23 - Cost-effectiveness evaluation of a low-cost, non-electric infant warmer for hypothermia in Rwanda

Publication Number: 2020.23

Santiago Reyes, Beth Israel Deaconess Medical Center, Boston, MA, United States; John Zupancic, Harvard Medical School, Boston, MA, United States; Anne R. Hansen, Boston Children's Hospital, Boston, MA, United States; Brian King, Harvard Medical School, Boston, MA, United States

Background: Neonatal hypothermia contributes significantly to morbidity and mortality among preterm and very low birth weight (VLBW) infants in low resource settings (LRS). A cluster-randomized stepped-wedge trial found that a low-cost, non-electric device, The Dream Warmer (DW), used in hypothermic of at-risk preterm or VLBW infants when Kangaroo Mother Care was unavailable, increased euthermia, with uncertain effects on mortality. Implementing such a novel intervention in LRS requires evaluation of its benefits, harms, and costs.

Objective: To evaluate the economic implications of the DW in LRS by estimating the cost per death averted and determining cost-effectiveness among newborns with, or at risk for, neonatal hypothermia.

Design/Methods: We conducted a retrospective, trial-based economic evaluation comparing costs and mortality rates among infants enrolled in the DW trial. We adopted a healthcare-payer perspective and a per-patient level analysis to estimate costs and effects. The primary effectiveness outcome was death before discharge. Costs were based on length of stay, published estimates of average NICU daily costs in Rwanda, and manufacturer-reported device costs (Costs in 2021 USD). The primary outcome of the analysis was the Incremental Cost-Effectiveness Ratio (ICER), expressed as additional costs per additional death averted. Uncertainty was addressed using nonparametric bootstrapping with replacement. Our base-case analysis used a willingness-to-pay (WTP) threshold of $30,000 per death averted, derived from Rwanda´s GDP per capita and average discounted lifespan. A cost-effectiveness acceptability curve (CEAC) explored the impact of varying the WTP threshold on cost-effectiveness probability.

Results: A total of 2,996 infants across 9 hospitals participated in the DW trial. Although survival did not differ significantly between groups (25.18 vs 32.08%, p=0.29), total hospitalization costs increased ($185,000 vs $138,000, p < 0.001) compared to the pre-intervention period, yielding a base-case ICER of $6903 per additional death averted (Table 1).  When sampling uncertainty was considered, there was a ~78.5% probability of cost-effectiveness at our designated WTP threshold of $30,000 per death averted (Figure 1) and 75% even at a WTP threshold of $20,000 per death averted (Figure 2).

Conclusion(s): Availability of the Dream Warmer was associated with higher costs but possibly improved survival to discharge. There was a ~78.5% probability of cost-effectiveness below a WTP threshold of $30,000 per death averted.

Incremental cost-effectiveness of the Dream Warmer intervention compared with standard care
Table 1. Incremental cost-effectiveness of the Dream Warmer (DW) intervention compared with standard care, expressed per 1,000 newborns. Costs are reported in 2021 USD. The table presents total and incremental costs, deaths averted before hospital discharge, incremental deaths averted, and the resulting incremental cost-effectiveness ratio (ICER).

Cost-Effectiveness Plane
Figure 1. The cost-effectiveness plane shows that ~75% of all simulations fall within the upper right quadrant, indicating that the intervention is associated with increased effectiveness at an increased cost. The dashed line represents a WTP threshold of $30,000, with 78.5% of simulations (blue dots) falling below and 21.5% (orange dots) exceeding this value. 

Cost-Effectiveness Acceptability Curve (CEAC)
Figure 2. The cost-effectiveness acceptability curve (CEAC) demonstrates that the probability of the intervention being cost-effective increases rapidly as the willingness-to-pay (WTP) threshold rises. At a WTP of approximately $30,000 per death averted (1x GDP), the intervention shows a 78.5% probability of being cost-effective, increasing to 82.9% at $60,000 (2x GDP) and approaching 83.9% at WTP values above $90,000 (3x GDP).