- Email: [email protected]
Costs and financing of routine immunization: Approach and selected findings of a multi-country study (EPIC)
Vaccine 33S (2015) A13–A20
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Costs and financing of routine immunization: Approach and selected findings of a multi-country study (EPIC) Logan Brenzel a,∗ , Darwin Young b , Damian G. Walker a a b
Bill & Melinda Gates Foundation, Seattle, Washington, United States UNC Gillings School of Global Public Health, Chapel Hill, NC, United States
a r t i c l e
i n f o
Keywords: Costs Financing Routine immunization Productivity Low-income
a b s t r a c t Background: Few detailed facility-based costing studies of routine immunization (RI) programs have been conducted in recent years, with planners, managers and donors relying on older information or data from planning tools. To fill gaps and improve quality of information, a multi-country study on costing and financing of routine immunization and new vaccines (EPIC) was conducted in Benin, Ghana, Honduras, Moldova, Uganda and Zambia. Methods: This paper provides the rationale for the launch of the EPIC study, as well as outlines methods used in a Common Approach on facility sampling, data collection, cost and financial flow estimation for both the routine program and new vaccine introduction. Costing relied on an ingredients-based approach from a government perspective. Estimating incremental economic costs of new vaccine introduction in contexts with excess capacity are highlighted. The use of more disaggregated System of Health Accounts (SHA) coding to evaluate financial flows is presented. Results: The EPIC studies resulted in a sample of 319 primary health care facilities, with 65% of facilities in rural areas. The EPIC studies found wide variation in total and unit costs within each country, as well as between countries. Costs increased with level of scale and socio-economic status of the country. Governments are financing an increasing share of total RI financing. Conclusions: This study provides a wealth of high quality information on total and unit costs and financing for RI, and demonstrates the value of in-depth facility approaches. The paper discusses the lessons learned from using a standardized approach, as well as proposes further areas of methodology development. The paper discusses how results can be used for resource mobilization and allocation, improved efficiency of services at the country level, and to inform policies at the global level. Efforts at routinizing cost analysis to support sustainability efforts would be beneficial. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Vaccines are one of the best buys in public health [1,2] and a good candidate for government financing [3]. Historically, the cost per fully immunized child (FIC) with measles, polio, DTP, and BCG vaccines ranges from less than $3 to $22, with variation between countries by facility type and ownership, volume of vaccines administered, wastage rates, type of delivery strategy,
∗ Corresponding author at: Bill & Melinda Gates Foundation, 1300 I (Eye) Street, NW, Suite 200, Washington, DC 20005. E-mail addresses: [email protected] (L. Brenzel), [email protected] (D. Young), [email protected] (D.G. Walker). http://dx.doi.org/10.1016/j.vaccine.2014.12.066 0264-410X/© 2015 Elsevier Ltd. All rights reserved.
and input prices [4,5–11]1 . Unit costs generally decline as output increases, but there is scant evidence on whether unit costs will rise as coverage increases beyond 80% [1,12,13]. Currently, all GAVI-eligible countries have introduced pentavalent vaccine into their routine immunization (RI) programs2 . Rotavirus, pneumococcal, and HPV vaccine introductions also are on the rise worldwide [14]. Total and unit costs of RI have increased with the introduction of new vaccines [15–17] but many of these studies have relied on secondary estimates of costs and financing. Secondary sources have been shown to underestimate the cost of
1 DTP refers to the diphtheria–tetanus–pertussis vaccine and BCG is the Bacillus Calmet–Guerin vaccine that protects against tuberculosis. 2 The pentavalent vaccine is a combination vaccine preventing pertussis, tetatnus, diphtheria, hepatitis B, and haemophilus influenza type b.
A14
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Table 1 EPIC study country characteristics (2011). Country
DTP3 coverage level (WUENIC) (%)
Total doses (million)
GNI per capita
Infant population
Benin Ghana Uganda Zambia Moldova Honduras
85 91 80 81 93 91
3.7 9.5 11.9 4.8 0.68 3.8
$720 $1,420 $470 $1,180 $1,980 $2,030
348,577 1011,012 1326,826 567,320 47,537 177,733
Sources: WHO-UNICEF Best Estimates (who.immunization.org); EPIC study reports; World Development Indicators, Ministry of Health Reports.
national immunization programs [11] and reliability of secondary sources has been questioned [18]. Facility-based costing surveys have the potential to more fully capture the costs and financing of immunization through more detailed analysis of input use. Given the pace and breadth of new vaccine introduction and universal coverage goals highlighted in the Global Vaccine Action Plan [19], as well as the emphasis by the global community on Universal Health Coverage (UHC), there is a critical need for updated cost and financing estimates for RI and new vaccine introduction. Understanding the costs of existing coverage levels is a first step toward adequate planning and budgeting for GVAP and UHC goals. Potential cost variation at the facility level merits further exploration, as such analyses can identify areas of program efficiency and value for money.
Immunization Financing & Sustainability Task Team4 . In addition to building the evidence-base, this exercise strengthened capacity of researchers and fostered collaboration between government, researchers, and international experts, ultimately developing a community of practice. Table 1 compares the characteristics of the EPIC countries. Coverage of the third dose of DTP vaccine (used as a proxy in this study for fully vaccinated child) ranged from 80% in Uganda to 93% in Moldova [22]. Two of the countries are low-income (Benin and Uganda); two countries are categorized as intermediate countries by GAVI (Ghana and Zambia); and two countries will graduate from GAVI support in the next few years (Moldova and Honduras). The infant target population ranges from less than 200,000 (Honduras) to nearly 1.5 million (Uganda). 3. Methods and common approach
2. Objective and scope of the EPIC studies This paper describes the methods, selected findings and policy focus the EPI Costing and Financing (EPIC) studies conducted between 2012 and 2013 in Benin, Ghana, Honduras, Moldova, Uganda and Zambia, supported by the Bill & Melinda Gates Foundation (BMGF). The purpose of the EPIC studies was to provide countries and the global community with updated and detailed cost estimates of RI programs with pentavalent vaccine in the schedule, as well as the incremental costs of other new vaccines. The EPIC studies were conducted by the Pan American Health Organization (Honduras); Curatio Foundation International (Moldova); Health and Development Africa (Uganda and Zambia); and Agence pour la Medicine Preventive (Ghana and Benin). Each organization selected countries that had introduced pneumococcal and/or rotavirus vaccines in either 2011 or 2012 as part of their response to a Request for Proposals (RFP) issued by the BMGF. The RPF focused attention on determining facility-level total and unit costs, and reasons for their variation; financial flow analysis of the RI program; new vaccine introduction cost; and, comparisons with other economic information and planning estimates. All studies received ethical clearance from respective governments. The EPIC studies estimated RI costs from a government perspective. RI was defined as immunization services provided on a regular basis either in public or non-governmental organization (NGO) facilities, or through community outreach services. The study excluded the cost of supplementary immunization activities (SIAs), and private, for-profit facilities. Introduction costs of HPV vaccines also were excluded as these were being evaluated elsewhere [20,21]. The studies benefited from advice of Steering Committee members3 and inputs from partners through the GAVI Alliance
3 Steering Committee members included Carol Levin (University of Washington); Ulla Griffiths (London School of Hygiene and Tropical Medicine); Raymond
Because EPIC studies were carried out by different organizations, a Common Approach was developed to harmonize and standardize methods for cost and financing estimation [23]. 3.1. Costing of routine immunization (RI) The EPIC studies evaluate both the economic and financial costs of the routine immunization program5 . Costing utilized an ingredients-based approach to data collection on quantities of inputs used, unit prices, and proportion of time or use for immunization, similar to other immunization costing methods [25–28]. The study evaluated recurrent inputs by line items such as personnel, vaccines, syringes and other supplies, per diem, fuel and transport, cold chain maintenance, maintenance of other equipment, and operating costs; and capital inputs, such as cold chain, vehicles, computers, and other capital inputs. In addition, the study examined costs by activity classification: facility-based service delivery, outreach, record-keeping and HMIS, supervision, training, social mobilization, supply chain and logistics, program management, surveillance, and other. A matrix between line items and activities was developed to facilitate analysis (Fig. 1).
