Scale-up of prevention programmes: sustained state-wide use of programme delivery software is explained by normalised self-organised adoption and non-adoption.

Background: Population-level health promotion is often conceived as a tension between "top-down" and "bottom-up" strategy and action. We report behind-the-scenes insights from Australia's largest ever investment in the "top-down" approach, the $45m state-wide scale-up of two childhood obesity programmes. We used Normalisation Process Theory (NPT) as a template to interpret the organisational embedding of the purpose-built software designed to facilitate the initiative. The use of the technology was mandatory for evaluation, i.e. for reporting the proportion of schools and childcare centres which complied with recommended health practices (the implementation targets). Additionally, the software was recommended as a device to guide the implementation process. We set out to study its use in practice.

Methods: Short-term, high-intensity ethnography with all 14 programme delivery teams across New South Wales was conducted, cross-sectionally, 4 years after scale-up began. The four key mechanisms of NPT (coherence/sensemaking, cognitive participation/engagement, collective action and reflexive monitoring) were used to describe the ways the technology had normalised (embedded).

Results: Some teams and practitioners embraced how the software offered a way of working systematically with sites to encourage uptake of recommended practices, while others rejected it as a form of "mechanisation". Conscious choices had to be made at an individual and team level about the practice style offered by the technology-thus prompting personal sensemaking, re-organisation of work, awareness of choices by others and reflexivity about professional values. Local organisational arrangements allowed technology users to enter data and assist the work of non-users-collective action that legitimised opposite behaviours. Thus, the technology and the programme delivery style it represented were normalised by pathways of adoption and non-adoption. Normalised use and non-use were accepted and different choices made by local programme managers were respected. State-wide, implementation targets are being reported as met.

Conclusion: We observed a form of self-organisation where individual practitioners and teams are finding their own place in a new system, consistent with complexity-based understandings of fostering scale-up in health care. Self-organisation could be facilitated with further cross-team interaction to continuously renew and revise sensemaking processes and support diverse adoption choices across different contexts.