Introduction

In the last two decades, Kenya has shown steady progress in reducing undernutrition. Between 1998 and 2014, the prevalence of stunting among children under five years decreased from 33% to 26%, wasting reduced from 6% to 4%, and underweight decreased from 22% to 11% [1]. Presently, Kenya is on course to meet four global nutrition targets on under-five stunting, wasting, overweight, and exclusive breastfeeding [2].

Globally, undernutrition costs the world at least US$ 761 billion per year, or US$ 2 billion per day, equivalent to 1% of global income [3]. While progress has been significant, undernutrition remains unacceptably high in Kenya. A 2019 Cost of Hunger in Africa study estimated that in 2014, the effect of undernutrition on health, education, and productivity amounted to about US$ 4.2 billion or roughly 7% of Kenya’s 2014 GDP [4].

In 2014, 26% of Kenya’s 7.22 million under-five population (approximately 1.8 million children) were stunted [1], and county-level stunting prevalence ranged from 15% in Nyeri to 46% in West Pokot [5]. High stunting levels were found in the Coast, Eastern, and Rift Valley regions, whereas the highest wasting prevalence was observed in the Northern Region [5]. Of Kenya’s 47 counties, 19 had a stunting prevalence above 30%, and wasting prevalence ranged from 0.2% in Siaya to 23% in Turkana [5]. High regional variation results from differences in poverty levels, food insecurity, education, access to water and sanitation, access to and use of health services and infrastructure, and governments’ budget restraints across regions. It is important to note that these estimates are from before the onset of the COVID-19 pandemic, which caused widespread unemployment, food insecurity, and disruptions to health systems [6], with facility closures [7] and treatment unavailability [8] contributing to an estimated additional 450–2,120 child deaths and 38–196 maternal deaths monthly in Kenya [9].

Unfortunately, the trends in gross official development assistance (ODA) and nutrition-financing have not followed this increasing need. Even at its highest, nutrition-related ODA was less than 1% of the total ODA in Kenya, despite high returns on investment [10]. A 2016 analysis by the World Bank estimated that if critical high-impact nutrition interventions are scaled up in Kenya, the benefit-cost ratio of the investment would be 22:1, i.e., every dollar invested in nutrition interventions would potentially result in $22 in economic returns [11].

In Kenya, public health care delivery is devolved; the 2010 Constitution gives Kenya’s 47 counties responsibility for delivering most health services while the national government retains leadership in developing health policy and regulation and in managing national referral facilities. Given the decentralized system of governance in the country, the significant subnational variation observed in trends of undernutrition, and the limited ODA allocated to Kenya, it has increasingly become essential that county governments spearhead financing nutrition interventions.

As a part of their response to this need, in 2018, the Government of Kenya developed its second National Nutrition Action Plan. The Kenya Nutrition Action Plan 2018-2022 (KNAP) utilized existing country-level policy and regional frameworks to outline the investment required to address malnutrition in all its forms and for all ages. The KNAP provided an umbrella framework and guidance for developing County Nutrition Action Plans (CNAPs) [1]. The costed CNAPs were created to tailor essential nutrition actions to meet the specific needs of individual counties and prioritize accordingly - as well as leverage additional county-level resources for nutrition. The high-impact nutrition interventions included in the CNAPs were the same as those identified in the 2016 World Bank report [11].

The CNAPs were created by the country’s two levels of government in collaboration with key line ministries, development partners, NGOs, and civil society organizations. The authors’ organization and other development partners supported county governments with technical and financial support for the CNAPs. In addition, County Investment Cases (CICs) were created to estimate the potential health impacts and economic benefits of investing in low-cost, high-impact direct interventions included in the CNAPs. The primary goal of the development of the CICs was to mobilize county governments to increase investment in nutrition actions by illustrating the connection to human well-being and economic development.

This complemented Kenya’s Universal Health Coverage (UHC) policy [12] which strives to ensure that all Kenyans can access quality health services without financial hardship, guided by the principle of “leaving no one behind.” Achieving this vision requires addressing key health determinants, with nutrition recognized as a foundational and cost-effective intervention to improve health outcomes across all populations. The UHC Policy 2020–2030 highlights the importance of multisectoral approaches and prioritizing preventive and promotive health strategies. In this context, investing in low-cost, high-impact nutrition programs at the county level offers a strategic opportunity to advance UHC goals.

Investment cases for development sectors, including health and nutrition, have been used increasingly globally and nationally in the last decade. Yet, few examples exist of the approach being applied to subnational decision-making. A secondary goal for this exercise was to develop a consistent data driven methodology to project health impacts and conduct benefit-cost analysis on investment of subnational nutrition plans. While many studies and analyses have tried to estimate the returns on investment in nutrition [13,14], each study applies a different methodology and uses data of varying degrees of quality, resulting in different conclusions. Applying a consistent methodology to estimate the economic returns on allocating resources for the implementation of CNAPs allows for better comparison across geographies.

