Assessment of readiness of the new setting for implementation of the intervention

When considering the adaptation of a proven intervention to a new setting, it is critical to understand which intervention components require high implementation fidelity and the readiness of the new setting to accommodate these. Adaptation of the SMARThealth intervention for each setting was preceded by careful assessment of the existing health policy landscape, disease specific guidelines, health care infrastructure and health workforce organization. For example, when implementation was planned in Indonesia following lessons learned in India (where there was less focus on codesign), there were concerted efforts to codesign and continuously improve implementation strategies to ensure best fit for context. This included a systematic analysis of the policy landscape relating to essential medicines supply to plan medication prescription and distribution at the community level, and also an assessment of community awareness of NCDs that led to the integration of an initial health awareness component in the intervention package [7]. A similar barrier analysis was also initially conducted in China to understand contextual constraints on the provision of best practice care [8]. Comparatively, an opportunity to implement the intervention in Sri Lanka was not pursued following an initial health system assessment because of an inherent separation of primary curative care and preventive services, lack of a relevant community workforce and an evolving IT infrastructure.

Integration with existing policies, systems and infrastructure

Based on these assessments, intervention components were integrated to the local context to maximize fidelity while ensuring minimal changes to the existing healthcare workflow. In Australia, the CDSS was designed to interface with existing patient management information systems used by GPs [9]. This did not change the clinical management workflow but provided additional information (regarding patient's NCD risk status) to support GPs in their patient management. Similarly, in Indonesia the SMARThealth mobile application was designed to interface with and send data to the local primary health care electronic medical record database (ePuskesmas). In India, a process evaluation of the initial intervention trial found that the effect of SMARThealth was likely diluted by other on-going initiatives, suggesting the intervention was not sufficiently integrated into existing policies and systems, receiving limited “buy in” from local healthcare system stakeholders [10].

Leveraging (and adequately training/supporting) an existing health workforce

One of the core intervention components for high implementation fidelity which was considered in our initial assessment of system readiness was the presence of a health workforce with capacity to deliver the intervention. In India, Indonesia and Thailand the presence of an existing community health care workforce delivering care through household visits provided task-sharing opportunities, avoiding the need to develop an entirely new cadre of worker. Patient screening and follow-up for the intervention was conducted by Kaders in Indonesia, Accredited Social Health Activists (ASHAs) in India and, currently, by Village Health Volunteers in Thailand. In China, this role was fulfilled by Family Health Promoters, voluntary family members who have received basic training in chronic disease management with responsibility for maintaining the health of their family [8]. In contrast, in the Australian context, local GPs used the CDSS as per the norms of clinical management in that country.

However, relying on an existing health workforce to deliver a new intervention has the potential to perpetuate prevailing challenges—including those associated with capacity, scope of practice and incentives. To ensure the SMARThealth mobile application appropriately met the needs and capabilities of frontline health workers in India, an “agile development methodology” was adopted [3, 11]. This included iterative development, testing with ASHAs who had little experience using information and communication technology, and refinement based on their feedback. In all SMARThealth implementation settings additional training was also provided to frontline health workers to build their capacity to manage NCD conditions and, where appropriate, higher level health workers (such as nurses) were engaged to provide monitoring and additional training and support. While this level of investment may not be feasible in all settings, the principle of adapting interventions in line with the capacity of the local workforce remains relevant.

Harnessing community engagement and local leadership for ownership and sustainability

A key reason for the success of the 2-year trial of the SMARThealth intervention in Indonesia was the high level of investment from local healthcare system stakeholders. This was facilitated by a local Expert Advisory Group led by champions from within the District Health Authority, together with representatives from the national health insurance agency and health care personnel. Before the initiation of the program at the village level, the District Health Authority visited village elders and representatives to promote the importance of NCD screening. Additionally, senior district government officials created a supportive policy and logistical environment for the implementation of the intervention by committing to improve existing supply chain issues by investing in the procurement of essential cardiovascular disease medicines to meet the likely demand. Local investment has continued in the district-wide scale up of the intervention, which is government led, to 390 villages and a targeted population of 1 million adult residents.

Implementation of SMARThealth in countries across a spectrum of health system contexts has provided important lessons emerging from both successes and failures in field trials. For policymakers seeking to accelerate their national NCD response, our experience suggests that contextual adaptation of proven interventions is an important option in their toolbox. However, substantial time and investment is required. The lessons learned and shared here are not intended to be definitive and exclusive, but to improve opportunities for success by ensuring that the adaptation of interventions is appropriate, integrated and accepted in their new implementation setting. However, we recognize that a limitation of these learnings is they have been accumulated by “researchers looking in” and based on pilot implementation and evaluation. The insights of health system administrators, policymakers, providers and community members are also critically needed, and are something that we hope to learn during SMARThealth scale up in Indonesia.