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

Postpartum hemorrhage (PPH), commonly defined as blood loss of at least 500 ml after childbirth, is the leading cause of maternal mortality worldwide and is particularly burdensome in low- and middle-income countries. PPH substantially increases healthcare costs relative to births without PPH and imposes severe economic consequences on households following maternal morbidity or mortality. Despite World Health Organization (WHO) recommendations for PPH prevention and treatment, adherence is poor in many low-resource settings due to late or missed detection of PPH and delayed or inconsistent administration of effective treatments. Visual estimation of blood loss is inaccurate and commonly underestimates bleeding, contributing to treatment delays, and key treatments such as tranexamic acid (TXA) are often administered late. The E-MOTIVE trial evaluated an intervention comprising a calibrated blood-collection drape for early, objective measurement of blood loss and a standardized first-response treatment bundle (uterine massage, oxytocic drugs, TXA, intravenous fluids, examination and escalation). This study reports the accompanying economic evaluation to determine whether early detection with a calibrated drape and bundled first-response treatment is cost-effective versus usual care from a healthcare system perspective.

The paper situates the work within evidence that PPH accounts for roughly 27% of maternal deaths globally, with disproportionate impact in LMICs. Prior studies show hospital care costs for PPH can be up to 2.8 times higher than non-PPH births, and maternal mortality leads to substantial long-term household economic burdens. WHO has issued guidance for prevention and treatment of PPH, including timely administration of uterotonics and TXA. However, adherence is hampered by inaccurate visual estimation of blood loss, leading to under-detection and treatment delays, and by sequential rather than bundled use of effective interventions. Empirical evidence from hospitals in Kenya, Nigeria, South Africa and Tanzania indicates TXA use is often delayed and reserved for severe cases requiring surgery. Technical consultations have advocated PPH care bundles to improve adherence to guidelines. Collectively, the literature supports objective blood-loss measurement and standardized, timely treatment, motivating an evaluation of resource implications and cost-effectiveness of implementing such strategies.

Design and setting: Economic evaluation conducted alongside the E-MOTIVE international, parallel cluster-randomized trial with a 7-month baseline control phase and a 7-month implementation phase, separated by a 2-month transition in intervention hospitals. Randomization was at the hospital (cluster) level using minimization to balance key prognostic factors (number of vaginal births per hospital, baseline prevalence of outcomes, oxytocin quality, and number of hospitals per country). Hospitals were secondary-level facilities in Kenya, Nigeria, South Africa and Tanzania (Pakistan sites were not randomized). A total of 78 hospitals across four countries provided data for analysis, encompassing 210,132 vaginal births (107,733 baseline; 98,722 implementation with source-verified blood loss data). Intervention and comparator: E-MOTIVE comprised a calibrated blood-collection drape with trigger lines for early detection of PPH and the WHO first-response treatment bundle: uterine massage, oxytocic drugs (oxytocin ± misoprostol/ergometrine per protocol), TXA (1 g IV), IV fluids, examination and escalation. Implementation supports included audit and feedback, local champions, restocked trolley/carry case, and on-site, simulation-based training. Usual care relied on visual estimation of blood loss, with components of the bundle administered sequentially and TXA often reserved for refractory bleeding; noncalibrated drapes (without trigger lines) were used solely to measure blood loss for study data collection. Outcomes for economic evaluation: Two outcomes were used: (1) severe PPH prevented, defined as blood loss ≥1,000 ml measured objectively on drapes within up to 2 hours postpartum; and (2) disability-adjusted life-years (DALYs) averted. DALYs combined years lived with disability (YLDs) for nonfatal PPH and years of life lost (YLLs) from maternal death due to bleeding. Disability weights from GBD 2019 were 0.324 for severe PPH (≥1,000 ml) and 0.114 for less severe PPH (<1,000 ml), applied over a 6-week postpartum duration. YLLs were based on country-specific female life expectancy (GBD life tables) with 3% discounting. Perspective, resource use, and costing: Healthcare system perspective. Prospectively collected resource use via eCRFs and REDCap included: calibrated drapes; uterotonics; TXA and its administration; IV fluids and giving set/cannula; hospital ward and ICU length of stay; transfer to higher-level facility; blood transfusions (assumed two units if transfused); laparotomy (assumed 80% of hysterectomy cost); hysterectomy; nonpneumatic anti-shock garment; uterine balloon tamponade (UBT; improvised in some settings or Ellavi device in South Africa); bimanual compression (assumed 30 minutes). Uterine massage and examination were treated as reprioritization of staff time (no additional cost). Unit costs were drawn from established international sources (USAID, UNFPA, MSH International Medical Product Price Guide, WHO-CHOICE, national salary/pay scales) and expert opinion where needed, adjusted to 2022 USD. Tradable goods included 25% for shipping/handling/distribution. Calibrated drape base price: 1.25 USD (2023) with adjustments applied. Country-specific indices (market-basket approach) were used to fill gaps and standardize across settings. Costs were not discounted due to short follow-up. Statistical analysis: Intention-to-treat, complete-case analysis using only source-verified blood loss data. Multilevel (mixed-effects) models accounted for clustering (random cluster and cluster-by-period effects) and adjusted for allocated exposure, time period, country, and randomization covariates. Severe PPH risk differences were estimated using binomial family with logit link and marginal standardization; mean differences in costs and DALYs used Gaussian family with identity link and nonparametric permutation tests due to skewness. Incremental cost-effectiveness ratios (ICERs) were computed for cost per severe PPH case averted and cost per DALY averted. Cost-effectiveness judged against weighted willingness-to-pay (WTP) thresholds: WHO per capita GDP (weighted mean 2,816 USD) and opportunity cost–based thresholds (weighted mean 1,690 USD). Uncertainty assessed via clustered bootstrapping (1,000 replications) to generate cost-effectiveness acceptability curves (CEACs). Sensitivity analyses: Deterministic analyses varied calibrated drape device cost (1.00, 0.75, 0.50 USD; 2023 prices). Multiple imputation for missing data (assuming missing at random with allowance for clustering) assessed robustness relative to the complete-case main analysis. Country-level analyses: Fully pooled, one-country costing CUAs applied country-specific unit costs and life expectancy to all trial participants, generating country-specific ICERs for Kenya, Nigeria, South Africa and Tanzania, compared against country-specific GDP-based and opportunity cost–based thresholds.

