Changing Risk Behaviours and Promoting Cognitive Health in Older

Intervention Tables

Changing risk behaviours and promoting cognitive health in older adults An evidence-based resource for local authorities and commissioners Prepared by the Cambridge Institute of Public Health, University of Cambridge 2 Evidence-based resource for changing risk behaviours in older adults: Intervention Tables

About Public Health England

Public Health England exists to protect and improve the nation’s health and wellbeing, and reduce health inequalities. We do this through world-class science, knowledge and intelligence, advocacy, partnerships and the delivery of specialist public health services. We are an executive agency of the Department of Health, and are a distinct delivery organisation with operational autonomy to advise and support government, local authorities and the NHS in a professionally independent manner.

Public Health England Wellington House 133-155 Waterloo Road London SE1 8UG Tel: 020 7654 8000 www.gov.uk/phe Twitter: @PHE_uk Facebook: www.facebook.com/PublicHealthEngland

This resource was prepared by the Cambridge Institute of Public Health, University of Cambridge. The systematic reviews underpinning the resource are co-funded by the Ageing Well Programme of the National Institute for Health Research (NIHR) School for Public Health Research and the Dementia Frailty End of Life theme of the NIHR Collaboration for Leadership in Applied Health Research and Care (CLAHRC) East of England.

Authors: Louise Lafortune, Sarah Kelly, Olawale Olanrewaju, Andy Cowan, Carol Brayne

For queries relating to this document, please contact: Louise Lafortune, [email protected]

© Crown copyright 2016 You may re-use this information (excluding logos) free of charge in any format or medium, under the terms of the Open Government Licence v3.0. To view this licence, visit OGL or email psi@ nationalarchives.gsi.gov.uk. Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned.

Published November 2016 PHE publications gateway number: 2016427

Evidence-based resource for changing risk behaviours in older adults: Intervention Tables 3

Contents

Table Index 4 Method 6 Intervention tables 10 • Multi-component 12 • Alcohol 64 • Smoking 82 • Diet 96 • Physical activity 132 • Cognitive stimulation 157 • Social 171 • Leisure activity 191 4 Evidence-based resource for changing risk behaviours in older adults: Intervention Tables

Table index

Multicomponent 1. Interventions with cognitive outcomes: complex multicomponent interventions 10 2. Interventions with cognitive outcomes: combined PA and cognitive stimulation 23 3. Interventions with cognitive outcomes: combined PA and diet 38 4. Interventions with cognitive outcomes: combined cognitive and social activation 41 5. Interventions to improve uptake/maintenance of healthy behaviours: combined PA/diet 46 6. Quality assessment for interventions reporting on uptake/maintenance of behaviours 58 7. Quality assessment for interventions reporting on cognitive outcomes 61 8. Quality assessment for studies reporting barriers and facilitators 63

Alcohol 9. Intervention that promote healthy drinking behaviours (RCTs) 64 10. Intervention that targeted a range of health behaviours with seperate alcohol outcomes 72 11. Included qualitative studies about barriers and facilitators 75 12. Older adults’ drinking habits in the context of ageing 79 13. Quality assessment for interventions reporting on healthy drinking behaviours 80 14. Quality assessment for interventions reporting barriers and facilitators 81

Smoking 15. Interventions with smoking cessation and reduction outcomes (RCTs) 82 16. Interventions with smoking cessation and reduction outcomes (non-RCTs) 86 17. Interventions that targeted a range of behaviours with separate smoking outcomes 88 18. Interventions targeted at training health professionals in smoking cessation services 90 19. Overview of qualitative studies reporting barriers and facilitators 91 20. Barriers to smoking cessation in older adults 93 21. Facilitators to smoking cessation in older adults 95 22. Quality assessment for interventions reporting on smoking cessation and reduction outcomes 96 23. Quality assessment for studies reporting barriers and facilitators 97

Diet 24. Interventions to increase uptake/maintenance of healthy diet behaviours (RCTs) 98 25. Interventions to increase uptake/maintenance of healthy diet behaviours (non-RCTs) 109 26. Interventions to increase uptake/maintenance of healthy diet behaviours (secondary analyses) 112 27. Intervention to increase uptake/maintenance of healthy diet behaviours with cognitive outcomes 113 28. Included qualitative studies about barriers and facilitators 124 29. Quality assessment for interventions reporting on uptake/maintenance of healthy diet 127 Evidence-based resource for changing risk behaviours in older adults: Intervention Tables 5

30. Quality assessment for interventions reporting on diet with cognitive outcomes 129 31. Quality assessment for interventions reporting barriers and facilitators 131

Physical activity 32. Interventions to improve uptake/maintenance of physical activity 133 33. Interventions to improve uptake/maintenance of physical activity with cognitive outcomes 137 34. Included qualitative studies about barriers and facilitators 151 35. Quality assessment for interventions reporting on uptake/maintenance of physical activity 155 36. Quality assessment for interventions reporting on physical activity with cognitive outcomes 156

Cognitive stimulation 37. Systematic reviews of cognitive training interventions with cognitive outcomes 161 38. Systematic reviews of cognitive training compared to physical activity 162 39. Systematic reviews of computerised cognitive training interventions 164 40. Other systematic reviews 166 41. Quality assessment of systematic reviews of cognitive stimulation interventions 169

Social 42. Systematic reviews or exposures for social participation interventions 171 43. Systematic reviews relevant to barriers/facilitators to social participation 173 44. Interventions to improve uptake/maintenance of socialparticipation with cognitive outcomes 174 45. Included studies about barriers and facilitators 186 46. Quality assessment of systematic reviews of social participation interventions 187 47. Quality assessment of intervention studies for social participation 189 48. Quality assessment of qualitative studies for social participation 190

Leisure 49. Interventions to improve uptake/maintenance of leisure activities 191 50. Interventions to improve uptake/maintenance of leisure with cognitive outcomes 193 51. Quality assessment of leisure activity intervention studies 199 52. Quality assessment of systematic reviews of leisure activity interventions 200 6 Evidence-based resource for changing risk behaviours in older adults: Intervention Tables

Method

The protocols for the three reviews underpinning this resource were registered on PROSPERO (Kelly 2015a; Lafortune 2015a; Lafortune 2015b) and the approach follows the PRISMA guideline (http://www.prisma-statement.org).

The overarching aim of the systematic evidence reviews is to identify which interventions to promote healthy behaviours and cognitive health in older adults are the most effective and cost- effective. The specific questions addressed in the reviews are:

1. What individual-level interventions targeting unhealthy behaviours in people in older age (55+ years) are effective for the primary prevention or delay of cognitive decline or dementia? 2. What individual-level interventions in people in older age (55+ years) are effective for increasing the uptake and maintenance of healthy behaviours? 3. What issues (barriers and facilitators) prevent or limit, or help and motivate the uptake and maintenance of healthy behaviours in people in older age (55+ years)?

Search strategies A structured search strategy was developed using a wide range of search terms covering the following concepts and domains: ageing and older adults; health behaviours and risk reduction relating to diet, physical activity, inactivity, alcohol, smoking; risk reduction relating to loneliness and isolation (i.e. leisure, social activities, participation), sun exposure, hearing and vision.

Databases searched include: MEDLINE, EMBASE, PsycINFO, CINAHL, Social Science Index, Cochrane Central Register of Controlled Trials (CENTRAL), The Cochrane Collaboration and Database of Systematic Reviews, Database of Abstracts of Reviews of Effectiveness (DARE), HTA and York CRD databases. The search strategies are easily accessible in the published PROSPERO protocols (Lafortune 2015a; Lafortune 2015b; Kelly 2015a).

Searching was conducted in two stages: 1) searching for systematic reviews in older age using a systematic review filter; 2) searching for published primary studies in older age. By including systematic reviews we captured the bulk of the scientific literature published before our cut-off date and avoided duplication of efforts. With the addiotnal primary studies, we captured the literature where there are no systematic reviews and/or where available systematic reviews are out of date.

Systematic reviews and primary studies published between January 2000 and December 2014 published in English and from countries of the Organisation for Economic Co-operation and Development (OECD) are included. However, studies were not excluded at the title/abstract Evidence-based resource for changing risk behaviours in older adults: Intervention Tables 7

screening stage on the basis of language so that the number of studies excluded on the basis of language can be measured and reported. Only studies and systematic reviews that have aimed to include people in older age (55 years and over) living in the community are included.

Types of studies For questions (1) and (2), we included systematic reviews and primary studies of any type of intervention studies. Study types include: randomised controlled trials; controlled clinical trials; controlled before and after studies; interrupted time series.

For question (3), systematic reviews and primary studies that reported qualitative data specific to barriers and facilitators to uptake or maintenance of modifiable healthy behaviours were included (i.e. physical activity, diet, smoking, alcohol, healthy life style, cognitive activities, leisure, participation, risk reduction related to loneliness and social isolation, sun exposure, hearing and vision health).

Participants/population Reviews and intervention studies that predominantly included populations with a mean age of 55+, living in the community and in a healthy condition, with pre-conditions such as high cholesterol, high blood pressure, overweight or obesity, or people on medication were included. Reviews and intervention studies from disadvantaged populations, minority groups and vulnerable communities werealso included. These covered low socio-economic status, ethnic minority groups, LGBT groups, travellers and other groups with protected characteristics under the equality and diversity legislation. Reviews and interventions specifically focused on populations with previous ill health such as stroke, coronary heart disease or asthma were excluded. Participants with previous health conditions however will have been included in some studies where there were no contra- indications.

Interventions/exposure Interventions targeting the following behaviours are included: • increase/maintain levels of physical activity or decrease sedentary lifestyles or maintain balance, strength and weight-bearing functions • improve/maintain good diet and nutrition (including components of diet e.g. fat intake, fruit and vegetable intake) • reduce/prevent/stop tobacco consumption • decrease/ prevent excessive alcohol consumption • maintain/increase cognitive, leisure and social activities, and participation • maintain hearing and vision • prevent excessive sun exposure or increase sun exposure in those with inadequate exposure 8 Evidence-based resource for changing risk behaviours in older adults: Intervention Tables

• promote/improve dental health • improve/modify multiple behavioural risk factors • remove barriers/facilitate uptake and maintenance of any unhealthy/healthy behaviours with demonstration of impact.

Interventions delivered in the following settings and using the following mode of delivery are included: • community settings (including, but not limited to, home, workplace, community and day centres, sheltered housing, primary care) • interventions at individual, family, community level • interventions in the private, public, voluntary or commercial sectors • interventions delivered by healthcare professionals, lay people, home carers, researchers, media, internet

Interventions in the following areas are excluded: • use of prescription drugs/medication (except for medication available ‘over the counter’ such as nicotine patches or gum for smoking cessation) • use of dietary supplements • management of existing disability, dementia, frailty and common non-communicable chronic disease • management of obesity, including medical and surgical interventions for obesity; • national policies, laws and taxation • screening • vaccination

No lower time limit was imposed for duration of intervention and follow-up.

Comparators/control Studies with any comparator or no comparator.

Data extraction, selection and coding Titles and abstracts were screened independently by two reviewers. Differences between reviewers’ results were resolved by discussion and, when necessary, in consultation with a third reviewer. If there was still doubt about the relevance of a study after discussion, it was retained. Full paper copies were obtained for all reviews identified by the title/abstract screening. Full paper screening was conducted independently by two reviewers. We extracted data on study design; population; intervention details, setting and delivery; comparators; type of outcome measures reported; outcome measures (measures of uptake and maintenance of healthy behaviour); design/ delivery of interventions and quantitative or qualitative data relating to implementation issues, Evidence-based resource for changing risk behaviours in older adults: Intervention Tables 9

barriers or facilitators and results.

Quality assessment The methodological quality of systematic reviews in Reviews (1) and (2) was assessed using the AMSTAR tool (www.Amstar.ca). A minimum of 10% of the reviews was fully and double quality assessed. Any discrepancy between reviewers was resolved by discussion. Reviews that adequately reported at least eight of the possible eleven AMSTAR criteria were rated as high quality; between five and seven criteria as moderate quality; reviews adequately reporting fewer than five criteria were considered to be of poor quality. Tables summarising the quality of each systematic review and primary study included in this resource are presented for each risk factors in the Intervention Tables included.

Data synthesis Findings are narratively synthesised and presented. Findings were initially tabulated to map the evidence in terms of included studies, country, age, population, interventions, comparators, outcome measures and effectiveness to map the level of evidence, quality and gaps.

Consultation Two two-and-a-half hour consultation events were held in Newcastle (n=12) and Cambridge (n=8) with representatives from local authorities, clinical commissioners, social care commissioners, public health consultants and clinicians to discuss the format of this resource, the level of detail required to inform decisions, the way messages are framed, and any other issues that felt relevant and important. Following a short presentation of the background for the project and the aims of the resource, facilitated group discussions took place using a draft version of the resource, detail findings for selected risk behaviours (i.e. smoking, alcohol, diet), examples of summary tables and key messages. Notes were compiled for each event and the research team implemented these changes. The revised version of the document was then reviewed by one of the public health consultants who attended the Cambridge workshop (Dr Angelique Mavrodaris, Consultant in Public Health, Public Health Directorate, Cambridgeshire County Council) as well as by members of the PHE team.

10 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 1: Multicomponent – Interventions with cognitive outcomes: complex multicomponent interventions

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Anstey 2015 RCT Australia 50-60 Cognitively healthy adults Multidomain web-based Follow-up: At end of 12- Overall dementia risk factor Body, Brain (mean age living independently, with intervention to address week intervention and at outcomes: In change scores Life (BBL) 55, SD access to a computer and multiple risk factors for AD 26 weeks from baseline at 12 weeks and 2.96) internet connection at using behaviour change 26 weeks post intervention, there home principles Lost to follow-up: BBL: was a significant reduction in risk 5.2%; BBL + FF: 1.7%; for both intervention groups but N=176 randomised Intervention 1: Body, Control: 5.0% not for the control group: BBL: 12 brain, life (BBL), N=58: weeks, p=0.019; 26 weeks post Gender: 52.8% female One module a week Outcome measurement: intervention: <0.001; BBL+ FF: delivered for 1 hour (7 Primary outcome was 12 weeks, p=0.007; 26 weeks Ethnicity: Not reported modules in total). The the ANU-ADRI validated post intervention, p=0.003 first 2 modules covered instrument measuring SES: Mean years general dementia risk and individual risk profile Individual protective factors: education 18.1 (3.56). protective factors. Modules for dementia. Risk There was a significant increase Author’s comment 3 to 7 covered PA, diet, factors included age, in cognitive engagement at 12 ‘relatively highly educated’ social engagement, sex, low education, weeks (odds ratio 2.64; 95% cognitive engagement, diabetes, history of CI 1.21–5.77; p=0.015) and 26 management of chronic brain injury, smoking, weeks postintervention (OR 4.49; conditions low pesticide exposure, 95% CI, 1.97–10.25; p =0.001). low social engagement, Also, a significant increase in Intervention 2: BBL plus and protective factors fish consumption at 26 weeks an additional face-to-face included were high PA, (OR 5.19; 95% CI, 1.51–17.91; component, N=58: Same high cognitive activity, p =0.009). ‘Consistent between intervention as BBL but fish consumption (3 or intervention groups’ (?) for both additionally received 5 more times/wk) and light outcomes face-to-face sessions in to moderate alcohol small groups to discuss consumption risk factors for dementia, goal setting, and barriers to behavior change

Active control, N=60. Received weekly emails with links to health-related website, videos and news items Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Carlson RCT (pilot) US Mean: In- Older adults Experience Corps® Follow-up: 4 to 8 months Cognitive outcomes: (age and 2008 tervention: places older volunteers education adjusted) Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 11 Experience (cluster 70.1 (6.42) N=149 randomised in public elementary Lost to follow-up: Corps randomised Control : schools in roles designed Intervention: 11.4%; Trail-making test by school) 68.4 (5.15) Baseline MMSE: to meet schools’ needs control: 17.2% Intervention: 24.96 (3.45); and increase the social, Part A: No sig diff between Control: 25.3 (2.60) physical, and cognitive Outcome measurement: groups activity of the volunteers memory, executive Setting: Community function (EF), and Part B: Significant difference Community-based psychomotor between intervention and control Gender: Intervention: program designed to Speed: groups at follow-up (p<0.05). 83%; Control 93% increase cognitive and Trail Making Test: Intervention group improved by physical Parts A and B 1.3 secs from baseline; control Ethnicity: Intervention: activity in a social, real- Word list memory: group declined 21.7 secs from 94% black; Control: 95% world setting Immediate recall baseline black Delayed recall Intervention (N=70): Rey-Osterrieth: Word list memory SES: Mean years Participants randomized to Copy score education: Intervention: EC trained in teams to Delayed recall No significant difference between 11.9 (2.54) ; Control: 11.2 help elementary school groups for immediate or delayed (2.66); 38% had less than children with reading recall high school education achievement, library support, and classroom Rey-Osterrieth CFT behaviour for 15 hr/week Copy score: No sig diff between groups Delayed recall: Significant difference between intervention and control groups at follow-up (p<0.05). Intervention group improved by 1.0 points and control group declined by 0.7 points 12 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Clare RCT (pilot) UK (Wales) 50+ Participants aged over Intervention: Behaviour Follow-up: 12 months Participation outcomes: Overall 2015 50 living and functioning change intervention in the whole sample, participants Mean independently in the to promote healthy Lost to follow-up: undertook 2.70 centre activi- The Agewell 68.21 (SD community, recruited ageing and reduce risk Information (control) 0%; ties (SD 2.50, range 0-13) and Trial 7.92; range through a community of dementia in later life Goal-setting: 12.5%; GS attended 34 sessions (range 51–84) Agewell Centre (aimed at promoting plus mentoring: 8.3% 35.62; 0-131). Between the 2 increased cognitive and goal-setting conditions, 137 goals N=75 randomised physical activity and Outcome measurement: were set (range 1-5; mean 2.85 improving mental and Montreal Cognitive +/- 1.2). 50 related to PA, 40 to Gender: 86.7% female physical fitness, diet and Assessment (MoCA), physical health and diet, 24 to health). immediate and delayed cognitive health and 7 to social Ethnicity: 100% white recall assessed with the engagement and 16 relating to a British or Irish 3 arms: Each arm involved California Verbal Learning mixture of categories a one-to-one interview Test (CVLT), and executive SES: 9.3% unskilled, with a researcher (up to 90 function with two subtests Cognitive outcomes 13.3% partly skilled, 9.3% minutes) from the Delis-Kaplan Montreal Cognitive Assessment: skilled manual, 28.0% Executive Function Significant increase from base- skilled non-manual, 34.7% Goal-setting (GS) (N=24): System, Trail-Making and line to follow-up in the goal-set- managerial and technical, Structured goal-setting Verbal Fluency ting with mentoring group (from 5.3% professional interview to identify up to 26.32 (2.64) to 27.23 (2.05), five goals they wished to p=0.03. No significant differences work on over the in the GS or information control coming year relating to group. physical activity, cognitive CVLT-II immediate recall: No sig- activity, physical health nificant differences from baseline and diet, and social to follow-up in any group engagement CVLT-II delayed recall: Signifi- cant increase in the information Goal-setting with control group from baseline to mentoring (GS+M) follow-up from 9.33 (3.21) to (N=24): GS as above plus 10.90 (2.73), p=0.02. No signifi- an additional 5 follow-up cant changes in the GS or GS+M mentoring phone calls group TMT T4-T2: No significant differ- Information (control) ences from baseline to follow-up N=27: Interview with in any group information about activities Verbal fluency: Significant and health and about increase in the GS group from Centre facilities 36.96 (14.58) to 43.47 (15.24, p<0.001 and in the information control group from 38.59 (10.87) to 41.74 (10.74), p=0.05.

Cost-effectiveness: Mean QA- LYs (SD); GS: 0.87 (0.17); GS+M 0.83 (0.24); control 0.85 (0.18) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Clark 2012 RCT US 60-95 Ethnically diverse older Intervention: Preventive Follow-up: 6 month Cognitive outcomes: No

Lifestyle (Mean age adults recruited from a lifestyle-based intervention significant intervention effect Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 13 oriented 74.9 (7.7) variety of community- occupational therapy was found for any cognitive based sites intervention. Weekly 2h Lost to follow-up: functioning outcome measures Well Elderly small group sessions and intervention 19.4%; control 2 RCT N=460 1rh individual sessions. 24.2% Other outcomes: Intervention Focus on involvement participants, relative to untreated 53% <$12 000 annual in activities, including Outcome measurement: controls, showed more income) physical and mental Three cognitive outcome favourable change scores on exercise variables, immediate indices of bodily pain, vitality, Gender: 65.9% female recall, delayed recall social functioning, mental health, Delivered by trained and recognition, were composite mental functioning, Ethnicity: 37.4% white; occupational therapists, measured by the word life satisfaction and depressive 32.4% black/African trained to deliver the list procedure. Selective symptomatology (ps<0.05) American; 20% Hispanic intervention attention was measured by or Latino; 3.9% Asian; median reaction time on Cost-effectiveness outcomes: 6.3% other Control: No treatment a widely used computer- The intervention group had a control group, but based visual search significantly greater increment SES: 53% <$12 000 underwent assessments task with lower scores in quality-adjusted life years annual income) indicating higher cognitive (p<0.02), which was achieved functioning; psychomotor cost-effectively (US $41 218/UK speed, assessed by the £24 868 per unit) Digit Symbol Substitution Task of the Weschler Adult Intelligence Scale Revised 14 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Cohen- RCT Israel 65+ Older adults with memory Health promotion Follow-up: 10 weeks Cognitive outcomes Mansfield complaints (N=15): structured intervention and follow-up No significant differences 2015 Note: In- Mean 73.5 format course including (p<0.05) between intervention terventions (SD 5.2) MMSE 24+ for inclusion lectures, discussions, Lost to follow-up: Health groups (health promotion; were not (Mean MMSE at baseline exercises, handouts and promotion: 20%; cognitive cognitive training; participation) conducted 28) homework; covering health training: 20%; participation: for any cognitive outcome except simultane- behaviours; 28.6%. Only completers for self-reported memory which ously N=44 randomised dementia and delirium; were analysed for results was higher in the group receiving communication; cognitive cognitive training Setting: Community activities to keep the Outcome measurement: mind fit; relationships, Global Cognitive Score There was a significant difference Gender: 72.7% female depression, and coping; assessed using the in change in Global Cognitive home and travel safety; MindStreams mild Score (GCS) between the Ethnicity: Not reported recreation and leisure; cognitive impairment cognitive training and wait-list medications and health assessment, a control groups (p<0.05). SES: Mean years care providers; physical computerized cognitive education: 14.82 (3.77); activity; and lifelong assessment. The Mini- Loneliness range 5-22 learning Mental No significant differences State Examination (p<0.05) between intervention Cognitive training and the self-report of groups (health promotion; (N=15): Memory training memory difficulties were cognitive training; participation) based on the previous also utilized. To assess for loneliness as an outcome. ACTIVE trial with a focus well-being, the UCLA on verbal episodic memory Loneliness Scale-8 There was a significant difference exercises was used. Health was in change in loneliness for evaluated by self-report the cognitive training group Participation-centred instruments compared to the wait-list control. course (N=14): Book club was used as a focus to deliver memory, cognitive and organisational strategies and using cognitive-behavioural principles. The course used external strategies (e.g. reading aid and daily planner), internal strategies (e.g. linking meaning to new information), and social interaction strategies (e.g., asking for help and sharing memory difficulties)

Control: Wait-list control (N=28) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Cohen- RCT Israel 65+ Older adults with memory Health promotion Follow-up: 10 weeks Cognitive outcomes Mansfield complaints (N=15): structured intervention and follow-up No significant differences 2015 Note: In- Mean 73.5 format course including (p<0.05) between intervention terventions (SD 5.2) MMSE 24+ for inclusion lectures, discussions, Lost to follow-up: Health groups (health promotion; were not (Mean MMSE at baseline exercises, handouts and promotion: 20%; cognitive cognitive training; participation) conducted 28) homework; covering health training: 20%; participation: for any cognitive outcome except simultane- behaviours; 28.6%. Only completers for self-reported memory which ously N=44 randomised dementia and delirium; were analysed for results was higher in the group receiving communication; cognitive cognitive training Setting: Community activities to keep the Outcome measurement: mind fit; relationships, Global Cognitive Score There was a significant difference Gender: 72.7% female depression, and coping; assessed using the in change in Global Cognitive home and travel safety; MindStreams mild Score (GCS) between the Ethnicity: Not reported recreation and leisure; cognitive impairment cognitive training and wait-list medications and health assessment, a control groups (p<0.05). SES: Mean years care providers; physical computerized cognitive education: 14.82 (3.77); activity; and lifelong assessment. The Mini- Loneliness range 5-22 learning Mental No significant differences State Examination (p<0.05) between intervention Cognitive training and the self-report of groups (health promotion; (N=15): Memory training memory difficulties were cognitive training; participation) based on the previous also utilized. To assess for loneliness as an outcome. ACTIVE trial with a focus well-being, the UCLA on verbal episodic memory Loneliness Scale-8 There was a significant difference exercises was used. Health was in change in loneliness for evaluated by self-report the cognitive training group Participation-centred instruments compared to the wait-list control. course (N=14): Book club was used as a focus to deliver memory, cognitive and organisational strategies and using cognitive-behavioural principles. The course used external strategies (e.g. reading aid and daily planner), internal strategies (e.g. linking meaning to new information), and social interaction strategies (e.g., asking for help and sharing memory difficulties)

Control: Wait-list control (N=28)

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Dannhauser Pilot in- UK Mean age People with MCI and a Physical, social and Follow-up: 12 significant Activity adherence rates were 2014 tervention 73.9 (8.3) sedentary lifestyle cognitive activities differences weeks high: physical activity = 71%; study – associated with reduced intervention and follow-up GCST = 83%; ICST = 67%. ThinkingFit crossover Mean MMSE at baseline dementia risk plus 4 week behavioural Program design but 26.3 (2.6) from N=61 pre-phase Cognitive outcomes not ran- Physical activity (3 x Significant effect (p<0.05) for domised N=67 30-45 min sessions/ Lost to follow-up: 67 cognition - improved working

week, mainly walking), started the programme memory [5.3/6.3 items]) Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 15 Gender: 58% male of group-based cognitive and were included in those who completed the stimulation (GCST): adult the analysis, and 63 Significant improvements at T2 programme education classes in arts completed more than 50% (post intervention) compared to and crafts (1 x week) of the activities offered. one but not both of T0 (baseline) Ethnicity: Not reported and individual cognitive (6% drop-out) and T1 (control pre-intervention stimulation (ICST) 3 x phase), were found for letter SES: Not reported 30 min/week: Training Outcome measurement: fluency and forward digit span was aimed at improving Neuropsychological specific cognitive functions outcome measures: Physical health and other such as attention, speed Halstead Trail Making outcomes of processing, working test (TMT) parts A and Significant effects (p <0 .05) memory, problem solving B; verbal- and category- for physical health outcomes and reasoning. Training fluency, and digit span (decreased BMI and systolic took place on the Lumosity forwards and backwards. blood pressure, [pre/post values programme that offers Life quality was measured of 26.3/25.9 kg/m2 and 145/136 different games and on the World Health mmHg respectively]), fitness puzzles and provides Organization Quality of Life (decreased resting and recovery continuous feedback of (WHOQOL)–BREF and heart rate [68/65 bpm and 75/69 performance and suggests the Alzheimer’s Disease bpm]) and quality of life games and puzzles to Cooperative Study MCI ensure balanced training Activities of Daily Living Scale (ADCS-MCI-ADL) Do something different everyday (DSD) 4 week pre-phase. Engagement and adherence were promoted by applying specific psychological techniques to enhance behavioural flexibility in an early pre-phase and during the course of the intervention

Control: Same participants during a 6 to 12 week run-in period without intervention 16 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Diamond RCT Australia 50+ Older adults ‘at risk’ of Intervention: Group- Follow-up: 7 week inter- Cognitive outcomes 2015 (Range 51- dementia, defined as based psychoeducation vention Intervention showed Healthy (Quasi 86) seeking help for new onset about cognitive strategies improvements compared to Brain Age- RCT? - Mean 66.5 cognitive impairment and/ and modifiable lifestyle Lost to follow-up: N=90 control for: ing Cogni- eligibility (8.6) or major depression, with factors relating to healthy randomised, N=64 ana- tive Training confirmed baseline MMSE score 24+. brain ageing, and lysed (23 excluded as did Verbal Memory: RAVLT % (HBA-CT) after rando- computerised cognitive not meet baseline criteria) z-score (p = 0.03) programme misation) N=90 initially randomised training. Intervention was and 3 in intervention group but N=8 in intervention conducted twice a week and none in control group Self-reported memory: Everyday group and N=15 in the for seven weeks. Each discontinued intervention Memory Questionnaire (EMQ) (p control group were group session comprised (due to personal reasons). = 0.03) subsequently excluded as a maximum of 10 they did not meet baseline participants and included Outcome measurement: Other outcomes inclusion criteria (based one-hour of Healthy Brain Battery of neuropsycho- Mood (p = 0.01), and sleep (p = on review of their medical/ Ageing logical tests, psychiatric 0.01). psychiatric assessment psychoeducation (covering and medical assessment None of the other outcomes and diagnosis) memory strategies, diet tests:- measured showed significant and exercise, using the RAVLT-15 Rey Auditory differences between intervention Setting: Clinic internet) and one-hour of Verbal Learning Test-total and control computer based cognitive learning over 5 trials; RAV- Gender: Intervention: 25% training LT%, Rey Auditory Verbal While the improvements in male; Control: 42.8% Learning Test-percent memory (p = 0.03) and sleep (p Control: Wait list control retention scores (i.e., (Trial = 0.02) remained after controlling Ethnicity: Not reported group (no contact from 7 / Trial 5)* 100); LOG- for improvements in mood, researchers but did receive MEM-I, total score for sto- only a trend in verbal memory SES: Mean years standard clinical care from ries A and B on Wechsler improvement was apparent after education: Intervention: their usual health-care Memory Scale-III Logical controlling for sleep 14.3 (3.4); Control: 13.7 professionals) Memory learning trials; (3.2) LOGMEM%, Wechsler Memory Scale-III Mean MMSE at baseline Logical Memory percent 28.4 (1.5); 81% of retention scores (LOG- participants met criteria for MEM-II delayed recall / MCI [LOGMEM-I story A+ second recall Story B]); RCFT, Rey Osterrieth Complex Figure Test; TMT-A, Trail Making Test Part A; TMT-B, Trail Making Test Part B; GDS, Geriatric Depression Scale (30-item); EMQ, Everyday Memory Questionnaire; PSQI, Pittsburgh Sleep Quality Index; ASS, Age Scaled Score. Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Diamond RCT Australia 50+ Older adults ‘at risk’ of Intervention: Group- Follow-up: 7 week inter- Cognitive outcomes 2015 (Range 51- dementia, defined as based psychoeducation vention Intervention showed Healthy (Quasi 86) seeking help for new onset about cognitive strategies improvements compared to Brain Age- RCT? - Mean 66.5 cognitive impairment and/ and modifiable lifestyle Lost to follow-up: N=90 control for: ing Cogni- eligibility (8.6) or major depression, with factors relating to healthy randomised, N=64 ana- tive Training confirmed baseline MMSE score 24+. brain ageing, and lysed (23 excluded as did Verbal Memory: RAVLT % (HBA-CT) after rando- computerised cognitive not meet baseline criteria) z-score (p = 0.03) programme misation) N=90 initially randomised training. Intervention was and 3 in intervention group but N=8 in intervention conducted twice a week and none in control group Self-reported memory: Everyday group and N=15 in the for seven weeks. Each discontinued intervention Memory Questionnaire (EMQ) (p control group were group session comprised (due to personal reasons). = 0.03) subsequently excluded as a maximum of 10 they did not meet baseline participants and included Outcome measurement: Other outcomes inclusion criteria (based one-hour of Healthy Brain Battery of neuropsycho- Mood (p = 0.01), and sleep (p = on review of their medical/ Ageing logical tests, psychiatric 0.01). psychiatric assessment psychoeducation (covering and medical assessment None of the other outcomes and diagnosis) memory strategies, diet tests:- measured showed significant and exercise, using the RAVLT-15 Rey Auditory differences between intervention Setting: Clinic internet) and one-hour of Verbal Learning Test-total and control computer based cognitive learning over 5 trials; RAV- Gender: Intervention: 25% training LT%, Rey Auditory Verbal While the improvements in male; Control: 42.8% Learning Test-percent memory (p = 0.03) and sleep (p Control: Wait list control retention scores (i.e., (Trial = 0.02) remained after controlling Ethnicity: Not reported group (no contact from 7 / Trial 5)* 100); LOG- for improvements in mood, researchers but did receive MEM-I, total score for sto- only a trend in verbal memory SES: Mean years standard clinical care from ries A and B on Wechsler improvement was apparent after education: Intervention: their usual health-care Memory Scale-III Logical controlling for sleep 14.3 (3.4); Control: 13.7 professionals) Memory learning trials; (3.2) LOGMEM%, Wechsler Memory Scale-III Mean MMSE at baseline Logical Memory percent 28.4 (1.5); 81% of retention scores (LOG- participants met criteria for MEM-II delayed recall / MCI [LOGMEM-I story A+ second recall Story B]); RCFT, Rey Osterrieth Complex Figure Test; TMT-A, Trail Making Test Part A; TMT-B, Trail Making Test Part B; GDS, Geriatric Depression Scale (30-item); EMQ, Everyday Memory Questionnaire; PSQI, Pittsburgh Sleep Quality Index; ASS, Age Scaled Score.

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Mendoza- RCT Mexico 60+ Healthy older adults “I am Active” program Follow-up: 8 weeks Cognitive function Ruvalcaba recruited from senior designed to promote intervention, followed up to Processing speed score: The in-

2015 Mean: In- centres active aging by increased 6 months tervention group had significantly Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 17 tervention: physical activity, healthy better processing speed at post- 70.45±6.37; N= 64 randomised nutritional habits, and Lost to follow-up: test and follow-up than control. Control cognitive functioning Intervention: 13.9%; 70.82±7.20 Gender: 93.5% female control: 9.1% Intervention group: Significantly in intervention group and Intervention (N=31): increased (P<0.001; for differ- 84.8% female in control The participants in Outcome measurement: ence from baseline at post-test group the intervention group Cognition: Working and follow-up) from 26.06 (9.99) received a “user manual” memory was assessed by at baseline to 31.16 (10.34) Ethnicity: Not reported specifically designed for the Digit Span Backward after the intervention and 30.52 this purpose. The program Subtest and processing (10.21) at follow-up SES: Mean years lasted 2 months, and speed by the Digit Symbol education: intervention: consisted of 2-hour group Subtest Control group: 19.87 (9.22) at 5.55±3.12; control sessions, held twice a baseline; 17.83 (9.18) after the 3.97±3.28 (p=0.054) week (i.e., 16 sessions Nutrition: Nutritional intervention and 19.41 (8.74) at in total). Each session status was measured follow-up (not significant) included physical activity using the Mini Nutritional and reality orientation with Assessment Working memory: No significant eight lessons including changes from baseline in working nutritional topics and eight memory performance were found sessions cognitive activity in either group

Control (N=33): Other outcomes: Remained on a wait At the end of the intervention list and participated in (8wks), significant improvement the program once the compared with the control group study was completed, for: physical activity (falls risk, participating in the balance, flexibility, self-efficacy), meantime in weekly social nutrition (self-efficacy and nutri- activities organized by the tional status), cognitive self-effi- senior center cacy, and quality of life (general, health and functionality, social and economic status)

At follow-up (6 months), improvements remained for self-efficacy for physical activity, self-efficacy for nutrition, nutritional status, quality of life 18 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Miller 2012 Non-ran- US Mean age: Older adults living at Intervention: 6-week Follow-up: 6 week Cognitive tests domised 80.9 (SD: the independent level of educational program on intervention Significant improvements from controlled 6.0 years) care in continuing care memory training, physical pre- (control condition) to post study retirement communities activity, stress reduction, Lost to follow-up: 18.3% intervention on recognition and reporting memory and healthy diet of word pairs (t[114] = 3.62, p complaints Outcome measurement: < 0.001); retention of verbal Control: Waiting list Objective cognitive information from list learning >24 on MMSE: mean control measures evaluated (t[114] = 2.98, p < 0.01); MMSE score 28.6 (1.3) changes in five domains: immediate memory (t[114] = immediate verbal memory, 6.56, p < 0.0001); delayed N=115 assigned to groups delayed verbal memory, memory (t[114] = 2.70, p < 0.02) retention of verbal Gender: 79% female information, memory No improvement was found for recognition, and verbal verbal fluency Ethnicity: 98% white fluency

SES: 32% had high school education; 44% bachelor or associate degree; 24% a graduate or professional degree Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Ngandu RCT Finland 60-77 Inclusion criteria were Intervention (N=631): Follow-up: 2 year Comprehensive

2015 CAIDE (Cardiovascular Diet, exercise, cognitive intervention neuropsychological test Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 19 Intervention Risk Factors, Aging and training, and vascular risk battery (NTB) Z score: FINGER group: 69·5 Dementia) Dementia Risk monitoring with advice and Lost to follow-up: 6% Estimated mean change in NTB Trial (4·6) Score of at least 6 points feedback intervention; 5% control total Z score at 2 years was and cognition at mean 0·20 (SE 0·02, SD 0·51) in the Control level or slightly lower than Control (N=629): General Outcome measurement: intervention group and 0·16 group: 69·2 expected for age health advice. Advice and (0·01, 0·51) in the control group. (4·7) feedback on metabolic and NTB total score: A Between-group difference in the Mean MMSE score at vascular risk factors also composite score based change of NTB total score per baseline 26.8 (2.0) given to the control group on results from 14 year was 0·022 (95% CI 0·002– tests (calculated as Z 0·042, p=0·030) N=1260 randomised scores standardised to the baseline mean Adverse events: Occurred Gender: Intervention: and SD, with higher in 46 (7%) participants in the 45%; Control 47% scores suggesting better intervention group compared performance) with six (1%) participants in Ethnicity: Not reported the control group; the most Secondary Outcomes: common adverse event was SES: Education: mean: Included NTB domain Z musculoskeletal pain (32 [5%] 10.0 (3.4) in both groups scores for executive individuals for intervention vs no functioning, processing individuals for control) speed, and memory. The executive functioning domain included category fluency test,19 digit span, concept shifting test (condition C), trail making test (shifting score B – A), and a shortened 40-stimulus version of the original Stroop test (interference score 3 – 2) 20 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Pitkala 2011 RCT (pilot) US Mean: In- Older adults Intervention (N=117): Follow-up: 3, 6 months for ADAS-Cog scale (3 months) tervention: Socially stimulating group cognitive outcomes. Mean changes (all participants) (cluster 70.1 (6.42) N=149 randomised intervention aimed at Lost to follow-up: randomised Control : enhancing interaction Intervention: 6.0%; control: I: -2.6points (95% CI -3.4 to -1.8) by school) 68.4 (5.15) Baseline MMSE: Interven- and friendships between 17.8% C: -1.6 points (95% CI -2.2 to tion: 24.96 (3.45); Control: participants and to -1.0) 25.3 (2.60) stimulate them socially. Outcome measurement: (p=0.023; F1,167.8 = 5.23) Groups were facilitated Cognition by Alzheimer’s Setting: Community by trained professionals. disease assessment scale Art group Participants were divided (ADAS-cog) and mental I: -2.4 points (95% CI -3.5 to -1.3) Gender: Intervention: into 3 groups depending function by 15D measure C: -1.8 points (95% CI -2.9 to 83%; Control 93% on their interests: and psychological -0.8) therapeutic writing (N=48); wellbeing and HRQoL at (p=0.017; F1,47.2 = 1.88) Ethnicity: Intervention: group exercises (N=92); 12 months 94% black; Control: 95% or art experiences (N=95) Exercise group black and then randomised to I: -3.2 points (95% CI -4.7 to -1.7) intervention or control C: -1.6 points (95% CI 2.6 to – SES: Mean years ed- within those groups. 0.5) ucation: Intervention: Intervention was once (p=0.60; F1,72.7 = 0.28) 11.9 (2.54) ; Control: 11.2 per week and usually (2.66); 38% had less than lasted for 6 hours and was Writing group high school education provided free of charge. I: -1.7 points (95% CI -2.7 to -0.7) Control (N=118): C: -1.2 points (95% CI -2.7 to Continued in normal 0.3) community care. They (p=0.033; F1,33.6 = 4.49) could participate in 15D index dimension of mental their normal hobbies function) over 12 months and activities but no I: +0.048 (95%CI: +0.013 to intervention was arranged +0.085) for them C: -0.027 (95%CI: -0.063 to +0.010) (p=0.004; t = 2.89, df=187)

Note: Cognitive outcomes also measured at 6 months but not reported here Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Tesky 2011 RCT Germany 50+ Older adults without AKTIVA intervention (n Follow-up: 8 week Participation in the group

AKTIVA dementia or cognitive = 126): Group programme intervention and post- program resulted in positive Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 21 study Mean: 72 impairment of cognitively stimulating test conducted 1 week effects on cognitive function (7) leisure activities (8 weekly after (week 9); 2 booster and attitude toward aging for N=307 randomised sessions and two booster sessions conducted at subassembly groups. Older sessions after a break of 4 27 to 28 weeks and then persons (≥75 years) showed Setting: months) follow-up tests conducted enhanced speed of information 29 weeks after start of processing (by TMT Version Gender: 72.3% female AKTIVA intervention plus intervention A) (F = 4.17*, p <.05); younger nutrition and exercise participants (< 75 years) showed Ethnicity: ‘German counselling (n = 84): Lost to follow-up: N=67 an improvement in subjective ethnicity’ reported (21.8%) withdrew from the memory decline (by MAC-Q) (F Control group (n = 97): No study. Those with impaired = 2.55*, p < .05). Additionally, SES: ‘Most participants intervention cognition (N=28) and AKTIVA enhanced the frequency had attended school for N=15 Turkish participants of activities for leisure activities about 10 years, and only a excluded from analysis for subassembly groups. The few participants after the programme. results of this study suggest had completed an that the AKTIVA program can academic university Outcome measurement: be used to increase cognitively education’ Cognitive outcomes: Mini- stimulating leisure activities in Mental Status Test the elderly. Further research is necessary to identify the long- ADAS-Cog: the cognitive term effects of this intervention part of the Alzheimer’s particularly with respect to the Disease Assessment prevention of dementia Scale

Part A and B of the Trail- Making Test

Clinical Dementia Rating (CDR) 22 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Wiegand RCT Canada 50-90 Community-dwelling older Intervention: A Follow-up: 5 weeks Cognitive/memory tests 2013 recruited adults without MCI or other multidimensional, intervention, and one Objective tests: (i) Name-learning medical conditions that evidence-based month after end of test: No significant differences Mean age: affect cognitive ability intervention, the Memory intervention (group x time); (ii) Fact-learning I: 70.3 (8.2); and Aging Program, that task: No significant differences C 72.1 (9.8) N=45 randomised provides education about Lost to follow-up: 2.4% (group x time) memory and memory Gender: 76.2% female change, training in the Outcome measurement: Lifestyle behaviours use of practical memory Pencil and paper tests and Relative to the control group, Ethnicity: Not reported strategies, and support questionnaires participants in the program for implementation of implemented more healthy SES: Mean years healthy lifestyle behaviour For name-learning and lifestyle behaviours by the end education 14.0 (2.3) in changes fact learning test, after of the program (p=0.015) and intervention group and presentation, participants maintained these changes 1 15.1 (2.9) in the control Control: Waitlist control, were asked to write down month later (p=0.007). The group participated in pre- and as many as they could most common behaviours post tests remember implemented were relaxation and cognitive engagement (post intervention) and PA (after 1 month)

Intention to seek medical attention for memory Intervention participants reported a decreased intention to seek unnecessary medical attention for their memory immediately after the program and 1 month later Table 2: Multicomponent – Interventions with cognitive outcomes: combined PA and cognitive stimulation

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Anderson- Cluster US 55+ Participants recruited from Intervention (N=38): Follow-up: 3 months Cognitive outcomes

