical Research Council: London. Cummings JL, Mega M, Gray K, et al. 1994. The neuropsychiatric inventory: comprehensive assessment of psychopathology in dementia. Neurology 44(12): 2308–2314. Department of Constitutional Affairs. 2007. Mental Capacity Act 2005: Code of Practice. TSO: Birmingham. Donaldson C, Tarrier N, Burns A. 1998. Determinants of carer stress in Alzheimer’s disease. Int J Geriatr Psychiatry 13(4): 248–256. Dunkin JJ, Anderson-Hanley C. 1998. Dementia caregiver burden: a review of the literature and guidelines for assessment and intervention. Neurology 51(1 Suppl 1): S53–60; discussion S65-57. Finkel SI. 2003. Behavioral and psychologic symptoms of dementia. Clin Geriatr Med 19(4): 799–824. Finkel SI, Costa e Silva J, Cohen G, Miller S, Sartorius N. 1996. Behavioral and psychological signs and symptoms of dementia: a consensus statement on current knowledge and implications for research and treatment. Int Psychogeriatr 8(Suppl 3): 497–500.
Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia Forbes D, Forbes S, Morgan DG, et al. 2008. Physical activity programs for persons with dementia. Cochrane Database Syst Rev 16(3), Art. No.: CD006489. DOI: 10.1002/14651858.CD006489.pub2 Goldberg DP. 1972. The Detection of Psychiatric Illness by Questionnaire. Oxford University Press: London. Goldberg DP, Gater R, Sartorius N, et al. 1997. The validity of two versions of the GHQ in the WHO study of mental illness in general health care. Psychol Med 27: 191–197. Heyn P, Abreu BC, Ottenbacher KJ. 2004. The effects of exercise training on elderly persons with cognitive impairment and dementia: a meta-analysis. Arch Phys Med Rehabil 85(10): 1694–1704. Iliffe S, Curry L, Kharicha K, et al. 2011. Developing a dementia research registry: a descriptive case study from North Thames DeNDRoN and the EVIDEM programme. BMC Med Res Methodol 11(1): 9. Kales HC, Valenstein M, Kim HM, et al. 2007. Mortality risk in patients with dementia treated with antipsychotics versus other psychiatric medications. Am J Psychiatr 164: 1568–1576. McCarney R, Warner J, Iliffe S, et al. 2007. The Hawthorne effect: a randomised, controlled trial. BMC Med Res Methodol 7: 30. Miller EA, Rosenheck RA, Schneider LS. 2010. Caregiver burden, health utilities, and institutional service costs among community-dwelling patients with Alzheimer disease. Alzheimer Dis Assoc Disord 24(4): 380–389. Nandy S, Parsons S, Cryer C, et al. 2004. Development and preliminary examination of the predictive validity of the falls risk assessment tool (FRAT) for use in primary care. J Public Health (Oxf) 26(2): 138–143.
Copyright # 2013 John Wiley & Sons, Ltd.
827 National Institute for Health and Clinical Excellence. 2006. Dementia: supporting people with dementia and their carers in health and social care. In Interventions for non-Cognitive Symptoms and Behaviour That Challenges in People with Dementia. National Institute for Health and Clinical Excellence: London. Rabins PV, Mace NL, Lucas MJ. 1982. The impact of dementia on the family. JAMA 248(3): 333–335. Schreiner AS, Morimoto T, Arai Y, Zarit S. 2006. Assessing family caregiver’s mental health using a statistically derived cut-off score for the Zarit Burden interview. Aging Ment Health 10(2): 107–111. Schur D, Whitlatch CJ. 2003. Circumstances leading to placement: a difficult caregiving decision. Lippincotts Case Manag 8(5): 187–195. Smith SC, Lamping DL, Banerjee S, et al. 2005. Measurement of health-related quality of life for people with dementia: development of a new instrument (DEMQOL) and an evaluation of current methodology. Health Technol Assess 9(10): 1–93, iii-iv. Studenski S, Duncan P, Chandler J, et al. 1994. Predicting falls: the role of mobility and nonphysical factors. J Am Geriatr Soc 42(3): 297–302. Thune-Boyle IC, Iliffe S, Cerga-Pashoja A, Lowery D, Warner J. 2012. The effect of exercise on behavioral and psychological symptoms of dementia: towards a research agenda. Int Psychogeriatr 24(7): 1046–1057. World Health Organisation. 1992. The ICD-10 classification of mental and behavioural disorders: diagnostic criteria for research. F00 - F09 Organic, Including Symptomatic, Mental disorders: Dementia. Geneva.