This paper aims to present both the quantitative results of the CICs for five counties that launched their CNAPs (Busia, Elgeyo-Marakwet, Makueni, Nandi, and Vihiga) and discuss the lessons learned from the process to add to the knowledge base on the value of and the methods for economic evaluation of nutrition interventions at the subnational level.

Methods

The CICs included two key steps - projecting the forward-looking health impacts of the proposed multi-year CNAPs, and conducting benefit-cost analysis and cost-effectiveness analysis. Ideally, investment cases for nutrition could also include a form of optimization analysis, which factors in budget constraints and/or priority outcomes to generate a recommended optimal set of interventions for investment and a few proposed realistic financing scenarios for policymakers and investors to consider. These two steps were not included in Kenya’s CNAP and CIC process.

The five counties—Nandi, Busia, Makueni, Vihiga, and Elgeyo Marakwet—were selected because they were the pioneers in spearheading the creation of County Nutrition Action Plans (CNAPs). These counties were the first to be ready to initiate the CNAP development process, positioning them ahead of other counties in the implementation timeline. County governments undertook a detailed exercise to develop fully costed CNAPs between 2019 and 2020. This activity was a real-world ingredients-based costing exercise designed to collect data on the costs, inputs and assumptions related to the price, quantity, and targets for every line item in the government’s budget related to nutrition. Kenya’s most recent DHS survey from 2014 (at the time of analysis in 2021) was used as the source of data for several demographic and health input parameters. The CNAPs were developed by county governments in collaboration with national governments and partner organizations. Details on the CNAPs, including methods used and assumptions made, are available in the published CNAP studies [15–19].

This CNAP costing data was reanalyzed to create the CICs to support advocacy for financing. The indicators of interest from the costing analysis were the unit cost per intervention per beneficiary, total cost per intervention, and total cost of each broad category of nutrition program.

No ethical clearance was required as this study utilized secondary, and anonymized data. The findings aim to promote equity and support underserved populations, aligning with public health objectives to improve nutrition outcomes in Kenya.

Targets for high impact direct nutrition interventions

The baseline and target coverage (%) figures from the CNAPs were used to estimate the cumulative persons reached by each CNAP intervention. The baseline and target coverage rates from the CNAPs for each HIDI are shown in Table 1, and we assumed a linear scaling of coverage between those years. The calculation of the unit costs of each intervention was based on the sum of the cost of the ingredients divided by the total number of target beneficiaries over the CNAP’s time frame. With 2018 as the baseline year as per the CNAPs, the costing and subsequent impact analyses were conducted for a 4-year timeframe from 2019 to 2022.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. CNAP targets for high-impact nutrition interventions (%).

https://doi.org/10.1371/journal.pgph.0004128.t001

Results

The five counties were chosen to create investment cases due to their high burden on undernutrition. Evidence shows that those who suffer from undernutrition and diseases are more likely to repeat grades in school and experience lower productivity levels into adulthood [30]. Table 2 presents the prevalence of undernutrition in the five focus counties in 2014 showing that all counties experienced high rates of stunting and anaemia in children. While county level data was not available for prevalence of anaemia in women, regional estimates suggest that the counties experienced between 40-55% anaemia rates among women.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Annual cases and costs of undernutrition by county.

https://doi.org/10.1371/journal.pgph.0004128.t002

The health impact of high impact nutrition interventions

Over the implementation period of the CNAPs, high-impact direct interventions are projected to prevent 1,835 child deaths across the five counties: 267 in Nandi, 276 in Vihiga, 650 in Busia, 340 in Makueni, and 302 in Elgeyo Marakwet. Additionally, it is projected that 115 maternal deaths will be prevented across the five counties: 23 in Nandi, 19 in Vihiga, 39 in Busia, 23 in Makueni, and 12 in Elgeyo Marakwet.

Cumulatively over four years and five counties, the projected health outcomes are estimated as approximately 19,000 cases of stunting averted, 4,700 cases of wasting averted, and 67,000 case-years of anaemia averted. In cases of anaemia, the protection is only assumed to last for one year, and ongoing supplementation would be essential to the long-term prevention of anaemia cases (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Projected cases averted due to reaching targets of interventions included in the CNAPs.

https://doi.org/10.1371/journal.pgph.0004128.g001

In total, these projected health outcomes averted by the CNAPs are equivalent to averting a total of 142,471 of DALYs across the five counties over four years, including 23,181 in Nandi, 21,151 in Vihiga 47,987 in Busia, 27,236 in Makueni, and 22,916 in Elgeyo Marakwet.

