https://orcid.org/0000-0002-0501-6170
Figures
Abstract
Methods
We analyzed data from the Centers for Disease Control and Prevention COVID-19 Vaccine Tracker to identify total number of adults aged 18 to 64 who received at least first dose of the COVID-19 vaccine or who were fully vaccinated in Massachusetts, Connecticut, Rhode Island, New Jersey, and Vermont during the study period of March 6 –July 31, 2021. Each of the five states contributed 148 days of a daily report on cumulative number of vaccinated people, comprising 740 state-days as the total sample size. We conducted multivariable, state-day level difference-in-differences (DID) regression using a negative binomial regression model that compared the change in outcomes for Massachusetts to those of four geographically adjacent comparison states without the lotteries, before and after the Massachusetts vaccine lottery announcement (June 15, 2021). Our analyses controlled for key state-level characteristics obtained from the American Community Survey as well as day fixed-effects to capture secular trends in the outcomes.
Results
Massachusetts COVID-19 vaccine lottery was not associated with a significant increase in the number of adults aged 18 to 64 who were fully vaccinated or received at least one dose of the vaccine, compared with other states [Full dose, incidence rate ratio (IRR): 1.04, 95% confidence interval (CI): 0.97 to 1.11, P > 0.05; At least one dose, IRR: 0.99, 95% CI: 0.93 to 1.06, P > 0.05].
Citation: Kim Y, Kim J, Li Y (2023) Did Massachusetts COVID-19 vaccine lottery increase vaccine uptake? PLoS ONE 18(1): e0279283. https://doi.org/10.1371/journal.pone.0279283
Editor: Jennifer Annette Campbell, Medical College of Wisconsin, UNITED STATES
Received: April 4, 2022; Accepted: December 2, 2022; Published: January 4, 2023
Copyright: © 2023 Kim et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: YL R01AG069733 National Institute of Health The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Vaccination is one of the most important tools to help end the COVID-19 pandemic. COVID-19 vaccines are highly effective at preventing serious infections, hospitalizations, and deaths from the virus [1–3]. However, after widespread COVID-19 vaccine uptake from late 2020 to the middle of April 2021, the vaccination rate has slowed down across the US [4].
Many states since then have implemented COVID-19 vaccine incentive programs to increase vaccination rates. On June 15 of 2021, the state of Massachusetts announced COVID-19 vaccine lottery to boost vaccine uptake in the state. While about 60% of residents have been vaccinated in Massachusetts by May of 2021, the pace of vaccinations has slowed since then [5]. The vaccine lottery had five weekly statewide drawings occurred running from July 26 to August 23, 2021 (Table 1). Massachusetts residents 18 years or older who received two doses of Pfizer/Moderna or one dose of Johnson and Johnson were eligible to enter the lottery to win one of five $1 million cash prizes. In addition, fully vaccinated residents between the ages of 12 and17 may enter for the chance to win one of five, $300,000 scholarship grants.
Previous research on the efficacy of financial incentives for increasing vaccine uptake such as COVID-19 vaccine lotteries is limited. For instance, a recent study reported that a lottery-based incentive in Ohio was not associated with increased COVID-19 vaccination rates among adults [6]. Experts have suspected that the lack of effectiveness of such programs may be due to the fact that a lottery-based incentive does not directly address vaccine hesitancy and access barriers to vaccinations [7]. Nevertheless, more evidence on the effectiveness of different state vaccine lotteries, or lack thereof, is needed to make definite conclusions on the efficacy of state COVID-19 vaccine lottery programs. To date, research on the effect of Massachusetts COVID-19 vaccine lottery on adult vaccination rates has not been explored.
We sought to examine whether Massachusetts’s large cash lottery drawings that benefit only a few lucky winners were able to motivate more people for vaccine uptake. In this study, we compared the change in the number of adult COVID-19 vaccinations before and after the Massachusetts vaccine lottery announcement with those of states that did not launch any COVID-19 vaccine lotteries that award large financial prizes.
Methods
This study was approved by the institutional review board of the University of Rochester, and informed consent was waived because this study was conducted using state-level datasets without any patient information involved.
Study sample
We analyzed data from the Centers for Disease Control and Prevention (CDC) COVID-19 Vaccine Tracker [8] to identify total number of people who received at least first dose of the COVID-19 vaccine or who were fully vaccinated (e.g., have second dose of a two-dose vaccine or one dose of a single-dose vaccine) based on where recipient lives. Data represent overall US COVID-19 vaccine deliveries and administration at national and jurisdiction levels at all vaccine partners (e.g., jurisdictional partner clinics, retail pharmacies, and long-term care facilities).
