Intergenerational Programmes bringing together community dwelling non-familial older adults and children: A Systematic Review

Abstract

Background:

Social isolation is associated with an increased risk of adverse health outcomes, including functional decline, cognitive decline, and dementia. Intergenerational engagement, i.e. structured or semi structured interactions between non-familial older adults and younger generations is emerging as a tool to reduce social isolation in older adults and to benefit children and adults alike. This has great potential for our communities, however, the strength and breadth of the evidence for this is unclear.

We undertook a systematic review to summarise the existing evidence for intergenerational interventions with community dwelling non-familial older adults and children, to identify the gaps and to make recommendations for the next steps.

Methods:

Medline, Embase and PsychInfo were searched from inception to the 28^th Sept 2020. Articles were included if they reported research studies evaluating the use of non-familial intergenerational interaction in community dwelling older adults. PROSPERO registration number CRD42020175927

Results:

Twenty articles reporting on 16 studies were included. Although all studies reported positive effects in general, numerical outcomes were not recorded in some cases, and outcomes and assessment tools varied and were administered un-blinded. Caution is needed when making interpretations about the efficacy of intergenerational programmes for improving social, health and cognitive outcomes.

Discussion:

Overall, there is neither strong evidence for nor against community based intergenerational interventions. The increase in popularity of intergenerational programmes alongside the strong perception of potential benefit underscores the urgent need for evidence-based research.

1. Background

Social isolation is associated with a cluster of negative outcomes in older adults. These include increased risk of poor physical and mental wellbeing, and cognitive decline and dementia in later life (Alun and Murphy, 2019; Lara et al., 2019; Penninkilampi et al., 2018; Siette et al., 2020). Conversely feeling younger and having more extensive social networks is associated with greater subjective wellbeing and cognitive health (Hajek and König, 2020; Siette et al., 2020). Two recent systematic reviews with several constituent meta-analyses provide further evidence, reporting consistent findings across different types of social measures (social networks, support, and activity) and cognitive assessments (Evans et al., 2019; Penninkilampi et al., 2018). Further, meta-analyses to examine the risk of bias show continued associations between higher levels of social isolation and increased risk of developing dementia or cognitive decline amongst studies with long follow-up (>10 years) thus implying no impact from reverse causality and further analyses show that this remains robust after adjustment for depression (Evans et al., 2019; Penninkilampi et al., 2018). The impact of social isolation and loneliness on health is an important and pressing issue and an area where research is needed (Mehrabi and Béland, 2020).

Social isolation is particularly relevant to our ageing society. Older adults are at increased risk of developing a range of diseases and conditions, including both dementia and the cognitive decline that precedes it (Fratiglioni et al., 2000). They are also at greater risk of experiencing social isolation due to decreased network size and activity levels (Fakoya et al., 2020). For cognitive health in particular, with no disease modifying drugs available, interest has grown in interventions that reduce exposure to risk factors for cognitive decline or dementia (Ngandu et al., 2015).This includes results from clinical trials in older adults where increasing social interaction has been associated with better cognitive performance, in particular, better performance on verbal tasks (Dodge et al., 2015; Mortimer et al., 2012; Pitkala et al., 2011).

One of the fastest growing types of interventions that may bring multiple benefits, including but not limited to reduced social isolation, is the bringing together of older adults and children to support active non-familial intergenerational interaction. This method of intergenerational interaction is routinely referred to as intergenerational practice. Intergenerational practice is defined as an active process that ‘aims to bring people together in purposeful, mutually beneficial activities which promote greater understanding and respect between generations and contributes to building more cohesive communities. Intergenerational practice is inclusive, building on the positive resources that the younger and older have to offer each other and those around them’. (Beth_-Johnson_Foundation, 2009; MacCallum J, 2006)

Intergenerational practice may be an apposite solution for modern society. Increasingly, family structures involve two working parents, geographically distant grandparents, and the use of childcare and aged care (Warren D, 2020). With an increasingly large older adult population and longer life expectancy, today’s population interacts on a uni-generational level more than ever before (United Nations Department of Economic and Social Affairs; Warren D, 2020). Intergenerational approaches are particularly relevant for older adults and children since both childhood and older adulthood are characterised by greater leisure time, need for close long-term connections, and a proclivity for learning and teaching (Uhlenberg, 2000). These qualities make for a unique, natural and reciprocal care partnership between children and older adults. Recent reviews of intergenerational interventions have reported potential benefits such as reduced age stereotyping and intergenerational solidarity, improved empathy and self-esteem in children, increased self-esteem, improved mood, reintegration into family and community life and, reduced loneliness in older adults (Giraudeau and Bailly, 2019; MacCallum et al., 2010). Despite this potential, rigorous research in this area is sparse and although over 90% of older adults reside in the community, research to date has focused almost entirely on older adults in residential aged care facilities (Giraudeau and Bailly, 2019; Radford et al., 2018). Thus we aimed to (i) systematically review and summarise the existing evidence for non-familial intergenerational interventions on cognitive and wider health outcomes for community dwelling older adults, and (ii) to identify the gaps and make recommendations for the next steps needed for robust and relevant intergenerational interventions in the future.

