Adopting a societal perspective, this research is based on an observational and comparative cost-effectiveness analysis. The analytical scope encompasses the costs related to the project, the beneficiaries, and the implementing partners, for a cohort of pregnant women followed within the Cotonou 1 - 4 health zone.

The selection of study sites was based on data from the baseline survey on the nutritional profile of pregnant women in the Cotonou 1 - 4 health zone [14]. Women were assigned to intervention and control sites prospectively and non-randomly on a geographical basis, determined by their consultation health center; this approach was utilized to prevent inter-group contamination. To ensure data comparability, two study sites were established by matching intervention sites (beneficiaries) with control sites based on geographic, socioeconomic, and ethnic diversity criteria. Each site was structured homogeneously to systematically include both a public and a private health facility. Thus, the first comparison site includes the Ahouansori health center and the John Hot clinic, while the second combines the Aïdjedo health center and the La Confiance clinic.

Before the intervention, an early prediction tool for low birth weight (LBW) risk was developed using key determinants, including socioeconomic and nutritional status, energy intake, occupation, and education level [15]. Participants were classified as “high-risk” if they scored ≥ 5 on the predictive tool. Inclusion criteria consisted of pregnant women residing in the study area who were at less than 14 weeks of gestation and had provided written informed consent. Exclusion criteria included multiple pregnancies, preterm birth, and the presence of severe chronic conditions (such as heart disease or renal failure). This tool accurately identified high-risk women within the study population. Using the initial survey database, we applied these criteria to select all eligible at-risk pregnant women. The final sample consisted of 80 participants, prospectively distributed across the study and comparison sites.

The intervention is based on a combined approach that integrates nutritional supplementation and educational support. The nutritional component consists of the weekly distribution of a kit containing fortified flour made from local products, protein biscuits, and oranges. This support begins in the third month of pregnancy and continues until delivery, covering a follow-up period of approximately six months.

The implementation of this program relies on rigorous, close support, including personalized follow-up and home visits. Nutritional education sessions are integrated into the selected health facilities. To ensure the participation of pregnant women and the effectiveness of the intervention, the midwives in charge of the centers were closely involved in the coordination.

The educational program is structured around a communication plan comprising five thematic sessions with specific objectives. Each session, lasting approximately one hour, is conducted in the local language and uses visual aids to optimize message comprehension. In addition to nutrition, the intervention includes a comprehensive health education component. This covers stress and sleep management, as well as adherence to prenatal care protocols, including the use of insecticide-treated bed nets (ITNs), monitoring of intermittent preventive treatment (Sulfadoxine/Pyrimethamine), deworming, and compliance with iron and folic acid supplementation.

The primary outcome of this study is the incidence of low birth weight (LBW) within the source population. To ensure data reliability, experienced nutritionists measure each newborn’s weight within 72 hours of delivery using high-precision SECA scales. For analysis, birth weights are treated as a dichotomous variable: infants weighing less than 2500 g are classified as LBW, while those at or above this threshold are categorized as having a normal birth weight.

We assessed costs by comparing the direct and indirect expenses of the intervention and control groups. Direct costs are subdivided into two categories: health costs and non-health costs. The reference year for calculation is 2024. The evaluation sources are based on project expenditure records and local market prices. The unit quantities per participant amount to 130 g/day of fortified flour, 70 g/day of biscuits, and 1 orange per day for an average duration of 180 days. Direct health costs encompass all operational expenses related to the intervention, including inputs for formulating flours and biscuits, the purchase of oranges, logistical costs to distribution sites, and the value of staff time dedicated to nutrition and health education activities. Direct non-health costs comprise ancillary expenses such as travel to distribution sites and logistics for home visits. Indirect costs, or lost productivity, were calculated by multiplying the time mothers or caregivers spent at education sessions by the average local daily wage. To ensure the internal consistency of the economic evaluation, participant time and productivity losses (indirect costs) were excluded from the primary Incremental Cost-Effectiveness Ratio (ICER) calculation and are reported separately.

The economic evaluation was based on a cost-effectiveness analysis to determine the incremental cost per case of low birth weight (LBW) averted and per disability-adjusted life year (DALY) saved, using a benchmark indicator that met WHO standards and was comparable to the thresholds. To do this, the incremental cost-effectiveness ratio (ICER) was established by dividing the cost difference (direct costs only) between the group receiving nutritional supplementation and the control group by the difference in observed effectiveness (the incremental number of low-birth-weight cases actually averted in the sample, then converted into DALYs based on a factor of 2.5 DALYs averted per LBW case prevented, to allow for the rigorous application of the GDP-based threshold). This ratio thus expresses the marginal cost per QALY gained through the support intervention compared to standard care. The cost-effectiveness ratio was compared to the WHO threshold: 1 × GDP per capita (798,974 FCFA) for “very cost-effective” and 3 × GDP per capita (2,396,922 FCFA) for cost-effective [16], while the costs related to productivity losses were treated separately in order to avoid biasing the calculation of the main ICER.