Hanley 2012 RCT independent living facilities Physical exercise intervention and follow-up Executive function: There Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 23 Mean: plus mental challenge were significant differences Am J Prev cybercycle 42.1% had clinical combined in a cybercycle Lost to follow-up: in change from baseline Med 75.7 (9.9); diagnosis of MCI at exergame (virtual reality– Intervention: 21.0% between intervention and control bike baseline in the cybercycle enhanced exercise control groups for all 3 tests of Reg Clin 81.6 (6.2) group and 34.1% in the that combines physical Control: 19.5%. executive function: Color Trails Trial control group exercise with computer- p=0.007; Stroop C p=0.05; simulated environments All participants at baseline Digits backward p=0.03. Overall Gender: Cybercycle: and interactive videogame included in analysis composite executive function 70.7% female; control bike features (e.g., the Wii Fit score p=0.002 86.8% female and PlayStation Move) Outcome measurement Primary cognitive Other cognitive outcomes Ethnicity: Not reported Control (N=41): Physical outcome: Executive There were no significant exercise alone (stationary function assessed via between group differences for SES: Mean years bike identical to the cyber Color Trails 2-1 difference changes from baseline for other education: Cybercycle bike except for virtual Score; Stroop C and Digit cognitive outcomes 12.6 (2.2); control bike reality control) Span Backwards. 14.8 (2.3) (Results above based on intent- Secondary cognitive to-treat analyses, controlling outcomes: Attention: for age, education, and cluster LDST, Letter Digit Symbol randomization). Test; Verbal fluency: COWAT, Controlled Cybercyclists had a 23% Oral Word Association relative risk reduction in clinical Test, categories; Verbal progression to MCI memory (immediate): RAVLT (sum 5 trials score), RAVLT immediate recall); Verbal memory (delayed) RAVLT delayed recall, Fuld delayed recall; Visuospatial skill: Figure copy, clock; visuospatial memory (delayed): figure delayed recall score;

Clinical diagnosis of MCI 24 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Barnes RCT (facto- US Mean age Inactive, community Home-based mental Follow-up: 12 weeks Composite cognitive scores 2013 rial design) 73.4 residing older adults with activity (1 h/d, 3 d/wk) intervention No significant difference between cognitive complaints plus class-based physical groups but improved significantly Mental activity (1 h/d, 3 d/wk) in 4 Lost to follow-up: from baseline in all groups Activity and N=126 groups: computer (aerobic) 28.1%; except MA-C/EX-1. The greatest eXercise computer stretching increase was in the most (MAX) trial Gender: 62.7% female MA-I/EX-1 (intensive control 19.5%; DVD intensive mental activity and computer mental activity/ (aerobic) 9.7%; DVD exercise group, but there was no Ethnicity: 34.9% Hispanic aerobic exercise) (stretching control): 25%. significant difference between Registered or non-white groups clinical trial MA-I/EX-C (intensive Outcome measurement: SES: 15.6 – 16.8 years computer mental activity/ Global cognitive change Change from baseline: depending on group (SD control stretching and based on a comprehensive MA-I/EX-1: 0.22 (0.12 to 0.33), 2.1-2.8) toning exercise) neuropsychological test p=<0.001 battery MA-I/EX-C: 0.17 (0.03 to 0.31), MA-C/EX-1 (mental activity p=0.01 control – exercise DVDs)/ MA-C/EX-1: 0.08 (0.004 to 0.17), aerobic exercise) p=0.06 MA-C/EX-C :0.16 (0.05 to 0.26), MA-C/EX-C (mental p=0.003 activity control – exercise DVDs)/control stretching Visuospatial function (UFOV): and toning exercise significant between group differences for divided attention and selective attention

No significant difference between groups for Verbal learning and memory (RAVL-T); verbal fluency, processing speed (DSST); executive function Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Fabre 2002 RCT France 60+ (range Older adults recruited Intervention: Combined Follow-up: 2 months Wechsler memory quotient: 60 to 76) from (leisure) clubs. None aerobic and cognitive intervention All 3 intervention groups Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 25 engaged in regular PA at training: incorporating 1 achieved a statistically significant baseline and 2 below Lost to follow-up: none difference in Wechsler overall reported memory quotient from baseline N=32 randomised Comparator 1: Aerobic to follow-up at 2 months (F=6.52, training alone: 2 Outcome measurement: DF=1, p<0.01). There was Gender: 15.6% male supervised one hour French BEC 96 no significant improvement exercise sessions per questionnaire to measure in the control group. Memory Ethnicity: Not reported week cognitive problems: recall, quotient improved by 9.2% in the learning, orientation, combined cognitive and aerobic SES: Mean years Comparator 2: Cognitive manipulation and mental training group compared to 8.5% education: 11-12 years training alone: One 90 problems, verbal fluency, in the aerobic training group min training session denomination and visual alone; 7.4% in the cognitive per week on 8 themes: reproduction. Wechsler training group alone; and 0.8% perceptive activities, memory scale: general in the control group. The mean attention, intellectual information, orientation, difference in pre- and post- structuration, association mental control, logical intervention memory quotient and imagination, language memory-immediate was significantly higher in the spatial marks, temporal recall, digit span in order combined aerobic and cognitive marks, associated and reverse, visual training group than in the aerobic recruitment reconstructions and paired or cognitive training groups alone associates learning. The (F=11.60, DF=3, p<0.001) Control: Followed usual combined results give a daily routine activities memory quotient score Wechsler subtests: All 3 without training but met as intervention groups showed many times as the other significant improvements in groups for leisure activities paired associates learning: e.g. painting, singing (F=5.47, DF=1, p<0.05) and logical memory-immediate recall ((F=4.31, DF=1, p<0.05) but no significant difference between the intervention groups

BEC 96 questionnaire: No changes after 2 months for any of the groups 26 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Hars 2013 Secondary Switzerland 65+ Community-dwelling older Intervention (N=66): Follow-up: 6 months Cognitive outcomes: analysis of adults at increased risk Structured music-based intervention No statistically significant an RCT Mean: Inter- of falling (1+ falls aged multitask exercise classes differences between intervention vention: 75 65 or over or balance (N=66). Group exercise Lost to follow-up: 15% and control for MMSE, clock (8); control impairment). classes once weekly intervention; 21% control. drawing test or overall FAB score 76 (6) for 1-h over 25 weeks All participants included in N=134 randomised that included multitask ITT analyses There was an improvement in the exercises of progressive sensitivity to interference subtest Setting: Community difficulty, sometimes Outcome measurement: of the FAB (adjusted between- involving the handling MMSE, the clock- group mean difference (AMD), Gender: 96% female of objects, performed drawing test, the frontal 0.12; 95% CI, 0.00 to 0.25; P = individually, in pairs or assessment battery (FAB) 0.047) Ethnicity: Not reported more. Such as: walking and the hospital anxiety following the piano music; (HADS-A) and depression Within-group analyses showed SES: Education: responding directly or scale an increase in MMSE score Intervention: 11% had oppositely to changes in (P = 0.004) and a reduction completed primary school music’s rhythmic patterns; in the number of participants education; 68% middle quick reaction exercises with impaired global cognitive school; 28% high school. and walking out of performance (i.e., MMSE Control: 19% primary; rhythmic patterns score ≤23; p = 0.003) in the 66% middle and 15% high intervention group school Control (N=68): Control group with delayed Other outcomes: intervention (N=68) Reduction in anxiety level (HADS-A; AMD, −0.88; 95% CI, −1.73 to −0.05; P = 0.039) for intervention participants, compared with the controls Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Hughes RCT US Mean: In- Older adults with MCI Intervention (N=10): Follow-up: 24 weeks Adherence: The Wii group 2014 tervention: recruited from a population Group-based interactive attended an average of 23.1 Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 27 78.5 (7.1); cohort study video gaming using Lost to follow-up: 10% (SD 1.1, range 21–24) sessions control: Nintendo Wii; 90 mins/ (1 died, 1 did not complete compared with 21.8 (SD 3.3, 76.2 (4.3) Mean MMSE at baseline week for 24 weeks follow-up assessment) range 14–24) in the control 27.1 (1.8) group; 18 participants attended The Wii Sports Outcome measurement: at least N=20 randomised games, including bowling, Cognitive function: 20/24 sessions; and 9 attended golf, tennis, and baseball Computerized Assessment all sessions Setting: Community were the core of the of Mild Cognitive sessions. From week Impairment (CAMCI) Cognitive function: No Gender: Intervention: 7, participants were Subjective cognitive ability: significant differences between 80% female; control: 60% introduced to new games Cognitive Self-Report intervention and control groups female for 15-20 mins of the Questionnaire-25 session. In weeks 10 and Social functioning: No significant differences for any Ethnicity: Intervention: 20 participants competed Cognitive Self-Report other outcomes measured 70% white; control: 90% in Wii tournaments to Questionnaire-25 white. encourage enhanced effort Other outcomes; mood, and social activation IADL, gait speed. SES: Mean years education: Intervention Control (N=10): Health 13.8 (2.4); control 13.1 education designed to (1.9) provide a source of passive cognitive stimulation in a socially matched setting ; 90 mins/week for 24 weeks. 28 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Gonzalez- Before and Spain 60+ Healthy elderly and people Computer-based Follow-up: 12 weeks Global cognitive function (MEC Palau 2014 after study with MCI recruited from cognitive and physical intervention (post test 35): significant pre-and post-test Mean 73.4 community centres training programme: one 1-2 weeks after end of differences in both healthy and Long (7.51) hour’s physical training intervention) MCI participants (p=0.04) Lasting N=50 (11 with MCI; 39 and 35 minutes’ cognitive Healthy: Pre: 30.91 (3.05); Post: Memories healthy people) training, 3 times a week, Lost to follow-up: 12.0% 31.84 (2.50) programme conducted at community MCI: Pre: 29.61 (3.53); Post: Gender: 80.5% female centres where participants Outcome measurement: 30.44 (3.8) usually went Battery of cognitive Verbal and episodic memory Ethnicity: Not reported measures that included: Significant pre- and post-test other than 100% Spanish Cognitive training: MMSE (Spanish version); differences in both healthy and Used Gradior cognitive Digit Span Test of the MCI participants SES: Mean years training software (a Wechsler Memory Scale III HVLT-R: education: 9.1 (3.2) neuropsychological (WMS III); Logical Memory Recognition assessment system and subtests of the WMS Healthy: Pre: 10.19 (1.19); Post: multi-domain cognitive III; The Color Trail Test 10.61 (1.45); training program including 1 and 2 (CTT 1 and 2); MCI: Pre: 10.06 (2.53); Post: attention, perception, Hopkins Verbal Learning 10.22 (0.943); p=0.006 episodic memory and Test Revised (HVLT-R); Free recall working memory tasks, Geriatric Depression Scale Healthy: Pre: 18.00 (5.69); Post: and incorporating feedback (GDS) 19.32 (5.02); and difficulty MCI: Pre: 14.50 (4.85); Post: 18.33 (6.61); p=0.004 Physical training: Used Delay recall Fit For All game platform Healthy: Pre: 4.71 (2.88); Post: that can help elderly 6.32 (2.77); people to keep fit and MCI: Pre: 3.00 (3.23); Post: 6.00 maintain their wellbeing (2.97); p<0.0001 through an innovative, low- WMS III logical memory cost ICT platform, such as Significant effects also seen for Wii Balance Board pre-test/post-test subtests of logical memory (p=0.02; p<0.0001) Attention CTT1: significant effect only for healthy participants (p=0.04) GDS Symptoms of depression decreased in both healthy and MCI participants Overall, study showed an improvement after the training in global cognitive function and verbal memory, which involved recognition, free recall and delay recall, in attention, in episodic memory and in symptoms of depression Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Kamegaya RCT Japan 65-87 Community-dwelling Physical and leisure Follow-up: 12 weeks Between group analyses: 2014 elderly people (30.8% with activity programme for the The intervention group showed Mean age amnestic MCI) prevention of cognitive Lost to follow-up: significant improvement on the 64.9 (5.9) decline, aimed at ‘enhanc- Intervention: 26.9% (11.5% analogy task of the Five-Cog test N=52 ing participants’ motivation completed <7 sessions (F1,38 = 4.242, P = 0.046) and to participate and support and were excluded); improved quality of life (F1,38 = Mean MMSE score at one another by providing Control: 3.7% 4.773, P = 0.035) compared to

baseline 27.7 (1.8). a pleasant atmosphere, the control group (n = 24) Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 29 empathetic communica- Outcome measurement: Gender: 90.4% female tion, praise, and errorless Other between group tests were support’. Programme was Cognitive outcomes: non-significant Ethnicity: Not reported provided as a community Five-Cog test, which service available to all evaluated the cognitive Within-group analyses: SES: Mean years community dwellers) domains of attention, Intervention group: significant education: 11.2 (2.2) memory, visuospatial changes for: character position Intervention (n=26): Re- function, language, and ref task, cued recall task, animal ceived intervention once a reasoning. Executive name listing, analogy task, week (2h programme) at function was evaluated by WDSST and Satisfaction in Daily a community centre, con- the Wechsler Digit Symbol Life (SDL: subjective measure of ducted by healthcare pro- Substitution Test (WDSST) QoL). Control group: significant fessionals, with the help of and Yamaguchi Kanji- differences for: character position senior citizen volunteers Symbol Substitution Test ref task, cued recall task, The physical activity pro- WDSST and grip strength gramme was the primary Other outcomes: content of the programme. Subjective health status, The exercise programme level of social support, included muscle-stretch- functional capacity, ing exercise in a sitting subjective quality of life, position (17 items), mus- and depressive symptoms cle-strengthening exercise were assessed with in a sitting position (3 a questionnaire. Grip items); and encouraged to strength test, timed up- perform exercise at home. and-go test, 5-m maximum Mean duration was 45 walking times test, and mins. Walking was recom- functional reach test were mended to participants as performed to evaluate a regular exercise physical function Leisure activities, such as cooking, handcrafts and competitive games, were included in the programme to stimulate cognitive function

Control (n=26): Did not attend a programme 30 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Legault RCT (pilot) US 70-85 Community-dwelling older Cognitive training Follow-up: 4 month Intervention attendance: Rates 2011 adults at risk for cognitive intervention and/or a phys- intervention were higher in the cognitive Mean 76.4 decline but without MCI ical training intervention training and combined groups: The Seniors Lost to follow-up: Cognitive training: 96%; Physical Health and Mean 3MSE score 94.8 Cognitive Training Cognitive training: 11.1%; training: 76%; combined 90%. Activity Re- (N=18): Sessions were Physical training 11.1%; (p=0.004) search Pro- N=73 randomised centre-based, word recog- Combined: 5.3%; Healthy gram Pilot nition tests conducted via aging control 5.6%. Cognitive outcomes (SHARP-P) Gender: 51% women computer, carried out with All intervention groups and the Study small groups of no more Outcome measurement: control showed improvements Ethnicity: Healthy than six individuals, and The Self-Ordered Pointing from baseline to follow-up in most training/cognitive training/ monitored by skilled train- Task to measure planning, cognitive outcomes (p≤0.05) physical activity training: ers. Training consisted of working memory and 6% African American; 94% four consecutive 10-12 min monitoring; Eriksen flanker No between-group statistically Caucasian; combined sessions per day, twice a test to measure response significant differences in 4-month intervention: 21% African day for two months, then inhibition; Task switching changes in composite scores American; 79% Caucasian one time per week for two test to measure attentional of cognitive, executive, and additional months flexibility; Trail making test episodic memory function SES: 75% had some post to measure alternating high school education Physical Activity training attention and executive This pilot study concluded that (N=18): Centre-based function; Four measures of future two-armed full-scale and home-based sessions episodic memory derived trials may require fewer than aimed primarily at aero- from the Hopkins Verbal 1,000 participants (continuous bic and flexibility training Learning Test and the outcome) or 2,000 participants with a targeted duration Logical Memory task from (categorical outcome) of 150 minutes/week, with the Wechsler Memory two center-based training Scale-III were also This pilot study is unlikely to be sessions per week for included sufficiently powered (see above) four months. The primary – SK comments focus was walking with the aim of improving cardiovascular fitness. Other endurance activity (e.g., stationary cycling) were used when regular walking was not appropriate

Combined intervention (N=19): Cognitive and physical training delivered on the same day (cognitive first)

Healthy Ageing Control (N=18): Healthy ageing education Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Linde 2014 RCT Germany 60-75 Community-dwelling older Comparison is combined Follow-up: 4 month All interventions led to an

adults recruited through exercise and cognitive intervention plus a further improvement in cognitive Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 31 Mean age newspapers and senior training versus exercise 3 months follow-up performance though only 2 out 67.1 (3.6) citizens associations. All training alone, cognitive of 7 cognitive outcomes were independent living and training alone or no Lost to follow-up: enhanced with different patterns functional intervention control Exercise training: 21.0%; in the 3 intervention groups Cognitive training: 38.9%; N=70 randomised Exercise training Combination exercise and Concentration: The exercise, (N=19): Moderate aerobic cognitive training: 5.9%: cognitive and combined 51% had not regularly endurance training and Control 18.8% training groups had improved participated in regular moderate strength training, concentration at the end of the sports or physical activity 60 mins, twice a week Note: participants lost at intervention compared to control, in the past year post-assessment or follow- however this was only significant Cognitive training up stages performed for the exercise training group Gender: 59% female (N=18): Primary element significantly worse on (p=0.01) was individual editing reasoning and spatial Ethnicity: Not reported of worksheets, with relation subtests than Cognitive speed: The exercise, some partner and group those who remained in the cognitive and combined SES: 47% had a university exercises. Focus on study training groups had improved degree short-term memory, visuo- concentration at the end of the spatial skills, information Outcome measurement: intervention compared to control, processing speed, Reasoning and spatial however this was only significant concentration and logical relations were assessed for the combined exercise and reasoning; once/week for with a subtest of the LPS cognitive training group (p=0.03) 30 mins 50+ performance test system; concentration All other cognitive outcomes Combined exercise measured with the d2:Test were not significantly different and cognitive training of attention; processing from the control (N=17): Consisted of both speed (Trail making test the exercise training and part A); cognitive speed cognitive interventions assessed with the digit- above and took place symbol substitution test; twice/week. Cognitive short-term memory with training was carried out the word-list test at the first session of the week, before the exercise training

Control (N=16): Waiting list control 32 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Maillot 2012 RCT France 65-78 Independently living older Intervention (N=16): Follow-up: 12 week Overall adherence to the inter- adults who reported never Exergaming intervention intervention vention was 97.5% Mean: playing video games and using the Nintendo Wii. interven- with a sedentary lifestyle Only games of physically Lost to follow-up: 6.2% Executive function tion 73.47 simulated sport were (n=1) in both groups All six measures of executive (4.10); con- Mean MMSE at baseline: selected on this exercise function showed significant trol 73.47 Intervention:28.67 (1.17); program: Wii Sports, Wii Outcome measurement: improved changes compared to (3.00) control 29.27 (0.88) Fit, and Mario & Sonic on Cognitive battery including: control (p<0.05) Olympic Games Executive control tasks N=32 randomised (Trail-Making test, Stroop Processing speed Control (n=16): No Color Word Interference Out of eight measures of pro- Gender: 84.4% training, no contact control test, Letter Sets test, cessing speed, all measures group Matrix Reasoning test2, showed significant improvement Ethnicity: Not reported Digit Symbol Substitution compared to control (p<0.05) test); visuo-spatial tasks SES: Mean years (Spatial Span test, Visuospatial tests education: intervention Directional Headings test, Changes in 4 visuospatial tests 11.2 (1.78); control 11.40 Mental Rotation test); were not significant except for (2.22) and processing-speed Directional headings (number) tasks subdivided into two categories: perceptual speed (Cancellation test and Number Comparison test) and psychomotor speed (the Reaction Time test and Plate Tapping test)

McDaniel RCT US 55-75 Older adults living in the Intervention 1: Combined Follow-up: 6 months Cognitive training only: Sig- 2014 community, those with less exercise and cognitive intervention and follow- nificant improvement in memory Mean age than 10th grade education intervention up (cognitive training 2 processing aspects of virtual (combined excluded months) week test. No significant im- group) Intervention 2: Exercise provement for cooking breakfast N= 96 randomised intervention only Lost to follow-up: 17.7% and memory for health informa- 64 (6) overall tion Gender: 60-71% female Intervention 3: Cognitive (depending on group) intervention only Outcome measurement: Exercise only: No significant ef- Primary outcome fect for any of the 3 tasks. Partic- Ethnicity: 71-96% white Control: Low intensity measures were 3 ipants in the two aerobic exercise (depending on group) home-exercise program for laboratory tasks that training groups achieved greater 6 months, and participated simulated everyday improvements in peak aerobic SES: Mean years in face-to-face health activities: Cooking power than the control group education 15-17 years education sessions for Breakfast, Virtual Week, 8 weeks (Months 5 and and Memory for Health Combination cognitive training 6, to correspond to the Information and exercise: No significant im- cognitive-training protocol) provement for any of the 3 tasks Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Oswald Quasi- RCT Germany 75-93 Older adults living Combined PA Follow-up: 5 years Cognitive function (compos- 2006 (some independently in the and cognitive or Lost to follow-up: 1) 1 year ite): Significant intervention vs exceptions Mean 79.5 community without psychoeducational 25%, 5 years 48.5%. 2) control differences over 5 years The SIMA from the (3.5) functional or cognitive intervention compared 1y 22.2%; 5 y 50%; 3) 1 for cognitive exercise (p<0.001); study randomisa- decline or hearing loss or to single PA or y 19.3%; 5 y 49.1% 4) 1y cognitive and physical exercise tion) visual impairment single cognitive or 9.4%, 5 y 53.1 % 5) 1y (p<0.001); psychoed alone psychoeducational 26.1%, 5y 59.1% 6) 1y (p<0.05) and psychoed and N=375 assigned training. Training every 5.8%, 5 y 48.5%. physical exercise (p<0.005) but week over 30 sessions, Outcome measurement not for physical group alone. Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 33 Gender: 64.8% female in small groups of 15-30, Cognitive Function: delivered by trained group Speed of information Cognitive impairment (com- Processing: Number con- Only significant differ- Ethnicity: Not reported leaders with standardised posite): nection test (NC-G), Maze ence was for cognitive and phys- written information test (MT-G), Digit symbol ical (p<0.001). All other groups, SES: 5.3% university manuals substitution test (DS-G) no significant difference. graduates; 14.4% German (Neuropsychological aging university entrance exam; 1) Cognitive and physical inventory NAI) Physical function: Only signifi- 39.2% secondary school training (N=32) Attention: Alters- (Ag- cant difference was for cognitive education; 41.1% primary ing concentration test); and physical (p<0.05). All other school education 2) Psychoeducational and Color word test (CWT-G) groups, no significant difference physical training (N=36) (Neuropsychological aging inventory NAI) Emotional status: Only signifi- 3) Cognitive training Primary memory: Memory cant difference was for cognitive alone (N=57): aimed at span (MS-G), Sentence and physical (p<0.05). All other information processing, test (ST) (Neuropsycho- groups, no significant difference attention and memory logical aging inventory NAI) Independent living: Significant 4) Physical training alone Secondary memory: Pic- difference for cognitive and (N=32): involved training of ture test (PT), Figure test physical and psychoed and phys- balance, perceptual (FT), Word list (WL), Word ical (p<0.05 for both). All other and motor coordination, pairs (WP) (Neuropsy- groups, no significant difference flexibility and overall level chological aging inventory of activity NAI) Everyday competence: Sig- Long term memory nificant difference for psychoed 5) Psychoeducational Information: (WAIS-Info) and physical (p<0.05)). All other training alone (N=115): (Wechsler adult intelli- groups, no significant difference aimed at strengthening gence scale, German individual resources to version), Word fluency Health status: Only significant cope with everyday life (Leistungsprüfsystem difference was for cognitive LPS). and physical (p<0.05). All other demands Reasoning: Similarities groups, no significant difference (WAIS-Sim) (Wechsler 6) Control (N=103): No adult intelligence scale, Well-being: Significant differ- treatment control group German version) ence for cognitive and phys- Cognitive Impairment: ical; psychoed and physical Interviewer rating (Sandoz a(p<0.001 for both) and physical clinical assessment geriat- alone (p<0.05). All other groups, rics SCAG). no significant difference 34 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Satoh 2014 Quasi-RCT Japan 65+ Older adults, physically Physical exercise with Follow-up: 1 year Visuo-spatial Kruskal-Wallace (semi-ran- and psychologically music compared to intervention test Funded by domly (range 65- healthy, living physical exercise alone Significant difference in change Yamaha classified) 84) independently Lost to follow-up: ExM scores across the groups Music Foun- Exercise with music 35%; Ex 30%; Control: (p=0.006) with the greatest dation Control Mean: Ex + N=119 randomised (ExM) (N=40): Received 10.2%. (High drop-outs in change in the ExM group 1.7 group music 73.1 music in harmony with ExM and Ex groups due to (1.8) compared to the Ex group Registered recruited (4.6); exer- Gender: 85 female ExM/ exercise. Exercise a major flood in the area) 0.57 (1.5) and control group 0.26 CT separately cise only: 87.5% female Ex/ 80% program included aerobic (1.5) from the 73.3 (4.8); female control movement, clapping, Outcome measurement: ExM and Ex control 73.5 training of lumbar, limbs LM: logical memory, No significant difference in groups (5.6) Ethnicity: Not reported and muscles, stretching, MMSE: Mini-Mental change scores across groups dancing and singing State Examination, for any other test. Within group SES: Mean years RCPM: Raven’s Coloured before and after analyses education: 10.3 (1.6) ExM/ Exercise alone (Ex) Progressive Matrices, showed significant improvement 10.8 (1.8) Ex/ 10.0 (2.1) (N=40): Heard only a standard, TMT: Trail- in visuo-spatial function in the control percussive beat in time to Making Test, VSRAD: ExM group, and significant the exercise (as above). Voxel-based Specific improvements in other batteries For the singing portion Regional analysis system in all three groups only read the lyrics aloud for Alzheimer’s Disease, without the melody WF: word fluency, Kruskal VSRAD: The VSRAD score Wallace visuo-spatial test significantly worsened in the Weekly 1 hour sessions – ExM and Ex groups (both groups 40 sessions over the year had high levels of drop-outs compared to controls) Control (C) (N=39): No intervention but underwent the same testing as ExM and Ex groups Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Shatil 2013 RCT US Range 65- Healthy older adults with Cognitive and/or mild Follow-up: 4 month Older adults in both cognitive 92 MMSE score 24+ aerobic training intervention and follow-up training groups (cognitive Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 35 Funding: training alone or combined not stated Mean: Gender: 65.5 to 71 % Combined cognitive and Lost to follow-up: cognitive training) showed but author 79-81 de- depending on group mild aerobic training Combined: 39.6%; CT only significant improvement in is CogniFit pending on (n=29): Both the cognitive 26.7%; PA 31.1%; Control: cognitive performance on employee group Ethnicity: Not reported training and the physical 31.0% hand-eye coordination, global activity training as visual memory (GVM), working No CTR SES: Those with some described below so these Outcome measurement: memory, long-term memory, college education and participants received twice Validated (in healthy speed of information processing, above: CT/No PA 78.8%; as many training sessions younger adults, visual scanning and naming CT/PA 58.6%; No CT/ as did the cognitive or against major standard No PA 79.3%; No CT/PA physical activity training neuropsychological tests), NB: There is a large amount 90.3% participants multi-domain computerized of data in table 9 – results cognitive evaluation for summarised above Cognitive training only older adults (CogniFit (n=33): CogniFit cognitive neuropsychological training programme for a evaluation: 15 evaluation total of 32h arranged in 48 tasks measuring a 40-minute sessions (three wide range of cognitive times weekly for 16 weeks) abilities such as focused with at least a 1-day and divided attention, interval between sessions inhibition, shifting, planning, working memory, Mild aerobic training and eye-hand coordination only (n=29): Senior exercise video including aerobic warm-up, cardiovascular workout seated and standing aerobic cool-down, strength training, and flexibility training

Control (n=31): Book reading activity 36 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Singh 2014 RCT US 70.1 (6.7) Adults with MCI n=100 Resistance training and/ Follow-up: 6 months PRT benefit or cognitive training intervention, 18 month 74% higher than the Combined Study of Setting: Clinic Cognitive Training (CT) follow-up. Group (p = 0.02) Mental and Intervention plus Sham Lost to follow-up: 6 Resistance Gender: 68% female Exercise (N=24): Com- months: Combined 11.1%; No significant between group Training puter-based exercises PRT 27.2%; CT 8.3%; differences for any other (SMART) Ethnicity: Not reported targeting memory, exec- control 11.1%. 12 months: cognitive outcomes study utive function, attention, Combined 18.5%; PRT SES: Not reported but and speed of information 54.5%; CT 16.7%; control Adverse events results adjusted for edu- processing using the 18.5%. All participants 6 adverse musculoskeletal cation. COGPACK program included in ITT analysis events over 18 months (3 falls Progressive resistance Outcome measurement: during assessments and 3 training (PRT) plus Sham Global cognitive function exacerbations of pre-existing Cognitive Training (N=22): (ADAS-Cog) Executive arthritis symptoms during Pneumatic resistance function, memory, speed strength testing/training, with 1 machines were used for attention tests: unresolved (exacerbation of an high intensity training for Executive function underlying rotator cuff tear) most major muscle groups (WAISIII) and verbal fluen- (chest press, leg press, cy (Controlled Oral Words Note: high drop-out in PRT seated row, standing hip Association Test group so those who remained abduction, knee exten- (COWAT) and Animal may have been those with better sion). Supervised by Naming) general including cognitive exercise physiologists and Memory tests included function who could cope with the physiotherapists auditory Logical Mem- tests Combined CT and PRT ory I (immediate) and II (N=27): Received both the (delayed) subtests of the CT intervention and PRT (WMS-III)25 and the List interventions as above se- Learning subsection of the quentially during the same ADAS-Cog, and visual via session Benton Visual Retention Control (N=27): Received Test-Revised 5th Edition both sham cognitive and (BVRT-R) sham exercise interven- Attention/speed via Sym- tions bol Digit Modalities Test Sham Cognitive Training (SDMT) Consisted of watching 5 Global Domain included all short National tests except List Learning, Geographic videos, as it was already included followed by a set of 15 within ADAS-Cog total questions (3/video) score. Executive Domain This has had minimal im- included WAIS Similarities pact in previous trials. and Matrices, COWAT, Sham Exercise and Animal Naming. Stretching and seated Memory Domain included calisthenics designed Logical Memory I and not to notably increase II, List Learning, and heart rate BVRT-R Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Yokoyama RCT Japan Mean ages: Sedentary elderly people Both groups received Follow-up: 12 week 3MS outcomes

2015 ST group 1-hour exercise training intervention There were significant Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 37 74.2 and N=27 randomised separately, supervised between group differences in PA and 3.4 yrs, by a trained instructor, Lost to follow-up: 7.4% attention, verbal fluency and cognitive DT Setting: University three times a week, for 12 overall. understanding, and similarities training group 74.2 research centre weeks between the intervention and and 4.3 yrs Outcome measurement: control group (p<0.05) Gender: Intervention Intervention (N=12): 91.7% female; control Cognitive-motor dual- Cognitive function The 3MS total score significantly 92.3% female task training, which Executive functions: improved from baseline requires dividing attention Modified Mini-Mental State compared to control (p<0.05) between cognitive tasks (3MS) examination and exercise (resistance and the Trail-Making Test and aerobic exercise). (TMT) Examples are: arithmetic tasks (subtraction of Biomarkers: one digit) or Shiritori, a Plasma amyloid β peptide Japanese word chain (Aβ) 42/40 ratio. game in which one player Other outcomes: has to say a word starting Physiological performance, with the last character anthropometry. of the word given by the previous player, carried out during thigh-raising exercise; or switching direction, walking either forward or backward, according to patterns of whistling

Control (N=13): Single- task training comprising aerobic and resistance exercise only. 38 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 3: Multicomponent – Interventions with cognitive outcomes: combined PA and diet

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Baker Secondary US 67.6/69.3 Population: N=49 older Intervention 1: High– Follow-up: 4 weeks For healthy adults, increased 2012 analysis interven- adults of which N=29 had saturated fat/high– intervention and follow-up hi–PA attenuated the effects RCT tion/control amnestic mild cognitive glycemic index diet of the HIGH diet on CSF Aβ42 Secondary impairment and N=20 were (HIGH): (fat, 45% Lost to follow-up: Not whereas in those with MCI, analysis of healthy controls [saturated fat,_25%]; reported (appears to be increased hi–PA potentiated the Bayer-Cart- carbohydrates, 35%-40% 0%) effects of the LOW diet. Authors er 2011 Setting: Veterans Affairs [glycemic index,_70]; and concluded that normal adults paper (DIET Medical Center clinical protein, 15%-20%) Outcome measurement: who engage in hi–PA are less section). research unit Diet composition vulnerable to the pathological Intervention 2: Diet with measured by self-reported effects of an unhealthy diet, while This Gender: 53.1% female a low-saturated fat/low– 3 day food intake record in MCI, the benefit of a healthy analysis glycemic index diet (LOW) diet on Aβ modulation is greatest examines Ethnicity: Not reported diet (fat, 25%; [saturated PA assessed using 7d when paired with hi–PA. Exercise the modu- fat,_7%]; carbohydrates, questionnaire may thus interact with diet to lating effect SES: Mean educational 55%- 60% [glycemic alter pathological processes that of PA on level across groups: 13-15 index,_55]; and protein, ultimately modify AD risk outcomes years 15%-20%) of diet intervention Both interventions were conducted in a group of participants with MCI and a group of healthy participants Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Komulainen RCT (ran- Finland 55-74 Random population 6 different groups Follow-up: 2 years Cognitive outcomes (2 yrs) 2010 domised by sample of older adults Aerobic exercise (interim data); study No statistically significant blocks) Mean age (N=234): prescribed an ongoing (to 4 years). differences between groups for Dose 65.6 (5.4) N=1410 randomised individualised aerobic changes in immediate memory, Responses to 66.9 (5.2) exercise programme at Lost to follow-up: 8.4% delayed memory, verbal to Exercise depending Setting: Community intensity 55–65% of VO- overall (at 2 years): performance, visual performance

Training on group 2max. Gradually increased control: 7.2%; Aerobic: or MMSE across the study Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 39 (DR’s EX- Gender: Not reported for during first 6 months. Then 7.3%; Resistance 11.0%; groups TRA) Study baseline sample subjects randomised into Diet: 8.5%; Aerobic + Diet: two subgroups with energy 6.8%; Resistance + Diet Other outcomes Ethnicity: Not reported expenditure of 1000–1500 9.4%. There was a mean increase in kcal/week (5 x 60 min/ moderate-to-heavy physical SES: Mean years week) or > 1500 kcal/week Outcome measurement: exercise in the four exercise education 10.7 (3.7) to (5 x 90 min/week) Cognitive function groups, but no change in the 11.4 (4.1) Resistance exercise assessed using the reference and diet only groups (P (N=236): Supervised, indi- Consortium to Establish < 0.001 between groups vidualised strength-training a Registry for Alzheimer’s VO2max remained unchanged in programme.Gradually Disease (CERAD) the groups that included increased during the first neuropsychological aerobic or resistance exercise, 6 months. Then subjects Tests. Sum scores but decreased in the reference randomised into two were calculated for five and the diet groups (P < 0.001 subgroups with the energy cognitive domains; between groups expenditure of 1000–1500 immediate memory (sum kcal/week (two sessions/ from three trials of Word Adverse events week) or > 1500 kcal/week List Memory Test), delayed Five adverse events were (three sessions/week).Plus memory (sum from Word reported (one angina pectoris aerobic exercise 150/180 List Recall Test, number of during a cycle ergometer test, min/week correctly identified one angina pectoris and three Diet (N=236): received words from Word List light-headedness episodes individually tailored coun- Recognition Test and during muscle strength training). selling by nutritionists for delayed Constructional Based on the first 6 months based Praxis Test), verbal medical examinations, physicians on Finnish nutrition recom- performance (sum from concluded that it was unlikely mendations (FNR). Verbal Fluency Test and any of these events were directly After 6 months, subjects Modified Boston Naming caused by the interventions, randomly allocated to fol- Test), visual performance and participants remained in the low the same guidelines or (sum from Constructional study to a special nutrition group Praxis Test and Clock Combined aerobic exer- Drawing Test) and MMSE cise and diet (N=234): Combined resistance ex- Other outcomes: ercise and diet (N=234): Physical exercise, VO2 Control group (N=236): max, adverse events Verbal advice on diet and physical activity 40 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Napoli 2014 RCT US 65+ Obese, older adults (BMI Intervention 1 (N=28): Follow-up: 1 year Modified Mini-Mental State 30+), sedentary with Combined dietary weight Examination: Scores improved Mean age stable body weight in the management programme Lost to follow-up: 13% more in the diet: 1.7 ( 0.4), 70 (SD 4) past year, and on stable and exercise. Participated did not complete study exercise: 2.8 ( 0.4), and diet- medications and with mild- in both weight-manage- but all included in ITT exercise 2.9 ( 0.4) groups than to-moderate frailty ment and exercise pro- analyses in the control group 0.1 (0.4) grammes described below (between-group P = 0.0001– Mean BMI at baseline separately from the other Outcome measurement: 0.04) 36-37 groups Modified Mini-Mental State Examination, Word List Scores in the diet-exercise group Gender: 57-67% female Intervention 2 (N=26): Fluency Test, Trail Making improved more than in the diet (depending on group) Diet/Weight loss only Tests Parts A and B, and group but not more than in the mprescribed a diet that Geriatric Depression Scale exercise group Ethnicity: 81-89% white provided a daily energy (GDS) Short Form. Impact (depending on group) deficit of 500–750 kcal/d. of Weight on Quality of Word Fluency Test: Scores Groups met with dietitians Life-Lite (IWQOL) improved more in the exercise SES: Mean years for food diary review, 4.1 (0.8) and diet-exercise 4.2 education: 15.3-16.9 caloric intake adjustments, (0.7) groups than in the control (depending on group) and behavioral therapy. group 20.8 (0.8); both p = 0.001 The goal was to achieve 10% weight loss for 6 mo Trail Making Test: Scores in and to maintain this weight the diet-exercise group 211.8 for the remaining 6 mo of (1.9) improved more than in the the study control group 20.8 (1.9) (P = 0.001); similar results were found Intervention 3 (N=26): for the Trail Making Test Part B Exercise only. Participants in the exercise group were IWQOL: Scores on the IWQOL counseled on maintain- improved more in the diet 7.6 ing a weight-stable diet. (1.6), exercise 10.1 (1.6), and They participated in a diet-exercise 14.0 (1.4) groups supervised progressive than in the control group 0.3 (1.6) multicomponent exercise (P = 0.0001–0.03); scores in the training programme. diet-exercise group improved Exercise sessions were more than in the diet group but 90 min (w15 min flexibility, not more than in the exercise 30 min aerobic, 30 min group resistance training, and 15 min balance exercises) 3 Weight Changes: As intended, times weekly body weight decreased in the groups that received the dietary Control (N=27): Re- weight loss intervention. Weight ceived general info about decreased similarly in the diet a healthy diet at regular and exercise 28.6 (3.8) and diet visits by staff only groups 29.7 (5.4) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Napoli 2014 RCT US 65+ Obese, older adults (BMI Intervention 1 (N=28): Follow-up: 1 year Modified Mini-Mental State 30+), sedentary with Combined dietary weight Examination: Scores improved Mean age stable body weight in the management programme Lost to follow-up: 13% more in the diet: 1.7 ( 0.4), 70 (SD 4) past year, and on stable and exercise. Participated did not complete study exercise: 2.8 ( 0.4), and diet- medications and with mild- in both weight-manage- but all included in ITT exercise 2.9 ( 0.4) groups than to-moderate frailty ment and exercise pro- analyses in the control group 0.1 (0.4) grammes described below (between-group P = 0.0001– Mean BMI at baseline separately from the other Outcome measurement: 0.04) 36-37 groups Modified Mini-Mental State Examination, Word List Scores in the diet-exercise group Gender: 57-67% female Intervention 2 (N=26): Fluency Test, Trail Making improved more than in the diet (depending on group) Diet/Weight loss only Tests Parts A and B, and group but not more than in the mprescribed a diet that Geriatric Depression Scale exercise group Ethnicity: 81-89% white provided a daily energy (GDS) Short Form. Impact (depending on group) deficit of 500–750 kcal/d. of Weight on Quality of Word Fluency Test: Scores Groups met with dietitians Life-Lite (IWQOL) improved more in the exercise SES: Mean years for food diary review, 4.1 (0.8) and diet-exercise 4.2 education: 15.3-16.9 caloric intake adjustments, (0.7) groups than in the control (depending on group) and behavioral therapy. group 20.8 (0.8); both p = 0.001 The goal was to achieve 10% weight loss for 6 mo Trail Making Test: Scores in and to maintain this weight the diet-exercise group 211.8 for the remaining 6 mo of (1.9) improved more than in the the study control group 20.8 (1.9) (P = 0.001); similar results were found Intervention 3 (N=26): for the Trail Making Test Part B Exercise only. Participants in the exercise group were IWQOL: Scores on the IWQOL counseled on maintain- improved more in the diet 7.6 ing a weight-stable diet. (1.6), exercise 10.1 (1.6), and They participated in a diet-exercise 14.0 (1.4) groups supervised progressive than in the control group 0.3 (1.6) multicomponent exercise (P = 0.0001–0.03); scores in the training programme. diet-exercise group improved Exercise sessions were more than in the diet group but 90 min (w15 min flexibility, not more than in the exercise 30 min aerobic, 30 min group resistance training, and 15 min balance exercises) 3 Weight Changes: As intended, times weekly body weight decreased in the groups that received the dietary Control (N=27): Re- weight loss intervention. Weight ceived general info about decreased similarly in the diet a healthy diet at regular and exercise 28.6 (3.8) and diet visits by staff only groups 29.7 (5.4)

Table 4: Multicomponent – Interventions with cognitive outcomes: combined cognitive and social activation

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Quasi-RCT US Range Older adults living in the Intervention (N=61): Follow-up: 20 week in- Mean participation: 17.3 weeks Morrow (field exper- 55–93 community or retirement Engaged lifestyle pro- tervention (Note: pre- and out of 20 (86.5%) 2007 iment). Mean: villages. gramme. Team-based, post-cognitive tests up Only processing speed showed Senior I: 73.6, N=81 randomised competitive programme of to 9 months apart). Only differential positive change in the Odyssey Assigned range creative problem solving processing speed showed experimental group relative to the programme those from 60–93 Setting: Community pro- involving repeated oppor- differential positive change control group; differential change retirement C: 70.2, gramme. tunities for engagement in the experimental group in divergent thinking reached a Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 41 villages to range with ill-defined problems relative to the control marginal level of significance the exper- 58–85 Gender: Not reported with multiple solutions. group; differential change Predisposition toward cog- imental (The programme was in divergent thinking nitive engagement. Change group Ethnicity: Not reported developed from a well-es- reached a marginal level (from pre-to-post intervention) because of tablished programme for of significance. Patterns of (SD) effort put SES: Mean years edu- children and young adults) change in working Intervention in to build cation: intervention: 16.1 memory, inductive rea- Mindfulness -0.02 relation- (0.4); control: 15.4 (0.7) Control (N=20): Wait-list soning, and visuo-spatial Need for cognition 0.11 ships. So control processing were also in Metamemory in adulthood (MIA) not properly the predicted direction but self-efficacy -0.06 randomised did not reach significance. Perceived activity level -0.11 Furthermore, individuals Processing speed 0.09 in the experimental group Working memory 0.12 showed differential positive Inductive reasoning 0.22 change in both mindful- Visuo-spatial processing 0.33 ness and need for Divergent thinking 0.29 cognition) Control Lost to follow-up: Inter- Mindfulness -0.40 vention: 20% (20 wks); Need for cognition -0.28 control: MIA self-efficacy -0.09 Outcome measurement: Perceived activity level 0.15 Cognitive outcomes Processing speed 0.70 1. Processing speed: Let- Working memory -0.06 ter and Pattern Compari- Inductive reasoning -0.29 son tasks and Finding As Visuo-spatial processing 0.01 and Identical Pictures Divergent thinking 0.11 2. Working memory: Let- Between group difference (p) ter–Number Sequencing Mindfulness 2.03 (0.02) 3. Inductive reasoning: Need for cognition 1.85 (0.03) Letter Sets and Figure MIA self-efficacy 0.18( >0.20) Classification. Perceived activity level -1.80 4. Visual-spatial process- (0.08) ing: Card Rotation and Processing speed 1.76 (0.04) Hidden Working memory 0.85 (0.20) 5. Divergent thinking (flu- Inductive reasoning 1.17 (0.12) ency): Word Association, Visuo-spatial processing 1.03 Ornamentation, (0.15) and Opposites Divergent thinking 1.32 (0.10) 42 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine-Mor- Dispositions reflective of Correlations Between Meas- row 2007 cognitive engagement: ures of Perceived Cognitive (continued) Mindfulness, need for cog- Engagement and Cognitive nition, memory self-effica- Components (p value) cy, activity (not specified?) Mindfulness Processing speed 0.12 (ns) Working memory 0.12 (ns) Inductive reasoning 0.20 (ns) Visuo-spatial processing 0.14 (ns) Divergent thinking 0.26 (p<0.05) Need for cognition Processing speed 0.11 (ns) Working memory 0.20 (ns) Inductive reasoning 0.26 (p<0.05) Visuo-spatial processing 0.30 (p<0.01) Divergent thinking 0.19 (ns) MIA self-efficacy Processing speed 0.21 Working memory 0.26 (p<0.05) Inductive reasoning 0.39 (p<0.01) Visuo-spatial processing 0.28 (p<0.01) Divergent thinking .11