2. Methods

Taking a focus relevant to the majority of the older adult population, that is, those who reside in the community, we used systematic review methodology (Grant and Booth, 2009) to identify the available evidence reporting on non-familial intergenerational practice (interventions or programmes) in community dwelling older adults and children. The databases Medline, Embase and PsychInfo were searched from inception to the 28^th Sept 2020. A two pronged approach was taken to searching the literature. A targeted search with search terms including ‘community’, ‘learning’, ‘child’, ‘preschool’, ‘kindergarten’, ‘nursery school’, ‘older adult’, ‘aged’, ‘elderly’ and ‘intergen*’ was supplemented with broader searches tailored to each database and using MeSH or equivalent terms. This ensured identification of the widest possible evidence base. Search terms were informed by prior literature (e.g. Ronzi, Orton, Pope, Valtorta, & Bruce, 2018) and consultation with experts in the field and searches were limited to English language and humans (Appendix 1 for details). We preferentially selected articles focused on intergenerational practice with mutual benefit rather than interventions designed with a predominance of benefit for one group or the other. Articles were included if they (I) reported on results from research studies evaluating the use of non-familial intergenerational interaction (pilot, cross sectional or longitudinal studies, quantitative or qualitative) (II) in community dwelling older adults (defined as those aged 65 and over or self-identifying as older adults) without a diagnosis of dementia, and children, (III) referred to structured activities where engagement was designed for both older and younger participants i.e. not where older adults were primarily in a supervisory or teaching role, and (IV) with some report of experimental or learning outcomes for the older and younger participants. We focused on complete reports of original research, excluding research protocols, reviews, editorials, commentaries and conference abstracts. Research reporting on familial intergenerational interaction was also excluded.

There were three independent analysts (RP, JP, NE). The lead analyst carried out the literature searches. All identified abstracts, or titles where abstracts were unavailable, were double reviewed and a list of potentially relevant references compiled independently by two analysts. For the supplementary searches an additional title screening stage was employed (Figure 1). These lists were compared, and differences were resolved by discussion. Once the list of possible references was agreed upon, full text articles were obtained, independently read, and assessed for relevance. Data were extracted by one of the analysts and checked against original manuscripts by a 2^nd analyst. Standard extraction tables were used and data was extracted on study design, aims, participants, recruitment source, frequency and duration of the intervention, analysis and outcomes. Narrative synthesis was used and risk of bias was evaluated. Where multiple supplementary articles were identified as reporting on the same study the data were extracted for the main articles and with the supplementary articles referenced.

To supplement the systematic review of the empirical research evidence, the identified articles were examined to (I) inform a complementary synthesis of the perceived potential benefits and risks and (II) to provide practical recommendations for intergenerational interventions and research. The review is registered on PROSPERO with the registration number, CRD42020175927.

3. Results

Searching yielded 328 abstracts from the targeted and 173 abstracts from the supplementary searches and 57 full texts were assessed. Twenty-four were excluded after full text screening, chiefly due to the selection of older adults in residential care, or where it was unclear whether those in residential care were included, selection of college students, where studies did not report on non-familial intergenerational interaction or did not report on experimental or learning outcomes. See flow chart supplementary Figure A1 for details. Twenty articles were included and extracted with a further 13 articles retained as reporting supplementary information relating to two of the included studies (Table 1). These two studies the Research of Productivity by Intergenerational Sympathy (REPRINTS) (Yasunaga et al., 2016) and the Baltimore Experience Corps (Carlson et al., 2008; Rebok et al., 2004) both focused around older adults helping children in schools with reading and other skills but were included as they presented themselves as designed to benefit both age groups.

Table 1.

