1.1 The Safer Births Bundle of Care (SBBC)

The SBBC is a continuous quality improvement intervention aiming at improving maternal and newborn outcomes at all stages of labour and early neonatal life [7]. The design of the SBBC intervention was developed based on the experience of the Helping Babies Breathe programme introduced in 2009 at Haydom Lutheran Hospital (HLH) in Tanzania, which was highly cost-effective [8]. The SBBC intervention employs clinical training coupled with simulation-based training using intermittent low-dose-high-frequency training to ensure competency and skills retention among healthcare providers [9]. In addition, innovative tools are used along with a bundle for monitoring foetal and newborn heart rates, facilitating newborn resuscitation, and for the simulation training of postpartum care for mothers and newborns [7,10]. In a feedback loop, mortality is recorded continuously for each facility and communicated back to the respective healthcare staff. The intervention targets labour management, newborn resuscitation, and management of bleeding after birth as main areas. To ensure scalability and that the program is rooted in the national health organization and local facilities, the initial training was organized as a training cascade, where eight national trainers trained 15 regional trainers who again train two so-called facility champions at each facility. These facility champions organize and supervise the simulation-based LDHF training and adjust it according to the data they receive via the feedback loop. Phase I of its implementation was conducted in Tanzania between 2021 and 2023. Thirty health facilities from five regions (Manyara, Tabora, Geita, Shinyanga and Mwanza) were selected based on the high burden of maternal and perinatal mortality and the high volume of deliveries [10]. All three levels of healthcare services, from primary (health centers) to secondary (district hospitals) and tertiary (regional referral hospitals), in both rural and urban settings, were represented. All facilities provided Comprehensive Emergency Obstetric and Basic Newborn Care services and had separate labour units and operating theaters. The collection of baseline data started in March 2021, while the implementation was rolled out in a stepped-wedge design between June 2021 and December 2023 [10]. No other relevant changes were made to the health systems or treatment recommendations during the study period.

Due to promising midterm results, an expansion of the programme has been launched and is currently performed as Phase II in a total of 150 facilities in the same regions, with the prospect of a nationwide expansion in Phase III [10,11]

Recently, outcome data of Phase I, based on almost 300,000 recorded mother–baby pairs, were published, reporting a significant decrease in perinatal and maternal mortality [10]. However, to inform decision-makers and potential funders, and to fully evaluate the SBBC Phase I intervention, its costs and cost-effectiveness have to be analyzed. This study thus presents the costs stratified into the start-up, running and research costs of the SBBC Phase I and cost-effectiveness analysis, where the cost per life saved, cost per life years gained, and cost per disability-adjusted life year (DALY) averted and the benefit–cost ratio (BCR) is determined.

2.1 Cost data collection

Cost data were based on secondary data collected retrospectively in March 2024. Data sources were the itemized budget of the project activities and the actual spending on the itemized activities as registered in the financial records. In addition, interviews were conducted with the principal investigator and the head of the HLH accountancy centre. Based on this information, costs were categorized into start-up, running and research costs. Start-up costs comprised costs incurred at the pre-implementation and preparatory phases of the SBBC Phase I intervention, such as initial training or information for stakeholders [12]. Running costs included among others refresher training, continuous low-dose-high-frequency simulation-based training and supervision. Research costs comprised costs directly related to the scientific evaluation of the SBBC intervention, such as data collection or scientific conferences.

Costs were calculated from a programme’s perspective. All costs were estimated in Tanzanian Shillings (TZS) and converted to 2020 USD using a mean exchange rate of TZS 2,305/USD provided by the Central Bank of Tanzania [13].

2.2 Effectiveness data collection

The number of lives saved was based on prospective data collection by two trained data collectors located at each facility throughout the study. For babies, the variables collected included, among others, fresh stillbirths, newborn deaths within 24 hours and seven days, foetal heart rate monitoring, foetal presentation, birth weight, sex and multiple pregnancies. From the mothers, data were collected on maternal death, age, parity and whether she was referred from another hospital.

2.3 Statistics

The number of perinatal and maternal lives saved was calculated based on the findings reported by Kamala and Ersdal et al., [8]. The estimated number of lives saved during the project period was calculated by taking the estimated risk differences between the baseline period and the implementation period and multiplying them by the number of births in the implementation period. The estimated risk differences were calculated as the differences in death incidences between the baseline period and implementation period adjusted for clinically relevant demographic and clinical factors as reported in Kamala and Ersdal et al., [8]. The number of perinatal life years gained was based on the life expectancy at birth in Tanzania as reported by the Tanzanian National Bureau of Statistics, 2022 [14]. The number of maternal life years gained was based on the remaining life expectancy for mothers reported by World Life Expectancy and the National Bureau of Statistics [14,15]. DALYs averted were calculated based on Fox-Rushby and Sassi [16,17], using the following parameters: Discounting rate (r) = 0.03 per year and age weighting with K = 1 and β = 0.04. Disability weights were not applied as we lacked data on morbidity.