Perceived activity level Processing speed 0.30 (p<0.01) Working memory 0.25 (p<0.05) Inductive reasoning 0.44 (p<0.01) Visuo-spatial processing 0.39 (p<0.01) Divergent thinking 0.39 (p<0.01) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Quasi-RCT US Interven- Older adults from the Intervention (N=107): Follow-up: 20 week Adherence to the programme: Morrow (field exper- 73.0, community and local Engaged lifestyle intervention Attendance at weekly meetings

tion: Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 43 2008 iment) range: retirement communities. programme. Team- was variable (6-20 session; Senior 59–93 based, competitive Lost to follow-up: mean 15.5 (SE4) Odyssey Assigned Control: N=181 randomised program of creative Intervention: 19%; control: programme those from 72.0, range: problem solving involving 15% completed post-test Cognitive tests (one-tailed retirement 58–91 Setting: Community repeated opportunities measures. (Note: n=23 tests) (appears villages to for engagement with dropped out of intervention Reliable for speed to be a the exper- Gender: Not reported ill-defined problems with group during the 20 weeks t(146)=1.81, p=0.036 different imental multiple solutions. (The but returned for post-test study from group Ethnicity: Not reported programme was developed measurement) Inductive reasoning t(146)=1.83, Stine- because of from a well-established p=0.034 Morrow effort put SES: Years of education programme for children Those who returned for 2007 above) in to build (mean): Intervention 16.3 and young adults) follow-up scored higher for Divergent thinking relation- (SE 4), Control 16.0 (SE speed of processing than t(147)=1.88, p=0.031 Note: ships. So 3). Control (N=74): Wait-list those who did not recruitment not properly control Working memory was randomised Outcome measurement: t(146)=1.01, p=0.136 conducted Longevity, resistance to over 2 dementia, and enhanced Visual- spatial processing years/ cognitive flexibility t(144)=.60, p=0.275 seasons so data Overall showed differential reported positive change among those in Stine- who participated in the cognitive Morrow intervention, t(149)=3.11, 2007 may p=0.001 be first wave of data that Control group vs. experimental may also be group reported in Self-efficacy, t(133) = −1.59, this trial Mindfulness, t(128)=0.81 Need for cognition, t(130)=0.68 44 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Quasi-RCT US 60 to 94 Healthy older adults from Engagement programme Follow-up: 16 weeks Adherence to the programmes: Morrow Mean: 72.6 the community and local (Senior Odyssey) intervention but pre-and Engagement participants 2014 Some par- retirement communities (N=188): A team-based post-test were 30 to 32 attended an average of 11.0 out Senior ticipants not competitive program in weeks apart of 16 Odyssey randomly N=461 randomised creative problem solving session (SD 4.8) and Training programme allocated (no explicit instruction) Lost to follow-up: participants completed an versus to meet Setting: Community Engagement: 16 %; average of 12.9 modules out of cognitive project Cognitive training Cognitive training: 12 %; 16 (SD _5.2); this difference in training deadlines Gender: Engagement (N=130): Home-based Wait-list: 12% adherence was significant, F(1, programme intervention: 71% inductive reasoning 317) 11.42, p < 0.001 female; Cognitive training training program Outcome measurement: Key results: intervention 77% female; (instruction and practice Processing Speed: Letter Those in the training condition Waiting list control: 76% explicit). and Pattern Comparison showed selective improvement female tasks and the Finding A’s in inductive reasoning. Training Wait-list control (N=143): task. participants showed more Ethnicity: Not reported No intervention but change than both Engagement participated in testing as a Reasoning: Letter and Wailist participants and SES: Years of education control for re-test effects Sets, Number Series, Engagement and Waitlist (mean): Engagement Letter Series, and Word participants did not change from intervention 15.7 (2.6); Series tasks and the each other Cognitive training Everyday Problem-Solving Those in the engagement intervention 15.2 (2.7); (EPS) task. condition showed selective Control 15.4 (2.5) improvement in divergent Visual-spatial thinking. Neither the Waitlist or processing (VSP): Card Training group had significant Rotation and Hidden re-test effects. Patterns Correlations between baseline characteristics and latent training Divergent Thinking: improvements: Alternate Uses task and Divergent Thinking: the Opposites task. Engagement Statistically significant (p<0.05) Verbal Episodic Memory: differences for age, Montreal was measured using Cognitive Assessment (MoCA), two indicators derived verbal, openness, need for from performance on the cognition, social network index Hopkins Verbal Learning Inductive Reasoning: Training Test; total number of Statistically significant differences words remembered over for MoCA, Verbal. three trials (HVLT-Tot) and Note: Both Engagement and the delayed recall score Training groups showed selective (HVLT-DR) improvement in individual skills that were practiced rather than broad cognitive improvements Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Zacharelli RCT 4 European 65+ Cognitively intact, Intervention: Computer- Follow-up: 3 and 6 Cognitive outcomes 2013 countries: elderly patients with MCl based cognitive training months Across all population groups Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 45 (Italy, and patients with mild and social activation (healthy, MCI and mild AD) there SOCIABLE Greece, Alzheimer’s disease (AD) program Lost to follow-up: 0% were significant differences programme Spain, (but reports outcomes for (analysis conducted on all between intervention and control Norway) all 3 groups together and Training program of 24 those randomised) for: global cognition, verbal each one separately) sessions of 60 minutes of memory, praxis, executive duration (individual or in Outcome measurement: function and language N=348 randomised group), twice per week for GLOBAL COGNITION 12 weeks REASONING Clock Cognitively healthy group: Gender: Not reported Drawing Test MEMORY- significant differences between Included cognitive training VERBAL- Digit Span intervention and control for Ethnicity: Not reported games with 3 levels of forward SHORT measures of verbal memory, difficulty, and a ‘book- MEMORY-VERBAL Rey praxis, executive function and SES: Not reported but of-life’ application (a Auditory Verbal; Learning language had to have a minimum personal diary, created by Test (RAVL)- immediate of 5 years education for the elderly, containing life MEMORY-VERBAL- Rey MCI group: Significant inclusion experiences, memories Auditory Verbal LONG differences between intervention and thoughts, to be shared Learning test (RA VL) and control for measures of with other users) – delayed MEMORY- global cognition, verbal memory VlSUOSP- Rey’s Complex and executive function Control: No intervention figure - LONG recall (delayed treatment) PRAXIS Rey’s Complex Mild AD group (excluded but EXECUTIVE Phonological shown for info): Significant Verbal FUNCTIONS differences between intervention Fluency EXECUTIVE and control for measures of Trail Making Test B global cognition, verbal memory FUNCTIONS EXECUTIVE and executive function Digit Span FUNCTIONS ATTENTION Trail Making Test LANGUAGE Naming Test 46 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables Table 5: Multicomponent – Interventions to improve uptake/maintenance of healthy behaviours: combined PA/diet

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Burke 2013, RCT Australia 60-70 Insufficiently active low to Intervention (N=248): Follow-up: 6 month Physical activity outcomes: Pasalich middle income older adults Low-cost, accessible, intervention plus further After controlling for demographic 2013 Mean: who resided in suburbs home-based physical 6 month follow-up at 12 and other confounding factors, I 65.80 activity and nutrition months (Pasalich 2013) the intervention group demon- The Physi- (2.95); C N= 478 randomised program that incorporated strated increased participation cal Activity 65.75 (3.19) goal-setting and social Loss to follow-up: 22.5% in strength exercise (p < 0.001), and Nutri- Gender: Intervention: support and included walking (p = 0.029) and vigorous tion for Sen- 52.8% male; control: a range of supportive Outcome measurement: activity (p = 0.015), together with iors (PANS) 50.8% male (of resources including written Self-reported; Modified fat significant reduction in mean programme completers) materials, pedometer and and fibre questionnaire for sitting time (p < 0.001) relative resistance band nutritional behaviours, and to controls. All pre-test/post-test Ethnicity: Not reported the International Physical comparisons were significantly Control (N=230): No Activity Questionnaire for better in the intervention group SES: low to medium; 21% intervention: completed PA (p<0.05) for strength exercise, of intervention group had postal questionnaires at walking, moderate activity, vigor- university education and baseline and post-program ous activity, sitting time whereas 17.6% of control none of these comparisons were significant for the control group 12 month follow-up: Sustained improvements were observed for strength exercises. However, mean walking time decreased below baseline levels for both groups. At post-program, the intervention participants had increased time spent in moderate activity (p > .05), which declined at follow-up (p < .05) Nutrition outcomes: Improve- ments in nutritional behaviours for the intervention group were also evident in terms of fat avoidance (p < 0.001), fat intake (p = 0.021) and prevalence of frequent fruit intake (p = 0.008). No sig differences for fibre intake behaviour and frequency of vegetable intake behaviour 12 month follow-up: Improve- ments in fibre intake, fat intake, fat avoidance, body mass index and waist-to-hip ratio were sustained at 12 months Costs: Reported but not cost-effectiveness Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Burke 2013, RCT Australia 60-70 Insufficiently active low to Intervention (N=248): Follow-up: 6 month Physical activity outcomes: Pasalich middle income older adults Low-cost, accessible, intervention plus further After controlling for demographic 2013 Mean: who resided in suburbs home-based physical 6 month follow-up at 12 and other confounding factors, I 65.80 activity and nutrition months (Pasalich 2013) the intervention group demon- The Physi- (2.95); C N= 478 randomised program that incorporated strated increased participation cal Activity 65.75 (3.19) goal-setting and social Loss to follow-up: 22.5% in strength exercise (p < 0.001), and Nutri- Gender: Intervention: support and included walking (p = 0.029) and vigorous tion for Sen- 52.8% male; control: a range of supportive Outcome measurement: activity (p = 0.015), together with iors (PANS) 50.8% male (of resources including written Self-reported; Modified fat significant reduction in mean programme completers) materials, pedometer and and fibre questionnaire for sitting time (p < 0.001) relative resistance band nutritional behaviours, and to controls. All pre-test/post-test Ethnicity: Not reported the International Physical comparisons were significantly Control (N=230): No Activity Questionnaire for better in the intervention group SES: low to medium; 21% intervention: completed PA (p<0.05) for strength exercise, of intervention group had postal questionnaires at walking, moderate activity, vigor- university education and baseline and post-program ous activity, sitting time whereas 17.6% of control none of these comparisons were significant for the control group 12 month follow-up: Sustained improvements were observed for strength exercises. However, mean walking time decreased below baseline levels for both groups. At post-program, the intervention participants had increased time spent in moderate activity (p > .05), which declined at follow-up (p < .05) Nutrition outcomes: Improve- ments in nutritional behaviours for the intervention group were also evident in terms of fat avoidance (p < 0.001), fat intake (p = 0.021) and prevalence of frequent fruit intake (p = 0.008). No sig differences for fibre intake behaviour and frequency of vegetable intake behaviour 12 month follow-up: Improve- ments in fibre intake, fat intake, fat avoidance, body mass index and waist-to-hip ratio were sustained at 12 months Costs: Reported but not cost-effectiveness

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Geller 2012 RCT (very US (Hawaii) Mean 72.2 Older adults recruited from Physical activity OR fruit Follow-up: 1 day Both programs were small, 2 community housing sites and vegetable programme intervention, followed up at implemented efficiently, and n=21) 2 weeks participants in both groups Mean BMI at baseline: Single day PA or fruit and improved their daily physical 26.14 (SD = 6.57) veg intervention based on Lost to follow-up: 29% activity but there were minimal Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 47 a decisional balance sheet from PA intervention and changes in fruit and veg N=21 randomised programme (a promotional 47% from fruit and veg consumption (p values not tool targeting the intervention reported) Gender: 76% female perceived pros and cons of behaviour adoption) Outcome measurement: Costs: The program required Ethnicity: Ethnically Self-reported using minimal staff involvement diverse population validated measures: PA: (≈30–40 min) and included including Japanese 23.8%; International Physical minimal paper costs ($0.05 [one Filipino 19.0%; Caucasian Activity questionnaire decisional balance sheet] × 19.0%; Native American (short); Daily fruit and 21 participants = $1.05). Cost- 4.8%; Native Hawaiian vegetable intake: National effectiveness not reported 4.8%; Hispanic 4.8%; Health and Nutrition others; 23.8% Examination Survey single Note: high fruit and vegetable item instrument consumption at baseline SES: 80% high school graduates

Jaacks RCT (multi- US 25+ (but People at high risk of Intervention: (Lifestyle): Follow-up: 1,5,6,9 years Reports outcomes separately for 2014 centre) reports developing diabetes 16 session core curriculum after randomisation people aged 60+ outcomes over the first 24 weeks, Diabetes separately N=3234 randomised followed by individual Lost to follow-up: >/= For those age 60+, there were Prevention for older counselling (at least 14% (not clear) trends towards increases in Program adults) (2 stages of randomisation monthly) with primary fruit and vegetable intake from and – n not clear for each goals of achieving and Outcome measurement: baseline over 9 years of follow- Diabetes group) maintaining weight loss of Self-reported, dietary up for the lifestyle arm compared Prevention >/= 7% initial body weight intake was assessed to the metformin or placebo Program Gender: 68% female and moderate intensity using a food frequency comparison groups but between Outcomes activity of >/=150 min/wk. questionnaire administered group differences were not Study Ethnicity: 54.7% Participants were advised by trained interviewers statistically significant Caucasian; 19.9% African- to reduce dietary fat intake American; 15.7% Hispanic; to <25% of total calories 5.3% American Indian; 4.4% Asian-American. Comparator: (Metformin)

SES: (yrs education): <13 Comparator: (Placebo) y 25.8%; 13-16 y 48.1%; >/=17 y 26.1%. 48 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Hageman RCT US 40-69 Rural women with Both intervention groups Follow-up: 12 months with Diet outcomes: Web-based and 2014 prehypertension. 18.4% (internet and print) further 12 month follow-up print-mailed groups improved Mean 56.4 were normal weight, received initial training to 24 months more than standard advice group (6.3) 37.7% were overweight about hypertension/ for % daily calories from fat (p and 43.9% were obese pre-hypertension with Lost to follow-up: 6% = 0.018 and p = 0.030) and discussion of strategies at 12 months; 11.4% at saturated fat (p = .049 and p = N=289 randomised to achieve targets for 24 months (no sig diff .013); daily servings of fruit and healthy eating and activity; between groups) vegetables (p =0 .008 and p <0 Gender: 100% female also blood pressure .005); and low fat dairy (p <0 monitors and pedometers. Outcome measurement: .001 and p = 0.002) Ethnicity: 97.9% Both groups also Self-reported, Web However, the standard advice Caucasian received telephone goal version of the 1998 Block group had greater decline setting counselling and Health Habit and History compared to both intervention SES: 40.8% had college newsletters Questionnaire (HHHQ) to groups for kilocalorie intake (p = degree or higher measure diet; Modified .024 and p = .027) and sodium (p Intervention Internet 7-Day Activity Interview = .030 and p = .026) (N=116): At the second instrument for PA Physical activity outcomes: session, those in the There were no significant internet group received differences between the web- training for using the based and print-mailed groups in DASH wellness for change on any of the outcomes women website including Other outcomes: Blood monitoring diet, PA and pressure: The 24-month blood pressure estimated marginal proportions of women achieving normotensive Intervention Print status were 47% for web-based, (N=115): At the second and 39% for both print-mailed session, received and standard advice groups, with instruction in tracking no group differences (p = .11 and of their blood pressure, p = .09, respectively) eating, and activity using Web-based and print-mailed paper logs groups improved more than standard advice group for waist Control (standard circumference (p = .017 and p = advice) N=48: .016, respectively) Greater improvements were observed in web-based versus standard advice groups in systolic blood pressure (p = 0.048) and estimated VO2 max (p = 0.037) There were no significant differences between the web- based and print-mailed groups in change on any of the outcomes Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Kelley RCT UK 65+ Outpatients from hospital Intervention (N=125): Follow-up: 2 weeks Diet outcomes: 50.4% set goals 2004 clinics Healthy living booklet to eat more healthily (e.g. ‘‘to eat Mean age designed to promote Lost to follow-up: 17% five portions of fruit and vegeta- of 81.7 N=252 randomised healthy eating and bles a day’’), and 67% of those years(5.6) physical activity amongst Outcome measurement: who set goals reported 100% Gender: 69% female older adults Based on Self reported changes in success in acting on them behavioural theory with diet and PA using single

Ethnicity: Not reported goal-setting prompts item questionnaire Physical activity outcomes: Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 49 Only 34% of intervention partici- Control (N=127): Patient pants set an activity goal (e.g. ‘‘a satisfaction questionnaire five minute walk everyday’’), and only 51% reported 100% success in achieving them

Kimura Cluster Japan 65-90 Community dwelling older Community-based social Follow-up: 3 months The mean attendance rate for the 2013 RCT adults health intervention to intervention intervention classes was 68.1% Mean age: improve dietary habits and (range 41-95%) I 74.3 ± 5.9/ Gender: 79.8% female promote physical activity Lost to follow-up: control C 74.3 ± among older adults 5.4%; intervention 0% Diet outcomes: Compared to 5.0 Ethnicity: Not reported baseline, post-intervention food Delivered in community Outcome measurement: intake frequency for 6 of 10 SES: Not reported centres Self-reported; Food intake food groups (meat, fish/shell- was assessed using a fish, eggs, potatoes, fruits, and Intervention group (3 questionnaire on food seaweed), FFS, and DVS were community centers; n = intake frequency covering significantly increased in the 57): Participated in social 1 week for changes in intervention group, and interac- health program “Sumida food intake frequency, tion effects of FFS and DVS were TAKE10!”, an educational food frequency score seen between the two groups. program incorporating the (FFS), dietary variety No significant differences were “TAKE10! for Older Adults” score (DVS); frequency of observed between baseline and program (eating regularly walking and exercise using post-intervention in the control from 10 food groups and a questionnaire group taking 10 min of physical activity at least 2–3 times Physical activity outcomes: per day), once every 2 Frequency of walking and ex- weeks for 3 months ercise remained unchanged in both groups, and no significant Control group (3 difference in improvement rate community centers; was seen between the groups n=35): No intervention but subsequently received Other outcomes: Self-rated same programme as a health was significantly increased crossover intervention in the intervention group. Appe- group tite and TMIG Index of Compe- tence score were unchanged in both groups 50 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Nahm 2010 RCT (pilot) US 55+ Community-dwelling older Intervention: (TSW) for Outcomes: included (1) Both groups showed significant adults with access to the older Social Cognitive knowledge (hip fractures improvement in most outcomes Mean age: internet and could use Theory (SCT)-based, and osteoporosis), (2) 69.3 (7.7) the internet and email Structured Hip Fracture self-efficacy and outcome For calcium intake, only the TSW years independently Prevention Website adults. expectations, and (3) group showed improvement. Included learning modules calcium intake and None of the group and time N=245 randomised about Osteoporosis, Falls exercise interactions were significant and Hip Fractures, Dietary/ Gender: 78.4% female Supplementary Calcium Follow-up: 2 weeks Intake, and Exercise. The intervention and follow-up Ethnicity: 91.0% white modules included text at 3 months material, video, audio, SES: 85.5% had some and self-assessment Lost to follow-up: college or higher education quizzes using in an older adult-friendly format; and Outcome measurement: relevant discussion boards Self-reported; Dietary calcium intake was Comparator: assessed using a 22-item Conventional website. measure derived from the Participants also Block-National Cancer completed 4 learning Institute Health Habits modules without the and History Questionnaire discussion boards. Content (HHHQ) that assesses included hyperlinks to frequency and portion relevant health websites size; Exercise behaviour was measured using the exercise dimension (6 items) of the Yale Physical Activity Survey (YPAS) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Oh 2014 RCT S. Korea 51–83 Post-menopausal older Intervention (N=21): Follow-up: 3 month Compared with the control group,

mean women without cognitive Lifestyle intervention to intervention the intervention group showed Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 51 66.2 years impairment recruited from improve bone health improvement in diet and exercise (8.2) a healthcare centre in a Lost to follow-up: 0% for after 12 weeks rural village Trained community nurse both groups practitioner provided the Diet outcomes: Intake of dairy, 75% (n¼31) of the 3-month intervention in Outcome measurement: calcium rich fish, nuts, and participants had a total of 24 sessions, Self-reported food diaries vitamin D-rich foods such as fish osteoporosis or osteopenia held 2 times/week at for diet; physical activity – oil and vegetables increased the healthcare center. single item question about significantly in the intervention Mean BMI at baseline 23.8 The intervention included number of days/week group individualized health participated in activity Gender: 100% female monitoring; group health The mean level of serum 25-OH- education; group exercise, Vit.D showed significantly greater Ethnicity: Not reported calcium–vitamin D increases in the intervention supplementation group than in the control group SES: Not reported at the study’s completion. Serum Control (N=20): levels of calcium changed little in Received an educational both groups booklet and were instructed to maintain their Physical activity outcomes: usual lifestyle behaviour Regular weekly exercise was more frequent in the intervention group than at baseline or among control group participants 52 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Plawecki RCT US Mean age Older, community-dwelling Intervention (N=35) : Follow-up: 8 weeks Calcium intake: At 8 weeks, 2013 was 65.5 people. 67% had no Community intervention there was no significant (9.6 years) history of osteoporosis, for improving bone health Lost to follow-up: difference between intervention almost all had previously behaviours targeting Intervention 11.4%; Control and control groups (p=0.70) in had a bone scan (81%) those older than 50 years. 8.8% total calcium intake using the 24- An 8-week, bone-health hour recall or CFFFQ (p=0.072) Gender: 83% female community program Outcome measurement: addressed risks and Calcium-Focused Food There was a significant increase Ethnicity: 90% white lifestyle changes within Frequency Questionnaire in total calcium from week 1 to the Health Belief Model (CFFFQ) - includes both week 8 (p=0.005 for 24-hour SES: 77% had some and Theory of Reasoned natural and fortified recall; p=0.027 for CFFFQ), with college eduction Action sources for usual calcium a significant increase for calcium frameworks intake. 24 hour recalls for from the fruit group (p=0.005) for daily intake the 24-hour recall and for calcium Control (N=34): Delayed from grains for the CFFFQ treatment control group Activity: An activity (p=0.042) log was used to record, number of steps using Those meeting or exceeding pedometers, the number the Recommended Dietary of heel drops (to measure Allowance (RDA) of calcium as ground force activity), time measured by 24-hour recall were: devoted to balance activity 26% at week 1, 44% at week 4, (including minutes of and 35% at week 8 (p=0.039) resistance band use) Vitamin D intake: While dietary vitamin D significantly increased (p<0.015) at each time point during the intervention for the treatment group, no significant difference was found comparing control to treatment at week 8 Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Resnick Before and US 65+ Older adults on Intervention: Motivational, Follow-up: 12 week There were significant decreases 2009 after study antihypertensive or lipid- educational and exercise intervention, follow-up to 4 in systolic (P = .02) and diastolic (feasibility Mean (SD) lowering medications/ sessions with techniques months blood pressure (P = .01) and PRAISEDD study) age was sedentary behaviour known to strengthen self- a nonsignificant trend toward intervention 76.4 (7.6) efficacy and outcome for Lost to follow-up: 9% improvement in cholesterol Gender: 64% female CVD prevention behaviors intake (P = .09). There were (People related to exercise, diet Outcome measurement: no changes in time spent in Reducing Ethnicity: 86% African and medication PA: Yale PA survey; moderate-level physical activity, Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 53 Risk and American cholesterol/sodium sodium intake, medication Improving 60-minute intervention intake: Block Brief Food adherence, or self-efficacy and Strength SES: Low income older sessions were held 3 Questionnaire outcome expectations across all through adults; Mean 11.0 (4.7) times per week for 12 3 behaviors Exercise, years education weeks. An interdisciplinary Diet and team implemented the Drug Adher- intervention. During the ence) first week,4 advanced practice nurses and a pharmacist were involved in delivering education. Further weekly sessions were implemented by a lay exercise trainer

Intervention conducted at a senior housing site

Silva-Smith RCT US 60+ Overweight/obese and Intervention (N=32): Follow-up: 8 week Statistically significant differences 2013 sedentary older adults Theory-based physical intervention in the 7-day physical activity self- Mean: I activity and healthy report were reported at post-test Promoting 71.3 (7.43) N=69 randomised eating intervention aimed Lost to follow-up: 10% in the intervention group (p<0.10) Older Adult- C 67.76 at reducing stroke risk intervention group; 9% but not for pedometer steps. Wellness (6.66) Gender: 81.2% female in factors. 8-week group control The dietary measures were not (POW) intervention group; 83.8% motivational intervention statistically significant at post- in control group Outcome measurement: test; however, the intervention Control (N=37): Biweekly Physical activity was group increased the quantity of Ethnicity: Intervention: newsletters by mail measured using a 7-day vegetable servings 75% white; 12.5% African pedometer and a 7-day PA American; control: 64.9% self-report. Diet by 5-pass white; 18.9% African 24-hour diet recall method American by a trained researcher

SES: Mean years education: Intervention 14.0 (2.7); control: 14.4 (3.0) 54 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Valente Secondary US 60-75 Overweight and obese Dietary education only: Follow-up: 10 weeks Diet outcomes: Significantly 2011 analysis of older adults. 77.7% were (DE) N=12 higher DASH diet index scores RCT mean 66.6 obese Lost to follow-up: None (p<0.01) post-intervention in (4.3) Dietary education plus reported DERT group compared to DE 25% were taking lipid- resistance training (DERT) lowering and 14.8% were N=15: Outcome measurement: Note: However, DERT group taking antihypertensive Resistance training Self-reported FFQ had higher energy intake post- medication and dietary education intervention and outcomes do not intervention seem to have been adjusted for Gender: 59.2% female energy intake

Ethnicity: 100% white Other outcomes: The DERT subjects had significantly SES: Not reported better triacylglycerol and apoB concentrations and DASH Diet Index scores than the DE subjects post-intervention. Improvements were seen within the DE group in energy intake, fat-free mass, and systolic blood pressure and within the DERT group in body weight, percentage of body fat, BMI, diastolic blood pressure, and oxidized low- density lipoprotein (all P , 0.05)

Conclusion: DERT was more effective than DE alone in improving DASH Diet Index scores Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Vrdoljak RCT (clus- Croatia 65+ (Mean Setting: General practice Lifestyle intervention that Follow-up: 18 months Outcomes reported separately 2014 ter) (multi- 72.3 (SD targeted a range of health for each health behaviour Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 55 centre, con- 5.2) Population: Croatian behaviours: PA, smoking, Lost to follow-up: Of ducted in citizens aged 65+ years alcohol, diet those completing the The only significant difference 59 general who visited their GP baseline survey for between the intervention and practices) for any reason (those Intervention (N=371): completed the follow-up control group at the end of with life expectancy < 6 Intensified intervention survey the study was for diet. More months, severe dementia, delivered by GPs. participants in the intervention severe mental illness, Intervention participants Outcome measurement: group reported eating the communication disability were counselled and Self-reported, Mediterranean diet, comparing to excluded) given a tailored life plan questionnaire the control group (x2 = 5.81, df = for adopting healthier 1, P = 0.02) Gender: 61% female behaviour. Each patient received educational No significant differences were Ethnicity: Not reported leaflets for their detected found between the intervention CV risk factors and a and control groups for physical SES: 61.7% below specific appointment was activity (chi-squared = 0.84, df average income; 21.0% given for the next follow-up = 1, p = 0.36), smoking (chi- average income and visit squared = 0.85, df = 2, p = 0.65), 17.3% above average alcohol consumption (chi- income Comparator (N=367): squared = 0.73, df = 1, p = 0.394) Usual care: GPs were not instructed to give any specific intervention 56 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Walker RCT (ran- US 50-69 Older rural women Intervention (N=115): Follow-up: 12 month Both groups significantly 2009, domised by Tailored PA and eating intervention with additional increased stretching and Walker site) Gender: 100% female newsletters based on the 12 month follow-up strengthening exercise and fruit 2010 Health Promotion model and vegetable servings and Ethnicity: Intervention: Lost to follow-up: decreased % calories from fat, Wellness 99.1% white non-Hispanic; Control (N-110): Generic Intervention 7.8%; control while only the tailored group for Women 0.9% Hispanic; Control: newsletter intervention 0.9% (though all at increased ≥ moderate intensity study 89.1% white non-Hispanic; baseline included in ITT activity and decreased % calories 6.4% Hispanic; 3.6% analysis) from saturated fat from baseline Native American or to 6 months Alaskan Outcome measurement: Self-reported, PA (7 day SES: Intervention: 25.2% activity record): Time college graduate or higher; engaged daily in moderate 45.2% some college; or greater intensity Control: 44.5% college activity and associated graduate or higher; 33.6% energy expenditure were some college measured by the Modified 7-Day Activity Recall; Diet self-reported daily servings of fruits, vegetables, and whole grain products and daily intake of dietary fat (% calories from total fat and saturated fat) were measured by the web- based version of the 1998 Block Health Habits and History Questionnaire (HHHQ) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Werkman RCT (clus- Netherlands Mean age Recent retirees Intervention (N=209): Follow-up: 12 months Diet and PA outcomes: 2010 ter) 59.5 years Multifaceted computer intervention with further 12 Physical activity and dietary Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 57 N=413 cluster randomised tailored one-year energy months follow-up behaviours improved in both the balance programme intervention and control group Gender: Intervention: Lost to follow-up: during the intervention period. 83.2% male; Control: Control (N=204): Intervention: 15.8%; Although these behaviours 87.2% male Received newsletters with Control: 16.2% changed more favourably in general information about the intervention group, these Ethnicity: Not reported the study only Outcome measurement: between-group-differences were Changes in the diet were not statistically significant SES: % reported as assessed with a validated, ‘low educational level’: semiquantitative food Other outcomes: Waist Intervention: 25%; Control frequency questionnaire circumference, body weight 23% (FFQ). PA assessed with and blood pressure decreased the Dutch version of the significantly in men of the PA scale for the elderly intervention and control group, (PASE) but no significant between- group-differences were observed at 12 or at 24-months follow- up. A significant effect of the programme was only observed on waist circumference (-1.56 cm (95%CI: -2.91 to -0.21)) at 12 month follow up among men with low education (n = 85) 58 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 6: Multicomponent – Quality assessment for interventions reporting on uptake/maintenance of behaviours

Selection bias Performance bias Attrition bias Detection bias Summary Risk of A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Bias all all all all Andersen- Un- Un- Yes Un- Yes Un- Un- Un- Yes No No Un- Un- Yes Yes Un- Un- Un- Unclear Hanley clear clear clear clear clear clear clear clear clear clear clear 2012 Anstey Yes Yes Yes Low Yes Un- Yes Low Yes Yes N/A Low Un- Un- Yes Yes Yes Low Low 2015 clear clear clear Barnes Yes Yes Yes Low Yes Yes Yes Low Yes Yes Yes Low Un- Yes Yes Yes Yes Low Low 2013 clear Carlson Un- Un- Un- Un- No Un- Un- Un- Un- Un- No Un- Un- Yes Yes Un- Un- Un- Unclear 2008 clear clear clear clear clear clear clear clear clear clear clear clear clear clear Clare 2015 Yes Un- Yes Un- Yes Un- No Un- Yes Yes Yes Low Un- Yes Yes Yes Yes Low Unclear clear clear clear clear clear Clark 2012 Yes Un- Yes Un- No Blind Blind Un- Yes Yes Yes Low Un- Yes Yes Yes Yes Low Unclear clear clear to de- to de- clear clear sign sign Cohen- Un- Un- Un- Un- Yes No No Un- Yes Un- Un- Un- Un- Yes Yes Yes Yes Low Unclear Mansfield clear clear clear clear clear clear clear clear clear 2015 Dannhaus- No No Yes High No No Un- High Yes N/A Un- Un- Un- Yes Yes Un- Un- Un- High er 2014 clear clear clear clear clear clear clear Diamond Un- Yes Yes Un- No No Un- Un- Yes Un- Un- Un- Un- Yes Yes Yes Yes Low Unclear 2015 clear clear clear clear clear clear clear clear Fabre 2006 Un- Un- Yes Un- No Un- Un- Un- Un- Un- Un- Un- Un- Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear clear clear Gonzalez- No N/A N/A High N/A No No High N/A N/A N/A N/A Un- Yes Yes Un- Un- Un- High Palau 2014 clear clear clear clear Hars 2013 Yes Yes Yes Low No No Un- Un- Yes Yes Yes Low Yes Yes Yes Yes Yes Low Unclear clear clear Hughes Yes Un- Yes Un- Yes Un- Un- Un- Yes Yes Yes Low Yes Yes Yes Un- Un- Un- Unclear 2014 clear clear clear clear clear clear clear clear Kamegaya Yes Un- Yes Un- No Un- Un- Un- Yes Un- Un- Un- Un- Yes Yes Un- Un- Un- Unclear 2014 clear clear clear clear clear clear clear clear clear clear clear clear Komulainen Yes Yes Yes Low Yes No Un- Un- Yes Yes Yes Low Yes Yes Yes Yes Yes Low Low 2010 clear clear Selection bias Performance bias Attrition bias Detection bias Summary Risk of A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Bias all all all all

Legault Un- Un- Yes Un- Yes Un- Un- Un- Yes Yes Un- Un- Un- Yes Yes Un- Un- Un- Unclear Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 59 2011 clear clear clear clear clear clear clear clear clear clear clear clear Linde 2014 Yes Yes Yes Low No No No High Yes No Yes Un- Yes Yes Yes Yes Yes Low Unclear clear Maillot 2012 Un- Un- Yes Un- No No Un- Un- Yes Yes Yes Low Un- Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear McDaniel Un- Un- Un- Un- Yes Un- Un- Yes Yes Yes Low Yes No No Yes Yes Un- Unclear 2014 clear clear clear clear clear clear clear Mendoza- Un- Un- Yes Un- No Un- Un- Un- Yes Yes Yes Low Un- Yes Yes Un- Un- Un- Unclear Ruvalcaba clear clear clear clear clear clear clear clear clear clear 2015 Miller 2012 Un- Un- Un- Un- No Un- Un- Un- Yes Yes Yes Low Un- Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear Napoli 2014 Yes Un- Yes Un- Un- Un- Un- Un- Yes Yes Yes Low Yes Yes Yes Yes Un- Low Unclear clear clear clear clear clear clear clear Ngandu Yes Un- Yes Un- Yes Yes Un- Un- Yes Yes Yes Low Yes Yes Yes Yes Yes Low Unclear 2015 clear clear clear clear Oswald No No Yes High N/A Un- Un- Un- Yes Yes Yes Low Yes Yes Yes Un- Un- Un- High 2006 clear clear clear clear clear clear Pitkala 2011 Yes Yes Yes Low No Un- Un- Un- Yes Yes No Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear Satoh 2014 No Un- Yes High Yes Un- Un- Un- Yes Yes Yes Low Yes Yes Yes Un- Un- Un- High clear clear clear clear clear clear clear Shah 2014 No No Yes High No No No High Yes No No Un- Un- Yes Yes No No Un- High clear clear clear Shatil 2013 Un- Un- Yes Un- No Un- Un- Un- Yes Yes Yes Low Un- Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear Singh 2014 Yes Yes Yes Low Yes Un- Un- Un- Yes No Un- Un- Yes Yes Yes Yes Yes Low Unclear clear clear clear clear clear Small 2006 Un- Un- Yes Un- No Un- Un- Un- Yes Un- Un- Un- Un- Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear clear Smith 2010 Un- Un- Un- Un- No Un- Un- Un- Yes Un- Un- Un- Un- Yes Yes Yes Yes Low Unclear clear clear clear clear clear clear clear clear clear clear clear 60 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Selection bias Performance bias Attrition bias Detection bias Summary Risk of A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Bias all all all all Stine- Un- Un- Yes Un- No Un- Un- Un- Un- Un- Un- Un- Un- Yes Yes Un- Un- Un- Unclear Morrow clear clear clear clear clear clear clear clear clear clear clear clear clear clear 2007 Stine- Un- Un- Un- Un- No Un- Un- Un- Yes No No Un- Un- Yes Yes Un- Un- Un- Unclear Morrow clear clear clear clear clear clear clear clear clear clear clear clear 2008 Stine- Un- Un- Yes Un- No Un- Un- Un- Yes Yes Yes Low Un- Yes Yes Un- Un- Un- Unclear Morrow clear clear clear clear clear clear clear clear clear clear 2014 Tesky 2011 Un- Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear Wiegand Yes Un- Yes Un- No Un- Un- Un- Yes Yes Yes Low Un- No No Un- Un- Un- Unclear 2013 clear clear clear clear clear clear clear clear clear Yokoyama Yes Yes Yes Low Yes Yes No Low Yes Yes Yes Low Un- Yes Yes No No Un- Unclear 2015 clear clear Zacharelli Un- Un- Un- Un- No Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear 2013 clear clear clear clear clear clear clear clear clear clear clear clear clear Table 7: Multicomponent – Quality assessment for interventions reporting on cognitive outcomes

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias

RCTs Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 61 Burke 2013, Un- Un- Yes Un- No Un- Un- Un- Yes Yes Yes Low Yes Yes Un- Un- Un- Un- Unclear Pasalich clear clear clear clear clear clear clear clear clear clear 2013 Geller 2012 Un- Un- Un- Un- Yes Un- Un- Un- Yes No No Un- Un- Un- Un- Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear clear clear Hageman Yes Yes Yes Low Yes Un- Un- Low Yes Yes Yes Low Yes Un- Yes Yes Yes Low Low 2014 clear clear clear Harari 2008 Yes Yes Yes Low No Un- Un- Un- Yes Yes Yes Low Yes Un- Yes Un- Un- Uncle- Unclear clear clear clear clear clear clear ra Jaacks Un- Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Un- Yes Un- Un- Un- Unclear 2014 clear clear clear clear clear clear clear clear clear clear clear clear clear Kelley 2004 Yes Un- Yes Un- Yes Un- Un- Un- Yes No Yes Un- No No No Un- Un- High Unclear clear clear clear clear clear clear clear clear Kimura Un- Yes Yes Un- No No No Un- Yes Yes Yes Low Yes Un- Un- No No Un- Unclear 2013 clear clear clear clear clear clear Nahm 2010 Yes Yes Yes Low Yes Yes Un- Low Yes Yes Yes Low Yes Un- Yes Un- Un- Un- Unclear clear clear clear clear clear Oh 2014 Yes Yes Yes Low No Un- Un- Un- Yes Yes Yes Low Yes Un- Yes Un- Un- Un- Unclear clear clear clear clear (diet), clear clear clear no (PA) Plawecki Un- Un- Un- Un- No No Un- Un- Yes Un- Un- Un- Yes Un- Yes Un- Un- Un- Unclear 2013 clear clear clear clear clear clear clear clear clear clear clear clear clear (diet) Yes (PA) Silva-Smith Yes Yes Yes Low Yes Un- Un- Un- Yes Yes Yes Low Yes Un- Yes Yes Yes Low Low 2013 clear clear clear clear Valente Un- Un- Yes Un- Yes Un- Un- Un- Yes Yes Yes Low Yes Un- Yes Un- Un- Un- Unclear 2011 clear clear clear clear clear clear clear clear clear clear Vrdoljak Yes Yes Yes Low Yes Yes Yes Low Yes Yes Yes Low Yes Un- Un- Yes Yes Un- Unclear 2014 clear clear clear 62 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Walker Un- Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Un- Yes Un- Un- Un- Unclear 2009/10, clear clear clear clear clear clear clear clear clear clear clear clear clear Yates 2012 Werkman Yes Un- Yes Un- No No No High Yes Yes Yes Low Yes Un- Yes No No Un- Unclear 2010 clear clear clear clear Non-randomised studies Resnick No No N/A High N/A Un- Un- Un- N/A No No Un- Yes Un- Yes Un- Un- Un- High 2009 clear clear clear clear clear clear clear clear Table 8: Multicomponent – Quality assessment for studies reporting barriers and facilitators

Theoretical Study Data Validity Analysis Ethics Overall Approach Design Collec- tion

1.1 1.2 2.1 3.1 4.1 4.2 5.1 5.2 5.3 5.4 6.1 6.2 Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 63 Clare 2015 Appropri- Clear Defensi- Appropri- Clear Not sure Not Reliable Not sure Not sure Yes Not sure Unclear ate ble ate report- – results NB: Mixed meth- ed fully to be ods. This sections in this fully deals with qualita- paper report- tive methods only ed in a separate paper

Jackson 2008 Appropri- Not sure Not sure Not sure Not sure Not sure Not re- Not re- Not re- Not re- Yes Not sure Unclear ate – used ported ported ported ported – little validated about question- the qual- naires itative analysis reported in this paper 64 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 9: Alcohol – Interventions to promote healthy drinking behaviours (RCTs)

Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Arean 2008 Multisite US 65+ Older primary care Intervention: Follow-up: 3 and Drinking declined in both Unclear Oslin 2006 RCT in ten patients with at-risk Integrated care: 6 months (post- intervention groups between primary Mean age drinking (assessed (N=280 randomised) randomisation). baseline and 6 months. However, PRISM-E care clinics 72.0 (SD by questionnaire), Mental health services there were no statistically study 5.3) identified from people integrated into primary Lost to follow-up (6 significant between-group who had a primary care care clinic (services months): Integrated differences in drinking or binge appointment for any on site including care: 18.1%; Enhanced episodes at 6 months reason psychotherapy, case referral: 14.6% management and Number of drinks per week N=560 randomised a brief behavioural Outcome declined from 18.1 (SD 10.6) at alcohol intervention measurement: Self- baseline in integrated care to Gender: 92% male based on harm reported questionnaire 11.8 (SD 11.8) at 6 months and reduction and MI) 17.5 (SD 11.3) at baseline in Ethnicity: 70% white; enhanced referral to 11.4 (SD 23.8% black; 3.5% Comparator: 10.7) at 6 months; p for between- Hispanic or Latino; Enhanced referral: group treatment difference 0.913 0.7% Asian (N=280 randomised) referred by primary care to a nurse or medical social worker model that linked patients to community-based services in a separate location (medication management, psychotherapy and alcoholics anonymous model treatment for heavy drinking) Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Cucciare RCT US Mean age: US Military veterans Intervention: (N=78) Follow-up: 3 and 6 Both interventions led to a Unclear

2013 59 (SD15) who screened positive Brief web-delivered months significant reduction in alcohol Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 65 for alcohol misuse intervention using quantity and frequency and (AUDIT-C) at a routine normative feedback Loss to follow-up: alcohol-related problems at primary care visit. comparing the Not reported, appears 6-month follow-up. No differential participant’s alcohol use to be zero (outcome treatment effects on outcomes N=167 randomised with age and gender data based on all those were observed between the matched peers plus randomised) two treatment groups. BAIs Gender: 88% male treatment as usual using normative feedback may (TAU+BAI) Outcome not have any additional benefit Ethnicity: 58% measurement: beyond usual treatment white; 10% black; 7% Comparator: (N=89) Self-reported (30 day Hispanic; 4% Asian Treatment as usual Timeline Follow Back Pacific Islander; 1% (TAU) (TLFB)) Native American