Study Characteristics

Study	Study Design	Goals	Older Adults (n) (age yrs)	Youth/ Children (n)(age yrs)	Inclusion criteria	Source	Results	Key findings
Sun et al 2019 Young-Old Link and Growth Intergenerational Program	Quasiexperimental design, with preand post-test	Examine effectiveness of an intergenerational programme. Aims: positive intergenerational attitude, improved sense of comfort while interacting in cross-age groups, positive intergeneration interaction.	I: n=73 M=72.54 (SD=7.18) yrs C: n=77 M=73.95 (SD=8.70) yrs	I: n=77 M=16.30 (SD=1.47) yrs C: n=84 M=16.65 (SD=1.05)yrs	Older adults with normal cognition, good mobility, capable of communication, age>=60. Children aged 12-18 years.	10 Community older adult social service units and 3 secondary schools, Hong Kong.	1. Change in intergenerational attitude: treatment x time Old: F(l, 147)=137.33, p <.001, n2p=.48 Young: F(l,156)=45.48, p <.001, n2p=.23 2. Intergenerational comfort: treatment x time Old: F(l,147)=6.70, p <.01, n2p=.04 Young: F(l,156)=8.26, p <.01, n2p=.05 3. Changes in intergenerational interaction - behaviours pre - post test: Visual attention Old: t=2.26, p<.005, d=0.29 Young: t=3.70, p<.001, d=.45 Initiate conversation Old: t=3.42, p<.0.01, d=0.46 Young: t=2.04, p<0.05, d=0.25 Touch Old, t=2.28, p<0.005, d=.30 Young: ns No pre/post-test changes for smile, assistance, and encouragement, in either group	Significant improvement in intergenerational attitude, comfort, and some measures of intergenerational interaction across groups.
Hanmore-Cawley & Scharf 2018	Quasiexperimental embedded mixedmethods design	To investigate the potential of intergenerational learning to develop civic literacy in young children.	N=34 ≥65 yrs	I: n=73 C: n=30 9–10 yrs	N/A	Community snowball sampling and 5 participating primary schools, Ireland.	Aggregate civic literacy, between groups x time: t (101)=9.68, p<.01, mean difference=0.74, 99%CI(.54, .94) Self reported pre-post tests: 1. Personal responsibility I: t(72)=-8.59, p<.01 C: t(29)=1.79, p=.084 2. Civic responsibility I: t(72)=-22.16, p<.01 C: t(29)=-0.39, p=.697 3. Leadership I: t(72)=-11.55, p<.01 C: t(29)=-0.64, p=.53	Through collaborating with older adults, regardless of activity, children’s civic literacy improved.
McConnell & Naylor 2016 International Physical Leadership program (IPAL)	Mixed-methods design	To determine the feasibility of an Intergenerational-Physical-Activity Leadership intervention. Programme: providing opportunities for older adults and students to engage in activities enhancing physical activity and intergenerational interaction	N=9 M=72.35 (SD=5.52) yrs	N=22 M=9.86 (SD=2.13) yrs	N/A	Community 55+ association, newsletter, posters in co-located space with, primary school, Canada.	Post intervention (all scores out of 5): • Satisfaction score Old: M=3.5 (SD=.58) Young: 4.45 (SD=.67) • Usefulness score Old: M=3.5 (SD=.58) Young: 3.72 (SD=.98) • Confidence score Old: 4.25 (SD=.5) Young: 4.23 (SD=1.15) • Physical activity and environmental context Greater organised activities, pre/post intervention: t (48)=-8.03, p<.001 Greater use of playground equipment pre/post intervention: t(48)=−9.13, p<.001 • Physical activity pre/post intervention: ns • Intergenerational interaction:% of interactions: Old 56%, Young 18% • Peer interaction: Old 20%, Young 65% • Positive affect/partner positive affect: 99–100% • Confidence Focus groups identified emerging themes: rewarding experience, enthusiasm, energy - as positive benefits and implementation facilitators	Both groups had high satisfaction levels, leadership, confidence, and intention to use learning, and displayed positive affect during intergenerational interactions. Modifications were recommended by study participants for future IPAL interventions
Kaplan 1997 Neighbourhood 2000 Programme	Case study: observation	To provide case study data indicating that i nter ge nerati onal community service programs developed along sound program planning principles invoke powerful communication and social support dynamics that impact participants and generate community improvements.	N=48 per project Unclear	Unclear 4th, 5th, or 6th graders	N/A	7 neighbour hoods, USA	No numerical results.	Participants learned to question negative age stereotypes, that generations are interdependent and that they have shared concerns about their communities. They learned critical thinking, communication, decision-making skills, and art of negotiation.
Peterat & Mayersmith 2006	Observational study	Create a space where i nter ge nerati onal conversation about land, food, community, society, and environmental issues would begin and be sustained in hope of reestablishing biophilic sensibilities’.	N=7 Unclear	N=7 Grade 7	N/A	Adults: community- based. Children: private girls’ school, Vancouver, Canada.	No numerical results	Older volunteers experienced rewards beyond those of grand parenting relationships. Cross generational reciprocal learning and connection took place. Physical outdoor environment generated teaching and learning.
Meschel & McGlynn 2010	Pre-post-test design (Randomisation of adolescents only)	Establish a cross-age program to assess youngsters’ attitudes and stereotypes of older persons, by improving the measurement of attitude by employing a consistent conceptual framework	I:n=17 >=60 yrs	I: n=21 C: n=21 Didactic C: 21 (older persons with no contact with young) 11–13 yrs	N/A	Older adults: from senior citizens centre. Students: 3 classrooms in a public middle school. Midwest, USA.	Attitudes • Young: condition x time: F (2,60)=3.47, p <.05 +ve change in intervention group • Old:, pre/post-test: t(16)=3.76, p<0.01) +ve change Life satisfaction • Young: condition x time: ns • Old: pre/post-test: t(16)= 2.16, p<.05. =ve change Young rated stereotypes: Condition x time: F(l, 60)= 4.16, p<.05	Attitudes in the intervention group and older adults improved. Older adult life satisfaction improved. Older persons attitudes became more positive after contact. Shift toward more negative stereotypes of older adults amongst students
