The costs of introducing a vaccine in sub-Saharan Africa: a systematic review of the literature The costs of introducing a vaccine 391

Purpose – This study provides a systematic literature review and categorization of the costs reported in the literature for the introduction of new vaccines, focusing on sub-Saharan Africa within LMICs, where vaccines are highly needed, financial resources are scarce and data are lacking and scattered. Design/methodology/approach – A systematic literature search of PubMed and Web of Science databases was conducted according to the PRISMA requirements. Searches also included the relevant grey literature. In total, 39 studieswere selectedand nine cost categories were investigated to build a comprehensive framework. Findings – The paper considers nine cost categories that cover the whole life of the vaccine, from its initial study to its full implementation, including for each of them the relevant subcategories. The systematic review, besides providing specific quantitative data and allowing to assess their variability within each category, points out that delivery, program preparation, administration and procurement costs are the most frequently estimated categories, while the cost of the good sold, costs borne by households and costs associated to AEFI areusuallyoverlooked.DatareportedonR&Dcostsandinvestmentintheproductionplantdiffersignificantlyamongtheselectedcontributions. Originality/value – The literature contributions on cost estimation tend to focus on a precise vaccine, a specific geographic area, or to adopt a narrow approach that captures only a subset of the costs. This article presents a rich and inclusive set of the economic quantitative data on immunization costs in limited-resource countries.


Introduction
Even before COVID-19, immunization has proven to be one of the most cost-effective medical interventions to defeat diseases worldwide (Berkley, 2014), having a great impact on human health and contributing to increase life expectancy and quality (WHO, 2013;Black, 2013). Indeed, vaccines are one of the most successful public health initiatives in eliminating or reducing the impact of infectious diseases (Timmis et al., 2017).
The recent pandemic has made the value of vaccines even clearer due to the substantial impact they had on mitigating COVID-19 outbreaks (Polack et al., 2020;Moghadas et al., 2021). Nowadays, the fast-track development and large-scale production of safe and effective COVID-19 vaccines (i.e. Pfizer BioNTech, Moderna, Janssen) would not have been possible without a renewed business vision and model characterized by both accelerated innovation and strong partnerships. The COVID pandemic has demonstrated that it is possible to develop, test and review multiple safe and effective vaccines against a new disease in less than a year. This should also be a paradigm to enhance access and rollout to vaccines, in particular in low-income and middle-income countries, since the COVID-19 pandemic may have worsened the already existing and wide gaps in healthcare services. The Ebola vaccine is another example of how solidarity, science and cooperation between international organizations and the private sector to save lives can help scaled rollout throughout sub-Saharan Africa.
Furthermore, as highlighted by Tatar and Wilson (2021), the COVID-19 campaign is a unique and valuable opportunity to advance public health and improve public trust in the healthcare system and decreasing global disparities in health outcomes.
However, despite this clear evidence, the goal of providing the needed vaccination programs, in line with the Sustainable Development Goal 3 (Good Health and Well-Being) of the UN-2030 Agenda, has not been achieved yet (UN, 2021), especially in lower-middle-income countries (LMICs), where vaccine-preventable diseases are still a relevant cause of morbidity and mortality. Furthermore, some infectious diseases are still lacking a vaccine: approximately 60% of the current all age burden of infectious diseases, there is no registered vaccine, especially in developing countries (Saul and O'Brien, 2017). Moreover, prevalence rates or the size of affected populations may be impeding the investment of discovery and industry development for many of the diseases that are currently missing a vaccine.
Therefore, the need of new vaccines and greater diffusion of those already existing is evident, especially in LMICs, where greater difficulties arise in facing the relative costs and there are more fragile health conditions. Hence, a reliable and well-founded forecast of costs is of fundamental importance and an essential prerequisite for further sustainability analysis, in order to decide among different competing health needs, prioritize the allocation of resources and attract the necessary funding. As highlighted by Bloom (2015): "looking at vaccination with an economic lens is meaningful because it communicates in the language of decision makers who have the power of the purse".
In line with this, a greater attention has been progressively paid both in estimating the immunization costs and in valuing vaccination's benefits. In 2008, WHO developed a special "Guide for Standardization of Economic Evaluations of Immunization Programmes" to provide guidance to those who conduct or critically appraise economic evaluations of immunization programmes at local, national and global levels (the document was updated in 2019). However, as underlined in the same preface of the WHO guide: "the emphasis is on what to do, rather than how to do it".
To support this last step, the literature is providing a growing number of examples. Often, however, they focus strictly on a precise vaccine (Nichol, 2003;Termrungruanglert et al., 2012) and a specific geographic area (Rose et al., 2017;Haidari et al., 2016) and they mainly adopt a narrow approach that captures only a subset of the costs (Gouglas et al., 2018) and benefits (B€ arnighausen et al., 2014) of vaccination campaigns.
The purpose of this research is to collect and cluster the main different quantitative information obtainable from the published literature to support economic evaluations of immunization programs. In particular, the present study focuses on the costs of vaccine introduction and aims at providing a systematic review of the costs reported in the literature for the introduction of new vaccines, focusing on sub-Saharan Africa within LMICs.
The reason for the choice of such a focus is because, in these countries, vaccines are highly needed; financial resources are scarce, and data are lacing and scattered. This makes it very difficult to elaborate complete cost estimations to support reliable sustainability analysis and, consequently, investments both in the development and in the delivery of vaccines.
Despite the information variety of the contributions analyzed, all the relevant cost categories with a significant impact on vaccine introduction were included. Hence, this article aims at providing an inclusive and quantitative review of the costs of vaccine introduction in limited-resource countries. The final aim is to make cost estimations more trustworthy and, thus, to provide valuable information to stakeholders involved in the introduction of a novel vaccine in sub-Saharan Africa.