Int J Geriatr Psychiatry 2014; 29: 819–827
The effect of exercise on behavioural and psychological symptoms of dementia: the EVIDEM-E randomised controlled clinical trial† David Lowery1,2, Arlinda Cerga-Pashoja1,2, Steve Iliffe2, Ingela Thuné-Boyle2, Mark Griffin2, James Lee5, Alex Bailey1, Rahul Bhattacharya4 and James Warner1,3 1
Central and North West London NHS Foundation Trust (CNWL), London, UK Primary Care and Population Health, University College London (UCL), London, UK 3 Imperial College London, London, UK 4 East London NHS Foundation Trust (ELFT), London, UK 5 Independent exercise instructor, Brighton, UK Correspondence to: Dr D. Lowery, E-mail: [email protected] 2
†
Trial Registration: ISRCTN01423159
Objective: The
objective of this study is to evaluate the effectiveness of a simple dyadic (person with dementia and their main carer) exercise regimen as a therapy for the behavioural and psychological symptoms of dementia. Method: A two arm, pragmatic, randomised, controlled, single-blind, parallel-group trial of a dyadic exercise regimen (individually tailored walking regimen designed to become progressively intensive and last between 20–30 min, at least five times per week). Community-dwelling individuals with ICD-10 confirmed dementia with the following: clinically significant behavioural and psychological symptoms, a carer willing and able to co-participate in the exercise regimen, and no physical conditions or symptoms that would preclude exercise participation were invited by mental health or primary care services into the study. Results: One hundred and thirty-one dyads were recruited to this study. There was no significant difference in Behavioural and Psychological Symptoms as measured by the Neuropsychiatric Inventory at week 12 between the group receiving the dyadic exercise regimen and those that did not (adjusted difference in means (intervention minus control) = 1.53, p = 0.6, 95% CI [ 7.37, 4.32]). There was a significant between-group difference in caregiver’s burden as measured by the Zarit Caregiver Burden Inventory at week 12 (OR = 0.18, p = 0.01, CI [0.05, 0.69]) favouring the exercise group. Conclusions: This study found that regular simple exercise does not appear to improve the behavioural and psychological symptoms of dementia, but did seem to attenuate caregiver burden. Further study to improve exercise uptake are needed. Copyright # 2013 John Wiley & Sons, Ltd. Key words: trial; dementia; exercise; physical activity; behavioural and psychological symptoms History: Received 12 June 2013; Accepted 17 November 2013; Published online 11 December 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/gps.4062
Introduction Dementia is associated with a cluster of non-cognitive symptoms and behaviours that are an integral part of the syndrome (Alzheimer, 1907). Commonly described as behavioural and psychological symptoms of dementia (BPSD), they include disturbances of perception, Copyright # 2013 John Wiley & Sons, Ltd.
thought content, affect or behaviour (Finkel, 2003). Over 80% of people with dementia experience BPSD at some point during the course of their illness (Aalten et al., 2003), and they are commonly associated with reduction in the quality of life for the person as well as their carers (Colerick and George, 1986); increase of caregiver stress (Finkel et al., 1996), costs of care Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
820
(Rabins et al., 1982); and premature institutionalisation (Cohen-Mansfield, 1995). Importantly, behaviours that challenge care givers—particularly agitation, psychosis and sleep disturbance, often lead to institutionalisation and increased service utilisation (Donaldson et al., 1998; Dunkin and Anderson-Hanley, 1998; Schur and Whitlatch, 2003; Miller et al., 2010). Although there is some evidence supporting the treatment of BPSD with antipsychotic drugs, there have been increasing concerns about the safety of these drugs for people with dementia. Several studies have indicated that there is a long-term risk of mortality, accelerated cognitive decline and impaired quality of life in patients with dementia who take antipsychotic medication (Kales et al., 2007; Ballard et al., 2009). The UK’s clinical regulatory body—the National Institute for Health and Clinical Excellence—recommends the use of antipsychotic medications is restricted to cases in which severe risk of harm or distress is likely (NICE, 2006). There are few other safe pharmacological interventions with a proven evidence base. Non-pharmacological alternatives are being increasingly explored, and exercise has appeared to be a promising candidate (Brodaty and Arasaratnam, 2012). As well as the physical benefit, exercise has been demonstrated to have a positive effect on cognitive symptoms and mood in other populations (Berger and Motl, 2000) and could provide other benefits such as a sense of empowerment. There have been few robust clinical trials of exercise, and conflicting metaanalyses (Heyn et al., 2004; Forbes et al., 2008). Heyn et al., reported a positive effect of exercise on BPSD (ES = 0.54; 95% CI, 0.36–0.72) (Heyn et al., 2004); however, this does not provide a full picture of the effectiveness of exercise for BPSD (Thune-Boyle et al., 2012). There is considerable heterogeneity among trialled interventions, and exercise is often combined with other behavioural interventions. Some exercise regimens are complex and required a degree of physical fitness that would preclude many older adults with complex physical problems and moderate or significant dementia from performing them; finally, many trials are relatively underpowered. Method
Trial design