Aday, Rice & Evans 1991 Aday, Sims, McDuffie &Evans 1996 Aday, Adey, Arnold & Bendix 1996 Intergenerational Partners Project (IPP)	Mixed- methods experimental and qualitative study 5 year follow-up	Aim to promote mutually beneficial interactions between seniors and younger students: to enhance learning and growth of children and to combat age stereotypes, lessen psychological distance felt by students towards the older persons, to improve wellbeing of frail older adults	N=25 71 (range 65–78) yrs	I: n=25 C: n=25 4th grade I: n=21 C: n=20	No restrictions on older adults	Older adults: senior citizens’ centre Children: Bright Elementary School, Chattanooga, USA	Children’s Perceptions of Aging and Elderly Inventory (CPAE): I vs C • Post-intervention: I: M=86.62 (SD 5.82), C: M=80.40 (SD 5.25) t=2.79, p < .011 • 1 year follow-up I: M=87.95 (SD 5.83), C: M= 81.52 (SD 5.96),t=2.12, p < .02Perceptions of elderly more favourable in I group. Semantic Differential Scale: I vs C • Post intervention I: M=111.68 (SD 12.80), C: M=102.52 (SD 11.57) t=2.57, p<.02 • 1 year follow up I: M=113.25 (SD 10.22), C: M=103.78 (SD 11.02), t=3.06, p<.01 Attitude towards elderly more favourable in the I group. • 5 year follow-up CPAE I: M=83.88, C: M=78.70, t=2.72, p< .01 Semantic Differential Scalel: M=107.75, C: M=98.80, t=3.39, p< .01 Year 1 to year 5 comparisonCPAE: t=1.10. p> .05, no significant decline in attitude over 5 yearsSMD: t=0.85, p> .05, no significant difference in attitude over 5 yearslntervention group only, pre and postCPAE:I (pre) M=77.08 (SD 5.8); I (post) M=85.20 (SD 7.5), t=2.72, p<.01 Student view of aging process (scored by expert):I (pre) M=8.92 (SD 3.80), I (post) M=11.4 (SD 2.75), t=4.90, p <. 01 Si gnifí cant improvement in how youth perceived elderly and aging process.	Intervention improved children’s perceptions of the elderly and this remained significant at 1-year follow-up. There was no significant decline or change in attitude over 5 years.
Morita & Kobayashi 2013	Structured observational study, using time sampling	To compare the changes in visual attention, facial expression, engagement/behaviour, and intergenerational conversation in older participants in performance-based and social-oriented IG programs to determine a desirable interaction style for older adults	Total: n=25 M=85 (SD=7.5) yrs	N~20 per group 5–6 yrs	Adults: Physically independent participants that could walk with assistance, normal cognitive function according to eligibility for long-term care insurance national criteria (independent and mostly independent)	Older adult day care centre and nursery school and kindergarten, Tokyo, Japan.	Constructive behaviour and intergenerational conversation both higher in social-orientated IG programming group: both p<.001 Smiles, between groups: p=0.129, ns Visual attention higher in performance-based IG programming group. p=0.025	Performance of older adults varied by the type of program they experienced.
Bales, Eklund & Siffin 2000 Learning and Growing Together Program	Pre- and postsurvey design. Qualitative analysis of 4th Graders journals.	To narrow the generation gap and foster relationships between young and old by assessing children’s descriptors, thoughts, and feelings of older adults, through a letter exchange program.	N=23 M=72.4 yrs (Older adults paired with 4th graders M=73.4 yrs)	2nd graders: n=22 M=7.9 yrs 4th graders: n=20 M=9.5 yrs 5th graders: n=21 M=10.9 yrs	N/A	Community-based older adults and children from parochial elementary school in the Midwest, USA.	Positive descriptors 2nd graders: χ2(3, N=122)=21.912, p < .000 4th graders: χ2(3, N=120)= 14.996, p < .002 5th graders: ns Negative descriptors decreased in all grades: ns Physical descriptors: ns 4. Other descriptors: ns	Significant increase in positive descriptors observed and nonsignificant decrease in negative descriptors
Griff et al 1996	Observational design, with interviews.	To extend the knowledge base of professional on intergenerational planning by differentiating between types of activities that can be used successfully with young children and different groups of older adults in intergenerational settings.	Frail: n=10 (5 in results) M=79 yrs Comm-unity: 5 M=70 yrs	N=10 M=4.0 yrs	Frail cognitively intact older adults in need of assistance with a limited number of daily living skills, community elders highly independent who maintain their own residences, and older adults diagnosed with Alzheimer’s disease in early or mid-stages of dementia.	Community older adults, older attendees and children at an ongoing life care facility (day care), Ohio, USA	No numerical results Short-term, gross motor, and simple activities ranked highest by frail older adults and children. Lower ranking activities included fine motor, cognitive and structured tasks.	No numerical results Short-term, gross motor, and simple activities ranked highest by frail older adults and children. Lower ranking activities included fine motor, cognitive and structured tasks.
Chua et al 2013	Long-itudinal study	To investigate effects of video game play on i nter generati onal perceptions compared with interaction in daily routine activities	I: Video game group: n=19 C: Non-video game group: n=18 Video game group: M=75.42 (SD=8.15) yrs Non-video game group: M=76.50 (SD=7.33) yrs	I: Video game group: n=19 C: Non-video game group: n=18 Video game group: M=16.74 (SD= 0.65) yrs Non-video game group: M=17.67 (SD= 0.84) yrs	N/A	2 older senior activity centres and two junior colleges and a polytechnic, Singapore	1.Intergroup anxiety a) pre-post test: reduction overtime, F(1.72)=9.98, p<0.001, n2=0.12 b) I vs C: greater reduction in I group, F(l,72)=4.04, p<0.05, n2=0.05 2.General attitudes towards other group a) pre-post-test: positive change over time F(l,72)=33.59, p<0.01, n2=.32 b) I vs C: positive change in I group F(l,72)=11.56, p < .01, n2=.14 3. Attracti on toward interaction partner (Social Dimension Attraction Scale) a) pre-post test: positive change in attraction F(l,72)=12.18, p < 0.01, n2=0.15 b) I vs C:greater attraction in I partners, F(l,72)=7.29, p < 0.01, n2=0.09 c) Interaction types and changes over time F(l,72)= 0.51, ns Mediation effect of attraction on a) intergroup attitudes: z=2.36, p < .05, R2=.32 b) inter group anxiety: z=− 2.35, p < 0.05	Significant increase in positive attitudes was driven by effect from the video game condition. Attitudes accounted for 32% of variation in inter group attitudes. Video games facilitated positive attitudes and reduced anxiety between the two age groups Increase in interpersonal attraction observed across all participants but interaction effects not significant. Reduced intergroup anxiety observed was driven by effect from the video game condition. Intergroup attraction accounted for 20% of variation in inter group anxiety.
Kuehne 1988	Ethno-graphic observational study	Study of interactions between intergenerational group members who participate in supervised activities in a particular program.	N=7 > 65 yrs	N=7 4–5 yrs	Older clients who are coherent and/or expressed interest in program; Children who are verbal and extroverted for their ages.	Adults: day care facility at a church, Chicago, US. Children: community centre.	Observed number of interactions between older adults and children: Positive interactions, e.g. giving affection,n= 462 Negative interactions e.g. aggression, n=14	Mostly positive interactions observed suggesting there is order and purpose in intergenerational group function.
