Solar-powered oxygen delivery to treat childhood pneumonia in low-resource settings: a randomised controlled non-inferiority trial and cost-effectiveness study

Abstract

Background: Pneumonia is the leading cause of paediatric mortality globally, causing more than 900 000 deaths every year. Access to oxygen is essential for the treatment of hypoxaemia in pneumonia, but such access remains limited in low-resource settings. Current methods of oxygen delivery are often unsustainable in low-income and middle-income countries (LMICs), since oxygen cylinders require a robust supply chain and operator training, and concentrators rely on consistent access to the power grid. Therefore, gaps remain in oxygen delivery in low-resource settings despite the potential to reduce pneumonia mortality by up to 35% if access were consistent. Through an initial proof-of-concept study, a randomised controlled trial, and further cost-effectiveness analysis, our team has compared outcomes for solar-powered oxygen delivery with that of standard oxygen cylinders for paediatric patients with pneumonia. Methods: In 2014, we installed two solar oxygen systems in Jinja and Kambuga, Uganda. These systems consist of solar panels, batteries, and an oxygen concentrator, and cost USD$15 420 and $17 328, respectively. In 2015, we performed a randomised controlled trial to compare outcomes for children with pneumonia who received solar oxygen with those who received standard cylinder therapy. The primary outcome was length of hospital stay with a secondary outcome of mortality. Furthermore, we performed a cost-effectiveness analysis using WHO-CHOICE guidelines, and recorded qualitative assessment of the systems from health care workers. The trial is registered at ClinicalTrials.gov (NCT02100865). Findings: We included 130 patients treated in Jinja and Kambuga between March, 2014, and May, 2015. Median length of stay for the solar group was 4·1 days, which was 0·41 days less than the cylinder group (95% CI −1·2 to 0·43), meeting the prespecified criterion for non-inferiority of within 1 day. Mortality was similar between the solar group (17%) and cylinder group (12%, p=0·62). Solar oxygen has a cost per disability-adjusted life year (DALY) saved of $27 per DALY, which is substantially less than the reported values for oxygen cylinders ($50 per DALY) and other pneumonia interventions such as the pneumococcal and Haemophilus influenzae tybe b vaccines. Maintenance costs are projected at $90 per month (including battery replacements) over the life of the system. Health-care workers reported that solar oxygen systems are easier to use than cylinders. Interpretation: Solar oxygen is a reliable and cost-effective option to provide oxygen therapy for children with pneumonia in low-resource settings. An expansion to 20 more sites in Uganda is being planned, with a stepped-wedge randomised controlled trial to assess mortality benefit. There will also be a focus on developing best practices for training staff to use and maintain these systems. Our team has engaged with local implementing partners, including health authorities, to ensure effective uptake of solar oxygen. Further reductions in cost and experience with maintenance are required to ensure effective widespread implementation. Funding: Grand Challenges Canada - Stars in Global Health.