Hutubessy (WHO); Stephen Resch (Harvard School of Public Health); and Mike Hanlon (Institute for Health Metrics). 4 IF&S TT members participating providing feedback on the study included Santiago Cornejo and Marya Paytna (GAVI Secretariat); Claudio Politi (WHO/HQ); Gian Gandhi and Tom O’Connell (UNICEF); Mike McQuestion (Sabin Institute); Niyazi Cakmak (WHO EURO); Amos Petu (WHO ESA); Alexis Saytalou (WHO WCA); Claudia Castillo (PAHO). 5 Economic costs represent the opportunity cost associated with using inputs in the routine immunization program as compared to their next best use. Financial costs refer to financial outlays made to support routine immunization. An economic cost analysis estimates the annualized value of capital investments and the value of donated goods and labor time; whereas, a financial cost estimate is based on the financial outlay for capital equipment and excludes the value of donated goods and services.
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Line Item/ Activity
Routine Facilitybased Service Delivery
RecordKeeping & HMIS
Supervision
Outreach Service Delivery
Train -ing
Social Mobilization& Advocacy
Surveill -ance
A15
Cold Chain Mainten -ance
Vaccine Collection, Distribution Storage
Program Management
Other
Salaried Labor Volunteer Labor Per Diem- Travel Allowances Vaccines Vaccine Injection & Safety Supplies Other Supplies Transport/ Fuel Vehicle Maintenance Cold Chain Energy Costs Printing Utilities, Communication Other Recurrent Cold Chain Equipment Vehicles Lab Equipment Other Equipment Other Capital Buildings Fig. 1. Matrix of costing line items and activities for routine immunization. Source: Brenzel, L. Common Approach for the Costing and Financing Analyses of Routine Immunization and New Vaccine Introduction Costs (EPIC). Working Paper, Mimeo (2014).
Costs were based on a multi-stage, stratified random sample of public and NGO primary health facilities in each country to obtain a representative sample (Table 2). The sampling frame was based on a purposive selection of regions, followed by random selection of districts [29]. Within each district, facilities were randomly selected from a complete list of facilities, with oversampling of rural facilities to better understand their cost structure. Sample size was estimated based on a two-stage sampling technique with correction for proportions, and a level of error of 10%, leading to a target sample size of approximately 50 facilities per country. There was some variation in approach, with the Honduras study having a sample of 71 facilities, and the Benin study 46 facilities. The EPIC studies resulted in a total sample of 319 facilities, of which 65% were in rural areas. Data collection was accomplished using a standardized questionnaire that was adapted and pretested in each country [23]. Data were collected on the total quantities of inputs, unit prices,
Table 2 EPIC studies facility sample by country. Country Benin Ghana Honduras Moldova Uganda Zambia Total Percent (%)
Total Facilities
Rural
Urban
46 50 71 50 49 53
26 38 31 42
20 12 40 8
38
15
319
175 65
95 35
Source: EPIC country reports, 2014. Note:the Uganda sampling frame was done on the basis of facility type.
facility outputs, facility characteristics, useful life of equipment and vehicles, and other factors to enable allocation of shared inputs to routine immunization. Labor time for RI was assessed through a cascade of time-related questions: • How many days per week does this staff usually work at this health facility? • How many hours per week does this staff usually work at this health facility? • What portion of these hours is spent on all routine immunization related activities? • What proportion of time is spent on each routine immunization activity (as specified in Fig. 1)? Other approaches to estimate labor time and costs, including time and motion studies and weekly activity diaries exist [30], but these options were not pursued given time and resource constraints, and potential for other biases to be introduced. While there are challenges in eliciting time responses from personnel, study teams incorporated validation techniques during field surveys to minimize bias. Other shared inputs, such as transport and vehicles, were allocated to RI on the basis of the share of distance traveled by vehicles. Building space was allocated based on the share of facility area used for RI activities including storage of vaccines. Questionnaires were used to obtain output information and qualitative data on facility management. In addition, administrative, managerial, and program-related outputs (doses and children vaccinated) and inputs such labor time, per diem, transport, vaccine storage and distribution, supervision, and management were collected at district, regional and national levels using standardized, pre-tested questionnaires.
A16
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Fig. 2. Cost aggregation method for the EPIC studies. Source: Brenzel, L. Common Approach for the Costing and Financing Analyses of Routine Immunization and New Vaccine Introduction Costs (EPIC). Working Paper, Mimeo (2014).
Each study recruited and trained enumerators to administer questionnaires, and pre-testing was a form of hands-on training. Each study team included field supervisors who regularly verified survey responses. Emphasis was on maintaining high quality and reliability of information collected. An Excel-database tool developed for analysis of facility-level data the EPIC studies [24]6 . This tool mimicked the facility questionnaire to facilitate data entry and estimated total and unit costs (per dose, child, and FIC) in both local currency units and $US 20117 . The database tool included validation checks to ensure data were entered properly and that answers were within a plausible range. The database tool also allowed for export of summary data into statistical software for further analysis. All facility results were weighted based on sampling weights, and capital costs were annualized based on useful life and a 3% discount rate. While the EPIC studies emphasized evaluation of facility cost variation, there was a need to generate national total and unit costs for policy recommendations. The approach was based on use of weighted average facility and administrative level costs. Facility, district, and regional costs were added to total national vaccine and RI program costs to generate the total national RI cost (Fig. 2)8 . Other methods including statistical approaches were described in the Common Approach and will be explored in the subsequent phase.
the EPIC studies, expenditures for capital equipment to fill gaps in the routine program six months prior to new vaccine introduction and up to six months after the initial period of introduction were evaluated, unless otherwise designated by the EPI manager. Studies estimated a series of incremental costs of new vaccines from a governmental perspective (Table 3). Incremental economic costs that valued additional time and use of cold chain capacity, donated goods and services would be useful for evaluating cost-effectiveness of NUVI. Incremental financial cost estimation, relevant for planning and budgeting for NUVI, relied on a different approach to capital evaluation and excluded donated goods and service. Incremental fiscal costs measured financial outlays to benchmark costs against needed subsidies. Costs were disaggregated into one-time investments, and ongoing NUVI costs. Investment costs were defined as those specific to the introduction event, such as initial training and social mobilization events. These were annualized and discounted using estimates of useful life of training (2–3 years) at a 3% rate. The WHO Vaccine Volume Calculator was used to estimate additional cold storage requirements o [34]. Ongoing costs of NUVI were defined as the additional recurrent costs to the RI program, including labor costs and maintenance.
3.3. Financial flows for routine immunization 3.2. Costing of new vaccine introduction (NUVI) Costing new vaccine introduction drew from the WHO Guidelines for estimating incremental costs of introducing new vaccines into the national immunization system [31], supplemented by methods used in recent studies [20,21,32,33]. The time period for costing new vaccine introduction is important to distinguish. For
6 The Moldova team developed a separate Excel-based analytical tool, and the Honduras team used the CostVac Tool, which was further developed under the EPIC studies. 7 FIC refers to a fully immunized child or those who have received the third dose of the pentavalent vaccine. 8 An alternative approach using the results of the determinants analysis was used in some studies, but the results were found to be similar to the averaging approach.
The EPIC studies measured the total financial envelope available for immunization, tracing flows from all sources to their intended use. All sources of financing, their relative importance, and purpose of support were evaluated. Information was collected from national and sub-national government, donor agencies (incountry and headquarters), non-governmental organizations, and other sources. The approach did not evaluate expenditures as these would be somewhat duplicative of the financial cost estimates previously generated. In addition, evaluating financial flows allows the government and partners to see the total available resources, whether they were actually spent or not. Expenditures are dependent upon the effectiveness and efficiency of public expenditure management systems and donor processes and procedures, and would under-estimate the total resources available for a program.
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
A17
Table 3 Cost estimation for new vaccine introduction for the EPIC studies. Line item
Economic costs
Financial costs
Fiscal costs
Salaried labor
Included to represent opportunity cost of time of existing staff involved in NUVI (FT and % of time) Economic Value Included Included Economic value of vaccines utilized Economic value of utilized supplies Economic value included
Labor costs of new staff hired to accommodate NUVI
Included if new staff needs to be hired
Excluded Included Financial costs of purchased vaccines Financial cost of purchased supplies Included
Excluded Included Financial cost of purchased vaccines Financial cost of supplies
Volunteer labor Per diems Vaccines Vaccine injection supplies Transportation and other transport costs Cold storage costs
Economic value of current cold chain volume – and/or – economic value of additional cold storage equipment and supplies purchased for NUVI (discounted annualized share of total cost) Economic value of vehicle use for NUVI activities
Vehicles
Financial costs of additional cold storage equipment and supplies purchase for NUVI using straight-line depreciation
Purchase of vehicles for NUVI
Financial cost of fuel and other transportation Payment of additional cold chain purchase for NUVI
Purchase of vehicles for NUVI
Source: Brenzel, L. Common approach for the costing and financing analyses of routine immunization and new vaccine introduction costs (EPIC). Working Paper, Mimeo (2014).