Hannon & Gueldner 2007	Randomised experimental design	To investigate short-term quality intergenerational contact on the attitude of children toward older adults, in a 9-hour intergenerational program and how demographic variable moderate attitudinal changes	N=17 65–97 yrs	N=67 (divided into Intervention and Control groups, numbers not given for separate groups) 6–12 yrs	Children 6–12 years old, able to speak and understand English, willing to participate in 9 hours of intergenerational programming and willing to be videotaped.	Summer camp and senior group of YMCA in Pennsylvania, USA.	Children’s attitudes towards ageing 1. Polozzo’s Revised Version of the Semantic Differential Pre/post-treatment difference: F=4.93, p=0.03 Intervention/Control group M=53.54 (SD =17.68)/ M=63.47 (SD=21.93), with intervention group having the more positive attitude. 2. Self-reported time spent with older adults and treatment group were predictors of children’s post-test attitude to older adults: F= 5.42, df =2/63, p = .007, R2=.147 3. Questionnaire: Whether the activities with older adults made a difference to their attitude: No difference; I: 6.98%, C: 51.28%, Made a difference: I: 58.14%, C: 25.64%, p<0.05	More positive attitude toward the elderly in intervention group than control group when baseline controlled for.
Teater 2016 Time after Time Programme	Cross-sectional study	To determine the perceptions of older adults on the contribution to their health, self-development and community building, and thoughts, feelings and learning, through participation in an intergenerational programme	N=70 M=74 (7.77) range 58–92	Number not stated	Not stated	Adult volunteers participating in Golden Oldies Charity singing groups. Children from 75 schools in	Higher mean = more positive response Index of Arts as Self Developing Activities: M=24.37, range 12–30 (potential overall score 6–30) Index of Arts as Self-Health Enhancers: M-25.91, range 20–30 (potential overall score 6–30) Index of Arts as Community Builders: M=17.71, range 13–20 (potential overall score 4–20) Age x Score; Time spent in IGP x Score; Type of contact with children x Score: all non-si gnifi cant Nearly 73% rated experience as excellent, 27.5% rated it as good. 4 emerging themes: happy to take part/ made to feel welcome/ student behaviour better than expected/ interacting between age groups worked well.	Older adults found participation beneficial, it provided greater understanding and respect between the generations and an opportunity for mixing with younger participants.
Yasunaga et al 2016 (Supplementary findings e. g. Sakurai 2016, Sakurai 2016, Murajama 2015, Fujiwara 2009) Research on Productivity through Intergenerational Sympathy (REPRINTS)	Non-randomised trial design	Health promotion programme to address older adults’ psychological, cognitive, social and physical needs through intergenerational approaches. Programme intervention – picture books to connect children and older adults, controls-conventional activities.	Baseline Intervention I:n= 67 C: n=74 Follow up 1 I: n=56 (Ia=39 intensive volunteers) C: n=66 Follow up 2 I: n=53 C: n=60 I: M=68.2 (SD=6) yrs C: M=68.7 (SD=4.8) yrs	I: n=470 (one school) 9–12 yrs	Independent in 6 basic activities of daily living.	Community via newspapers/newsletters, elementary schools in 3 target cities, Japan	• Old: self-rated health la vs C:: group x time: p <.01 • Intergenerational interaction, frequency of exchanges with children la vs C, group x time p <.01 • Children’s emotional image of older adults: Group (high/low frequency of exchanges with older adults) x time: Subscale evaluation: high frequency group score remained high, low frequency group score reduced, p<0.05Subscale potency/activity, ns • Long term older adult mental health coherence, sense of meaningfulness x time: improved, p<.05 • Depressive mood, mediating effect meaningfulness, p<.05	Positive effects on self-rated older adult health and children’s sense of participation in community shown for continuous and regular programme.
Rebok et al 2004 (reported on children) Carlson et al 2008 and Hong 2010 (reported on older adults) Hong & Morrow-Howell 2010 Baltimore Experience Corps (supplementary findings e.g. Fried 2004, Carlson 2015, Gruenwald 2016, Fried et al 2013)	Randomised pilot trial (children randomised by school)	Older adults received training and volunteered in schools to help children with reading, conflict resolution etc: Aim: to evaluate whether older at risk adults increase their physical activity levels and cognitive performance as a result of participating in the program, and whether child academic achievement improved, over one year.	I: n=70 (FU: 62) C: n=58 (FU 48) I: M=70.1 (SD=6.42) C: M=68.4 (SD=5.15) yrs	I:n= 688 C: n=506 M=8.3 (range 5.3–12.5) yrs	Older adult volunteers ≥60 years, able to travel to the schools, MMSE ≥24 if 12 years of education else MMSE 20–23 and completion of the trail making test (TMT) and willing to commit 15 hrs/week: Children attending schools and in grades kindergarten, 1, 2,3.	Older volunteers recruited through mailings, radio, and community events. Children attending 6 public elementary schools in Baltimore, USA.	Older adults • Trail Making Test: Part A: I and C both improved over time: p<0.05. • Part B: I improved by 8%, C slowed by 13%, p<0.05. • Rey Osterrieth: Delayed Recall: I improved by 9%, C slowed by 10%, p<0.05 • Word List Memory: Immediate Recall and Delayed Recall, and Rey Osterrieth: Copy Score ns.. Children • Alphabet recognition: Grade x treatment x time Grade: F(l,401)-41.1, pcO.OOlTreatment F(l, 401)=8.54, p<0.004Time F (1, 401)=8.1. p<0.005Grade x Time F(l,401)=9.67, p<0.002Treatment x Time F (1,401)2.5, p=0.11 ns • Reading subtest of CTBS School level comparison of grades 1–3 over 3 years t(4)= 3.8, p>0.44 ns • MSPAP reading scores School level analysis, grade 3 only, t(4)−3.8, p<0.02	No strong results but some indication that such an intervention may be positive for both adults and children.
	Pre/post design	Aim: to compare the health of older adults in the program for two years matched with those from a Control in Health and Retirement Study	I: 185 recruited, 167 in study C: 167	N/A	Adults >50 years in the program for the first time in 2006/7	As above	Depression: pre and post I: 5.08 (SD 2.28); 4.14 (SD 3.15), t=4.68, p<.001 Depression decreased in I and became worse in C group Functional limitations, pre and post I: 2.06 (SD 2.16); 1.71 (SD 1.61), t=2.02, p<.05 C: 2.48 (SD 2.26); 2.80 (SD 2.45), t=2.32, p<.05 Limitations decreased in I and became worse in C group Self-rated health: pre and post I: 3.64 (SD 0.96): 3.63 (SD 0.94), t=0.69, ns C: 3.44 (SD 0.85); 3.38 (SD 0.98), t=0.97. ns After adjustment for pre-test health status, socio-economic status, volunteer history and boosted propensity score: I experienced more positive changes in functional limitations and depressive symptoms at post-test (both p<0.0001) than C. Health status difference p<0.09.