Methods
Authors have been involved in studying the existing literature regarding economic evaluations of introducing a vaccine in LMICsespecially in sub-Saharan Africasince 2017, when they have been engaged in the S-AFRIVAC project (under the coordination of Achille Sclavo Foundation and the scientific direction of Rino Rappuoli), whose focus was the introduction of a vaccine against invasive non-typhoidal salmonella (iNTS) disease in sub-Saharan Africa. In particular, the authors were involved in drawing up the sustainability plan of the project.
For this reason, a remarkable amount of materials have been analyzed throughout the almost four years of study, following a well-defined process (Muka et al., 2020;Tawfik et al., 2019) summarized in Table 1.
Data collected from these contributions concern several diseases, among which meningitis, rotavirus, cholera, malaria and typhoid. Despite the variety of the contributions analyzed, an effort to embrace all the relevant cost categories that have a significant impact on the introduction of a new vaccine in sub-Saharan countries was made.

Results
Economic quantitative data on immunization costs were extracted from the selected contributions. Data were divided into cost categories that cover the whole life of the vaccine, from its initial study to its full implementation, including procurement costs. For each category, subcategories were identified in order to rationalize the different cost estimations that can be found in the literature, understand their peculiarities and seize their comparability. The division in subcategories has been carried out in such a way as to provide detailed quantitative economic data on the most relevant components.
The main cost categories included in this review were identified distinguishing them according to the particular phase of the introduction process to which they refer. Table 2 describes the nine cost categories included. Within the 39 selected contribution, the most analyzed and discussed cost categories are delivery, administration and procurement costs (they are present in 14, 19 and 15 contributions, respectively), while little quantitative information can be retrieved on the cost of the good sold, costs borne by households and on costs associated to adverse events following immunization (they are present in 1, 2 and 1 contribution, respectively). Costs related to R&D, program preparation and investment in the production plant can be found in 8, 7, 3 contributions respectively. Table 3 briefly describes the main findings for each cost category, while Table 4 summarizes the quantitative data collected (reported as in the reviewed articles, without converting them back to the same year, in order to leave freedom of elaboration and computation to stakeholders who will benefit from the cost classification and data collected of this systematic review). Supplementary files from 2 to 9 collect the information for each cost The costs of introducing a vaccine category from the selected papers. These results are separately discussed in the following sections.

