Background
In the last two decades, several vaccines have been developed that target a range of infectious diseases of global public health importance [
1]. However, the list price of these vaccines in high income countries is substantially greater than for traditional vaccines [
2]. Hence governments and external funders with limited resources often have to trade off purchasing vaccines against investing in other preventative and therapeutic interventions. In many high income countries, assessment of the relative cost-effectiveness of competing strategies is a key element to the adoption of new health technologies [
3], alongside other considerations such as effectiveness, disease burden and equity impact. Such analyses usually focus on the health sector impact of vaccination, weighing the cost of an immunization program against the morbidity and mortality prevented, as well as potential savings derived from a reduction in health care utilization.
However, it has been suggested that these cost effectiveness analyses may present too narrow a perspective of the benefits of vaccination [
4,
5] and other child health programs [
6]. This is particularly relevant to many low and middle income countries (LMICs), where decisions about health care spending are often made in the context of wider development goals, and not on issues restricted to the health care sector alone [
7]. Furthermore, many such countries do not have locally established decision rules to facilitate the optimal allocation of resources within a fixed health care budget [
8], so wider considerations such as the impact of immunization on economic growth or national budgets are often highly relevant.
Aside from their effectiveness in reducing disease and mortality, the economic benefits of vaccination have usually been measured in terms of the averted costs of medical care. Sometimes consideration is also made of the immediate productivity loss to patients (as a result of illness or death) and their carers. However, in the longer term, it has been suggested that vaccines can increase lifetime productivity due to better physical health, decreased chance of cognitive impairment caused by some vaccine preventable diseases, and better educational outcomes through increased school attendance [
4,
9]. It has also been suggested that reduced childhood mortality may encourage mothers to decrease the number of planned births, hence increasing household investment per child and enabling greater labor force participation [
4,
10]. In the long run, decreasing fertility may improve the dependency ratio and increase investment in physical and human capital [
11,
12].
On a community level, vaccination is associated with positive externalities due to its effect on the pathogen-host ecology [
2,
13]. For instance, vaccination can decrease the infection risk of unvaccinated people since they will no longer be at risk of becoming infected by vaccinated people (herd immunity). The use of vaccines may also reduce the need for antimicrobial use against affected organisms, hence reducing antimicrobial resistance. Vaccination programs may also improve the financial sustainability and affordability of healthcare programs in LMICs. Their use as part of a treatment cluster, or in combination with other infrastructure projects (such as water management systems) to maximize community health outcomes, offers opportunities for cost sharing between programs. Additionally, introduction of vaccination to disease endemic areas may boost private demand for vaccines within communities, which could enable partial cost recovery by health departments [
14‐
22].
Traditional health economic evaluations only consider the sector-specific impact of interventions on the health care sector. However, large outbreaks of diseases such as cholera or pandemic influenza can affect other sectors of the economy such as manufacturing, tourism and transport [
2,
9,
23,
24]. In countries with a high disease burden, vaccination may modulate such outbreaks, and have a substantial impact on demand, supply, trade and investment in the wider economy.
Recently, several frameworks have been proposed by which these wider benefits of vaccination can be categorised [
4,
9]
. However, the extent to which these broader benefits are considered in current economic evaluations of vaccines is unclear. This systematic review aims to identify the extent to which broader economic benefits are already incorporated within existing economic evaluations of vaccines in LMICs, and to examine the validity of the underlying methodologies and associated outcomes employed in estimating their effect. Through the use of a narrative approach, we investigate how inclusion of these broader effects could impact on the performance of healthcare programs, and also explore the potential for incorporating them within metrics which can be adequately interpreted by decision makers looking to compare returns on investment (ROI) against those in other sectors.
