Cost analysis of vaccine-preventable diseases surveillance in Ethiopia

Abstract

Background:

This study evaluated the cost of the vaccine-preventable diseases (VPD) surveillance system in Ethiopia to inform budgeting for disease detection, prevention, and control, and planning for transitions from donor funding.

Methods:

This cross-sectional retrospective, bottom-up micro-costing study collected data on resource utilization to conduct VPD surveillance during one fiscal year (2018–2019) in Ethiopia. The study covered 16 VPDs and costs from the government and partners perspective.

Results:

The estimated economic costs of VPD surveillance in Ethiopia were US$69.11 million or US$0.70 per capita, while financial costs were US$26.83 million or US$0.27 per capita. The largest economic cost was for labor (41.0%) and the largest financial cost was for supplies (36.9%). Resources were mostly allocated to integrated, general disease surveillance (economic: 54.6%; financial: 50.7%), followed by surveillance for measles and rubella (economic: 15.9%; financial: 22.5%) and polio (economic: 12.4%; financial: 15.7%). The main funder was the Ethiopian Ministry of Health (economic: 78.8%; financial 62.6%), including donor resources channeled through the ministry.

Conclusions:

Human resources were primary drivers of economic costs. Although the Ethiopian VPD surveillance system uses shared resources across multiple VPDs, resulting in scale efficiencies, and the Ethiopian government covers most of its economic costs, a substantial portion relies on direct external donor support.

Background

Surveillance for vaccine-preventable diseases (VPDs) delivers numerous benefits: detecting outbreaks rapidly; identifying unreached and under-immunized populations; monitoring progress toward disease elimination and eradication goals; assessing disease burden to inform policy decisions on vaccine introduction and program implementation strategies for optimal use; tracking circulating strains to guide vaccine choice, schedules, and development; and evaluating vaccine impact (see Appendix for more detail).¹

The importance of VPD surveillance systems is affirmed in the global Immunization Agenda 2030 (IA2030)² and the accompanying Global Strategy on Comprehensive VPD Surveillance,¹ which present a strategy to achieve comprehensive, high-functioning, and sustainable surveillance systems at country, regional, and global levels to effectively detect and respond to VPD outbreaks.¹ In many low- and middle-income countries, infectious disease surveillance has been supported through funding from the Global Polio Eradication Initiative (GPEI) for staff and operational support for polio surveillance.³ Upon achieving polio-free certification status and transitioning out of GPEI support, countries will experience reduced donor funding. It is expected that countries gradually take ownership of VPD surveillance systems, including costs, and plan for sustainable operation at an optimal level.⁴

Despite the critical role of VPD surveillance, there is limited information on the costs of establishing surveillance for new diseases and maintaining and strengthening existing surveillance platforms.^5,6 A 2018 study commissioned by the World Health Organization (WHO) estimated the expenditure on VPD surveillance in 63 low- and middle-income countries using comprehensive multi-year plans (cMYPs), which are country-level five-year budget projections of resource needs for national immunization programs.⁶ This study reported substantial variation across countries in the amount spent per capita on VPD surveillance, as well as the surveillance categories included in their plans. Furthermore, in 2019, a systematic review commissioned by WHO found only 11 VPD surveillance costing studies, conducted primarily in Africa and not assessing all VPDs, and concluded that these data are insufficient to determine country-level and global costs of VPD surveillance for different diseases and in different settings.⁵ An improved understanding of country-level surveillance costs will facilitate planning and budgeting, allow national programs to budget sufficient funds for VPD surveillance activities, and support advocacy efforts at country and international levels to solicit supplemental donor funds, where required, to implement optimal surveillance.

This article presents findings from a country-level costing analysis in Ethiopia, which has experienced a significant VPD burden in the African Region (Box 1). The analysis objective was to estimate the economic and financial costs of the VPD surveillance system in Ethiopia in 2018–2019 to inform planning and budgeting for sustaining quality VPD surveillance activities.

Box 1. Country profile and VPD surveillance system in Ethiopia.