Our study sample consisted of adults aged 18 to 64 who received at least one dose of the vaccine or who were fully vaccinated in Massachusetts, or in four comparison states (Connecticut, Rhode Island, New Jersey, and Vermont) during the study period of March 6 –July 31, 2021. The four comparison states were chosen because of their geographic proximity to Massachusetts and because they did not launch the COVID-19 vaccine lotteries that awarded large cash prizes during the study period [9]. These comparison states were previously used to evaluate a major health reform in Massachusetts in the literature [10]. We limited the study period to July 31 because government moved forward requiring employees to be vaccinated late July or early August [11, 12], potentially affecting adult vaccinations in Massachusetts and the comparison states. This study period provided sufficient pre- and post-Massachusetts vaccine lottery announcement (June 15, 2021) periods for each state to keep track of total number of residents who are vaccinated. Each of the five states contributed 148 days (March 6 to July 31) of a daily report on cumulative number of people who received vaccine, thus comprising 740 state-days as the total sample size.
Outcome measures
In this study, our primary outcome measure was the daily cumulative number of fully vaccinated adults aged 18 to 64 in either Massachusetts or each of other four comparison states, where full vaccination was defined as receiving two doses of Pfizer/Moderna or one dose of Johnson and Johnson. Our secondary outcome measure was the daily cumulative number of those who received at least one dose of the COVID-19 vaccine in either Massachusetts or other comparison states.
Statistical analysis
We conducted multivariable, state-day level difference-in-differences (DID) regression that compared the change in outcomes for Massachusetts to those of four comparison states before and after the Massachusetts vaccine lottery announcement. The effect of Massachusetts lottery program on the daily cumulative number of vaccinations was estimated using a negative binomial regression model with exposure of the number of state population and a log link function. Since the outcomes were the daily cumulative number of vaccinated adults, the mean of outcomes increases over time and the variance of outcomes increases faster over time. Thus, the outcomes in this study were overdispersed. Therefore, we chose the negative binomial regression model over the Poisson regression model to allow the conditional variance of the outcome to be greater than its conditional mean, providing greater flexibility in model fitting [13]. The model structure is as follows:
(1)
The dependent variable in Eq (1) (Ysd) was cumulative number of vaccinated adults in state s in date d. The variable (MAs) was a binary indicator which was equal to the value 1 for Massachusetts and the value 0 for four comparison states. The variable (Postd) was a binary indicator which was equal to the value 1 for the post-announcement period (June 15—July 31, 2021) and the value 0 for the pre-announcement period (March 6 –June 14, 2021). The coefficient β1 estimated the expected difference in log count of vaccinated adults between Massachusetts and other four comparison states before the announcement, whereas β2 estimated the expected difference in log count of vaccinated persons in the comparison states before and after post-announcement. The main parameter of interest was the coefficient on the interaction term (β3), which estimated the expected change in log count of vaccinated adults in Massachusetts, compared to those in four comparison states, before and after lottery announcement.
A positive and significant coefficient on the interaction term would indicate that Massachusetts COVID-19 vaccine lottery was significantly associated with increased number of people vaccinated compared with other four comparison states. All results are reported as incidence rate ratios (IRRs). Data were analyzed using Stata version 16 (StataCorp) [14]. A two-sided p < 0.05 was considered statistically significant.
Covariates
The regression presented in Eq (1) included a set of variables that could have affected changes in the number of vaccinated persons (represented by ). Our analyses controlled for key state-level characteristics including percentage of non-Hispanic white population, percentage of those speaking language other than English, and median household income. These variables were obtained from the 2020 American Community Survey, conducted by the U.S. Census Bureau [15]. We further controlled for day fixed-effects to capture secular trends in the outcomes.
Results
Fig 1 presents the daily cumulative number of fully vaccinated adults in Massachusetts and other four comparison states. Massachusetts experienced a sharp increase in the number of people fully vaccinated until late May of 2021 but the rate of vaccinations has slowed down since early June despite the vaccine lottery announcement. Although the numbers of fully vaccinated adults differed by states, similar trends were observed in Massachusetts and other comparison states.
Abbreviations: MA, Massachusetts; CT, Connecticut; RI, Rhode Island; NJ, New Jersey; VT, Vermont. Note: Vertical dash line indicates Massachusetts COVID-19 Vaccine announcement date. Source: Author’s analysis of the Centers for Disease Control and Prevention COVID-19 Vaccine Tracker.