Ettner 2014 RCT US 60+ At-risk older drinkers Intervention: (N=546) Follow-up: 3, 6, and 12 At risk drinking: At 6 and 12 Unclear (cluster living in the community Personalised reports, months after baseline months, there were significantly Project randomised Mean age (identified by telephone educational materials, greater reductions in at-risk SHARE trial of 31 71 (SD 7.3 and a baseline mailed drinking diaries, Loss to follow-up (12 drinking in the intervention – Senior prima- years) survey) physician advice months): Intervention: groups compared to control: 6 Health and ry care during office visits, and 19.6% Control: 4.7% months: (60% vs. 72%; p

Costs: Average variable costs per patient were $31 for screening and $79 for intervention 66 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Fink 2005 RCT (ran- US 65+ mean Older primary care 3 arms of trial - Follow-up: 12 months Drinks/week: Unclear domised 76.6 (SD patients who reported at Combined report versus In the combined report group by site – 3 6.2 years) least one alcoholic drink Patient report versus Loss to follow-up: compared to usual care, alcohol sites ran- in the last 3 months Usual care Usual care 6.7%; consumption decreased by 1.14 domised to Patient report 6.1%; drinks per week (p<0.05). There 3 groups) At baseline, 21% were Combined report Combined report 6.6% was no statistically significant harmful drinkers, and Intervention: N=6 difference (p<0.05) between the 26% were hazardous physicians and N=212 Outcome patient report intervention and drinkers patients received measurement: usual care in their changes in reports of patients’ Computerized drinking from baseline to follow- N=711 randomised, drinking classifications Alcohol-Related up N=665 completed. and patients also Problems Survey received education (CARPS) Lower-risk drinking: Gender: 54% female The patient report and combined (completers) Patient report report interventions were each intervention: N=245 associated with greater odds Ethnicity: 88% non- patients only received of lower-risk drinking at follow- Hispanic white; 4% reports and education, up than usual care (OR 1.59 Hispanic; 7% Asian but their five physicians and 1.23, respectively, P<0.05 American; 1% African did not receive reports for each). Similar results were American (completers) obtained when intention to treat Comparator: Usual analysis was used care N=238: Neither the 12 participating physicians nor their patients received reports, and the patients did not receive any education during the study Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Gordon RCT (N=45 US 65+ (76% Setting: Primary care Both intervention Follow-up: 1 year All 3 groups decreased the High

2003 ran- were 66-75 groups described as number of drinks per month, Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 67 domised) years; 23% Population: Hazardous brief interventions. Loss to follow up: increased the number of days (post hoc were 76-85 alcohol drinking elderly 0% (from baseline abstained, and reduced the analysis From 12 years and N=45; (analysis also ME Intervention: assessment to follow- number of days per month they by age of primary 2% over 85 compared with younger Motivational up) drank. However, there were no the Early care offices years) hazardous drinkers, Enhancement (N=18) significant differences between Lifestyle 13,438 pa- N=256) BA Intervention: Brief Outcome the intervention groups and Modifica- tients were advice (N=12) measurement: Self- standard care tion (ELM) screened of Gender: 87% male Comparator: Standard reported, Timeline programme, whom 2702 (elderly); 67% male Care (N=12) Follow Back (TLFB) The ME group decreased drinks/ Maisto were elderly (non-elderly) questionnaire month from 60.7 at baseline to 2001) (180 were The brief interventions 29.6 at 6 months to 34.4 at 12 hazardous Ethnicity: 69% white, (ME and BA) were months. The BA group decreased drinkers) 29% African-American, delivered by extensively drinks/month from 126.9 at 2% other trained researchers. baseline to 66.9 at 6 months to The ME intervention 58.6 at 12 months. The standard was more intense, care group decreased drinks/ longer and more month from 61.9 at baseline to frequent than the BA. 50.1 at 6 months to 48.3 at 12 Standard care may months. have included the usual range of services There were no significant in primary care or no differences between the elderly discussion of alcohol group and a younger group for all use problems 3 interventions 68 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Hansen RCT Denmark Mean age Heavy drinkers Intervention: N=391. Follow-up: 6 (N=670) Consumption among women Unclear 2012 60 (men); (identified from a Brief Motivational and 12 months (N=616) decreased from a mean baseline 59 (women) population survey) Interview (mean level of 20.6 to 15.0 drinks/week duration 11 mins) Loss to follow-up: (6 for the control group (95% CI: Range 48- 12,364 participants followed up by months): Intervention 13.5–16.5) and 14.1 drinks/week 65 in a Danish health telephone booster (5 12.5%; control 13.9% for the BMI (95% CI: 12.9–15.2) examination survey mins), plus leaflets and (12 months) after 6 months. Consumption identified. 1026 heavy information sheet about Intervention 19.2%; among men decreased from a drinkers of whom N= local alcohol treatment control 21.3% mean baseline level of 31.8 to 772 were included and delivered by BMI trained 24.0 drinks/week for the control randomised research team Outcome group and 23.1 drinks/wk for the measurement: Self- BMI (95% CI: 21.1–25.1) after 6 Gender: 49% men/51% Comparator: N=381. reported, internet-based months. (12 month data shown women in BMI group; Control group received questionnaire graphically in the paper but was 54% men/46% women the same leaflets and similar to 6 month outcomes) in control group information sheet as the intervention group The intervention effect of the BMI Ethnicity: Not reported was −1.0 drinks/week, but there were no significant differences between groups (intention to treat analysis) (95% CI: −2.15 to 0.23)

There were also no significant differences by gender Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Kuerbis RCT (Pilot) US 50+ Setting: Primary care Intervention: (N=44) Follow-up: 3 months Drinks per week decreased in the Unclear 2015 Brief mailed intervention intervention group from 15.6 (8.8) Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 69 Mean age Population: Individuals with personalised Loss to follow-up: to 12.1 (7.0); in the control group Co- 64.7 (SD aged 50 and older who mailed feedback Intervention 13.6%; from 14.4 (7.0) to 13.5 (6.0). morbidity 8.4 years) were identified as at-risk outlining their specific Control: 2.4% Between group difference was Alcohol drinkers according to risks associated with not significant. CARET risk score Risk the Comorbidity alcohol use, educational Outcome also declined in both groups. In Evaluation Alcohol Risk Evaluation booklets about alcohol measurement: Self- the intervention group it declined Study Tool (CARET) reported, Comorbidity from 2.6 (1.6) to 1.6 (1.7) and Control group: (N=42). Alcohol Risk Evaluation from 2.3 (1.3) to 2.1 (1.4) in the N=86 randomised No intervention, Tool (CARET) control group, but the intervention received $5 gift card group had a statistically Gender: 66% male significantly greater decline than the control group (p<0.01). At Ethnicity: 88% white 3 months, fewer intervention (non-Hispanic); 9% group participants than controls Hispanic; 2% other were at-risk drinkers (66% vs 88%; OR 0.32, p=0.05), binge drinking (45% vs 68%; OR 0.33, p=0.03), using alcohol with a medical or psychiatric condition (3% vs 17%; OR 0.28, p=0.12), or having symptoms of such a condition (29% vs 49%; OR 0.38, p=0.07) 70 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Moore 2011 RCT US 55+ Older at-risk drinkers in Intervention: (N=310) Follow-up: 3 and 12 The intervention group consumed Low Addiction primary care identified Received a months fewer drinks (in the past 7 days) Mean age by the Comorbidity personalised report, at 3 months (rate ratio 0.79 (0.70 Healthy 68.4 (SD Alcohol Risk Evaluation booklet on alcohol and Loss to follow-up: to 0.90; p<0.001) and at 12 Living as 6.9 years). Tool (CARET). aging, drinking diary, (3 months) months (rate ratio 0.86 (0.76 to you Age Age range: advice from the primary Intervention 21%; 0.98; p<0.05) study 55-89 N=631 randomised care provider and Control 4%; (HLAYA) telephone counseling (6 months) Intervention There was a lower proportion of Gender: 71% male from a health educator 28%; at-risk drinkers in the intervention at 2, 4 and 8 weeks Control 7% group compared to the control Ethnicity: 87% white group at 3 months: 49.6% vs (non-Hispanic); 8% Control: (N=321) Outcome 61.2%, odds ratio 0.45 (95% Hispanic/Latino; 3% Received a booklet measurement: Self- CI 0.28, 0.81; p<0.01). At 12 other on healthy behaviour reported, Comorbidity months the proportion of at-risk during an office Alcohol Risk Evaluation drinkers in the intervention group appointment Tool (CARET) and compared to control was 54.1% Timeline Follow Back vs 59.9%, but the difference was (TLFB) questionnaire not statistically significant. Odds ratio 0.75 (95% CI (0.42 to 1.36). Similarly, at 3 months, there was less heavy drinking in the intervention group: odds ratio 0.45 (0.21 to 0.96; p<0.05) but this was not significant at 12 months Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Watson RCT (mul- UK 55+ Older hazardous Brief, minimal Follow-up: 12 months Both groups reduced alcohol Unclear 2013 ti-centre) (England alcohol users in primary intervention vs stepped consumption between baseline and Mean age care scoring >8 on the care. Minimal intervention: and 12 months. There were no AESOPS Scotland) 63.0 (SD Alcohol Use Disorders 11% at 6 months; 11.8% significant differences in average trial 5.8 years) Identification Test (10- Intervention: (N=266) at 12 months drinks/day (ADD) between the

item) Those in the stepped groups at 12 months Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 71 (AUDIT) care arm initially Stepped care received a 20-minute intervention: 9.8% at Average drinks/day (ADD) N=529 randomised session of behavioural 6 months; 13.2% at 12 In the stepped care group change counselling, months drinks/d decreased from 3.38 Gender: 80.3% male with referral to step (SD 2.14) at baseline to 2.45 (SD 2 (motivational Outcome 1.83) at 6 months to 2.56 (SD Ethnicity: Not reported enhancement therapy) measurement: 2.09) at 12 months. and step 3 (local Self-reported, In the minimal intervention group specialist alcohol AUDIT-Consumption drinks/d decreased from 3.41 services) if indicated. (AUDIT-C), DPI (SD 2.19) at baseline to 2.81 (SD Sessions were recorded (Drinking Problems 2.03) at 6 months to 2.49 (SD and rated to ensure Index); Quality- 1.93) at 12 months. treatment fidelity adjusted life-years At 6 months the mean difference (QALYs) (for cost–utility between the groups (drinks/d) Comparator: analysis derived from was -0.073 (–0.156 to 0.011); p (N=263) The minimal European Quality of = 0.088 intervention group Life-5 Dimensions); and At 12 months the mean received a 5-minute health and social care difference between the groups brief advice intervention resource use (drinks/d) was 0.025 (–0.062 to with the practice 0.112); p = 0.575 or research nurse involving feedback Screening costs: Mean of the screening screening cost for every results and discussion participant recruited into the trial regarding the health was £5.52 (2010 costs) consequences of continued hazardous Intervention costs: No alcohol consumption statistically significant difference in costs between groups at 6 and 12 months. At 12 months, participants in the stepped care group incurred fewer costs, with a mean difference of –£194 (95% CI –£585 to £198), and had gained 0.0117 more QALYs (95% CI –0.0084 to 0.0318) than the control group 72 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 10: Alcohol – Interventions that targeted a range of health behaviours with seperate alcohol outcomes

Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Harari 2008 RCT UK 65+ Functionally Intervention: Multi- Follow-up: 1 year Over the range of health Unclear Mean age: independent domain health behaviours or preventative health This study 74 community-dwelling promotion study using Loss to follow-up: 0% care measures in older adults referenced older adults the mailed HRA-O (numbers analysed at examined, there was minimal in Ander- followed by computer- baseline and follow-up improvement in any health son et al. Gender: 56.0 % female generated individualised the same) behaviour or uptake (except VINTAGE in intervention group written feedback to pneumococcal vaccination) project and 52.9% female in participants and GPs. Outcome control Health behaviours measurement: Self- At 1 year follow-up, there was addressed: PA, diet, reported, health risk no significant difference between Ethnicity: Not reported smoking, alcohol, seat appraisal for older groups in people reporting ‘no or belts when driving. persons (HRA-O) moderate’ alcohol use - 80.2% Preventive care: BP, questionnaire of those in the intervention group cholesterol, blood and 79.7% of those in the control glucose, faecal occult group (OR: 1.1 (95% CI 0.8, 1.3), blood test, influenza p=0.63) or pneumococcal vaccinations, dental, vision, hearing, mammography checks

Comparator: No intervention Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Schonfield Non-ran- US Mean age: Setting: Screening Intervention: Brief Follow-up: Time from Scores on the SMAST-G High 2010 domised 75 conducted at health intervention (1-5 1-hour baseline to intervention significantly decreased Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 73 before and fairs, retirement sessions): future goals discharge not reported; (t108=6.09; P<.001) from initial Florida Brief after inter- communities and to improve quality of life, then optional 30-day screen to discharge. The means Interven- vention senior housing health habits (exercise and 90-day follow-up between baseline and discharge tion and sites. Interventions and use of tobacco, were significantly Treatment conducted at home, alcohol, medications, Loss to follow-up: different but the means from for Elders ageing services sites or and drugs), motivational 53.3% of those who discharge to 30-day follow-up (BRITE) medical settings interviewing, education received intervention were not significantly different project about consequences of did not complete Population: Older drinking and reasons to intervention discharge. Among those who screened adults who screened cut down positive for alcohol problems on positive for alcohol Outcome the baseline SMAST-G screen, misuse The programme was measurement: self- only 18.9% were still positive at delivered by trained reported, telephone or discharge and follow-up (N=244 screened counsellors including in-person interviews, positive for alcohol and addictions specialists, including the 10- received intervention) nurses, social workers, item Short Michigan and mental health Alcoholism Screening Gender: 69.5% female counsellors with specific Test, Geriatric Version Ethnicity: 76.2% white; BRITE training (SMAST-G) 17.1% black; 5.7% multiracial; 0.3% Asian 74 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention and Follow-up Results Risk of design (years) and setting comparator and outcomes bias Vrdoljak RCT multi- Croatia 65+ Setting: General Lifestyle intervention Follow-up: 18 months Outcomes reported separately High 2014 centre, con- Mean age: practice that targeted a range of for each health behaviour, ducted in 72.3 (SD health behaviours: PA, Loss to follow-up: including alcohol 59 general 5.2 years) Population: Croatian smoking, alcohol, diet Of those completing practices citizens aged 65+ the baseline survey for There was no significant years who visited Intervention (N=371): alcohol (N=104), 97.1% difference between groups their GP for any Intensified intervention completed the follow-up for alcohol consumption (chi- reason (those with life delivered by GPs. survey. However, only squared = 0.73, df = 1, p = 0.394) expectancy <6 months, Intervention participants 29% of participants at at the end of intervention severe dementia, were counselled and baseline completed the severe mental illness, given a tailored life alcohol questions Note: it may not have been the communication plan for adopting same people who completed disability excluded) healthier behaviour. Outcome alcohol questions at baseline and Each patient received measurement: Self- follow-up. (Also no significant Gender: 61% female educational leaflets for reported, questionnaire differences between the their detected CV risk intervention and control groups Ethnicity: Not reported factors and a specific for physical activity (chi-squared appointment was given = 0.84, df = 1, p = 0.36), smoking for the next follow-up (chi-squared = 0.85, df = 2, p = visit 0.65) at the end of the study)

Comparator (N=367): Usual care: GPs were not instructed to give any specific intervention Table 11: Alcohol – Included qualitative studies about barriers and facilitators

Study Study Country Age Population Objective Quality design (years) Aira 2008 Qualitative Finland 75+ N=699 home-dwelling elderly living in the community To describe alcohol use - (interviews) as self-medication by Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 75 (Part of GeMS and (83.4% aged Gender: 30.5% male; 69.5% female people aged over 75 study: Geriatric quantitative 75-84; 16.6% years Multidisciplinary assessment aged 85+) SES: 18.0% had >9 years education; 52.0% had 4-9 years Strategy for Good of alcohol use education; 17.4% <4 years education Care of the Elder- using AUDIT- ly) questionnaire Alcohol consumption: 48.5% had used alcohol in past year; 19.7% had used alcohol for medicinal purposes in past year. Similar in males and females

Dare 2014 Qualitative Australia 65-74 N=20 men and N=22 women who were living in either private To identify relationships ++ study (in-depth residences or (secular, resident-funded) retirement villages between social interviews) Mean age: 69.7 engagement, setting and (SD 3.3 years) Gender: 47.6% male; 52.4% female alcohol use

SES: Participants from areas classified as having higher levels of socioeconomic advantage and >50% had a post- school qualification

Alcohol consumption: Over 75% drinking alcohol over 4 days/wk. Average daily consumption of alcohol (standard drinks): Private home; men 1.89 (1.4); women 1.21 (0.8); Retirement village: men 3.13 (4.4); women 1.68 (1.1)

Haarni 2010 Qualitative Finland 60-75 N=31 Urban older adults Life experience and - (interviews) alcohol: 60-75 year olds (Too Much is Gender: 48.4% male; 51.6% female relationship to alcohol Always Too Much - Alcohol and Ageing SES: 14 had passed the matriculation examination, eight project) the middle school, and eight had at least the equivalent of elementary school studies

Alcohol consumption: People who had abstained from alcohol all their life were excluded. Reports: ‘the study included many kinds of alcohol consumers’ 76 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Objective Quality design (years) Johannessen 2015 Qualitative Norway 65+ N=16 older adults that received in-home nursing service or Older adults’ experience ++ (interviews) home-help services (N=14 were widows or widowers) with and reflections on Mean age: 81 use and misuse of alcohol Age range: Gender: 37.5% male; 62.5% female and psychotropic drugs 65-92 SES: Not reported

Alcohol consumption: 15 (of 16) had used alcohol

Joseph 2012 Qualitative (in- Canada 44-74 Older male cricket players of Afro-Caribbean origin (friendly, Alcohol and older + depth formal non-league) and spectators (male and female). N=27 formal Caribbean-Canadian men interviews, Mean age: 61 interviews conducted but data collected in a range of ways observation, casual Gender: Predominantly male conversation, informal SES: Not reported interviews) Alcohol consumption: Most of the participants appeared to drink heavily but not specifically reported

Kim 2009 Qualitative Canada 60+ N=19 elderly Korean immigrants residing in Canada (14 men, To explore drinking + (focus group) 5 women) culture, alcohol and Mean age 72 alcohol use in older (SD 5.94 years) Gender: 26.3% male; 73.7% female Korean immigrants in Age range: Canada 62-83 SES: Not reported

Alcohol consumption: Drinking alcohol was a criterion for recruitment. 63.2% drank more than once a week Study Study Country Age Population Objective Quality design (years) Millard 2008 Qualitative UK (Scotland) 65+ N=90 staff and managers providing home, day and residential Alcohol and service gaps -

(focus groups) care to elderly clients in homecare for older Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 77 adults: including how Gender: Not reported for staff/managers or elderly clients client’s alcohol problems were identified, role of SES: Not reported for staff/managers or elderly clients home care provider, barriers to seeking help Alcohol consumption: Not reported for staff/managers or elderly clients

Tolvanen 2005 Qualitative Finland 90+ N=181 participants who mainly lived in their own homes Alcohol in life story + Vitality 90+ project (interviews) though some were in service housing or in nursing homes interviews with Finnish people aged 90 Gender: 33.5% male; 76.5% female (of interviews that or over discussed alcohol)

SES: Not reported

Alcohol consumption: 63% of men and 34% of women currently used alcohol; 23% of men and 6% of women had previously used alcohol earlier but no longer drank; and 13% of men and 24% of women did not drink at all (based on those who reported consumption in interviews)

Ward 2011 Qualitative UK Range: mid 50s Aimed to include a diverse range of older adults. Included Older adultss + (N=21 to late 80s people living in their own homes, in sheltered housing and in perspectives on alcohol interviews and hostels use in later life N=3 focus groups with Gender: 61.9% male, 29.1% female (interviews) older adults) SES: Not reported

Alcohol consumption: Not specifically reported but all participants appeared to consume alcohol 78 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Objective Quality design (years) Wilson 2013 Qualitative UK (North East 50+ Older adults recruited through Age UK and regional services To understand older ++ England) Range: for alcohol problems who had experience of drinking alcohol adults reasoning about Interviews 50-95 at any level of consumption drinking in later life and (N=24) and 3 how this interacted with focus groups Recruited a range of patterns of consumption, including health concerns (N=27 people) occasional minimal drinkers, moderate and heavy drinkers, previously dependent drinkers and 2 currently dependent drinkers

Gender: 50% male; 50% female (interviews); 22.2% male; 77.8% female.

SES: Not reported

Alcohol consumption: Those recruited had a range of patterns of consumption including occasional minimal drinkers, moderate and heavy drinkers, previously dependent drinkers and 2 currently dependent drinkers Table 12: Alcohol – Older adults’ drinking habits in the context of ageing

Aira Dare Haarni Johannes- Joseph Kim Millard Tolvanen Ward Wilson sen Social/Relaxation Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 79 Alcohol as a social lubricant X X X X X X Drinking for relaxation X X X As a treat/something special X X Fun and enjoyment X X X Quality of life X Part of social environment X X Access issues Cost and availability X Health-related aspects Drinking for medicinal purposes X X X X X Drinking for relaxation X X X In the context of ageing X X Ill health/drinking behaviour X X Health risks X X X X X X Driving X Other Drinking to deal with negative issues X X X X X Positive versus negative alcohol X X X X X identities Self-regulating strategies X X X 80 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 13: Alcohol – Quality assessment for interventions reporting on healthy drinking behaviours

Population Method of allocation to intervention Outcomes Analysis Summa- Over- (or comparison) ry all 1.1 1.2 1.3 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 3.1 3.2 3.3 3.4 3.5 3.6 4.1 4.2 4.3 4.4 4.5 4.6 5. 5.2

Arean 2008 + + + ++ + - - + ++ NR ++ ++ + ++ NA ++ ++ ++ ++ ++ ++ NR ++ ++ + ++ - + Cucciare + + + ++ ++ - + ++ - - ++ ++ ++ ++ NA ++ ++ ++ ++ ++ ++ - ++ + ++ - + + 2013 Ettner 2014 ++ ++ ++ ++ ++ ++ + ++ ++ + + ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + ++ ++ ++ ++ ++ ++ Fink 2005 ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ NR ++ ++ ++ ++ ++ ++ ++ Gordon ++ + + + ++ NR - ++ NA + NR ++ NA ++ NA ++ ++ ++ ++ ++ - - ++ + - + - - 2003 Hansen ++ ++ + ++ ++ ++ - ++ NR ++ ++ NA + ++ NA ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + + 2012 Harari 2008 ++ + + ++ ++ ++ - + + NR ++ ++ ++ - NA + + ++ ++ ++ ++ NR ++ ++ ++ + + + Kuerbis ++ + + ++ ++ ++ - ++ ++ + ++ ++ NA ++ NA ++ ++ ++ ++ ++ - - ++ ++ ++ - + + 2015 Moore 2011 ++ ++ ++ ++ ++ ++ + ++ ++ + ++ ++ NA ++ NA ++ ++ ++ ++ ++ ++ + ++ ++ ++ ++ ++ ++ Oslin 2006 + + - + + - - + NR NR ++ ++ + + NA + + ++ + - - - ++ + ++ + - + Schonfield + - + - + NA NA + NA NA - + + ++ NA NA ++ NA + NA NA NA NA ++ ++ - + - 2010 Vrdoljak ++ + + + + NR + + ++ + NR ++ ++ + NA + + ++ + + - NR + - - - + - 2014 Watson ++ ++ ++ ++ ++ ++ - ++ + + ++ ++ ++ ++ NA ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ 2013 Table 14: Alcohol – Quality assessment for interventions reporting barriers and facilitators

Theoretical Study Data Trustworthiness Analysis Ethics Overall approach collection

1 2 3 4 5 6 7 8 9 10 11 12 13 14 Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 81 Aira 2008 Appro- Clear Not sure Appro- Unclear Unclear Not sure Not re- Poor Not re- Not sure Partially Inade- Appro- - priate priate ported ported relevant quate priate Dare 2014 Appro- Clear Defensi- Appro- Clear Not sure Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ priate ble priate ous vincing vant quate priate Haarni 2010 Appro- Clear Not sure Inade- Not de- Unclear Not sure Not sure Not sure Not sure Not sure Partially Not sure Not re- - priate quately scribed relevant ported reported Johanessen Appro- Clear Defensi- Appro- Clear Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ 2015 priate ble priate ous vincing vant quate priate Joseph 2012 Appro- Unclear Not sure Appro- Clear Clear Not sure Not sure Rich Not sure Not sure Partially Ade- Not re- + priate priate relevant quate ported Kim 2009 Appro- Clear Defensi- Appro- Clear Clear Not sure Rigor- Rich Reliable Not sure Partially Not sure Not re- + priate ble priate ous relevant ported Millard 2008 Appro- Mixed Not sure Inade- Not de- Unclear Not sure Not re- Not re- Not re- Not sure Rele- Not sure Not re- - priate quately scribed ported ported ported vant ported reported Tolvanen 2005 Appro- Mixed Defensi- Appro- Clear Not sure Not sure Rigor- Rich Not sure Con- Rele- Ade- Not re- + priate ble priate ous vincing vant quate ported Ward 2011 Appro- Clear Defensi- Appro- Clear Clear Not sure Not re- Not sure Not re- Con- Rele- Ade- Not re- + priate ble priate ported ported vincing vant quate ported Wilson 2013 Appro- Clear Defensi- Appro- Clear Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ priate ble priate ous vincing vant quate priate 82 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 15: Smoking - Interventions with smoking cessation and reduction outcomes (RCTs)

Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Hall 2009 RCT US 50+ Smokers of >/= 10 1. Participants Follow-up: 6, 12, 18 E-CBT had significantly higher Unclear cigarettes per day: completed a 12-week and 24 months (post- cigarette abstinence rates than Mean age treatment that included randomisation). ST (no further treatment), odds 56.7 (SD Mean cigarettess/day: group counselling, ratio (OR) 1.27; 5.9) 20.5 (SD 8.7) nicotine replacement Lost to follow-up: wk 95% CI 1.52, 1.05, P = 0.01 and therapy (NRT) and 24 4.0%; wk 52 7.0%; E-NRT (OR 1.22; 95% CI 1.45, Mean years of regular bupropion wk 104 13.4% 1.03, smoking: 37.8 (8.2) 2. Participants, P = 0.02 over 2 years. There independent of smoking Outcome were no significant differences for N=402 randomised status, were then measurement: Primary E-NRT or E-combined compared randomised to follow-up outcome variable was to ST Gender: 60% male conditions: 7-day point prevalence There was no significant (i) N=100 standard cigarette abstinence difference between E-CBT and Ethnicity: 76.9% treatment (ST; no verified biochemically E-combined (OR 1.18; 95% CI Caucasian further treatment) (2 assays) at weeks 24, 1.40, 0.99, P = 0.06) (ii) N=99 extended NRT 52, 64 and 104. No differential effects by gender SES: 21.9% had a grad (E-NRT; 40 weeks of degree; 30.5% college nicotine gum availability Note: Participants graduates; 35.5% some (iii) N=99 extended were paid $25 for college; 12.1% high cognitive behavioural each assessment they school graduates or therapy (E-CBT; 11 completed less cognitive behavioural sessions over a 40- week period) OR (iv) N=104 E-CBT plus E-NRT (E-combined; 11 cognitive behavioural sessions plus 40 weeks nicotine gum availability) Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Joyce 2008 Cluster US 65+ Older adults voluntarily Interventions: Follow-up: 6 and 12 Telephone counselling Quitline High

RCT enrolled on the Usual care: self-help months post-baseline with nicotine patch was most Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 83 Medicare Stop Smoking educational materials effective Programme (N=2230) Lost to follow-up: Reimbursement for 23.6% did not complete Quit rates at 12 months were:- Years smoking: 66-69% provider counselling either 6 or 12 month Quitline +nicotine patch: 19.3% >50 years (N=829) assessments (95% CI 17.4-21.2) Reimbursement for Provider counselling: 14.1 % Heavy smokers: 27- provider counselling Outcome (11.7-16.5) 30% smoked 25+ cigs/ with pharmacotherapy measurement: 7 day Provider counselling + day (N= 2605) self-reported cessation pharmacotherapy: 15.8% (14.4- Telephone counselling 17.2) N=7354 randomised Quitline with nicotine Usual care: 10.2% (9.0-11.5) patch (N=1690) Gender: 39-42% male (depending on group)

Ethnicity: White: 89-95%; black: 3-7%; other: 2-6%(depending on group)

SES: Income: lowest: 18-21%; low: 19- 21%; medium: 43- 47%; higher: 13-18% (depending on group) Education: less than high school: 16-22%; high school: 36-39%; college: 41-47% 84 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Kim 2005 RCT Korea 50+ In younger and older Agency for Health Care Follow-up: 5 months More effective for younger versus Unclear smokers but conducts Policy and Research older smokers Mean age: sub-group analysis by guideline tailored for a Lost to follow-up: intervention age (50+ versus <50 Korean population 1.7% overall For the whole group: 14% of the 51.6 (SD years) intervention group and 9% of the 13.0); con- Intervention: Outcome control group had validated quits trol: 53.1 Currently smoking one N=132 who were willing measurement: after 5 months: risk ratio (1.56, (SD 13.5) cigarette a day or more. to quit within one month Smoking cessation 95% CI 0.89-2.73) received part 2A of defined as absolutely Participants were AHCPR guideline: no smoking since the However, when only those aged referred by physicians including being helped last quit attempt, self- >50 were included there was no regardless of their to set a quit date, sign a reported by telephone difference between the groups: willingness to quit stop-smoking contract, interview and verified N=112 in intervention group provided with self-help by CO analyser (7ppm and N=123 in the control group; N=401 randomised material. cut-off or less) 13.4% of the intervention group (66% of intervention N=68 who were not and 13.0% of the control group group and 64% of willing to quit within had validated quits: risk ratio control group willing to one month received (1.03, 95% CI 0.53-1.99) quit within 1 month) an intervention based on relevance, risks, Gender: 94.5% male rewards and repetition (intervention), 90.5% with phone follow-ups male (control) (also based on AHCPR)

Ethnicity: Korean (no Comparator: other ethnicity details N=201 were told to quit reported) smoking by their own free will without any SES: 54% in further assistance intervention group and 51% of control group were high school graduates Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Orleans RCT US 65+ Older smokers Intervention: Follow-up: 6 and 12 At 6 months, those in the High 2000 who applied for a Participants received months post-baseline intervention group were more Mean age transdermal nicotine a copy of a tailored likely to report 7 day point Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 85 Note: not 72 prescription self-help quitting guide Lost to follow-up: prevalence abstinence than clear if this for older smokers (Clear 14% (6 months) ; 22% those in the comparator group. study has Mean smoking duration Horizons) and a series (12 months) (40% vs 33%; p<0.05) been peer- >50 yrs of seven personalised reviewed computer generated Outcome However, at 12 months, the Mean cigarettes/day: 22 mailings (received over measurement: differences were not statistically a period of 10 days to 6 Self-reported (phone significant (33% vs 31%) N=470 months after baseline), interviews) and delivered through a Gender: predominantly state prescription plan female (specific data not reported) Comparator: Usual care: a fact sheet on Ethnicity: patch-assisted predominantly white (no quitting specific data reported)

SES: majority had not completed high school (no specific data reported) 86 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 16: Smoking - Interventions with smoking cessation and reduction outcomes (non-RCTs)

Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Pothirat Non- Thailand 60+ Community-dwelling Behavioural group Follow-up: 3, 6, 12 The continuous abstinence rate High 2015 randomised elderly smokers with therapy (n=120) versus months (CAR) of the behavioural therapy controlled Mean age: smoking history of at educational programme group at the end of the study trial Behavioural least 10 pack years (n=104). Lost to follow- (month 12) was significantly 67.2 (SD and who aimed to quit up: Intervention higher than the education group 6.7); smoking within a month Intervention: (behavioural): 3 months: (40.1% vs 33.3%, p=0.034). Educational Behavioural group 2.5%; 6 months 1.7%; Similar results were also found 66.9 (6.5) Mean smoking duration therapy 12 months 1.7%. throughout all follow-up visits 51yrs (SD 13) Same educational at month 3 (57.3% vs 27.0%, programme as the Outcome p<0.001) and month 6 (51.7% vs Mean cigarettes/day: 12 other group plus a measurement: 25%, p<0.001) (SD 0.8) total of 9 hours over self-reported and 3 days of behavioural biochemical test All study participants therapy including (exhaled carbon smoked roll-your-own demonstrations of the monoxide < 10 ppm) native unregulated health consequences cigarettes (crudely cut of smoking, coping tobacco mixed with and social skills ground tamarind pod) training, self-control, cognitive-behavioural N=224 interventions, reinforcement and Gender: Behavioural relaxation group therapy 40% male; Education 49% Comparator: male Educational programme Ethnicity: Not reported 2hr education program that included a SES (based on lecture on the health income): 96% low SES; consequences of 4% moderate SES smoking. Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Tait 2006 Non- Australia 68+ Community-dwelling Participants in a larger Follow-up: 6 months At 6 months, 88.5% of the High

randomised older smokers (>/= study were interviewed: intervention group had made Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 87 controlled 5 cigarettes/day those considering Lost to follow-up (6 an attempt to quit, and 31% of trial equivalent) quitting received months): Intervention: the intervention group versus intervention 3.6%; Comparator: 10% none from the control group N=215 eligible reported not smoking for the past Intervention (N=165): Outcome 30 days, however a confirmed Mean years of regular Brief intervention, measurement: ECO reading

Comparator (N=50): No intervention (those with no plans to quit smoking i.e. continuing smokers) 88 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 17: Smoking - Interventions that targeted a range of behaviours with separate smoking outcomes

Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Harari 2008 RCT UK 65+ Functionally Intervention: Multi- Follow-up: 1 year Over the range of health Unclear Mean age independent domain health behaviours or preventative health 74 community-dwelling promotion study using Loss to follow-up: 0% care measures in older adults older adults the mailed HRA-O (numbers analysed at examined, there was minimal followed by computer- baseline and follow-up improvement in any health Gender: 56.0% female generated individualised the same) behaviour or uptake (except in intervention group written feedback to pneumococcal vaccination) and 52.9% female in participants and GPs. Outcome control Health behaviours measurement: Self- At 1 year follow-up, there was addressed: PA, diet, reported, health risk no significant difference between Ethnicity: Not reported smoking, alcohol, seat appraisal for older groups in people reporting ‘no belts when driving. persons (HRA-O) current tobacco use’- 90.9% of Preventive care: BP, questionnaire those in the intervention group cholesterol, blood and 89.6% of those in the control glucose, faecal occult group (OR: 1.2 (95% CI 0.9, 1.6), blood test, influenza p=0.36) or pneumococcal vaccinations, dental, vision, hearing, mammography checks.

Comparator: No intervention Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Vrdoljak RCT Croatia 65+ Mean Setting: General Lifestyle intervention Follow-up: 18 months Outcomes reported separately High 2014 (multicentre age 72.3 practice that targeted a range of for each health behaviour, Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 89 conducted (SD 5.2) health behaviours: PA, Loss to follow-up: including smoking in 59 Population: Croatian smoking, alcohol, diet. Of all respondents, general citizens aged 65+ N= (94%) completed There was no significant practices) years who visited Intervention (N=371): the baseline survey difference between groups for their GP for any Intensified intervention for smoking and 59% smoking (chi-squared = 0.85, reason (those with life delivered by GPs. completed the follow-up df = 2, p = 0.65) at the end of expectancy <6 months, Intervention participants survey. intervention severe dementia, were counselled and severe mental illness, given a tailored life Outcome Note: it may not have been the communication plan for adopting measurement: Self- same people who completed disability excluded) healthier behaviour. reported, questionnaire alcohol questions at baseline and Each patient received follow-up At baseline, 69.6% were educational leaflets for Also no significant differences non-smokers, 23.5% their detected CV risk between the intervention and former smokers and factors and a specific control groups for physical 6.8% were smokers appointment was given activity (chi-squared = 0.84, df = for the next follow-up 1, p = 0.36), alcohol (chi-squared Gender: 61% female visit. = 0.73, df = 1, p = 0.394) at the (whole sample) end of the study) Comparator (N=367): Ethnicity: Not reported Usual care: GPs were not instructed to give any specific intervention. 90 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 18: Smoking - Interventions targeted at training health professionals in smoking cessation services

Study Study Country Age Population Intervention Follow-up Results Risk of design (years) and comparator and outcomes Bias Kerr 2011 RCT UK (Scot- Mean age: N=57 members of the Intervention: 1-day, Follow-up: 1 week and Statistically significant improve- Unclear (Kerr 2007) land) intervention primary care team who brief, smoking cessa- 3 months after training ment in knowledge and attitudes 46/ control work with older adults tion training aimed at of the intervention group. Prac- 44 (health (N=47 nurses and N=9 providing the knowl- Loss to follow-up: tice also improved but differences profession- allied health profes- edge and skills aimed N=73 randomised, were not significant als) sionals recruited from to deliver effective brief N=57 completed train- 7 community health interventions. Training ing; N=54 (94.7% of and social care partner- was tailored to over- those trained) complet- ships) come key barriers and ed 1 week follow-up; delivered by a profes- N= 52 (91.2%) 3 month Gender: Not reported sional experienced in follow-up the delivery of smoking Ethnicity: Not reported cessation training Outcome measurement: Self-reported, Comparator: No using a specifically de- training veloped and validated questionnaire. Qualita- tive data also collected from N=8 Table 19: Smoking - Overview of qualitative studies reporting barriers and facilitators

Study Study Country Age Population Objective Quality design (years) Huddlestone 2015 Qualitative UK (North-East 65+ N=7 older smokers registered at a general practice in a large To explore issues around -

(face to face England) Mean age 69 city engagement of older Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 91 (Abstract only, semi-structured smokers with smoking but reports main interviews) Gender: Not reported cessation support results) delivered in primary care Ethnicity: Not reported

SES: Not reported

Kerr 2006 Qualitative UK (Scotland) 65+ N=20 current (N=13) and former smokers (N=7) recruited Older current and ++ (semi- through general practices and a forum for older smokers former smokers’ views structured on smoking, stopping interviews) Gender: 46% male, 54% female smoking, and smoking cessation resources and Ethnicity: Not reported services

SES: Of the smokers, the majority were in the highest categories of deprivation (N=8 DEPCAT 6/7; N=3 DEPCAT 4/5; N=2 DEPCAT 3). Former smokers were less likely to be in the highest deprivation categories (N=1 DEPCAT 6/7; N=5 DEPCAT 4/4; N=1 DEPCAT 3)

The scoring system ranges from DEPCAT 1 (the most affluent postcode sectors) to DEPCAT 7 (the most deprived)

Kerr 2007 Qualitative UK Range < 25 A sample of N=41 health visitors, district nurses, practice Exploration of the ++ (semi- to 64 nurses and general practitioners working in primary care, with knowledge attitudes and structured contact with older adults (>/= 65 years). Recruited through practice of the primary interviews) 33 general practices to ensure diversity in level of socio- care team in relation to economic deprivation and geographical location smoking cessation to identify barriers to the Gender: 32% male, 68% female effective provision of intervention to older adults Ethnicity: Not reported

SES: Recruited from practices in a diverse range of socioeconomic areas 92 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Objective Quality design (years) Lundqvist 2006 Qualitative Sweden Age range N=9 middle-aged and elderly women. N=5 were smokers and To identify attitudes ++ (interviews) 47-70 N=4 ex-smokers. Aimed to recruit a wide range in terms of: and barriers to smoking social background, education, marital status and profession cessation among middle- aged and elderly women Gender: 100% female

Ethnicity: Not reported

SES: N=3 were blue collar workers, 2 were well-paid administrators, 2 had sickness pensions and 2 were retired

Medbo 2011 Qualitative Norway 58+ N=18 elderly smokers (N=5) and ex-smokers (N=13). N=11 Aim was to gain insights ++ (semi- had serious chronic diseases such as myocardial infarction, that may help general structured (Range 58-80) angina, stroke, COPD. Years smoking ranged from 10 to >60 practitioners understand interviews) why people smoke, and Gender: 65% male, 35% female why smokers quit and maintain quitting to inform Ethnicity: Not reported interventions to stop smoking SES: Not reported

Mohammadnezhad Qualitative Australia 50+ N=20 older Greek-Australian smokers living in a metropolitan To understand reasons for ++ 2015 A/B (semi- area (smoking defined as those who had smoked at least 100 smoking and attitudes to (A: BMC Public structured Mean age 64.6 cigarettes during his/her lifetime and were currently smoking). quitting; and sociocultural Health, B: Int J face-to-face (SD 10.0) 65% were suffering from diseases such as cancer or heart factors that can influence Eviron Res Public interviews, disease smoking and smoking Health) with a Greek cessation behaviour translator) The mean years of smoking were 45.5 years (SD = 10.8 years)

Gender: 60% male, 40% female

Ethnicity: All self-identified as Greek-Australians residing in Australia

SES: 60% had completed high school level of education Table 20: Smoking - Barriers to smoking cessation in older adults

Kerr 2007 Huddlestone Lundqvist Mohammadn- Mohammadn- Medbo ezhad A ezhad B Health and quality of life

Loss of enjoyment/pleasure X X Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 93 Loss of perceived stress reduction/relaxation X X Loss of ‘boredom relieving’ qualities X Potential weight gain X X Belief the damage has already been done X Adverse life events X Addiction to nicotine X X Low knowledge about related diseases and risks X Sociocultural Difficulty in engaging longstanding smokers X High acceptability in some cultural groups X X Social arrangements X ‘Loss of special sense of group belonging’ X Integration into social activities X Social networks/ influence of other people X X Psychological Low self-efficacy X X Lack of motivation X Lack of acknowledgement of health problems X X Fatalism X Fear of craving X Smoking as a life-long habit X Smoking as a personal choice X Smoking cessation routes and services Potential health risks/side effects of NRT X X 94 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Kerr 2007 Huddlestone Lundqvist Mohammadn- Mohammadn- Medbo ezhad A ezhad B Belief NRT not compatible with other health prob- X lems Lack of confidence in NRT X X X Focus on disease within healthcare system X Lack of confidence in available support X Lack of follow-up from medical staff, pharmacists, X etc. (e.g. NRT) Disease focus of hospital-based programmes X Lack of support from some health professionals X Table 21: Smoking - Facilitators to smoking cessation in older adults

Kerr 2007 Huddlestone Lundqvist Mohammadn- Mohammadn- Medbo ezhad A ezhad B Health and quality of life

Improved health/prevention of ill health X X X X Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 95 Free from smell of smoking X Sense of freedom after cessation X Financial issues Saving money X Sociocultural Increasing social unacceptability X X Care, household and occupational X responsibilities Family pressure/encouragement X X

Social networks/ influence of other people X X Psychological Will to stop smoking X Smoking cessation routes and services Lack of knowledge about services X Health promotion rather than focus on disease X Key role of support from primary care team X Input of health professionals X X Consistent advice from primary care physicians X Support tailored to individual needs X X Choice of smoking cessation routes X Tobacco as a ‘safety net’ X Different diversion to help stop smoking X 96 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 22: Smoking – Quality assessment for interventions reporting on smoking cessation and reduction outcomes