The financing need for high impact direct nutrition interventions

The total financing need for the CNAPs was approximately USD $ 17 million on average per county over four years. The costs of the interventions grouped together under the HIDI category made up the highest cost category for each county, close to 35% of the total cost (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Total financing need for each county over four years by intervention type (USD $ Millions).

https://doi.org/10.1371/journal.pgph.0004128.g002

Within HIDI, IYCF and IMAM programs made up most of the costs. These interventions include overlapping activities and use similar equipment, and are often carried out together in Kenyan counties. For example, children under 5 years are monitored on their height and weight and overall health as part of routine IYCF growth monitoring programs. Most cases of acute malnutrition are identified during screening in the routine growth monitoring and IYCF programs. On average, across the counties, IYCF and IMAM interventions made up about 70% of the total cost of high-impact direct interventions captured in the CNAPs. Notably, there was significant variation in the cost allocation of the other high-impact direct interventions. For example, WIFAS intervention was allotted 1% of the total high-impact direct budget for Nandi, Vihiga, and Makueni, and assigned 7% and 23% of the total cost respectively for EMC and Busia’s HIDI budget. Similarly, VAS programs made up 0.3% of the total cost for Vihiga, 6-7% for Busia, Nandi and Makueni, and about 12% of EMC's total HIDI budget. S2 Fig presents the cost breakdown of high-impact direct interventions except for IYCF and IMAM interventions to provide further detail.

Table 3 presents the unit costs of high-impact direct interventions by county, compared with national-level unit costs [11] and global unit costs [32]. Generally, the unit costs align with global estimates from the literature, but there are a few significant outliers. These discrepancies can be attributed to two primary factors. First, measurement errors during the costing process may have introduced inaccuracies. Second, counties independently negotiated commodity prices with suppliers, leading to significant cost variation—some counties secured lower prices, while others faced higher costs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Calculated unit costs, cost-effectiveness, and cost-benefit values of high-impact direct interventions.

https://doi.org/10.1371/journal.pgph.0004128.t003

Additionally, differences in the proposed reach of interventions played a critical role in unit cost variation. For counties like Nandi and Vihiga, where the proposed reach increased dramatically (e.g., from 0% to 75% coverage), unit costs were lower due to economies of scale. Conversely, in counties where the proposed reach was relatively small (e.g., from 90% to 95% coverage or 0% to 25% coverage), the total reach was smaller, but costs remained similar to those of other counties, resulting in higher unit costs.

The lack of standardized commodity costs, combined with discrepancies in program scale and proposed reach, inherently shaped the data. These subnational unit cost calculations, therefore, provide a nuanced and context-specific understanding of intervention costs while highlighting the challenges of working with real-world, diverse datasets.

Discussion

The results of the County Investment Case (CIC) analyses in Kenya, across five counties, describe a consistent pattern showing that high-impact direct nutrition interventions will drastically save and improve lives in a cost-effective manner. The actual health impact of the CNAP investment could be possibly be much higher if the nutrition sensitive, other public health, and enabling environment investments (i.e., infrastructural costs, such as buying equipment, building breastfeeding rooms in facilities) were included in the analysis. Furthermore, the investment is expected to generate savings in the health and education sector and increase productivity in the long run. The default discount scenario (3%), with many conservative assumptions included, suggests that the CNAPs will generate a high projected benefit-cost ratio that ranges between $5-14 for every dollar invested. This aligns with broader evidence from similar studies, which found benefit-cost ratios in the range of 5.2 - 19.4 in different contexts, respectively, for similar nutrition interventions [34,35] and, for Kenya, a benefit-cost ratio of 5.8 for priority nutrition packages as estimated in the WBG Investment Framework 2024 [36]. Our analysis adds to the literature by focusing on subnational planning that considers differences between counties, supports decision-making in Kenya’s decentralized system, and uses a consistent method to compare the economic benefits of CNAP investments across regions.

The CNAP process in Kenya is the first time such detailed costing investment data have been made available at the county level in Kenya. While there is much that can be improved in the CNAP process, this first iteration was important for policymakers’ engagement in the nutrition planning and financing process. This analysis also demonstrated that there is a significant amount of variation at the subnational level in terms of the burden of undernutrition, the priorities selected by subnational policymakers and stakeholders, and the unit cost of nutrition interventions. As a result, there is variation in the cost-effectiveness across counties, which suggests that each county requires a unique optimal mix of interventions to include in their financing strategy.

The authors must note that while the analysis presented in this paper was conducted in 2019 with data from Kenya Demographic Health Survey (KDHS) 2014, newer data is now available through the KDHS 2022.The newer data shows that the burden of stunting has significantly decreased in the 5 focus counties, though prevalence of wasting and low birthweight still remain high.

