Background: Training of primary care practitioners is one of the most implemented interventions in medical international development programmes targeting non-communicable diseases (NCD). Yet in many cases their effectiveness is below expectations. One potential cause of this is that they struggle to account for local context, especially when working with ethnic minorities. Here we begin to address this gap through a qualitative case-study of how local contextual factors have impacted the success of a World Health Organization (WHO) healthcare training programme on Type 2 diabetes with an ethnic minority group in rural central Vietnam.
Design: A qualitative case-study collected data during 2018. We conducted 25 semi-structured interviews, two focus groups, and participant observation with patients, healthcare professionals, and members of a local non-governmental organisation involved in the programme. We used thematic coding to identify important contextual factors and how they helped or hindered programme delivery. Next, we synthesised each of these themes in a narrative style, drawing on the rich detail provided by respondents.
Results: We found that, despite using a notionally decentralised approach, the effectiveness of the training was hindered by social, political, and economic determinants of health which influenced the inhabitants’ relations to healthcare and diabetes. Particular barriers were the political perceptions of minorities, their economic access to services, the healthcare prejudices toward ethnic rural populations and the rigidity of medical training.
Conclusions: Given the similarity of our case with other WHO NCD programmes, we view that our findings are of wider relevance to global public health policy and practice. We suggest that better recognising and addressing local contextual factors would make such programmes more polyvocal, grounded, and resilient, as well as enabling them to better support long-term transformative change in public health systems. We conclude by discussing methods for implementing this in practice.
