746 - Pre-Hospital Pulse-Oximetry and Supplemental Oxygen Utilization in Malawi: An Exploratory Cost-Effectiveness Analysis

Publication Number: 746.6507

James Newton, University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, NC, United States; Michael Hawkes, University of British Columbia Faculty of Medicine, Vancouver, BC, Canada; Kenneth Smith, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States

Background: Pneumonia is the leading cause of death globally in children aged 0-5 years, claiming over 700,000 lives annually, the majority from low-resource regions. Strengthening oxygen systems can reduce in-hospital mortality by 50%. However, an estimated majority of deaths occur outside of hospital settings where supplemental oxygen is seldom available. Earlier access to oxygen may be life-sustaining, though research is currently lacking on the efficacy or cost of pre-hospital oxygen in low-resource environments.

Objective: We performed an exploratory cost-effectiveness analysis to examine whether pulse-oximetry and oxygen, when applied in the management of pediatric pneumonia in rural, pre-hospital settings, are likely to be cost-effective interventions according to WHO standards when compared to no pulse-oximetry or oxygen.

Design/Methods: We designed a decision analytic model utilizing TreeAge Pro software (Figure 1), populated with parameters derived from literature review (Table 1). Malawi was selected as a representative country for our model given availability of relevant data. Outcomes were in terms of the Incremental Cost-Effectiveness Ratio (ICER), equating to the direct and indirect costs of our interventions divided by the efficacy, units being USD (2021) per Disability Adjusted Life-Year (DALY) averted. Outcomes were compared against a conservative Willingness to Pay (WTP) threshold of 1x the GDP per capita of Malawi per DALY averted. Because the efficacy of pre-hospital oxygen is unknown, we first confirmed the cost-effectiveness of pre-hospital pulse-oximetry. Two-way sensitivity analysis was then performed for pre-hospital oxygen in tandem with pulse-oximetry to examine ranges of oxygen cost and efficacy for which our interventions remained cost-effective.

Results: Our model yielded an ICER for pre-hospital pulse-oximetry use of $35/DALY averted compared to no pulse-oximetry use, considered cost-effective against a WTP threshold of $588/DALY. When analyzed in combination with pulse-oximetry, we found a baseline WTP cost threshold for pre-hospital oxygen of $71 per patient, with an increased WTP allowance of $4.53 for every 1% reduction in mortality consequent to pre-hospital oxygen use.

Conclusion(s): We conclude pulse-oximetry is likely cost-effective in low-resource, pre-hospital environments like those of Malawi. We suggest ranges of cost and efficacy for which pre-hospital oxygen is likely to be cost-effective in tandem with pulse-oximetry. Further clinical research is warranted to evaluate the effectiveness of pre-hospital oxygen in reducing pediatric pneumonia mortality.

Figure 1.
Schematic diagram of decision analysis model. The square represents the decision to make pre-hospital pulse-oximetry +/- oxygen available in pre-hospital care. Circles represent chance nodes, with assigned probabilities for each of the associated occurrences. Triangles represent terminal nodes, with their payoffs being patient survival vs death and the associated costs of care for each scenario. The symbol [+] indicates a collapsed subtree, hidden for brevity. Each hidden subtree is identical in structure to its alternative pathway, differing only in assigned values. Definitions of abbreviations: Peds PNA = pediatric pneumonia; POx = pulse-oximetry; O2 = oxygen.

Table 1.
Table 1. Parameter estimates for cost-effectiveness of pulse-oximetry and oxygen in rural, pre-hospital settings. (CDS = Clinical Danger Signs; LMIC = Low and Middle-Income Country.)
1 All nominal costs adjusted to real costs in 2021 USD.
2 * Denotes absence of confidence intervals. Relative ranges of +/- 25% used in such cases.