2.4 Cost-effectiveness analysis

The cost-effectiveness ratio was calculated by dividing the costs by the number of lives saved, life years gained and DALYs averted when comparing the pre- and post-implementation data of the SBBC Phase I intervention. International Dollars were equated to USD. The BCR was calculated based on the value of statistical life years gained of 2,983 USD as stated by the World Health Organization (WHO) [18].

2.5 Sensitivity analysis

To assess the robustness of this cost-effectiveness analysis, a one-way sensitivity analysis (SA) was conducted by altering the following parameters (Table 3):

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Sensitivity analysis results.

https://doi.org/10.1371/journal.pone.0323234.t003

1. We assumed that the number of lives saved was the lower limit of the 95% confidence interval (CI) of perinatal and maternal lives saved, respectively the higher limit.
2. We assumed that labour costs were double, respectively trebled.
3. We assumed that there were no research costs.

2.6 Ethical considerations

The SBBC Phase I project has received formal approval from the National Institute of Medical Research (NIMR) for Tanzania with Ref. NIMR/HQ/R.8c/Vol. I/2060 dated 14 June 2022 and amendment Ref. NIMR/HQ/R.8b/Vol. I/1040 dated 8 August 2022 [19]. The extension was made to Ref. NIMR/HQ/R.8a/ Vol.IX/3458, dated 26th February 2025. The CUHAS Directorate of Research and Innovations also authorized the publication of a Manuscript with Research Clearance Number. CREC/832/2024 of 20th March 2025. Local permission to use the costs and outcome data was sought and obtained from HLH management. No consent was sought from the study population as this study used outcome data as secondary data. However, data confidentiality was ensured.

3.1 Lives saved, life years gained and DALYs averted

A total of 297,755 mother–baby pairs were enrolled during SBBC Phase I implementation, and of these, 281,165 mothers and 277,734 babies were included in the final analysis of cost-effectiveness evaluation.

The total estimated number of lives saved during SBBC Phase I was 580 (95% CI 225, 935) perinatal lives, 382 (95% CI 138, 626) maternal lives, and 962 combined lives saved. Based on the life expectancy at birth in Tanzania in 2022 [14] of 65·5 years the number of potential life years gained for newborns was 37,914 years. The calculation of maternal life years gained was based on the average age of reproduction in this study of 25·8 years and the remaining life expectancy for women in Tanzania at this age of 48·1 years, resulting in 18,374 life years gained [15]. Thus, the total number of life years gained was 56,288. The total number of DALYs averted was 31,221 when age weighting and discounting were applied.

3.2 Costs of the SBBC intervention

All the costs incurred were borne by the SBBC Phase I project, which was fully funded by the World Bank through the Global Facility Financing. Funds were administered by UNICEF, whereby nine percent of the total funds were charged as overhead. The total cost budgeted for the implementation of SBBC Phase I was 4,491,204 USD, while the actual spending was 4,489,848 USD, showing an average of 0·03% deviation between budget and expenditure. However, within the cost items, higher deviations occurred. For example, the cost of national and international travel of faculty was reduced by 9·8% due to the pandemic. Funds not spent on some items such as travel expenses were transferred to other items that needed more funds than the amount budgeted.

Table 1 presents the costs drivers according to resource use. Salaries and staff benefits amounted to 35·2% of all project cost, followed by per diems and allowances for training which summed up 18%. Thus, 53·2% of all expenditure was allocated to labour costs. These were followed by office and equipment maintenance (16.6%), travel costs (9.4) and office administration (9.2%). Other costs included scientific documentation and research publications (4.7%), conference/meeting expenses (4.1%) and so forth in reduced proportions Figs 1 and 2.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Costs for Phase I of the SBBC program, stratified according to costs item.

https://doi.org/10.1371/journal.pone.0323234.t001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. The four main cost drivers in all cost categories of start-up, running and research costs in the order of higher to lower resource use.

https://doi.org/10.1371/journal.pone.0323234.g001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Costs stratified into start -up cost, running cost, and research cost.

https://doi.org/10.1371/journal.pone.0323234.g002

3.3 Costs per life saved, life years gained, DALYs averted and BCR

Table 2 presents the costs per life saved, life years gained and DALYs averted. Costs per life saved were 4,669 USD, and costs per life year gained were 82·5 USD for combined perinatal and maternal outcomes. Costs per DALY averted (combined) were 31·2 USD. To rank the costs of an intervention per DALY, the WHO discerns three categories of cost-effectiveness: highly cost-effective (less than gross domestic product (GDP) per capita); cost-effective (between one and three times GDP per capita); and not cost-effective (more than three times GDP per capita) [18]. As the GDP per capita in Tanzania was 1,193 USD in 2022, the SBBC programme can be ranked as highly cost-effective [19]. Based on the value of a statistical life year of 2,983 USD, the BCR of Phase I was 36.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Cost-effectiveness parameters of the SBBC phase 1 program.