Country Profile

In 2022, Ethiopia had 123 million inhabitants,⁷ with a birth cohort of 3.9 million.⁸ Approximately a quarter of Ethiopians (24% in 2016) live under the national poverty line⁸ with an average gross national income per capita of US$1020 in 2022.⁹ In 2023, Ethiopia was considered a Gavi (the Vaccine Alliance) priority country, with national-level coverage of the third dose of diphtheria, tetanus toxoid, and pertussis vaccine (DTP3) at 72% (2023).¹⁰

The surveillance system in Ethiopia helps detect outbreaks, identify under-vaccinated groups for informing immunization outreach strategies, and inform vaccine introduction decisions and vaccine impact assessments. In addition to the government’s health infrastructure, the WHO Expanded Program on Immunization (EPI) Unit established an integrated VPD surveillance network which is composed of 42 surveillance medical officers (SMOs) located in nine regions and two administrative cities. Each SMO covers at least two zones, which have several districts and woredas. The SMOs conduct case investigations, monitoring and supervision for acute flaccid paralysis (AFP), neonatal tetanus, measles and rubella surveillance, in addition to supporting routine immunization and supplementary immunization activities (SIA).

Ethiopia has experienced substantial VPD burden (Appendix Table A.I), including recurrent measles outbreaks, as well as periodic outbreaks of other VPDs, such as cholera for which vaccines are not included in the routine immunization program.^11–13

The country is at risk of poliovirus importation due to low routine oral poliovirus third dose (bOPV3) vaccine coverage in border woredas (equivalent to districts) (national coverage of 72% in 2023¹⁰); low national coverage of the first dose of the inactivated polio vaccine (IPV1) (72% in 2023,¹⁰); and population movements, civilian conflict, political instability, and conflicts in neighboring countries. Excluding 2018, and 2020–2021 due to the COVID-19 pandemic, Ethiopia has met the minimum standards of polio surveillance indicators at national level for over a decade (Appendix Table A.II), but there are subnational gaps in surveillance.^14–16 In addition, the cMYP 2016–2020 highlighted various weaknesses of the VPD surveillance system: inadequate triangulation of information, weak data verification efforts at each administrative level, and weak links between VPD surveillance and the Ethiopian Federal Ministry of Health’s (EFMoH) Expanded Programme on Immunization (EPI). Such weaknesses can lead to delayed detection of outbreaks; more costly outbreak responses (because outbreaks are more widespread with delayed detection); and preventable health and economic burden to households, health systems, and government.¹⁷ Furthermore, the civil conflict that started in 2020 has blocked surveillance and outbreak response in certain zones of the country.

VPD Surveillance in Ethiopia

In 2019, the EFMoH included 15 VPDs in its disease surveillance system (Appendix Table A.III). This system included surveillance for outbreak-prone VPDs such as polio, measles, rubella, yellow fever, meningitis, typhoid, and cholera, as well as sentinel site surveillance for influenza, rotavirus, and invasive bacterial VPD (e.g. Streptococcus pneumoniae, Haemophilus influenzae type b).

Ethiopia achieved polio-free status certification in June 2017. However, there were circulating vaccine-derived poliovirus type 2 (cVDPV2) outbreaks between 2018 and 2022.^18–20 The current VPD surveillance system depends partially on GPEI funding, workforce, and infrastructure. Ethiopia was budgeted to receive an annual average of US$3.4 million in GPEI funding for the period 2017–2023 (Appendix Table A.IV). The GPEI Transition Independent Monitoring Board assessed that halting GPEI funding to Ethiopia could result in rapid deterioration of the VPD surveillance system of the country, leading to delayed disease detection and increased risk for larger-scale outbreaks, reduced immunization program performance, and lack of evidence for immunization policy and program strategy decisions.¹⁵

Methods

This study was a program costing analysis, characterized as a cross-sectional, retrospective, bottom-up, microcosting ingredients-based study with primary and secondary data collection on resource utilization and costs. The time frame and analytic horizon were the 12 months in the Government of Ethiopia’s fiscal year 2018–2019, from 1July 2018 through 30 June 2019. It included all administrative levels in Ethiopia: federal, regional, zonal, woreda, and health center/hospital.

The study took the perspective of the public health sector for VPD surveillance, composed of the Ethiopian Public Health Institute (EPHI)—Public Health Emergency Management Division (PHEM), EFMoH (including the EPI), WHO Ethiopia (Vaccine-Preventable Disease Surveillance Unit), and the CORE Group implementing partners forum. Therefore, only costs (monetary expenses and in-kind resources) incurred by these entities were collected. Additionally, to focus on the cost of routine VPD surveillance, patient care, vaccination, and shipping costs to international laboratories were excluded from the analysis because they were out of the direct scope of the study. The cost of outbreak investigation and response was excluded from the analysis as well due to potential annual variation attributable to differing annual incidences of VPDs.²¹

Each cost was captured and analyzed according to five dimensions: resource input, support function, administrative level, funding source, and VPD (Box 2).