Methods
A systematic review was conducted to identify economic evaluations capturing the benefits of vaccination apart from (i) health effects (morbidity and mortality averted), (ii) health treatment costs saved and (iii) short term productivity losses due to being ill or caring for someone ill. For the purpose of this review, “narrow benefits” are defined as health effects, healthcare costs and short-term productivity losses to patients and caregivers. These benefits are typically incorporated into traditional economic evaluations. They are generally short-term (lasting not much longer than the duration of the illness and its sequelae) and are restricted to the vaccinated individual and closely related individuals (such as caregivers). “Broader” benefits are defined as potential benefits of vaccination aside from these; they typically involve longer term effects and/or wider externalities other than individuals vaccinated and their caregivers.
The framework proposed in Bärnighausen
et al.[
4] also differentiates between narrow and broad benefits on these grounds. However, following our review we expanded Bärnighausen’s framework to encompass other externalities and macroeconomic-level benefits proposed in recent literature [
9]. In our classification, benefit categories were grouped under three broad headings depending on whether they related to health gains (narrow), productivity gains or community externalities (see Table
1). Productivity gains related to outcomes were divided into narrow short-term outcomes (immediate disease or death) and long-term outcomes (effects of better health leading to improved physical, educational and cognitive outcomes).
Table 1
Categorized list of the benefits of vaccination
A. Health-related benefits
|
A1. Health gains | Reduction in morbidity and mortality | Cases averted |
Deaths averted |
Disability-adjusted life years saved |
A2. Health care savings | Reduction in cost of health care borne by the public sector or private individuals | Costs saved |
B. Productivity-related benefits
|
B1. Productivity gains related to care | Reduction in lost days of work due to sickness or caring for a sick patient | Value of productivity gained |
B2. Productivity gains related to short term outcomes | Reduction in lost days of work due to sickness or death of sick patient | Value of productivity gained |
Lifetime earnings |
B3. Productivity gains related to long term outcomes | Increased lifetime productivity because better health improves cognition, educational attainment and physical strength | Educational outcomes |
Cognitive outcomes |
Lifetime earnings |
B4. Productivity gains related to household behaviour | Economic improvements due to changes in household choices such as fertility and consumption/saving as a result of improved child health and survival | Productivity |
Female labour participation |
Household investment per child |
Dependency ratio |
C. Community externalities
|
C1. Ecological effects | Health improvements in unvaccinated community members as a result of ecological effects such as herd immunity and reduced antibiotic usage. | Indirect vaccine protection |
Prevalence of antibiotic resistance |
C2. Equity | More equal distribution of health outcomes | Distribution of health outcomes |
C3. Financial sustainability | Improved financial sustainability of health care programs as a result of synergies with vaccination programs and/or stimulation of private demand. | Financial benefits |
Private demand estimates |
C4. Macroeconomic impact | Changes in the national economy or individual sectors of the economy. | Gross domestic product |
Production by economic sector |
We reviewed MEDLINE, EconLit and the National Health Service Economic Evaluation Database (NHS-EED) for English language articles published between January 1990 and July 2011. Search filters for identifying papers specific to vaccination, economic outcomes and LMICs were then applied in combination. In addition, specific keywords associated with the various benefit categories were also included in the search in combination with these filters. A separate search was conducted for the broader benefits identified by Bärnighausen
et al.[
4] (category B3-4 and C1-4 in Table
1), using category specific keywords. Full search terms can be found in the additional file [see Additional file
1. Grey literature reports and working papers published online were also identified through discussions with subject experts involved in a World Health Organization consultation on the broader economic impact of vaccination (Toronto, 2011). To ensure reproducibility of search results, grey literature reports not publicly available on the internet were excluded. Reference lists in identified articles were scanned to identify additional studies missed during the search (snowball method).
Each study identified through the database search was initially categorized on the basis of its title and abstract to determine its possible relevance to the review. Studies could be economic evaluations, reviews, epidemiological studies (observational or experimental), contingent valuation studies or descriptions of measurement tools. Articles were then read to ensure that they met the following criteria: (i) in humans, (ii) not studies of travel vaccines, (iii) discussed or measured at least one of the benefits of vaccination listed in categories B3-4 or C1-4 of Table
1, (iv) conducted in LMICs, (v) discussed both cost and benefits of vaccines, (vi) discussed economic consequences of vaccine use (and not purely epidemiological studies looking at the impact of vaccination on infection or disease).
