10th Biennial ACSPRI Social Science Methodology Conference in conjunction with the Australian Social Network Analysis Conference (ASNAC)

Social isolation is an established dementia risk factor, linked to 5% of dementia cases globally. Candidate risk mechanisms include direct cognitive benefits, mental health benefits, and social influence improving health behaviour. Health promotion messaging calls individuals to socialise to reduce their dementia risk, with heterogeneity among recommendations. Our sequential mixed-methods study aimed to more effectively promote social engagement to reduce dementia risk. Participants were older adults (aged 50+) engaged in the Island Study Linking Aging and Neurodegenerative Disease (ISLAND), a Tasmanian longitudinal cohort study that monitors, and educates about, dementia protective behaviour.

To investigate participants’ understanding and enactment of existing messages promoting social engagement to reduce dementia risk, semi-structured interviews were conducted with a purposively sampled ISLAND participant subset (n=27). Sampling aimed to achieve breadth of coverage across individual characteristics (example: gender, education, age), social circumstances (example: living alone or cohabiting, working or retired), and social preferences (extraversion). Interview transcripts were analysed using reflexive thematic analysis.

Interview participants understood messages about socialising for dementia risk reduction within broader knowledge frameworks. These broader knowledge-frames included experiential knowledge of social engagement’s impact on cognition, preferred forms and previous experiences of social engagement, and broader priorities and constraints. Specificity, perceived trustworthiness, and individual relevance of health promotion information impacted the weight placed on this within broader knowledge-frames. Participants described enjoyably enacting health promotion messages when they perceived sufficient congruence between health promotion information and their broader knowledge-frame, with some individuals having greater incongruence tolerance. When incongruence tolerance was exceeded, participants described either disregarding or conflictedly enacting health promotion information. For some, this caused anxiety. Participants discussed social resilience and adaptability as facilitating enjoyable enactment of social activity when facing changing social circumstances.

To improve message specificity, and potentially increase individual relevance, we examined the relationship between social engagement and cognition across people with different social preferences (extraversion) using multiple mixed-effects linear regression models and a pooled cross-sectional design (980 observations, 622 participants). All models were adjusted for age, gender, and education level, along with project timepoint, repeated cognitive assessments, and multiple observations per-participant. Online tools measured cognition (age-standardised Cambridge Neuropsychological Test Automated Battery (CANTAB) paired associate learning (episodic memory) and spatial working memory (executive function)); extraversion (10item Big Five Inventory); and social engagement (Lubben Social Network Scale friends, family, and neighbours subscales; two loneliness questions). Across the extraversion continuum, episodic memory scores were significantly better among participants who reported more engagement with family (p(main effect)=.004, gradient[confidence interval]: .017[.005-.029]; p(interaction)=.11). There was a relationship between engagement with neighbours and episodic memory only when considering differences across the extraversion continuum (p(interaction)=.038); better episodic memory scores were associated with more engagement with neighbours only among less extraverted participants (extraversion 25th percentile gradient[confidence interval]: .020 [.005-.035]; extraversion 75th percentile gradient[confidence interval]: .000[-.017-.018]). In gender stratified analyses, an additional significant interaction showed a stronger relationship between family engagement and episodic memory scores among more extraverted men (interaction p(men)=.028; extraversion 25th percentile gradientconfidence interval: .012[-.002-.025]; extraversion 75th percentile gradientconfidence interval: .024 [.009-.039]). No significant associations were found for executive function.

We identified requisites for enjoyable enactment of messages promoting socialising to reduce dementia risk and generated evidence that indicates engagement with neighbours may be beneficial among less extraverted individuals, while engagement with family may have benefit for all, particularly more extraverted men. Incorporating this evidence into messages promoting socialising to reduce dementia risk may improve specificity and perceived message relevance, promoting enjoyable enactment.

Social Network Analysis (SNA) has demonstrated contagion effects across a wide range of behaviours. The Island Study Linking Aging and Neurodegenerative Disease (ISLAND) is an ongoing longitudinal study with over 3000 actively engaged Tasmanians aged 50 and above, who report annually on social engagement and other health behaviours. ISLAND participants who report more social engagement report healthier behaviour across alcohol consumption, smoking, physical activity, cognitive activity, diet, blood pressure, body weight, cholesterol, blood sugar, and diet, with evidence that social engagement is an upstream driver of many of these behaviours. In this study, we aimed to understand participants’ experiences of the relationship between social engagement and other healthy behaviours.

We engaged a purposively sampled subset of ISLAND project participants (n=27, 90% of 30 invitees) in semi-structured interviews, analysed through reflexive thematic analysis. Purposive sampling ensured a wide range of individual characteristics (e.g. gender, education, age), social circumstances (e.g. living alone or cohabiting, working or retired), and social preferences (introversion/extraversion).
At one extreme, some participants experienced health choices as independent agents, with no known social influences impacting their behaviour. At the other extreme, some experienced health choices as highly socially driven or constrained, with health behaviour being dependant on the actions and opinions of others. Most participants located themselves at intermediate points in this continuum: experiencing some influence through knowledge, motivation, and/or behaviour co-participation, while still experiencing autonomy in their final health decisions. Some participants positioned themselves differently when different health behaviours were discussed; for example, dietary choices may be more socially influenced, while choices to attend medical appointments may be more individually determined. Where influence was experienced, a wide range of influential actors were reported, including friends, family members, and animal companions.

This longitudinal cohort provides the opportunity to quantitatively study the role of social influence in prompting a range of healthy behaviours among older adults. The understandings developed in this initial exploratory project may provide the foundation for further exploration using social network analysis.