Video Abstract
Mandatory vaccination has been effective in maintaining high vaccination coverage in countries such as the United States. However, there are no peer-reviewed analyses of the association between mandates and both coverage and subsequent incidence of vaccine-preventable disease in Europe.
Using data from the European Centre for Disease Prevention and Control and the World Health Organization, we evaluated the relationship between country-level mandatory vaccination policies and (1) measles and pertussis vaccine coverage and (2) the annual incidence of these diseases in 29 European countries. Multivariate negative binomial and linear regression models were used to quantify these associations.
Mandatory vaccination was associated with a 3.71 (95% confidence interval [CI]: 1.68 to 5.74) percentage point higher prevalence of measles vaccination and a 2.14 (95% CI: 0.13 to 4.15) percentage point higher prevalence of pertussis vaccination when compared with countries that did not have mandatory vaccination. Mandatory vaccination was only associated with decreased measles incidence for countries without nonmedical exemptions (adjusted incidence rate ratio = 0.14; 95% CI: 0.05 to 0.36). We did not find a significant association between mandatory vaccination and pertussis incidence.
Mandatory vaccination and the magnitude of fines were associated with higher vaccination coverage. Moreover, mandatory vaccination was associated with lower measles incidence for countries with mandatory vaccination without nonmedical exemptions. These findings can inform legislative policies aimed at increasing vaccination coverage.
Outbreaks in Europe are in part due to decreased vaccine coverage. Vaccine mandates have been used successfully in countries such as the United States to establish and maintain high vaccine coverage; however, the impact of such laws in Europe is unknown.
In evaluating current vaccine mandates in Europe, we found that mandatory vaccination and the magnitude of fines were associated with higher vaccination coverage in European countries.
Vaccination is 1 of the most useful tools for preventing infectious disease and reducing morbidity and mortality.1 Availability of vaccines has had a substantial impact on rates of vaccine-preventable diseases across the world.1 However, achieving high vaccination rates is challenged by less-than-optimal acceptance. In countries such as the United States and Australia, vaccine mandates, such as those requiring vaccination for entry into school, have been used to increase and sustain vaccination rates.2,3
In Europe, mandatory vaccination policies are heterogeneous across countries. Countries vary not only in the presence or absence of a mandate but also in the implementation and enforcement of the mandates as well as in the consequences faced by individuals who fail to comply with their country’s policy.4,5 Multiple countries in Europe, including France and Italy, have recently faced outbreaks of vaccine-preventable diseases, such as measles, and in response have changed their policies, whereas other countries are considering introducing immunization mandates.6–8 The effectiveness of these legal and administrative policies and their possible consequences has not been assessed in a peer-reviewed analysis. In the face of these changes, an evaluation of existing vaccine mandates in Europe could inform policy choices in Europe and beyond.
In this study, we aimed to determine if (1) the vaccination rates for pertussis-antigen vaccines and measles, mumps, and rubella vaccines were associated with mandatory vaccination and the availability of nonmedical exemptions to these mandates; (2) the incidence of pertussis and measles was associated with mandatory vaccination and the availability of nonmedical exemptions; and (3) if the magnitude of the financial penalties faced by noncompliant parents was associated with the rates of vaccination against pertussis and measles.