Table 2 presents descriptive statistics for state-level characteristics in 2020. Among the states, New Jersey had the largest population, followed by Massachusetts. Percentage of non-Hispanic white population was highest in Vermont but lowest in New Jersey. Percentage of those speaking language other than English was lowest in Vermont. Lastly, median household income was higher in Massachusetts and New Jersey than other states.
Table 3 shows the results of multivariable state-day level DID estimates of Massachusetts COVID-19 vaccine lottery on the number of vaccinations. After controlling for state-level covariates and day effects, Massachusetts COVID-19 vaccine lottery was not associated with a significant increase in the number of adults aged 18 to 64 who were fully vaccinated or received at least one dose of the vaccine, compared with other adjacent states [Full dose, IRR: 1.04, 95% confidence interval (CI): 0.97 to 1.11, P > 0.05; At least one dose, IRR: 0.99, 95% CI: 0.93 to 1.06, P > 0.05]. The regression confirmed that changes in the number of people who received vaccinations in Massachusetts after the lottery announcement were closely similar to those of comparison states.
Discussion
There was insufficient evidence in our analyses that the number of adult COVID-19 vaccinations between Massachusetts and other four comparison states differed before and after Massachusetts COVID-19 vaccine lottery. The COVID-19 vaccine lotteries that awarded large but uncertain financial prizes for a few selected winners may not be an effective means to encourage people to be vaccinated.
Our findings are consistent with the current literature on the association of states COVID-19 vaccine lotteries with vaccine uptake. Similar to our findings, a prior study [6] has found that after the state of Ohio announced a lottery system to pay randomly selected vaccine recipients up to $1 million, this lottery-based incentive was not associated with increased adult COVID-19 vaccination rates in Ohio. In contrast, that study found that after the lottery announcement, the slower decline in vaccinations in the US than in Ohio could be explained by expanded vaccine eligibility to adolescents, which was associated with an increase in adult vaccinations [6]. Similar to our findings, another recent article [16] showed that using data from the Johns Hopkins University Vaccine Tracker between April 28 and July 1, 2021, the large cash lotteries tied to COVID-19 vaccination in 19 states were not associated with increased number of vaccinations, when compared to states that did not have COVID-19 vaccine lotteries. Massachusetts was one of the 19 states but comparing the effects of Massachusetts COVID-19 vaccine lottery to the rest of the US without lotteries may not be an ideal comparison given different state-level characteristics and geographic proximity. Also, given that Massachusetts announced the lottery on June 15, 2021, two weeks before the end of the observation period in the study using the Johns Hopkins University Vaccine Tracker, it was likely to be too short to observe meaningful changes in vaccination rates in Massachusetts. This study contributes to this limited literature on the lottery-based incentive effects on the COVID-19 vaccine uptake in Massachusetts and provides further evidence that significant association was not found between the lottery-based incentive and the number of vaccinations.
Our findings demonstrate that lottery-style drawings may not be a promising strategy to address access barriers to vaccinations. Several approaches might be considered to boost vaccine uptake. One option is to use guaranteed small financial incentives that help offset expenses related to utilities such as electricity and water. For instance, a recent article [7] reported that in a pilot program in North Carolina, guaranteed a $25 cash card to adults who either received or drove someone to receive their first dose of COVID-19 was associated with slower decline in vaccination rate. A guaranteed small financial incentive may be far more effective, and less costly, in encouraging people to be vaccinated than drawings. Also, it is important to build up positive attitude toward COVID-19 vaccine to increase vaccine uptake. In particular, vaccine misinformation is rapidly spread by social media leading to vaccine hesitancy [17]. Therefore, health misinformation associated with COVID-19 vaccine should be addressed across social media platforms to boost vaccine uptake.
We acknowledge the limitations of our study. First, there may be potential uncontrolled characteristics that confound the estimated effect of Massachusetts COVID-19 vaccine lottery. For example, we do not have data on individual-level characteristics that may be associated with vaccine hesitancy. Second, the study findings are subject to the accuracy of data from the CDC COVID-19 Vaccine Tracker [8]. Finally, our analyses were based on daily report of number of vaccinations across five states during the study period, thus we end up having small sample size (n = 740), leading to potential loss of power to observe statistically significant association [18, 19].
Conclusions
In conclusion, this study found that vaccination rates were not significantly different for Massachusetts compared to other adjacent states before and after Massachusetts COVID-19 vaccine lottery announcement, suggesting the lack of effect of the lottery on boosting vaccine uptake rate. Alternative approaches with guaranteed financial incentives or that directly address vaccine hesitancy should be considered in future policy interventions.
References
- 1.