We conducted a pragmatic, randomised, controlled, single-blind, parallel-group trial of a dyadic exercise regimen (tailored walking) for community-dwelling individuals with dementia and their carers. The trial was designed following the UK Medical Research Council’s guidance on the development and testing of complex interventions (Craig et al., 2008). Individuals with a diagnosis of dementia or suspected dementia who had a carer willing and able to be a co-participant in the exercise regimen were eligible for recruitment. We registered and reported the trial design prior to recruitment (Trial Registration: ISRCTN01423159) (Cerga-Pashoja et al., 2010). The study opened for recruitment in January 2010 and final follow-up was in December 2012. After screening and consent, participants were randomly allocated (ratio 1 : 1) to one of two arms: the treatment arm, which received an individually tailored regimen of walking in addition to TAU; or the control arm, which received TAU only. A simple randomisation protocol was performed by an independent randomisation officer using a computer algorithm. Allocation was communicated to participants, carers, the participant’s General Practitioner and the therapist; other study personnel were kept blind to randomisation status until completion of the analysis. Blinding of group allocation from participants was impossible, but we adopted several strategies to minimise de-blinding of research staff, and these are discussed in the published trial protocol (CergaPashoja et al., 2010). Funding was provided by the UK’s National Institute for Health Research (RP-PG-0606-1005), sponsorship by CNWL NHS Foundation Trust, and ethical approval by Outer North East London Research Ethics Committee (ref 09/H0701/67). The study was overseen by an independent steering group including a data and safety monitoring subcommittee. Informed consent was sought from all participants. When an individual lacked capacity to consent, assent was sought and an appropriate consultee was asked for an opinion on the individual’s involvement and assent was periodically reassessed (Department of Constitutional Affairs, 2007).
Objectives Study population
The primary objective was to determine the effectiveness of a dyadic exercise regimen delivered through a programme of incremental walking for treating BPSD compared with treatment as usual (TAU). Copyright # 2013 John Wiley & Sons, Ltd.
This study was conducted in several inner city, urban and semi-rural locations in and around London. Participants were recruited either directly from the Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia
North Thames Dementias and Neurodegenerative Diseases Research Network’s dementia research register (Iliffe et al., 2011), by self-referral or via primary clinical services or specialist mental health services (e.g., memory assessment and community mental health). Participants with a clinical diagnosis of dementia (confirmed using ICD-10 Diagnostic Criteria for Research (1992)) with at least one significant BPSD symptom defined by the Neuropsychiatric Inventory (NPI) (NPI minimum score of frequency = 2, and severity = 2) (Cummings et al., 1994) were eligible for the trial. For the purpose of the trial, we excluded the NPI domains of delusions and hallucinations from screening and outcome as we did not anticipate exercise would affect these symptoms. A risk assessment protocol to check individuals’ suitability for the exercise programme included the falls risk assessment tool and timed unsupported standing tool (Studenski et al., 1994; Nandy et al., 2004) and primary care physician case note review.
821
Interventions
Physical exercise was delivered as an individually tailored regimen of walking designed to become progressively more intensive and last between 20–30 min. This was facilitated by one registered exercise professional qualified in instructing physical activity and exercise (National Vocational Qualification Level 3) and delivered to participants in the treatment arm of the trial in and around their own home. To help distinguish between a therapist effect operating through social contact and the effect of physical activity, the intervention was divided into two 6-week epochs. The exercise therapist progressively withdrew support at the end of the first epoch, with the expectation that the dyad would perform the exercise regimen regularly and independently at least five times per week (Table 1). Outcomes were measured at the end of both epochs. All participants were asked to record their daily activities throughout the 12 weeks of participation using a diary designed for this study. The intervention group
Table 1 Administration of outcome measures Day
Administrator
0
Research worker
0 1–2 3–4 5–8 6–8 13–17 17–28 40–46
Randomisation Exercise therapist Exercise therapist Exercise therapist Exercise therapist Exercise therapist Exercise therapist Independent researcher
40–46
Research worker
40–46 80–88
Exercise therapist Independent researcher
80–88
Research worker
80–88 90–99 182–196
Exercise therapist Research worker Research worker
Intervention/assessment ICD-10 DemQol MMSE GHQ NPI ZBI Demographics Medication CSRI Vital signs Allocation and diaries RPE timed walk, diarya RPE timed walka Telephone contact—advice and guidancea RPE timed walka Telephone contact—advice and guidancea Telephone contact as neededa NPI Diary reminder DemQol GHQ ZBI Medication Vital Signs RPE timed walka NPI Diary reminder DemQol GHQ ZBI Medication Vital Signs RPE timed walka Collection of diaries Telephone contact Check for change of status (residence and mortality)
ICD, International Classification of Diseases; MMSE, Mini Mental State Examination; NPI, Neuropsychiatric Inventory; CSRI, Client Service Receipt Inventory; RPE, Rating of Perceived Exertion; GHQ, General Health Questionnaire; ZBI, Zarit Caregiver Burden Inventory. a Treatment group only.
