Abstract
{ "background": "The deployment and sustainability of point-of-care diagnostic devices in resource-limited settings face significant engineering challenges, including harsh environmental conditions, intermittent power, and limited technical support infrastructure. These factors critically impact device reliability and clinical utility.", "purpose and objectives": "This working paper aims to analyse field failure modes of deployed diagnostic devices and to propose a novel, context-adapted biomedical engineering framework for their lifecycle management, focusing on predictive maintenance and local technical capacity building.", "methodology": "We conducted a mixed-methods analysis of device service logs and environmental sensor data from multiple health facilities. Failure time data were modelled using a Weibull proportional hazards regression: $h(t|X) = \\frac{\\beta}{\\eta} \\left( \\frac{t}{\\eta} \\right)^{\\beta-1} \\exp(\\gamma1 X1 + \\gamma2 X2)$, where $X1$ represents mean ambient humidity and $X2$ power fluctuation frequency. Technician interviews provided complementary data on repair challenges.", "findings": "Power quality was the predominant predictor of failure, with a hazard ratio of 2.3 (95% CI: 1.7 to 3.1) for facilities experiencing daily voltage sags. A central theme from interviews was the critical shortage of training on device-specific fault diagnosis, not just generic repair skills.", "conclusion": "Device reliability is intrinsically linked to local infrastructure constraints, necessitating an engineering approach that integrates device design, environmental hardening, and locally executable maintenance protocols.", "recommendations": "Implement mandatory pre-deployment power quality assessments. Develop fault-tree analysis guides tailored to common device models. Establish a centralised digital platform for sharing maintenance solutions among technicians.", "key words": "Medical devices, maintenance engineering, health technology management, predictive maintenance, global health", "contribution statement": "This paper introduces a novel integrated framework that couples statistical reliability modelling with locally sourced ethnographic data to generate actionable maintenance protocols for field