Box 2. Overview of methods.

This study estimated the costs to operate the VPD surveillance program for 12 months in Ethiopia. Retrospective data were collected over the Ethiopian government’s fiscal year 2018–2019, from 1 July 2018 through 30 June 2019. Spending related to conducting VPD surveillance was reviewed to understand the source and destination of funds. Information was gathered from all levels of administration in Ethiopia: federal, regional, zonal, woreda, and health facilities (health center/hospital). The data originated from interviews with staff implementing the VPD surveillance program as well as from existing documents.

This study used the perspective of the Ethiopian government and its implementing partners.

Costs included in the analysis were all financial costs (i.e. monetary outlays) and economic costs (i.e. all financial costs plus the monetary value of in-kind/existing resources deployed to conduct surveillance) of the VPDs included in the study by the sampled cost centers at different administrative levels. The values of in-kind or existing resources were estimated using their monetary value, if known; otherwise, the commercial value of a similar item was used for analysis.

Costs excluded from the analysis were those not paid for by the in-country partners (external quality assurance and quality control provided by global and regional reference labs), shipping costs to international laboratories, patient costs, vaccination costs, and cost of outbreak investigation and response. The last three activities were excluded because the study focused on the cost of routine VPD surveillance.

Each cost was captured and analyzed according to five dimensions: resource input, support function, administrative level, funding source, and VPD. Resource inputs included labor; equipment; building and utilities; supplies; transport and vehicles; fuel; per diem, lodging, and travel costs; contracted services; and other. The support function categorization followed the guidelines from the Global Strategy on Comprehensive VPD Surveillance¹ and included governance, program management, workforce capacity, field logistics and communication, laboratory, supervision, data management and use, and coordination. The administrative level categorized costs by whether they were incurred at federal, regional, zonal, woreda, or health center/hospital levels. The funding source was defined as the organization that purchased or owned the resource based on records and interviews; funds were not traced to their original source (e.g. domestic government revenues or external donor funds within government expenditure). Finally, costs were also classified by the VPD targeted by surveillance.

The sampling units were cost centers (offices, laboratories, or health facilities) to which the surveillance activity costs can be allocated. This study included all cost centers at federal level, a purposive sample of sites at regional level, and a mix of purposively and randomly selected sites at sub-regional levels (Appendix and Table A.V). Results were extrapolated from sampled to similar non-sampled sites to obtain country-wide estimates (Box 3).

All VPDs in the Ethiopian surveillance system with vaccines that were included [or recommended for inclusion by the National Immunization Technical Advisory Group (NITAG)] in the national immunization program of the country were included in this study. The VPDs included in the study were: polio (AFP), measles and rubella (grouped together) and congenital rubella syndrome, neonatal tetanus, pediatric meningococcal meningitis, yellow fever, cholera, human rabies, diphtheria, pertussis, typhoid fever, smallpox (included as legacy disease, no vaccine used), rotavirus, and invasive-bacterial VPDs (Streptococcus pneumoniae and Haemophilus influenzae type b, grouped together). Surveillance for influenza was not included in the analysis as influenza vaccine is not included in the national immunization program of Ethiopia (Appendix Table A.III).

Data were collected from March 21 to May 27, 2022, by the Ohio State Global One Health LLC Ethiopia (GOH LLC Ethiopia). Microsoft Excel-based tools on tablets were used to collect data.

Cost data were collected either in local currency (Ethiopian Birr) or U.S. dollars, depending on the source of information, and adjusted to March 2023 US$ values based on local monthly inflation and exchange rates.^22,23

All VPDs in the Ethiopian surveillance system with vaccines in the national immunization program of the country were included in the study (influenza surveillance was excluded as influenza vaccine is not included in the Ethiopian national immunization program) (Appendix Table A.III).

Sampling

The sampling units were cost centers. A cost center was defined as a unit (office, laboratory, or health facility) to which the surveillance activity costs can be allocated. The sampling strategy for this study included all cost centers at federal level, a purposive sample at regional level, and a mix of stratified random and purposive sample of sites at sub-regional level (Appendix and Table A.V).

Data collection and analysis

Cost data were collected between 21 March and 27 May 2022, by GOH Ethiopia, adjusted by exchange rate and inflation, and presented in March 2023 US$ values.

