Implementation fidelity to the World Health Organization package of essential non-communicable disease interventions (PEN) during scale-up in Nepal’s primary healthcare system: a mixed methods study

BACKGROUND: The World Health Organization package of essential non-communicable disease interventions (PEN) comprises evidence-based interventions for primary healthcare systems in low- and middle-income countries (LMICs). Implementation fidelity, defined as the degree to which an intervention is delivered as intended, is crucial for intervention effectiveness. Yet, across LMICs, evidence on fidelity to PEN is scarce and mostly limited to pilot studies. METHODS: We evaluated fidelity to PEN during national scale-up in Nepal using a convergent parallel mixed methods study design based on the Conceptual Framework for Implementation Fidelity. Data were collected in 2021 from 105 randomly selected primary healthcare facilities across all seven provinces of Nepal. Using direct observation of 172 non-communicable disease (NCD) patient visits, we quantified health service providers' adherence to PEN protocol 1 (hypertension and diabetes management) and protocol 2 (lifestyle counselling). We calculated overall fidelity scores and examined their potential determinants in multivariable regression models using generalized estimating equations. Moreover, we conducted semi-structured interviews with 47 providers and 33 NCD patients to elicit their views on factors affecting fidelity. We analyzed verbatim interview transcripts using thematic analysis. RESULTS: PEN-trained health service providers managing hypertension and diabetes patients (protocol 1) mainly focused on asking about medication (76% of patient visits), measuring blood pressure (82%), and adjusting medication (85%). Blood glucose was measured in 28% of eligible patient contacts. Other care steps, including cardiovascular risk estimation, were mostly skipped. Lifestyle counselling for NCD patients (protocol 2) most frequently entailed salt restriction advice (41%) and instructions on medications (62%) and follow-up appointments (69%). Overall fidelity scores for protocols 1 and 2 were 20.8 and 22.1 out of 100, respectively. In multivariable regressions, prior PEN training was associated with 25% (95%-CI: 1-54%; p = 0.04) and 32% (95%-CI: -4-80%; p = 0.08) higher fidelity scores for protocols 1 and 2, respectively. Qualitative data suggested that providers and patients were generally engaged by the intervention, though various factors impede fidelity, including negative user experiences and limited availability of medical supplies. Strategies to improve fidelity to PEN that participants suggested were intervention manuals, supervision and monitoring visits, brief refresher trainings, and community outreach programs. CONCLUSIONS: Implementation fidelity to PEN during scale-up in Nepal was often limited to a few diagnostic and therapeutic actions. Next to enhancing user experience, ensuring a reliable supply of diagnostic equipment and medicines will likely be crucial to sustainably improve NCD care. Fidelity assessments in resource-limited settings should examine the availability of key implementation resources.