To test result robustness, we performed a univariate deterministic sensitivity analysis by varying key parameters, specifically direct unit costs and estimated productivity losses, by ±20%. This approach aligns with CHEERS (Consolidated Health Economic Evaluation Reporting Standards) guidelines, ensuring the evaluation meets international publication standards [17].

The research protocol received approval from the Health Sciences Research Ethics Board (CERS) under number 003-2023/CERS. Written informed consent was obtained from each participant before inclusion. They were informed that their refusal to participate would not entail any harm or risk to them.

The intervention reduced the incidence of low birth weight (LBW) to 0.025, representing 39 normal-weight births out of 40 deliveries (only 1 case of LBW). For comparison, the control group, receiving the standard nutrition and health education intervention, had an incidence of 0.125, or 35 normal-weight births out of 40 (5 cases of LBW). This improvement represents an absolute decrease in the incidence of LBW of 10% in favor of the intervention group, which corresponds to an incremental number of 4 cases of low birth weight actually avoided in our sample (5 cases in the control group minus 1 case in the intervention group). These results are confirmed by a relative risk of 0.2 with a 95% confidence interval between 0.023 and 0.58, demonstrating a significant 80% reduction in the risk of LBW among women who received the new intervention (Table 1).

Table 1. Effectiveness of the intervention.

The total intervention cost was 3,273,000 FCFA, split between direct and indirect expenses. Direct costs accounted for 70.66% (2,313,000 FCFA) of the budget, driven largely by input procurement, specifically flour, biscuits, and oranges, which totaled 1,983,000 FCFA (85.73% of direct costs). The remaining 14.26% of direct funds supported operations, including 130,000 FCFA for logistics and 200,000 FCFA for facilitator per diems. Indirect costs, representing participant time, were estimated at 960,000 FCFA. In contrast, the control group’s standard program cost significantly less at 1,330,000 FCFA, covering facilitation, midwifery follow-up, and participant time compensation. Consistent with our methodological approach, indirect costs related to participant time have been excluded from the primary ICER calculation presented below.

The cost-effectiveness analysis indicates that the standard nutritional intervention costs 1,330,000 FCFA to achieve 35 normal-weight births. When comparing the food support program to the nutrition education program alone, the difference in direct costs amounts to 983,000 FCFA (2,313,000 FCFA versus 1,330,000 FCFA). The Incremental Cost-Effectiveness Ratio (ICER), calculated based on the actual number of cases prevented, is 245,750 FCFA per additional low-birth-weight case avoided (983,000 FCFA divided by the 4 additional cases prevented within the sample).

When applying a standard health indicator conversion, assuming 2.5 DALYs saved per averted case, for a total of 10 DALYs, the incremental cost-effectiveness ratio is 98,300 FCFA per DALY averted. This figure represents the additional cost required to prevent one extra low-birth-weight case or to achieve an additional QALY/DALY. Although this cost exceeds that of the base intervention, it remains highly competitive, as it is more than eight times lower than the threshold defined by the WHO for interventions considered “highly cost-effective” (798,974 FCFA).

The univariate sensitivity analysis, summarized in the table below, assesses the robustness of the Incremental Cost-Effectiveness Ratio (ICER) in response to ±20% variations in key input parameters.

The results indicate that the intervention’s economic viability is primarily driven by production costs, with the highest sensitivity observed in the formulation costs of biscuits and flours. Specifically, this parameter exerts a significant -35.3% impact on the Cost/DALY ratio, causing it to fluctuate between 95,700 FCFA and 100,900 FCFA. This classifies it as a high-impact variable for the model, followed by the cost of orange procurement, which shows moderate variation (-6.05%).

Conversely, the system demonstrates resilience to distribution factors. Variations in logistics costs result in only marginal impacts on the final outcome (-2.64). Classified as having low impact, this parameter suggests that transport-related uncertainties do not significantly compromise the intervention’s overall efficiency, confirming the model’s stability in the face of logistics-related fluctuations (Table 2).

Table 2. Sensitivity analysis results.