Note. ‘d’ denotes Cohen’s d where effect sizes of ≥0.8 are considered large; ‘I’, denotes Interaction group and ‘C’ denotes control group; M’ demotes mean; ”MMSE’ denotes Mini-Mental State Examination; ns’ denotes not significant; ‘n2p’ denotes partial Eta-squared; ‘N/A’ denotes not available; ‘SD’ denotes standard deviation

Table 1 and supplementary Table A1 show the characteristics of the included studies.

The vast majority were designed to focus on outcomes related to the attitudes, behaviours, and perceptions of the older or younger groups rather than on social isolation and health outcomes. For both of these studies those that did include health-related measures assessed aspects of physical, mental or cognitive function and self-rated health (Carlson et al., 2008; Hong and Morrow-Howell, 2010; Morita and Kobayashi, 2013; Rebok et al., 2004; Yasunaga et al., 2016). There were also two studies that provided outcome data on the impact of intergenerational practice but were designed primarily to develop understanding of intergenerational practice delivery (Griff et al., 1996; McConnell and Naylor, 2016).

Study recruitment employed newspaper or newsletter advertisements, snowball sampling and tapping into day centre or social service units for the older adults. Child participants were recruited primarily from schools (11 of 16 programmes) a kindergarten (Morita and Kobayashi, 2013), a summer camp (Hannon and Gueldner, 2008) and the community (Griff et al., 1996; Kaplan, 1997; Kuehne, 1988).

Study size also varied, from 14 (Kuehne, 1988; Peterat and Mayer--Smith, 2006) to over 1000 (Carlson et al., 2008; Rebok et al., 2004) and follow up from four weeks (Hannon and Gueldner, 2008) to two years (Yasunaga et al., 2016).