Discussion
This cost review takes a broad perspective, considering also those costs that accompany and follow the actual administration of the vaccine, such as the costs borne by household. The results found point out that quantitative data are lacking and more research and quantitative studies are highly needed in order to account for the costs that have to be faced to introduce a vaccine in LMICs. Some cost categories are particularly characterized by an indefinite estimation. For instance, for the R&D category, this review clearly identifies a relevant gap in the literature: few studies contain quantitative data and most of them just provide total costs without Search strategy A systematic literature search of both scientific and economic databases was conducted according to the PRISMA requirements (Preferred Reporting Items or Systematic reviews and Meta-Analyses; Moher et al., 2009). Figure 1 shows the different phases of the search process Web of Science and PubMed databases were systematically searched and the following restrictions were applied: English was the only language selected; the dates of the publications ranged from January 1, 2001 to July 15, 2021; only peer-reviewed articles were included; only studies related to LMICs were included, due to the research question we intend to address. Moreover, articles pertaining to non-humans, abstract-only papers as preceding papers, conference, editorial or articles without available full text available were excluded. The list of excluded studies is available upon request The search was preformed using appropriate text words and thesaurus terms for papers relating to the costs of immunization program in LMICs, particularly in sub-Saharan Africa. As for keywords, for both databases, we combined "vaccination cost/s", "immunization cost/ s" and "vaccine cost categories" with each of the following research strings: "Africa", "sub-Saharan Africa" and "LMICs". Thus, no search was performed using one keyword only, since our research question concerns the introduction of a vaccine in low-middle income countries and, in particular, in sub-Saharan Africa We collected all references in a single file for each database (2.269 references for PubMed and 848 for Web of Science). After merging the references in a unique file, 481 duplicates were removed. The titles first and the abstracts afterward of each reference were screened for relevance by the authors, who judged the relevance of the abstract if the title was found to be relevant. The authors worked independently and then the references screened were collected and compared. The overlapping references were selected to be considered for the next step, the analysis of the full text. The non-overlapping references were discussed and a common final decision was reached. Our search strategy returned 441 articles to be screened for fulltext review. We excluded 411 articles since they were not relevant for the purpose of the present study and/or they did not provide quantitative reliable information about the costs of introducing a vaccine in poor countries. We were left with 30 articles eligible for our study. The references of the 30 selected articles were carefully screened and so doing 4 more contributions were selected to be included Due to the comprehensive approach this study aims at embracing, by providing quantitative information on different cost categories in order to cover the different phases required by the vaccine introduction, searches were also undertaken by hand focusing on references from identified papers, authors' own collections and review articles. Google and the bibliography of the main international agencies (as WHO, GAVI, UNICEF etc.), as well as reports from the World Bank, were searched for the relevant grey literature and 5 crucial contributions were included Study outcomes Data were extracted and stored in an Excel sheet containing the economic information reported in each publication, on the basis of their availability. Variables collected included article information (title, authors, year and journal of publication) and economic quantitative information on the costs of the immunization programs Table 1. The research process distinguishing among the different phases (Serdobova and Kieny, 2006;Phelps et al., 2017;Pronker et al., 2013). Since each phase has its own peculiarity (for instance, failure risk lies mostly in early stages, while most costs are incurred during Phase III), it would be important to have more precise costs, achieving more transparency and disclosure. Moreover, it is not always clear whether the costs provided incorporate failed attempts or not. Actually, detailed Vaccine R&D consists in identifying candidate vaccines, test safety and efficacy and evaluate efficacy in human beings. Vaccine R&D is a lengthy -it lasts more than 10 years (Serdobova and Kieny, 2006) − complex and risky process, involving high opportunity costs (Lightet al., 2009). R&D activities are followed by preclinical and clinical phases that can be divided into (Light et al., 2009;UNICEF, 2018): • phase I (preclinical), enclosing adverse effects, safety, consistency, trial and efficacy assessment studies; • phase II (clinical), with age calculation, target population dosing and clinical trial studies; • phase III (advanced development), including dose optimality tests, larger samples and scaling up.

Investments in production plant
They include all costs supported in order to establish the vaccine production, including a full new plant or an already existing one.
3 Cost of goods sold They concern all costs concerning the production of the goods sold and costs incurred throughout the product cycle life of a vaccine They concern the supply through the health system network, the storage and transport within the country to the different vaccination sites, aiming at preserving stability and reaching the entire population (WHO, 2013). They include all cold chain costs for storing vaccines and transportation logistics. They are particularly significant in Equatorial countries, where climate issues are relevant and communication routes are poorly developed. Vaccine delivery costs can be divided into service delivery costs, including personnel, program management, training, social mobilization and disease surveillance, and vaccine delivery costs, including storage, cold chain equipment, vehicles and transport. 10 5

Vaccination program preparation costs
They include micro-planning activities, training of personnel, community sensitization, social mobilization and other costs prior to vaccine implementation (Mogasale et al., 2016). Micro-planning activities mainly consist of initial planning, remuneration and allowances for planners and management staff, costs sustained for planning meetings, equipment needed for planning and transportation costs. The training of personnel requires updated training material and should be targeted both to front-line health workers and to managers: hence, technical knowledge about vaccines and their administration, management competences and communication skills are needed (Berlier et al., 2015). Community sensitization activities includes all the actions undertaken to enhance communication on the disease, contributing to target society and Stakeholder engagement. These are linked to social mobilization activities and organization of promotion campaigns. All communications activities in support of a novel immunization program need to be implemented in a strategic way in order to support in facing the challenge of how to present the need for a vaccine, as recently showed by the MenAfrivac case (Berlier et al., 2015).