Table 4 Sample Coding of Health Care Functions for RI and NUVI for the EPIC Studies. Classification of health care functions HC (functions) Code
Sub-code
HC.1 HC.6 HC.6.1 HC.6.1.1 HC.6.2 HC.6.2.1 HC.6.2.2 HC.6.2.3 HC.6.2.4 HC.6.2.5 HC.6.2.6 HC.6.2.6 HC.6.2.7 HC.6.5 HC.6.5.1 HC.6.5.2 HC.7 HC.99
Description Curative care Preventive care Information, education and counseling programs Social mobilization, advocacy Immunization programs Facility-based routine immunization service delivery Outreach routine immunization service delivery Training Vaccine collection, storage and distribution Cold chain maintenance Supervision Program management Other routine immunization program activity Surveillance EPI Surveillance Record-keeping and HMIS Governance and health system financing and administration Not disaggregated
Source: Brenzel, L. Common approach for the costing and financing analyses of routine immunization and new vaccine introduction costs (EPIC). Working paper, Mimeo (2014).
Complementary facility costing information at was paired with analysis of funding flows as needed. The System of Health Accounts coding [35,36] was used to analyze financial flows in a harmonized manner across the six countries. These codes were disaggregated further to be consistent with the matrix of line items and activities developed for the Common Approach (Table 4)9 . Data collection was facilitated through
9 The SHA codes are organized around the following: (1) Revenues (Financing Sources-FS): classifies funding sources at country level; (2) Health Care Financing (HF): classifies intermediary financing agents; Health Providers (HP): classifies
standard pre-tested data collection formats, and analysis was conducted using an Excel-based tool. Countries were encouraged to collect and analyze data for more than one year to evaluate trends. All findings were converted to $2011. Each team produced a mapping of financial flows for RI. 3.4. Health facility RI productivity performance The EPIC studies focused on a comparison of facilities through a graphical analysis where facilities were plotted into quadrants according to high/low unit cost per dose based on median values (X-axis) and high/low doses administered (Y-axis). These plots revealed a spread of facility types across the performance spectrum. These plots could be used to identify those facilities with the same level of output but widely different costs. Facilities that can provide the same level of output for less cost could be interpreted as being more efficient. Additional productivity indicators such as doses administered per full-time equivalent; wastage rates for different vaccines; and, drop-out rates, were compared across facilities. Statistical analysis also sought to uncover the possible reasons for variation in performance levels to provide information to managers on how to improve performance and efficiency of service delivery. 4. Selected findings of the EPIC studies We report some of the findings from the EPIC studies across countries to illustrate trends. Additional detail and analysis can be found in other manuscripts in this Supplement. First, there is a wide range of total facility immunization economic costs within the country samples, with nearly 200-fold differences in Moldova and 60-fold differences in Uganda. Benin had the smallest range between high and low cost facilities. Second, differences in total facility cost between countries are related to variation in wage rates and labor time spent on immunization. Fig. 3 illustrates the main cost drivers of facility costs for the EPIC studies and shows that the proportion of labor costs appears to
funding by type of facility; Health Care Functions (HC): classifies RI activities; Health Care Provision (FP): classifies RI line item.
A18
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Fig. 3. Cost drivers at facility level for the EPIC studies. Source: author’s calculations.
be related to level of economic development. In Moldova, 65% of total facility cost was for labor. Ghana has a relatively high labor share because of the role of volunteer labor (5% of total cost) and higher wage rates. All of the countries had a similar immunization schedule (pentavalent, oral polio, measles and BCG vaccines). Third, average unit costs of RI at the facility level are higher than previously estimated. Table 5 shows the cost per dose ranges from $2 in Benin to $18 in Moldova, and the cost per infant varies from $19 to over $300. The cost per FIC ranges from $25 to $332, with an average of $46.50 in low income countries (Benin and Uganda), and $58.50 in socio-economically intermediate countries (Ghana and Zambia). The average cost/FIC for countries in Africa is $46.50. In Moldova, non-labor cost per dose is $6 compared to $18 with labor, and the non-labor cost per infant is $112 compared to $317. More detailed examination of cost variation will be undertaken in the next phase of EPIC. The majority of total national RI costs was for service delivery with an average of 15% for administration and management costs above the facility-level. Moldova, which had the highest immunization coverage, had an above-facility share of 18%. Further
investigation as to the relationship between management and immunization performance would be useful. Finally, governments in the EPIC study sample are financing approximately 70% of the RI program on average, ranging from 45% in Benin and Uganda to 95% in Moldova. Total available financing ranged from approximately $9 million to $50 million in 2011.
5. Policy conclusions and lessons learned The EPIC studies provide a unique set of high quality data on RI program costs, updating the evidence-base to include pentavalent vaccines in the immunization schedule. The studies reveal that governments are financing a larger total amount and share of RI than previously estimated [17]. The major cost drivers continue to be labor, followed by vaccines. The EPIC studies also showed a wide range in facility-level total costs within countries, as well as variation in unit costs between countries. The extent of this variation may mean it is no longer possible to refer to a single point estimate of unit costs. Further analysis of factors contributing to cost variation can be useful for
Table 5 Facility-based routine immunization economic costs and national-level financing ($2011). Country
Benin Uganda Ghana Zambia Honduras Moldova
Facility Weighted Average
Share of above facility costs (%)b
Total available financing (m)c
Share of government financing (%)
$25
6
$10.12
45
$61
$44
16
$34.3
45
$36
$78
$51
12
$50
85
$7
$60
$123
$66
26
$38.9
82
$8
$113
$217
$128
14
$49.1
64
$18
$317
$672
$332
18
$8.8
95
Routine immunization economic cost
Cost/dose
Cost/infant
Cost/infant ($PPP)
Cost/FICa
$14,994 ($6,026–$50,472) $23,470 ($1,911–$112,753) $16,460 ($4,385–$99,789) $28,286 ($6,260–$64,019) $12,069 ($2,165–$221,235) $11,943 ($565–$112,548)
$2
$19
$42
$5
$20
$5
Source: authors’ calculations rounded to nearest whole number. a FIC = fully immunized child represented by number of children receiving the third dose of DTP-containing vaccine. b Reflects administrative and program management costs at higher than facility levels. c Figures reflect total resources potentially available and not actual expenditures made.
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
identifying ways in which performance of the RI program can be improved [38]. In all countries, the value of labor time represented an important share of total costs, ranging from 19% in Benin to 65% in Moldova. Labor time is a shared cost and data collection required enumerators to probe to determine both the allocation to immunization, and within immunization, the allocation to various activities. Because of the exclusive focus on immunization, time allocations may be overestimated relative to other services. Future studies could explore alternative methods, such as diaries or observation, or estimate time allocations (and costs) across the full range of services provided. While most of the study teams followed the methods described in the Common Approach, there was some variation due to Ministry of Health requests. The sample size of 50 facilities was intended to be as representative as possible, with a reasonable level of power. Facility surveys are costly to conduct, and analysis can be time consuming. Achieving the right balance between feasibility and generalizability is important. Further work regarding facility sampling for costing studies and approaches for routine cost data collection and analysis are needed. The policy implications of having accurate cost and financing information are several. For instance, these data can be used to improve planning of resource requirements and financing needs at the country level. Understanding the incremental delivery costs associated with new vaccine introduction will be important for updating donor policies related to subsidizing new vaccine introduction, and for domestic and external resource mobilization for routine programs. The EPIC studies have provided much-needed information on the costs and financing of RI. The study design aimed to generate a representative sample for each country, though the results can only be interpreted as indicative for other settings. Further investigations are warranted, as there were no studies from Eastern Mediterranean, Southeast or East Asian countries in the first wave. Further analysis could be undertaken with the existing data sets both within country and across countries, particularly with respect to evaluation of productivity and determinants of costs. Finally, there is scope to refine the methods and make them useful for regular data collection and analysis by the national immunization program.
