Abstract
Well-funded e-Health pilot projects in low- and middle-income countries (LMICs) consistently fail to achieve sustainable integration into routine practice. Existing frameworks address single dimensions of this problem, lacking integrated, multi-level perspectives and actionable roadmaps. This paper presents the development and empirical validation of the STOE (Social-Technological-Organisational-Environmental) Framework and its complementary Implementation Architecture, a dual model for resource-constrained public healthcare settings. We employed a sequential explanatory mixed-methods design across three public healthcare facilities in Western Kenya. Following a deductive-inductive process, we deductively constructed the STOE Framework by synthesising Normalisation Process Theory, the Technology-Organisation-Environment framework, and the Four-in-Balance Model. This a priori model was then inductively tested through quantitative surveys (n=108 healthcare providers) and qualitative interviews (n=21 key informants). Empirical data confirmed the framework's validity. Implementation failures stemmed from structural disconnects across nested STOE contexts, not isolated technical deficits. Three disconnect patterns emerged: (1) environmental policy uncertainty paralysing organisational decision-making (E→O); (2) technologies procured without frontline user input leading to systematic social rejection (S↔T); and (3) fragmented funding creating unsustainable donor-dependent pilots that collapsed post-completion (E→O→S cascade). The validated dual framework offers a pragmatic two-stage approach for resource-constrained health systems: diagnose systemic weaknesses using the STOE lens, then target interventions precisely using the architectural blueprint. This moves the field beyond checklist-style models toward integrated diagnosis and action.