6
Vaccine administration costs They refer to administration costs sustained to provide vaccination to individuals. (Mogasale et al., 2016). They include materials and supplies (i.e. cups for water, soap for hand wash, vaccination cards and supplies), site preparation (i.e. venue rental, cleaning and vaccination booth setup), vaccine administration (i.e. personnel costs, per diem and other allowances, including food and refreshment), supervision, monitoring and waste management. Transportation and logistics to vaccination field sites may also be included in vaccine administration costs (Mogasale et al., 2016).

7
Vaccine procurement costs They refer to the vaccine price, costs of freight, insurance, taxes, clearance charges and pre-clearance storage cost (Mogasale et al., 2016).

8
Costs borne by households They refer to the costs that households have to face to receive the immunization.
r e t f A 9 Costs associated with AEFI (Adverse Events Following Immunization) They concern reactions that can affect the health of vaccine recipients (Mogasale et al., 2016) and should include, for example, the salary of the personnel involved, transportation, delivery costs, medicine and laboratory costs. The costs of introducing a vaccine R&D, preclinical, early clinical and registration costs With regards to R&D costs (8 out of 39 contributions selected contained details on costs related to vaccine R&D), most of the studies just report total costs, which range from US $100 million estimated for Ebola vaccines (Phelps et al., 2017) to 2 billion (Oyston and Robinson, 2012) that refer to some common estimates for a single vaccine, based on development costs for new drugs previously elaborated (Light et al., 2009) However, only little R&D costs estimation is present in the literature (Waye et al., 2013). Furthermore, no clear detail is provided about the methodology used to come up with the total figure, pointing out the need to improve the R&D costs data. The only exception among the selected articles is represented by the study from Light et al. (2009)

Investments in production plant and
Cost of goods sold The next two categories are usually overlooked when estimating cost of vaccine introduction (respectively, 3 and 1 out of the 39 contributions contain related information). As for production plant, there are two main scenarios to be considered: one in which the vaccine requires the construction of a new production plant and the other in which only an incremental cost for an already existing plant is required. A range between US $50 to 700 million is reported (Plotkin et al., 2017) and includes facilities and equipment capitalized costs that depreciate over time, considering a period of 10 (Aguado et al., 2015) and 15 years (Light et al., 2009). According to Light et al. (2009), in the first scenario, the cost is about US $200 million. Examined studies do not provide cost estimations for the second scenario As for the cost of goods sold, precise estimations are lacking. The per-dose cost of goods sold to develop a meningococcal group A vaccine for a US manufacturer is estimated between US $0.35-1.35, considering volumes of 25-50 million doses annually and depending on composition, doses/vials number and formulation type (Aguado et al., 2015) Vaccine delivery costs Among the 39 analyzed studies, 14 contain quantitative information on delivery costs. Most of them provide overall costs, without distinguishing between service and vaccine delivery costs, but dividing them according to the following variables: vaccine type, doses, immunization campaigns, country and immunization site (facilities, urban and rural settings) As reported in Table 4, when estimated as a unique figure, mass vaccination campaigns costs range from US $0.30 to 2.09 (Cravioto et al., 2012). Range estimation is much wider for routine immunization: US $0.50 is the lowest bound (Diop et al., 2015), while US $8.07 is the highest (Sch€ utte et al., 2015) Unit delivery cost per fully immunized individual with mass vaccination campaign against cholera ranges from US $0.53 to 2.09 (Cravioto et al., 2012). Unit delivery cost per fully immunized child of new and underutilized vaccines is significantly higher and it is also affected by the hospital where immunization takes place. Estimations for Ghana range from US $23.64 for urban reproductive and child health facilities to US $80.94 for rural community health-based planning and services facilities . Some articles, focusing on vaccine supply chain per dose costs, report a range between US $0.43 and 1.21 . Others provide specific components estimations, such as cold chain (Constenla, 2015;Vaughan et al., 2020;Memirie et al., 2020), transportation (Cravioto et al., 2012), vehicles, gas price per liter and personnel costs (Vaughan et al., 2020;Portnoy et al., 2015). Some contributions interestingly distinguish between urban and rural facilities (Sch€ utte et al., 2015): delivery costs to rural centers, on average, are almost three times the costs of urban centers. Only one study explicitly provides delivery costs as an estimated incremental cost per dose (US $0.50), assuming that existing immunization program staff are involved (Diop et al., 2015) (continued ) Table 3.