Copyright # 2013 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
822
diary contained an additional visual analogue scale, the rating of perceived exertion (RPE) (Borg, 1982). Perceived exertion is based on the physical sensations a person experiences during exercise, including increased heart rate, respiration or breathing rate, sweating, and muscle fatigue. Although this is a subjective measure, an exertion rating may provide a reasonable estimate of the actual heart rate during exercise (Borg, 1982). The therapist trained participant dyads in its use and validated their ratings. Participants were encouraged to extend (or reduce in some circumstance) the level of intensity to between 12 and 14 on the RPE scale. At this level, individuals should be able to respond verbally to questions, but may have to pause to take another breath more frequently than they would normally. Beyond 14, individuals would struggle to respond verbally and maintain pace. Study endpoints
Both the treatment and control arms were assessed for all outcomes at 6 and 12 weeks after randomisation. Further telephone contact took place at 26 weeks to assess additional adverse events, mortality, change in domiciliary status and adherence to exercise regimen (Table 1). The primary endpoint was a between-group difference of proportions of people with a reduction of three or more points on the composite NPI score (Cummings et al., 1994) at 12 weeks follow-up. This was chosen as a clinically meaningful change in BPSD symptoms. Secondary outcome measures were: mean differences in NPI score, the General Health Questionnaire (GHQ) (Goldberg, 1972), DemQOL-Proxy (DemQol) (Smith et al., 2005), and Zarit Caregiver Burden Inventory (ZBI) (Bedard et al., 2001). Participants’ level of exercise and compliance with the intervention were measured by carers using daily diaries and the RPE scale. Analytic methods
Assuming 80% of patients in the control group will have BPSD symptoms measured by the NPI at the 12 week follow-up and based on an anticipated between-group (control minus intervention) absolute risk difference of 30% in the proportion of people with BPSD, we calculated that a sample size of n = 116 participants would provide 90% power to detect this difference with a 5% (two-sided) significance level. To allow for anticipated attrition, the target sample was set at n = 146. Copyright # 2013 John Wiley & Sons, Ltd.
A statistical analysis plan was written and locked before data analysis began. Descriptive statistics are presented as numbers and percentages for categorical variables and means with standard deviations for continuous data. Where assumptions for normality and linearity were met, analysis of covariance adjusting for baseline scores was used for between-group comparisons of continuous data (intervention-control). Analysis was conducted on an intention-to-treat basis with available data. At endpoint, participants were dichotomised according to a clinically significant reduction (Baseline—Week 12) of BPSD (Composite NPI score change ≥ 3 points); and validated thresholds for ‘caseness’ relating to psychological wellbeing (GHQ) (0–24; optimal cut-off ≥23) (Goldberg et al., 1997) and for care giver burden (ZBI) (0–28; optimal cut-off ≥24) (Schreiner et al., 2006); change in domiciliary status and mortality. Binary logistic regression was used to analyse binary outcomes. All analysis was performed blind by the researcher and overseen by the study statistician using the statistics package SPSS (v20) (IBM, Armonk, New York, USA). Results Description of the sample
One hundred and thirty-one participant dyads were randomised to either receive the exercise regimen in addition to TAU (treatment), or to receive TAU only (control). Eighty nine percent of dyads (n = 116) completed the trial, with seven and eight dyads lost from the control group and the intervention groups, respectively (Figure 1). Descriptive statistics for the baseline demographic and the outcome data are presented for each group (Table 2). The groups were relatively similar in terms of age, type of dementia and other characteristics (Table 2). Behavioural and psychological symptoms as measured by the NPI were similar in prevalence to those reported in the original validation paper (Cummings et al., 1994). At baseline, 36 carers were at or above the validated threshold for ‘caseness’ for psychological morbidity on the GHQ and 25 for care giver burden on the ZBI. Analysis of primary outcome
Although the intervention group were 41% more likely to reach a clinically significant reduction of three or more points of NPI composite score by week 12, there was no significant difference between the groups (OR = 1.41, p = 0.36, 95% CI [0.67, 3.01]). Using Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia
823 Screened n = 474*
Excluded n = 82 No BPSD (n=39) Falls Risk (n=29) No carer (n=8) GP Excluded (n=4) No dementia (n=2) Declined n=252
Randomised n = 131
Allocated to Treatment as Usual N = 64
Allocated to Treatment as Usual N = 67
Lost to follow up Withdrew
Included in Primary Analysis N = 57
n=3 n=4
Lost to follow up Withdrew Died
n=1 n=5 n=2
Included in Primary Analysis N = 59
* Confirmation of ability to form a dyad was part of screening process. Therefore, numbers screened are presented as individuals with dementia only.