Future CNAP processes in Kenya, and around the world, could be improved in several ways if the following limitations are addressed: First, the authors do not account for externalities such as inter-county migration rates, any disruptions to the health system (such as a pandemic), and effects of climate change on nutrition, while calculating the health impacts of the activities included in the CNAPs. The focus counties have some of the highest rates of net negative migration in the country (the highest being 10%) [37]. However, even accounting for this net negative migration across the counties, the interventions maintain high benefit-to-cost ratios, ranging from $4-12 return per dollar invested. Investing in public nutrition would still yield a positive return. Second, the strength of county-level analyses depends on the availability and quality of data at the county level. Certain analyses did not have data available for key inputs into the models at the county level, and therefore used proxies from national level. Using county level data will generate a better overall picture of the state of undernutrition in the counties. Third, the authors assumed each case of stunting and wasting averted as unique, i.e., the analysis did not account for the possibility that the same child may be likely to be stunted or wasted more than once during the CNAP timeframe. Fourth, the analysis showed high variation in unit costs across the counties. While a common costing framework was applied in the project, in-depth exploration of the costing data shows that in many cases, the methods for costing were not consistently applied across counties, and input costs varied drastically across counties. Integrating a review and feedback process may decrease some of the variation in costs and improve the quality of CNAPs altogether. Fifth, the CNAP process highlighted variation in costs for commodities, which justified several counties coordinating themselves in 2021 to do pooled procurement of commodities at an optimal cost via the Kenya Medical Supplies Authority. Further comparative analysis of the costing data across counties could lead to other important lessons learned and cost-saving approaches. Sixth, the CNAPs included activities for a wide-spectrum of nutrition actions, not only evidence-based high-impact low-cost nutrition interventions. Future CNAP processes could include either a prioritization process to identify the most cost-effective nutrition actions to invest in before the costing exercise, or an optimization analysis at a later stage that generates recommendations for an investment strategy given budget constraints or specific goals from policymakers, and readjust aspirational targets for scaling set earlier to more realistic levels.

In addition, the CNAPs filled an essential gap in the nutrition planning process for counties, however it led to a plan that was too aspirational. For example, the total costs of the CNAPs presented in this analysis make up about 20-22% of the total fiscal budget of many counties. Within this, the estimated costs of the activities amount to, on average, 60% of the counties’ health budgets [38–42], with significant variation observed across counties. In Kenya, counties aim to allocate 15% of their total budgets to the health sector, and at present, most counties are not able to meet this criterion. Particularly in increasingly constrained fiscal environments, future CNAPs must focus on allocating investments to high-impact nutrition interventions that are of the most value to the county and should include another step of review to ensure that the CNAP targets are plausible and that the subsequent financing need is realistically affordable for county policymakers.

Conclusion: application of CICs in future

The CICs are a powerful tool that can be used for resource mobilization for nutrition at the county level. Between 2020 and 2022, the authors’ organization first used the CICs to build a case on the consequences of not investing and the projected returns on high impact nutrition interventions in discussions with decision-makers at county level to lobby for increased allocation of country budgets to nutrition. As a result of these deliberations, county governments ratified a three-year joint financing agreement with the authors’ organization for implementation of low-cost high impact nutrition interventions at scale.

There is high potential for applying CICs for knowledge sharing, advocacy, and policymaking to inform and influence investments in nutrition. It is often difficult to present the scale of the loss to a region’s productivity due to undernutrition and to quantify the potential impact to be gained from investing in nutrition. The CICs fill this crucial gap. Additionally, given that most existing analyses often focus on national-level results, the CICs highlight how the need for resources and potential health impacts vary considerably across different contexts. It may be recommended to also to integrate subnational county nutrition action planning (CNAPs and CICs) into the national nutrition planning process (i.e., KNAP 2023-2027) to promote bidirectional cohesiveness in the strategy, financing and implementation as seen in some other countries such as the Philippines [43]. Integration of nutrition planning processes at various government levels with a standard approach along with incorporating improved costing, prioritization and optimization feedback loops would be recommended to increase the allocative efficiency of nutrition investments in the future. In recent years, there has been an increasing focus on generating data for decision-making and financing, so this analysis for Kenya provides a framework and case study for undertaking future CICs as part of a domestic resource mobilization process in alignment with the current global focus.

Supporting information

Acknowledgments

The authors would like to acknowledge colleagues at the Nutrition International country office in Kenya and government partners who supported this research’s conceptualization, data collection, application, and knowledge translation. We are grateful to Richard Morgan and Joel Spicer for their critical feedback.