Use of financial incentives to increase adult vaccination coverage: A narrative review of lessons learned from COVID-19 and other adult vaccination efforts

Highlights • Impact of cash transfers in incentivizing adult vaccination is marginal and their effectiveness in addressing vaccine hesitancy is inconclusive.• Lottery programs do not appear to have a consistent meaningful impact on vaccination.• Non-cash incentives do not appear to have any positive effects on vaccination.


Introduction
Access to medicines is often characterized as influenced by four factors: accessibility, availability, affordability, and acceptability [1]. In December 2020, the first coronavirus disease 2019 (COVID- 19) vaccine was introduced in the United States (US) through emergency use authorization (EUA) [2]. In 18 months, eleven COVID-19 vaccines are now authorized for use by the World Health Organization (WHO) [3], and more than 66 percent of the world's population has received at least one vaccine dose [4].
While COVID-19 vaccines are available and affordable in most countries, accessibility is an ongoing challenge in many areas of the world, leading to calls for increased investment in vaccine delivery [5]. In addition, there remain individuals and groups who are ''hesitant" towards vaccinations -meaning they display indecision or uncertainty [6,7]. While this phenomenon has been observed for other vaccine-preventable diseases, it is also true for COVID-19 [7][8][9].
The WHO classified vaccine hesitancy as among the ten biggest global health threats in 2019 [10]. Some hesitancy is due to contextual factors, including trust in expertise and authority, or religious and political beliefs. Some is tied to individuals' concerns about specific vaccines, potential side effects, or even fear of needles. Reasons for hesitancy can also be more rooted in community, geography, or social networks and can be correlated with ethnic or socio-economic background [7,11]. More recently, hesitancy has been fueled by social media and a so-called ''info-demic". For COVID-19 specifically, a survey in 23 countries found vaccine hesitancy associated with a lack of trust in COVID-19 vaccine safety and science and skepticism about its efficacy [6].
Several factors have demonstrated the potential to shift attitudes and behavior of vaccine-hesitant individuals. These include trust between patients and primary care providers, clarity about vaccine safety and efficacy, honesty around side effects, and an explanation of the role of vaccination in terms of both community https://doi.org/10.1016/j.jvacx.2022.100225 2590-1362/Ó 2022 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). and individual protection levels [12]. For childhood immunizations, strategies to overcome parental vaccine hesitancy and strengthen vaccine confidence canter around the key role of primary care in promoting vaccination [13].
While not specific to addressing hesitancy, there is also a wide body of literature that many health care interventions, including vaccines, provider-based ''pay for performance" and conditional cash transfers (CCTs), can incentivize care-seeking behavior in both high-and low-income country settings [14].
For COVID-19, many governments and municipalities have put in place incentives aimed at individuals to encourage them to get vaccinated [15]. These included direct cash transfers, lottery tickets, and non-financial incentives, such as doughnuts, blenders, marijuana, herring, or even cows [16][17][18]. In New York City, residents were offered a range of items -from a $100 pre-paid debit card, to free amusement park tickets, to a trip to the Statue of Liberty [19,20].
In this narrative review, we assess the evidence on offering cash transfers and other incentives for increasing adult vaccination uptake. We focus on COVID-19, influenza, hepatitis B, maternal tetanus, and human papillomavirus (HPV). We excluded studies of routine childhood immunizations because the target of those incentives is usually the caregivers rather than the patients themselves.

Methods
We conducted a narrative review of the literature using MED-LINE, PubMed, and Cochrane databases. The narrative review approach was chosen to enable a thematic approach and inclusion of different types of interventions, outcomes, and populations [21]. The review was limited to peer-reviewed articles published in English and Arabic between 1 January 2012 and 9 February 2022. Relevant articles were identified by conducting an abstract/title search with any of the following key terms: (Incentiv*; cash transfer) AND (COVID-19; SARS-CoV-2; Coronavirus; influenza; flu; maternal/pregnan*/wom*n tetanus) AND (Vaccin* or immunis* or immuniz*). We also hand searched the reference lists of included articles to identify additional relevant publications and searched preprint studies in MedRxiv.
Identified articles were screened by two reviewers independently with the following inclusion criteria: 1) immunization targets the adult population and 2) material or financial incentives are offered. We excluded articles that focused on 1) punitive or negative incentives (e.g., sanctions, restrictions of movement); 2) routine childhood immunization where incentives, when offered, are provided to caregivers rather than patients; 3) incentivizing behavior at the facility level (e.g., a hospital). We also excluded study protocols, opinion papers, and modeling studies. We then synthesized evidence from the articles as described below.