Methods
Data Collection Methods
Analyses were conducted using publicly available data on vaccine mandate policies and vaccine rates by the first author. We focused our analysis on the 29 countries that were analyzed in the Vaccine European New Integrated Collaboration Effort (VENICE).9 This included the member states of the European Union (at the time of the study) as well as Iceland and Norway.9
Vaccination Rates
Vaccination rates were collected for 2006 to 2015 for measles and 2006 to 2016 for pertussis for these countries from the Global Health Observatory, a World Health Organization (WHO) collection of health-related statistics for its 194 member states.10 These were the most recent available data at the time the analysis was conducted. We elected to focus on measles and pertussis for multiple reasons. First, recent measles outbreaks have been 1 of the main drivers of policy changes in Europe. Both measles and pertussis have relatively high infectivity, meaning that even small changes in vaccine coverage can impact disease incidence. Lastly, both are diseases for which there is good surveillance of vaccine coverage and disease incidence. This country-specific information was compiled from reports by the health ministries or departments of WHO member states and was reported by the WHO on a global level. Immunization rates are collected as 1 of the Core Health Indicators used by the WHO. Although immunization rates are self-reported by the individual countries, the WHO does provide guidance and training to health systems on what and how to report.10
Measles and Pertussis Case Counts
Both measles and pertussis case counts were acquired from the WHO’s centralized information system for infectious diseases. This system collects, analyzes, and presents data on infectious diseases in the WHO European Region, compiled from reports submitted by member states.11 The population estimates used were from Eurostat, the statistical office of the European Union. These population statistics are provided annually by member states from the country’s statistical authorities and are consolidated by Eurostat, which also ensures that the methodology is harmonized across countries to ensure comparable data.12
Vaccine Mandates
Information on whether vaccination was mandatory was obtained from a combination of sources. First, we used the VENICE 2010 survey, which indicated the 2010 status of vaccination requirements. The VENICE 2010 project surveyed experts working within immunization programs in participating countries. These experts were asked about the requirements for their country and were given a definition of “mandatory” and “recommended” to avoid misinterpretation.9 We then evaluated any changes in country-level vaccine mandate policies between 2010 and 2015 using the Vaccine Scheduler, a tool maintained by the European Centre for Disease Prevention and Control that presents the vaccine schedules of all countries in the European Union, including information on recommended versus mandated vaccinations.13
For those countries with vaccine mandates, we obtained additional information on the mandates and any available exemptions by conducting a review of country-specific vaccine policy on health ministry Websites. We collected information on the presence of nonmedical exemptions and the potential penalties faced for noncompliance, including the amount of any financial penalties, from each country's health ministry Web sites and vaccine policy.9,13–20 For accurate comparison among countries, we transformed these financial penalties using Purchasing Power Parities (PPPs) on the actual individual consumption scale, calculated and provided by Eurostat.21 PPPs are a measure of how the units of national currency of a country compare in value to a standardized unit of currency and market, which Eurostat set as 1 euro in the EU. These PPPs were calculated and aggregated by using a set of annual prices for certain products and the breakdown of expenditure on gross domestic product (GDP) provided by participating countries.21
Data Analysis
All analyses were conducted by using SAS version 9.4 (SAS Institute, Inc, Cary, NC) and R version 3.3.3. We used multivariate models to conduct a longitudinal analysis of the association between different vaccine policies, vaccination rates, and disease incidence over time in the countries of interest.
Vaccination Rates and Associations With Vaccine Policy
A linear mixed-effects model was fit to estimate the association between the vaccination rates (outcome) and mandatory vaccination policies (primary exposure) in our countries of interest. We used an autoregressive correlation structure to account for within-country correlation over time. Using propensity score methods for covariate adjustment,22,23 we adjusted for the following covariates: percentage of the population living in urban areas, proportion of adults aged 24 to 65 with at least a lower-secondary education, GINI coefficient, universal health care service index, percentage of the population <14 years of age, and median age of the country’s residents. Covariate selection was informed by the existing literature on vaccine uptake and disease incidence as well as previous evaluations of the relationship between vaccine policy and disease incidence.24–28 The GINI coefficient represents the wealth distribution of the citizens of a country and is a commonly used measure of inequality.29 The universal health care index combines 16 health service coverage indicators for reproductive health, maternal and child health, infectious diseases, noncommunicable diseases, and service capacity and access into a single metric.30 The resulting regression coefficients comparing the presence of a mandatory vaccination policy, a nonmedical vaccination exemption, and a financial penalty were interpreted as the percentage point difference associated with each of these binary conditions.
Disease Incidence and Associations With Policy
We analyzed disease incidence for both measles and pertussis and the association between disease incidence and vaccine mandate policy for the years 2006 to 2015 (for measles) and 2006 to 2016 (for pertussis) for the countries of interest. To estimate the rates of pertussis and measles infections and associations with country mandates, a negative binomial model was fit with the assumption that the number of cases for each year was distributed as a Poisson random variable with mean μ and an offset for the population of the country. A negative binomial model was used to account for overdispersion. We used an autoregressive correlation structure to account for within-country correlation of the incidence of the disease over time. We used propensity score methods for covariate adjustment to account for demographic variables that have been previously shown to be associated with vaccine uptake.22,23 Incidence rate ratios were derived from the models as a measure of association between incidence and the vaccination requirements of the country. Incidence rate ratios are interpreted as a relative difference in the incidence of the disease associated with the presence of vaccine mandates in a country.