Conflict of interest statement The authors declare no conflict of interest. The EPIC studies and preparation of this manuscript were supported by the Bill & Melinda Gates Foundation. However, the views expressed represent those of the authors only.
Acknowledgements We would like to thank all of the members of the study teams who provided inputs into the development of the methods and approaches used for the EPIC studies. The contributions of Steering Committee members, most notably Carol Levin and Ulla Griffiths, were critical to the successful outcome of the EPIC studies. Finally, we appreciate the comments of two independent reviewers.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine. 2014.12.066.
A19
References [1] Brenzel L, Halsey N, Miller M, Wolfson L, Fox-Rushby J. Vaccine Preventable Diseases. In: Jamison D, Bremen J, Measham A, Alleyne G, Claeson M, Evans D, Jha P, Mills A, Musgrove P, editors. Disease control priorities in developing countries. Oxford: Oxford University Press; 2006. [2] WHO, UNICEF. World Bank State of the world’s vaccines and immunization. 3rd ed. Geneva: World Health Organization; 2009. [3] Musgrove P. Public spending on health care: how are different criteria related. Health Policy 1999;47:207–23. [4] Creese AL, Sriyabbaya N, Casabal G, Wiseso G. Bull World Health Organ 1982;60(4):621–32. [5] Brenzel L, Claquin P. Immunization programs and their costs. Soc Sci Med 1994;39(4):527–36. [6] Kaddar Miloud, Levin Ann, Dougherty Leanne, Maceira Daniel. Costs and financing of immunization programs: findings of four case studies. In: Special Initiatives Report 26. Bethesda, MD: Partnerships for Health Reform Project, Abt Associates Inc.; 2000. [7] Phonboon K, Shepard DS, Ramaboot S, Kunasol P, Preuksaraj S. The Thai expanded programme on immunization: role of immunization sessions and their cost-effectiveness. Bull World Health Organ 1989;67(2):181–8. [8] Robertson RL, Davis JH, Jobe K. Service volume and other factors affecting the costs of immunizations in the Gambia. Bull World Health Organ 1984;62(5):729–36. [9] Shepard DS, Robertson RL, Cameron CS, Saturno P, Pollack M, Manceau J. Costeffectiveness of routine and campaign vaccination strategies in Ecuador. Bull World Health Org: 1989;67:649–62. [10] Mosquiera D, Walker NR, Penny ME, Clark CF, Lanata AD, Sanderson CFB, et al. Variation in the costs of delivery routine immunization services in Peru. Bull World Health Organ 2004;82(9):676–82. [11] Creese AL, Domínguez-Uga MA. Cost-effectiveness of immunization programs in Colombia. Bull Pan Am Health Organ 1987;21(4):377–94. [12] Bishai D, McQuestion M, Chaudhry R, Wigton A. The cost of scaling up vaccination services in the world’s poorest countries. Health Aff 2006;25(2):348–56. [13] Gavi, The Vaccine Alliance. Geneva, Switzerland. www.gavi.org (last accessed August 2014). [14] Wolfson LJ, Gasse F, Lee-Martin SP, Lydon P, Magan A, Tibouti A, et al. Estimating the costs of achieving the WHO-UNICEF Global Immunization Vision and Strategy, 2006–2015. Bull World Health Organ 2008;86:27–39. [15] Lydon P, Levine R, Makinen M, Brenzel L, Michell V, Milstein J, et al. Introducing vaccines in the poorest countries: what did we learn from the GAVI experience with financial sustainability? Vaccine 2008;26:6706–16. [16] L. Brenzel, C. Politi. Historical Analysis of the Comprehensive Multi-Year Plans in GAVI-Eligible countries (2004–2015). World Health Organization. http://apps.who.int/immunization financing/analysis/Historical cMYP Analysis 2012.pdf. [17] Khan M, Khan S, Walker D, Fox-Rushby J, Cutts F, Akramuzzaman S. Cost of delivering child immunization services in urban Bangladesh: a study based on facility-level surveys. J Health Popul Nutr 2004;22(4):404–12. [18] Nader A, de Quadros C, Politi C, McQuestion M. An analysis of government immunization program expenditures in lower and lower middle income countries (2006–12). Health Policy Plan 2014, http://dx.doi.org/10.1093/ heapol/czu002. [19] World Health Organization. Global vaccine action plan. http://www.who.int/ immunization/global vaccine action plan/en/. [20] Levin Carol E, Van Minh Hoang, Odaga John, Rout Swampa Sarit, Ngoc Diep Nguyen Thi, Menezes Lysander, et al. Delivery cost of human papillomavirus vaccination of young adolescent girls in Peru, Uganda and Viet Nam. Bull World Health Organ 2013;91(8):585–92. [21] Raymond Hutubessy, Levin Ann, Wang Susan, Morgan Winthrop, Ally Mariam, Broutet Nathalie, et al. A case study using the United Republic of Tanzania: costing nationwide HPV vaccine delivery using the WHO Cervical Cancer Prevention and Control Costing Tool. BMC Med 2012;136, http://dx.doi.org/10.1186/1741-7015-10-136. [22] World Health Organization. Immunization vaccines & biologicals. http://www. who.int/mediacentre/factsheets/fs378/en/ (last accessed on August 12, 2014). [23] Brenzel L. Common approach to the costing and financing analyses for routine immunization and New Vaccine Introduction (NUVI). Mimeograph. Bill & Melinda Gates Foundation; 2014 (weblink). [24] Young D. Guidelines and database tool for estimating the costs and financing of routine immunization. Mimeograph. Bill & Melinda Gates Foundation; 2013 (weblink). [25] World Health Organization. WHO guide for standardization of economic evaluations of immunization programmes. Geneva, Switzerland: World Health Organization; 2008. [26] USAID. Asia/near east bureau guidance for costing of health service delivery projects, background and guidance, REACH, project. Arlington, VA: USAID; 1989. [27] World Health Organization. Immunization costing & financing: a tool and user guide for comprehensive multi-year planning (cMYP). Geneva, Switzerland: World Health Organization; 2008. [28] Brenzel L. Selecting an essential package of health care services using costeffectiveness analysis: a manual for professionals in developing countries. In: Data for decision making project. Department of Population Sciences and International Health, Harvard School of Public Health; 1996.
A20
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
[29] Measure Evaluation. Sampling manual for health facility surveys. Washington, D.C.: Measure Evaluation; 2001. http://gametlibrary.worldbank.org/ FILES/665 Sampling%20Manual%20for%20Health%20Facility%20Surveys.pdf. [30] Bratt JH, Foreit J, Chen PL, West C, Janowitz B, de Vargas T. A comparison of four approaches for measuring clinician time use. Health Policy Plann 1999;14(4):374–81. [31] World Health Organization. Guidelines for estimating costs of introducing new vaccines into the national immunization system. Geneva, Switzerland: World Health Organization; 2002. [32] Griffiths U, Korczak VS, Ayalew D, Yigzaw A. Incremental system costs of introducing combined DTwP–hepatitis B–Hib vaccine into national immunization services in Ethiopia. Vaccine 2009;27:1426–32. [33] Clinton Health Access Initiative (CHAI). Facility-based unit costing for anteretroviral treatment in five sub-Saharan African Countries. Mimeograph. Clinton Health Access Initiative (CHAI); May 2011.
[34] World Health Organization Vaccine volume calculator spreadsheet. World Health Organization. Geneva, Switzerland. http://www.who.int/ immunization delivery/systems policy/logistics/en/index4.html. [35] Organization of Economic Cooperation and Development (OECD). 2011 System of health accounts (SHA). Paris, France. [36] Diener A, Morgan D. Deriving expenditures by disease under the System of Health Accounts (SHA) 2011 framework: a systematic analysis. Lancet 2013;381:S41, http://dx.doi.org/10.1016/S0140-6736(13)61295-7. [38] LaFond A, Kanagat N, Steinglass R, Fields R, Sequeira J, Mookherji S. Drivers of routine immunization coverage improvement in Africa: findings from district-level case studies. Health Policy Plann 2014 2014, http://dx.doi.org/10.1093/heapol/czu011.
