Cost-effectiveness and sustainability of improved hospital oxygen systems in Nigeria

Abstract

Introduction: Improving hospital oxygen systems can improve quality of care and reduce mortality for children, but we lack data on cost-effectiveness or sustainability. This study evaluated medium-term sustainability and cost-effectiveness of the Nigeria Oxygen Implementation programme.

Methods: Prospective follow-up of a stepped-wedge trial involving 12 secondary-level hospitals. Cross-sectional facility assessment, clinical audit (January-March 2021), summary admission data (January 2018-December 2020), programme cost data.

Intervention: pulse oximetry introduction followed by solar-powered oxygen system installation with clinical and technical training and support.

Primary outcomes: (i) proportion of children screened with pulse oximetry; (ii) proportion of hypoxaemic (SpO₂ <90%) children who received oxygen. Comparison across three time periods: preintervention (2014-2015), intervention (2016-2017) and follow-up (2018-2020) using mixed-effects logistic regression. Calculated cost-effectiveness of the intervention on child pneumonia mortality using programme costs, recorded deaths and estimated counterfactual deaths using effectiveness estimates from our effectiveness study. Reported cost-effectiveness over the original 2-year intervention period (2016-2017) and extrapolated over 5 years (2016-2020).

Results: Pulse oximetry coverage for neonates and children remained high during follow-up (83% and 81%) compared with full oxygen system period (94% and 92%) and preintervention (3.9% and 2.9%). Oxygen coverage for hypoxaemic neonates/children was similarly high (94%/88%) compared with full oxygen system period (90%/82%). Functional oxygen sources were present in 11/12 (92%) paediatric areas and all (8/8) neonatal areas; three-quarters (15/20) of wards had a functional oximeter. Of 32 concentrators deployed, 23/32 (72%) passed technical testing and usage was high (median 10 797 hours). Estimated 5-year cost-effectiveness US$86 per patient treated, $2694-4382 per life saved and $82-125 per disability-adjusted life year-averted. We identified practical issues for hospitals and Ministries of Health wishing to adapt and scale up pulse oximetry and oxygen.

Conclusion: Hospital-level improvements to oxygen and pulse oximetry systems in Nigerian hospitals have been sustained over the medium-term and are a highly cost-effective child pneumonia intervention.

Keywords: Child health; Health economics; Health services research; Paediatrics; Pneumonia.

Figure 1

Oxygen-related clinical practice change among children (top) and neonates (bottom) admitted to 12 hospitals in southwest Nigeria. Pulse oximetry coverage and oxygen provision expressed as the proportion of all children/neonates unless otherwise specified. ‘Indication for oxygen’ is expressed as the proportion of those given oxygen who had SpO₂<90% or WHO emergency signs. Error bars showing 95% CIs. Note: the oxygen coverage to patients with hypoxaemia indicator is biased in the preintervention period due to extremely low pulse oximetry coverage.