Discussion
This review shows that several broader categories of economic benefits associated with vaccination in LMICs are being captured in primary studies and quantified in economic evaluations. Our search identified 26 studies which assessed at least one broader benefit of an immunization program.
The first category involves productivity gains because of reduced disease risk. Of these, productivity gains related to care and to short-term outcomes are more commonly captured in economic evaluations. Productivity gains related to long-term outcomes and to household behavior are not commonly estimated or incorporated in economic evaluations [
4]. There may be a number of reasons for this. Firstly, evidence for many of the proposed benefits of vaccination is still limited. Vaccine trials do not routinely incorporate cognitive, educational and behavioral outcomes. Furthermore, the final endpoint in such an analysis is the productivity of an adult of working age who either was or was not vaccinated as a child. Such an endpoint may occur too long after the vaccination event to be feasibly collected. There are also ethical problems in keeping trial participants unvaccinated once a vaccine has been shown to be safe and effective. Hence existing evidence is largely based around retrospective observational cohorts, which may be biased since they rely on controls with unvaccinated individuals.
Secondly, the economic modeling framework within which such benefits can be incorporated is still unclear. While many of the reviewed studies were able to quantify the productivity-related benefits of vaccination, apart from care-related and short-term outcome-related productivity effects this was not routinely incorporated into summary statistics such as cost-effectiveness or cost-benefit ratios. The most common means of incorporating productivity gains was the human capital approach, which values the remaining years of productive employment a person can have up to retirement based on national average incomes [
43]. However, this method only captures productivity loss due to death or disability, and not due to cognitive or educational deficits as a result of childhood illness. Furthermore, the method has been criticized in its own right, since it assumes that the work a sick or deceased worker does cannot be replaced by someone currently unemployed or by simply reassigning it to other workers [
44]. Possibly for this reason, lost income as a result of premature death was rarely incorporated in the reviewed studies, even when economic evaluations used the human capital approach to value productivity loss. Other methods (such as the friction cost approach) are less likely to overestimate productivity loss due to illness, but are equally difficult to use to capture the effects of long-term outcomes [
44].
None of the studies accounted for behavior related productivity gains such as potential reductions in fertility as a result of improved child survival. In principle, lifetime benefit models (such as by Connolly and Constenla [
27] and Bloom
et al.[
11]) are able to quantify the effect of demographic changes such as dependency ratio improvements or productivity changes from increased female workforce participation. However, primary studies are still needed to explore the link between vaccination and household decisions such as childbearing and investment spending.
Value of statistical life methodology, used by Ozawa
et al.[
22], offers a different approach by valuing lives based on individuals’ willingness to pay to reduce the risk of death rather than the economic output accrued to a year of life. This method often produces estimates of monetary value of prevented deaths that are far greater than those estimated using other methods (see for example Molinari
et al.[
45]). While this does not mean it is wrong, it does raise issues of comparability with existing economic evaluations and burden estimates conducted using more conservative methods. Another challenge in applying this methodology on a global level is the difficulty in knowing the extent to which to apply normative values on life (which mostly originate from high income countries) across national borders. Traditional cost-effectiveness analyses sidestep this problem by valuing health equally across the world in non-monetary terms; differences in decision making across countries hence stem from differences in thresholds representing societal willingness to pay for health (which reflect differences in budgets and national priorities) and in the direct economic impact of disease.
Ecological externalities, particularly herd immunity, were the most common broader benefit to be incorporated into economic evaluations. This is likely to be because the techniques for such analyses (such as the use of “transmission dynamic models”) are already well-established in the epidemiological and health economics fields [
13]. However, many LMICs lack the capacity and/or surveillance systems to perform and parameterize these sophisticated analyses [
46]. For example, infection transmission models capturing the indirect effects of vaccination may require data on behavioral and contact patterns which are not readily available in many LMICs. Besides herd immunity, negative ecological externalities may also occur such as replacement of strains of pathogens eliminated by vaccination with other strains unaffected by vaccination [
13].