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
Costs and financing of routine immunization: Approach and selected findings of a multi-country study (EPIC) Logan Brenzel a,∗ , Darwin Young b , Damian G. Walker a a b
Bill & Melinda Gates Foundation, Seattle, Washington, United States UNC Gillings School of Global Public Health, Chapel Hill, NC, United States
a r t i c l e
i n f o
Keywords: Costs Financing Routine immunization Productivity Low-income
a b s t r a c t Background: Few detailed facility-based costing studies of routine immunization (RI) programs have been conducted in recent years, with planners, managers and donors relying on older information or data from planning tools. To fill gaps and improve quality of information, a multi-country study on costing and financing of routine immunization and new vaccines (EPIC) was conducted in Benin, Ghana, Honduras, Moldova, Uganda and Zambia. Methods: This paper provides the rationale for the launch of the EPIC study, as well as outlines methods used in a Common Approach on facility sampling, data collection, cost and financial flow estimation for both the routine program and new vaccine introduction. Costing relied on an ingredients-based approach from a government perspective. Estimating incremental economic costs of new vaccine introduction in contexts with excess capacity are highlighted. The use of more disaggregated System of Health Accounts (SHA) coding to evaluate financial flows is presented. Results: The EPIC studies resulted in a sample of 319 primary health care facilities, with 65% of facilities in rural areas. The EPIC studies found wide variation in total and unit costs within each country, as well as between countries. Costs increased with level of scale and socio-economic status of the country. Governments are financing an increasing share of total RI financing. Conclusions: This study provides a wealth of high quality information on total and unit costs and financing for RI, and demonstrates the value of in-depth facility approaches. The paper discusses the lessons learned from using a standardized approach, as well as proposes further areas of methodology development. The paper discusses how results can be used for resource mobilization and allocation, improved efficiency of services at the country level, and to inform policies at the global level. Efforts at routinizing cost analysis to support sustainability efforts would be beneficial. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Vaccines are one of the best buys in public health [1,2] and a good candidate for government financing [3]. Historically, the cost per fully immunized child (FIC) with measles, polio, DTP, and BCG vaccines ranges from less than $3 to $22, with variation between countries by facility type and ownership, volume of vaccines administered, wastage rates, type of delivery strategy,
∗ Corresponding author at: Bill & Melinda Gates Foundation, 1300 I (Eye) Street, NW, Suite 200, Washington, DC 20005. E-mail addresses: [email protected] (L. Brenzel), [email protected] (D. Young), [email protected] (D.G. Walker). http://dx.doi.org/10.1016/j.vaccine.2014.12.066 0264-410X/© 2015 Elsevier Ltd. All rights reserved.
and input prices [4,5–11]1 . Unit costs generally decline as output increases, but there is scant evidence on whether unit costs will rise as coverage increases beyond 80% [1,12,13]. Currently, all GAVI-eligible countries have introduced pentavalent vaccine into their routine immunization (RI) programs2 . Rotavirus, pneumococcal, and HPV vaccine introductions also are on the rise worldwide [14]. Total and unit costs of RI have increased with the introduction of new vaccines [15–17] but many of these studies have relied on secondary estimates of costs and financing. Secondary sources have been shown to underestimate the cost of
1 DTP refers to the diphtheria–tetanus–pertussis vaccine and BCG is the Bacillus Calmet–Guerin vaccine that protects against tuberculosis. 2 The pentavalent vaccine is a combination vaccine preventing pertussis, tetatnus, diphtheria, hepatitis B, and haemophilus influenza type b.
A14
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Table 1 EPIC study country characteristics (2011). Country
DTP3 coverage level (WUENIC) (%)
Total doses (million)
GNI per capita
Infant population
Benin Ghana Uganda Zambia Moldova Honduras
85 91 80 81 93 91
3.7 9.5 11.9 4.8 0.68 3.8
$720 $1,420 $470 $1,180 $1,980 $2,030
348,577 1011,012 1326,826 567,320 47,537 177,733
Sources: WHO-UNICEF Best Estimates (who.immunization.org); EPIC study reports; World Development Indicators, Ministry of Health Reports.
national immunization programs [11] and reliability of secondary sources has been questioned [18]. Facility-based costing surveys have the potential to more fully capture the costs and financing of immunization through more detailed analysis of input use. Given the pace and breadth of new vaccine introduction and universal coverage goals highlighted in the Global Vaccine Action Plan [19], as well as the emphasis by the global community on Universal Health Coverage (UHC), there is a critical need for updated cost and financing estimates for RI and new vaccine introduction. Understanding the costs of existing coverage levels is a first step toward adequate planning and budgeting for GVAP and UHC goals. Potential cost variation at the facility level merits further exploration, as such analyses can identify areas of program efficiency and value for money.
Immunization Financing & Sustainability Task Team4 . In addition to building the evidence-base, this exercise strengthened capacity of researchers and fostered collaboration between government, researchers, and international experts, ultimately developing a community of practice. Table 1 compares the characteristics of the EPIC countries. Coverage of the third dose of DTP vaccine (used as a proxy in this study for fully vaccinated child) ranged from 80% in Uganda to 93% in Moldova [22]. Two of the countries are low-income (Benin and Uganda); two countries are categorized as intermediate countries by GAVI (Ghana and Zambia); and two countries will graduate from GAVI support in the next few years (Moldova and Honduras). The infant target population ranges from less than 200,000 (Honduras) to nearly 1.5 million (Uganda). 3. Methods and common approach
2. Objective and scope of the EPIC studies This paper describes the methods, selected findings and policy focus the EPI Costing and Financing (EPIC) studies conducted between 2012 and 2013 in Benin, Ghana, Honduras, Moldova, Uganda and Zambia, supported by the Bill & Melinda Gates Foundation (BMGF). The purpose of the EPIC studies was to provide countries and the global community with updated and detailed cost estimates of RI programs with pentavalent vaccine in the schedule, as well as the incremental costs of other new vaccines. The EPIC studies were conducted by the Pan American Health Organization (Honduras); Curatio Foundation International (Moldova); Health and Development Africa (Uganda and Zambia); and Agence pour la Medicine Preventive (Ghana and Benin). Each organization selected countries that had introduced pneumococcal and/or rotavirus vaccines in either 2011 or 2012 as part of their response to a Request for Proposals (RFP) issued by the BMGF. The RPF focused attention on determining facility-level total and unit costs, and reasons for their variation; financial flow analysis of the RI program; new vaccine introduction cost; and, comparisons with other economic information and planning estimates. All studies received ethical clearance from respective governments. The EPIC studies estimated RI costs from a government perspective. RI was defined as immunization services provided on a regular basis either in public or non-governmental organization (NGO) facilities, or through community outreach services. The study excluded the cost of supplementary immunization activities (SIAs), and private, for-profit facilities. Introduction costs of HPV vaccines also were excluded as these were being evaluated elsewhere [20,21]. The studies benefited from advice of Steering Committee members3 and inputs from partners through the GAVI Alliance
3 Steering Committee members included Carol Levin (University of Washington); Ulla Griffiths (London School of Hygiene and Tropical Medicine); Raymond
Because EPIC studies were carried out by different organizations, a Common Approach was developed to harmonize and standardize methods for cost and financing estimation [23]. 3.1. Costing of routine immunization (RI) The EPIC studies evaluate both the economic and financial costs of the routine immunization program5 . Costing utilized an ingredients-based approach to data collection on quantities of inputs used, unit prices, and proportion of time or use for immunization, similar to other immunization costing methods [25–28]. The study evaluated recurrent inputs by line items such as personnel, vaccines, syringes and other supplies, per diem, fuel and transport, cold chain maintenance, maintenance of other equipment, and operating costs; and capital inputs, such as cold chain, vehicles, computers, and other capital inputs. In addition, the study examined costs by activity classification: facility-based service delivery, outreach, record-keeping and HMIS, supervision, training, social mobilization, supply chain and logistics, program management, surveillance, and other. A matrix between line items and activities was developed to facilitate analysis (Fig. 1).