Selection bias Performance bias Attrition bias Detection bias Summary Risk of A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Bias all all all all RCTs Hall 2009 Y Y Y Low Y Un- Un- Un- Y Y Y Low Y Y Y Un- Un- Un- Unclear clear clear clear clear clear clear Joyce 2008 Y Un- N Un- Y Un- Un- Un- Y Un- Un- Un- Y N Un- Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear Kerr 2011 Un- Un- Un- Un- Y N N Un- Y Un- Un- Un- Y N Un- N Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear Kim 2005 Y Y Y Un- Un- Un- Un- Un- Y Un- Un- Un- Un- Y Y N Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear Orleans Un- Un- Un- Un- Un- Un- Un- Un- Un- Un- Un- Un- Y Un- Un- Un- Un- Un- High 2000 clear clear clear clear clear clear clear clear clear clear clear clear clear clear clear clear clear Vrdoljak Y Un- Un- Un- Un- Y Y Un- Y Un- Un- Un- Y N N Un- Un- High High 2014 clear clear clear clear clear clear clear clear clear clear Non-randomised studies Pothirat N N N High Un- Un- Un- Un- Y Un- Un- Un- Y Y Y Un- Un- Un- High 2015 clear clear clear clear clear clear clear clear clear clear Tait 2006 N N N High N Y Y Un- Y N N Un- Y Y Y Un- Un- Un- High clear clear clear clear clear/ Low Table 23: Smoking – Quality assessment for studies reporting barriers and facilitators

Theoretical Study Data Trustworthiness Analysis Ethics Overall approach collection

1 2 3 4 5 6 7 8 9 10 11 12 13 14 Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 97 Huddlestone Appro- Unclear Not sure Not Not de- Unclear Not sure Not Not Not Not sure Rele- Not sure Not - 2015 priate sure/ scribed sure/not sure/not sure/not vant sure/not inade- reported reported reported reported quately reported Kerr 2006 Appro- Clear Defensi- Appro- Clear Clear Not sure Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ priate ble priately ous vincing vant quate priate Kerr 2007 Appro- Clear Defensi- Appro- Appro- Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ priate ble priately priately ous vincing vant quate priate Lundqvist 2006 Appro- Clear Defensi- Appro- Unclear Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ priate ble priately ous vincing vant quate priate Medbo 2011 Appro- Clear Defensi- Appro- Appro- Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ priate ble priately priately ous vincing vant quate priate Mohammadn- Appro- Clear Defensi- Appro- Clear Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ ezhad 2015 A priate ble priately ous vincing vant quate priate Mohammadn- Appro- Clear Defensi- Appro- Clear Clear Reliable Rigor- Rich Reliable Con- Rele- Ade- Appro- ++ ezhad 2015 B priate ble priately ous vincing vant quate priate 98 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 24: Diet – Interventions to increase uptake/maintenance of healthy diet behaviours (RCTs)

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Appleton RCT N. Ireland, 65+ Community-dwelling older Fruit-tasting sessions Follow-up: 5 weeks (to For all participants (n=95), 2013 UK adults involving familiar fruits and end of intervention) There were no significant novel fruit products and differences in fruit or fruit and Setting: Community dishes. The interventions Lost to follow-up: 0% vegetable intake between groups based church and social were designed to be reported but data imputed groups social, enjoyable activities from mean of 2 other In low consumers of fruit (one to be undertaken in timepoints for missing portion/day or less) (n=38) N=95 randomised community-based values Fruit consumption was groups, which could be significantly higher in the Gender: 28-34% across implemented by non- Outcome measurement: repeated exposure groups groups health professionals Self-reported, 24 hr food compared to the single exposure recalls with additional group (groupxtime interaction: Ethnicity: Not reported 3 intervention groups: prompts from researchers F(3,138)=3.36, p=0.02) (1) Single exposure (E1): SES: Not reported fruit sampling on one Fruit intakes increased occasion significantly in the repeated Only (n=39) exposure groups (E5, E5+) (2) Repeated exposure (t(30) 5·79, P<0·01), but did not (E5): fruit sampling on one change in the E1 group (t(16) occasion per week for 5 0·29, P=0·78). No differences weeks (n=38) were found between E5 and E5+ (3) Repeated exposure groups (F(3,87) 1·22, P=0·31). Plus (E5+): fruit sampling Similar effects were also found on one occasion per week for fruit and vegetable intakes for 5 weeks and provision of one portion of fruit per d to be consumed at home (n=18) Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Atienza RCT (pilot US 50+ (mean Population: N=27 Examined effect of using Lost to follow-up: (per 1000 kcal) in the intervention

2008 study) age in randomised. a hand-held computer for Intervention: 16%; Control group compared to controls Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 99 control 58 increasing vegetable and 31% (from 1.5 at baseline to 2.5 y and int Setting: Community whole-grain intake over 8 servings/day; p=0.02). Control 63 y) weeks. Outcomes: Servings of group remained constant at 1.9 N=36 randomised Intervention (n=20): vegetables/day; dietary servings/d ( from graph) Instruction session and fibre from grains (as a Gender: Intervention hand-held computer proxy for wholegrain Dietary-fibre intake from grain 69% female; control 70% programmed to monitor intake) assessed by FFQ sources per 1000 kcal was not female their vegetable and whole- statistically significant between grain intake twice per groups Ethnicity: Intervention: day and to provide daily 88% white; control 90% individualised feedback, white goal-setting and support Control (n=16): Standard, SES: Education (yrs): age-appropriate written Intervention: 16.8 (2.2); nutrition education control 16.7 (2.5) materials 100 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 100 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Babatunde RCT US 50+ (mean Population: Community Intervention: (n=59) Follow-up: 6 weeks (end The experimental group 2011 age 70.2; dwelling older, black Received 6-weekly of intervention) had a statistically significant range 50 to adults. Male and female osteoporosis education improvement in calcium intake 92) programme, developed Lost to follow-up: 15.4 % compared to those in the control Setting: Churches with a theoretical (Wilks λ 0.47, F1,108 and community-based background, with a focus Outcomes: Self-reported, 122.97, P < .001, η2 0.53). There organisations on modifying health beliefs Dietary calcium intake was an average increase of 556 and self-efficacy with measured using the mg in dietary calcium intake from N=110 randomised regard to osteoporosis Random Assessment a mean score of 874 (SD324) at prevention Method (RAM) calcium baseline to 1430 (SD331) mg/ Gender: 90% female checklist, a type of food day at the end of the program in Control: (n=51) Wait- frequency questionnaire the experimental group Ethnicity: 100% black (of list control (received the African descent) intervention later) In the wait-list control group, calcium intake slightly SES: Years of schooling decreased from a mean of 817.6 ranged between 8 and 20 (SD326.7 to 778.2 (SD 369.3) years (mean 13.1 yrs), and mg/day 73% had more than a high school diploma

Barr 2000 RCT US 55 to 85 People in good general Intervention: The milk Follow-up: 4 week Compared with controls, years health, 5 years or more group were instructed to baseline period and 12 participants in the milk post menopause for add 3 x 8oz servings of week intervention supplemented group significantly women, consuming 1.5 skim or 1% fluid milk to increased energy, protein, servings or fewer of dairy their usual consumption Lost to follow-up: 3% in cholesterol, vitamins A, D and products daily and willing of dairy products and to intervention group and 1% B12, riboflavin, pantothenate, to consume 3 additional follow their usual diet in in control group calcium phosphorus, magnesium, servings of fluid milk daily other ways. No instruction zinc and potassium to adjust for the added Outcome measurement: N=204 randomised energy from milk was self-reported 3-day food The milk group gained 0.6 kg given records checked by more than the control group dietitians (p<0.01), however weight gain Control: Instructed to was less than predicted. Authors maintain their usual diet suggest likely there was some including consumption of compensation for the added fewer than 1.5 servings/ energy from milk day Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Babatunde RCT US 50+ (mean Population: Community Intervention: (n=59) Follow-up: 6 weeks (end The experimental group 2011 age 70.2; dwelling older, black Received 6-weekly of intervention) had a statistically significant range 50 to adults. Male and female osteoporosis education improvement in calcium intake 92) programme, developed Lost to follow-up: 15.4 % compared to those in the control Setting: Churches with a theoretical (Wilks λ 0.47, F1,108 and community-based background, with a focus Outcomes: Self-reported, 122.97, P < .001, η2 0.53). There organisations on modifying health beliefs Dietary calcium intake was an average increase of 556 and self-efficacy with measured using the mg in dietary calcium intake from N=110 randomised regard to osteoporosis Random Assessment a mean score of 874 (SD324) at prevention Method (RAM) calcium baseline to 1430 (SD331) mg/ Gender: 90% female checklist, a type of food day at the end of the program in Control: (n=51) Wait- frequency questionnaire the experimental group Ethnicity: 100% black (of list control (received the African descent) intervention later) In the wait-list control group, calcium intake slightly SES: Years of schooling decreased from a mean of 817.6 ranged between 8 and 20 (SD326.7 to 778.2 (SD 369.3) years (mean 13.1 yrs), and mg/day 73% had more than a high school diploma

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Bernstein Parallel US 69+ Population: Community Intervention: Follow-up: 6 months Compared to the control group, 2002 RCT dwelling elderly Personalised home- intervention and follow-up the nutrition group increased (randomly based nutrition education their intake of fruits by 1.1 (SEM assigned N=70 randomised intervention (n=38) Lost to follow-up: 0% 0.2) servings/day - from 2.8 to Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables by gender including behaviour (follow-up data obtained 3.9, p<0.01; vegetables 1.1 (SEM and age in Setting: Home-based modification techniques for all those at baseline) 0.2) from 2.3 to 3.4 servings/d, blocks of 4) designed to increase fruit, p<0.001; milk/dairy by 0.9 Gender: Intervention: vegetable and calcium- Outcomes: Self-reported (SEM 0.2) / day, from 3.0 to 3.9, 78.9% female; Control rich food consumption, FFQ and biochemical p<0.001 81.2% female delivered by home visits, blood nutrient and telephone calls and carotenoid levels No adverse effects reported Ethnicity: Predominantly monthly letters white 97.2%; 2.8% African At the end of the study, the American Control: Home-based nutrition group reported exercise intervention consuming 7.3 servings of fruit SES: Not reported but (n=32) and vegetables and 3.9 servings economically diverse of dairy products daily while to participants were targeted exercise control group reported for recruitment 6.2 fruit/veg servings and 3.1 dairy servings/day (p<0.001)

The nutrition group reported eating more citrus fruit, orange vegetables, tomatoes, other vegetables, cheese and milk compared to the control group (p<0.06)

There was no significant change in intake from other food groups 101 102 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 102 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Carcaise- RCT US 18-72 years Population: Individuals Low-intensity, physician- Follow-up: 1, 6 and 12 Fruit and vegetables (servings/d) Edinboro (reports from a rural and high endorsed dietary education months (post-intervention). in those aged 56+ yrs. 2008 separate minority population intervention versus no outcomes intervention Lost to follow-up: 18% Control (SD)/ Intervention(SD) The Rural for those N=754 randomised Baseline: 2.76(1.40)/2.77(1.48) Physician aged 56+) (N=623 provided at least Intervention: Designed to Outcome measurement: 1 month: 2.96(1.61)/3.78(1.93); Cancer one follow-up measure) improve dietary behaviour; self-reported, 5-item fruit p<0.001 between groups Prevention using tailored feedback and vegetable subscale of 6 months: 3.21(1.73)/3.72 (1.70); Setting: and self-help dietary the fibre FFB and the Food p<0.001 between groups intervention. Participants Frequency Questionnaire 12 months: 3.17(1.69)/3.33 Gender: (of those that received a mailed (FFQ) (1.69), p=0.13 between groups completed follow-up) personalised statement of 65.3% female their dietary behaviour and theory-based, low-literacy Ethnicity (%): White 60.5, nutrition information based African American 36.8, on local advice about Other 2.7 the target community in the form of 4 self- SES: 49% percent of help booklets, and the participants had less than intervention was endorsed or equal to a high school by their physician education, 24% had a Comparator: No college degree intervention

Foldi 2005 Before and US 65+ Elderly men Intervention: Video on 67% of participants increased after study osteoporosis prevention amount of calcium ingested Setting: Veteran’s Affairs and 10% began calcium Medical Centre supplementation. Significantly more met the RDA of calcium after watching the video Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Francis RCT US Range 54- N=59 randomised Intervention (N=28): Diet Follow-up: 90 days Intervention and control Mini 2009 (designed 83 education program based Nutritional Assessment Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 103 as an RCT Gender: 100% female on social marketing theory. Lost to follow-up: 1.7% scores improved (P=0.0001). but does The intervention group Intervention group consumed not report Ethnicity (%): Not received two individual Outcome measurement: more fiber than control (P=0.013) results for reported registered dietitian-led in- 3-day food records and reduced sodium intake between home education sessions and Mini Nutritional (P=0.02). Controls reduced group SES: Not reported (all assessment energy differences were literate) Control (N=30): (P=0.01) and cholesterol intakes so should Received education (P=0.029), likely because of the really be material mailings as the decreased food intake interpreted main intervention group as a before – the only difference Notes: and after between groups was the A: control group significantly study) number of contacts with lowered energy intake the dietitian B: does not report between group differences, only within group differences

Greene RCT US 60+ Population: Community- Intervention: Received a Follow-up: 12, 24 months Participants had a high intake of 2008 dwelling older adults behavior-specific fruit and fruit and vegetables at baseline Mean 74.7 vegetable manual based Lost to follow-up: 34.7% (5.2–7.4 servings), and this SENIOR (SD 6.4) Setting: Community on the Transtheoretical (24.4% did not complete increased by 1.0–2.2 servings a Project model of behaviour the study and 10.3% had day in the intervention group and (Study of Gender: 72.9% female change, newsletters, one missing data point) by 0.7–1.4 servings a day in the Exercise computer based expert control group and Ethnicity (%): 79.5% system reports, and Outcome measurement: Nutrition White-Caucasian; 13.2% coaching calls from trained Daily servings of fruit and Absolute values differed by in Older Hispanic-Portuguese counsellors vegetables assessed assessment tool used Rhode using brief food frequency Islanders) SES: 19.5% had <12 yrs Control: Received either a instruments e.g. NCI Fruit Between-group differences education; 38.5% were manual about exercise or and Vegetable Screener were significant at the 12 and NB: high school graduates; a fall-prevention manual, and the 5 A Day Program 24 month assessments (for all Comparison 19.9% were college neither of which included screener; and dietary assessment measures used) is exercise graduates information about nutrition recall telephone interviews versus in a subset of participants nutrition 104 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 104 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Kristal RCT US 18-69 Population: Participants Intervention (N=440): Follow-up: 3, 12 months Reports outcomes separately for 2000 recruited were enrollees of a tailored, multiple- those age 55-69 but reports consumer-owned health component self-help Lost to follow-up: 13.5% Puget separate maintenance organisation intervention designed completed both 3 and 12 Fruit and veg: increased by 0.61 Sound outcomes to lower fat intake month follow-ups servings/d (p<0.001) Eating for 55+ N=1459 randomised and promote fruit and Fat-related diet habits: (summary Patterns overall; N=440 completed vegetable consumption. Outcome measurement: score): decreased by 0.10 Trial aged 55-69 Consisted of a computer Telephone based surveys. P<0.01) generated personalised Change in fat intake Note: not clear if the above Setting: Community letter, motivational measured with fat-related are based on between group (home-based) phone-call, self-help diet habits questionnaire; comparison manual, computer- change in fruit and Gender: 50.9% male generated behavioural vegetable intake measured For all participants (i.e age 18 to feedback based on FFQ with 6 item FFQ 69) Ethnicity (%): 85.9% analysis, newsletters and white; 4.5% black; 5.8% supplementary materials) Fruit and vegetables (svg/day): Asian; 3.0% Hispanic; 0.8% other Control: No intervention The intervention group (n=601) increased servings from 3.62 SES: overall a range of (SD 1.49) by 0.41 (SD 1.88) at 3 household incomes months and by 0.47 (SD 1.83) at 12 months

The control group (n=604) increased servings from 3.47 (SD 1.41) by 0.08 (SD 1.63) at 3 months and by 0.14 (SD 1.80) at 12 months

Between-group difference p<0.0001

Also reports cost analyses of the programme in 1998 costs (but not cost-effectiveness outcomes) Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Lara 2015 RCT (pilot UK 50-84 Healthy men and women Intervention: Three-week Follow-up: 3 week The intervention was rated RCT) aged 50+ brief MD intervention intervention – assessed at as acceptable. No significant Mean 66 with two levels of dietary end of intervention differences were observed (SD 9) N=23 advice: between groups 1 and 2. Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables Lost to follow-up: 0%. Analysis of the combined sample Mean BMI 27 at baseline Group-1: Attended a 2h showed significant increases educational group session Outcome measurement: from baseline in fish intake (P = Setting: Community on the MD Self-reported, 3 day food 0.01) and MD score (P = 0.05) diaries Gender: Not reported Group-2: Attended a The cost of food intake during 2h group session and intervention was not significantly Ethnicity: Not reported received additional different from baseline support. Additional SES: Not reported support was provided during telephone calls on days 3, 11 and 16 of the intervention, and providing participants with additional written materials including MD menus and recipes 105 106 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 106 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Moynihan RCT UK 65-85 years Older adults living in Intervention: Peer-led Follow-up: 1 year Immediately following the 2006 (cluster: 32 Mean 76 sheltered housing in community based food intervention no significant groups of 9 (range 71- socially deprived areas clubs (N=97) Lost to follow-up: changes in diet were observed (Food people) 80) N=201 at baseline; Standards Diet at baseline high in Control: Did not attend a N=94 immediately after After 1 year, percentage of Agency saturated fat, low in fibre food club intervention and N=72 at energy from carbohydrate was N09015) and fruit and vegetables one year after intervention significantly greater (2.4% more and low in vitamin D than people who Details Outcome measurement: had not attended a club) based on At baseline 76% of questionnaire summary subjects were overweight Changes in other dietary report only or obese and only 0.5 % of outcomes or blood nutrient levels – full report subjects were underweight were not significantly different requested between groups, though there from FSA Gender: 14% male were trends towards higher 22-3-2016. fruit and sugar intakes in the No pub- Ethnicity (%): Not intervention group. At 1 year, lished paper reported significantly higher Vit D in the found. control group SES: Not reported No significant changes in measurements of body fatness

No effect on other aspects of diet, or on knowledge, attitudes or physical health

Costs Training of one peer educator: £700; average cost of running a Food Club £130 per week. This cost could be reduced if peer educators did more than one Food Club: if each peer educator did 5 Food Clubs the average weekly cost per club would be approx £70 (2006 costs) Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Moynihan RCT UK 65-85 years Older adults living in Intervention: Peer-led Follow-up: 1 year Immediately following the 2006 (cluster: 32 Mean 76 sheltered housing in community based food intervention no significant groups of 9 (range 71- socially deprived areas clubs (N=97) Lost to follow-up: changes in diet were observed (Food people) 80) N=201 at baseline; Standards Diet at baseline high in Control: Did not attend a N=94 immediately after After 1 year, percentage of Agency saturated fat, low in fibre food club intervention and N=72 at energy from carbohydrate was N09015) and fruit and vegetables one year after intervention significantly greater (2.4% more and low in vitamin D than people who Details Outcome measurement: had not attended a club) based on At baseline 76% of questionnaire summary subjects were overweight Changes in other dietary report only or obese and only 0.5 % of outcomes or blood nutrient levels – full report subjects were underweight were not significantly different requested between groups, though there from FSA Gender: 14% male were trends towards higher 22-3-2016. fruit and sugar intakes in the No pub- Ethnicity (%): Not intervention group. At 1 year, lished paper reported significantly higher Vit D in the found. control group SES: Not reported No significant changes in measurements of body fatness

No effect on other aspects of diet, or on knowledge, attitudes or physical health

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Taylor-Davis RCT US 60-75 Population: Medicare Intervention 1: Home- Follow-up: Not clear The only significant difference 2000 recipients selected based educational (approx. 3 months) between groups was one food Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 107 Mean 69.4 randomly from a Medical intervention. Five mailed, behaviour subscale ‘avoid fats’ +/- 3.2 Centre patient database biweekly nutrition Lost to follow-up: 19.6% where those who received newsletters based on the newsletters only (Intervention Note: self-rated interest in nutrition communication Outcome measurement: 1) performed significantly better nutrition was significantly model and adult learning Self-reported, food (p<0.05) than controls higher for participants than theory behaviour scale: 81 non-participants item, 5-point Likert scale, There were no significant Intervention 2: Nutrition with food behaviour differences between any of Setting: Home-based newsletters as intervention subscales:modify meat; the groups for all other food 1 with telephone follow-up avoid fat; replace fat; behaviour scales examined – Gender: 57% male substitute fat; increase ‘replace fat’, ‘substitute fat’ and Ethnicity (%): 100% white Control: No intervention fibre ‘increase fibre’ (completed pre and post- SES: Approx half had test surveys only) at least a high school education; a third had received training or education beyond school 108 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 108 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Patterson RCT (40% US 50-79 Population: Post- Intervention: Group Follow-up: 5 years Mean difference in total energy 2004 randomised menopausal women dietary intervention intake between the intervention (Women’s to int; 60% enrolled from clinical targeted at reducing Lost to follow-up: 10% and control women for total fat: Health to control) centres with baseline fat total dietary fat (to 20% was 10.9 % at Year 1, decreasing Initiative intake >/=32% of energy of energy), reducing Outcome measurement: to 9.0 at Year 5; saturated Dietary consumption saturated fat (to 7% of Food Frequency fat 4.0 at Year 1 and 3.5% at Modification energy) and increasing Questionnaire Year 5. The difference in fruit Trial) N=19542 randomised to fruit and vegetable and vegetable intake between intervention group and servings (to five or more intervention and controls was N=29294 to control group daily) and grain servings -1.2 servings per day at 1 Year (to six or more daily). and -1.3 servings per day at 5 Setting: Clinic Each group received 18 years and for grains it was -0.8 sessions in the first year and -0.5 servings/day at Year 1 Gender: 100% female to promote dietary and and Year 5 behaviour change and Ethnicity (%): 82.8% subsequent quarterly All of these differences white; 9.9% African maintenance sessions were statistically significant American; 3.3% Hispanic; (P<0.001) 2.3% Asian/Pacific Control: Usual diet Islander; 0.4% American Indian/Alaskan Native SES: 21.7% <12 years education; 39.6 13-15 years; 10.7% 16 years; 28.0% >16 years Table 25: Diet – Interventions to increase uptake/maintenance of healthy diet behaviours (non-RCTs)

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Hermann Before and US 55+ The project was marketed Eight session food and Follow-up: 8 weeks Significant increases in the

2000 after study to the public via the nutrition promotion intervention, assessed at number of servings consumed Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 109 Mean 69 media including television programme. The end of intervention from the Food Guide Pyramid Oklahoma (SD 8) and radio public service programme was designed food groups were achieved. ‘Healthy announcements, news to provide clear information Lost to follow-up: Not Statistically significant increases Living’ releases and flyers that participants could clear. N=76 completed were found for mean servings program understand and apply programme but number from the: bread, cereals, rice Volunteers were recruited to their daily lives. Used recruited not reported. and pasta group increased from to the programme nutrition education 88.1% completed post-test 3.8 to 4.9/d; vegetable group theories including nutrition recalls increased from 2.7 to 3.4/d; milk, N=67 completed pre and information, nutrition yoghurt and cheese increased post dietary recalls promotion and behaviour Outcome measurement: from 1.4 to 2.3/d change strategies. Focus self-reported, 24hr dietary was on the total diet rather recall questionnaire Non-significant changes were than individual foods (analysed by registered seen in the fruit group – mean dietitian) servings increased from 2.4 to 2.8/d and the meat, poultry, fish, dry beans, egg and nut group – increased from 1.9 to 2.0/d

The mean servings/d from fats, oils and sweets significantly decreased from 2.7 to 1.8/d 110 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Keller Before and Canada Mean age: Relatively healthy Intervention: (Surveys Follow-up: 3 years Those at nutritional risk dropped 2006 after study 72 .4 yrs members of a seniors completed by 247 at from 56.7 to 38.5%; (chi-squared (however, at baseline recreation centre baseline and 251 at follow- Lost to follow-up: N/A = 16.0 p<0.001); however, there (Evergreen note: not and 74 +/- up) Community health (different participants) was no difference in nutritional Action same 7.4 yrs at Setting: Community education programme risk for those that did or did not Nutrition) people follow-up) recreation centre for older developed involving Outcome measurement: participate in the programme assessed adults researchers, nutrition SCREEN assessment, at baseline educators and healthy self-reported validated 57.4% consumed few fruits and and follow- Gender: 69.1% female older adults. Programme tool for overall nutritional vegetables at baseline compared up) aimed at: improving risk and frequency of to 42.7% at follow-up (chi- Ethnicity (%): Not fruit and vegetable consumption of individual squared = 12.2, p=0.001) reported consumption; promoting food items positive attitudes towards Frequency of eating: at baseline SES: 45.8% had post- eating, nutrition and 6.6% ate fewer than 3 times/day secondary education health; overcoming compared to 3.3% at follow-up cooking difficulties (chi squared =3.2, p=0.05) and making cooking interesting; nutrition and Of those who participated in the food information in relation programme, 37.3% compared to disease. Emphasis was to 53.7% who did not participate on activities that were had low consumption of fruit and social, relevant, and user- vegetables (chi-squared =4.0, friendly p=0.045) Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Rousset Non-ran- France 65-75 Healthy elderly people Intervention (N=35): Follow-up: 2 weeks after For control group participants,

2006 domised living at home Nutrition information intervention protein consumption (by body Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 111 controlled program targeting protein mass unit) in the second survey study 47% of elderly people consumption in elderly Lost to follow-up: was 6 g/day less than baseline (53% of women and 40% people. Education Attrition from intervention (p=0.01 for women; p=0.10 for of men) had a body mass messages about protein group: 27.6%; from control men) index <24.5, and only two delivered by nutrition 2.0%. 96 randomised, men and one woman were professionals (physician, 13 withdrew (12 from The control group consumed less obese (body mass index registered dietitian and intervention (as did not protein from meat (P=0.03) and >30) scientist) in a lecture and want to or have time to dairy products (P=0.006) than the discussion. Participants participate in the lecture; 1 participants in the intervention Setting: Community were informed of their from control) group actual protein intake Gender: 54.9% female based on their baseline Outcome measurement: However, protein intake in the survey and the level of 7 day food questionnaire, intervention group was higher Ethnicity (%): Not their nutrition and sensory quantities self-assessed (P=0.02) at follow-up than reported knowledge; the role of using photographs baseline. Protein consumption proteins in muscle and by body mass unit increased SES: Not reported bone maintenance and the significantly in women (P=0.02) immune system, and given and slightly but not significantly protein details for different (P=0.35) in men. Fish foods and menu ideas consumption was higher in the second dietary survey (P=0.03). Control (N=47): No Less hard cheese (P=0.01) and intervention but completed more cottage cheese (P=0.06) baseline questionnaires were consumed on food consumption and attitudes 112 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Table 26: Diet – Interventions to increase uptake/maintenance of healthy diet behaviours (secondary analyses)

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Yates RCT US 50-69 Women from two rural Intervention: Tailored Follow-up: One year The tailored newsletter group 2012 (randomi- counties in Nebraska, newsletter. 18 mailed intervention with follow-up was successful in improving sation by Mean 58 English speaking and tailored or standard at 18 and 24 months their healthy eating behaviour Secondary geographic (SD 5.5) able to use a computer to newsletters were sent to compared to the standard analysis of area - two complete online surveys. women every 2 weeks Lost to follow-up: N=225 newsletter group during the Wellness for counties to Those in the maintenance for the first 6 months and randomised one-year intervention, at the Women trial intervention stage for consumption of every 4 weeks for the next end of the intervention, and or control) fat, fruits and vegetables, 6 months during the follow-up phase. (Hageman and grain intake as Family support at the end of 2014 is measured by the Stages The content of the the intervention was positively main paper) of Healthy Eating tailored newsletters was associated with healthy eating Questionnaire individualized in relation at the end of the intervention. (Initial study to: Personal goals, most Perceived barriers had the examined Setting: Community current assessment of strongest impact on healthy physical benefits, barriers, self- eating behaviour at all time activity and Gender: 100% female efficacy, and interpersonal points. Compared to participants healthy support, and biomarker in the standard newsletter group, eating Ethnicity (%): Not results regarding those in the tailored newsletter behaviours reported (check) eating. The daily dietary group perceived more family – only components emphasized support and fewer barriers for eating SES: Not reported (check) in the intervention were: 2 healthy eating at the end of the behaviours servings of fruit; 3 servings intervention (mediation effects). reported of vegetables (1 dark Based on these findings, both here) green or deep yellow); 6 family support and perceived servings of grains (3 whole barriers should be central grains); <30% of calories components of interventions from fat; and <10% of focused on healthy eating calories from saturated fat behaviour in rural midlife and older women Control: Standard newsletter. The standard newsletters contained general information about healthy eating that is currently available from organisations such as the American Heart Association and American Cancer Society Table 27: Diet – Interventions to increase uptake/maintenance of healthy diet behaviours with cognitive outcomes

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Bayer- RCT (ran- US Mean: N=49 older adults of which Intervention 1: high– Follow-up: 4 weeks Cognitive test outcomes Carter 2011 domised intervention N=29 had amnestic mild saturated fat/high– intervention post-baseline Delayed visual memory improved by dietary 67.6; con- cognitive impairment glycemic index diet Lost to follow-up: Not for both groups (healthy and interven- trol 69.3 and N=20 were healthy (HIGH): (fat, 45% reported (appears to be aMCI) after completion of the tion) controls [saturated fat,_25%]; 0%) LOW diet (p=0.04). In the HIGH

carbohydrates, 35%-40% group differences were not sig- Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 113 Setting: Veterans Affairs [glycemic index,_70]; and Outcome measurement: nificant. No diet-related changes Medical Center clinical protein, 15%-20%) Diet composition were observed for immediate research unit measured by self-reported memory, executive or motor Intervention 2: diet with 3-day food intake record speed domains Gender: 53.1% female a low– saturated fat/low– Biomarker outcomes (demen- glycemic index diet (LOW) tia-related) Ethnicity: Not reported diet (fat, 25%; [saturated CSF Aβ42: For the aMCI group, fat,_7%]; carbohydrates, the LOW diet increased CSF SES: Mean educational 55%- 60% [glycemic Aβ42 concentrations (contrary to level across groups: 13-15 index,_55]; and protein, the pathologic pattern of lowered years 15%-20%) CSF Aβ42 typically observed in Alzheimer disease). Authors Both interventions were discuss a ‘tipping point’ process conducted in a group to explain the results. For healthy of participants with MCI adults, the LOW diet decreased and a group of healthy CSF Aβ42 whereas the HIGH participants diet increased CSF Aβ42 CSF apolipoprotein E (has a role Note 1: all food was in Aβ clearance) concentration supplied by the research was increased by the LOW diet team and menus were and decreased by the HIGH diet designed by a research for both groups F2-isoprostanes nutritionist (quantitative bimarker of CNS free radical injury that are elevat- Note 2: The difference ed in AD patients). Overall, the between diets is not just LOW diet reduced and the HIGH in GI – it is also in fat and diet increased CSF F2-isopros- CHO content. Protein tane concentrations. Both groups content was the same showed lowered F2-isoprostane between diets. Energy concentrations with the LOW diet content of diets was but only healthy adults showed designed to maintain pre- increased concentrations with the study weights HIGH diet (healthy control vs a MCI group change scores in the HIGH condition p=0.04) Other dementia biomarkers No differential effects were observed for CSF Aβ40, tau-protein or p-tau 114 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Martinez- RCT Spain 55-80 Older adults at high risk Intervention 1: MedDiet Follow-up: 6.5 years Cognitive tests Lapiscina (conduct- of cardiovascular disease supplemented with extra intervention and follow-up After adjusting for multiple 2013 J Nutr, ed in a Mean: (i.e. presence of type- virgin olive oil (EVOO) comparisons, there were no Health, randomly MedDi- 2 diabetes,or at least 3 Lost to follow-up: statistically significant differences Aging selected et+EVOO: major risk factors such as Intervention 2: Advice on MedDiet+EVOO: 4.2% between the groups (p<0.05) sample 67.2 (5.6); hypertension, dyslipidemia, Mediterranean diet and MedDiet+Nuts: 7.8% Randomly from a co- MedDi- smoking and obesity) provision of mixed nuts Low fat: 6.3% MCI selected hort study) et+nuts: Odds ratio for MCI was subsample 67.3(6.0); N=285 randomly selected Control: Advice on low-fat Outcom measurement: significantly lower in the 67.5 (5.7) of which N=271 agreed to diet Global cognitive MedDiet+EVOO group compared PREDIMED- participate functioning using the Mini- to control (0.34, 95% CI 0.12 to NAVARRA Participants in all groups Mental State Examination 0.97). OR for the MedDiet+Nuts study Setting: Delivered received intensive (MMSE) and a Spanish compared to control was not in primary care but education to increase version of the Clock significant participants community- adherence to their Drawing Test (CDT) dwelling allocated intervention Dementia Note: did not measure Only 5 cases of dementia Gender: 44.8% male cognition at baseline reported so difficult to make (original trial was between group comparisons Ethnicity: Not reported designed for CVD disease outcomes) Note: Cognition not measured at SES: Mean baseline years education: MedDiet+EVOO: 8.9 (2.0); MedDiet+nuts: 8.6 (2.8); 8.7 (3.3) Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Martinez- RCT (ran- Spain 55-80 As above As above Follow-up: 6.5 years Cognitive tests

Lapiscina domisation intervention and follow-up Mean MMSE and CDT scores Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 115 2013 J done prior N= 522 were significantly higher for Neurol to agree- Lost to follow-up: Not participants allocated to a Neurosurg ment to N=268 underwent clear (randomisation done MedDiet pattern supplemented Psychiatry participate) comprehensive cognitive prior to agreement to with either EVOO or mixed nuts, assessment to identify MCI participate) compared with the control group PREDIMED- or dementia NAVARRA Outcome measurement: MedDiet+EVOO vs control study MMSE, clock drawing (adjusted) test (CDT). Incidence of MMSE: +0.62 95% CI +0.18 to (different dementia and MCI +1.05, p=0.005 participants CDT +0.33 (95% CI +0.003 to analysed) 0,67, p=0.048)

MedDiet+Nuts vs control (adjusted) MMSE: +0.57 (95% CI 0.11 to 1.03), p=0.015 CDT: +0.33 (95% CI +0.003 to +0.67), p=0.048

Dementia incidence (cases from n=268) MedDiet+EVOO: 12 MedDiet+Nuts: 6 Low fat: 17

MCI incidence (cases from n=268 MedDiet+EVOO: 18 MedDiet+Nuts: 19 Low fat: 23 116 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Valls-Pedret RCT Spain 55-80 Older adults at high risk Intervention 1: MedDiet Follow-up: 4.1 years In multivariate analyses adjusted 2015 JAMA of cardiovascular disease supplemented with extra (median) for confounders, those in the Mean: (i.e. presence of type-2 virgin olive oil (EVOO). MedDiet plus olive oil group Subco- MedDi- diabetes, or at least 3 Received quarterly Lost to follow-up: scored significantly better on the hort of the et+EVOO: major risk factors such as education sessions and MedDiet+EVOO:18.1% RAVLT (p=0.049) and Color Trail PREDIMED 67.9 (5.4); hypertension, dyslipidemia, received supplemental MedDiet+Nuts: 23.8% Test part 2 (p=0.04), compared to study MedDi- smoking and obesity) foods at no cost Control: 34.5% control group et+nuts: Different 66.7 (5.3); N=447 randomised Intervention 2: MedDiet Outcome measurement: No between-group differences sub-cohort 65.5 (5.8) supplemented with mixed Validated FFQ collected for other cognitive tests were of the above Setting: Delivered nuts. Received quarterly during a face-to-face found (MMSE, Paired associates, 2 studies in primary care but education sessions and interview with a trained verbal fluency, digit span forward, which did participants community- received supplemental dietitian digit span backward, colour trail include dwelling foods at no cost test part 1) baseline cognitive Gender: 52.1% female Control: Advice to reduce test evalua- dietary fat (personalised tion Ethnicity: Not reported advice and group sessions at the same frequency Much SES: Mean as the MedDiet groups better study, years education: after 3 years but for the reports full MedDiet+EVOO: 6.8 (3.0); first 3 years just received nutrient MedDiet+nuts: 7.6 (3.3); a yearly visit and low fat and energy 7.1 (2.8) advice leaflet) intake analysis Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Diet Components Crews 2008 RCT (paral- US 60+ Healthy, cognitively intact Intervention: Dark Follow-up: 6 weeks Cognitive tests lel design) older adults (MMSE 24+) chocolate and cocoa No significant group (dark Mean: in- (N=51). 37-g dark Lost to follow-up: chocolate and cocoa or placebo)- tervention: N=101 randomised chocolate bar and 8 Intervention: 11.7%; by-trial (baseline, midpoint, and 68.8 (8.6); ounces (237 mL) control: 8.0% end-of-treatment assessments)

control: Mean MMSE score at of an artificially sweetened interactions were found Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 117 68.7 (8.0 baseline: intervention: 28.5 cocoa beverage Outcome measurement: (1.42); control 28.6 (1.4) Selective Reminding Test Other outcomes Control: Placebo products Wechsler Memory Scale-III Also no sig diff for other Gender: 57.8% (similar) (N=50) Wechsler Adult Intelligence Haematological or blood Scale-III pressure variables examined Ethnicity: NR Digit symbol Trail Making However, the midpoint and SES: Mean years of Stroop Color-Word Test end-of-treatment mean pulse education: intervention rate assessments in the dark 15.5 (2.7); control 15.2 chocolate and cocoa group were (2.8) significantly higher than those at baseline and significantly higher than the midpoint and end-of treatment rates in the control group

Nadeem Data from UK (North- Trials Participants from the Intervention: Both studies Follow-up: 8 weeks Serum amyloid A (SAA) 2014 2 x RCTs ern Ireland) recruited FAVRIT study were aimed to increase fruit and (FAVRIT); 16 weeks Concentration of SAA in HDL3 combined 40-65 and hypertensive (systolic vegetable consumption (ADIT) decreased in the FAVRIT cohort (secondary 65-85 blood pressure (BP) range (p=0·049) and those in HDL2 analysis) years 140–190 mmHg; diastolic In the FAVRIT (Fruit and Lost to follow-up: NR in and HDL3 decreased in the ADIT BP range 90–110 mmHg), Vegetable Randomised this paper cohort (P=0·035 and 0·032) aged 40-65 Intervention Trial) participants were Outcome measurement: Other biomarkers In the ADIT (Ageing and randomised to receive a Determination of serum The concentrations of hsCRP, Dietary Intervention Trial) 1-, 3- or 6-portion F&V/d amyloid A concentrations. IL-6 and E-selectin were older subjects (65–85 intervention for 8 weeks The concentrations of unaffected by increasing F&V years) SAA in serum, HDL2 and intake in both studies (P.0·05 for In the ADIT (Ageing and HDL3 were determined all comparisons) Dietary Intervention Trial using an ELISA procedure older subjects (65–85 (KHA0011; Invitrogen Life The authors concluded that years) were randomised Technologies these results indicate that SAA to receive a 2- or 5-portion responds to increased F&V F&V/d intervention for 16 intake, while other inflammatory weeks markers remain unresponsive, and this leads to changes in HDL2 and HDL3, which may influence their antiatherogenic potential 118 Evidence-basedresourceforchangingriskbehavioursinolderadults:Intervention Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Drinks/juices delivering nutritional components which could be part of usual diet Bookheimer RCT US Mean: in- Non-demented, older Intervention: Follow-up: 4 weeks After 4 weeks, only the 2013 tervention: right-handed subjects with Pomegranate juice (N=15) pomegranate group showed a 63.1 (8.0); self-reported age-related and instruction in low Lost to follow-up:12.5% significant improvement in the control memory complaints. Mean polyphenol diet. (Restrict Buschke selective reminding 62.0 (7.8) MMSE in both groups at their intake of several fruits Outcome measurement: test of verbal memory (total baseline: 28 (1.5). and vegetables, onions, Buschke selective items recalled and consistent tea, chocolate, and dried reminding test of verbal longterm retrieval). Furthermore, N=32 randomised beans, for one week prior memory test; fMRI scans compared to the placebo group, to the baseline visit andfor during cognitive tasks the pomegranate group had Gender: Intervention: the duration for the study). increased fMRI activity during 73.3% female; control verbal and visualmemory 76.9% female Control: Placebo juice tasks; plasma metabolites of (N=13) and instruction in pomegranate juice biomarkers low polyphenol diet Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Desideri RCT (paral- Italy High/ Elderly people with MCI Cocoa flavonol Follow-up: 8 weeks Cognitive tests: 2012 lel design) Int/Low: consumption: a cocoa 71.2±4.9 N=90 randomised drink was consumed once Lost to follow-up: In MMSE Cocoa, Cog- 65 – 80 daily for 8 low flavonols group, one At the end of the follow-up nition and 71.3±4.5 Setting: Conducted via weeks at 3 different levels: person discontinued period, Mini Mental State Aging (Co- 65 - 82 clinic but participants living intervention; intermediate: Examination was similar in the 3 Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 119 CoA) Study 71.0±4.5 in community. 990 mg (high flavanols), 2 people discontinued; treatment groups (P=0.13) 64 – 81 (N=30) high: none discontinued. (Funded by Gender: 52.2% ITT analysis so all Trail making test Mars) 520 mg (intermediate participants included The time required to complete Ethnicity: Not reported flavanols), (N=30) Trail Making Test A and Trail Outcome measurement: Making Test B was significantly SES: Not reported 45 mg (low flavanols), Cognitive function was (P_0.05) lower in subjects (N=30) assessed by Mini Mental assigned to high flavanols State Examination, Trail (38.10_10.94 and 104.10_28.73 The dairy-based cocoa Making Test A and seconds, respectively) drinks used in this B, and verbal fluency test and intermediate flavanols study were specially (40.20_11.35 and 115.97_28.35 designed so that they seconds, respectively) in were indistinguishable in comparison with those assigned taste and appearance, to low flavanols (52.60_17.97 calorically balanced, and 139.23_43.02 seconds, and contained similar respectively) macronutrient, mineral, theobromine, and Verbal fluency test caffeine content, varying Similarly, verbal fluency test significantly only in the score was significantly (P_0.05) content of cocoa flavanols better in subjects assigned to high flavanols in comparison with those assigned to low flavanols (27.50_6.75 versus 22.30_8.09 words per 60 seconds)

Other outcomes Insulin resistance, blood pressure, and lipid peroxidation also decreased among subjects in the high-flavanol and intermediate-flavanol groups. Changes of insulin resistance explained _40% of composite z score variability through the study period 120 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 120 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Kean 2015 RCT UK 67 yrs Healthy, older adults Intervention (N=37 Follow-up: 8 weeks Global cognitive function was crossover): Flavanol significantly better after 8-wk 60–81 y N=37 enriched orange juice Lost to follow-up: consumption of flavanone-rich (mean 6 (100% orange juice). High- 2.7% (1 person from low juice than after 8-wk consumption SD age: flavanone (305 mg) 100% flavonone arm) of the low-flavanone control. No 66.7 6 5.3 orange juice significant effects on mood or y) Outcome measurement: blood pressure were observed Control (N=37 crossover): 8 weeks An equicaloric low- flavanone (37 mg) orange- flavored cordial (500 mL)

(Flavonone was naturally occurring in the drink)

Krikorian Non-ran- US Mean 76.2 Older adults with early Intervention N=9: Wild Follow-up: 12 weeks Significantly improved paired 2010 domised (± 5.2) memory changes blueberry juice intervention and follow-up associate learning (V-PAL) controlled performance, F(1,13) = 5.58, p study N=9 randomised. Control N=7: Non- Lost to follow-up: = 0.03, although improved recall (non-simul- blueberry matched juice on the CVLT for the blueberry taneous Gender: 55.6% male. Outcome measurement: juice subjects did not achieve control) statistical significance, F(1,13) = Ethnicity: NR 2.27, p = 0.12

