1. Interventions that explicitly use social networks show promise

A major systematic review and meta‐analysis (Hunter et al., 2019) found that interventions which explicitly use social network mechanisms (for example, recruiting influential individuals, reshaping network ties, promoting peer‐to‐peer diffusion) are associated with favourable short‐term (< 6 months) and somewhat longer‐term (>6 months) effects, especially in sexual health outcomes.

More broadly, interventions showed positive signals for outcomes such as alcohol misuse, smoking cessation and HbA1c for diabetes, though effect sizes varied and many studies had methodological limitations.

In summary, social network-based interventions are feasible, and show meaningful effects, but are still under-utilised and under-theorised.

2. The mechanisms matter: influence, social support, diffusion, network embedding

Our collective work emphasises that simply delivering information to individuals is not sufficient. What matters is how that information flows through social ties, how behaviours get modelled, how social norms change, and how network position influences exposure and uptake. For example:

• Interventions that identify and train key “influence agents” or “peer leaders” can shape network dynamics and behaviour change.
• Mapping a person’s social network or community network and then linking them to resources or support can augment connectedness, reinforcing behaviour change.
• Community- or group-based interventions that create new ties, or strengthen existing networks (rather than simply existing one‐to‐one dyads) may generate broader diffusion and longer-lasting impact.

In doing so, our work bridges behavioural science (individuals) and network science (structures, ties, positions).

3. Application across different health domains and settings

Our research spans multiple domains including physical activity, healthy eating, community‐based programmes, sexual health, and chronic disease management, and covers different settings (schools, communities, online platforms, peer networks).

For example, we examined how a community‐based physical activity programme helped to foster social cohesion via changes in network ties over time (temporal network analysis).

In adolescent physical activity, studies using trained peer “influence agents” were deployed, though results were mixed and highlighted important lessons about recruitment, engagement, and measurement.

Some pilot feasibility studies explored online social network interventions in children or young people, mapping both behaviour change and network engagement.

By covering such diversity, we aim to uncover mechanisms that generalise across contexts, and also pinpoint context‐sensitive features (e.g., age, population, platform, network type).

4. Challenges and future directions

Our research highlights several important caveats and steps forward:

• Many interventions include multiple components (e.g., education + peer support + network mapping) and isolating the unique contribution of the network component remains challenging.
• Measures of social networks (structure, position, density, cohesion) are often absent or simplistic, limiting insight into how network change mediates outcomes.
• Intervention‐engagement and retention pose a considerable challenge, especially in digital or adolescent contexts.
• Heterogeneity in effect sizes, populations, behaviours, and follow-up durations means that generalisability is challenging and still emergent rather than established.
• There is a growing need to design interventions with network theory built in (rather than retrofitting network concepts), to test causal mechanisms (who influences whom, how ties change, what the ripple/diffusion effects are) and to integrate network and behaviour change measurement.
• Across chronic disease settings, the evidence on network/support interventions remains weak or very low certainty highlighting the need for more high‐quality RCTs in those domains.

5. What this means for practice

For practitioners designing health behaviour interventions, our research suggests a few guiding principles:

1. Do not just deliver information to individuals but consider the social context (who they talk to, whom they trust, which peer networks they’re part of).
2. Use social network mapping or peer‐leader selection to identify leverage points (influencers, connectors, brokers).
3. Facilitate the creation or strengthening of active ties (not just static ones). For example, peer groups, buddy systems, community meet‐ups, online forums.
4. Consider diffusion effects: interventions may spread beyond the direct recipients via their networks.
5. Monitor network‐level metrics (e.g., tie formation, density changes, centrality of key actors) alongside behavioural outcomes.
6. Social network interventions are complementary to individual‐level behaviour change techniques, but they hold under-used potential.