Hutubessy (WHO); Stephen Resch (Harvard School of Public Health); and Mike Hanlon (Institute for Health Metrics). 4 IF&S TT members participating providing feedback on the study included Santiago Cornejo and Marya Paytna (GAVI Secretariat); Claudio Politi (WHO/HQ); Gian Gandhi and Tom O’Connell (UNICEF); Mike McQuestion (Sabin Institute); Niyazi Cakmak (WHO EURO); Amos Petu (WHO ESA); Alexis Saytalou (WHO WCA); Claudia Castillo (PAHO). 5 Economic costs represent the opportunity cost associated with using inputs in the routine immunization program as compared to their next best use. Financial costs refer to financial outlays made to support routine immunization. An economic cost analysis estimates the annualized value of capital investments and the value of donated goods and labor time; whereas, a financial cost estimate is based on the financial outlay for capital equipment and excludes the value of donated goods and services.
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Line Item/ Activity
Routine Facilitybased Service Delivery
RecordKeeping & HMIS
Supervision
Outreach Service Delivery
Train -ing
Social Mobilization& Advocacy
Surveill -ance
A15
Cold Chain Mainten -ance
Vaccine Collection, Distribution Storage
Program Management
Other
Salaried Labor Volunteer Labor Per Diem- Travel Allowances Vaccines Vaccine Injection & Safety Supplies Other Supplies Transport/ Fuel Vehicle Maintenance Cold Chain Energy Costs Printing Utilities, Communication Other Recurrent Cold Chain Equipment Vehicles Lab Equipment Other Equipment Other Capital Buildings Fig. 1. Matrix of costing line items and activities for routine immunization. Source: Brenzel, L. Common Approach for the Costing and Financing Analyses of Routine Immunization and New Vaccine Introduction Costs (EPIC). Working Paper, Mimeo (2014).
Costs were based on a multi-stage, stratified random sample of public and NGO primary health facilities in each country to obtain a representative sample (Table 2). The sampling frame was based on a purposive selection of regions, followed by random selection of districts [29]. Within each district, facilities were randomly selected from a complete list of facilities, with oversampling of rural facilities to better understand their cost structure. Sample size was estimated based on a two-stage sampling technique with correction for proportions, and a level of error of 10%, leading to a target sample size of approximately 50 facilities per country. There was some variation in approach, with the Honduras study having a sample of 71 facilities, and the Benin study 46 facilities. The EPIC studies resulted in a total sample of 319 facilities, of which 65% were in rural areas. Data collection was accomplished using a standardized questionnaire that was adapted and pretested in each country [23]. Data were collected on the total quantities of inputs, unit prices,
Table 2 EPIC studies facility sample by country. Country Benin Ghana Honduras Moldova Uganda Zambia Total Percent (%)
Total Facilities
Rural
Urban
46 50 71 50 49 53
26 38 31 42
20 12 40 8
38
15
319
175 65
95 35
Source: EPIC country reports, 2014. Note:the Uganda sampling frame was done on the basis of facility type.
facility outputs, facility characteristics, useful life of equipment and vehicles, and other factors to enable allocation of shared inputs to routine immunization. Labor time for RI was assessed through a cascade of time-related questions: • How many days per week does this staff usually work at this health facility? • How many hours per week does this staff usually work at this health facility? • What portion of these hours is spent on all routine immunization related activities? • What proportion of time is spent on each routine immunization activity (as specified in Fig. 1)? Other approaches to estimate labor time and costs, including time and motion studies and weekly activity diaries exist [30], but these options were not pursued given time and resource constraints, and potential for other biases to be introduced. While there are challenges in eliciting time responses from personnel, study teams incorporated validation techniques during field surveys to minimize bias. Other shared inputs, such as transport and vehicles, were allocated to RI on the basis of the share of distance traveled by vehicles. Building space was allocated based on the share of facility area used for RI activities including storage of vaccines. Questionnaires were used to obtain output information and qualitative data on facility management. In addition, administrative, managerial, and program-related outputs (doses and children vaccinated) and inputs such labor time, per diem, transport, vaccine storage and distribution, supervision, and management were collected at district, regional and national levels using standardized, pre-tested questionnaires.
A16
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Fig. 2. Cost aggregation method for the EPIC studies. Source: Brenzel, L. Common Approach for the Costing and Financing Analyses of Routine Immunization and New Vaccine Introduction Costs (EPIC). Working Paper, Mimeo (2014).
Each study recruited and trained enumerators to administer questionnaires, and pre-testing was a form of hands-on training. Each study team included field supervisors who regularly verified survey responses. Emphasis was on maintaining high quality and reliability of information collected. An Excel-database tool developed for analysis of facility-level data the EPIC studies [24]6 . This tool mimicked the facility questionnaire to facilitate data entry and estimated total and unit costs (per dose, child, and FIC) in both local currency units and $US 20117 . The database tool included validation checks to ensure data were entered properly and that answers were within a plausible range. The database tool also allowed for export of summary data into statistical software for further analysis. All facility results were weighted based on sampling weights, and capital costs were annualized based on useful life and a 3% discount rate. While the EPIC studies emphasized evaluation of facility cost variation, there was a need to generate national total and unit costs for policy recommendations. The approach was based on use of weighted average facility and administrative level costs. Facility, district, and regional costs were added to total national vaccine and RI program costs to generate the total national RI cost (Fig. 2)8 . Other methods including statistical approaches were described in the Common Approach and will be explored in the subsequent phase.
the EPIC studies, expenditures for capital equipment to fill gaps in the routine program six months prior to new vaccine introduction and up to six months after the initial period of introduction were evaluated, unless otherwise designated by the EPI manager. Studies estimated a series of incremental costs of new vaccines from a governmental perspective (Table 3). Incremental economic costs that valued additional time and use of cold chain capacity, donated goods and services would be useful for evaluating cost-effectiveness of NUVI. Incremental financial cost estimation, relevant for planning and budgeting for NUVI, relied on a different approach to capital evaluation and excluded donated goods and service. Incremental fiscal costs measured financial outlays to benchmark costs against needed subsidies. Costs were disaggregated into one-time investments, and ongoing NUVI costs. Investment costs were defined as those specific to the introduction event, such as initial training and social mobilization events. These were annualized and discounted using estimates of useful life of training (2–3 years) at a 3% rate. The WHO Vaccine Volume Calculator was used to estimate additional cold storage requirements o [34]. Ongoing costs of NUVI were defined as the additional recurrent costs to the RI program, including labor costs and maintenance.
3.3. Financial flows for routine immunization 3.2. Costing of new vaccine introduction (NUVI) Costing new vaccine introduction drew from the WHO Guidelines for estimating incremental costs of introducing new vaccines into the national immunization system [31], supplemented by methods used in recent studies [20,21,32,33]. The time period for costing new vaccine introduction is important to distinguish. For
6 The Moldova team developed a separate Excel-based analytical tool, and the Honduras team used the CostVac Tool, which was further developed under the EPIC studies. 7 FIC refers to a fully immunized child or those who have received the third dose of the pentavalent vaccine. 8 An alternative approach using the results of the determinants analysis was used in some studies, but the results were found to be similar to the averaging approach.
The EPIC studies measured the total financial envelope available for immunization, tracing flows from all sources to their intended use. All sources of financing, their relative importance, and purpose of support were evaluated. Information was collected from national and sub-national government, donor agencies (incountry and headquarters), non-governmental organizations, and other sources. The approach did not evaluate expenditures as these would be somewhat duplicative of the financial cost estimates previously generated. In addition, evaluating financial flows allows the government and partners to see the total available resources, whether they were actually spent or not. Expenditures are dependent upon the effectiveness and efficiency of public expenditure management systems and donor processes and procedures, and would under-estimate the total resources available for a program.
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
A17
Table 3 Cost estimation for new vaccine introduction for the EPIC studies. Line item
Economic costs
Financial costs
Fiscal costs
Salaried labor
Included to represent opportunity cost of time of existing staff involved in NUVI (FT and % of time) Economic Value Included Included Economic value of vaccines utilized Economic value of utilized supplies Economic value included
Labor costs of new staff hired to accommodate NUVI
Included if new staff needs to be hired
Excluded Included Financial costs of purchased vaccines Financial cost of purchased supplies Included
Excluded Included Financial cost of purchased vaccines Financial cost of supplies
Volunteer labor Per diems Vaccines Vaccine injection supplies Transportation and other transport costs Cold storage costs
Economic value of current cold chain volume – and/or – economic value of additional cold storage equipment and supplies purchased for NUVI (discounted annualized share of total cost) Economic value of vehicle use for NUVI activities
Vehicles
Financial costs of additional cold storage equipment and supplies purchase for NUVI using straight-line depreciation
Purchase of vehicles for NUVI
Financial cost of fuel and other transportation Payment of additional cold chain purchase for NUVI
Purchase of vehicles for NUVI
Source: Brenzel, L. Common approach for the costing and financing analyses of routine immunization and new vaccine introduction costs (EPIC). Working Paper, Mimeo (2014).