Figure 1 Participant Dyad Flow. Table 2 Socio-demographic statistics (baseline) Control (n = 64) n Person with Dementia Age Gender: female Ethnicity: white Marital status: married FRAT Score: ≥2 Living: private residence Years Education MMSE Alzheimers Years since diagnosis: ≤2 years Carer Age Gender: female FRAT score: =0 Relationship: ‘partner or spouse’ Carer distress (NPI)
Mean ± SD (range) or frequency (percentage)
Intervention (n = 67) n
Mean ± SD (range) or frequency (percentage)
64 64 64 63 64 64 60 64 59 61
78 ± 7.4 (58–99) 39 (60.9%) 50 (78.1%) 45 (71.4%) 36 (56.3%) 57 (89.1%) 11.92 ± 5.9 (0–36) 14.9 ± 8.7 (0–29) 38 (64.4%) 38 (64.4%)
67 67 67 67 67 67 63 67 65 59
79 ± 6.8 (64–97) 35 (52.2%) 56 (83.6%) 46 (68.6%) 27 (40.3%) 59 (88.1%) 12.1 ± 4.1 (6–23) 16.3 ± 7.4 (0–30) 44 (67.7%) 35 (54.7%)
59 64 64 64 64
60.9 ± 17 (22–88) 39 (60.9%) 46 (71.9%) 35 (54.7%) 11.9 ± 8.1 (0–39)
53 67 67 67 67
65.4 ± 14.9 (27–89) 50 (74.6%) 40 (59.7%) 42 (62.7%) 11.8 ± 8.9 (0–38)
FRAT, Falls Risk Assessment Tool; MMSE, Mini Mental State Examination; NPI, Neuropsychiatric Inventory.
Copyright # 2013 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
824
analysis of covariance adjusting for baseline covariates, no significant difference was found between the groups of mean NPI at week 12 (intervention-control) (β = 1.53, p = 0.6, 95% CI [ 7.37, 4.32])(Table 3).
week) was achieved by 30.77% of the treatment group, and prescribed intensity (RPE, 12–14) was achieved in 53.25% of the walks. Safety and adverse events
Analysis of secondary outcomes
Caregiver’s burden as measured by the ZBI doubled from 16% at baseline to 32% by week 12 for the control group participants but decreased from 23 to 17% for those in the intervention group; (OR = 0.18, p = 0.01, CI [0.05, 0.69]). There were no statistically significant differences between the groups at week 12 in carers’ mental health (measured using the GHQ), carers’ distress as measured by the NPI and quality of life of participants with dementia (measured using the Demqol-proxy), as shown in Table 3. There were no significant differences between groups at week 26 in change of domiciliary status (OR = 0.39, p = 0.42, CI [0.38, 3.85]), or mortality (OR = 1.76, p = 0.55, CI [0.28, 11.08]). Treatment
There were no significant differences between the groups of self-reported walking time (minutes) at baseline (β = 5.2, p = 0.09, 95% CI [ 11.33, 0.93]), week 6 (β = 0.55, p = 0.82, 95% CI [ 5.32, 4.21]) and week 12 (β = 1.18, p = 0.697, 95% CI [ 4.81, 7.17]). Prescribed frequency of walks (five times per
There were six reported falls (two in the control arm and four from the intervention), but none required hospitalisation. There were also five reported deaths, three in the intervention arm and two in the control. Two deaths were reported during the intervention period and the other three at long-term follow-up. All adverse events were reviewed by a senior clinician and were deemed not to be related to trial participation. Discussion Previous studies have reported that exercise is beneficial in reducing some BPSD, especially depressed mood and agitation and may also improve sleep and reduce ‘wandering’. However, research evaluating the efficacy of exercise as a therapeutic tool for other important symptoms such as anxiety, apathy and repetitive behaviours is limited (Thune-Boyle et al., 2012). This study shows that physical exercise tailored to participant-carer dyads, designed to become progressively intensive and lasting between 20–30 min per day made no significant difference to NPI score after 12 weeks. This contradicts the metaanalysis reported by Heyn et al., (Heyn et al., 2004).