SES: Mean (± SD) educational level was 15.6 (± 1.5) years Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Krikorian RCT US Mean: 78.2 Older adults with memory Intervention N=5: 100% Follow-up: 12 weeks Item acquisition across 2010 Br J (SD 5) decline but not dementia Concord grape juice intervention and follow-up learning trials: (California Verbal Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 121 Nutr with daily consumption Learning test) for subjects in N=12 randomised (very between 6 and 9 ml/kg Lost to follow-up: None the Concord grape juice group small trial) reported compared to those receiving Control N=9: Placebo placebo Gender: 66.7% male beverage that contained Outcome measurement: (F(1, 8)=5·55; P=0·04; Cohen’sf no juice or natural California Verbal Learning =0·28) Ethnicity: NR polyphenol but was Test was administered to formulated to look and assess verbal learning and Delayed verbal recall and SES: Mean years taste like grape juice retention, and the Spatial spatial memory: No statistically education: 14·1 (2·9) and to have the same Paired Associate Learning significant differences between carbohydrate and energy Test was used to evaluate grape juice subjects and control composition non-verbal memory. for delayed verbal recall (P=0·10; Mood also measured as a Cohen’s f = 0·33) and spatial potential covariate memory (P=0·12; Cohen’s f=0·67)

Adverse effects: Greater frequency of urination and aversion to the taste of juice or placebo that developed over time reported (no. of participants affected not reported) 122 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 122 Evidence-based Tables

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Mastroiaco- RCT Italy Elderly individuals without Cocoa flavonol Follow-up: 8 weeks MMSE vo 2015 clinical evidence of consumption: a cocoa The changes were not different Cocoa, cognitive dysfunction drink was consumed daily Lost to follow-up: In Cognition for 8 wk: low flavonols group, two Trail making test (TMT) and Aging N=90 randomised people discontinued Mean changes (6SEs) in the (CoCoA) Containing 993 mg [high intervention; intermediate: time required to complete the Study Setting: Conducted via flavanol (HF)] (N=30) 1 person discontinued; TMT A and B after consumption clinic but participants living high: one person of the HF (28.6 6 0.4 and 216.5 (Funded by in community 520 mg [intermediate discontinued. ITT analysis 6 0.8 s, respectively) and IF Mars) flavanol (IF)] (N=30) so all participants included (26.7 6 0.5 and 214.2 6 0.5 s, Gender: 58.9% female respectively) drinks significantly 48 mg [low flavanol (LF)] Outcome measurement: (P , 0.0001) differed from that Ethnicity: Not reported (N=30) Cognitive function was after consumption of the LF assessed at baseline drinks (20.8 6 1.6 and 21.1 6 0.7 SES: Not reported and after 8 wk by using s, respectively) the Mini-Mental State Examination (MMSE), the Verbal fluency test (VFT) Trail Making Test (TMT) Similarly, VFT scores significantly A and B, and the Verbal improved among all treatment Fluency Test (VFT) groups, but the magnitude of improvement in the VFT score was significantly (P ,0.0001) greater in the HF group (7.7 6 1.1 words/60 s) than in the IF (3.6 6 1.2 words/60 s) and LF (1.3 6 0.5 words/60 s) groups

Other outcomes Significantly different improvements in insulin resistance (P , 0.0001), blood pressure (P , 0.0001), and lipid peroxidation (P = 0.001) were also observed for the HF and IF groups in comparison with the LF group. Changes in insulin resistance explained w17% of changes in composite z score (partial r2 = 0.1703, P , 0.0001) Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Mastroiaco- RCT Italy Elderly individuals without Cocoa flavonol Follow-up: 8 weeks MMSE vo 2015 clinical evidence of consumption: a cocoa The changes were not different Cocoa, cognitive dysfunction drink was consumed daily Lost to follow-up: In Cognition for 8 wk: low flavonols group, two Trail making test (TMT) and Aging N=90 randomised people discontinued Mean changes (6SEs) in the (CoCoA) Containing 993 mg [high intervention; intermediate: time required to complete the Study Setting: Conducted via flavanol (HF)] (N=30) 1 person discontinued; TMT A and B after consumption clinic but participants living high: one person of the HF (28.6 6 0.4 and 216.5 (Funded by in community 520 mg [intermediate discontinued. ITT analysis 6 0.8 s, respectively) and IF Mars) flavanol (IF)] (N=30) so all participants included (26.7 6 0.5 and 214.2 6 0.5 s, Gender: 58.9% female respectively) drinks significantly 48 mg [low flavanol (LF)] Outcome measurement: (P , 0.0001) differed from that Ethnicity: Not reported (N=30) Cognitive function was after consumption of the LF assessed at baseline drinks (20.8 6 1.6 and 21.1 6 0.7 SES: Not reported and after 8 wk by using s, respectively) the Mini-Mental State Examination (MMSE), the Verbal fluency test (VFT) Trail Making Test (TMT) Similarly, VFT scores significantly A and B, and the Verbal improved among all treatment Fluency Test (VFT) groups, but the magnitude of improvement in the VFT score was significantly (P ,0.0001) greater in the HF group (7.7 6 1.1 words/60 s) than in the IF (3.6 6 1.2 words/60 s) and LF (1.3 6 0.5 words/60 s) groups

Study Study Country Age Population Intervention Relevant outcomes and Results design (years) and comparator follow-up Ongoing trials – published protocols and registered clinical trials Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 123 Feasibility RCT US Older adults with mild Intervention: Ketogenic, Follow-up: 12 week trial Ongoing and Efficacy cognitive impairment modified Atkins diet (MAD). of Dietary (MCI) or early Alzheimer’s Outcomes: Achieving Interven- disease (AD) living in the Control: Non-ketogenic a Healthy Eating Index tions for community control diet based on (>=85) at 3 consecutive Cognitive the National Institute on follow-up visits Impairment Also to determine the role Aging’s recommendations in Older of apolipoprotein E (ApoE) for senior nutrition. Efficacy for Cognition Adults genotype in participants’ Composite Memory Score response to the MAD (mean z-score on delayed recall of HVLT-R and BVMT-R) Efficacy for Function Change in MDS-HC Instrumental Activities of Daily Living Score

Hawaii De- RCT US 65+ Older adults with MCI Nutrition intervention to Nutrition intervention to Ongoing? – no further details mentia Pre- prevent cognitive decline prevent cognitive decline found vention Trial (HADEPT) Knight 2015. RCT Australia 65+ Healthy, older adults with Intervention: Follow-up: 6 months PROTOCOL ONLY – Results not BMC Geri- normal cognitive function Mediterranean diet. Cretan intervention yet published atrics and proficient in English MedDiet (i.e. vegetables, language fruits, olive oil, legumes, Outcomes: MedLey fish, whole grain cereals, Cognitive function, study N=166 randomised nuts and seeds and low psychological wellbeing consumption of processed foods, dairy products, red meat and vegetable oils)

Control: Usual lifestyle and diet 124 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 124 Evidence-based Tables

Table 28: Diet – Included qualitative studies about barriers and facilitators

Study Study Country Age Population Objective Quality design (years) Qualitative studies in older adults Drummond 2006 Qualitative Australia Not reported N=50 Older community-dwelling World War II and Vietnam Ageing men’s (semi- male war veterans living in both urban and rural areas understanding of nutrition structured and implications for health interviews and Gender: 100% male focus groups) SES: Not reported

Ethnicity: Not reported

McKie 2000 Individual UK (Scotland) 75+ (47% aged N=152 Random sample of older adults from patient lists To examine food semi-structured 75-79) of GPs in 3 areas from rural and urban areas. 49% were consumption patterns and interviews and widowed, 54% lived in lone households perceptions of dietary 24 hour dietary advice of older adults recall Gender: 68% female

SES: 49% reported income below social security benefit rate (1997)

Ethnicity: Not reported

Pettigrew 2012 Individual Australia 40+ but N=111 Mid-life and older adults in both metropolitan and To explore the diet-related interviews stratified by regional areas beliefs and behaviours of (N=20) and age group midlife and older adults focus groups Gender: Male and female included (ratio not reported) to inform future healthy (N=12) eating interventions SES: Not reported targeting this group

Ethnicity: Not reported Study Study Country Age Population Objective Quality design (years) Qualitative studies conducted in participants of dietary interventions in older adults Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 125 Hammarstrom Thematic Sweden Mean 60, N=12 Middle-aged and older women participating in dietary To explore barriers and 2014 structured range 49 to 71 weight loss interventions (from both intervention groups facilitators to weight interviews (for the total including drop outs from the dietary interventions) loss in older adults intervention participating in a diet group) Gender:100% female intervention

SES: Not reported other than overall ‘quite highly educated’

Ethnicity: Not reported other than ‘all Swedish born’

Hyland 2007 Individual UK (England) Mean age for N=20 Peer educators’ semi-structured intervention perceptions of training Peer Led Food interviews participants Gender:18.2 % male and implementing a Club Study and group was 76, range community-based nutrition discussions 71-86. Peer SES: ‘varied in background’ intervention for older educators 60+, adults though age not Ethnicity: Not reported reported other than ‘slightly younger’

Moynihan 2006, Focus groups UK (England) Mean 76 N=97 Community living older adults completed interviews Attitudes towards eating 2007 and semi- (range 71-80) N=160 Completed dietary behaviour questionnaire more healthily and structured perceived barriers to Peer-led food club in-depth Gender: 14% male healthy eating, attitudes study interviews and towards participation in questionnaire SES: Not reported peer-led food club for dietary behaviour Ethnicity: Not reported 126 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 126 Evidence-based Tables

Study Study Country Age Population Objective Quality design (years) Penn 2008 Individual UK (England) Mean age 64 N=15 Participants in the European Diabetes Prevention Study To understand the semi-structured (range 47 to (EDIPS) who maintained behaviour change over 3-5 years. experience of those PA, diet interviews 74) participants who Gender: 53.3% male maintained behaviour change in the European SES: 5 (33.3%) were or previously employed in professional Diabetes Prevention or related occupations; 1 (6.7%) administrative; 2 (13.4%) Study, an RCT of diet and skilled trade; 6 (40%) process operatives or elementary exercise interventions occupations in people with impaired glucose tolerance Ethnicity: Not reported

Sandison 2008 Qualitative – UK (Scotland) 56-65, mean N=61 Older women who were participants in the dietary arm To investigate the questionnaire age 61 of an RCT to increase fruit and vegetable intake. experiences of women Note: The primary using open and who increased fruit and RCT is not some closed Gender: 100% female veg consumption as referenced so questions part of an RCT aiming not clear which SES: Not reported to increase fruit and veg primary trial this is consumption for bone a sub-study of Ethnicity: Not reported health Table 29: Diet – Quality assessment for interventions reporting on uptake/maintenance of healthy diet

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 127 RCTs Appleton No No No High Yes Un- Un- Un- Yes Yes Yes Low Un- Yes Un- No No Un- High 2013 clear clear clear clear clear clear Atienza Yes Un- Yes Un- Yes Un- Un- Un- Yes Yes Yes Low Un- Yes Un- No Un- Un- Unclear 2008 clear clear clear clear clear clear clear clear clear Babatunde Un- Un- Yes Un- Yes Un- Un- Un- Yes Yes Yes Low Un- Yes Un- No Un- Un- Unclear 2011 clear clear clear clear clear clear clear clear clear clear Barr 2000 Un- Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Un- Yes Un- Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear clear clear Bernstein Un- Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Un- Yes Yes Un- Unclear 2002 clear clear clear clear clear clear clear clear clear clear clear Carcaise- Un- Un- Un- Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Un- Un- Un- Un- Unclear Edinboro clear clear clear clear clear clear clear clear clear clear clear clear clear clear 2008 Francis Un- Un- Un- Un- No Un- Un- Un- Yes Un- Un- Un- Yes Yes Un- Yes Yes Un- Unclear 2009 clear clear clear clear clear clear clear clear clear clear clear clear Greene Un- Un- Un- Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Un- Un- Un- Un- Unclear 2008 clear clear clear clear clear clear clear clear clear clear clear clear clear clear (by group) Kristal Yes Un- Un- Un- Un- Un- Un- Un- Un- Un- Un- Un- Yes Yes Un- Un- Un- Un- Unclear 2000 clear clear clear clear clear clear clear clear clear clear clear clear clear clear clear (by (by (by group) group) group) Lara 2015 Yes Yes Yes Low Yes Yes Yes Low Yes Yes Yes Low Yes Yes Un- Un- Un- Un- Low (brief clear clear clear clear int) Moynihan Yes Un- Yes Low Yes No No Un- Un- Un- Un- Un- Yes Yes Un- Yes Yes Un- Unclear 2006 clear clear clear clear clear clear clear clear (by cluster) Patterson Un- Un- Un- Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Un- Un- Un- Un- Unclear 2004 clear clear clear clear clear clear clear clear clear clear clear clear clear clear 128 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 128 Evidence-based Tables

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Rousset Un- Un- Un- Un- Yes Un- Un- Un- Yes No No Un- Yes Yes Un- Un- Un- Un- Unclear 2006 clear clear clear clear clear clear clear clear clear clear clear clear Taylor- Un- Un- Yes Un- Yes No Un- Un- Yes Un- Un- Un- Un- No Un- Un- Un- Un- Unclear Davis 2000 clear clear clear clear clear clear clear clear clear clear clear clear clear Non-randomised studies Hermann N/A N/A N/A High N/A N/A N/A High N/A N/A N/A High Un- Yes Un- No N/A Un- High 2000 clear clear clear Keller 2006 N/A N/A N/A High N/A N/A N/A High N/A N/A N/A High Yes No No No N/A Un- High clear Table 30: Diet – Quality assessment for interventions reporting on diet with cognitive outcomes

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Bayer- Un- Un- Un- Un- Yes Yes Un- Low Yes Yes Yes Low Un- Yes Yes Yes Yes Low Unclear Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 129 Carter 2011 clear clear clear clear clear clear Un- Un- No Un- Yes Yes Un- Low Yes Un- Un- Un- Un- Yes Yes Yes Yes Low Unclear Bookheim- clear clear clear clear clear clear clear clear er 2013 Crews Yes Yes Yes Low Yes Yes Yes Low Yes Yes Un- Low Un- Yes Yes Yes Yes Low Low 2008 clear clear Desideri Yes Yes Yes Low Yes Yes Un- Low Yes Yes Yes Low Un- Yes Yes Yes Yes Low Low 2012 clear clear Kean 2015 Yes Yes Un- Low Yes Yes Yes Low Yes Yes Yes Low Un- Yes Yes Yes Yes Low Low clear clear (shown for cognition but not de- mog etc.) Martinez- Un- Yes Un- Un- Yes No Yes Un- Yes Un- Un- Un- Yes Yes Yes Yes Yes Low Unclear Lapiscina clear clear clear clear clear clear clear 2013 J (ran- (cog- (only Nutr, donly nition those Health, se- not who Aging lected meas- partic- from ured ipated an at ana- RCT base- lysed) but line) method for 2nd nr) 130 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 130 Evidence-based Tables

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Martinez- Un- Yes Un- Un- Yes No No Un- Yes Un- Un- Un- Yes Yes Yes Yes Yes Low Unclear Lapiscina clear clear clear clear clear clear clear 2013 J (ran- (cog- (only Neurol donly nition those Neurosurg se- not who Psychiatry lected meas- partic- from ured ipated an at ana- RCT base- lysed) but line) method for 2nd nr) Mastroia- Yes Yes Yes Low Yes Yes Un- Low Yes Yes Yes Low Un- Yes Yes Yes Yes Low Low covo 2015 clear clear Nadeem Yes Un- Yes Un- Yes No No Un- Yes No Un- Low Un- Yes Yes Yes Yes Low Unclear 2014 clear clear/ clear clear clear Low Valls- Yes Un- Yes Un- No No No High Yes No Un- Un- Yes Yes Yes Yes Yes Low Unclear Pedret clear (small clear/ (?) – clear clear 2015 JAMA diffs Low slight in age differ- and ences lipid in FU values time but unlikely to favour results) Non-randomised studies Krikorian N/A N/A Un- High Yes Yes Un- Low Yes Un- Un- Un- Un- No Un- Un- Un- Un- High 2010 clear clear clear clear clear clear clear clear clear clear Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Martinez- Un- Yes Un- Un- Yes No No Un- Yes Un- Un- Un- Yes Yes Yes Yes Yes Low Unclear Lapiscina clear clear clear clear clear clear clear 2013 J (ran- (cog- (only Neurol donly nition those Neurosurg se- not who Psychiatry lected meas- partic- from ured ipated an at ana- RCT base- lysed) but line) method for 2nd nr) Mastroia- Yes Yes Yes Low Yes Yes Un- Low Yes Yes Yes Low Un- Yes Yes Yes Yes Low Low covo 2015 clear clear Nadeem Yes Un- Yes Un- Yes No No Un- Yes No Un- Low Un- Yes Yes Yes Yes Low Unclear 2014 clear clear/ clear clear clear Low Valls- Yes Un- Yes Un- No No No High Yes No Un- Un- Yes Yes Yes Yes Yes Low Unclear Pedret clear (small clear/ (?) – clear clear 2015 JAMA diffs Low slight in age differ- and ences lipid in FU values time but unlikely to favour results) Non-randomised studies Krikorian N/A N/A Un- High Yes Yes Un- Low Yes Un- Un- Un- Un- No Un- Un- Un- Un- High 2010 clear clear clear clear clear clear clear clear clear clear

Table 31: Diet – Quality assessment for interventions reporting barriers and facilitators

1. Theoretical 2. Study 3. Data 4. Validity 5. Analysis 6. Ethics Overall approach design collection 1.1 1.2 2.1 3.1 4.1 4.2 5.1 5.2 5.3 5.4 6.1 6.2 Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 131 Drummond 2006 Appropriate Clear Not sure Appropriate Un- Not Not Not Not sure Adequate Not Not - clear sure sure re- re- reported ported ported Hammarstrom 2014 Appropriate Clear Defensible Appropriate Clear Not Rich Relia- Convincing Adequate Yes Clear ++ sure ble

Hyland 2007 Appropriate Clear Defensible Appropriate Clear Relia- Rich Not Convincing Adequate Yes Clear ++ ble sure

McKie 2000 Appropriate Clear Defensible Appropriate Clear Not Rich Not Convincing Adequate Not Not + sure re- re- reported ported ported Moynihan 2006/2007 Appropriate Clear Defensible Appropriate Clear Relia- Rich Not Convincing Adequate Yes Clear ++ ble sure

Penn 2008 Appropriate Clear Defensible Appropriate Clear Relia- Rich Not Convincing Adequate Yes Clear ++ ble sure

Pettigrew 2012 Appropriate Clear Defensible Appropriate Clear Relia- Not Relia- Convincing Adequate Yes Clear ++ ble sure ble

Sandison 2008 Appropriate Clear Defensible Not sure Un- Not Not Not Convincing Adequate Yes Not + clear sure sure re- clear ported 132 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 132 Evidence-based Tables

Table 32: Physical Activity – Interventions to improve uptake/maintenance of physical activity

Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Asikainen 28 RCTs in NR 50-65 Postmenopausal Nine studies Not reported PA uptake and PA uptake / Maintenance / 2004 total; RCTs with women aged 50 used walking; Maintenance: Mean AE: Walking (mean drop out sufficient quality to 65 years; four studies drop out, mean was 13%, mean attendance in more than 25 2632 women in used combined attendance. Other four studies was 84%. Mean participants and total. Healthy, aerobic exercise Physiological and injury rate reported in six studies less than 35% of sedentary or had (walking + QoL measures was 3%). Combined Aerobics drop outs some leisure flexibility + relevant to PA (Mean dropout rate was 12%; PA at entry into one of cycling, uptake (Short Attendance rate reported in one study; swimming, term): Health-related study was 77% in home based Special dance); Nine fitness (bodyweight; exercise and 53% in group Population: studies used proportion of body fat based exercise; Incidence of Women with combined of total bodyweight injury in this study varied from diseases or risk resistance (F%); bone mineral 23% (high intensity exercise factors such as training and density (BMD); bone and 13% for low intensity dyslipidaemia, aerobics; two mineral content exercise). Combined aerobics hypertension, studies used (BMC); various and resistance training (Mean obesity or resistance tests on muscle dropout rate was 15% and osteoporosis training with performance, mean attendance was 67%. weight machines. flexibility, balance Mean attendance was higher Five other and coordination; in exercise groups with more studies used maxi- mal oxygen aerobic component compared other resistance consumption with resistance training; Mean training of five to (VO2max); resting injury rate was 6%). Resistance nine exercises blood pressure (BP); training with weights (mean total cholesterol dropout rate was 16% and mean (TC); high-density attendance was 90%; Injury lipoprotein-cholesterol; rate was 33%). High impact low-density resistance training (mean lipoprotein-cholesterol; attendance rate was 68%, triglycerides; blood injury rate: 8%). One single glucose and insulin). resistance, back extensor Adverse Outcome: or jumping exercise (mean Injury Rates attendance: 91%, injury rate: 2-3% Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Clegg 2012 6 RCTs in total Interna- The medi- Frail older adults; Home based NR PA uptake and PA uptake and Maintenance: tional (3 in an age 987 participants. exercise. One Maintenance: Median completion rate Western was 83 The majority of intervention Completion and reported in six studies was Europe, 2 years participants were included a single adherence rates 83% (65%-88%). Median in USA and (range 78- female (median component of Other Physiological adherence rate reported in 1 in New 88) 79% female, progressive and QoL measures three studies was 78% (66%- Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 133 Zealand) range 50–88%). resistance relevant to PA 89%). Other Physiological Three of the exercise. Two uptake (Short term): and QoL measures relevant trials recruited combined Measures of mobility to PA uptake (Short term): less than 100 progressive (TUG), HRQoL (EQ- One high-quality trial reported subjects; only resistance 5D) and ADL (Barthel improved disability in those with two recruited exercises with Index), muscle moderate but not severe frailty. more than 200 one or more strength, balance, Meta-analysis of long-term subjects. additional depression, bone care admission rates identified Special components strength a trend towards reduced risk Population: of flexibility, (pooled risk ratio, 0.89; 95% Frail OAP; balance, walking confidence interval, 0.55–1.45). limited mobility or range of Improved gait speed was requiring the motion exercises. reported in one trial, a trend use of a walking Two interventions towards improved gait speed aid; sedentary; were complex was reported in one further trial, in receipt of interventions and gait speed did not improve home care and combining in two. Improvements in ADL housebound but multiple exercise were reported in one trial; no able to get out of components with improvements in ADL were chair and bed an occupational reported in the other three trials. intervention. One Three trials measured muscle study used an strength using upper and lower electronic device body strength or grip strength. that counted the One trial reported improved lower number of sit-to- body strength. There was no standsstands improvement in either upper or lower body strength in one trial. No improvement in grip strength was recorded in the study that measured this outcome. No improvement in general physical performance was reported in one trial. Improved balance was reported in one trial but there was no effect on balance in three trials. There was no effect on depression, bone density or flexibility 134 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 134 Evidence-based Tables

Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Chase 2013 20 RCTs Interna- Study age Community- Five studies Controls or usual PA uptake and PA uptake and Maintenance: tional ranges dwelling adult employed care control Maintenance: Two cognitive-based were from subjects aged 60 behavioural groups PASE, pedometer interventions using Motivational 66.30 to years or older. interventions and accelerometer, Interviewing strategies 81.70 years 3148. Sample only; twelve modified 7-day activity demonstrated success in size ranged studies combined interview, activity significantly improving PA from 33 to 966. cognitive and diary, leisure-time behaviour among participants. Two studies had behavioural PA questionnaire All studies using ‘supervised 100% female interventions expressed in energy exercise sessions alone’ participants; expenditure units, demonstrated non-significant one study had 7-day PA recall findings in differences in PA no female instrument, Baecke PA behaviour between treatment participants. scale, Yale PA survey, and control groups at outcome. Three studies Auckland Heart Study Inconsistent success in did not report PA questionnaire, increasing PA behaviour was gender. In other Modified CHAMPS, observed among studies studies female Flemish PA combining Cognitive-Behavioural participants computerised Interventions; however most made up questionnaire studies using combination between 16.0% cognitive-behavioural and 72.0% of the interventions reported successful sample long-term findings. One cognitive-behavioural–based intervention demonstrated long-term PA behaviour change results, with evidence of continued higher levels of PA from baseline up to 2 years beyond the end of a study. There were no significant changes in PA outcomes from post-test to follow-up period in three studies (Kelly, 2004, Talbot, 2003, Bird, 2011) Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Con 2003a 43 primary studies NR Mean Community- Unspecified in NR PA uptake and PA uptake and Maintenance: participant dwelling paper Maintenance: Overall Interventions targeted to PA ages in the individuals PA and episodic exclusively are more effective primary 60 years or older. exercise than those targeting multiple studies 33,000 aging behaviour behaviours. Studies without ranged adults health education were more from 60 to effective than those reporting 77.2 years that they taught health benefits. Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 135 MA findings strongly support the importance of some self-monitoring system to increase adults’ PA. Meta-analysis revealed that the intensity of the intervention, in terms of contact time between the activity professionals and elders, is important. Intense direct contact with staff more than doubled the effect size of the interventions. In contrast, mailed or telephone interventions made no difference in outcomes. Although staff contact time is expensive, the profound effect on elders’ physical activity behaviours makes this an important aspect of programming. Elders who ex- ercised at centres as compared to home based activity, were much likely to continue PA. The most effective part of an activity prescription is making specific intensity recommendations. OAP were more likely to increase their PA when recommendation was for moderate intensity activity than when low intensity activity was suggested. Interventions delivered to groups were consid- erably more effective than that delivered to individuals. Interven- tions that recommend walking are probably more effective than programme without a walking suggestion. Walking is easily accessible and may be perceived as ‘natural 136 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 136 Evidence-based Tables Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Con 2003b 17 RCTs NR Mean 6,391 subjects Walking. Also NR PA uptake and Main- PA uptake and Maintenance: subject age 6/17 studies tenance: Overall PA 7/10 studies with theory based of 65 and looked at overall and episodic exercise intervention reported positive older PA behaviour (<= 6 findings. 5/7 studies which used months post-test). Ex- social cognitive framework ercise maintenance (> reported positive results. 2/3 6 months post-test) studies that used TTM reported positive outcomes. 4/5 that used combined models reported significant treatment effects. 3/5 studies with supervised centre based exercise reported positive treatment effect. 7/12 studies without supervised exercise reported greater exercise in treatment group than in the control group. Further, 4/6 studies with individualised interventions reported greater exercise in the treatment groups than the control groups. Interventions delivered to individuals were about equally likely to result in positive (6/11) and negative findings (5/11). Non-PA intervention used included motivational strategies and behavioural change techniques such as social support, stimulus control, self-regulation, health education etc. Self-monitoring and health education were most commonly used in studies. Mixed results on the association between these BCTs and PA uptake. 3/5 studies that delivered interventions in subjects’ homes reported positive results. Each of the four studies that delivered interventions in aggregate community sites such as senior centres or churches reported more exercise by experimental subjects than control Maintenance: 5/17 studies with >=6 months follow-up reported significance maintenance of PA after end of intervention Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Cyarto 2004 8 intervention (5 Internation- The age of Participants aged Progressive NR PA uptake and PA uptake and Maintenance: All RCTs, 2 qua- al (4 USA, participants 60+. Sample Resistance Maintenance: the general practice interventions Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 137 si-experiments, 1 1 Australia, ranged sizes ranged Training Questionnaires produced some positive impact pre-post) 2 UK, 1 from 40 from small (Walking, measuring PA, on PA levels. Study that used Belgium) to over 90 studies with strength training, exercise logs. Other community mass media years, with only 17 and 20 flexibility, Physiological and communication reported approx- participants, to balance and co- QoL measures 23% increase in walking and imately one with 719. ordination) relevant to PA proportion of those achieving 30 even Most studies had uptake (Short term): minutes of activity a day. Most representa- between 75 and Accelerometers, of the studies reported positive tion of the 300 participants heart rate monitors, results. Maintenance: CHAMPS ‘young’ old measurement of II study reported maintenance (mean age CV-risk factors and of FU after 12 months. A 50-60) the direct observation of notable feature of this study ‘mid old’ participants was the long-term involvement (mean age of the local community in the 60-70), and project through a local advisory the ‘older committee old’ (mean age over 70)

de Vries 3/18 RCTs relevant NR The age of Community- Physical exercise No exercise, low- PA uptake and PA uptake and Maintenance: 2012 to PA uptake & the study dwelling older therapy (Strength intensity exercise Maintenance: PASE, Three studies evaluated their maintenance population adults with training, balance self-reported PA, intervention on the level of varied from impaired training, YPAS, FAI, number physical activity. Two of these 60 to 85 mobility, physical functional and of walks. Other studies were pooled in a meta- years disability and/or task-related Physiological and analyses, which showed no multi-morbidity. training, QoL measures exercise on PA level (SMD: 0.08, Special endurance and relevant to PA uptake 95% CI: -0.21, 0.31, I2: 0%) Population: Frail mobility training (Short term): WS None of these studies found and with mobility timed chair stands, a significant effect. However, problems SF-36, IADL, TUG, results show that physical 6MWT, 400MWT exercise therapy has a positive effect on mobility 138 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 138 Evidence-based Tables

Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Fairhall 15 RCTs Interna- Aged 60+ 3,616 Exercise The effects of PA uptake and PA uptake and Maintenance: 2011 tional participants interventions that the intervention Maintenance: The pooled estimate of the aimed to reduce were compared Participation in life role effect of interventions including falls in older with placebo, measured by Scales exercise indicated a small adults (strength, alternate therapy with ICF components improvement in participation balance, Tai Chi) or usual care such as Adelaide (Hedges’ g = 0.16, 95% Activity Profile, PASE, confidence interval = 0.04–0.27, Older American’s P = 0.006). Meta-regression Resources and showed multifactorial intervention Services, Nottingham with an exercise component Extended ADL Index, had a larger effect than exercise Lawton’s IADL Scale, intervention alone, but the Late Life Function and difference was not statistically Disability Index, The significant (effect on Hedges’ g = Groningen Activity 0.22, 95% CI = −0.05 to 0.50, P Restriction Scale, = 0.10) Frenchay Activities Index, Falls Handicap Inventory

French 2014 16/25 relevant NR The overall Community- Lifestyle PA NR PA Uptake and PA Uptake and Maintenance: intervention stud- mean dwelling adults (gardening, Maintenance: (Pooled ES) BCT Interventions ies (1 feasibility, 1 age of 60 years or walking); Change in Physical had a small effect on PA pre-post, 1 cluster partici- over. The mean Exercise Activity measured in (d=0.14, 95% CI 0.09, 0.2, RCT, 13 RCTs) pants was number of (Aerobic class, ‘d’ Cohen ES p<0.001). Effect size ranged 69 years participants in gym, jogging), from d= -0.02 to 0.63. (study the comparisons others (Individual results) 3 BCTs were means included in the significantly associated with ranged self-efficacy higher PA behaviour: ‘barrier from 60 to analysis was identification / problem solving’, 84 years) 247 (range 5 to ‘provide rewards contingent on 1,011); the mean successful behaviour’, ‘model number included / demonstrate behaviour’. 10 in the physical BCTs were associated with lower activity analysis PA behaviour. The greatest was 349 difference in effect size occurred when the following BCTs were present: ‘provide normative information about others’ behaviour’, ‘provide information on where and when to perform behaviour’ and ‘plan social support / social change’ Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Geraedts 9/24 relevant Interna- 55+ 5,328 Remote Exercise or Non- PA Uptake and PA Uptake and Maintenance: 2013 studies (8 RCTs & tional participants aged feedback on Exercise Maintenance: Results show that PA programs 1 pre-post) 55+ home-based PA Walking speed, with frequent and non-frequent (Phone, Text) strength, balance, remote feedbacks are equally peak VO2, 7- day PA as effective in enhancing recall, accelerometer, physical capacity measures as

6 MWT, TUG, supervised exercise without Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 139 adherence rate, remote feedback. Maintenance: compliance rate Adherence to interventions using remote feedback was higher in the control groups in studies where intervention groups were compared to TAU. Adherence to interventions using remote feedback seems mostly acceptable-to-good, with rates in intervention groups varying between 32.1 and 91%.One study compared text messaging to a phone strategy and found that texting led to a significantly higher adherence than phone

Hobbs 2013 21 intervention Interna- The mean 10,519 at risk The majority of Received usual PA Uptake and PA Uptake and Maintenance: studies (3 cluster tional age of adults aged 55 to interventions care; exercise Maintenance: Five Interventions to promote physical RCTs, 2 pre-post (USA, participants 70 years. ‘At were multimodal alterative; some trials used pedometers activity were effective at 12 and 16 RCTs) Europe, was 60.7 risk’ participants and provided studies included deriving step-count months (standardized mean NZ, Japan, years (SD were reported physical activity information and one trial used an difference (SMD) = 1.08, 95% Australia, = 4.4; as having at and lifestyle leaflet and accelerometer deriving confidence interval (CI) = 0.16 Canada) range 55 to least one of the counselling newsletters vector magnitude. to 1.99, pedometer step-count, 67.6) following disease Twenty trials estimated approximating to an increase risk factors: PA duration by self- of 2,197 steps per day; SMD = hypertension, report questionnaires 0.19, 95% CI = 0.10 to 0.28, self- impaired glucose reported as minutes reported physical activity duration tolerance, of PA or energy outcome), but not at 24 months overweight/ expenditure. Four based on a small subset of trials. obese, trials assessed PA Further analysis by O’Brien hyperlipidaemia, using both objective (2015) shows that increasing dyslipidaemia, and self-report the number of BCTs in PA family history, methods promoting intervention does not metabolic enhance long term effectiveness. syndrome or Interventions aiming to promote osteopenia PA should consider using BCT feedback in order to enhance effects 140 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 140 Evidence-based Tables

Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Muller 2014 16 interven- Interna- 50+ Healthy, Non-face to face NR PA Uptake and PA Uptake and Maintenance: tion-studies (2 tional community- PA Maintenance: Of the 16 studies, 14 reported cluster RCTs, 2 (USA, Neth- dwelling older Self-reported significant improvements in PA quasi-experiments, erland, Aus- adults (≥ 50 questionnaires / over the respective study periods 1 non-randomised tralia, New years) instruments, (1 week to 24 months). Only CT & 11 RCTs) Zealand accelerometer, one reported a non-significant weekly time spent in decrease of PA in terms of daily PA, weekly energy calorie expenditure and time expenditure spent in moderate or greater PA over the previous week. Maintenance: PA levels were maintained after the intervention stimulus was removed in all but one study

Neidrick 8/11 relevant RCTs Interna- 50+ NR Standardized NR PA Uptake and PA Uptake and Maintenance: 2012 tional intervention to Maintenance: 7/10 studies found that PA (USA, promote PA Self-reported promotion intervention was Australia, questionnaires / effective in increasing PA. Canada, instruments, PASE, 1/10 found that generalised Europe, Physical Activity Recall health behaviour modification England) (PAR), Active Australia without PA component was not Physical Activity effective. 1/10 studies found that Questionnaire, Dutch supplementing verbal advice Short Questionnaire with written advice did not show to Assess Health a significant effect on PA uptake. Enhancing Physical Maintenance: Limited evidence Activity (SQUASH), to show effect on long term Pedometer adherence Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Muller 2014 16 interven- Interna- 50+ Healthy, Non-face to face NR PA Uptake and PA Uptake and Maintenance: tion-studies (2 tional community- PA Maintenance: Of the 16 studies, 14 reported cluster RCTs, 2 (USA, Neth- dwelling older Self-reported significant improvements in PA quasi-experiments, erland, Aus- adults (≥ 50 questionnaires / over the respective study periods 1 non-randomised tralia, New years) instruments, (1 week to 24 months). Only CT & 11 RCTs) Zealand accelerometer, one reported a non-significant weekly time spent in decrease of PA in terms of daily PA, weekly energy calorie expenditure and time expenditure spent in moderate or greater PA over the previous week. Maintenance: PA levels were maintained after the intervention stimulus was removed in all but one study

Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Nigg 2012 14/18 relevant NR 55+ NR Single Health NR PA Uptake and PA Uptake and Maintenance: RCTs Behaviour Maintenance: Of the 12 SHBC studies Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 141 Change Self-reported evaluating PA or exercise, Interventions questionnaires / participants generally improved (SHBC) and instruments, their level of activity at FU (6-12 Multiple Health accelerometer, months). MHBC showed mixed Behaviour weekly time spent in results; one study found that the Change PA, weekly energy combination of PA and fruit and Interventions expenditure vegetable consumption improved (MHBC) only nutritional outcomes but not PA behaviour at FU. The other showed improvement in both PA and weight loss behaviour

Stevens 6 RCTs (5 RCTs & Internation- Five of the Adults aged 50 Tailored PA NR PA Uptake and PA Uptake and Maintenance: 2014 1 Cluster RCT) al (USA, studies and above (aerobic, strength Maintenance: PASE; 2/6 studies reported a statistically NZ, Austral- reported and balance time to reach target significant increase in physical ia, Canada, the mean exercises) of >= 90 mins / week activity levels. Two studies UK) age of par- of MVPA; Auckland showed no significant increase in ticipants, Heart Study Exercise activity which Questionnaire; ranged frequency and from 65 to duration of walking 74; four and vigorous exercise recruited a greater number of females 142 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 142 Evidence-based Tables Study Study design Country Age Population Intervention Control Relevant outcomes/ Findings (years) measures Van der Bij 2/6 studies report- Internation- Mean age Community- Home based PA, NR PA Uptake and PA Uptake and Maintenance: 2002 ed a statistically al (USA, ranging dwelling, healthy Group based PA Maintenance: Home based PA (Participation significant increase Europe) from 51-88 and inactive and Educational Participation rates rate 86%-93%). Participation in in physical activity years older adults. PA longer term interventions was levels. Two studies Large majority lower than in short-term Rx. 2 showed no signif- were white, well- studies reported a decline in PA icant increase in educated and level after end of intervention. activity had moderate to Group based PA (Mean participa- high incomes tion in short term duration Rx (< 1 year) = 84%, 55-100%). Inter- ventions in nursing or residential homes achieved high participation rates (mean= 87%). Partici- pation rate in Rx > 1 year ranged between 63-84% mean=75%). All studies reported higher PA levels than baseline and that PA levels in intervention groups were significantly higher than control. 1/3 studies achieved higher PA level at 10 year follow-up. Education PA (35%-96%). Participation rate declined with increase in intervention sessions. All educational interventions showed significant increase in PA compared to control in short term (< 1 year). Overall high participation rates are achievable with short term PA interventions (< 1 year) but not found for long-term intervention. Possible explanation for inverse relationship between the participation rate and length of intervention are lack of interest, motivation, enjoyment, time or perceived benefit. It appears group-based and education interventions were effective in increasing PA in short-term. Long-term education was ineffective. Insufficient data to show long-term effectiveness of group based interventions. Home and group based interventions appear to be comparable Table 33: Physical Activity – Interventions to improve uptake/maintenance of physical activity with cognitive outcomes

Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Angevaren 11 Randomised Con- Interna- Age range Sedentary, frail Cycling, No activities, Outcomes: Overall, the effects of aerobic

2008 trolled Trials (RCTs) tional (55-91) participants walking, jogging, stretching exer- Cognitive speed, exercise on cognitive function Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 143 with age-related strengthening cise verbal memory compared with any other illness exercise, function, visual intervention were significant for resistence, memory function, speed (SMD random effects weight training, working memory, 0.26, 95% CI 0.04, 0.48, P=0.02), aerobics executive and visual attention speed functions, (SMD random effects 0.26, perception, face 95% CI 0.02, 0.49, P=0.03). recognition, The effects of aerobic versus cognitive inhibition, no intervention were positive for auditory attention, auditory attention (WMD random motor function. effects 0.52, 95% CI 0.13, 0.91, Outcome P<0.01) and motor function Measures: Ross (WMD random effects 1.17, 95% Information CI 0.19, 2.15, P=0.02). Aerobic Processing training appears effective in the Assessment, short term for cognitive delay or Wechsler Adult prevention. Aerobic training was Intelligence Scale not superior to strength training (WAIS), Randt memory test story recall

Balsamo 8 RCTs USA, Average Mixed population Structured ADLs (bingo, Outcomes: (Cog- Although 5/8 studies showed 2013 Australia, age = 74.8 of normal cogni- physical patchwork, nitive function) Ex- higher cognitive response Brazil, years tive older adults, exercise, sewing), balance, ecutive, short-term than controls, evidence was France AD and MCI, 1 stretching, stretching memory, attention, inconclusive due to lack of study only female study strength training, long term episodic power walking, daily memory. Outcome living activities, Measures: ADAS- strength training, Cog, MMSE, cardiorespiratory Wechsler Adult training, music Intelligence Scale therapy III, Wechsler Mem- ory Scale-revised, Toulouse-Pieron’s concentration attention Test, Rey Osterrieth Com- plex Figure 144 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 144 Evidence-based Tables

Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Carvalho 10/27 RCTs Asia, USA, Older Mixed population PA (resistance Flexi-tone (tri- Outcomes: PA confers a protective effect 2014 Europe population (could not as- training, aerobic, weekly training of Cognitive status/ on cognition in elderly subjects. of age ≥ 60 certain gender strength, balance 10 mins warm- function, Brain 26/27 of all studies showed years proportion of and flexibility; up, 25-30 mins Volume. Outcome positive association between PA some studies); combination of strength, Measures: River- and cognition while 9/10 RCTs sedentary; flexibility and mead Behavioural showed a positive association independently balance, 10 min Memory Test, between PA and cognition ambulatory, living cool down); No Wechsler Adult independently intervention; Intelligence Scale, Education to Direct and Indirect improve lifestyle Digit Span, Mem- and PA; Balance ory Complaints and tone training: Scale, Cambridge stretching, Cognitive Test, range of Wechsler Adult In- motion, balance telligence Scale III, exercises, Wechsler Memory and relaxation Scale-Revised, technique; Social Toulouse-Pieron Interaction; concentration one weekly attention test, training session Ray-Osterrieth consisting of complex figure, warm-up and Freed and Cued stretching Selective Re- exercises, but no minding Test, overload training Trail Making Test, and Stroop Test, MMSE, 3MS ADAS-Cog, Brain volume using MRI, Neuropsycholog- ical battery test, Reaction time tests including simple reaction time, 8-choice reaction time, 8-choice incompatible reaction time, and Go/ No-Go reaction Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Carvalho 10/27 RCTs Asia, USA, Older Mixed population PA (resistance Flexi-tone (tri- Outcomes: PA confers a protective effect 2014 Europe population (could not as- training, aerobic, weekly training of Cognitive status/ on cognition in elderly subjects. of age ≥ 60 certain gender strength, balance 10 mins warm- function, Brain 26/27 of all studies showed years proportion of and flexibility; up, 25-30 mins Volume. Outcome positive association between PA some studies); combination of strength, Measures: River- and cognition while 9/10 RCTs sedentary; flexibility and mead Behavioural showed a positive association independently balance, 10 min Memory Test, between PA and cognition ambulatory, living cool down); No Wechsler Adult independently intervention; Intelligence Scale, Education to Direct and Indirect improve lifestyle Digit Span, Mem- and PA; Balance ory Complaints and tone training: Scale, Cambridge stretching, Cognitive Test, range of Wechsler Adult In- motion, balance telligence Scale III, exercises, Wechsler Memory and relaxation Scale-Revised, technique; Social Toulouse-Pieron Interaction; concentration one weekly attention test, training session Ray-Osterrieth consisting of complex figure, warm-up and Freed and Cued stretching Selective Re- exercises, but no minding Test, overload training Trail Making Test, and Stroop Test, MMSE, 3MS ADAS-Cog, Brain volume using MRI, Neuropsycholog- ical battery test, Reaction time tests including simple reaction time, 8-choice reaction time, 8-choice incompatible reaction time, and Go/ No-Go reaction

Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Chang 2012 10 RCTs NR OAP mean Healthy adults Resistance Health Outcomes: Designs including loads from

age ≥ 65 without cognitive Exercise (Otago, Education; Cognition. 60% to 80% 1RM, approximately Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 145 impairment or resistance Flexibility and Outcome seven movements in two sets specific disease, training and Relaxation; measures: with 2 minutes rest between sets adults with an balance, ST with balance and WAIS-R, TMT- at least twice per week for 2–12 average age aquatic exercise, toning Word-list memory months (usually 6 months), might older than 65 callisthenic test, word-list recall positively affect cognition in older years training with test, verbal-fluency adults aquatic exercise, test, modified Aerobic exercise Boston naming and diet test, constructional praxis & clock- drawing test B, Stroop CW, COWAT, WMS-R, Auditory oddball task, WAIS III, Toulouse-Pieron’s concentration attention test, Mental arithmetic, computerized mirror drawing task, Rey- Osterrieth complex-figure test 146 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 146 Evidence-based Tables Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Coelho 5 RCTS and 1 NR Mean age Older women Physical exercise Stretching Outcomes: Aerobic exercise increases 2013 non-randomised CT = 66.2 (no-frail and pre- (Resistance- exercise; Peripheral serum BDNF in older adults years frail); mean age training; Aerobic Unspecified; no and plasma BDNF = 66.2 years; OP exercise control (brain-derived neu- with MCI; elderly (treadmill, rotrophic factor) subjects with stationary, concentrations; glucose toler- bicycle, or Cognitive function; ance criteria for elliptical trainer); depression. Out- pre-diabetes or Nordic walking; come Measures: newly diagnosed; Gymnastics; Blood analysis patients with Acute aerobic (Plasma/ELISA, major depression exercise Serum/ELISA). and healthy OP (treadmill) MMSE, GDS, Spatial memory paradigm, Episodic memory performance (auditory verbal learning test), Symbol-digit modalities, verbal fluency, stroop, trails B, task switching, story recall, and list learning, HAMD (Hamilton Rating Scale for Depression), and Dem Test

Colcombe 18 intervention studies NR Young old Community- Aerobic Fitness Any Outcomes: Significant difference in overall 2003 (4 non-randomised (55-65), dwelling, supervised Cognitive Function ES (Overall ES for Intervention CTs, 1 pre-post and 13 middle-old “normal” older aerobic training, (Speed, group was 0.478 (SE=0.029, RCTs) (66-70) adult; Sedentary combined visuospatial, n=101, P<.01), and Overall ES and old-old clinical aerobic training controlled- for control group was 0.164 (71+) populations processes and (SE=0.028, n=96. P<.05). of one kind or executive-control Aerobic fitness training improved another, ranging processes). cognitive performance in the from depressed Outcome older adults and markedly in persons to Measure: MA of the executive processing of the geriatric mental ES brain. The mid-old and old-old patients and reported to have benefited more individuals with from fitness training compared cardiopulmonary with the young-old participants obstructive disorder Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Gates 2013 14 RCTs NR 65-95 Predominantly Physical exercise Any (social visits, Outcomes: Overall, exercise training had

years female, with (isolated no contact, Cognitive function minimal but positive effect on Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 147 cognitive moderate education (Executive verbal fluency (ES=0.17; 95% impairment, frail intensity aerobic programs, function, memory CI= 0.04, 0.30). Aerobic training elderly exercise, low normal and and information effective on global cognition in intensity walking, recreational processing). three studies [(ES=0.74; 95% CI: resisted training, activities, sham, Outcome 0.43, 1.05), (ES=0.56; 95%CI: combined active control) Measures: 0.19. 0.92), (ES=0.69; 95%CI: training, Tai MMSE, ADASCog, 0.03, 1. 32)]. Isolated resistance Chi, supervised CAMCOG, training produced significant aerobic training, WAIS-R. MA of ES effects on memory [(ES=3.37, and combined was performed 95%CI: 2.27, 4.74), (ES=0.54: aerobic training. 95%CI: 0.02, 1.08)]. Aerobic Aerobic training exercise training did not make up half of improve executive function studies) compared with other reviews

Ohman 8/22 relevant RCTs NR Age range: 60% female; Physical exercise Social visits or Outcomes: There were some positive effects 2014 50-86 mean MMSE (aerobic normal social Cognition; on one or several domains of score of 24 exercise, activities; Outcome cognition, global cognition, strength training, educational Measures: executive function or attention balance, dual material; MMSE, ADAS- tasking, walking, stretching; health Cog, Symbol digit, hand and face education verbal fluency, exercises, Tai Stroop and task Chi, treadmill, switching, delayed stationary recall, CDR-SUB, bicycle, elliptical Stroop Test, WMS- trainer LM 148 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 148 Evidence-based Tables

Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Patterson 12 intervention studies Interna- NR NR Physical activity NR Outcome: 58% of intervention studies 2010 (6 non-randomised tional (strengthening, Cognitive function demonstrated small positive CTs and 6 RCTs) included 1 aerobic training) (speed, visual effects on at least one measure non OECD memory, visual of cognitive function. These studies reproduction, studies employed moderate verbal memory, intensity aerobic physical activity motor function, interventions; however, it is working memory, difficult to quantify the actual executive volume of exercise used in each function, cognitive intervention inhibition and auditory attention. Outcome Measure: Varied (simple reaction time, choice reaction time, Wechsler Memory Scales, Benton visual retention test, Randt memory test, Wechsler memory scale, finger tapping, digit span tests, face recognition, verbal fluency, problem solving, word comparison, Stroop test, letter search, visual search, Digit span forward Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Sherder 8 RCTs (5 RCTs NR Age ranges NR Walking Flexibility, Outcomes: Studies suggest that walking

2014 involving normal from 55- balance, Executive improved executive functions in Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 149 cognition, & 3 with 73 years strengthening, Function; cognitively intact older persons cognitive impairment) for NC; toning, social Outcome who have lived a sedentary life 75-86 for visits, no Measures: Spatial participants treatment Word Memory, with spatial switching, cognitive Trial Making Test, impairment Stroop test, Verbal Fluency test, Digit span

Tseng 2011 12 RCTs NR Mean age Older adult Physical exercise No treatment, Outcomes: Trials showed a positive effect = 71.5 participants with (walking, stretching, Cognitive Function; for exercise on cognition when years and without treadmill running, normal Outcome the exercise regimen lasted for cognitive extremity daily activities, Measures: MMSE, 6 weeks and occurred at least impairment stretching educational WAIS III WMS-R, three times per week for 60 were 50%, exercise, weight materials, social ADAS-Cog, minutes respectively, and bearing strength visits, vitamin B CERAD, CDR, only 16.7% of training, and supplements SCWT, WCST, the trials focused swimming) AVLT, VFT, DSST on female participants 150 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 150 Evidence-based Tables

Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Uffellen 21 RCTs NR >=55 years Age of the study Physical exercise Yoga, Exercise Outcomes: This review suggests that 2008 populations (Aerobic; (strengthening, Cognition function. different kinds of exercise ranged from Strength; balance, Outcome may benefit cognitive function 55 to 94 years Flexibility; flexibility) Measures: WAIS, irrespective of baseline cognitive in cognitively Balance or a WIAS-R, TMT, status. Five out of the 15 healthy combination of Verbal fluency studies in cognitively healthy populations the above test, Stroop colour subjects observed significant and from 67 Word Tests, Visual beneficial effects on some of the to 99 years in reproduction, included measures for cognition. populations with digit span; visual Significant effects were also cognitive decline. reproduction, observed in 5 out of the 8 studies In both groups, verbal memory, in subjects with cognitive decline. the majority of verbal pairs test In cognitively healthy adults, participants were (mental status test improvements were observed women (Strub and Black), in memory (Corsis block- based on WMS tapping test, Rey – Osterrieth figure, face recognition, digit span) information processing abilities (organization, auditory processing), and executive function (word fluency). Effective interventions in this group included aerobic exercise (n = 2); strength exercise alone or combined with balance exercise (n = 1); and all-round exercise including aerobic, strength, balance, and flexibility training (n = 1) Study Study design Country Age Population Intervention Control Relevant out- Findings (years) comes/measures Uffellen 21 RCTs NR >=55 years Age of the study Physical exercise Yoga, Exercise Outcomes: This review suggests that 2008 populations (Aerobic; (strengthening, Cognition function. different kinds of exercise ranged from Strength; balance, Outcome may benefit cognitive function 55 to 94 years Flexibility; flexibility) Measures: WAIS, irrespective of baseline cognitive in cognitively Balance or a WIAS-R, TMT, status. Five out of the 15 healthy combination of Verbal fluency studies in cognitively healthy populations the above test, Stroop colour subjects observed significant and from 67 Word Tests, Visual beneficial effects on some of the to 99 years in reproduction, included measures for cognition. populations with digit span; visual Significant effects were also cognitive decline. reproduction, observed in 5 out of the 8 studies In both groups, verbal memory, in subjects with cognitive decline. the majority of verbal pairs test In cognitively healthy adults, participants were (mental status test improvements were observed women (Strub and Black), in memory (Corsis block- based on WMS tapping test, Rey – Osterrieth figure, face recognition, digit span) information processing abilities (organization, auditory processing), and executive function (word fluency). Effective interventions in this group included aerobic exercise (n = 2); strength exercise alone or combined with balance exercise (n = 1); and all-round exercise including aerobic, strength, balance, and flexibility training (n = 1)

Table 34: Physical Activity – Included qualitative studies about barriers and facilitators

Study Study design Country Age Population PA type Barriers Facilitators Limitations (years) Barnett 5 qualitative studies Interna- NR All Recreational Lack of time for Lack of time for The qualitative 2012 tional participants PA recreational PA; recreational PA; Low evidence is limited by had been Low perceived value perceived value of the small number of

retired for of recreational PA recreational PA and studies available and Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 151 between six and preference for preference for produc- the months and productive/meaningful tive / meaningful PA. limited socioeconomic 5.6 years PA Health properties of diversity of study PA motivate adoption/ participants, increase of recrea- who were mostly from tional PA but do not relatively affluent guarantee long-term backgrounds maintenance; Lifelong PA habits influence recreational PA patterns after retirement; Rec- reational PA provides a new daily routine; Recreational PA offers new personal challeng- es; Recreational PA provides opportunities for social interactions

Boehman 5 qualitative studiess USA 50+ Community Population- (Personal) Health, (Personal) Health, There were only five 2013 dwelling – based falls Lack of motivation, fa- enjoying the activity, articles that focused people living prevention tigue, time factors, lack self-motivation, body on rural locations, all independently exercise of knowledge about image, previous of which were from the in their home programs exercise, low self-effi- exercise experience, United States, four and not an cacy, feelings and per- exercise, exercise focused on women aged care ception about exercise, knowledge. (Social) only and the total facility previous exercise ex- Social support, population was 326 perience, body image, social contact, people fear. (Social) Lack of recommendation to social support, family exercise, role models. and household commit- (Environmental) ments. (Environmen- Accessible facilities tal) Poor built environ- and programs, ment, lack of access to available transport, programs and facilities, conducive built safety concerns, dogs, environment, low/ traffic, weather, lack of reasonable costs transportation, costs 152 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 152 Evidence-based Tables

Study Study design Country Age Population PA type Barriers Facilitators Limitations (years) Bunn 2008 24 studies (4 RCTs, 1 Interna- 55+ Older popu- Falls (General) Fatalism/ (General) Information Potential for survey, 1 cross-sectional tional lation prevention attributing falls to that falls can publication bias to & 18 qualitative) programme external causes/lack be preventable; affect results due to of knowledge about Communicating non-RCTs in review. effectiveness of falls life-enhancing Study quality not used prevention; Perception aspects of strategies, to weight or exclude that physical deterio- e.g. maintaining studies may also affect ration inevitable with independence and results age; Lack of relevant control; Accessible, information in appropri- appealing information ate formats/language; format, from a variety Provision of ‘one size of sources and in fits all ’ advice. Advice different languages; seen as common Choice of interventions sense/patronising; Low for different people and self-efficacy. Fear of lifestyles; High self- loss of independence/ efficacy; Personalised risk taking ability; No modifications; perception of need for Emphasis on help (no previous falls); social aspects Provoking fear of falling of interventions. by using scare tactics; (Exercise) Previous Social stigma: associa- exercise ‘habit’, Making tion with old age/frailty; exercise fun/enjoyable/ Differing agenda of sociable, Good older adults and health leadership/facilitation, professionals. (Exer- Motivation/information cise) No previous ex- about physical ercise ‘habit’; Physical and psychological discomfort/unpleasant benefits of exercise, sensations associated Programmes tailored with exercise; Under- to needs or lifestyle, lying beliefs about Convenient scheduling/ personality type (e.g. reasonable pricing/ too lazy, no willpower); good access and Self-perception: too transport. (Home old to exercise. (Home modifications / assistive modifications / as- devices) Facilitate sistive devices) Dislike feeling of ownership of of interventions seen interventions, shared as intrusive/didactic; decision making, Stigma of devices as- Referral from health- sociated with old age care professional Study Study design Country Age Population PA type Barriers Facilitators Limitations (years) Child 2012 12 qualitative studies Interna- OA Community- Falls (Practical considera- tional dwelling older prevention tion) Cost of accessing adults programme programme; ease of access to falls-prevention intervention i.e. ability to drive, availability and cost of

transport, car parking Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 153 facilities, cold weather; Time (Adapting for community) social and cultural acceptability of assistive devices, types of exercise, and fatal- istic attitudes towards falling (Psychosocial) transforming identity (independence, confidence and QoL)

Cunning- 1 qualitative, 2 cross- NR Seniors Community PA Safety of footpaths, Inconsistent findings ham 2004 sectional and 3 surveys dwelling Access and and mixed results in seniors convenience of primary studies of the facilities, proximity relationship between to services, heavy PA and environmental traffic, safety (dogs, factors crime), noise, adequate lighting, public transportation, litter

Dunsky 6/7 relevant surveys Interna- 45+ Adults and PA and sports Injury Main finding was that 2012 tional older adults the information on rate of injuries in purpose- ful physical and sports activities in advanced age was too limited for reliable conclusions. While there is some information regarding the rate of injuries, it is not presented relative to the extent of activities, either in the number of active people or intensity of the activity 154 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 154 Evidence-based Tables

Study Study design Country Age Population PA type Barriers Facilitators Limitations (years) Horne 2012 10 qualitative studies Interna- 60+ OA from PA (Communication) (Communication) tional South Asian Obtaining accurate Who provides advice; Community information; Lack of Positive reinforcement; information; Lack (Relationship) of support and Facilitative relatives; encouragement; Group, peer and Language barriers. community support; (Relationship) Instructor support. Overprotective family; (Beliefs) Collectivist Dependence on norms; (Environment) social support; Group Engaging in community norms. (Beliefs) activities; Walking Concepts of ageing; outdoors was a Lack of knowledge preferable form of PA and understanding about the benefits of exercise and keeping active; Unfamiliarity of gym-based exercise; Role of fate and lack of personal control. (Environment) Migration; Not socialised to spend time outdoors or doing sport; Lack of culturally sensitive facilities; Lack of knowledge of geographical area and facilities; Obligations to others and contribution to community activities Evidence-based resource for changing risk behaviours in older adults: Intervention Tables 155 ing High High High High High High High High Poor Good Good Good Good Good Good Good Rank - Y Y Y Y Y Y Y Y Y Y Y Y Y N N N 11 Y Y N N N N N N N N N N N N N N 10 9 Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y 8 Y Y Y Y Y Y Y Y N N U U N N U N 7 Y Y Y Y Y Y Y Y N N U N N N U N 6 Y Y Y Y Y Y Y Y Y Y Y Y Y N N N 5 Y Y N N N N N N N N N N N N N N 4 Y Y Y Y N N N N N N N N N N N N Appropriate method to combine findings; 10. Publication bias; 11. Conflict of interest Appropriate method to combine findings; 10. Publication bias; 3 Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y 2 Y Y Y Y Y Y Y Y Y Y Y Y Y U N N 1 Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Author (Year) Conn 2003a Clegg 2012 Asikainen 2004 Conn 2003b Chase 2013 de Vries 2012 de Vries Cyarto 2004 Fairhall 2011 French 2014 Geraedts 2013 Hobbs 2013 Muller 2014 Neidrick 2012 Stevens 2014 Nigg 2012 Van der Bij 2002 Van Table 35: Physical Activity – Quality assessment for interventions reporting on uptake/maintenance of physical activity 35: Physical Table AMSTAR (High= 8-11, Good= 5-7, Poor= <5) (High= 8-11, AMSTAR Key: 1. ‘a priori design; 2. duplicate study selection and data extraction; 3. comprehensive literature search; 4. Status of publication as an inclusion criterion; 5. List studies (included and excluded provided)?; 6. Characteristics of the included studies provided 7. Scientific quality assessed documented; 8. of the included studies considered in formulating conclusions; 9. 156 Evidence-based resource for changing risk behaviours in older adults: Intervention Tables ing High High High High High High High Poor Poor Good Good Good Rank - Y Y Y Y Y Y Y Y Y N N N 11 Y Y Y N N N N N N N N N 10 9 Y Y Y Y Y Y Y Y N N N N 8 Y Y Y Y Y Y Y Y Y Y Y N 7 Y Y Y Y Y Y Y N U N N N 6 Y Y Y Y Y Y Y Y Y Y N N 5 Y Y Y N N N N N N N N N 4 Y Y N N N N N N N N N N Appropriate method to combine findings; 10. Publication bias; 11. Conflict of interest Appropriate method to combine findings; 10. Publication bias; 3 Y Y Y Y Y Y Y Y Y Y Y Y 2 Y Y Y Y Y Y Y Y N U N N 1 Y Y Y Y Y Y Y Y Y Y Y N Author (Year) Angevaren 2008 Balsamo 2013 Carvalho 2014 Chang 2012 Coelho 2013 Colcombe 2003 Gates 2013 Ohman 2014 Paterson 2010 Scherder 2014 Tseng 2011 Tseng Van Uffellen 2008 Uffellen Van Table 36: Physical Activity – Quality assessment for interventions reporting on physical activity with cognitive outcomes 36: Physical Table AMSTAR (High= 8-11, Good= 5-7, Poor= <5) (High= 8-11, AMSTAR Key: 1. ‘a priori design; 2. duplicate study selection and data extraction; 3. comprehensive literature search; 4. Status of publication as an inclusion criterion; 5. List studies (included and excluded provided)?; 6. Characteristics of the included studies provided? 7. Scientific quality assessed documented; 8. of the included studies considered in formulating conclusions; 9. Table 37: Cognitive Stimulation – Systematic reviews of cognitive training interventions with cognitive outcomes

Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Gross 2012 N=35 intervention studies International 60+ Cognitively normal, Memory training Objective memory performance

(limited to English community dwelling older interventions (non- outcomes; intervention or Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 157 language) adults pharmacological) participant characteristics that influence effectiveness

Mean standardized difference in pre-post change between memory-trained and control groups was 0.31 standard deviations (SD; 95% confidence interval (CI): 0.22, 0.39)

The pre-post training effect for memory-trained interventions was 0.43 SD (95% CI: 0.29, 0.57) and the practice effect for control groups was 0.06 SD (95% CI: -0.05, 0.16)

10 distinct memory training strategies were identified from studies. From meta-analyses, training multiple strategies was associated with larger training gains (p=0.04), although not statistically significant after adjusting for multiple comparisons.

There was no difference in effect from training in any particular strategy, by age of participants, session length, or type of control condition 158 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 158 Evidence-based Tables

Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Kelly 2014 N=31 RCTs International 50+ (most Community dwelling older Cognitive training: Cognitive training Ag Res Rev A included adults without known ‘specifically designed From meta-analysis studies age cognitive impairment training programmes that Compared to active controls, >65) provide guided practice on cognitive training improved a standard set of cognitive performance on measures of tasks, aimed at improving executive function (working performance in one or memory, p = 0.04; processing more cognitive domains’ speed, p < 0.0001) and composite measures of cognitive General mental function (p = 0.001) stimulation: interventions that promote increased Compared to no intervention, engagement in mentally cognitive training improved stimulating activities. (e.g. performance on measures of reading, playing music or memory (face-name recall, p = playing chess) 0.02; immediate recall, p = 0.02; paired associates, p = 0.001) and subjective cognitive function (p = 0.01)

Data were not available for face- name recall, paired associates, verbal fluency, reasoning, or everyday functioning.

Mental stimulation Due to heterogeneity and lack of available data, meta-analysis not conducted for this outcome

Mental stimulation had significantly larger effects compared to ‘no intervention’ controls on four out of eight memory measures; nine out of 17 measures of executive function; and one out of three composite measures of cognitive function Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Kelly 2014 In individual trials, one found no Ag Res Rev A significant differences between Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 159 (continued) mental stimulation and active control groups on four measures of memory, four measures of executive function, one composite measure of cognitive function and one measure of subjective cognitive performance. One study found that acting class participants significantly outperformed singing class controls in two measures of memory and two measures of executive function. The authors concluded that ‘more research is required to determine if general mental stimulation can benefit cognitive and everyday functioning. Transfer and maintenance of intervention effects are most commonly reported when training is adaptive, with at least ten intervention sessions and a long-term follow-up. Memory and subjective cognitive performance might be improved by training in group versus individual settings.’ ‘The impact of cognitive training on everyday functioning is largely under investigated’. Transfer of training effects were reported in nine out of ten trials: five reported transfer to untrained tasks within the same domain 160 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 160 Evidence-based Tables Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Li N=20 intervention studies International NR for the People with MCI (no age Cognitive interventions People with MCI benefited of which N=17 were SR but in specified) significantly more from the included in meta-analysis all studies, cognitive intervention than MCI mean age control group in overall cognition is 55+ (Q= 16.21, p < .001), overall self-ratings (Q= 6.92, p = .009), episodic memory (Q= 13.96, p < .001), executive function/WM (Q= 5.40, p = 0.02). The effect sizes of separated domains in MCI intervention group were all larger than the MCI control group, although the differences did not reach significance, MMSE (Q= 2.47, p = 0.12), semantic memory (Q= 1.18, p = 0.28), attention/processing speed (Q= 2.27, p = 0.13), visuospatial ability (Q= 1.26, p = 0.26), language (Q= 0.49, p = 0.49), self-rated memory (Q= 3.08, p = 0.08), depression (Q= 0.08, p = 0.78), anxiety (Q= 0.02, p = 0.88), ADL (Q= 0.23, p = 0.63)

Martin 2011 RCTs (N=36) of which 24 International 60+ Healthy older adults and Cognitive training No studies reported dementia (Cochrane review) studies relating to memory (limited to English older adults with MCI. interventions targeting outcomes (though only looked for training were pooled in or German Eligible studies were from specific domains of in MCI studies) meta-analysis language) any setting cognitive functioning such as memory, attention, or No studies provided adverse speed effects information

Note: most of the From meta-analyses included studies included For healthy older adults, focused on memory immediate and delayed verbal training interventions, recall improved significantly and very few on speed (p<0.05) through training improvements or compared to a no treatment training of executive control condition functioning For individuals with mild cognitive impairment significant training gains were obtained for treatment compared to no contact control in immediate (p=0.04) recall and delayed recall (p=0.05) Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Reijnders 2013 N=35 intervention studies. International NR Healthy older adults and Cognitive training Evidence that cognitive training

(27 RCTS and 8 clinical (limited to English people with MCI interventions can be effective in improving Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 161 studies) published language) various aspects of objective between 2007 and 2012. cognitive functioning; memory N=6 studies were in performance, executive people with MCI, the rest functioning, processing speed, were in healthy older attention, fluid intelligence, and adults subjective cognitive performance

Tardif 2011 N=14 cognitive International NR for Healthy elderly Cognitive intervention Nine out of 14 studies targeted intervention programme (limited to English review but participants - 13/14 programmes memory as the principal cognitive studies with a control or French) all included studies recruited in the function to train or stimulate. group (9 RCTs, 1 studies >55 community, 2 recruited Face-name associations, mental controlled study, 2 from retirement homes imagery, paired associations, and quasiexperimental, 2 used (1 of these recruited from the method of loci were the main a within subject crossover both) techniques taught to participants. design) Improvements were observed on at least one outcome measure in each study included

Teixeira 2012 N=7 intervention studies International 60+ Community-dwelling older Non-pharmacological Cognitive function, including adults with MCI interventions including aspects of memory or executive cognitive interventions function (N=6) and 1 PA intervention All 6 included studies reported at least one significantly improved measure of cognitive function. All 6 reported improved episodic memory and one study also reported improved executive function

The authors noted ‘contradictions and divergences’ in the available studies, so it was not possible to establish a systematic protocol for effective cognitive training interventions 162 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 162 Evidence-based Tables

Table 38: Cognitive Stimulation – Systematic reviews of cognitive training compared to physical activity

Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Hindin 2012 N=42 intervention studies International 55+ (mean Healthy community living Cognitive tasks or aerobic Untrained cognitive outcomes (25 reported extended (restricted to 69.2 y) older adults (included exercise including choice reaction time, cognitive practice English language) interventions which memory, executive function. interventions and 17 took place in a range Both extended cognitive practice aerobic interventions) of settings including and aerobic interventions laboratories, gymnasiums, produced significant effects but Extended practice in the home and outdoors they did not differ in magnitude. cognitive training is Better study quality was completing many trials associated with larger effect of basic tasks such sizes as phoneme span or choice response time Extended practice effect size with or without strategy (ES) =0.33, 95% CI 0.13 , 0.52; instruction (hundreds Aerobic intervention ES = 0.33, to thousands). Only 95% CI 0.10 , 0.55) strategies likely to generalise to other untrained outcomes were included

Karr 2014 N=46 intervention studies International 65+ Healthy people, people Compares cognitive Executive function, participant (23 PA, 21 cognitive (limited to English (mean age with dementia or MCI training versus PA and intervention characteristics training (CT) and 2 both language studies) 74 (5.7) for that produce the best outcomes PA and CT) PA studies Studies of PA had mean and 73.3 MMSE 27.1 (2.2) and CT (4.3) for CT studies had mean MMSE Both PA and cognitive training 26.0 (3.9) significantly improved executive functions but no significant Gender: PA: 30.4% male; difference in effect size between CT: 40.9% male the two interventions

Note: some evidence of publication bias – for CT more recent publications produced lower effect sizes Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Law 2014 N=8 RCT or non- International 60+ Older adults with Combined cognitive and Cognitive outcomes randomised trials (5 in (but limited to and without cognitive exercise interventions Overall, the results of the Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 163 people with cognitive English language impairment (included reviewed studies in both impairment and 3 in publications) studies in people with populations with and without cognitively healthy dementia but reported cognitive impairment were people). Of the 3 in outcomes separately conflicting. Among the three cognitively healthy people, for cognitively healthy studies with cognitively healthy 1 was an RCT and 2 were participants so only those participants, one study found no non-RCT interventions outcomes are reported) effects while two studies revealed significant effects. Fabre et al. (2002) reported a significant effect on memory performance (d = 1.29). Oswald et al. (2006) found significant effects, with within-group effect sizes reported, on general cognitive functions (d+ = 1.14), subjective rating of cognitive impairment (d+ = 0.59) and functional status (d+ = 0.27), which were sustained at 5-years follow-up 164 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 164 Evidence-based Tables

Table 39: Cognitive Stimulation – Systematic reviews of computerised cognitive training interventions

Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Bleakley 2013 N=12 interventions (5 International (only 65+ Older adults Interactive computer Cognitive function (or physical RCTs, 2 controlled trials English language games (involving a outcomes). Secondary outcomes: and 5 other studies studies) physical component compliance, enjoyment and (observational) e.g. aerobic, strength, adverse effects balance, or flexibility components) Only 2 studies reported cognitive outcomes, the rest reported physical outcomes. Both studies were non-controlled (before and after studies, Rosenberg 2010, Studenski 2010)

Both studies found positive but non-significant effects on cognitive function

Kueider 2012 N=38 intervention studies International 55+ Participants without AD Computer-based cognitive Cognitive domain-specific (RCTs, non-randomised (limited to English or MCI training computerised programs and pre-post design language) interactive gaming, video Reported pre-post training studies) games, virtual reality. effect sizes for intervention (N=21 classic cognitive groups ranged from 0.06 to 6.32 based tasks; N=9 for classic cognitive training neuropsychological interventions, 0.19 to 7.14 for software; N=8 video neuropsychological software games) interventions, and 0.09 to 1.70 for video game interventions. Most studies reported older adults did not need to be technologically savvy in order to successfully complete or benefit from training. The authors reported that ‘findings are comparable or better than those from reviews of more traditional, paper-and-pencil cognitive training approaches suggesting that computerized training is an effective, less labour intensive alternative’ Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Lampit 2014 N= 52 RCTs International (no 60+ Cognitively healthy older Computerised cognitive The overall effect size (Hedges’ language limits) adults (without dementia training g, random effects model) for Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 165 or cognitive impairment) computerised cognitive training Studies had to report versus control was small and >/= 4h of computerised statistically significant, g = 0.22 cognitive training to be (95% CI 0.15 to 0.29). Small to included; compared to moderate effect sizes were found active or passive control for nonverbal memory, g = 0.24 condition (95% CI 0.09 to 0.38); verbal memory, g = 0.08 (95% CI 0.01 to 0.15); working memory (WM), g = 0.22 (95% CI 0.09 to 0.35); processing speed, g = 0.31 (95% CI 0.11 to 0.50); and visuospatial skills, g = 0.30 (95% CI 0.07 to 0.54). No significant effects were found for executive functions and attention. Analyses of effective design factors showed that home-based administration was ineffective compared to group- based training, and that more than three training sessions per week was ineffective versus three or fewer. There was no evidence for the effectiveness of working memory training, and only weak evidence for sessions less than 30 min 166 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 166 Evidence-based Tables

Table 40: Cognitive Stimulation – other systematic reviews

Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Huckans 2013 N=14 RCTs International NR Community-dwelling older Cognitive rehabilitation Short-term (<1 month) and long- (limited to English adults with MCI therapies defined as any term (>/= 1 month) impact on language) systematic behavioural objective cognitive performance, therapy specifically including attention/information designed to improve processing, memory, executive cognitive performance function and global cognition

Jean 2010 N=15 intervention NR NR People with mild amnestic Cognitive training Objective or subjective measures studies (5 were RCTs, cognitive impairment of memory 8 quasiexperimental (A-MCI) Some programs focused studies, 3 pre-test-post- only on memory, whereas No meta-analysis conducted test between group other programs used controlled trials, 2 multifaceted approaches 44% of objective measures of between group pre-test- targeting two or more memory significantly improved, post-test, 3 pre-test post- cognitive functions. Eight when compared with 12% test trials with only one were offered in groups, of objective measures of group and 2 were single and seven took place cognition other than memory case studies. Number of on an individual basis participants in studies was 49% of subjective measures of from 1 to 193 but Most memory, quality of life, or mood included studies had small significantly improved after sample sizes (n<30) training Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Metternich 2008 N=14 RCTs included of International Mean age People with subjective Non-pharmacological Subjective memory and which 13 were included (limited to English, 53+ for memory complaints or interventions (The objective memory. EC followed in meta-analyses and 10 Dutch, German or include desire to improve their categories of intervention by combined interventions, was were included in meta- French) studies memory performance. were: conventional most effective for subjective analysis of the primary Studies exclusively memory memory. MT or physical and outcome measure: investigating groups of training (MT; n = 8), mental training combined were subjective memory. patients with objective expectancy change (EC: not efficient. On objective Studies were only memory deficits e.g. MCI, cognitive restructuring, memory, MT was the only Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 167 included if they included AAMI were excluded psychoeducation, etc.; n = efficient intervention a measure of objective 5), combined interventions memory (n = 5), physical training Objective memory (n = 1), and physical and No significant effects for EC mental training combined versus WL, P, CT or MT. (n = 2) Significant effects for MT versus WL and P, as well as MT versus CT; No significant effects for MT versus PT or CPT, for CT versus WL/P and for PT versus WL or combined physical and mental training versus WL/P

Subjective memory The meta-analyses show significant effects for EC training over WL and P conditions, and no significant effects for EC versus CT or MT. There were no significant effects for MT versus WL and P. A significant effect for CT over WL (p = 0,047), but not versus P, was observed. There was a weak trend towards significance for CT over MT (p = 0.098). There was no significant effect for combined physical and mental training versus WL. There were no data for comparisons regarding PT CT = combination of conventional memory training and expectancy change; PT = physically oriented training; CPT = combination of conventional memory training and physically oriented training; MT = conventional memory training; P = placebo; WL = waitlist 168 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 168 Evidence-based Tables

Study Included studies and Country Age (for Population Intervention Relevant outcomes design inclusion) or exposure Papp 2009 RCTs (N=10) International NR for Community-dwelling Cognitive training Cognitive outcomes. (limited to English review healthy elderly (people language) inclusion with MCI, Alzheimer’s From meta-analysis: but all disease or dementia were The weighted mean effect studies excluded) size (Cohen’s d) of cognitive mean age intervention across all outcome >>55+ measures after training was 0.16 (95% confidence interval, 0.138 to 0.186)

The authors reported a lack of consensus on what constitutes the most effective type of cognitive training, insufficient follow-up times, a lack of matched active controls, and few outcome measures showing changes in daily functioning, global cognitive skills, or progression to early AD

Valenzuela 2009 N=7 RCTs International 50+ Healthy older adults. Cognitive exercise training Neuropsychological performance Am J Geriatr Soc (no language People with dementia or restrictions) cognitive impairment were excluded

Zehnder 2009 N=24 intervention studies International 60+ Older adults – healthy and Memory training Cognitive outcome measures (limited to English with MCI or German) Table 41: Cognitive Stimulation – Quality assessment of systematic reviews of cognitive stimulation interventions

AMSTAR (High= 8-11, Moderate= 5-7, Low= <5)

Key: 1. ‘a priori design; 2. duplicate study selection and data extraction; 3. comprehensive literature search; 4. Status of publication as an inclusion criterion; 5. List of studies (included and excluded provided)?; 6. Characteristics of the included studies provided? 7. Scientific quality of the included studies assessed and documented; 8. Scientific quality of the included studies considered in formulating conclusions; 9. Appropriate method to combine findings; 10. Publication bias; 11. Conflict of interest Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 169

Author (Year) 1 2 3 4 5 6 7 8 9 10 11 Ranking Bleakley 2013 Not re- Y Y Y Y for Y Y Y Y N Y High ported included studies; N for excluded Gross 2012 Not re- Y Y Y Included N N N Y Y Y for the Moderate ported studies SR; N for shown in individual meta- studies analysis; excluded not reported Hindin 2012 Not re- Not re- Y Y Y for N (only Y Y Y Y Y Moderate ported ported included refs in S1) studies (Appendix S1), N for excluded Karr 2014 Y - re- Y for data Y Y Y Y Y Y Y Y Y High ports prior extrac- inclusion tion; not criteria reported for study inclusion Kelly 2014 Not re- Y Y Y Y Y Y N Y Y Y for the High (Ag Res Rev 15, 28- ported SR; N for 43) individual studies 170 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 170 Evidence-based Tables Author (Year) 1 2 3 4 5 6 7 8 9 10 11 Ranking Kueider 2012 N Not re- Y Y Y for Y N N Y Y Y Moderate ported included studies; N for excluded Lampit 2014 Y Y Y Y Y for Y Y N Y Y Y High included studies; N for excluded Li 2011 Not re- Y Y Not re- Y for Y N N Y Y Y for Moderate ported ported included review; studies; N for some individual excluded studies referenced Martin 2011 Y Y for Y Y Y Y Y N Y Y Y for SR; High study N for inclusion, individual unclear studies for data extraction Reijnders 2013 Not re- Y for data Y Y Y for Y Y Y Y Y Y for SR; High ported extraction; included N for unclear studies; individual for study N for studies inclusion excluded Stott 2011 Y - Not re- Y Y Y for Y Y Y Y N Y High specifies ported included a priori studies; inclusion N for criteria excluded Tardif 2011 Not re- Not re- Y (but Y Y for Y N Y Y N N Low ported ported very limit- included ed search studies; terms) N for excluded Teixeira 2012 Not re- Not re- Limited Unclear Y for Y N N N/A N Y for SR; Low ported ported search included N for terms studies; individual some studies excluded referenced Table 42: Social – Systematic reviews or exposures for social participation interventions

Study Included study and Country Age Population Intervention Relevant outcomes design (years) or exposure Bickerdike 2014 N=7 systematic reviews International Not report- People who are, or were at risk Overview of SRs Loneliness, social examining interventions (limited to English ed of loneliness and social isolation of interventions for isolation and related University of York for loneliness and language) loneliness and social outcomes (incl. measures Evidence briefing social isolation N=7 isolation of health services

(York CRD) systematic reviews utilisation and associated Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 171 examining interventions costs) for loneliness and social isolation

Berger 2013 N=13 intervention International NR Older adults with low vision Occupational therapy Social and leisure studies (9 RCTs, 1 non interventions participation randomised controlled trial, 3 other intervention studies)

Cattan 2005 N=30 intervention studies International - but ‘Older Older adults ‘regardless of race, Health promotion inter- Social isolation and/or included. 16 RCTs. 10 included studies adults’ as gender, physical disability or ventions to prevent social loneliness non-randomised and 4 conducted in USA defined in ability’ isolation and loneliness other intervention studies (N=17) Canada individual among older adults and Europe reviews

Choi 2012 N= 6 controlled International (Lim- NR Older adults Computer mediated social Loneliness outcomes Healthcare Info intervention studies ited to studies in support for managing generally measured on included of which 4 were English or Korean) (mean age loneliness a loneliness scale (and RCTs. 5 studies reported of all includ- depression) loneliness as an outcome ed studies 65+)

Dickens 2011 N=32 (16 RCTs and International NR Most studies were conducted in Interventions designed to Social isolation and 16 quasi-experimental (limited to English community dwelling older. Also alleviate social isolation loneliness studies) language studies) included studies were in people and loneliness with existing disease

Findlay 2003 N=17 included International NR Older adults – inclusion criteria Interventions to reduce Loneliness, social intervention studies (N=6 (English language). not specified. Generally studies social isolation isolation, measures of RCTs) Included studies are in people living in the personal contact, and a conducted in the community with and without range of other outcomes US (8), Australia support such as mortality etc (3), Canada (2), The Netherlands (2), Italy (1), Sweden (1). 172 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 172 Evidence-based Tables

Study Included study and Country Age Population Intervention Relevant outcomes design (years) or exposure Forsman 2011 N=69 intervention studies: N=17 trials from 65+ Older adults Psychosocial Primary outcomes related n=56 RCTS and n=13 Europe, N=52 non- who did not meet diagnostic interventions for the to mental health, although non-randomised European (majority Mea of criteria for any mental disorder promotion of mental social capital, social controlled studies from US) included (e.g. dementia) health or prevention of network or social support studies depression were also included as In N=32 trials, participants outcomes lived independently, in senior communities or received home health services; n=15 trials, participants lived in nursing homes or other institutions; n=22 studies the participants’ living status was not clear

Hagan 2014 N=17 intervention studies International NR Older adults Social therapeutic Loneliness, social (9 controlled trials; 3 (limited to English interventions primarily isolation, social networks, before and after studies; language) N=9 studies were conducted in aimed at reducing social support 2 pilot studies and 3 people living in the community; loneliness or social evaluations) N=7 studies in supported living isolation Ten of the 17 studies communities or nursing care measured loneliness and N=1 was in an unspecified using discrete validated setting scales. The rest used broader non-explicit measures

Heaven 2013 14 articles, reporting 11 Studies from highly 55-70 Older adults in the retirement Interventions to extend Participants’ perception of separate studies that developed nations transition or support social roles for social roles (or to aspects evaluated 7 different only included older adults of their health or well- interventions being). Minimum follow-up (Interventions: US period of 3 months (Only studies with a n=5; Japan n=1; control or comparison The Netherlands group were included, only (n=1) 3 were RCTs)

Morris 2014 N= 18 intervention studies International 45+ Older adults living at home. Smart technologies Social connectedness (N=12 RCTs, 3 non- (limited to English Varying participant characteristics randomised controlled language) studies, 3 before and after studies) Study Included study and Country Age Population Intervention Relevant outcomes design (years) or exposure Forsman 2011 N=69 intervention studies: N=17 trials from 65+ Older adults Psychosocial Primary outcomes related n=56 RCTS and n=13 Europe, N=52 non- who did not meet diagnostic interventions for the to mental health, although non-randomised European (majority Mea of criteria for any mental disorder promotion of mental social capital, social controlled studies from US) included (e.g. dementia) health or prevention of network or social support studies depression were also included as In N=32 trials, participants outcomes lived independently, in senior communities or received home health services; n=15 trials, participants lived in nursing homes or other institutions; n=22 studies the participants’ living status was not clear

Table 43: Social – Systematic reviews relevant to barriers/facilitators to social participation

Study Included study and Country Age Population Intervention Relevant outcomes design (years) or exposure Nef 2013 N=18 interventions and International 55+ Older adults Social networking sites Use of SNS and B/F in

qualitative studies (limited to English relation to maintenance of Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 173 language) social networks 174 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 174 Evidence-based Tables Table 44: Social – Interventions to improve uptake/maintenance of social participation with cognitive outcomes

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Carballo- Non-ran- Spain 65-85 Elderly people with Intervention: ‘Non- Follow-up: 9 months Results reported separately for García 2013 domised Mean 77 (6) or without cognitive pharmacological therapy’ those with and without cognitive controlled impairment who applied delivered 1 hour a day/2 Lost to follow-up: impairment. trial to participate in social or days a week in groups Outcome assessment: cultural activities provided of no more than 20, for 9 Cognitive outcomes: Normal ageing group (no free by municipal senior months. The Miniexamen cognitive impairment) (N=240 centres in central Madrid Consisted of: Cognitive Cognoscitivo (MEC) and total) stimulation exercises the adapted and validated (N=240 with normal ageing according to group Spanish-language Intervention (N=217) and N=77 with cognitive abilities. Exercises covered version of the Short Mean (SD) impairment) a wide range of cognitive Portable Mental Status MEC cognition processes: attention, Questionnaire (SPMSQ) Pre 31.49 (2.533) Setting: Community perception, memory, Post 32.10 (2.298), language, inhibition, Other outcomes: Also: Improvement Gender: 86.1% female planning, reasoning, General mental state, arithmetic, drawing, etc. emotional well-being, SPMSQ cognition Ethnicity: Not reported quality of life, and daily life Pre 1.16 (1.226) Group dynamics tasks activities Post 0.56 (0.907) SES: Educational level: designed to strengthen Improvement Without cognitive social skills, the impairment:28.3% expression of positive Control (N=23) functionally illiterate; feelings, and interaction Mean (SD) 41.7% primary studies; between participants MEC cognition 12.5% secondary studies; Pre 30.52 (2.556) 17.5% university education General topics such as Post 28.43 (3.396) With cognitive impairment: depression, anxiety, etc. Decline 71.4% functionally illiterate; 20.8% primary Art therapy conducted SPMSQ cognition studies; 2.6% secondary in partnership with a Pre 1.26 (1.453) studies; 5.2% university museum, and team- Post 2.09 (1.379) building workshops were Decline also used to increase social participation and Effect of intervention vs control prevent isolation MEC cognition F=9.091 Control group: Other P< .001 leisure, social or cultural N2p = 0.134 activities SPMSQ cognition F= 8.366 P< .001 N2p = 0.125

Significant effect by age (F =2.583, P < .05), younger participants benefitted the most Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Carballo- Non-ran- Spain 65-85 Elderly people with Intervention: ‘Non- Follow-up: 9 months Results reported separately for García 2013 domised Mean 77 (6) or without cognitive pharmacological therapy’ those with and without cognitive controlled impairment who applied delivered 1 hour a day/2 Lost to follow-up: impairment. trial to participate in social or days a week in groups Outcome assessment: cultural activities provided of no more than 20, for 9 Cognitive outcomes: Normal ageing group (no free by municipal senior months. The Miniexamen cognitive impairment) (N=240 centres in central Madrid Consisted of: Cognitive Cognoscitivo (MEC) and total) stimulation exercises the adapted and validated (N=240 with normal ageing according to group Spanish-language Intervention (N=217) and N=77 with cognitive abilities. Exercises covered version of the Short Mean (SD) impairment) a wide range of cognitive Portable Mental Status MEC cognition processes: attention, Questionnaire (SPMSQ) Pre 31.49 (2.533) Setting: Community perception, memory, Post 32.10 (2.298), language, inhibition, Other outcomes: Also: Improvement Gender: 86.1% female planning, reasoning, General mental state, arithmetic, drawing, etc. emotional well-being, SPMSQ cognition Ethnicity: Not reported quality of life, and daily life Pre 1.16 (1.226) Group dynamics tasks activities Post 0.56 (0.907) SES: Educational level: designed to strengthen Improvement Without cognitive social skills, the impairment:28.3% expression of positive Control (N=23) functionally illiterate; feelings, and interaction Mean (SD) 41.7% primary studies; between participants MEC cognition 12.5% secondary studies; Pre 30.52 (2.556) 17.5% university education General topics such as Post 28.43 (3.396) With cognitive impairment: depression, anxiety, etc. Decline 71.4% functionally illiterate; 20.8% primary Art therapy conducted SPMSQ cognition studies; 2.6% secondary in partnership with a Pre 1.26 (1.453) studies; 5.2% university museum, and team- Post 2.09 (1.379) building workshops were Decline also used to increase social participation and Effect of intervention vs control prevent isolation MEC cognition F=9.091 Control group: Other P< .001 leisure, social or cultural N2p = 0.134 activities SPMSQ cognition F= 8.366 P< .001 N2p = 0.125