Results
Of the 29 European countries included in this study, 7 mandated vaccination (Fig 1 A and B).14–20 Among these countries, only the Czech Republic and Latvia offered processes to acquire a nonmedical vaccination exemption.14,16 These processes required parents to either receive information (either online or through their health care provider) about vaccination or provide a written refusal of the vaccination as well as discuss the decision with their child’s health care provider.14,16 In 6 of the 7 countries that had mandatory vaccination policies in place, if parents failed to meet the requirements of their country’s regulations and requirements, they faced a financial penalty. Of these 7 countries, Latvia was the only one that did not have a financial penalty for refusing to vaccinate without obtaining a nonmedical exemption.16 In Latvia, health care providers are compelled to collect the signature of individuals who refuse vaccination and are required to inform the individual of the health consequences of failing to vaccinate.31
A, Annual reported incidence per 100 000. The mean annual reported pertussis incidence by country (2006–2016) among the entire population is shown. B, Annual reported incidence per 100 000. The mean annual reported measles incidence by country (2006–2016) among the entire population is shown. a Slovakia penalizes per parent, so if it is a 2-parent household, the total financial penalty would be €662. €, euro; Ft, forint; лв, lev; Kč, koruna; zł, złoty.
A, Annual reported incidence per 100 000. The mean annual reported pertussis incidence by country (2006–2016) among the entire population is shown. B, Annual reported incidence per 100 000. The mean annual reported measles incidence by country (2006–2016) among the entire population is shown. a Slovakia penalizes per parent, so if it is a 2-parent household, the total financial penalty would be €662. €, euro; Ft, forint; лв, lev; Kč, koruna; zł, złoty.
When adjusted to the euro, the country with the highest possible financial penalty was Hungary, where parents could face a financial penalty of up to 500 000 forints (∼€1600 or ∼$1800) in 2016 if they failed to follow vaccination requirements.17 The country with the lowest financial penalty was Bulgaria, where parents could be fined a maximum of 300 lev (∼€150 or ∼$170) in 2016 if they failed to comply with vaccination requirements (Fig 1A).17
Vaccination Rates and Associations With Vaccine Policy
Mandatory vaccination was associated with a difference of 3.71 percentage points for measles vaccination and 2.14 percentage points for pertussis vaccination when compared with countries that did not have mandatory vaccination (measles: 3.71 [95% confidence interval (CI): 1.68 to 5.74]; pertussis: 2.14 [95% CI: 0.13 to 4.15]; Table 1, Fig 2). Mandatory vaccination without exemption was associated with a difference of 3.80 percentage points for measles and 2.07 percentage points for pertussis vaccination (measles: 3.80 [95% CI: 1.23 to 6.37]; pertussis: 2.07 [95% CI: −1.74 to 5.89]).
Median percentage and interquartile ranges of vaccination rates for countries with and without mandatory vaccination. a The horizontal line indicates the average vaccination rate across all countries and years included in the study.
Median percentage and interquartile ranges of vaccination rates for countries with and without mandatory vaccination. a The horizontal line indicates the average vaccination rate across all countries and years included in the study.
Among countries that impose a fine for noncompliance by parents, every €500 increase in the maximum possible penalty was associated with an increase of 0.8 points for measles vaccination coverage (95% CI: 0.50 to 1.15; P ≤.0001) and an increase of 1.1 percentage points for pertussis vaccination coverage (95% CI: 0.95 to 1.30; P ≤.0001).
Disease Incidence and Associations With Policy
For measles, Bulgaria had the highest incidence at 32.9 cases per 100 000 individuals per year (Fig 1B). Countries with mandatory vaccination experienced lower incidence rates of measles compared with countries without mandatory vaccination, although the 95% CI crossed the null (adjusted incidence rate ratio [aIRR] = 0.38; 95% CI: 0.13 to 1.16). Countries that had mandatory vaccination without the possibility of nonmedical exemptions did have 86% lower disease incidence than countries without mandatory vaccination (aIRR = 0.14; 95% CI: 0.05 to 0.36). The presence of a financial penalty was associated with lower incidence rates of measles (aIRR = 0.14; 95% CI: 0.05 to 0.39).
Norway had the highest incidence of pertussis at 73.6 cases per 100 000 individuals per year (Fig 1A). Pertussis incidence in countries that had mandatory vaccination was 70% lower than in countries that did not have mandatory vaccination, although the 95% CI crossed the null (aIRR = 0.30; 95% CI: 0.07 to 1.30). Countries with mandatory vaccinations without the possibility of nonmedical exemption had incidence rates similar to countries without mandatory vaccination policies (aIRR = 1.04; 95% CI: 0.18 to 6.32) (Table 2). The presence of a financial penalty was associated with lower incidence rates of pertussis (aIRR = 0.42; 95% CI: 0.18 to 0.98).