Table 3 Between group comparison at weeks 6 and 12 Outcome NPI GHQ ZBI
NPI Demqol NPI carer distress
Week
Control frequency (percentage)
Intervention frequency (percentage)
Baseline 6 12 Baseline 6 12 Baseline 6 12
64/64 (100%) 32/56 (57.1%) 33/57 (57.9%) 46/63 (73%) 24/57 (42.1%) 24/56 (43%) 52/62 (83.9%) 14/55 (25.5%) 18/56 (32%)
67/67 (100%) 39/62 (62.9%) 39/59 (66.1%) 48/67(71.6%) 16/62 (25.8%) 17/55 (31%) 50/65 (76.9%) 14/59 (23.7%) 10/57 (17.5%)
6 12 6 12 6 12
Mean ± SD 26.6 ± 17.5 25.6 ± 16.6 101.1 ± 14.9 101 ± 13.5 11.07 ± 7.2 9.98 ± 5.9
Mean ± SD 25.7 ± 20.5 23.9 ± 20.6 103.6 ± 11.9 104 ± 10 11.5 ± 8.5 10.9 ± 9.3
OR†
95% CI
1.27a 1.41a
0.61 : 2.66 0.67 : 3.01
0.52 0.36
0.42 b 0.59 b
0.18 : 1.00 0.24 : 1.43
0.05 0.19
0.48 b 0.18 b
0.14 : 1.67 0.05 : 0.69
0.25 0.01
β* 0.81 1.53 1.27 2.62 0.06 1.14
6.08 : 4.45 7.37 : 4.32 2.33 : 4.86 0.78 : 6.02 2.25 : 2.14 1.31 : 3.58
0.76 0.60 0.49 0.09 0.96 0.76
p
†
OR of aImprovement (reduction) of composite NPI score of 3 or more points between baseline and weeks 6 and 12 bScoring at or above the validated threshold for ‘caseness’ at weeks 6 and 12. *Analysis of covariance of the adjusted difference in means (intervention-control) at weeks 6 and 12. The denominator of week 6 and week 12 data varies as result of participant availability.
Copyright # 2013 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia
Although overall BPSD were lower at week 12 in the group that received the exercise intervention in comparison with the control group, this difference was not statistically significant. This may be related to the poor adherence to the full exercise regimen, and/or the duration of the programme; longer follow-up periods and methods for increasing adherence might result in a significant effect, and this warrants further investigation. A large proportion of the TAU group experienced an improvement in BPSD over the course of the study, which may be a Hawthorne effect attributable to the engagement of research staff with participants (McCarney et al., 2007), and/or the natural fluctuation of BPSD. An alternative explanation is confounding due to changes of exercise behaviour in the control group. More objective measurement of control participants’ exercise behaviour may have allowed us to test for this, but would have brought its own methodological challenges (see succeeding text). Caregiver’s burden as measured by the ZBI decreased significantly in the intervention arm but increased in the control arm. This finding is potentially important because caregiver burden mediates decisions to relocate the person with dementia to a care home (Dunkin and Anderson-Hanley, 1998) and is associated with greater use of services (Miller et al., 2010); consequently, it may provide a useful proxy measure of both a longer-term outcome and future service use. It seems somewhat counter-intuitive that declining caregiver burden was attenuated in the treatment group given the absence of effect for BPSD. Although, this may have been as a consequence of the beneficial effects of exercise direct to the carer, rather than because of any change in the person with dementia. Feelings of mastery (or lack of in the control group), because of an anticipated positive effect of the intervention is one possible explanation; this could have profound ethical implications for the design of future research and should be explored along with other potential mechanisms and long-term effects. This will also be important for optimising this intervention and the development of other interventions. Strengths and limitations of this study include strengths, the relatively large sample and low attrition rates providing sufficient power to detect an effect (positive or negative) of clinically significant magnitude; the recruitment of a clinically relevant sample in terms of diagnosis, severity of disease and mobility allowing greater generalisability; use of well validated and blinded outcome measures; the use of a pragmatic exercise programme, which does not require specialist Copyright # 2013 John Wiley & Sons, Ltd.
825
training or equipment; and efforts to control for the impact of therapist contact. This study has shown that an acceptable exercise intervention for people with dementia can be developed and tested. The limitations of the study include the single-blind design and the weakness of self-report in measuring exercise. The single-blind design had the inherent and unavoidable risk of affecting control participants’ exercise behaviour. Additionally, there was also a risk that participants would divulge information about the group allocation to the researcher. We attempted to reduce this risk through the inclusion of a second, independent, researcher who collected primary outcome data (NPI) at 6 and 12 weeks. Assessing physical activity with minimum intrusion to study participants is challenging and we relied on participants’ self reports through daily diaries. Self-report is prone to information bias, especially when participants are not blinded to their group allocation (Bauhoff, 2011). We gave considerable thought to the issue of measuring physical activity, in particular, GPS receivers, actigraphs and pedometers. We decided not to use such instruments to avoid excessive intrusion and prevent use of such instruments affecting exercise behaviour in their own right. Phasing out the effect of psychosocial contact (from the exercise therapist) from the exercise intervention appeared to allow a relatively successful control for contact between participants and the exercise therapist—results were consistent across both time periods. In addition, this was a complex intervention in which the social contact between the person with dementia and their carer formed an important component. Dyadic walking may have encouraged psychosocial support from the carer, and this may partly explain the changes we observed for caregiver burden. Conclusions A regular simple exercise such as walking does not appear to be effective for reducing BPSD but does appear to help reduce carer burden. However, we cannot be sure, particularly as increasing physical activity in people with dementia and their carers is difficult. The intervention group were 41% more likely to experience a clinically important reduction of NPI score, but this was not a significant difference. Given that less than half the intervention group achieved the prescribed exercise frequency and intensity targets, further studies attempting to increase uptake are warranted, as are longer studies to measure the effects of reduced carer burden. Int J Geriatr Psychiatry 2014; 29: 819–827
