M-Pesa was designed for a country with 10 million bank accounts and 30 million mobile subscribers. The constraints — limited banking infrastructure, unreliable electricity, users without formal financial literacy — were not obstacles to be designed around. They were the design brief. The resulting system is more resilient, more accessible, and more widely trusted than banking products built for high-infrastructure environments, and it now operates in fourteen countries (Safaricom 2024). This is not an accident. It is a reproducible dynamic — and it applies to healthcare innovation in ways the global health community has been slow to absorb.
The Gap in the Literature
A substantial body of scholarship examines reverse innovation — the process by which innovations developed for resource-constrained settings are later adopted in high-income countries (Govindarajan and Trimble 2012; Syed et al. 2012). Less examined is the specific question of what design principles produce innovations that travel in this direction, and whether those principles can be made explicit and intentional rather than incidental to constraint. For digital health tools specifically, the relationship between offline-first design architectures and the broader principle of constraint-as-advantage has not been systematically articulated in the East African literature. This article proposes that articulation.
What Designing for Kenya Actually Requires
Healthcare innovation for Kenyan contexts must operate in environments that the majority of global health digital products were not designed for. The KNBS 2022 Kenya Integrated Household Budget Survey documented that approximately 36 percent of Kenyan households had no piped water at home, and significant proportions remain without reliable electricity (Kenya National Bureau of Statistics 2022). Community health posts that serve as first points of contact operate without electronic health records, often without reliable connectivity, and with community health promoters managing caseloads without digital decision support. A technology that functions in these conditions — offline, on a basic handset, with minimal training, producing outputs useful to a community health promoter in the field — is a technology that will work almost anywhere. The constraint is a quality assurance mechanism.
Bangladesh's BRAC Model: Scale Achieved, Interoperability Gap Identified
Bangladesh's BRAC operates the largest non-governmental community health worker network in the world, supporting over 4,300 frontline workers with a digitised offline-first mHealth platform built on the WHO-endorsed OpenSRP architecture (Transform Health 2022). It is frequently cited as the model for designing digital health tools for low-connectivity environments, and in important respects it is a genuine success: the platform has digitised longitudinal household health tracking for tens of millions of people who would otherwise be invisible to any health information system at all.
The most useful evidence, however, is not the success story. A 2025 scoping review of mHealth tools for community health workers across low- and middle-income countries found that the majority of platforms studied adopted offline-first connectivity design — and that persistent data-syncing errors and data loss occurred regardless, because the underlying problem was not connectivity at the point of data entry but interoperability with national health information systems further upstream (Patwary and Islam 2025). Offline-first design solves the symptom. It does not solve the structural gap between a community-level tool and the national system it is meant to feed. This is the specific gap that Kenya's digital health architecture should be designed to close from the outset.
A solution designed for the most constrained environment is tested more rigorously than one built for optimal conditions. That rigour is the export value.
Three Steps Available Now
First, any digital health tool procured for Kenya's Community Health Strategy should be required, as a procurement condition, to demonstrate a defined data-interoperability standard with the Kenya Health Information System — not merely offline functionality at the point of capture — closing the gap the BRAC literature identifies.
Second, the Ministry of Health can commission a structured review of Kenya's existing community-level mHealth pilots specifically for sync-failure and data-loss rates, publishing findings before further national scale-up.
Third, KEMRI can establish a design-for-constraint review panel that any digital health tool intended for community health promoter use must pass before procurement, testing specifically for offline resilience, low-literacy usability, and basic-handset compatibility.
The point-of-care diagnostic tools developed for African settings without laboratory infrastructure are now being adopted in rural emergency departments in high-income countries facing specialist shortages. Kenya is not designing for the margins of global health. It is designing for the majority.