Although the sub-national study sample was not selected to be statistically representative at country level, sampling weights were used to extrapolate the results to non-sampled sites within the same region (Box 3).

Box 3. Weighting and extrapolation strategies.

Weighted costs at each administrative level were first estimated for the location where they occurred (e.g. health facility, woreda, zone) and then aggregated at regional level. For the randomly sampled cost centers, each cost was multiplied by the product of the administrative-level weights for that cost center level and higher levels within the region. For example, for weighting costs from a health facility, those costs were multiplied by [(facility weight) x (woreda weight) x (zone weight)]. The sampling weight of the cost center was determined by the inverse probability of selection of that cost center. Probability of selection was calculated considering all characteristics used in the sample selection, such as urban–rural strata and facility type. For the purposively selected cost centers, costs across cost centers of the same type in the same administrative level and geographic area (e.g. health facility type, woreda, zone, region) were averaged and multiplied by the administrative-level weights for that cost center level and higher levels. In this case, the sampling weight was determined by the number of cost centers of the same type within an administrative level and geographic area, also considering rural–urban strata. For example, for two health centers sampled in a rural woreda, costs were averaged between the two health centers and multiplied by the total number of health centers in that rural woreda, and then by the weights at higher administrative levels (woreda and zone). Within a sampled region, weighted costs were summed to estimate the total regional weighted costs.

To estimate costs at country level, the weighted costs of the sampled regions were extrapolated to non-sampled regions. MOH and EPHI experts were consulted to group the regions based on similar characteristics, ultimately using wealth quintiles as of 2019.²⁴ Costs from Addis Ababa were extrapolated to Dire Dawa and Harari; those from Somali to Afar, Benishangul-Gumuz, and Gambela; and those from Amhara to Oromia, Tigray, and Southern Nations, Nationalities, and Peoples Region.

Data were analyzed in Microsoft Excel and Stata. Costs were estimated at country level, per region, per capita, and per each of the five dimensions used to classify costs.

Results

Overview

The total countrywide weighted financial and economic costs of the VPD surveillance system in Ethiopia for the fiscal year 2018–2019 were US$26 832 811 and US$69 105 310, respectively. The total financial cost per capita and total economic cost per capita were US$0.27 and US$0.70, respectively (Fig. 1). The largest financial and economic costs per capita were in the group of city regions of Addis Ababa, Dire Dawa, and Harari (US$1.15 and US$2.99, respectively).

Figure 1. Total Weighted Costs of Ethiopia’s VPD Surveillance System by Aggregated Regions and Federal Level (March 2023 US$) (July 2018-June 2019).

Abbreviations: SNNP = Southern Nations, Nationalities and Peoples.

* = Extrapolated region for weighted results.

By one cost dimension

By administrative level, the largest costs were incurred by health facilities (financial: US$9 509 653; 35.4%; economic: US$40 017 506; 57.9%) followed by the federal level (financial: US$7 681 796; 28.6%; economic: US$8 815 815; 12.8%) (Table 1). Appendix Tables A.VI, A.VII, A.VIII, and A.IX present results disaggregated by administrative level and each of the other four cost dimensions.

Table 1. Ethiopia GPEI Surveillance Budget by Year (nominal US$).

Objective 1: Poliovirus detection and interruption, Activity: Surveillance
	Budget Year							Annual Average
Sub-activity	2017	2018	2019	2020	2021	2022	2023
Laboratory	0	0	0	0	0	0	0	0
Surveillance and running costs	3,296,000	3,395,000	3,395,000	2,500,000	2,187,000	2,187,000	2,187,000	2,735,286
Technical assistance	0	517,000	591,000	591,000	1,011,000	1,011,000	1,011,000	676,000
Total budget for surveillance	3,296,000	3,912,000	3,986,000	3,091,000	3,198,000	3,198,000	3,198,000	3,411,286

Source: GPEI Budget 2017–2023 (10).

Laboratories, a key component in the VPD surveillance system, totaled US$4 385 419 in financial costs (16.3% of total costs) and US$4 875 881 in economic costs (7.1% of total costs). The national reference laboratory under EPHI accounted for the largest share of costs among laboratories [95.0% of financial and 92.1% of economic costs (Appendix Fig. A.1)], consistent with its role as the national reference laboratory processing samples from regions across the country.