Study population and follow-up: 78 hospitals (14 Kenya, 38 Nigeria, 14 South Africa, 12 Tanzania). Total vaginal births analyzed with source-verified blood loss data: 206,455 (98% follow-up). Primary results: Severe PPH occurred in 786/48,678 (1.6%) in E-MOTIVE vs 2,129/50,043 (4.3%) in usual care; adjusted risk difference −2.6 percentage points (95% CI −3.1 to −2.1). Mean DALYs per patient: 0.00767 (SD 0.394) in E-MOTIVE vs 0.01158 (SD 0.454) in usual care; adjusted DALY difference −0.00266 (95% CI −0.00814 to 0.00287). Mean per-patient total cost: 45.14 USD (SD 107.93) in E-MOTIVE vs 43.19 USD (SD 126.84) in usual care; adjusted cost difference 0.30 USD (95% CI −2.31 to 2.78). ICERs: 11.83 USD per severe PPH case averted; 113.91 USD per DALY averted—well below both the weighted GDP-based WTP threshold (2,816 USD) and the weighted opportunity cost–based threshold (1,690 USD). CEAC: For WTP per DALY averted > ~1,500 USD, probability of cost-effectiveness exceeds 80%. Resource utilization: Despite lower overall PPH rates in E-MOTIVE (8.5% vs 16.7%), oxytocin, TXA, and IV fluids were administered more often, reflecting improved detection and bundled response. Usual care incurred more blood transfusions, slightly longer hospitalization, and more additional interventions; more severe PPH cases required additional physician time. Sensitivity and country-level results: Reducing the calibrated drape cost to 1.00 USD made E-MOTIVE approximately cost-neutral and more effective (dominant on point estimates), with greater cost savings at 0.75–0.50 USD. Additional sensitivity analyses (costing assumptions, multiple imputation) did not materially change conclusions. Country-level pooled costing analyses indicated E-MOTIVE was cost-effective in all four countries versus both thresholds; in South Africa, E-MOTIVE was less costly than usual care (dominant on point estimate).

The economic evaluation demonstrates that implementing objective early detection of PPH with a calibrated drape coupled with a standardized first-response treatment bundle substantially reduces severe PPH and likely reduces disability burden at a very low incremental cost, yielding highly favorable ICERs relative to accepted thresholds. These findings directly address the central question of whether such an intervention represents good value for money in resource-constrained settings; the results indicate that E-MOTIVE is a cost-effective and pragmatic strategy for improving maternal outcomes. The CEAC suggests high probability of cost-effectiveness at plausible WTP levels, and deterministic analyses show that modest reductions in drape price could make the intervention cost-saving, improving affordability and scalability. Implementation insights suggest the intervention’s support strategies (audit and feedback, champions, trolley/carry case, on-site training) can be integrated with minimal incremental costs and without additional staffing. Given the direction of effects on severe PPH and maternal deaths due to bleeding, a societal perspective—including patient and household costs—would likely yield even more favorable cost-effectiveness. While the complex cluster design required careful multilevel modeling and adjustments for baseline imbalances, the analytical approach accounted for clustering and multinational context, and sensitivity analyses support robustness. The country-level pooled costing provides indicative local relevance, acknowledging that exact country-specific CEAs would require larger samples and local data. Overall, E-MOTIVE offers an efficient, scalable approach aligned with WHO recommendations, with potential to reduce maternal morbidity and mortality in LMICs.

Early detection of postpartum blood loss using a calibrated drape combined with a bundled first-response treatment is cost-effective compared with usual care in secondary-level hospitals across Kenya, Nigeria, South Africa and Tanzania. At approximately 114 USD per DALY averted and 12 USD per severe PPH case averted, the E-MOTIVE intervention represents a worthwhile use of constrained health budgets and aligns with WHO guidance. Modest reductions in drape device costs could render the intervention cost-saving. Future research should: (1) assess the cost-effectiveness and environmental impact of sustainable/reusable blood loss monitoring devices; (2) undertake comprehensive country-specific CEAs with detailed bottom-up costing; (3) include societal costs and long-term outcomes to capture broader economic impacts; and (4) integrate equity analyses to understand distributional effects and potential to reduce health disparities.

Key limitations include: (1) device cost scenarios focused on current calibrated drapes; sustainable or reusable alternatives were not evaluated; (2) pragmatic design precluded exhaustive bottom-up costing for all resources, increasing uncertainty in some unit costs; (3) several unit costs required assumptions or secondary sources; although prespecified and tested, these may affect precision; (4) societal costs to patients and households were not captured, likely underestimating overall benefits; (5) complex cluster trial structure necessitated modeling assumptions and adjustments; (6) fully country-specific CEAs were not feasible due to power constraints; instead, pooled clinical data with country-specific costing were used; (7) equity impacts were not quantified; (8) primary analyses relied on source-verified blood loss data with 2% missingness; although multiple imputation sensitivity analyses supported robustness; (9) implementation strategy costs were not formally quantified but are believed minimal based on post-trial experience.