Equity, affordability and financial sustainability are often important considerations for vaccine introduction in LMICs. However, trials do not to routinely collect economic information by socioeconomic strata. Such data would enable analyses of the impact of vaccines on equity or financial sustainability of other interventions, as well as elucidate any methodological difficulties with incorporating them into standard cost-effectiveness frameworks. This is evident in the case of certain spill over effects of immunization programs. Reuse of immunization infrastructure, better surveillance, human resource gains, and improved drug procurement systems, are some potential benefits of immunization programs which are rarely incorporated into cost effectiveness studies [
47]. Other frameworks such as return on investment analyses or optimization modeling may be able to address some of these issues more effectively, and provide more comprehensive representation of the societal value of vaccination.
None of the identified studies took a macroeconomic approach to evaluating the impact of vaccination, as proposed by a recent World Health Organization guidance document [
9]. Such an approach has been used to evaluate the impact of vaccination during an influenza pandemic in several European countries, using computable general equilibrium models [
48], but an LMIC application or an application to situations beyond pandemics (such as an endemic disease) has yet to be published.
This review has limitations, because it was designed to give a broad qualitative overview of existing literature rather than to enable detailed quantitative synthesis. Only one person was responsible for study selection, and the studies were not weighted by quality scores. However, the permissive inclusion criteria ensured that a variety of study types and measurement techniques were reviewed, including those employing unconventional tools and techniques. Also, our study was restricted to methods applied to LMICs only, since it was motivated by decision making in these settings, where stakeholders often require information not provided by conventional economic evaluation methods. As a result, it may not have captured novel approaches being developed or applied in high income settings.
Most (22/26) of the reviewed studies included only a single category of broader economic impact. There may be several reasons for this. Most (16/26) studies were observational, willingness to pay, return on investment or cost of illness studies rather than full economic evaluations (cost-effectiveness or cost-benefit analyses) attempting to capture all important costs and outcomes of an intervention. However, even the full economic evaluations presented a limited number of broader categories of impact, even though a comprehensive range of traditional (narrow) measures were presented, such as direct medical costs, cases avoided or lives saved. This may be due to the current novelty of these measures, as well as the lack of comprehensive guidelines for economic evaluations about which of these broader measures should be reported and in what way. A further difficulty is the complexity of the relationships between vaccination, health and broader economic outcomes, which require a range of types of evidence and techniques to quantify. Given the difficulty with both measurement and interpretation, it may be impractical to develop a single composite measure capturing all relevant economic benefits of vaccination. Instead, several evaluation techniques may need to be implemented to obtain a representative set of outcome measures. However, there is little guidance about the way several categories of benefits estimated using different techniques can be combined in the same evaluation.
Acknowledgements
We would like to thank the following participants of a WHO consultation on the broader economy impact of vaccination in Toronto, Canada, 13–14 July 2011 for helpful discussions: Arnab Acharya, Til Bärnighausen, Ricardo Brandao, Anaïs Colombini, Dagna Constenla, Adrienne Goebbels, Mira Johri, Ann Levin, Arindam Nandi, Jennifer O’Brien, Maarten Postma, Baudouin Standaert, Aparnaa Somanathan, Stéphane Verguet and Damian Walker. We also thank Sachiko Ozawa for helpful comments. Finally, we would like to thank the GAVI Alliance for their financial support. RH is a staff member of the World Health Organization. The views expressed are that of the author and do not necessarily represent the views of the World Health Organization.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MJ, RH and RD conceived the study. The search filters were developed by RD and reviewed by MJ. Initial categorisation of search results based on title and abstract and the subsequent review of articles for inclusion was conducted by RD. The final list of included articles was reviewed by MJ and RH. RD reviewed the shortlisted articles with input from MJ, RH, IvdP and SE. RD summarised the literature and wrote the first draft of the manuscript with input from MJ and RH. All authors contributed to the writing of the manuscript. All authors read and approved the final manuscript.