D. Lowery et al.
826
Conflict of interest None to declare. The study sponsor had no role in the study design, in the collection, analysis and interpretation of data, in the writing of the report and in the decision to submit this paper for publication. Key points
• •
This study does not provide any evidence that exercise is an effective therapeutic intervention for the behavioural and psychological symptoms of dementia. Dyadic exercise supported by up to 3 hours therapist input significantly attenuates worsening caregiver burden.
Ethics This study adhered to and obtained the requisite approvals necessary to conduct research with patients in the NHS. Ethical approval was given by Outer North East London Research Ethics Committee (ref 09/H0701/67). Permissions (CSP20305) were obtained to recruit participants from the following Sites: Central & North West London NHS Foundation Trust, Surrey and Borders Partnership NHS Foundation Trust, West London Mental Health Trust, East London NHS Foundation Trust; and Participant Identification Centres: Camden & Islington NHS Foundation Trust and Barnet, Enfield and Haringey Mental Health NHS Trust. Acknowledgements We gratefully acknowledge the contribution of all study participants and their carers. The study was supported by a proactive and valuable steering group (Chaired by Dr Craig Ritchie); a wide number of recruitment hubs (North Thames DeNDRoN, Surrey and Borders Partnership NHS Foundation Trust—Dr Nilforooshan, and West London Mental Health NHS Trust—Dr Regan); and several volunteers who kindly provided their time in exchange for experience of working on a research trial (Mr Aeron-Smith, Miss Dimitrova, Miss Theobald, Miss Pulford, Ms Mystris, Ms Sula, Dr Jordanova, Dr Rajenthran and Dr Holzer). The study was supported by Central and North West London NHS Foundation Trust. This report presents independent research funded by the National Institute for Health Research under Copyright # 2013 John Wiley & Sons, Ltd.
its Programme Grants for Applied Research scheme (RP-PG-0606-1005). The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. Author contributions J. W. conceived of the idea of the study and together with S. I. obtained funding for it. J. W., along with D. L. and A. P., had primary oversight of the design and day to day management of the study. A. P. was responsible for data collection, A. P., D. L. and M. G. led the analysis of data. D. L. wrote the first draft of this paper. All other authors, along with the steering group, made essential contributions to the design of the study in line with their individual expertise. All authors contributed to the drafting and/or critical revision of this manuscript for important intellectual content, approval of the final version, and take public responsibility for its content. References Aalten P, DeVught ME, Lusberg R, et al. 2003. Behavioural symptoms in dementia: a factor analysis of the neuropsychiatric inventory. Dement Geriatr Cogn Disord 15: 99–105. Alzheimer A. 1907. Über eine eigenartige Erkrankung der Himrinde. Allgemeine Zeitschrift fur Psychiatrie und Psychisch-gerichtliche Medizin. Ballard C, Hanney ML, Theodoulou M, et al. 2009. The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. Lancet Neurol 8(2): 151–157. Bauhoff S. 2011. Systematic self-report bias in health data: impact on estimating cross-sectional and treatment effects. Health Serv Outcomes Res Method 11(1–2): 44–53. Bedard M, Molloy DW, Squire L, et al. 2001. The Zarit Burden Interview: a new short version and screening version. Gerontologist 41(5): 652–657. Berger BG, Motl RW. 2000. Exercise and mood: a selective review and synthesis of research employing the profile of mood states. J Appl Sport Psychol 12(1): 69–92. Borg G. 1982. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 14(5): 377–381. Brodaty H, Arasaratnam C. 2012. Meta-analysis of nonpharmacological interventions for neuropsychiatric symptoms of dementia. Am J Psychiatry 169(9): 946–953. Cerga-Pashoja A, Lowery D, Bhattacharya R, et al. 2010. Evaluation of exercise on individuals with dementia and their carers: a randomised controlled trial. Trials 11: 53. Cohen-Mansfield J. 1995. Assessment of disruptive behavior/agitation in the elderly: function, methods, and difficulties. J Geriatr Psychiatry Neurol 8(1): 52–60. Colerick EJ, George LK. 1986. Predictors of institutionalization among caregivers of patients with Alzheimer’s disease. J Am Geriatric Soc 34: 492–498. Craig P, Dieppe P, et al. 2008. Developing and Evaluating Complex Interventions: new Guidance. Medical Research Council: London. Cummings JL, Mega M, Gray K, et al. 1994. The neuropsychiatric inventory: comprehensive assessment of psychopathology in dementia. Neurology 44(12): 2308–2314. Department of Constitutional Affairs. 2007. Mental Capacity Act 2005: Code of Practice. TSO: Birmingham. Donaldson C, Tarrier N, Burns A. 1998. Determinants of carer stress in Alzheimer’s disease. Int J Geriatr Psychiatry 13(4): 248–256. Dunkin JJ, Anderson-Hanley C. 1998. Dementia caregiver burden: a review of the literature and guidelines for assessment and intervention. Neurology 51(1 Suppl 1): S53–60; discussion S65-57. Finkel SI. 2003. Behavioral and psychologic symptoms of dementia. Clin Geriatr Med 19(4): 799–824. Finkel SI, Costa e Silva J, Cohen G, Miller S, Sartorius N. 1996. Behavioral and psychological signs and symptoms of dementia: a consensus statement on current knowledge and implications for research and treatment. Int Psychogeriatr 8(Suppl 3): 497–500.