Centers for Disease Control and Prevention. COVID-19 Vaccines Work 2021 [cited 2022 January 24]. Available from: https://www.cdc.gov/coronavirus/2019-ncov/vaccines/effectiveness/work.html
- 2. Tenforde MW, Self WH, Adams K, Gaglani M, Ginde AA, McNeal T, et al. Association between mRNA vaccination and COVID-19 hospitalization and disease severity. Jama. 2021;326(20):2043–54. pmid:34734975
- 3. Thompson MG. Effectiveness of a Third Dose of mRNA Vaccines Against COVID-19–Associated Emergency Department and Urgent Care Encounters and Hospitalizations Among Adults During Periods of Delta and Omicron Variant Predominance—VISION Network, 10 States, August 2021–January 2022. MMWR Morbidity and Mortality Weekly Report. 2022;71.
- 4.
BBC. Covid: Where and why is the US vaccine rollout slowing down? 2021 [cited 2022 January 10]. Available from: https://www.bbc.com/news/57889961
- 5.
The Boston Globe. Baker announces ’VaxMillions’ lottery program, offering $1 million cash prizes to boost vaccination rates 2021 [cited 2022 January 10]. Available from: https://www.bostonglobe.com/2021/06/15/metro/mass-have-special-lottery-sweepstakes-fully-vaccinated-individuals-eligible-five-1-million-prizes/?p1=Article_Inline_Text_Link&p1=Article_Inline_Text_Link
- 6. Walkey AJ, Law A, Bosch NA. Lottery-based incentive in Ohio and COVID-19 vaccination rates. Jama. 2021;326(8):766–7. pmid:34213530
- 7. Wong CA, Pilkington W, Doherty IA, Zhu Z, Gawande H, Kumar D, et al. Guaranteed financial incentives for COVID-19 vaccination: a pilot program in North Carolina. JAMA Internal Medicine. 2022;182(1):78–80. pmid:34694349
- 8.
Centers for Disease Control and Prevention. COVID-19 Vaccinations in the Unites States, Jurisdiction 2021 [cited 2022 January 10]. Available from: https://data.cdc.gov/Vaccinations/COVID-19-Vaccinations-in-the-United-States-Jurisdi/unsk-b7fc
- 9.
The New York Times. These are the U.S. states trying lotteries to increase Covid vaccinations 2021 [cited 2022 January 10]. Available from: https://www.nytimes.com/2021/07/03/world/covid-vaccine-lottery.html
- 10. Miller S. The effect of insurance on emergency room visits: an analysis of the 2006 Massachusetts health reform. Journal of public Economics. 2012;96(11–12):893–908.
- 11.
The White House. FACT SHEET: President Biden to Announce New Actions to Get More Americans Vaccinated and Slow the Spread of the Delta Variant 2021 [cited 2022 January 10]. Available from: https://www.whitehouse.gov/briefing-room/statements-releases/2021/07/29/fact-sheet-president-biden-to-announce-new-actions-to-get-more-americans-vaccinated-and-slow-the-spread-of-the-delta-variant/
- 12.
U.S.News. Governors Start Mandating the COVID-19 Vaccine for State Employees 2021 [cited 2021 January 10]. Available from: https://www.usnews.com/news/best-states/articles/2021-07-29/these-governors-are-mandating-the-covid-19-vaccine-for-government-employees
- 13. Yang S, Berdine G. The negative binomial regression. The Southwest Respiratory and Critical Care Chronicles. 2015;3(10):50–4.
- 14. StataCorp L. Stata Statistical Software: Release Vol. 16. 2019.
- 15.
United States Census Bureau. American Community Survey Data [cited 2022 January 10]. Available from: https://www.census.gov/programs-surveys/acs/data.html
- 16. Dave D, Friedson AI, Hansen B, Sabia JJ. Association Between Statewide COVID-19 Lottery Announcements and Vaccinations. JAMA health forum. 2021;2(10):e213117. pmid:35977158
- 17. Kim J, Kim Y, Li Y. Source of information on COVID‐19 vaccine and vaccine hesitancy among US Medicare beneficiaries. Journal of the American Geriatrics Society. 2022;70(3):677–80.
- 18. Zhang F, Wagner AK, Ross-Degnan D. Simulation-based power calculation for designing interrupted time series analyses of health policy interventions. Journal of clinical epidemiology. 2011;64(11):1252–61. pmid:21640554
- 19. Ledolter J, Kardon RH. Focus on data: statistical design of experiments and sample size selection using power analysis. Investigative ophthalmology & visual science. 2020;61(8):11. pmid:32645134