Table 4 Sample Coding of Health Care Functions for RI and NUVI for the EPIC Studies. Classification of health care functions HC (functions) Code
Sub-code
HC.1 HC.6 HC.6.1 HC.6.1.1 HC.6.2 HC.6.2.1 HC.6.2.2 HC.6.2.3 HC.6.2.4 HC.6.2.5 HC.6.2.6 HC.6.2.6 HC.6.2.7 HC.6.5 HC.6.5.1 HC.6.5.2 HC.7 HC.99
Description Curative care Preventive care Information, education and counseling programs Social mobilization, advocacy Immunization programs Facility-based routine immunization service delivery Outreach routine immunization service delivery Training Vaccine collection, storage and distribution Cold chain maintenance Supervision Program management Other routine immunization program activity Surveillance EPI Surveillance Record-keeping and HMIS Governance and health system financing and administration Not disaggregated
Source: Brenzel, L. Common approach for the costing and financing analyses of routine immunization and new vaccine introduction costs (EPIC). Working paper, Mimeo (2014).
Complementary facility costing information at was paired with analysis of funding flows as needed. The System of Health Accounts coding [35,36] was used to analyze financial flows in a harmonized manner across the six countries. These codes were disaggregated further to be consistent with the matrix of line items and activities developed for the Common Approach (Table 4)9 . Data collection was facilitated through
9 The SHA codes are organized around the following: (1) Revenues (Financing Sources-FS): classifies funding sources at country level; (2) Health Care Financing (HF): classifies intermediary financing agents; Health Providers (HP): classifies
standard pre-tested data collection formats, and analysis was conducted using an Excel-based tool. Countries were encouraged to collect and analyze data for more than one year to evaluate trends. All findings were converted to $2011. Each team produced a mapping of financial flows for RI. 3.4. Health facility RI productivity performance The EPIC studies focused on a comparison of facilities through a graphical analysis where facilities were plotted into quadrants according to high/low unit cost per dose based on median values (X-axis) and high/low doses administered (Y-axis). These plots revealed a spread of facility types across the performance spectrum. These plots could be used to identify those facilities with the same level of output but widely different costs. Facilities that can provide the same level of output for less cost could be interpreted as being more efficient. Additional productivity indicators such as doses administered per full-time equivalent; wastage rates for different vaccines; and, drop-out rates, were compared across facilities. Statistical analysis also sought to uncover the possible reasons for variation in performance levels to provide information to managers on how to improve performance and efficiency of service delivery. 4. Selected findings of the EPIC studies We report some of the findings from the EPIC studies across countries to illustrate trends. Additional detail and analysis can be found in other manuscripts in this Supplement. First, there is a wide range of total facility immunization economic costs within the country samples, with nearly 200-fold differences in Moldova and 60-fold differences in Uganda. Benin had the smallest range between high and low cost facilities. Second, differences in total facility cost between countries are related to variation in wage rates and labor time spent on immunization. Fig. 3 illustrates the main cost drivers of facility costs for the EPIC studies and shows that the proportion of labor costs appears to
funding by type of facility; Health Care Functions (HC): classifies RI activities; Health Care Provision (FP): classifies RI line item.
A18
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
Fig. 3. Cost drivers at facility level for the EPIC studies. Source: author’s calculations.
be related to level of economic development. In Moldova, 65% of total facility cost was for labor. Ghana has a relatively high labor share because of the role of volunteer labor (5% of total cost) and higher wage rates. All of the countries had a similar immunization schedule (pentavalent, oral polio, measles and BCG vaccines). Third, average unit costs of RI at the facility level are higher than previously estimated. Table 5 shows the cost per dose ranges from $2 in Benin to $18 in Moldova, and the cost per infant varies from $19 to over $300. The cost per FIC ranges from $25 to $332, with an average of $46.50 in low income countries (Benin and Uganda), and $58.50 in socio-economically intermediate countries (Ghana and Zambia). The average cost/FIC for countries in Africa is $46.50. In Moldova, non-labor cost per dose is $6 compared to $18 with labor, and the non-labor cost per infant is $112 compared to $317. More detailed examination of cost variation will be undertaken in the next phase of EPIC. The majority of total national RI costs was for service delivery with an average of 15% for administration and management costs above the facility-level. Moldova, which had the highest immunization coverage, had an above-facility share of 18%. Further
investigation as to the relationship between management and immunization performance would be useful. Finally, governments in the EPIC study sample are financing approximately 70% of the RI program on average, ranging from 45% in Benin and Uganda to 95% in Moldova. Total available financing ranged from approximately $9 million to $50 million in 2011.
5. Policy conclusions and lessons learned The EPIC studies provide a unique set of high quality data on RI program costs, updating the evidence-base to include pentavalent vaccines in the immunization schedule. The studies reveal that governments are financing a larger total amount and share of RI than previously estimated [17]. The major cost drivers continue to be labor, followed by vaccines. The EPIC studies also showed a wide range in facility-level total costs within countries, as well as variation in unit costs between countries. The extent of this variation may mean it is no longer possible to refer to a single point estimate of unit costs. Further analysis of factors contributing to cost variation can be useful for
Table 5 Facility-based routine immunization economic costs and national-level financing ($2011). Country
Benin Uganda Ghana Zambia Honduras Moldova
Facility Weighted Average
Share of above facility costs (%)b
Total available financing (m)c
Share of government financing (%)
$25
6
$10.12
45
$61
$44
16
$34.3
45
$36
$78
$51
12
$50
85
$7
$60
$123
$66
26
$38.9
82
$8
$113
$217
$128
14
$49.1
64
$18
$317
$672
$332
18
$8.8
95
Routine immunization economic cost
Cost/dose
Cost/infant
Cost/infant ($PPP)
Cost/FICa
$14,994 ($6,026–$50,472) $23,470 ($1,911–$112,753) $16,460 ($4,385–$99,789) $28,286 ($6,260–$64,019) $12,069 ($2,165–$221,235) $11,943 ($565–$112,548)
$2
$19
$42
$5
$20
$5
Source: authors’ calculations rounded to nearest whole number. a FIC = fully immunized child represented by number of children receiving the third dose of DTP-containing vaccine. b Reflects administrative and program management costs at higher than facility levels. c Figures reflect total resources potentially available and not actual expenditures made.
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
identifying ways in which performance of the RI program can be improved [38]. In all countries, the value of labor time represented an important share of total costs, ranging from 19% in Benin to 65% in Moldova. Labor time is a shared cost and data collection required enumerators to probe to determine both the allocation to immunization, and within immunization, the allocation to various activities. Because of the exclusive focus on immunization, time allocations may be overestimated relative to other services. Future studies could explore alternative methods, such as diaries or observation, or estimate time allocations (and costs) across the full range of services provided. While most of the study teams followed the methods described in the Common Approach, there was some variation due to Ministry of Health requests. The sample size of 50 facilities was intended to be as representative as possible, with a reasonable level of power. Facility surveys are costly to conduct, and analysis can be time consuming. Achieving the right balance between feasibility and generalizability is important. Further work regarding facility sampling for costing studies and approaches for routine cost data collection and analysis are needed. The policy implications of having accurate cost and financing information are several. For instance, these data can be used to improve planning of resource requirements and financing needs at the country level. Understanding the incremental delivery costs associated with new vaccine introduction will be important for updating donor policies related to subsidizing new vaccine introduction, and for domestic and external resource mobilization for routine programs. The EPIC studies have provided much-needed information on the costs and financing of RI. The study design aimed to generate a representative sample for each country, though the results can only be interpreted as indicative for other settings. Further investigations are warranted, as there were no studies from Eastern Mediterranean, Southeast or East Asian countries in the first wave. Further analysis could be undertaken with the existing data sets both within country and across countries, particularly with respect to evaluation of productivity and determinants of costs. Finally, there is scope to refine the methods and make them useful for regular data collection and analysis by the national immunization program.