https://doi.org/10.1371/journal.pone.0323234.t002

3.4 Sensitivity analysis

A sensitivity analysis (SA) of the number of lives saved was performed by Kamala and Ersdal et al. to assess the impact of modelling choices and assumptions regarding missing outcome data [8]. The results of the one-way SA by altering the outcome and cost parameters as outlined under methods are presented in Table 3. In a worst-case scenario with treble salary costs and a lower limit of lives saved, the programme costs would result in 810 USD per DALY averted and thus still be ranked as highly cost-effective, while costs in the best-case scenario of a higher limit of lives saved and no research costs would amount to 72 USD per DALY.

4.1 Study strengths and limitations

The SBBC Phase 1 was a multi-centre, prospective and observational study, with high-quality outcome data collected. In addition, it provided high statistical power due to a large sample size of almost 300,000 mother–newborn dyads observed [10]. Finally, the study settings were heterogeneous including health facilities in rural, semi-urban and urban settings, thus increasing the transferability of the study findings to other settings.

However, as discussed in more detail by Kamala and Ersdal et al., the study might suffer from incomplete or under-reported baseline data such as the number of fresh stillbirths, which could affect the evaluation of the effectiveness, and result in an underestimation of the number of lives saved [10,24,25]. The collection of cost data was performed retrospectively, based on the recordings of the accounting centre, and thus important information might not be captured, though the results from various sources were cross-checked to ensure validity. In addition, this study is merely a partial economic evaluation as it only deals with the programme’s perspective, not capturing all costs arising to the health care provider like staff time and utilities, and leaving behind the societal part of economic evaluation. As we could not disaggregate the costs into costs allocated to newborn versus maternal care, the presented cost-effectiveness results for perinatal and mothers separately are of a rather theoretical value. As we lacked data on morbidity, DALYs were calculated based solely on years of life lost due to premature mortality (YLL) based on life expectancy in Tanzania, while we did not include years of healthy life lost due to disability (YLD). Instead, we assumed that there was no additional morbidity caused by saving the lives of newborns and mothers and no morbidity averted. This might have resulted in both an over- or underestimation in DALYs averted. Our findings of costs per DALY averted have thus to be treated with caution. However, our approach is in the same range as previous studies, where the number of DALYs averted in low-income countries was estimated as either half of the life years gained or 32 DALYs per life saved [5]. Also, a one-way SA might underestimate the overall uncertainties of parameters and ignore the correlation between parameters and hence not provide information about which parameters critically impact cost-effectiveness. Lastly, as the study lacks information on the costs of standard care, an incremental cost-effectiveness ratio could not be provided.

4.2 Policy implication

The SBBC Phase I appeared to be feasible in terms of financial perspectives and it is aligned with the national programmes for reproductive and child health regarding training design, health commodities, technology, and the health information system as well as other health system building blocks [6]. In addition, our results show that Phase I was highly cost-effective as the SBBC combines different interventions, thus lowering the costs [10]. The bundle addresses not only the challenges of the national reproductive and child health programme of the Ministry of Health in the United Republic of Tanzania but also the global efforts to achieve universal health coverage and sustainability goals 3.1 and 3.4 [26]. The bundle is user-friendly in the different levels of the healthcare system such as health centers, district hospitals and regional referral hospitals, and particularly in rural settings where most of the deliveries take place [22]. Rural settings are likely to have poorer coverage of quality health services and a higher share of home deliveries due to inadequate healthcare provision such as skilled staff or lack of the necessary infrastructure and equipment [10,27]. The design of the bundle that enables the local staff to perform continuous simulation-based training to retain their skills renders the programme more self-sufficient and thus less dependent on continuous financial aid [25]. The training cascade with a strong anchorage at the local health facilities with low-cost and easy-to-maintain training and treatment equipment will facilitate the implementation of the bundle in other LMICs. As of today, it is implemented in Nigeria, but outcome data is not yet available.

4.3 Conclusion

The analysis of the SBBC Phase I intervention showed that it was highly cost-effective. In addition, it was feasible from a financial perspective and aligned with Tanzania’s national programme for reproductive and child health. Currently, the bundle is being scaled up in Phase II, financed by the World Bank. Phase II implements the full bundle, including a training cascade, simulation-based training, the introduction of innovative tools, and a feedback loop based on local outcome data. However, as the scale-up also extends to health facilities with a lower burden of maternal and perinatal mortality and lower volumes of deliveries, efficiency may be altered. A nationwide scale up is planned, but awaits clarification of funding availability. Health-economic evaluations during these scale-ups must provide information on compliance with the original programme content, its clinical and financial sustainability, and the robustness of its cost-effectiveness.