Int J Geriatr Psychiatry 2014; 29: 819–827
Exercise for behavioural symptoms of dementia Forbes D, Forbes S, Morgan DG, et al. 2008. Physical activity programs for persons with dementia. Cochrane Database Syst Rev 16(3), Art. No.: CD006489. DOI: 10.1002/14651858.CD006489.pub2 Goldberg DP. 1972. The Detection of Psychiatric Illness by Questionnaire. Oxford University Press: London. Goldberg DP, Gater R, Sartorius N, et al. 1997. The validity of two versions of the GHQ in the WHO study of mental illness in general health care. Psychol Med 27: 191–197. Heyn P, Abreu BC, Ottenbacher KJ. 2004. The effects of exercise training on elderly persons with cognitive impairment and dementia: a meta-analysis. Arch Phys Med Rehabil 85(10): 1694–1704. Iliffe S, Curry L, Kharicha K, et al. 2011. Developing a dementia research registry: a descriptive case study from North Thames DeNDRoN and the EVIDEM programme. BMC Med Res Methodol 11(1): 9. Kales HC, Valenstein M, Kim HM, et al. 2007. Mortality risk in patients with dementia treated with antipsychotics versus other psychiatric medications. Am J Psychiatr 164: 1568–1576. McCarney R, Warner J, Iliffe S, et al. 2007. The Hawthorne effect: a randomised, controlled trial. BMC Med Res Methodol 7: 30. Miller EA, Rosenheck RA, Schneider LS. 2010. Caregiver burden, health utilities, and institutional service costs among community-dwelling patients with Alzheimer disease. Alzheimer Dis Assoc Disord 24(4): 380–389. Nandy S, Parsons S, Cryer C, et al. 2004. Development and preliminary examination of the predictive validity of the falls risk assessment tool (FRAT) for use in primary care. J Public Health (Oxf) 26(2): 138–143.
Copyright # 2013 John Wiley & Sons, Ltd.
827 National Institute for Health and Clinical Excellence. 2006. Dementia: supporting people with dementia and their carers in health and social care. In Interventions for non-Cognitive Symptoms and Behaviour That Challenges in People with Dementia. National Institute for Health and Clinical Excellence: London. Rabins PV, Mace NL, Lucas MJ. 1982. The impact of dementia on the family. JAMA 248(3): 333–335. Schreiner AS, Morimoto T, Arai Y, Zarit S. 2006. Assessing family caregiver’s mental health using a statistically derived cut-off score for the Zarit Burden interview. Aging Ment Health 10(2): 107–111. Schur D, Whitlatch CJ. 2003. Circumstances leading to placement: a difficult caregiving decision. Lippincotts Case Manag 8(5): 187–195. Smith SC, Lamping DL, Banerjee S, et al. 2005. Measurement of health-related quality of life for people with dementia: development of a new instrument (DEMQOL) and an evaluation of current methodology. Health Technol Assess 9(10): 1–93, iii-iv. Studenski S, Duncan P, Chandler J, et al. 1994. Predicting falls: the role of mobility and nonphysical factors. J Am Geriatr Soc 42(3): 297–302. Thune-Boyle IC, Iliffe S, Cerga-Pashoja A, Lowery D, Warner J. 2012. The effect of exercise on behavioral and psychological symptoms of dementia: towards a research agenda. Int Psychogeriatr 24(7): 1046–1057. World Health Organisation. 1992. The ICD-10 classification of mental and behavioural disorders: diagnostic criteria for research. F00 - F09 Organic, Including Symptomatic, Mental disorders: Dementia. Geneva.
Int J Geriatr Psychiatry 2014; 29: 819–827