Conflict of interest statement The authors declare no conflict of interest. The EPIC studies and preparation of this manuscript were supported by the Bill & Melinda Gates Foundation. However, the views expressed represent those of the authors only.
Acknowledgements We would like to thank all of the members of the study teams who provided inputs into the development of the methods and approaches used for the EPIC studies. The contributions of Steering Committee members, most notably Carol Levin and Ulla Griffiths, were critical to the successful outcome of the EPIC studies. Finally, we appreciate the comments of two independent reviewers.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine. 2014.12.066.
A19
References [1] Brenzel L, Halsey N, Miller M, Wolfson L, Fox-Rushby J. Vaccine Preventable Diseases. In: Jamison D, Bremen J, Measham A, Alleyne G, Claeson M, Evans D, Jha P, Mills A, Musgrove P, editors. Disease control priorities in developing countries. Oxford: Oxford University Press; 2006. [2] WHO, UNICEF. World Bank State of the world’s vaccines and immunization. 3rd ed. Geneva: World Health Organization; 2009. [3] Musgrove P. Public spending on health care: how are different criteria related. Health Policy 1999;47:207–23. [4] Creese AL, Sriyabbaya N, Casabal G, Wiseso G. Bull World Health Organ 1982;60(4):621–32. [5] Brenzel L, Claquin P. Immunization programs and their costs. Soc Sci Med 1994;39(4):527–36. [6] Kaddar Miloud, Levin Ann, Dougherty Leanne, Maceira Daniel. Costs and financing of immunization programs: findings of four case studies. In: Special Initiatives Report 26. Bethesda, MD: Partnerships for Health Reform Project, Abt Associates Inc.; 2000. [7] Phonboon K, Shepard DS, Ramaboot S, Kunasol P, Preuksaraj S. The Thai expanded programme on immunization: role of immunization sessions and their cost-effectiveness. Bull World Health Organ 1989;67(2):181–8. [8] Robertson RL, Davis JH, Jobe K. Service volume and other factors affecting the costs of immunizations in the Gambia. Bull World Health Organ 1984;62(5):729–36. [9] Shepard DS, Robertson RL, Cameron CS, Saturno P, Pollack M, Manceau J. Costeffectiveness of routine and campaign vaccination strategies in Ecuador. Bull World Health Org: 1989;67:649–62. [10] Mosquiera D, Walker NR, Penny ME, Clark CF, Lanata AD, Sanderson CFB, et al. Variation in the costs of delivery routine immunization services in Peru. Bull World Health Organ 2004;82(9):676–82. [11] Creese AL, Domínguez-Uga MA. Cost-effectiveness of immunization programs in Colombia. Bull Pan Am Health Organ 1987;21(4):377–94. [12] Bishai D, McQuestion M, Chaudhry R, Wigton A. The cost of scaling up vaccination services in the world’s poorest countries. Health Aff 2006;25(2):348–56. [13] Gavi, The Vaccine Alliance. Geneva, Switzerland. www.gavi.org (last accessed August 2014). [14] Wolfson LJ, Gasse F, Lee-Martin SP, Lydon P, Magan A, Tibouti A, et al. Estimating the costs of achieving the WHO-UNICEF Global Immunization Vision and Strategy, 2006–2015. Bull World Health Organ 2008;86:27–39. [15] Lydon P, Levine R, Makinen M, Brenzel L, Michell V, Milstein J, et al. Introducing vaccines in the poorest countries: what did we learn from the GAVI experience with financial sustainability? Vaccine 2008;26:6706–16. [16] L. Brenzel, C. Politi. Historical Analysis of the Comprehensive Multi-Year Plans in GAVI-Eligible countries (2004–2015). World Health Organization. http://apps.who.int/immunization financing/analysis/Historical cMYP Analysis 2012.pdf. [17] Khan M, Khan S, Walker D, Fox-Rushby J, Cutts F, Akramuzzaman S. Cost of delivering child immunization services in urban Bangladesh: a study based on facility-level surveys. J Health Popul Nutr 2004;22(4):404–12. [18] Nader A, de Quadros C, Politi C, McQuestion M. An analysis of government immunization program expenditures in lower and lower middle income countries (2006–12). Health Policy Plan 2014, http://dx.doi.org/10.1093/ heapol/czu002. [19] World Health Organization. Global vaccine action plan. http://www.who.int/ immunization/global vaccine action plan/en/. [20] Levin Carol E, Van Minh Hoang, Odaga John, Rout Swampa Sarit, Ngoc Diep Nguyen Thi, Menezes Lysander, et al. Delivery cost of human papillomavirus vaccination of young adolescent girls in Peru, Uganda and Viet Nam. Bull World Health Organ 2013;91(8):585–92. [21] Raymond Hutubessy, Levin Ann, Wang Susan, Morgan Winthrop, Ally Mariam, Broutet Nathalie, et al. A case study using the United Republic of Tanzania: costing nationwide HPV vaccine delivery using the WHO Cervical Cancer Prevention and Control Costing Tool. BMC Med 2012;136, http://dx.doi.org/10.1186/1741-7015-10-136. [22] World Health Organization. Immunization vaccines & biologicals. http://www. who.int/mediacentre/factsheets/fs378/en/ (last accessed on August 12, 2014). [23] Brenzel L. Common approach to the costing and financing analyses for routine immunization and New Vaccine Introduction (NUVI). Mimeograph. Bill & Melinda Gates Foundation; 2014 (weblink). [24] Young D. Guidelines and database tool for estimating the costs and financing of routine immunization. Mimeograph. Bill & Melinda Gates Foundation; 2013 (weblink). [25] World Health Organization. WHO guide for standardization of economic evaluations of immunization programmes. Geneva, Switzerland: World Health Organization; 2008. [26] USAID. Asia/near east bureau guidance for costing of health service delivery projects, background and guidance, REACH, project. Arlington, VA: USAID; 1989. [27] World Health Organization. Immunization costing & financing: a tool and user guide for comprehensive multi-year planning (cMYP). Geneva, Switzerland: World Health Organization; 2008. [28] Brenzel L. Selecting an essential package of health care services using costeffectiveness analysis: a manual for professionals in developing countries. In: Data for decision making project. Department of Population Sciences and International Health, Harvard School of Public Health; 1996.
A20
L. Brenzel et al. / Vaccine 33S (2015) A13–A20
[29] Measure Evaluation. Sampling manual for health facility surveys. Washington, D.C.: Measure Evaluation; 2001. http://gametlibrary.worldbank.org/ FILES/665 Sampling%20Manual%20for%20Health%20Facility%20Surveys.pdf. [30] Bratt JH, Foreit J, Chen PL, West C, Janowitz B, de Vargas T. A comparison of four approaches for measuring clinician time use. Health Policy Plann 1999;14(4):374–81. [31] World Health Organization. Guidelines for estimating costs of introducing new vaccines into the national immunization system. Geneva, Switzerland: World Health Organization; 2002. [32] Griffiths U, Korczak VS, Ayalew D, Yigzaw A. Incremental system costs of introducing combined DTwP–hepatitis B–Hib vaccine into national immunization services in Ethiopia. Vaccine 2009;27:1426–32. [33] Clinton Health Access Initiative (CHAI). Facility-based unit costing for anteretroviral treatment in five sub-Saharan African Countries. Mimeograph. Clinton Health Access Initiative (CHAI); May 2011.
[34] World Health Organization Vaccine volume calculator spreadsheet. World Health Organization. Geneva, Switzerland. http://www.who.int/ immunization delivery/systems policy/logistics/en/index4.html. [35] Organization of Economic Cooperation and Development (OECD). 2011 System of health accounts (SHA). Paris, France. [36] Diener A, Morgan D. Deriving expenditures by disease under the System of Health Accounts (SHA) 2011 framework: a systematic analysis. Lancet 2013;381:S41, http://dx.doi.org/10.1016/S0140-6736(13)61295-7. [38] LaFond A, Kanagat N, Steinglass R, Fields R, Sequeira J, Mookherji S. Drivers of routine immunization coverage improvement in Africa: findings from district-level case studies. Health Policy Plann 2014 2014, http://dx.doi.org/10.1093/heapol/czu011.