Results
The initial search yielded 617 articles. After title and abstract screening, we conducted a full-text appraisal of 110 articles, excluded duplicates, and identified 26 articles that met our inclusion criteria ( Table 1). The rationale for exclusion is detailed in the PRISMA flow diagram in Fig. 1. Three of the articles were systematic reviews. We extracted information on the purpose of the study, key findings as well as a type of incentive, vaccine, target group, and country. We grouped the findings according to whether the incentive was financial or material. The majority of studies were from the United States (US) (13). Studies were also identified from Nigeria (2), Germany (2), and one from each of the United Kingdom (UK), Singapore, Mexico, India, Sweden, and Australia.
No studies on COVID-19-related incentives were identified from low-or middle-income countries. While we identified two relevant studies in MedRxiv, we did not include them as they have not yet been peer-reviewed.

Conditional cash transfers
We identified twelve studies, including two systematic reviews, on conditional cash transfers for incentivizing uptake of hepatitis B, HPV, influenza, or maternal tetanus vaccinations and found evidence that cash transfers (or the promise thereof) can increase vaccination uptake or stated intent to be vaccinated.
A study by Yue et al. in Singapore randomly sampled 4,000 people aged 65 and older to take part in an experimental design where some participants were offered a partial subsidy (shopping vouchers worth 10/20/30 SGD (approx. 7.5/15/23 USD) following receipt of a seasonal influenza vaccine which they paid for themselves) [45]. Increasing incentive from 10 to 20 SGD was associated with improved vaccine uptake, while no additional gains were found with a further increase to 30 SGD. Non-working people were more sensitive to the amount offered than those working, being over 2.3-2.8 times more likely to get vaccinated if incentivized [45].
Clark et al. assessed incentives required to promote influenza vaccine uptake among 66 college students in Pennsylvania, USA [43]. Among students that had previously been offered a vaccine but had not yet been vaccinated, 85 % of participants responded that they were willing to receive the influenza vaccine for 20 USD or less, 70 % for 5 USD or less, and half required less than 1 USD. All participants in the study were willing to accept a vaccine for a certain price, and for the overwhelming majority, this was less than 100 USD. Participants were not asked if they would take the vaccine without remuneration [43].
Bronchetti et al. also focused on influenza vaccines for college students, investigating the effect of social networks and financial incentives [41]. They found that a financial incentive (30 USD) offered within two weeks of vaccine uptake was effective at increasing vaccine uptake among college students, with an 11 % difference between the study and control group [41].
Two systematic reviews have been conducted on the use of incentives to increase uptake of hepatitis B vaccine among people who inject drugs, both of which concluded that financial incentives are effective at increasing coverage. Herrmann et al., based on 5 randomized control trials (RCTs) and 1 historical trial, reported incentives being associated with an absolute increase of between 21 % and 36 % in vaccine uptake compared to control [47]. Tressler and Bhandari, based on 3 RCTs, found that the pooled incentive group had seven times the odds of completing a 3-dose vaccine course compared to control [35].
Mantzari et al. investigated the effects of financial incentives on HPV vaccine uptake in the UK. Girls 16-18 years old were offered vaccination (control) or vaccination with a voucher upon completion of the three-dose vaccination course (treatment) [37]. Incentives were associated with increased uptake of the first dose and completion of the course among both those invited for the first time and those who had been previously invited for vaccination but not attended. (Among those invited for the first time, OR 1.63 for receiving the first dose in the treatment group compared to control, OR 2.15 for completing the course. Among those with the previous non-attendance, OR 2.65 for the first dose, OR 4.28 for completing the course.) However, the uptake in the treatment group was still below the national target, suggesting that the incentive may have been more effective at reaching the undecided than the hesitant [36]. A randomized control trial by Caskey et al. found that 36 % of adolescents receiving incentives completed the full HPV vaccination course and 75 % received one or two doses, compared to 13 % completing the course and 47 % receiving one or two doses in the control group [36]. The offer of an incentive for HPV did not result in a difference between the experimental and control groups vis-à-vis seeking influenza vaccination at a later date [36].
In one of four studies identified from a low-or middle-income country, Salinas-Rodríguez and Manrique-Espinoza assessed the effect of cash transfer on tetanus, pneumococcal, and influenza vaccine uptake among the elderly in the 731 poorest rural communities in 13 Mexican states using a cross-sectional design matching cases to controls [46]. 44 000 households were surveyed, and vaccination status was self-reported. Those incentivized were more likely to have each of the three vaccinations (influenza 46 % vs 41 %; pneumococcal 52 % vs 45 %; tetanus 79 % vs 71 %) [46]. The amount of the cash transfer was not reported nor was there any analysis of cost-effectiveness more broadly.
Three studies looked at the use of conditional cash transfers on the uptake of maternal tetanus vaccines. A pilot program in Nigeria, where the incentive was part of a broad range of interventions from antenatal to postnatal care, demonstrated a significant increase of 21.66 vaccinations per 100 000 catchment population per month over the baseline (9.23 to 34.08) [39]. Another also from Nigeria, where a payment was made at the clinic upon receipt of the vaccine, showed a positive effect of a financial incentive when the incentive was large enough to compensate for the cost of travel to the clinic [40]. A study of a program in India, also offering an incentive for a range of maternal health interventions, including tetanus vaccine during antenatal care visits, similarly demonstrated a small positive effect on uptake [38].