By VPD syndrome, the largest share of weighted total costs countrywide was for the category general (not specific to one VPD) (financial: 50.7%; economic: 54.6%), followed by measles and rubella surveillance (financial: 22.5%; economic: 15.9%) and polio surveillance (financial: 15.7%; economic: 12.4%) (Appendix Table A.VI).

By resource input, the largest financial and economic costs were for supplies (36.9%) and full-time labor (41.0%), respectively (Appendix Table A.VII). By support function, the largest share of financial costs was for field logistics and communication (26.3%), followed by laboratory (24.0%), while the largest share of economic costs was for workforce capacity (25.2%), followed by supervision (20.4%) (Appendix Table A.VIII).

By funding source, the EFMoH at federal level provided the largest amount of resources (financial 51.7%; economic 68.5%), followed by WHO (financial 36.7%; and economic 15.6%) (Table A.IX), inclusive of any external donor funding channeled through these implementing organizations.

By two cost dimensions

Although the resource input allocation across VPDs varied widely, there were some patterns. Supplies; vehicles; and per diem, travel, and lodging were the three resource inputs with the largest financial cost across most VPDs (Fig. 2 and Appendix Table A.X). Measles and rubella (61.6%), meningococcal meningitis (63.5%), and yellow fever surveillance (71.3%) allocated more than 60% of their financial resources to supplies, while neonatal tetanus (59.0%) with no lab component, and cholera (68.6%) spent more than half of their financial resources on vehicles. Polio surveillance spent 44.0% of its financial resources on per diem, lodging, and transport.

Figure 2. Total Weighted Costs of the Laboratories of Ethiopia’s VPD Surveillance System [March 2023 US$, (%)] (July 2018-June 2019).

Abbreviations: SNNP = Southern Nations, Nationalities and Peoples.

* = Extrapolated laboratory for weighted results.

When considering economic costs, full-time labor, supplies, and vehicles were the three resource inputs with the largest share across most VPDs. Surveillance for yellow fever (58.4%), rotavirus (71.3%), CRS (72.7%), human rabies (72.1%), and the category ‘other VPDs’ (68.0%) used more than half of their economic resources in full-time labor. Polio (29.5%), neonatal tetanus (39.5%), and cholera surveillance (42.7%) allocated approximately a third or more of their economic resources to vehicles. Measles and rubella surveillance spent the highest proportion of its economic resources on supplies (34.0%).

The support functions of field logistics and communication, laboratory, and supervision were those with the largest share of financial resources across VPDs, while field logistics and communication, supervision, and workforce capacity comprised the largest share of economic costs (Fig. 3 and Appendix Table A.XI). Surveillance for polio (financial 60.7%; economic 49.2%), neonatal tetanus (financial 67.3%; economic 36.6%), and cholera (financial 66.2%; economic 33.8%) used the largest share of their resources in field logistics and communications. Laboratory functions absorbed most of the financial resources of measles and rubella (61.0%), yellow fever (72.0%), and invasive bacterial VPDs (62.8%). Surveillance for CRS (31.0%), human rabies (30.1%) and other VPDs (35.0%) used more than 30% of their respective financial resources on supervision. In contrast, workforce capacity constituted the largest share of economic costs for yellow fever (34.6%), rotavirus (35.4%), CRS (45.4%), human rabies (46.4%), and other VPDs surveillance (36.8%). Finally, supervision used more than 20% of the respective economic resources for neonatal tetanus (22.9%), rotavirus (26.3%), and other VPDs surveillance (23.7%).

Figure 3. Total Weighted Costs of VPD Surveillance System by VPD and Resource Input (Top-three resource inputs, March 2023 US$) (July 2018-June 2019).

Notes: The category “Other” within each pie chart includes the sum of the costs of the seven other resource inputs that did not rank in the top three.

By funding source, the EFMoH covered the largest share of financial costs for general surveillance (56.7%), measles and rubella (54.0%), neonatal tetanus (66.3%), cholera (92.4%), CRS (75.8%), human rabies (68.7%), and other VPDs surveillance (81.5%), while WHO covered the largest share for polio (75.1%), yellow fever (72.4%), invasive bacterial VPDs (69.8%), and rotavirus (48.6%) surveillance (Fig. 4 and Appendix Table A.XII). When considering the economic resources, the EFMoH covered the largest share of all VPDs, except for polio surveillance, which was mainly funded by WHO (40.1%). Although the Government of Ethiopia was reported to be the source of the largest amount of resources, identifying the original source of funding to the government was beyond the scope of this study; it is possible that some resources reported as funded by EFMoH may have been supported by funding originally provided by external donors.

