Methodology

Summary of the methodology used to estimate the potential impact on lives saved of fully optimized community health platforms in fifteen countries of the Roadmap

• Present the maximal potential impact from fully scaling up a comprehensive package of interventions that are proven to be implemented successfully at the community level

• Use well-established literature and real-world policies to guide assumptions

• Develop a replicable methodology which could be applied in future analyses to include additional countries

• Connect the results with other global and national-level targets

1. Timeframe: The timeframe for the analysis is 2020 to 2030, to enable comparisons with the SDG targets.

2. Impact metrics: The main quantitative impact measure is the cumulative number of lives saved between 2020 and 2030, which can be shown as maternal death averted, neonatal death averted, child death averted, and still births averted. In addition, we also calculated the estimated mortality rate reductions by 2030 and the corresponding contributions to reaching SDG 3.1 and SDG 3.2 targets.

3. Tool used: The tool used to assess impact is the most current version of the Lives Saved Tool, or LiST (version 5.76). LiST is a well-established modeling tool that researchers and policy makers can use to estimate country-specific maternal, child, and pregnancy outcomes based upon changes in population-level coverage of interventions, while considering the country’s underlying health status, cause-specific mortality distribution, and the best available estimates of intervention effectiveness. LiST has a defined list of interventions and is constantly being updated to contain the latest country-specific data. This data is abstracted from the most recent nationally-representative household surveys and sources (e.g., DHS, MICS, etc.).

4. Package of interventions to be delivered at the community level: We defined and applied a comprehensive “universal package” of interventions in this analysis. This universal package was developed by referencing published literature (the two listed below in particular) and reviewing available national strategies/policy documents that included specific definitions of intervention packages for delivery at the community level.

5. Coverage target for interventions delivered at the community level: We set the target coverage rate as at least 90% by 2030 for each intervention, to be in sync with the commonly used coverage rate for Universal Health Coverage (UHC).

To define a comprehensive package of intervention while staying grounded in existing literature and real-world policies, we applied the rule of thumb of including all interventions contained in the Lives Saved Tool that were found in at least one of the following sources:

• Chou, V. B., Friberg, I. K., Christian, M., Walker, N., & Perry, H. B. (2017). "Expanding the population coverage of evidence-based interventions with community health workers to save the lives of mothers and children: an analysis of potential global impact using the Lives Saved Tool (LiST)". In: Journal of Global Health, 7(2), 020401. doi:10.7189/jogh.07.020401

• Black, R., Walker, N., Laxminarayan, R., Temmerman, M. (2016). "Reproductive, Maternal, Newborn, and Child Health: Key Messages of this Volume". In: Disease Control Priorities (third edition): Volume 2, Reproductive, Maternal, Newborn, and Child Health, edited by R. Black, M. Temmerman, R. Laxminarayan, N. Walker. Washington, DC: World Bank.

• The most comprehensive packages among the 15 Roadmap countries, specifically Ethiopia and Liberia

The package of interventions used in this analysis is show in Figure 3.

While we aimed to define a comprehensive package of interventions, we were limited by several factors (with the availability of interventions in the LiST being the largest constraint) and did not estimate the possible benefits of certain interventions often carried out at the community level or found in government policies. These interventions include the promotion and provision of family planning services, management of labor and delivery in low risk women by skilled attendants, activities to reduce gender-based violence and female genital mutilation, education and screening of non-communicable diseases, support to the disabled and the elderly, and activities related to early childhood development, to name a few.

In particular, management of labor and delivery in low risk women by skilled attendants was not included in the analysis because the intervention as defined in the LiST will soon be updated and likely to be disaggregated by signal functions.

The decision to not include family planning in this analysis is mainly due to the challenge of setting a specific share of unmet need addressed associated with community-level service delivery vis-à-vis other parts of the healthcare system. While the effective provision of several contraception methods at the community level is well demonstrated, the rest of the health care system (especially primary care level) is also essential to ensuring the opportunity for women to obtain a method that suits their needs. Setting too high of a target (e.g., 90%) would overestimate the impact of family planning and underestimate the impact of other interventions in the comprehensive package due to fewer pregnant women and newborns under such a scenario. It is possible, however, to consider including family planning in future analyses for specific countries, when governments can weigh in on the contribution of community health platforms in their respective countries toward reducing unmet need for family planning.

For seven countries, existing documents on national policies or strategies included definitions of specific packages of interventions to be implemented at the community level (Figure 4). We conducted an initial analysis for each country using these government-defined packaged and coverage rates (where available).

In most cases, we found the results to be +/-20% of the estimated impact on lives saved using the comprehensive "universal package", given most government-defined packages are also quite complete. The comparison by country is shown in the table below. More detailed information on the results and assumptions used is available upon request. Further work is needed to refine these estimates in consultation with governments.