Just under half the studies could be classified as primarily observational or case study designs (Aday et al., 1996a; Aday et al., 1991; Aday et al., 1996b; Griff et al., 1996; Kaplan, 1997; Kuehne, 1988; Morita and Kobayashi, 2013; Peterat and Mayer-Smith, 2006; Teater, 2016) with a similar number employing a more explicitly quasi-experimental approach (Bales et al., 2000; Chua et al., 2013; Hanmore-Cawley and Scharf, 2018; McConnell and Naylor, 2016; Meshel and McGlynn, 2004; Sun et al., 2019). Three studies specified having taken a mixed methods approach (Aday et al., 1991; Hanmore-Cawley and Scharf, 2018; McConnell and Naylor, 2016) and three employed randomised and non-randomised trial designs (Carlson et al., 2008; Hannon and Gueldner, 2008; Rebok et al., 2004; Yasunaga et al., 2016). The studies using trial designs also varied. Hannon and Gueldner (Hannon and Gueldner, 2008) recruited 17 older adults and randomised 67 children aged 6–12 years from summer camp to a control or active group, with a 9 hour intergenerational program. In contrast, the work by Rebok (Rebok et al., 2004) and Carlson (Carlson et al., 2008) was based around the much larger and longer running Baltimore Experience Corps program in the USA (Fried et al., 2004; Fried et al., 2013; Glass et al., 2004; Tan et al., 2014) where older adults volunteered to go into schools to support child learning and life skills. In this case the pilot trial involved cluster randomisation by school (6 schools) with both adults and children allocated to intervention or control. Results suggesting benefit (see Table 1) are supported by further articles from this project relating to different subgroups and broader outcomes such as imaging (Brydges et al., 2020; Carlson et al., 2009; Carlson et al., 2015; Gruenewald et al., 2016) The article from Yasunaga (Yasunaga et al., 2016) is similar in that it reports outcomes for a non-randomised trial where older adult participants who went into schools were matched with a control group, who did not, as part of the larger Research on Productivity through Intergenerational Sympathy (REPRINTS) project from Japan. Similarly to the Baltimore initiative, there are several additional articles on related outcomes adding support (Fujiwara et al., 2009; Murayama et al., 2019; Murayama et al., 2015; Sakurai et al., 2018; Sakurai et al., 2016).

3.1. Quality assessment and risk of bias

Because variation in study design precluded the use of a traditional risk of bias assessment a bespoke tool was constructed for use at a study level based around key areas of risk. For example, this included making an assessment of the appropriateness of the chosen research design to answer the questions selected, the way the data was collected in practice, data handling and analysis. The overall risk of bias in the included studies is low-moderate (Table A2 shows our risk of bias ratings and explanatory comments), however there were particular areas of greater concern at a study level. These included a lack of clarity around the impact of selective recruitment, attrition, small sample sizes, unblinded outcome assessment, and short follow up limiting generalisability.

When we review the studies overall and consider risk of bias across the evidence base the lack of rigorous and replicable studies providing robust statistical and numerical estimates of the impact of intergenerational practice means that we do not have sufficient quality or quantity of data to draw conclusions supporting or recommending intergenerational practice. In particular caution is needed when making interpretations about the efficacy of intergenerational programmes for improving social, health and cognitive outcomes.

3.2. Themes within the evidence

Despite the lack of robust empirical evidence for intergenerational practice in these population groups the studies included in our review overwhelming reported positive potential benefits. These potential positives, and the potential pitfalls, challenges, requirements and recommendations for intergenerational activity and future research are further summarised in table 2. They included potential benefits in terms of enhanced self-esteem, self-worth and perceived usefulness, alongside reduced loneliness, anxiety and boredom, improved affect and empathy. Other potential benefits included improved communication, appreciation for diversity and transfer of knowledge. Potential pitfalls included a risk of reinforcing negative age stereotypes and participants disliking programmed activities. Practical barriers were also mentioned frequently including access to appropriate space, equipment, travel support and the potential for attrition.

Table 2.

Summarising The Potential Benefits And Practical Recommendations Expressed In The Literature On Intergenerational Interventions^a

Potential positives	Potential pitfalls	Challenges and barriers to set-up and delivery	Recommendations/requirements for setup and delivery	Recommendations for research
Internal benefits • Enhanced self esteem • Enhanced/improved affect • Enhanced empathy Interactive/external benefits • Enhanced communication • Increased interaction • Appreciation for diversity • Potential calming effect/Reduced caregiving stress • Reduced behavioural issues/better able to regulate own behaviour Group benefits • Transfer of experience knowledge know-how • Formal and informal learning • Social acceptance • Greater willingness to help • Active solidarity to those in difficulty • Expanded social roles, networks, supporting community • Develop humane and sustainable society • Community building and cohesion	Outcomes • Older adults may express dislike aspects of a programme or of children’s behaviour, noise etc: • Risk of reinforcing negative age-related stereotypes Logistics/programme • Use of generationally or developmentally inappropriate activities • Assumption that bringing people together is enough/ serendipitous programme • Insufficient support for delivery • Lack of participants • Infrequent sessions or excessively time-intensive sessions Community • Social concern re intergenerational contact	Practical barriers • Facilities/space/equipment • Paperwork/regulatory requirements • Accessibility • Travel expense/difficulty level/ distance • Weather Participant concerns • Last minute cancellations due to illness etc: • Older adults: Lack of confidence and concerns re lack of ability, energy, stamina, intolerance of loud or misbehaving children Fidelity issues • Absences/cancellations • Need for volunteers to help – e.g. parents if planning trips outside venue • Co-ordination across multiple schedules	Key themes • Ensure rights and respect due to all • Co-operation • Close not casual contact • Pleasant and rewarding contact • Functional interactions with a goal directed programme and common goals • Meaningful activities/ feeling of being useful • Voluntary and equal status Shared status of participants i.e. everybody has the capacity to contribute and benefit • Opportunity for friendship • Reciprocal and collaborative Practical recommendations • Provide advance information to each group about the other • Use of biographical information to support older adult clients’ needs • Children and older adult should be active agents in planning and review of intergenerational programme and experience • Possibly aim for a staff ratio 1:5 and a neutral space • Older adults always accompanied by staff • Geographical proximity (opening up geographically and socially’ to the community) • Policy in place for illness (including contagious illness) • Regular and frequent interaction • Family support is critical • Same number in older adult and child groups • Maybe pair older adults and children based on prior knowledge of their characters • Team building. Emphasize group identify and pair identity to build cohesion • Consider same gender pairs Other • Clear vision and leadership	• Theoretically informed studies • Larger sample sizes • Longitudinal assessment tracking participants beyond the assessment period • Use of rigorous quantitative and qualitative measures • Process evaluation • Some assessment of costs