Non-cash transfers
Evidence of non-cash incentives for hepatitis B, HPV, influenza, or maternal tetanus vaccinations was limited. We found only one systematic review and one cross-sectional study, both focusing on seasonal influenza vaccinations for health care workers. According to these studies, non-cash incentivization did not increase vaccine uptake. A systematic review by Lytras et al. (11 studies in total) on interventions to increase seasonal influenza vaccine uptake in health care workers found that incentives (including gifts, perks, raffles at the individual level or free drinks, bonus/rewards for meeting target at the group level) did not significantly affect influenza vaccination uptake [44]. Cheema et al. surveyed health care workers on the effect of a one-hour time off incentive for influenza vaccination and found no association with this influencing their decision to get vaccinated [42].

Conditional cash transfers
We identified six studies on COVID-19 vaccination, four of which focused on 'intent to vaccinate' with a hypothetical offer of a vaccine and two assessing actual cash transfer programs. The studies reported a positive effect of financial incentivization, with an increase in uptake ranging from 4.2 % to 9 %. The effect was more significant among those that identified themselves as ''undecided" than those who ''refused." One study found a significant increase in first-dose uptake but no difference in course completion [26].
A randomized survey conducted prior to the availability of vaccines by Robertson et al. in the USA of one thousand American adults in December 2020 on the intent to vaccinate for COVID-19 demonstrated financial incentives would yield an 8 % increase in uptake [29]. While the incentives proposed in this study were dramatically larger than in other studies (1000, 1500, or 2000 USD), the size of the incentive did not significantly affect the outcome. The middle-income group was most responsive to an incentive. For Black and Latino respondents, the largest incentive (2000 USD) was counter-productive [29].
A study in Germany in November 2020, also prior to the availability of vaccines, asked randomly selected participants (nonprobabilistic sample; quota representative) about intent to vaccinate for COVID-19, and did not find any effect of a potential financial incentive, even after controlling for participant financial status [31]. A later nationally representative survey conducted in March 2021, also on intent to vaccinate, showed that financial incentives would have an impact and that doubling the incentive from 25 Euros to 50 Euros corresponded to a doubling effect on vaccine uptake [27]. However, the effect size was noted only for those who were declared as ''undecided" (at 5 percentage points). Those who ''refused" were less likely to respond to the incentive suggesting it may be better to focus incentives on the ''undecided" [27].
Two studies on cash transfers conducted following the introduction of COVID-19 vaccines showed a positive impact. In Sweden, Campos-Mercado et al. conducted a randomized control trial between May and July 2021 with over 8000 participants, offering 200 Swedish kroner (24 USD) for vaccination within 30 days of the vaccine becoming available to them [24]. This was associated with a 4.2 % increase in COVID-19 vaccination rates from the baseline of 71.6 %. A similar increase was reported for ''intention" to get vaccinated. The effect was noted as similar across all socioeconomic groups [24].
A non-randomized trial by Wong et al. using a difference-indifferences approach reported findings of a two-week pilot for a COVID-19 vaccination in four counties in North Carolina, USA [33]. A 25 USD cash card was given to adults who either received or drove someone to receive their first dose of COVID-19. Incentives were associated with a higher vaccine uptake rate. 41 % reported that the cash card was an important reason for vaccination, more so if Hispanic or other non-white and if from the lower-income groups. About 9 % reported they would not have been vaccinated if the cash card had not been offered, and 15 % waited to get vaccinated until they found an event that gave a cash card or other incentive [33].

Non-cash transfers
We identified six studies on the effects of lottery entry on vaccine coverage, all assessing the impact at a population level, which showed limited to no effect. No studies were identified on the effects of other types of material incentives for increasing uptake of COVID-19 vaccines.
Studies by Barber and West and Sehgal et al. found modest improvements (approx. 1.5 % and 0.98 %, respectively) comparing vaccination rates in Ohio to a synthetic control from other states, weighted to match Ohio's population (''synthetic Ohio") [23,30]. In Ohio, the Vax-A-Million campaign offered a ticket to a weekly prize of one million USD during five weeks in May-June 2021 for the residents that had received at least one dose of a COVID-19 vaccine. In absolute terms, between the lottery announcement and the end date, the increase in vaccination coverage was modest, from approximately 42 % to 47 % [30].
A cross-sectional study by Acharya et al. using a difference-indifferences analysis estimated an aggregate 2.1 % increase in vaccination coverage associated with lottery programs in the US (11 states with a program vs 28 states without) [22]. However, when analyzed separately by state, the results were mixed, with a positive association in some states but not in others [22].
Dave et al., also using a difference-in-differences method, compared states implementing incentives against those that did not, and found that lottery-based incentives in US states were not associated with a statistically significant effect on COVID-19 vaccination rates either before or after the announcement of the drawing [25]. A study by Law et al. comparing 15 states using lotteries with 31 non-lottery states did not find a significant effect when compared with the pre-lottery trend [28]. Results from Walkey et al. similarly found no difference when comparing firstdose vaccination rates in Ohio and with states without lotteries [28].