Main findings for each cost category
Vaccination program preparation costs Seven out of the 39 contributions analyzed contain information on program preparation costs. As shown in Table 4, estimation for overall program preparation costs range between US $0.42 and 1.22 (Mogasale et al., 2016). Some articles provide specific information on the subcategories described in Table 2. In particular, costs for microplanning ranges from US $0.040 per dose in supplementary activities for measles (Cravioto et al., 2012)  Vaccine administration costs Nineteen studies out of those selected contain information on materials and supplies costs, either in percentage or in monetary terms. As reported in Table 4 No specific information can be retrieved on site preparation costs, while detailed costs are provided for personnel's costs. Some contributions contain useful data on hourly wage in specific African countries (Portnoy et al., 2015). The personnel's cost per dose ranges from US $0.05, which is the minimum immunization-specific, routine recurrent personnel cost per dose, estimated for 94 countries by Portnoy et al. (2015) to

The costs of introducing a vaccine
In some studies, vaccine administration costs are included into operational costs (Carias et al., 2015). Values range from US $0.30 for operational cost per dose of cholera vaccines in Bangladesh, including logistics, social mobilization, training, monitoring and surveillance, and personnel costs, to US $3.61 for typhoid conjugate vaccines in Kenya, both for routine and catch-up vaccination ( Vaccine procurement costs Fifteen studies report information on vaccine FOB price (Freight on Board, i.e. the vaccine price before shipping, insurance and customs costs are added). In the initial phases of production, when there is often only one producer, the public sector has limited information and the initial price tends to be high. Then, price declines overtime due to improvements in production capacity, economies of scale and potential competition . In developing countries, vaccine price strategy differs from the one usually applied in developed countries, where the price derives from the amount the market will bear rather than from the cost (Diop et al., 2015) As Table 4  Specific information on customs, insurance and freight can be found in IVI (International Vaccine Institute) Report (Cravioto et al., 2012). IVI includes infectious disease experts, health professionals and humanitarian leaders and its mission reads: "Discover, develop and deliver safe, effective and affordable vaccines for global public health" (https://www.ivi.int/).
Hence, there is a clear lack of information in the literature concerning additional costs to be applied to vaccine price in order to be able to compute full procurement costs, the so-called CIF (Cost, Insurance and Freight). These additional costs per dose of oral cholera are estimated as being equal to $0.22 for shipping and handling, corresponding to 15% of the vaccine price Mogasale et al. (2016) report estimations on total procurement costs for fully immunized individual against oral cholera in Guinea, South Sudan and Uganda, pointing out that this category accounts for the highest proportion of total vaccination program costs. Results range from $0.29 for Dukoral oral cholera vaccination, in rural Uganda, in a refugee camp and in a pre-emptive vaccination setting, to $7.96 for Shancol oral cholera vaccination, in South Sudan, in the case of a pre-emptive vaccination setting Costs borne by households This review includes also the costs borne by households to obtain the vaccination that, although being a minor cost category, should be considered in order to provide a more cost accountability (Cravioto et al., 2012). This cost category, mainly referring to time and transportation needed to reach the vaccination point and get the vaccination, tends to be overlooked because of poor data and information availability. Among the analyzed contributions, the 2012 IVI Report (Cravioto et al., 2012) estimates such cost as being equal to US $0.11. Other studies attempt at analyzing the travel time for households to the nearest public hospital, which is a useful piece of information to assess the costs borne by households (Marsh and Rouhani, 2018) Costs associated with AEFI In the contributions analyzed, Adverse Events Following Immunization are widely described and their occurrence explained (Enwere et al., 2015;Gessner and Halsey, 2017) but a quantitative evaluation of the costs associated with their management is lacking. The latter should include salary of trained staff engaged in monitoring, food and transportation costs, medical and advisory service provision and laboratory investigation (Mogasale et al., 2016). The only estimation found is computed for India (Mogasale et al., 2016) as reported in Table 4   Table 3.   . . . . . .   The costs of introducing a vaccine R&D costs are not easily disclosed by pharmaceutical companies for obvious reasons of protection of competition. There is a clear trade-off between informative needs of researchers and decision makers on the one hand and pharmaceutical companies on the other. Sometimes this is covered by referring to drug development, but there is no clear-cut agreement on the relationship between vaccine and drug R&D costs (Light et al., 2009;Oyston and Robinson, 2012). Moreover, clinical phases appear to be different in required resources, while remaining phases seem more similar in process but not in resource intensity (Waye et al., 2013). Furthermore, technological and scientific progress poses a challenge on the generalizability of vaccine R&D cost estimation, making the understanding of the peculiarities of each vaccine even more relevant. As a result, R&D costs should be reported more frequently and openly, possibly using a common methodology (Light et al., 2009). As for production plant costs, more research is needed, especially for incremental costs, as the setting up of a completely new plant to produce a new vaccine is infrequent. Furthermore, the reviewed literature appears to be lacking also in information on the cost of goods sold.