Discussion
Mandatory vaccination policies in Europe were associated with higher rates of vaccination for both measles and pertussis. Although mandatory vaccination overall was not associated with a lower incidence of either measles or pertussis, when there was no option of a nonmedical vaccination exemption, the incidence of measles was significantly lower. Moreover, the presence of a financial penalty for nonvaccination was associated with a lower incidence of both pertussis and measles. Countries with a fine for noncompliance experienced incidence rates of less than half that of countries without mandatory vaccination for both measles and pertussis. Among these countries with a financial penalty, a €500 increase in the maximum penalty parents could face for failing to comply with their country’s vaccination regulations was associated with a 1.1% higher vaccination rate for pertussis and a 0.8% higher vaccination rate for measles.
Financial penalties could be an effective measure in a comprehensive mandatory vaccination program, although they need to accompany other measures, including a stable vaccine supply and vaccine education, to reach those who refuse to vaccinate. Financial penalties have been used in non-European countries, with some success. For example, in Australia, 6 months after the implementation of “No Jab No Pay” (a no vaccination, no family tax benefits program) in the beginning of 2016, “fully immunized coverage” of children at 1 year of age and 5 years of age had both reached record highs.2 However, individuals who objected to vaccination were approximately twice as likely to reside in areas of higher socioeconomic resources. In this scenario, financial disincentives may not be adequate because those more likely to refuse vaccination may also have the ability to bear the consequences of the financial disincentives.32
Association of Mandatory Vaccination, the Availability of a Nonmedical Exemption Option, and the Penalty of Noncompliance With Rates of Vaccination Against Measles (2006–2015) and Pertussis (2006–2016)
| No. Country-Years | Difference in Vaccination Rate, % (95% CI) | P | |
|---|---|---|---|
| Measles | |||
| Mandatory vaccination, yes or no | 287 | 3.71 (1.68 to 5.74) | .0003 |
| Mandatory vaccination without nonmedical exemption, yes or no | 267 | 3.80 (1.23 to 6.37) | .0037 |
| Financial penalty for noncompliance per €500 | 59 | 0.80 (0.5 to 1.15) | <.0001 |
| Pertussis | |||
| Mandatory vaccination, yes or no | 316 | 2.14 (0.13 to 4.15) | .04 |
| Mandatory vaccination without nonmedical exemption, yes or no | 294 | 2.07 (−1.74 to 5.89) | .28 |
| Financial penalty for noncompliance per €500 | 65 | 1.1 (0.95 to 1.30) | <.0001 |
| No. Country-Years | Difference in Vaccination Rate, % (95% CI) | P | |
|---|---|---|---|
| Measles | |||
| Mandatory vaccination, yes or no | 287 | 3.71 (1.68 to 5.74) | .0003 |
| Mandatory vaccination without nonmedical exemption, yes or no | 267 | 3.80 (1.23 to 6.37) | .0037 |
| Financial penalty for noncompliance per €500 | 59 | 0.80 (0.5 to 1.15) | <.0001 |
| Pertussis | |||
| Mandatory vaccination, yes or no | 316 | 2.14 (0.13 to 4.15) | .04 |
| Mandatory vaccination without nonmedical exemption, yes or no | 294 | 2.07 (−1.74 to 5.89) | .28 |
| Financial penalty for noncompliance per €500 | 65 | 1.1 (0.95 to 1.30) | <.0001 |
All models are adjusted for GDP per capita, GINI, urban population, universal health care coverage, median age, percentage of the population <14 years of age, and education level.
Association of Mandatory Vaccination, the Availability of a Nonmedical Exemption Option, and the Penalty of Noncompliance With the Incidence of Measles (2006–2015) and Pertussis (2006–2016)
| Mandatory Vaccination | No. Countries | Pertussisa Incidence Rate Ratio (95% CI) | Measles Incidence Rate Ratio (95% CI) |
|---|---|---|---|
| No mandatory vaccination | 22 | Reference | Reference |
| Mandatory vaccination | 7 | 0.30 (0.07 to 1.30) | 0.38 (0.13 to 1.16) |
| Mandatory vaccination without nonmedical exemptions | 5 | 1.04 (0.18 to 6.32) | 0.14 (0.05 to 0.36) |
| Monetary fine | |||
| No mandatory vaccination | 22 | Reference | Reference |
| Mandatory vaccination and monetary penalty for noncompliance | 6 | 0.42 (0.18 to 0.98) | 0.14 (0.05 to 0.39) |
| Mandatory Vaccination | No. Countries | Pertussisa Incidence Rate Ratio (95% CI) | Measles Incidence Rate Ratio (95% CI) |
|---|---|---|---|
| No mandatory vaccination | 22 | Reference | Reference |
| Mandatory vaccination | 7 | 0.30 (0.07 to 1.30) | 0.38 (0.13 to 1.16) |
| Mandatory vaccination without nonmedical exemptions | 5 | 1.04 (0.18 to 6.32) | 0.14 (0.05 to 0.36) |
| Monetary fine | |||
| No mandatory vaccination | 22 | Reference | Reference |
| Mandatory vaccination and monetary penalty for noncompliance | 6 | 0.42 (0.18 to 0.98) | 0.14 (0.05 to 0.39) |
Models are adjusted for GDP per capita, GINI, urban population, universal health care coverage, median age, percentage of the population <14 years of age, and education level.
