A cost-effectiveness analysis of early detection and bundled treatment of postpartum hemorrhage alongside the E-MOTIVE trial

Timely detection and treatment of postpartum hemorrhage (PPH) are crucial to prevent complications or death. A calibrated blood-collection drape can help provide objective, accurate and early diagnosis of PPH, and a treatment bundle can address delays or inconsistencies in the use of effective interventions. Here we conducted an economic evaluation alongside the E-MOTIVE trial, an international, parallel cluster-randomized trial with a baseline control phase involving 210,132 women undergoing vaginal delivery across 78 secondary-level hospitals in Kenya, Nigeria, South Africa and Tanzania. We aimed to assess the cost-effectiveness of the E-MOTIVE intervention, which included a calibrated blood-collection drape for early detection of PPH and a bundle of first-response treatments (uterine massage, oxytocic drugs, tranexamic acid, intravenous fluids, examination and escalation), compared with usual care. We used multilevel modeling to estimate incremental cost-effectiveness ratios from the perspective of the public healthcare system for outcomes of cost per severe PPH (blood loss ≥1,000 ml) avoided and cost per disability-adjusted life-year averted. Our findings suggest that the use of a calibrated blood-collection drape for early detection of PPH and bundled first-response treatment is cost-effective and should be perceived by decision-makers as a worthwhile use of healthcare budgets. ClinicalTrials.gov identifier: NCT04341662.


Mean per-patient DALYs
-0.00266 (-0.00814 to 0.00287) 85.76 * Adjusted difference between severe PPH risks is presented in percentage points, and differences between mean values are presented in the unit of the values.Adjusted for number of vaginal births per hospital, time period, country, the proportion of patients with a clinical primary-outcome event at each hospital and the quality of oxytocin at each hospital during the baseline phase and for clustering using random cluster and clusterby-period effects.** Confidence intervals were constructed using non-parametric permutation tests, by finding the upper and lower boundaries of the intervention effect that would lead to a two-sided P value less than the 5% level (1000 replications).

Supplementary Table 4. Results of deterministic sensitivity analyses (continued) 6) Reduced the number of units of whole blood required for transfusion to 1 units
Adjusted * Adjusted difference between severe PPH risks is presented in percentage points, and differences between mean values are presented in the unit of the values.Adjusted for number of vaginal births per hospital, time period, country, the proportion of patients with a clinical primary-outcome event at each hospital and the quality of oxytocin at each hospital during the baseline phase and for clustering using random cluster and clusterby-period effects.** Confidence intervals were constructed using non-parametric permutation tests, by finding the upper and lower boundaries of the intervention effect that would lead to a two-sided P value less than the 5% level (1000 replications).

88.66
Risk of severe PPH -2.6 (-3.1 to -2.1) 9.87 *Adjusted difference between severe PPH risks is presented in percentage points, and differences between mean values are presented in the unit of the values.Adjusted for number of vaginal births per hospital, time period, country, the proportion of patients with a clinical primary-outcome event at each hospital and the quality of oxytocin at each hospital during the baseline phase and for clustering using random cluster and clusterby-period effects.

Methods
We conducted a budget impact analysis (BIA) for each participating country, based on country-level costeffectiveness estimates, to assess the potential implications of introducing the E-MOTIVE intervention on public healthcare systems expenditure over a 5-year period.
[4][5] Based on discussions within the E-MOTIVE study team, we assumed a linear growth in implementation of the E-MOTIVE intervention in public hospitals from 10% in Year 1 (2024) to 100% in Year 3 (2026).The initial training required for E-MOTIVE implementation was considered a one-off cost before its integration into standard training practices.We estimated training costs based on staff time required, in alignment with table A1, and included a 10% additive adjustment to account for overhead costs and training materials.
For the E-MOTIVE trial, training was conducted at two levels: training of trainers (ToT) and on-site training.Based on discussion with the E-MOTIVE study team, we assumed that one doctor and two midwives attended ToT for 2.5 days and facilitated the training of 20 midwives at their facility.On-site training, led by 1 trainer for 10 midwives, lasted for 1.5 days.Hospitals in the E-MOTIVE trial had approximately 20 midwives and 2000 vaginal births annually: training costs were apportioned accordingly to determine a per-delivery training cost.These costs were incorporated for all deliveries in Year 1, and subsequent uptakes in Years 2 and 3 respectively.Given the assumption that all public hospitals implemented the E-MOTIVE intervention by Year 3, no training costs were applied for Years 4 and 5.
In sensitivity analyses, we explored the potential budget impact of reducing the device cost of calibrated drapes to 1 USD, 0.75 USD, 0.50 USD and 0.25 USD respectively, and decreasing the duration of training by 50%.
All costs are reported in 2022 USD and are not discounted, as the BIA focuses on evaluating the actual expected expenditure within a specific budgetary timeframe.

Results
Supplementary Tables 6-9 show the potential budget impact of introducing the E-MOTIVE intervention in public hospitals in the participating countries over a 5-year period.The base case analysis suggests introducing E-MOTIVE in South Africa could decrease costs to the public healthcare system.In Kenya, Nigeria and South Africa, where the calibrated drape is more costly relative to hospital services, a modest increase in budget would be required to achieve substantially improved PPH-related outcomes.However, in the years following the completion of the initial training to deliver the intervention, the budget impact decreases.
Sensitivity analyses suggest that as the cost of the calibrated drape is reduced, feasibly due to expanded production, the E-MOTIVE intervention becomes substantially more affordable.
It should be noted that these potential budget impact estimates were derived from country-level cost analyses, which for pragmatism are based on whole-trial clinical and utilisation data.Although fully countryspecific data were not applied, we believe they provide a useful indication of the potential implications of introducing the E-MOTIVE intervention on public healthcare system expenditure.Nevertheless, these estimates should be interpreted with caution.

4 ) 5 )
Assigned cost to 2 minutes of midwife time for uterine massage and 5 minutes for examination of the genital tract Adjusted difference (E-MOTIVE -usual care)Increased the number of units of whole blood required for transfusion to 3 units Adjusted difference (E-MOTIVE -usual care)

Table 7 . Potential budget impact of introducing the E-MOTIVE intervention in
Drape device cost reported is prior to adjustment to 2022 USD and 25% adjustment for shipping and handling.

Table 8 . Potential budget impact of introducing the E-MOTIVE intervention in South Africa (2022
Drape device cost reported is prior to adjustment to 2022 USD and 25% adjustment for shipping and handling.

Table 9 . Potential budget impact of introducing the E-MOTIVE intervention in Tanzania (2022 USD)
Drape device cost reported is prior to adjustment to 2022 USD and 25% adjustment for shipping and handling.