(continued )
Moving to delivery costs, the review points out that they should be distinguished between service and vaccine delivery costs and the latter from distribution costs that can be defined as the cost to transport the vaccine from the site of production to the country of vaccination, when the two are different. In the case of limited-resource countries, these are often covered, totally or partially, by GAVI (Portnoy et al., 2015): hence, in their computation the amount paid by GAVI and the eventual co-payment requested to the country is critical. In some studies, delivery costs are not distinguished and are included in routine immunization costs as operational costs (Antill on et al., 2017) or simply shown as percentages of total immunization costs. Moreover, delivery cost data are often fragmented, and, consequently of variable quality (Vaughan et al., 2019).
The evaluation of program preparation costs would definitely benefit from clearer estimations of incremental costs. This would contribute to the implementation of appropriate donor policies and to domestic and external resource mobilization for vaccination programs . The objectives of immunization programs definitely require a careful planning of activities, in order to guarantee successful implementation, especially for new vaccines. Many studies confirm the growing importance of non-vaccine delivery costs, among which program preparation costs (Gandhi et al., 2013;Lydon et al., 2014). This category includes different subcategories: it is important to evaluate costs separately, in order to estimate their different weights and requirements in economic and financial terms. Moreover, the challenge is that they tend to be country specific and differ according to the vaccination campaign (routine vs. catch up).
Estimation of administration costs is complex because their different subcategories present specific characteristics to be taken into account. For example, many cost items are likely to represent incremental rather than full costs Brenzel et al., 2015). No personnel unit will exclusively work on a new vaccine, and hence, percentage time's allocation should be provided. Similarly, surveillance and monitoring activities require incremental costs in terms of additional staff costs, training, data management, communications and transport (WHO, 2019;Brenzel et al., 2015). Moreover, several data sources are needed for this cost estimations. Information on salary levels can be obtained from health ministry payrolls, while other data through surveys onlyi.e. the percentage of time allocated to the new vaccine administration (WHO, 2019). Volunteer labor should also be included and computed in terms of opportunity costs . Furthermore, vaccine administration costs closely depend on the adopted vaccination strategy but not much on the vaccine type. Other key cost considerations include wastage rates and the need to be able to minimize the number of doses remaining in vials. Unfortunately, specific cost estimations on waste management are lacking, as they are often included in overall vaccine administration costs. Moreover, the fact that the vaccination is integrated with existing Expanded Program on Vaccination (EPI) schedules affects this cost category significantly, since the vaccine could be co-administered, leading to a cost reduction. Being part of the EPI schedules improves vaccination coverage by enhancing and facilitating its distribution, reduces inequalities in accessing immunization and leads to a decrease in child mortality. This is particularly relevant in the sub-Saharan Africa region where under-five mortality rate is the highest globally. Vaccines that are not part of the EPI system require different approaches to distribution (e.g. school-based clinics instead of health facility-based vaccine days, as in the case of the HPV).
The present review points out that a remarkable amount of work has been done on estimation of vaccine prices and also taking a long-run perspective, while more data need to be collected on the computation of additional costs that affect the overall vaccine price. Actually, while Freight on Board (FOB) estimations are available for different vaccines and countries, data on the Cost, Insurance and Freight (CIF) vaccine process are lacking. Moreover, vaccine development can/could be quite understandably affected by price considerations: the foreseen profitability of the immunization can be crucial in determining which vaccines are actually developed. However, for diseases that are unique to poor countries, this business model might not be applicable. In this regard, MenAfrivac was a successful example of how a group of African countries can come together and stipulate what price point (US$0.50/dose) they are willing to pay for a given vaccine. This greatly affected the development phase and ultimately the composition of who was involved in the effort, establishing a public-private partnership that led to relevant public health achievements in fighting MenA. It could be an inspiring model of action for future developments in global public health . The vaccine price is a relevant component of the overall cost computation and it is also an important parameter for potential sensitivity studies.
The estimation of the costs borne by households appears to be very challenging. The component concerning transport costs is quite country specific and differs in rural villages with respect to urban settings. Moreover, it is extremely difficult to reasonably estimate the distance households have to face, on average, to reach the vaccination point. Surveys should be administered to gain more insights and information on the costs endured by households. The same applies to the evaluation of the costs associated with AEFI management that suffer from the lack of data and the objective difficulty of distinguishing adverse events from coincidental health events that can randomly follow immunization but are unrelated to it (WHO, 2013). Actually, they are generally overlooked and we found no quantitative estimation for this category referred to sub-Saharan countries. The costs of introducing a vaccine Summing up, among the various cost categories mapped in these contributions, most studies deal with delivery, administration and procurement costs, while only a lower number of publications were found on investments in production plant, costs borne by households to get the vaccination and costs due to AEFI, although they should be taken into account in order to provide a complete cost accountability for new vaccines in developing countries. Moreover, a clear definition of the components of cost categories is highly needed, in order to enhance comparability among different settings.