Figure 4. Heat Map of Percentage of Total Weighted Costs of Ethiopia’s VPD Surveillance System by VPD and Support Function (July 2018-June 2019).

Considering resource inputs used for each support function, supplies represented the largest financial expense for laboratory (92.9%), program management (67.5%), and data management and use (33.4%), while per diem, travel and lodging was the most expensive resource input for field logistics and communication (50.6%) and workforce capacity (63.3%) (Appendix Fig. A.2 and Table A.XIII). Regarding economic costs, full-time labor was the highest resource input for coordination (86.4%), data management and use (80.8%), governance (69.5%), and workforce capacity functions (82.1%), while supplies were the highest for laboratory (82.4%) and program management (38.0%).

Among all the resource inputs, most financial resources funded by the EFMoH went to supplies (45.9%) and most of its economic resources went to full-time labor (51.2%) (Appendix Fig. A.3 and Table A.XIV). WHO’s financial and economic resources were primarily allocated to per diem, travel, and lodging (financial: 43.4% and economic: 39.6%) followed by supplies (financial: 35.1% and economic: 32.0%). The regional governments spent 42.5% of their financial resources on vehicles and 30.2% of their economic resources on full-time labor.

Analyzing the funders by support function, the EFMoH spent the largest share of financial resources on program management (27.5%) and field logistics and communication (20.4%), and the largest proportion of economic costs on workforce capacity (30.2%) and supervision (22.8%) (Appendix Fig. A.4 and Table A.XV). In the case of WHO, these were field logistics and communication (financial 42.8%; economic 42.3%) followed by laboratory (financial 37.3%; economic 34.5%).

Discussion

Main findings of this study

This study estimated the economic and financial costs of VPD surveillance in Ethiopia in 2018–2019 as US$0.70 and US$0.27 per capita, respectively, in 2023 US$(or $0.48 and $0.19 per capita in 2015 US$). Total health expenditure in Ethiopia for 2019 was approximately US$39 per capita (2023 US$), and the government (including all levels) covered ∼22% of this expenditure.⁷ Expenditures by non-governmental entities were larger than those from government, at 78%, of which domestic private sources and external donors comprised 33% and 44%, respectively.⁷ This composition may suggest that the government has limited fiscal capacity to cover healthcare and to increase its expenditure on VPD surveillance, and this pattern may continue in the short-term. The estimated VPD surveillance costs in our study were equivalent to 2% of total health expenditure per capita in Ethiopia in 2019.

What is already known on this topic?

Financial costs estimated in this study are higher than the projected financial expenditure from the WHO African Region VPD Surveillance Investment Case of US$0.00–0.13 per capita (2023 US$; or US$0–0.10 per capita in 2015 US$) estimated from the information reported in the cMYPs for the lowest funded country group, which included Ethiopia.¹⁷ The estimates from this study are also higher than those from studies included in a previous systematic review of VPD surveillance costs, which ranged from US$0.03–US$0.16 per capita (in 2016 US$), mostly for single VPDs.⁵ Although this discrepancy may be due to methodological differences in the estimations (e.g. cMYPs are projected resource needs rather than actual resource use, and include fewer activities than those included in this study), it suggests that VPD surveillance could be more expensive than previously known and demonstrates the need for tailored studies at country level that reflect more accurately its costs covering all its components.

When compared with another VPD surveillance cost study from Nepal,²⁵ which follows a similar (but not identical) methodology, the estimated economic cost per capita in Ethiopia is approximately three times larger (US$0.70 in Ethiopia vs. US$0.26 in Nepal, 2023 US$) (Table 2). In both cases, the country governments covered most of the economic costs (79% in Ethiopia and 56% in Nepal), followed by WHO (16% in Ethiopia and 44% in Nepal). Labor was the resource input with the largest proportion of economic costs in both studies (41% in Ethiopia and 51% in Nepal). In Ethiopia, the costliest support function (i.e. program activity) was workforce capacity (25%), which includes personnel costs for active surveillance visits, while active surveillance visits were the largest cost category in Nepal (18%). The discrepancy in total amounts could be due to the specific epidemiologic conditions of each country, the VPDs included in each surveillance system, the standard operating procedures for the respective VPD surveillance systems, and differences in sampling strategy and personnel remuneration structure among others.²¹

Table 2. Surveillance performance indicators against Global Polio Eradication Initiative and Measles & Rubella Partnership targets.

