ObjectiveGay and bisexual men who have sex with men (GBMSM) bear a disproportionate burden of sexually transmitted infections (STIs). Most STIS are asymptomatic and people infected won’t know to seek care unless they are told about their exposure. Contact tracing, is the process of identifying and contacting sex partners of people with STIs for testing and treatment. Contact tracing is sometimes particularly challenging amongst GBMSM because of the kinds of sexual relationships which GBMSM enjoy. These include one-off partners who are particularly important for transmission dynamics as they contribute disproportionately to onwards transmission. The effectiveness of contact tracing interventions within sexual health are patterned by sexual-partner type. Contact tracing and management for ‘one-off’ partners is an on-going public health challenge. Low motivation amongst index patients, high resource burden on health care professionals and problems with contactability are key barriers to contact tracing. Using insights from complex adaptive systems thinking and behavioural science, we sought to develop an intervention which addressed both the upstream and down-stream determinants of contact tracing and change the system in which many inter-dependent contact tracing behaviours are embedded.SettingUK community-recruited GBMSM, stakeholders, sexual hcp, dating app providersMethodUsing the MRC complex intervention framework and insights from the INDEX study, a three-phase intervention development process was adopted to specify intervention content. Phase one consisted of an inter-professional and community-member stakeholder event (n=45) where small mixed groups engaged in exploratory systems-mapping and the identification of ‘hot spots’ for future intervention. Phase two used a series of focus groups with GBMSM (n=28) and interviews with representatives from key dating app providers (DAPs) (n=3) to further develop intervention ideas using the theoretical domains framework, the behaviour change wheel and the behaviour change technique taxonomy. In Phase 3 we again worked with key stakeholders expert health care professionals (HCPs) (n=5) and key workers from community-based organisations (CBOs) (n=6) to hone the intervention ideas and develop programme theory using the APEASE criteria.ResultsThe co-produced intervention levers change simultaneously across the system within which contact tracing is embedded. Multiple change-agents (i.e., GBMSM, CBOs, HCPs) work together, sharing an overall vision to improve sexual health through contact tracing. Each make relatively modest changes that over time, synergistically combine to produce a range of multiple positively-reinforcing feedback loops to engender sustainable change around contact tracing. Agreed intervention elements included: a co-ordinated, co-produced mass and social media intervention to tip cultural norms and beliefs of GBMSM towards enabling more contact tracing and to challenge enduring STI- and sex-related stigmas; complementary CBO-co-ordinated, peer-led work to also focus on reducing STI stigma and enabling more contact tracing between one-off partners; priming GBMSM at the point of STI diagnosis to prepare for contact tracing interactions and reduce HCP and sexual health-service burden; changes to SHS environments and HCP-led interactions to systematically endorse contact tracing; changing national audits and monitoring systems to directly address one-off partner targets; delivering bespoke training to HCPs and CBO staff on one-off partners and the social and cultural context of GBMSM; DAPs active involvement in mass and social media promoting appropriate contact tracing messaging.ConclusionOur combination of multiple data sources, theoretical perspectives and diverse stakeholders have enabled us to develop an expansive, complex intervention that is firmly based in the priorities of those it will affect, and which has a solid theoretical foundation. Future work will assess if and how it will be possible to evaluate it. The resulting intervention is profoundly different than other ways of enhancing contact tracing, as it simultaneously addresses multiple, multi-levelled, upstream and social determinants of contact tracing.