The key recommendations drawn from the published literature can be divided into themes and practicalities. Themes include the importance of respect and equality for participants and co-operation with reciprocal meaningful and goal-directed collaboration. Practicalities focus on use of team-building techniques, provision of advance biographical information (age, sex, likes, dislikes, life history etc), co-design of activities, provision of sufficient staff support and geographical proximity to the community (Beynon et al., 2013; Cartmel et al., 2018; Gallagher and Fitzpatrick, 2018; Giraudeau and Bailly, 2019; Jarrott, 2011; Jarrott and Bruno, 2007; Kuehne and Melville, 2014; Radford et al., 2018; Sun et al., 2018).

4. Discussion

Using systematic review methodology we have shown that despite the increasing popularity of intergenerational programmes involving older adults, and the majority of older adults being community based, there is little research in this population. Furthermore, the limited research that does exist is subject to risk of bias and lacks strong empirical outcome data. There is a consensus that intergenerational programmes are likely to have positive outcomes for reducing social isolation and increasing cognitive functioning and health, but that they may also bring other benefits in psychological well-being, personal development, growth, meaningfulness, and empowerment with consequent direct and indirect impact on our societies. However, there is not only a lack of detail around the empirical evidence of benefit (for individuals, for society) but there is a lack of best practice evidence for the delivery of such programmes. Although related reviews reinforce the potential benefits (Ronzi et al., 2018) they also serve to demonstrate a lack of evidence for intergenerational programmes with older adults in aged care, meaning opportunities for extrapolation are limited (Giraudeau and Bailly, 2019; Jarrott, 2011; Radford et al., 2018). Going forwards, policy and practice decisions require evidence.

One aim of this systematic literature review was to identify the gaps and make recommendations for the next steps and it is clear that further research in this area is both required and important from an individual and a societal perspective. Given our ageing population and the majority of older adults residing in the community, clinical trial and observational research in this area is both timely and relevant with potential benefits spanning perspectives from personal fulfilment and societal gain to healthy ageing, healthcare resource use and economic considerations. From a societal perspective, the UN declaration on the rights of older adults explicitly states that ‘Older persons should remain integrated in society, participate actively in the formulation and implementation of policies that directly affect their well-being and share their knowledge and skills with younger generations’, that ‘Older persons should be able to seek and develop opportunities for service to the community and to serve as volunteers in positions appropriate to their interests and capabilities’ and that they ‘should be able to pursue opportunities for the full development of their potential’ and ‘have access to the educational, cultural, spiritual and recreational resources of society’ (United Nations, 1991). Similarly, the UN convention of child’s rights argues that it is in the best interest of the child to: “Make educational and vocational information and guidance available and accessible to all children” supporting the “development of the child’s personality, talents and mental and physical abilities to their fullest potential” (United Nations, 1990). For these reasons and in the context of geographically mobile populations and reduced intergenerational grandparent/grandchild interaction, it is essential that further research is conducted to explore whether intergenerational engagement is of significant value to both children and older adults globally, or whether it is only socially desirable.

This is also important from a healthcare perspective whereby an intergenerational programme may help protect cognition, either via increased social interaction, or, more likely by a combination of greater social, physical, and cognitive activity with the nature of the interaction lending itself to a naturalistic multi-domain intervention. Given the likely interactive and complex relationships between the risk factors for cognitive decline and the suggestion of benefit shown by a previous multi-domain trial (Ngandu et al., 2015) this could carry huge potential.

Although our exclusion of non-English language articles and the immaturity of the evidence base limits the scope of findings of the review but at the same time reveals key gaps where research is needed, the systematic nature of our review is a strength and whilst we may have missed some findings in the grey or in the non-English language literature, these would need to have been robust strong studies to change our conclusions relating to the overall lack of evidence and need for further research.

4.1. Conclusions

Overall there is, as yet, neither strong evidence for nor against community based intergenerational interventions.

Our review highlights clear gaps, an overall lack of strong empirical evidence for the hypothesized benefits; to individuals, (older adults, children), to society, and a lack of evidence about the best practice method of delivery. Alongside this, an apparent perception of benefit and increase in popularity of intergenerational programmes (see (Playgroup New South Wales)) underscore the urgent need for rigorous empirical testing. If such programmes deliver their implied benefits a case could be made for their widespread implementation and they may become more of the norm within society. To take the first step we need to know if and how such programmes can capitalise on their potential.