VPD	Description	2018	2019	2020	2021	2022	2023
Polio	Non-polio AFP ratea	2.59	2.89	3.06	4.03	4.46	3.47
	Percentage of adequate specimensb	77%	91%	58%	56%	93%	95%
Measles and rubella	Discarded non-measles/rubella cases per 100,000 total populationc	1.6	1.57	1.16	1.31	3.5	2.66

Abbreviations: VPD = Vaccine Preventable Disease; AFP= acute flaccid paralysis; N/A = Not available.

What this study adds?

Although these findings cannot be generalized to all contexts, they suggest that VPD surveillance costs vary significantly by country setting and could be more expensive than previous estimates based on country budgets alone. Findings highlight the need for conducting country-level studies to better understand the specific cost drivers of the country system. Even in countries receiving considerable amounts of external donor support for the health sector, such as Ethiopia, funds and in-kind resources from country governments may provide a substantial proportional support to VPD surveillance. In Ethiopia, WHO currently provides significant support to the country’s VPD surveillance system, which must be considered for planning donor funding transition. The findings of this study may aid the Ethiopian government in budgeting for the transition and making investments to strengthen the VPD surveillance system.

Future research may include programmatic indicators of the VPD surveillance system (e.g. timeliness of case detection, completeness of reporting, data accuracy) linking them to costs. The COVID-19 pandemic demonstrated the value of surveillance, and strong VPD surveillance systems can provide the foundation for rapid detection and response to future pandemic threats, among other critical ‘always-on’ use cases for immunization optimization. Understanding whether higher costs correspond to higher quality of VPD surveillance will inform value-for money assessment and will provide insights on the amount of additional resources that would be needed for strengthening the system.

Furthermore, the large shares of resources allocated to general VPD surveillance demonstrated that certain surveillance functions are integrated across VPDs, which may suggest cost savings due to reduced duplication of pathogen-specific activities. Future research could further explore potential cost savings from integration across VPDs and of VPDs within broader communicable disease surveillance systems. Additional cost-saving strategies for potential examination include implementing tiered testing protocols that use lower-cost screening tests prior to confirmatory diagnostics, negotiating bulk purchasing agreements for laboratory supplies, integrating specimen transport network across disease programs, adopting electronic reporting systems to streamline data management, and using remote training programs to minimize travel and venue costs.

Limitations

The results of this study should be interpreted in light of several limitations. First, costs were collected in 2022, and respondents may have underestimated or overestimated their retrospective allocation of time or resource shares for the fiscal year 2018–19 due to recall bias. Second, the sample was statistically non-representative at country-level, and it included a small sample size of cost centers across administrative levels. This aspect limits the generalizability of our results both to the broader country level context and to settings outside of Ethiopia. Third, due to ongoing civilian conflict in Ethiopia, there was limited accessibility to certain regions and/or certain woredas and health facilities that affected the sampling selection; it is unknown how this limitation affected the analysis. Fourth, prices of certain laboratory supplies, such as sheep’s blood for bacterial culture, could not be determined and were excluded from the analysis, resulting in an underestimation of the costs of supplies. Although, this value is estimated to be modest based on the nature of the resource inputs. Fifth, there could be a potential underestimation of the costs of capital assets (vehicles, buildings, and equipment) as certain dates of purchase for older items were unknown, in which case it was assumed that the purchase occurred in July 2012, the earliest date for which there were local consumer price indices available. Sixth, for resources reported as funded by EFMoH, respondents were not able to identify if those funds originated from domestic government revenues or from external donors. Such information on original source of funds may reframe the dependence of VPD surveillance on external donor resources, particularly from GPEI. Finally, there was no simultaneous collection of programmatic data (e.g. VPD incidence, annual number of cases) at subnational levels for contextualization of the quality and quantity of surveillance corresponding to the collected costs.

Conclusions

The costs of the Ethiopian VPD surveillance system are substantial and exceed previous country and regional estimates. Skilled human resources, important for maintaining quality of VPD surveillance, were major drivers of economic costs and require sustained workforce investment. Although the Ethiopian VPD surveillance system uses shared resources across multiple VPDs, resulting in scale efficiencies, and the Ethiopian government covers most of its economic costs, a considerable portion of those costs still relies on external donor support. Such estimates may aid Ethiopia in planning for donor funding transition and inform strategic investment decisions in surveillance system strengthening from domestic resources. Aligning country-level investments with WHO’s Global Strategy on Comprehensive VPD Surveillance will advance progress towards an efficient, high-functioning system that contributes towards global health security.