Conclusions
Economic valuations of immunization programmes are one of the key issues to be considered before deciding to introduce a vaccine (WHO, 2014). This calls for further studies to offer a useful support to decision makers in evaluating the major economic challenges linked to new vaccines introduction in LMICs.
In line with this, our paper aims at providing a systematic literature review of the costs that have to be faced when a vaccine is introduced in sub-Saharan Africa. In these countries, due in particular to the strong necessity of new vaccines to defeat diseases and to the competing health needs (Madhi and Rees, 2018) and the fairly widespread need to identify different financing strategies for their introduction, it is highly relevant for all stakeholders involved to know the costs of this essential medical intervention to support economic valuations of each immunization programme.
Thirty-nine contributions containing quantitative economic data have been selected. In order to elaborate a complete cost accountability, nine cost categories were identified. Our analysis takes into account also those costs that accompany and follow the actual administration of the vaccine, such as the costs borne by household to get the vaccination. To the best of our knowledge, this is the first review on this topic to take such a broad perspective and include all these categories.
Our contribution points out that quantitative data are lacking and more research on quantitative studies is highly needed. Some cost categories are characterized by an unclear cost estimation: for instance, for the category of research and development most studies report the total figure without distinguishing among the different phases and components (Serdobova and Kieny, 2006;Phelps et al., 2017;Pronker et al., 2013). Moreover, most studies deal with delivery, administration and procurement costs, while a lower number of publications were found on the cost of the good sold and the costs borne by households to get and benefit from the vaccination, although they should be included in order to provide a complete cost accountability for new vaccines in developing countries. Precisely, these costs are less vaccine-specific and, thus, they could be even more helpful in supporting cost evaluations of immunization programmes. Similarly, the costs referred to AEFI are generally overlooked and not included in cost computations: actually, we found no quantitative estimation for this category referred to sub-Saharan countries.
Moreover, broadly speaking, a clearer definition of the costs estimated should be provided, in order to enhance the comparability among different settings. Finally, policy makers and international agencies could definitely benefit from more data on cost subcategories.
Although the number of relevant contributions of this review could have been larger, our study provides relevant points of reference for scholars and decision makers when assessing costs associated to the introduction of a new vaccine in poor countries. It aims at rationalizing the different cost estimation found in the literature, by providing quantitative cost identification, including categories that are usually overlooked. Thus, it paves the way for extensive future research and the elaboration of sustainability plans that are highly needed when evaluating the opportunity of introducing vaccines in developing countries, providing sound and reliable estimations for cost-effectiveness analysis.