Some European countries are considering making vaccinations mandatory, including implementing financial penalties for noncompliance, even if they do not have a history of such policies. Italy and France have both made childhood vaccinations mandatory since the beginning of 2017 in response to decreasing vaccination rates and outbreaks of multiple vaccine-preventable infectious diseases, including measles, although the Italian mandate was later repealed.6,7 For countries with mandatory vaccination, there can still be outbreaks of disease due to the vaccination rate still not reaching the herd immunity threshold or geographic clusters of unvaccinated individuals.24 The WHO Regional Office for Europe recently released a Guide to Tailoring Immunization Programs, offering a process through which to identify obstacles and motivators to immunization in areas of low vaccination rates and for how to design interventions tailored to meet those factors.33
Previous examination of vaccine mandates in Europe via a nonpeer-reviewed analysis by the Action Plan on Science in Society Related Issues in Epidemics and Total Pandemics (ASSET) did not report a relationship between national policies on mandatory vaccination and vaccination coverage in Europe. Although the ASSET study examined most of the countries that were included in our study, there are several differences between the ASSET analysis and our study. For example, whereas our study accounted for confounders in the regression models, the ASSET analysis ostensibly did not use adjusted models (because there were no adjusted analyses reported). Specifically, our study accounted for various factors that are associated with vaccine uptake, including education level, distribution of wealth in the country, and health care accessibility. Moreover, our study also examined a longer period of time (2006–2016) than did the ASSET study (2007–2013), providing a higher statistical power for various analyses. Whereas the ASSET study only examined the presence of a mandate in a country as a binary variable, we also examined the effects of allowing for nonmedical exemption on vaccine-preventable disease incidence and of the magnitude of a fine for noncompliance.
There are a few limitations to our study that should be considered. First, this study is ecological in nature; therefore, we cannot draw a direct link between individual vaccine refusals and vaccine policies. However, given that policy is implemented at the population level, a country-level analysis is an appropriate way to evaluate the potential impact of a policy. Second, although the WHO’s Global Health Observatory and the centralized information system for infectious diseases are the best available sources for immunization rates and disease incidence, both are reliant on reporting by physicians and officials in each country. For disease incidence, case counts rely on individuals seeking medical care in the case of disease, their physician reporting the case to the local health authority, and these local health authorities properly reporting these cases to the national health ministry or other entities who would eventually report to the WHO or the European Centers for Disease Control. However, this is true for countries with mandates and without; therefore, the associations in our analysis are likely to be an underestimate due to nondifferential misclassification. Similarly, although we conducted a review to identify the best available sources of information on country-level vaccine policies, it is possible that we either did not or were unable to find all details on a given policy. Therefore, we were unable to account for a lag in effect that a change in policy may have. Additionally, there is also the possibility that the availability of nonmedical vaccination exemptions as well as the magnitude of the financial penalties faced by those who fail to comply were measured by using information that might not have reflected the entire period studied. Using resources from 2006 to 2018, we were not able to find any indications that any countries substantively changed their mandatory vaccinations from 2006 to 2016.7,9 Lastly, both vaccine coverage and disease incidence are complicated outcomes dependent on a number of factors. We either did not or could not account for all possible factors that may influence this relationship, such as the geographic proximity of the selected countries or overall population size, instead choosing to prioritize those factors that had been previously associated with these relationships in the literature.
