The question
How do social and digital determinants of health influence the implementation and equity outcomes of digital health interventions? The review covered empirical studies of mHealth, telemedicine and platform-based interventions in rural low- and middle-income settings — environments where constraints are sharpest and easiest to observe.
Four findings that generalise
1. Adoption takes more than digital capability
Digital literacy matters, but uptake is equally constrained by health literacy, economic conditions and interpersonal dynamics such as patient–clinician trust. Usability improvements alone don't fix weak trust or missing intermediaries.
2. Socioeconomic conditions shape capacity, not just willingness
Groups differ not only in whether they want to adopt but in whether they can sustain use. Uniform rollout assumptions miss this — and produce uneven results despite strong execution.
3. Determinants are dynamic
Capabilities and constraints shift with training, awareness and service availability. One-time readiness assessments go stale; implementation needs ongoing monitoring.
4. The local health system is a binding constraint
Infrastructure, governance, financing and social conditions vary by setting — interventions that don't adapt to institutional context fail to scale or sustain.
What a decision-maker should take from it
Prioritise the determinants that constrain feasibility, not just demand signals. Account for intermediaries and trust, align with existing care pathways, treat inequity as a structural performance constraint, and pair quantitative KPIs with contextual indicators.
What it demonstrates about how I work
Synthesising a heterogeneous, multi-context evidence base into decision-relevant structure: framing an ambiguous multi-level problem, connecting individual through institutional factors, and writing for non-academic readers without losing methodological rigour.
Method: PRISMA-guided systematic review with deductive thematic analysis using a structured framework. Out of scope: clinical efficacy, algorithmic performance, cost-effectiveness modelling. Full detail in the PDF.
← Back to all work