Discussion
In this review, we explored cash and non-cash incentives offered for the adult population to improve vaccination coverage. While we found evidence of cash transfers increasing both the coverage and intention to be vaccinated, very few studies considered these effects at a population level and the ones that did found that the improvements were limited to a few percentage points in vaccination coverage. While the evidence is limited, findings on the experience with financial incentives and COVID-19 vaccines are largely consistent with findings from other adult vaccination such as hepatitis B, HPV, maternal tetanus, and influenza as well as the literature on the effectiveness of conditional cash transfers. According to evidence to date, lottery programs do not appear to have a meaningful impact on vaccination for COVID-19, with effects ranging from none to a 2.1 % increase in coverage, and no evidence was identified of positive effects of other non-cash incentives for COVID-19 or other adult vaccines. Further, most studies were conducted in the US with only four from low-and middleincome countries, none of which were on COVID-19.
Of note, for all vaccines, incentives were found to be more effective for the first dose than the second dose [26]. The reason behind this difference was not explored in any of the studies. There were also no studies identified evaluating the extent to which incentives would affect booster shots or other annual vaccination. As ''point in time," studies did not assess the extent to which payments ''now" may affect health-seeking behavior in the future, including for vaccines that are offered without incentives. Only one study explored the impact of incentivizing one vaccination (HPV) and found no adverse impact vis-a-vis later vaccination for seasonal influenza [36]. Perhaps most surprising, there was no evidence presented in any of the studies on the extent to which incentives serve to address the concerns of those who are hesitant or even increase uptake among this specific subset of the population.
Data from other conditional cash transfer programs do show that when targeted at low-income populations, incentives can be highly cost-effective as well as lifesaving [48][49][50]. For COVID-19 vaccines, only one study reported on findings related to cost or marginal cost [22]. No analysis was identified on the extent to which programs were either cost-effective or cost-saving.
Some studies raise ethical concerns that financial incentives for vaccination could be construed as coercive [51,52], and that in politically-divided contexts, government-promoted incentives might generate a backlash among those who are already hesitant, heightening suspicion of vaccination programs [51,53]. Considering existing problems with vaccine hesitancy in many populations, this is an important concern to bear in mind and plan for. There is also a question about whether the use of financial incentives as a tool to promote vaccination uptake for COVID-19 may result in a societal expectation of these in the future, affecting both COVID-19 and other vaccination programs and one study did point out a delay based on an expectation of an incentive [33].
In terms of study design, the choice of the narrative review allowed us to conduct a comprehensive overview of the published literature. While we have borrowed some methods from the practice of a systematic review (e.g. PRISMA statement), because of the limited number of studies available and the breadth of scope of the research question, we were unable to conduct any grading of evidence or combined statistical data analysis that could help reduce bias in the data and conclusion [54]. With regard to limitations, our search focused on the adult population. Expanding it to routine or other childhood immunization could have yielded more results, particularly from low-and middle-income countries. The number of databases and languages used may have also limited the findings. There are also likely to be incentive schemes implemented in low-and middle-income countries but not described or investigated in peer-reviewed literature.

Conclusion
Given the paucity of evidence, and the seriousness of the potential unintended consequences, it is remarkable how many governments, states, and cities offered incentives to increase vaccination coverage and did not embark on any type of implementation research to evaluate program effectiveness. Equally puzzling, while we may have evidence that some programs work, we do not necessarily understand why or for whom.
Moving forward, it will be important to monitor the impact of these incentive programs over time at both the individual and population levels. It will also be important to unpack the theory of change of these programs and understand whether their impact is more on those who are truly hesitant or those who are undecided. It will also be important to assess their cost-effectiveness against other known and effective ways to improve coverage, such as increasing trust with providers [55], increasing points of access and community outreach, and reducing out-of-pocket costs [56,57], and providing clear information and reminders [58].
While we did not explore the role of provider-based incentives (''pay for performance") for COVID-19 vaccines, these have shown promise to increase coverage for other types of adult vaccination and should be further explored [59][60].
CRediT authorship contribution statement

Ethical Approval
Ethical approval for this type of study is not required by our institute.

Data availability
No primary data were used for the research described in the article.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.