Complex multi-levelled factors shaped PrEP implementation. Nine specific areas of the PrEP care cascade involved in initiation and uptake of PrEP were both amenable to change and prioritised for improvement. The corresponding barriers and facilitators were multi-levelled and interdependent. Many were psychosocial, relating directly to the way staff or patients thought and felt; others related to the organisation of services, wider issues of access to support and training, and factors relating to the environmental infra-structure of services. Using tools from implementation science, we systematically generated highly specific, theoretically informed and evidence-based ways of optimising PrEP implementation in the future. Examples include: provision of PrEP in diverse settings to reach all in need; co-produced, culturally sensitive training resources for healthcare professionals, with focused content on non-daily dosing; meaningful collaborative working across all stakeholders.

To date, several attempts have been made to conceptualise the implementation of PrEP but these have been largely broad and descriptive, typically categorising the whole of PrEP care into four or five large steps within a continuous, linear ‘care cascade’ (25-28). Published studies have tended to focus on using these high-level steps to audit or quantify PrEP implementation, seeking to identify and understand key points of attrition within particular populations and associated health care systems (29). There are numerous examples of PrEP prescribing guidance (15,30-31), but fewer published studies specifically address the implementation of PrEP routine care pathways and services. A scoping review of PrEP delivery models (32) created a comprehensive inventory of existing models, but did not specifically focus on delivery of the detailed steps of the PrEP cascade within the models described. A review of PrEP implementation identified multiple barriers to PrEP uptake, some of which mirrored those we described (33). The authors proposed multilevel interventions to target these barriers but acknowledge that proposed interventions do not always align to specific barriers.

In contrast, no work to date has used conceptualisations of the care cascade as a starting point for systematic, focussed service improvement whilst explicitly using theory and evidence to enhance implementation. We directly addressed this gap by taking a single key step of the PrEP care cascade, the ‘uptake and initiation of PrEP’, and focussed on it as an area in need of intervention development to enhance future implementation. We derived recommendations (interventions) directly from the barriers and facilitators at each priority area.

Some recommendations warrant additional comment. In relation to ‘engaging HCPs with PrEP as an acceptable approach to HIV prevention’, we highlight the need to address both structural and psychosocial issues. We also emphasise the importance of considering financial and other resources as well as the timescale for implementation. These factors are likely to be central to HCP engagement which in turn is central to patient uptake. We also recommend a multileveled national infrastructure to promote, coordinate, and monitor HCP engagement with PrEP and highlight how these structural initiatives could be bolstered by a range of local initiatives such as engaging staff through local “PrEP champions”. The barriers these recommendations are designed to overcome were strikingly similar to those reported in a number of studies within Pinto et al’s recent review (33).

In relation to ‘potential PrEP users accurately reporting their HIV risk behaviour…”, we found that depending on the cultural context, it may be important to educate and persuade HCP about the ‘bigger picture’ of PrEP provision and overcome any residual moralism and stigma relating to sex, homophobia, or racism which has also been described in other studies (33-34). Stigma is well recognised as a potent barrier to accessing HIV testing, prevention and care and it also might inhibit the full disclosure of HIV acquisition risk factors such as stigmatised sexual behaviours or partner numbers relevant to PrEP offer and uptake. Stigma may also apply to and inhibit the taking of PrEP itself (35-37). We recommend close partnership work between sexual health services, CBOs and PrEP users to enable sensitive, culturally appropriate conversations around PrEP, and to help HCPs improve their cultural competencies. The strongly supported health care and community-level “PrEP-positive” ethos described by our participants seems highly appropriate and would need to be extended to all settings in which PrEP may be provided in the future, particularly those in which sexual health is less familiar.

Our findings suggest that the ‘PrEP eligibility criteria’ which were used by HCPs to help identify people who might benefit most from PrEP (26), should be reframed and understood as needs-based approaches to HIV prevention, conveying the pros and cons of PrEP so that it can be extended to all who could benefit. This could largely remove the issue that criteria are less sensitive for identifying people from certain groups or racial backgrounds as also reported in other countries (38).

A large epidemiological analysis published after this study showed that Scottish implementation models strongly favour GBMSM and have limited reach into other key vulnerable populations (6,14). In parallel, the characteristics of people newly diagnosed with HIV in Scotland have changed since the introduction of PrEP and now people are more likely to have acquired HIV though heterosexual sex and to be non-white indigenous than in the pre-PrEP era (14,39), similar to findings from Australia (40). As noted in our recommendations and by others, reaching all groups that could benefit from PrEP is essential (9); Several studies provide explanations for low PrEP uptake in some key vulnerable populations. Among women of colour in the UK, important factors were low awareness of PrEP, feelings of stigma related to HIV itself and attending sexual health clinics, and a preference for trusted community settings for discussion about HIV testing and prevention (35,41)). Among people who inject drugs in Scotland, awareness of PrEP was low but some would find PrEP appealing if provided within familiar settings such as outreach drug services (42). Very few trans people have accessed PrEP in Scotland (12). International studies suggest that the need for PrEP among this group is high but important barriers to access preclude uptake (36,43). Restricting PrEP provision to sexual health clinics probably deters some trans people who could benefit (44). Additional or tailored recommendations to enhance PrEP uptake and initiation for people from vulnerable populations are needed as evidence accrues.

We used a novel, rigorous approach to developing recommendations which is not typical of approaches to enhancing implementation. The resulting recommendations are anchored in the evidence and theory-driven (22) and are specified using a standardised language to describe intervention content in detail (i.e., intervention functions and behaviour change techniques (23)). Typically, the initial stages of the PrEP care cascade involve a complex patient journey, marked by setting-specific interactional dynamics and a series of interdependent joint and individual behaviours. Our adoption of a behavioural lens, and the subsequent systematic development of highly specific ways to enhance implementation, meant we re-conceptualised this patient journey as a series of distinct and sequential behaviours. This approach led to costs and benefits; where we gained through this behavioural specificity and our ability to specify future recommendations in great detail (e.g. behavioural change techniques), there was duplication of effort to detail shared antecedents to the varied behaviours.

We focussed on one national context and although findings are likely to be generalisable to similar settings, it is uncertain how recommendations might apply in very different contexts. In particular, as all PrEP care was free of charge, participants did not face the financial barriers reported from some settings (45). Very few people in Scotland on PrEP are not GBMSM (13) and our findings lack specificity for other groups. A high proportion of PrEP user participants had a university qualification and while representative of those on PrEP in Scotland, the sample under-represents those with lower health and PrEP literacy who may have other needs and preferences for accessing PrEP care. Furthermore, the COVID-19 pandemic led to a reconfiguration of some sexual health and PrEP services and our findings may be more or less relevant as a result. Our evaluation took place relatively early in the PrEP programme which probably magnifies early stage issues which become less important as familiarity increases.

To support individuals and populations to fully benefit from PrEP we must overcome the considerable challenges of large-scale implementation (31). Here, we combined qualitative data from multiple viewpoints and used multiple analytic tools to systematically detail useful insights concerning initiation and uptake from the first two years of Scottish PrEP implementation. To our knowledge, we present the first evidence-based and theory-informed recommendations which can be used flexibly across a range of settings to improve PrEP initiation and uptake. Our findings will inform future Scottish implementation of PrEP (46) and could usefully contribute to the global public health priority of elimination of HIV transmission by 2030 (31,47).