Significant effect by age (F =2.583, P < .05), younger participants benefitted the most

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Carballo- Cognitive impairment group García 2013 (N=77 total) Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 175 (continued) Intervention (N=63) MEC cognition Pre 27.19 (3.868) Post 26.43 (4.578) Decline

SPMSQ cognition Pre 2.52 (1.804) Post 2.14 (1.983) Decline

Control (N=14) MEC cognition Pre 27.21 (4.710 Post 22.21 (5.041) Decline

SPMSQ cognition Pre 2.93 (2.336) Post 5.29 (2.091) Decline

Intervention Vs control MEC cognition F = 7.451, P < .001. N2p = 0.293 SPMSQ cognition F = 6.754, P < .001. N2p = 0.273 176 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 176 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Carlson RCT (pilot) US Mean: In- Older adults Experience Corps® Follow-up: 4 to 8 months Cognitive outcomes: (age and 2008 tervention: places older volunteers education adjusted) Experience (cluster 70.1 (6.42) N=149 randomised in public elementary Lost to follow-up: Corps randomised Control : [N=21 dropped out schools in roles designed Intervention: 11.4%; Trail-making test by school) 68.4 (5.15) immediately following to meet schools’ needs control: 17.2% Part A: No significant difference randomization] and increase the social, between groups physical, and cognitive Outcome measurement: Part B: Significant difference Baseline MMSE: activity of the volunteers Cognitive outcomes: between intervention and control Intervention: 24.96 (3.45); Memory, executive groups at follow-up (p<0.05). Control: 25.3 (2.60) Community-based function (EF), and Intervention group improved by program designed to psychomotor 1.3 secs from baseline; control Setting: Community increase cognitive and Speed: group declined 21.7 secs from physical activity in a social, Trail Making Test: baseline Female: Intervention: real-world setting Parts A and B 83%; Control 93% Word list memory: Word list memory Intervention (N=70): Immediate recall No significant difference between Ethnicity: Intervention: Participants randomized Delayed recall groups for immediate or delayed 94% black; Control: 95% to EC trained in teams to Rey-Osterrieth: recall black help elementary school Copy score children with reading Delayed recall Rey-Osterrieth CFT SES: Mean years achievement, library Copy score: No sig diff between education: Intervention: support, and classroom groups. 11.9 (2.54) ; Control: 11.2 behaviour for 15 hr/week Delayed recall: Significant (2.66); 38% had less than difference between intervention high school education Control (N=58): Wait-list and control groups at follow-up control (p<0.05). Intervention group improved by 1.0 points and control group declined by 1.1 points Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Cohen- RCT Israel 65+ Older adults with memory Health promotion Follow-up: 10 weeks Cognitive outcomes Mansfield complaints (N=15): structured intervention and follow-up No significant differences 2015 Note: In- Mean 73.5 format course including (p<0.05) between intervention terventions (SD 5.2) MMSE 24+ for inclusion lectures, discussions, Lost to follow-up: groups (health promotion; were not (Mean MMSE at baseline exercises, handouts and Health promotion: 20%; cognitive training; participation) conducted 28) homework; covering health cognitive training: 20%; for any cognitive outcome except simultane- behaviours; dementia and participation: 28.6%. Only for self-reported memory which ously N=44 randomised delirium; communication; completers were analysed was higher in the group receiving Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 177 cognitive activities to keep for results cognitive training Setting: Community the mind fit; relationships, depression, and coping; Outcome measurement: Loneliness Gender: 72.7% female home and travel safety; Global Cognitive Score No significant differences recreation and leisure; assessed using the (p<0.05) between intervention Ethnicity: Not reported medications and health MindStreams mild groups (health promotion; care providers; physical cognitive impairment cognitive training; participation) SES: Mean years educa- activity; and lifelong assessment, a for loneliness as an outcome tion: 14.82 (3.77); range learning computerized cognitive 5-22 assessment. The Mini- There was a significant difference Cognitive training Mental State Examination in change in loneliness for (N=15): and the self-report of the cognitive training group Memory training based memory difficulties were compared to the wait-list control on the previous ACTIVE also utilized. To assess trial with a focus on verbal well-being, the UCLA episodic memory exercises Loneliness Scale-8 was used. Health was Participation-centred evaluated by self-report course (N=14): Book instruments club was used as a focus to deliver memory, cognitive and organisational strategies and using cognitive- behavioural principles. The course used external strategies (e.g. reading aid and daily planner), internal strategies (e.g. linking meaning to new information), and social interaction strategies (e.g., asking for help and sharing memory difficulties)

Control: Wait-list control (N=28? but unclear, changing numbers) 178 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 178 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Hughes RCT US Mean: In- Older adults with MCI Intervention (N=10): Follow-up: 24 weeks Adherence: The Wii group 2014 tervention: recruited from a population Group-based interactive (post-intervention) & at 1 attended an average of 23.1 78.5 (7.1); cohort study video gaming using year (SD 1.1, range 21–24) sessions control: Nintendo Wii; 90 mins compared with 21.8 (SD 3.3, 76.2 (4.3) Mean MMSE at baseline week for 24 weeks Lost to follow-up: 10% (1 range 14–24) in the control Intervention: 27.2 (1.9) died by 1-year follow-up, group; 18 participants attended Control: 27.1 (1.8) The Wii Sports games, 1 did not complete post- at least 20/24 sessions; and 9 including bowling, golf, intervention assessment) attended all sessions N=20 randomised tennis, and baseball were the core of the sessions. Outcome measurement: Cognitive function: No Setting: Community From week 7, participants Cognitive function: significant differences between were introduced to new Computerized Assessment intervention and control groups Gender: Intervention: games for 15-30 mins of Mild Cognitive 80% female; control: 60% of the session. In weeks Impairment (CAMCI) No significant differences for any female 10 and 20 participants Subjective cognitive ability: other outcomes measured competed in Wii Cognitive Self-Report Ethnicity: Intervention: tournaments to encourage Questionnaire-25 70% white; control: 90% enhanced effort and social Social functioning: white interaction Cognitive Self-Report Questionnaire-25 SES: Mean years Control (N=10): Health Other outcomes; mood, education: Intervention education designed to IADL, gait speed 13.8 (2.4); control 13.1 provide a source of (1.9) passive cognitive stimulation in a socially matched setting ; 90 mins week for 24 weeks Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Park 2014 RCT US 60 to 90 Older adults without 6 lifestyle conditions: Follow-up: 3 months Episodic memory: sig diffs Synapse cognitive impairment comparing productive intervention and follow-up between photo group and Project Mean 71.7 engagement with receptive placebo (p=0.01) and for dual MMSE at baseline 26+ engagement (did not Lost to follow-up: condition and placebo (p=0.05). actively acquire new skills) No significant effects were found Setting: Community which involve different Outcome measurement: for other conditions (quilt or cognitive function Processing speed, social) compared to placebo Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 179 Gender: 73.9% female Participated in all assessed using digit- conditions for an average comparison tasks with No significant difference between Ethnicity: 14.2% ‘minority’ of 16.51 hr a week for three, six, and nine items the two receptive-engagement (not described) participants 3 months (except for no conditions (p = 0.59), and the treatment control) Mental control, assessed three productive-engagement SES: Only those with 10th Productive engagement: using Flanker Center conditions did not differ from one grade + education were Photo (N=29): novice Letter, Flanker Center another (p = 0.19. included participants learned digital- Arrow, and Flanker Center photography and computer Symbol tasks and the Productive engagement (the skills using photo-editing Cogstate Identification quilt, photo, and dual conditions) software Task caused a significant increase in Quilting (N=35): novice episodic memory compared with participants learned how to Episodic memory, receptive engagement (the social design and sew quilts assessed using the and placebo conditions) Dual condition (N=42): immediate recall participants spent half of section of the modified The three productive- the 3 months doing quilting Hopkins Verbal Learning engagement groups were and the other half on Task, the Cambridge superior in episodic memory photography Neuropsychological when compared with the social Receptive engagement: Test Automated Battery group alone Social condition (N=36): (CANTAB) Verbal participants engaged Recognition Memory Task in on-site, facilitator-led and the long-delay section In summary, the evidence social interactions, field of the modified Hopkins suggests that sustained effort trips, and entertainment. Verbal Learning Task. to acquire a demanding new Placebo condition skill improves episodic memory; (N=39): Visuospatial processing, no evidence suggesting Tasks off site that assessed using the that socializing, information appeared to be beneficial CANTAB exchange, and novelty alone to cognition but had no Spatial Memory Task, facilitate cognitive function. substantiated link to the CANTAB Stockings cognitive improvement of Cambridge Task, and Participants in the photo and dual (e.g., listening to classical a modified version of conditions exhibited a significant music, completing word- Raven’s Progressive improvement in episodic memory, meaning puzzles). Matrices whereas the effect was not No treatment control significant for those in the (N=40) quilt condition (p = 0.11) but was in the direction of facilitation 180 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 180 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Pitkala 2011 RCT Finland 75+ Population: Older adults Intervention (N=117): Follow-up: 3, 6 months ADAS-Cog scale (3 months) Mean: 80 with subjective feelings of Socially stimulating group for cognitive outcomes Mean changes (all participants) (3.6 loneliness intervention aimed at I: -2.6points (95% CI -3.4 to -1.8) enhancing interaction Lost to follow-up: C: -1.6 points (95% CI -2.2 to 80.2% of intervention and friendships between Intervention: 6.0%; control: -1.0 group and 78.8% of control participants and to 17.8% (p=0.023; F1,167.8 = 5.23) group lived alone. stimulate them socially. Art group Groups were facilitated Outcome measurement: I: -2.4 points (95% CI -3.5 to -1.3) Mean MMSE at baseline: by trained professionals. Cognition by Alzheimer’s C: -1.8 points (95% CI -2.9 to Intervention: 26.9 (2.4); Participants were divided disease assessment scale -0.8) control (26.6) (2.6). into 3 groups depending (ADAS-cog) and mental (p=0.17; F1,47.2 = 1.88) on their interests: function by 15D measure Exercise group N=235 randomised therapeutic writing (N=48); and psychological I: -3.2 points (95% CI -4.7 to -1.7) group exercises (N=92); wellbeing and HRQoL at C: -1.6 points (95% CI 2.6 to – Setting: Day care centres or art experiences (N=95) 12 months 0.5 and then randomised to (p=0.60; F1,72.7 = 0.28) Gender: Female intervention or control Writing group Intervention: 74.4%; within those groups. I: -1.7 points (95% CI -2.7 to -0.7) control: 72.9% Intervention was once C: -1.2 points (95% CI -2.7 to per week and usually 0.3) Ethnicity: Not reported lasted for 6 hours and was (p=0.033; F1,33.6 = 4.49) provided free of charge 15D index dimension of mental SES: % with economic function) over 12 months status good or moderate: Control (N=118): I: +0.048 (95%CI: +0.013 to Intervention 97.2; control: Continued in normal +0.085) 96.2 community care. They C: -0.027 (95%CI: -0.063 to could participate in +0.010) their normal hobbies (p=0.004; t = 2.89, df=187) and activities but no intervention was arranged Note: Cognitive outcomes also for them measured at 6 months but not reported here. Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Quasi- US Range 55- Older adults living in the Intervention (N=61): Follow-up: 20 week Mean participation: 17.3 weeks Morrow RCT (field 93 community or retirement Engaged lifestyle intervention (Note: pre- out of 20 (86.5%) 2007 experiment) Mean: villages programme. Team- and post cognitive tests up Only processing speed showed Senior I: 73.6, based, competitive to 9 months apart). differential positive change in the Odyssey Assigned range N=81 [initially 85] programme of creative experimental group relative to the programme those from 60–93 randomised problem solving involving Lost to follow-up: control group; differential change retirement C: 70.2, repeated opportunities Intervention: 18% (20 in divergent thinking reached a villages to range Setting: Community for engagement with wks); control: 12% marginal level of significance. Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 181 the exper- 58–85 programme ill-defined problems with Predisposition toward imental multiple solutions. (The Outcome measurement: cognitive engagement group Gender: Not reported programme was developed Cognitive outcomes Change (from pre-to-post because of from a well-established 1. Processing speed: intervention SD) effort put Ethnicity: Not reported programme for children Letter and Pattern Intervention in to build and young adults) Comparison tasks and Mindfulness -0.02 relation- SES: Mean years Finding As and Identical Need for cognition 0.11 ships. So education: intervention: Control (N=24): Wait-list Pictures Metamemory in adulthood (MIA) not properly 16.1 (0.4); control: 15.4 control 2. Working memory: self-efficacy -0.06 randomised (0.7) Letter–Number Perceived activity level -0.11 Sequencing Processing speed 0.09 3. Inductive reasoning: Working memory 0.12 Letter Sets and Figure Inductive reasoning 0.22 Classification and Visuospatial processing 0.33 Everyday Problem Solving Divergent thinking 0.29 4. Visual-spatial Control processing: Card Rotation Mindfulness -0.40 and Hidden Patterns Need for cognition -0.28 5. Divergent MIA self-efficacy -0.09 thinking (fluency): Perceived activity level 0.15 Word Association, Processing speed 0.70 Ornamentation, Working memory -0.06 and Opposites, FAS and Inductive reasoning -0.29 Alternate Uses Visuospatial processing 0.01 Divergent thinking 0.11 Dispositions reflective of Between group difference (p cognitive engagement: value) Mindfulness, need for Mindfulness 2.03 (0.02) cognition, memory self- Need for cognition 1.85 (0.03) efficacy, activity MIA self-efficacy 0.18( >0.20) Perceived activity level -1.80 (0.08) Processing speed 1.76 (0.04) Working memory 0.85 (0.20) Inductive reasoning 1.17 (0.12) Visuospatial processing 1.03 (0.15) Divergent thinking 1.32 (0.10) 182 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 182 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Correlations Between Morrow Measures of Perceived 2007 Cognitive Engagement and Senior Cognitive Components (p Odyssey value) programme (cont...) Mindfulness Processing speed 0.12 (ns) Working memory 0.12 (ns) Inductive reasoning 0.20 (ns) Visuospatial processing 0.14 (ns) Divergent thinking 0.26 (p<0.05) Need for cognition Processing speed 0.11 (ns) Working memory 0.20 (ns) Inductive reasoning 0.26 (p<0.05) Visuospatial processing 0.30 (p<0.01) Divergent thinking 0.19 (ns) MIA self-efficacy Processing speed 0.21 Working memory 0.26 (p<0.05) Inductive reasoning 0.39 (p<0.01) Visuospatial processing 0.28 (p<0.01) Divergent thinking .11 Perceived activity level Processing speed 0.30 (p<0.01) Working memory 0.25 (p<0.05) Inductive reasoning 0.44 (p<0.01) Visuospatial processing 0.39 (p<0.01) Divergent thinking 0.38 (p<0.01) Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Correlations Between Morrow Measures of Perceived 2007 Cognitive Engagement and Senior Cognitive Components (p Odyssey value) programme (cont...) Mindfulness Processing speed 0.12 (ns) Working memory 0.12 (ns) Inductive reasoning 0.20 (ns) Visuospatial processing 0.14 (ns) Divergent thinking 0.26 (p<0.05) Need for cognition Processing speed 0.11 (ns) Working memory 0.20 (ns) Inductive reasoning 0.26 (p<0.05) Visuospatial processing 0.30 (p<0.01) Divergent thinking 0.19 (ns) MIA self-efficacy Processing speed 0.21 Working memory 0.26 (p<0.05) Inductive reasoning 0.39 (p<0.01) Visuospatial processing 0.28 (p<0.01) Divergent thinking .11 Perceived activity level Processing speed 0.30 (p<0.01) Working memory 0.25 (p<0.05) Inductive reasoning 0.44 (p<0.01) Visuospatial processing 0.39 (p<0.01) Divergent thinking 0.38 (p<0.01)

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Quasi-RCT US Interven- Older adults from the Intervention (N=107): Follow-up: 20 week Adherence to the programme: Morrow (field exper- tion: 73.0, community and local Engaged lifestyle intervention Attendance at weekly meetings Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 183 2008 iment) range: retirement communities programme. Team- was variable (6-20 session; 59–93 based, competitive Lost to follow-up: mean 15.5 (SE4) Senior Assigned Control: N=181 randomised program of creative Intervention: 19%; control: Odyssey those from 72.0, range: problem solving involving 15% did not completed Cognitive tests (one-tailed programme retirement 58–91 Setting: Community repeated opportunities post-test tests) (appears villages to for engagement with to be a the exper- Gender: Not reported ill-defined problems with Those who returned for Different in change significant for different imental multiple solutions. (The follow-up scored higher speed study from group Ethnicity: Not reported programme was developed for speed of processing at t(146)=1.81, p=0.036 Stine-Mor- because of from a well-established pretest than those who did row 2007 effort put SES: Years of education programme for children not return Inductive reasoning t(146)=1.83, above) in to build (mean): Intervention 16.3 and young adults). p=0.034 relation- (SE 4), Control 16.0 (SE 3) Outcome measurement: Note: ships. So Control (N=74): Wait-list Cognitive ability: Divergent thinking recruitment not properly control composite measure of fluid t(147)=1.88, p=0.031 was con- randomised ability ducted over Not significant for Working 2 years/ Measures of dispositions memory seasons reflective of habitual t(146)=1.01, p=0.136 so data cognitive engagement: reported in Mindfulness, need for Visual-spatial processing Stine-Mor- cognition, memory self- t(144)=.60, p=0.275. row 2007 efficacy may be Overall showed differential first wave positive change among those of data that who participated in the cognitive may also be intervention, t(149)=3.11, reported in p=0.001 this trial Control group vs. experimental group Self-efficacy, t(133) = −1.59 Mindfulness, t(128)=0.81 Need for cognition, t(130)=0.68 184 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 184 Evidence-based Tables Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Stine- Quasi-RCT US 60 to 94 Healthy older adults from Engagement programme Follow-up: 16 weeks Adherence to the programmes: Morrow Mean: 72.6 the community and local (Senior Odyssey) intervention but pre- and Engagement participants 2014 Some par- retirement communities (N=188): A team-based post-test were 30 to 32 attended an average of 11.0 Senior ticipants not competitive program in weeks apart out of 16 session (SD 4.8) and Odyssey randomly N=461 randomised creative problem solving Training participants completed programme allocated (no explicit instruction) Lost to follow-up: an average of 12.9 modules out versus to meet Setting: Community Engagement: 16 %; of 16 (SD -5.2); this difference in cognitive project As above Cognitive training: 12 %; adherence was significant, F(1, training pro- deadlines Gender: Engagement Wait-list: 12% 317) 11.42, p < 0.001 gramme intervention: 71% Cognitive training female; Cognitive training (N=130): Home-based Outcome measurement: Pre-to-post change effect intervention 77% female; inductive reasoning Cognitive Outcomes: sizes: Waiting list control: 76% training program Processing Speed: Letter Results presented graphically female (instruction and practice and Pattern Comparison explicit) tasks and the Finding A’s Key results: Ethnicity: Not reported task Those in the training condition Wait-list control (N=143): Reasoning: Letter Sets, showed selective improvement SES: Years of education No intervention but Number Series, Letter in inductive reasoning. Training (mean): Engagement participated in testing as a Series, and Word Series participants showed more intervention 15.7 (2.6); control for re-test effects tasks and the Everyday change than both Engagement Cognitive training Problem-Solving (EPS) and Waitlist participants and intervention 15.2 (2.7); task Engagement and Waitlist Control 15.4 (2.5) Visual-spatial processing participants did not change from (VSP): Card Rotation and each other. Hidden Patterns Those in the engagement Divergent Thinking: condition showed selective Alternate Uses task and improvement in divergent the Opposites task thinking. Neither the Waitlist or Verbal Episodic Memory: Training group had significant Was measured using re-test effects two indicators derived from performance on the Correlations between baseline Hopkins Verbal Learning characteristics and latent Test; total number of training improvements words remembered over three trials (HVLT-Tot) and Divergent Thinking: the delayed recall score Engagement: Statistically (HVLT-DR) and proportion significant (p<0.05) differences of correct propositions for age, Montreal Cognitive recalled in an immediate Assessment (MoCA), verbal, sentence free-recall task openness, need for cognition, social network index

Inductive Reasoning: Training: Statistically significant differences for MoCA, Verbal Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Tesky 2011 RCT Germany 50+ Older adults without AKTIVA intervention (n = Follow-up: 8 week Cognitive outcomes AKTIVA dementia or cognitive 126): Group programme intervention and post- Older persons (≥75 years) Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 185 study Median: 72 impairment of cognitively stimulating test conducted 1 week showed enhanced speed of (7) leisure activities (8 weekly after (week 9); 2 booster information processing (by TMT N=307 randomised sessions and two booster sessions conducted at Version A) (F = 4.17*, p <0.05); sessions after a break of 4 27 to 28 weeks and then younger participants (< 75 years) Setting: Community months) follow-up tests conducted showed an improvement in 29 weeks after start of subjective memory decline (by Gender: 72.3% female AKTIVA intervention plus intervention MAC-Q) (F = 2.55*, p < 0.05). [73.1% of analysed sample nutrition and exercise (N=152)] counseling (n = 84): Lost to follow-up: N=67 Frequency of leisure activities (21.8%) withdrew from the Additionally, AKTIVA enhanced Ethnicity: German & Control group (n = 97): No study. Those with impaired the frequency of activities for Turkish ethnicity in initial intervention cognition (N=28) and leisure activities for subassembly sample; only German N=4 Turkish participants groups ethnicity in analysed excluded from analysis sample after the programme.

SES: ‘Most participants Outcome measurement: had attended school for Cognitive outcomes: Mini- about 10 years, and only Mental Status Test a few participants had completed an academic ADAS-Cog: the cognitive university education’ part of the Alzheimer’s Disease Assessment Scale

Part A and B of the Trail- Making Test

Clinical Dementia Rating (CDR)

Self-report questionnaires 186 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 186 Evidence-based Tables

Table 45: Social – Included studies about barriers and facilitators

Study Study Country Age Population Objective Quality design (years) Qualitative studies Goll 2015 Qualitative UK 60+ Older adults (N=15) identified by voluntary organisations in Barriers to social ++ (semi- urban and multicultural areas of London, identified by staff participation among lonely structured Range 62-100; as accessing none/few social opportunities or experiencing older adults interviews) mean 79 (12) loneliness/social isolation

All participants lived alone, except one who lived with her husband who had severe dementia

Gender: 66.7 % female

Ethnicity: 73.3% (N=11) white British; 13.3% (N=2) Black Caribbean; 6.7% (N=1) Central Asian; 6.7% (N=1) South-East European

SES: Standard Occupational Classification (SOC): 1 (management, N=2); 2 (professional, N=4); 3 (technical, N=1); 4 (administrative, N=2); 5 (skilled trades, N=2); 8 (operative, N=2); 9 (elementary, N=2)

Barriers and facilitators to a specific intervention Cattan 2011 Qualitative UK Range mid-50s N=40 in total. Older service recipients (N=27) of a national Qualitative evaluation ++ (interviews) to early-90s telephone befriending scheme for isolated and/or lonely of a national telephone older adults from 8 project sites across the UK; and service befriending scheme volunteers (N=6); combined recipients and service volunteers (N=7)

People who work with older adults Anderson 2009 Qualitative US Of stylists: N=40 hairstylists from 31 randomly selected salons To understand - J Appl Ger (survey with relationships and informal open-ended 18-30y 12.5% Gender: 85% female social support from questions). 31-45y 25.0% hairdressers with older Also descriptive 45-60y 50.0% Ethnicity: 80.0% white, 17.5% African American, 2.5 % other adult clients stats & 61y and above, frequencies 12.5% SES: High school or GED 32.5%; some college 52.5%; associates degree 12.5% Table 46: Social – Quality assessment of systematic reviews of social participation interventions

AMSTAR (High= 8-11, Good= 5-7, Poor= <5)

Author (Year) 1 2 3 4 5 6 7 8 9 10 11 Rank- ing Berger 2013 N N Y Y Y for N N N N/A N N Low (Arbesman 2013 methods) included studies; N for excluded Bickerdike 2014 Over- Unclear view of the above SRs – no assessment tool Cattan 2005 N Y for YN Y Y for Y Y Y N/A N Partly High study, included for SR; unclear studies; N for for broken included abstract link for studies screen- exclud- ing/data ed extraction Choi 2012 N N (not Y Y Y Y Y Y Y Y N High Healthcare Info reported) Dickens 2011 Y Y Y Y Y for Y Y Y Y N Y High Protocol included not re- studies; ferred to N for ex- in main cluded text 188 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 188 Evidence-based Tables

Author (Year) 1 2 3 4 5 6 7 8 9 10 11 Rank- ing Findlay 2003 N N Y Y Y for Y N N N/A N N Low (not re- included ported) studies; N for excluded Forsman 2011 N Y Y Y Y for Partly N (not Y Y N Y Moder- included (some reported ate studies; data in for indi- N for ex- me- vidual cluded ta-anal- studies) ysis) Hagan 2014 N N Y Y Y for Y N N N/A N N Low included studies; N for excluded Heaven 2013 N Y Y Y Y for Y Y Y N/A N Y High included studies; N for excluded Morris 2014 N Y Y Y Y for Y Y Y N/A N Y (for High included SR), studies; N for N for ex- included cluded studies Nef 2013 Y Y Y for Y N N N/A N Y Moder- (au- included ate thor’s studies; roles N for ex- refer to cluded protocol but not in Meth- ods) Author (Year) 1 2 3 4 5 6 7 8 9 10 11 Rank- ing Findlay 2003 N N Y Y Y for Y N N N/A N N Low (not re- included ported) studies; N for excluded Forsman 2011 N Y Y Y Y for Partly N (not Y Y N Y Moder- included (some reported ate studies; data in for indi- N for ex- me- vidual cluded ta-anal- studies) ysis) Hagan 2014 N N Y Y Y for Y N N N/A N N Low included studies; N for excluded Heaven 2013 N Y Y Y Y for Y Y Y N/A N Y High included studies; N for excluded Morris 2014 N Y Y Y Y for Y Y Y N/A N Y (for High included SR), studies; N for N for ex- included cluded studies Nef 2013 Y Y Y for Y N N N/A N Y Moder- (au- included ate thor’s studies; roles N for ex- refer to cluded protocol but not in Meth- ods)

Table 47: Social – Quality assessment of intervention studies for social participation

Selection bias Performance bias Attrition bias Detection bias Summary A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Risk of all all all all Bias Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 189 RCTs Carballo- N/A N/A N/A High N/A No N/A N/A N/A N/A N/A Un- Yes Yes Yes Un- Un- Un- High García clear clear clear clear 2013 Carlson Un- Un- Un- Un- No Un- Un- Un- Yes Yes Un- Un- Yes Yes Yes Un- Un- Un- Unclear 2008 (Fried clear clear clear clear clear clear clear clear clear clear clear clear 2004) Cohen- Un- Un- Un- Un- Yes No No Un- Yes Un- Un- Un- Un- Yes Yes Yes Yes Low Unclear Mansfield clear clear clear clear clear clear clear clear clear 2015 Hughes Yes Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear 2014 clear clear clear clear clear clear clear clear clear clear clear Park 2014 Un- Un- Yes Un- Yes No No Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear Pitkala 2011 Yes Yes Yes Low No Un- Un- Un- Yes Yes No Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear Stine- Un- Un- Yes Un- No Un- Un- Un- Un- Un- Un- Un- Un- Yes Yes Un- Un- Un- Unclear Morrow clear clear clear clear clear clear clear clear clear clear clear clear clear clear 2007 Stine- Un- Un- Un- Un- No Un- Un- Un- Yes No No Un- Un- Yes Yes Un- Un- Uncle- Unclear Morrow clear clear clear clear clear clear clear clear clear clear clear aer 2008 Stine- Un- Un- Yes Un- No Un- Un- Un- Yes Yes Yes Low Un- Yes Yes Un- Un- Un- Unclear Morrow clear clear clear clear clear clear clear clear clear clear 2014 Tesky 2011 Un- Un- Yes Yes Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear 190 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 190 Evidence-based Tables

Table 48: Social – Quality assessment of qualitative studies for social participation

1. Theoretical 2. Study 3. Data 4. Validity 5. Analysis 6. Ethics Overall approach design collection 1.1 1.2 2.1 3.1 4.1 4.2 5.1 5.2 5.3 5.4 6.1 6.2 Anderson 2009 Appropriate Clear Not sure Inadequately Clear Not Poor Not Not sure Not sure Not Not Low reported sure sure re- clear ported Cattan 2011 Appropriate Clear Defensible Not clear Un- Relia- Not Relia- Not sure Adequate Yes Clear Moderate (unclear who clear ble sure ble conducted interviews) Goll 2015 Appropriate Clear Defensible Appropriate Clear Relia- Rich Relia- Convincing Adequate Yes Clear High ble ble Table 49: Leisure – Interventions to improve uptake/maintenance of leisure activities

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Matsuka Before and US 70-92 Older adults living in senior Intervention: Follow-up: 6 months Social functioning: 2003 after study apartment complexes. Program focused on the Scores on the SF-36 Health Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 191 Designing importance of participation Lost to follow-up: Survey were significantly higher a Life of N=65 in meaningful social and 40% (65 participated in social functioning following Wellness community occupations but complete pre- and participation in the program: Program Setting: Community for better quality of post test data was only pre: 69.55 (30.32); post: 77.24 life and strategies to available for 39) (24.81); p

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Zingmark RCT (pilot) Sweden Range 77- Generally well older adults, Interventions to increase Follow-up: 3, 12 months Leisure engagement outcomes 2014 82 without apparent cognitive occupational engagement All groups showed a decline in or communication prob- (participation in a wide Lost to follow-up: IG: leisure engagement between Mean: 79 lems and without major range of activities e.g. 7.3%; AG: 6.1%; DG: baseline and 3 and 12 months activity restrictions; who physical, social and leisure 17.1%; Control: 15.2% lived alone, without home activities). All interventions All participants included in At 3 months, all intervention care in an urban area were based on concepts ITT analysis. groups showed less decline in of healthy eating and leisure engagement than the At baseline, participants health promotion and were Outcome measurement: control group but significance of performed a mean of 8 delivered by occupational Leisure engagement between group differences not leisure activities therapists. All interventions measured using Modified reported involved discussion of NPS interest checklist N=177 randomised meaningful activities and (MNPS); This covers 20 Effect sizes were small. coping with age-related checklists and leisure Statistical significance of Setting: Community activity restrictions to engagement was a difference between groups was improve or maintain composite of self-report of not reported Gender: 82% women engagement in valued whether they performed activities the activity, wanted to At 12 months, the Individual Ethnicity: Not reported perform the activity and intervention and the Discussion Individual intervention importance for wellbeing group maintained less decline SES: Post high school (IG) (N=41): using a client- ratings in leisure engagement than the education: centred, goal orientated control group but significance of CG: 26.7%; IG: 31.7%; approach. Also involved Activities of daily living between-group differences not AG: 30.6%; DG: 42.1% engagement in activities (ADL) ability reported when identified ADL Activity group (AG) All the intervention groups and (N=49): OT-led group control showed decline from format and included baseline to T3 engagement in activities such as walking with Statistical significance of a pedometer, cooking, difference between groups was visiting a coffee shop not reported

Discussion group (DG) (N=41): participated in one lecture and discussion group about engagement and healthy ageing. OT led group format

Control group (CG) (N=46): No intervention Table 50: Leisure – Interventions to improve uptake/maintenance of leisure activities with cognitive outcomes

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Volunteering Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 193 Carlson RCT (pilot) US Mean: In- Older adults Experience Corps places Follow-up: 4 to 8 months Cognitive outcomes: (age and 2008, tervention: older volunteers in public education adjusted) Fried 2004 (cluster 70.1 (6.42) N=149 randomised elementary schools in Lost to follow-up: (no cog out- randomised Control : roles designed to meet Intervention: 11.4%; Trail-making test comes) by school) 68.4 (5.15) Baseline MMSE: schools’ needs and control: 17.2%. Part A: No nificant difference Experience Intervention: 24.96 (3.45); increase the social, between groups Corps Control: 25.3 (2.60) physical, and cognitive Outcome measurement: activity of the volunteers memory, executive Part B: Significant difference Setting: Community function (EF), and between intervention and control Community-based psychomotor groups at follow-up (p<0.05). Gender: Intervention: program designed to Speed: Intervention group improved by 83% female; Control 93% increase cognitive and Trail Making Test: 1.3 secs from baseline; control female physical activity in a social, Parts A and B group declined 21.7 secs from real-world setting Word list memory: baseline Ethnicity: Intervention: Immediate recall 94% black; Control: 95% Intervention (N=70): Delayed recall Word list memory black Participants randomized Rey-Osterrieth: No significant difference between to EC trained in teams to Copy score groups for immediate or delayed SES: Mean years help elementary school Delayed recall recall education: Intervention: children with reading 11.9 (2.54); Control: 11.2 achievement, library Rey-Osterrieth CFT (2.66); 38% had less than support, and classroom Copy score: No significant high school education behaviour for 15 hr/week difference between groups Delayed recall: Significant Control (N=58): Wait-list difference between intervention control and control groups at follow-up (p<0.05). Intervention group improved by 1.0 points and control group declined by 0.7 points 194 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 194 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Learning new leisure skills e.g. digital photography, quilting, music education and playing an instrument Bugos 2007 RCT US 60-85 Musically inexperienced Intervention (N=16): Piano Follow-up: 6 months (end The experimental group community-dwelling older instruction: A broad-based intervention) and 3 months significantly improved adults music education program, later (9 months from performance on the Trail Making including instruction with baseline) Test and Digit Symbol measures N=31 randomised progressive difficulty in as compared to healthy controls, musical performance, Lost to follow-up: but not other cognitive measures Setting: Community technical motor/dexterity Intervention 23.8%; control exercises, and music 16.7% Gender: Intervention: theory. Lessons were half 81.0% female; control: an hour each week plus Outcome measurement: 77.8% female independent practice for Digit Symbol, Digit Span, a minimum of 3 hours a Block Design, Letter Ethnicity: Not reported week Number Sequencing, and Trail Making SES: Mean years Control (N=15): No education: Intervention: intervention 16.5; control 16.3 Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Park 2014 RCT US 60 to 90 Older adults without 6 lifestyle conditions: Follow-up: 3 months Episodic memory: significant Synapse cognitive impairment comparing productive intervention and follow-up differences between photo Project Mean 71 engagement with receptive group and placebo (p=0.01) and MMSE at baseline 26+ engagement (did not Lost to follow-up: for dual condition and placebo actively acquire new skills) (p=0.05). No significant effects Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 195 Setting: Community which involve different Outcome measurement: were found for other conditions cognitive function Processing speed, (quilt or social) compared to Gender: 73.9% female Participated in all assessed using digit- placebo conditions for an average comparison tasks with Ethnicity: 14.2% ‘minority’ of 16.51 hr a week for three, six, and nine items But this doesn’t look right from (not described) participants 3 months (except for no Mental control, assessed graph? treatment control) using Flanker Center SES: Only those with 10th Productive engagement: Letter, Flanker Center grade + education were Photo (N=29): novice Arrow, and Flanker Center included participants learned digital- Symbol tasks and the photography and computer Cogstate Identification skills using photo-editing Task software Quilting (N=35): novice Episodic memory, participants learned how to assessed using the design and sew quilts immediate recall Dual condition (N=42): section of the modified participants spent half of Hopkins Verbal Learning the 3-months doing quilting Task, the Cambridge and the other half on Neuropsychological photography Test Automated Battery Receptive engagement: (CANTAB), Verbal Social condition (N=36): Recognition Memory Task participants engaged and the long-delay section in on-site, facilitator-led of the modified Hopkins social interactions, field Verbal Learning Task trips, and entertainment Placebo condition Visuospatial processing, (N=39): Tasks at home assessed using the that appeared to be CANTAB Spatial Memory beneficial to cognition but Task, the CANTAB had no substantiated link Stockings of Cambridge to cognitive improvement Task, and a modified (e.g., listening to classical version of Raven’s music, completing word- Progressive Matrices meaning puzzles) No treatment control (N=40) 196 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 196 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Leisure activities in general Tesky 2011 RCT Germany 50+ Older adults without AKTIVA intervention (n Follow-up: 8 week Participation in the group AKTIVA dementia or cognitive = 126): Group programme intervention and post- program resulted in positive study Mean: 72 impairment of cognitively stimulating test conducted 1 week effects on cognitive function (7) leisure activities (8 weekly after (week 9); 2 booster and attitude toward aging. Older N=307 randomised sessions and two booster sessions conducted at adults (≥75 years) showed sessions after a break 27 to 28 weeks and then enhanced speed of information Setting: Community of 4 months). Sessions follow-up tests conducted processing (by TMT Version A) included information and 29 weeks after start of (F = 4.17*, p < 0.05); younger Gender: 72.3% female education about dementia intervention participants (< 75 years) showed and risk factors and the an improvement in subjective Ethnicity: ‘German importance of a healthy, Lost to follow-up: N=67 memory decline (by MAC-Q) (F = ethnicity’ reported for those active lifestyle, awareness (21.8%) withdrew from the 2.55*, p < 0.05) included in analysis and motivational factors. study. Those with impaired Participants were cognition (N=28) and Additionally, AKTIVA enhanced SES: ‘Most participants instructed and motivated to N=15 Turkish participants the frequency of activities for had attended school for perform more cognitively excluded from analysis leisure activities about 10 years, and only stimulating leisure after the programme. a few participants had activities such as reading, N=208 analysed completed an academic playing games or playing university education’ music as part of their daily Outcome measurement: routine Cognitive outcomes: Mini- Mental Status Test AKTIVA intervention plus nutrition and exercise ADAS-Cog: the cognitive counselling (n = 84): in part of the Alzheimer’s addition to the cognitively Disease Assessment stimulating leisure Scale activities, participants could engage in e.g. Part A and B of the Trail- walking, yoga, gymnastics Making Test and healthy eating sessions Clinical Dementia Rating (CDR) Control group (n = 97): No intervention Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Crosswords Murphy RCT Ireland 57-90 Community-dwelling Intervention: Crossword Follow-up: 4 week After 4 weeks, the crossword 2014 older adults with normal group (N=19): asked to intervention and follow-up group had significantly higher Mean: 71.5 cognition attempt one crossword Phonemic Verbal Fluency (PVF) (SD = 8.3) daily for 4 weeks. A mix- Lost to follow-up: Not scores and for the cluster size N=37 randomised ture of simple and cryptic reported component

crosswords from a national Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 197 Setting: Community newspaper was used Outcome measurement: [F (1, 35) = 3.17, p = 0.042, Phonemic verbal fluency partial eta squared = 0.083] with Gender: 73.0% female Control (N=18): asked to (PVF): measured using a moderate effect size keep a diary in which they Controlled Oral Word Ethnicity: Not reported daily recorded three things Association Test (COWAT) for which they felt grateful SES: No difference in education between groups Note: no measures were but level not reported taken to ensure compliance by either the crossword or gratitude group

Interactive gaming (Note this only includes interactive games – solely video games will probably be under cognitive training)

Hughes RCT US Mean: In- Older adults with MCI Intervention (N=10): Follow-up: 24 weeks Adherence: The Wii group 2014 tervention: recruited from a population Group-based interactive attended an average of 23.1 78.5 (7.1); cohort study video gaming using Nin- Lost to follow-up: 10% (SD 1.1, range 21–24) sessions control: tendo Wii; 90 mins/week (1 died, 1 did not complete compared with 21.8 (SD 3.3, 76.2 (4.3) Mean MMSE at baseline for 24 weeks follow-up assessment) range 14–24) in the control 27.1 (1.8) The Wii Sports games, group; 18 participants attended including bowling, golf, Outcome measurement: at least 20/24 sessions; and 9 N=20 randomised tennis, and baseball were Cognitive function: attended all sessions the core of the sessions. Computerized Assessment Setting: Community From week 7, participants of Mild Cognitive Cognitive function: No were introduced to new Impairment (CAMCI) significant differences between Gender: Intervention: games for 15-20 mins of Subjective cognitive ability: intervention and control groups 80% female; control: 60% the session. In weeks 10 Cognitive Self-Report female and 20 participants com- Questionnaire-25 No significant differences for any peted in Wii tournaments Social functioning: other outcomes measured Ethnicity: Intervention: to encourage enhanced Cognitive Self-Report 70% white; control: 90% effort and social activation Questionnaire-25 whit. Control (N=10): Health Other outcomes; mood, education designed to pro- IADL, gait speed SES: Mean years vide a source of passive education: Intervention cognitive stimulation in a 13.8 (2.4); control 13.1 socially matched setting; (1.9) 90 mins/week for 24 weeks 198 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 198 Evidence-based Tables

Study Study Country Age Population and setting Intervention Relevant outcomes and Key results design (years) and comparator follow-up Meditation Gard 2014 Systematic Interna- Not Review aimed to examine Studies covering a range Range of outcomes N=12 studies included, of which review tional reported the effects of meditation on of meditation techniques relevant to ageing included 6 were RCTs. Preliminary cognitive decline were included: including positive effects on attention, mantra-based meditations memory, executive function, Studies in any population such as TM, Kirtan Kriya processing speed, and general included. Included studies yogic meditation (KKYM), cognition were reported across were in people with and meditations that studies. However, most studies healthy cognitive function, involve focusing on a had a high risk of bias and small poor cognitive function visual object,to Buddhist- sample sizes and dementia caregivers based mindfulness (normal function) practices, including Shamata-Vipassana,81 Setting, gender, ethnicity Zen, and other forms. and SES not generally reported for relevant Two studies did not clearly intervention studies describe the meditation techniques Table 51: Leisure – Quality assessment of leisure activity intervention studies

Selection bias Performance bias Attrition bias Detection bias Summary Risk of A1 A2 A3 Over- B1 B2 B3 Over- C1 C2 C3 Over- D1 D2 D3 D4 D5 Over- Bias all all all all Evidence-based resourceforchangingriskbehavioursinolderadults:Intervention Tables 199 RCTs Bugos 2007 Un- Un- Yes Un- No Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear clear Carlson Un- Un- Un- Un- No Un- Un- Un- Yes Yes Un- Un- Yes Yes Yes Un- Un- Un- Unclear 2008 (Fried clear clear clear clear clear clear clear clear clear clear clear clear 2004) Hughes Yes Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear 2014 clear clear clear clear clear clear clear clear clear clear clear Murphy Un- Un- Yes Un- Yes Un- Un- Un- Yes Un- Un- Un- Un- Un- Un- Un- Un- Un- Unclear 2014 clear clear clear clear clear clear clear clear clear clear clear clear clear clear clear Park 2014 Un- Un- Yes Un- Yes No No Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear Tesky 2011 Un- Un- Yes Yes Yes Un- Un- Un- Yes Un- Un- Un- Yes Yes Yes Un- Un- Un- Unclear clear clear clear clear clear clear clear clear clear clear clear Zingmark Yes Un- Yes Un- Yes No No Un- Yes Un- Yes Un- Yes Un- Yes Yes Yes Un- Unclear 2014 clear clear clear clear clear clear clear Non-randomised studies Matsuka N/A N/A N/A High N/A Un- Un- Un- N/A No No Un- Yes Un- Yes Un- Un- Un- High 2003 clear clear clear clear clear clear clear clear 200 Evidence-based resource for changing risk behaviours in olderadults:Intervention risk behaviours resource forchanging 200 Evidence-based Tables

Table 52: Leisure – Quality assessment of systematic reviews of leisure activity interventions

AMSTAR (High= 8-11, Good= 5-7, Poor= <5)

Author (Year) 1 2 3 4 5 6 7 8 9 10 11 Rank- ing Gard 2014 N Y Y Y Y Y Y Y N/A N/A Y +