Supplementary Material

Supplementary data

Supplementary data is available at the Journal of Public Health online.

Figure 5. Heat Map of Percentage of Total Weighted Costs of Ethiopia’s VPD Surveillance System by VPD and Funding Source (July 2018-June 2019).

Figure 6. Heat Map of Total Weighted Costs of Ethiopia’s VPD Surveillance System by Resource Input and Support Function (March 2023 US$) (July 2018-June 2019).

Figure 7. Heat Map of Total Weighted Costs of Ethiopia’s VPD Surveillance System by Resource Input and Funding Source (March 2023 US$) (July 2018-June 2019).

Figure 8. Heat Map of Total Weighted Costs of Ethiopia’s VPD Surveillance System by Support Function and Funding Source (March 2023 US$) (July 2018-June 2019).

Table 3. Number of Sampled Cost Centers by Administrative Level.

Cost center level	Total number of cost centers by administrative level (2018–2019)	Number of cost centers sampled	Sampled cost centers as a percentage of total cost centers in administrative level
Federal	4	4	100%
FMoH - EPI	1	1	100%
EPHI: PHEM Office and National Laboratories	1	1	100%
WHO Office	1	1	100%
Core Group	1	1	100%
Regional	19	6	32%
Health Bureau	11	3	27%
Laboratory	3	2	67%
Core Group Office	5	1	20%
Zonal	109	9	8%
Zonal Health Departments	93	8	9%
Core Group Office	16	1	6%
Woreda	807	7	1%
Woreda Health Offices	727	6	1%
Core Group Office	80	1	1%
Health Facility	4,298	24	<1%
Hospital	313	10	3%
Health Center	3,985	14	<1%
Total	5,237	50	<1%

Abbreviations: FMoH: Federal Ministry of Health; EPI: Expanded Program in Immunization; EPHI: Ethiopian Public Health Institute; PHEM: Public Health Emergency Management; WHO; World Health Organization.

Table 4. Total Weighted Costs of Ethiopia’s VPD Surveillance System by Administrative Level of Cost Centers (March 2023 US$) (July 2018-June 2019).

Cost Center Level	Financial Costs		Economic Costs		N (number of sampled cost centers)
	Total	Share of Total (%)	Total	Share of Total (%)
Federal	7,681,796	28.6	8,815,815	12.8	4
Regional	2,096,590	7.8	3,506,666	5.1	6
Zonal	3,477,021	13.0	7,855,124	11.4	9
Woreda	4,067,751	15.2	8,910,999	12.9	7
Health Facility	9,509,653	35.4	40,017,506	57.9	24
Hospital	2,086,742	7.8	8,853,116	12.8	10
Health Center	7,422,911	27.7	31,164,390	45.1	4
Total at Country Level	26,832,811	100.0	69,105,309	100.0	50

Table 5. Comparison of Results from the Ethiopia and Nepal VPD Surveillance Costing Studies.

Country-wide category	Ethiopia (FY 2018–2019)	Nepal (FY 2016–2017)*
Total economic cost of VPD surveillance (in Mar 2023 US$)	69.1 million US$	7.1 million US$
	0.70 US$ per capita	0.26 US$ per capita
Funders	Ethiopian Government (79%)	Government of Nepal (56%)
	WHO (16%)	WHO (44%)
	USAID (2%)
	CDC (<1%)
	UNICEF (<0.1%)
	Personal/Self-paid, unknown, and others (3%)
Resource input with highest economic cost	Personnel (41%)	Personnel (51%)
Program activity with highest economic cost	Workforce capacity (25%)	Mission activities (active surveillance visits) (18%)
VPD with highest economic cost (excluding general VPD surveillance)	Measles (35%, excluding general)	Not reported in publication

^*Huang, X. X., Bose, A. S., Gupta, B. P., Rai, P., Joshi, S., Gautam, J. S., Tinkari, B. S., Vandelaer, J., Cohen, A. L., & Patel, M. K. (2021). Vaccine preventable diseases surveillance in Nepal: How much does it cost?. Vaccine, 39(40), 5982–5990.

Data availability

The data for this study were provided by the Ethiopian Public Health Institute by permission. Data will be shared upon request to the corresponding author with permission of the Ethiopian Public Health Institute.