Conclusions
Although our findings support the use of mandatory vaccination and financial penalties to ensure the maintenance of the necessary high vaccination rates, our analysis involved existing mandates. The introduction of new mandates should be accompanied by careful surveillance of the impact on both vaccine acceptance and disease outcomes. Vaccine mandates shift the balance of convenience in favor of vaccination and, when accompanied by robust vaccine safety assurance and vaccine communications programs, have the potential to play a substantial role in decreasing the negative impacts of vaccine-preventable diseases.
Ms Vaz conducted the literature search, created the figures, conducted data collection, analysis, and interpretation and was primarily responsible for the writing of the manuscript; Ms Ellingson assisted in the literature search, data analysis and interpretation, and writing and editing of the manuscript; Prof Weiss and Dr Jenness contributed to data analysis and interpretation and the editing of the manuscript; Dr Bardají assisted in data interpretation and the editing of the final manuscript; Dr Bednarczyk contributed to the data analysis and interpretation and the writing and editing of the manuscript; Dr Omer conceptualized and designed the study and contributed to data analysis and interpretation, writing, and editing; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
Ms Vaz’s current affiliation is Department of Epidemiology, Gillings School of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC.
Ms Ellingson’s current affiliation is Department of Epidemiology of Microbial Diseases, School of Public Health, Yale University, New Haven, CT.
Dr Omer’s current affiliations are Department of Epidemiology of Microbial Diseases, School of Public Health; Department of Infectious Disease, School of Medicine; and Yale Institute of Global Health, Yale University, New Haven, CT.
FUNDING: No external funding.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/2019-2436.
Comments
Implement vaccine mandates with caution, but don't dismiss the evidence (Response from authors to Nandy et al.)
Regression analyses should be interpreted in the context of their limitations and not be used to draw causal conclusions. Given that this is precisely what we do in our manuscript, we were surprised by Nandy et al.’s interpretation that “mandatory vaccination led to increased vaccination coverage,” injecting causal assumptions where none are made. The authors also question our methodology and yet only state that they “re-ran” our analyses with no additional detail. Therefore, we cannot comment on the accuracy and robustness of their analysis or provide any direct comparison between their unshared analysis and our own peer-reviewed approach.
Despite the inability to draw causal conclusions from regression analyses, we would argue that observational studies are a key component of exploring and understanding the relationships between policies and health outcomes and should not be excluded from discussions of policy impact. Such analyses are frequently used by organizations like UNICEF to support policy recommendations, as seen in the Lancet supplement published as part of the Global Breastfeeding Initiative.1
Observational studies, in conjunction with all available evidence, should be assessed when making a policy recommendation. We do not claim that our study alone provides definitive evidence of the effectiveness of vaccine mandates nor that it should be interpreted that way. However, vaccine mandates are not always just a “quick political fix,” as Dr. Nandy and colleagues call them. Whatever political considerations UNICEF may have, our role as academic entities is to collect and present evidence and we would cite the existing body of literature on the effectiveness of vaccine mandates in certain contexts, of which our study is one piece.2
However, vaccine mandates are not a panacea, should not be implemented in isolation and are not always the correct approach. We have argued previously that care must be taken when considering the implementation of a vaccine mandate, particularly in contexts where trust in government is low or the resources are not available to make vaccines easily accessible to the population.3 Furthermore, although we did find a positive association between fines and vaccine uptake, we have argued that stakeholders should seriously consider the consequences of penalties for non-compliance.3 Although our approaches were different, our conclusions are in line with those of the Sabin Vaccine Institute (disclosure – Dr. Omer serves on the board of the Sabin Vaccine Institute), which found that legislation can be effective at increasing vaccine coverage but needs to be paired with other key functions.4
Although it is important to identify and address the underlying causes of low vaccine uptake and vaccine hesitancy, as Brewer et al. found in a review of interventions to increase vaccine coverage, interventions such as vaccine requirements that “intervene on behavior directly without trying to change individual’s thoughts and feelings” are remarkably effective in comparison with interventions targeted at changing attitudes alone in high-income countries (p. 185).5 In order to attain our common goal of universal high vaccine coverage, we should consider all options, including vaccine mandates, provided they are implemented judiciously and in settings where they do not exacerbate inequities.
1. Victora CG, Bahl R, Barros AJD, et al. Breastfeeding in the 21st century: epidemiology, mechanisms, and lifelong effect. The Lancet. 2016;387(10017):475-490.
2. Increasing Appropriate Vaccination: Vaccination Requirements for Childcare, School and College Attendance. Community Preventive Services Task Force;2016.
3. Omer SB, Betsch C, Leask J. Mandate vaccination with care. Nature. 2019;571:469-472.
4. Institute SV. Legislative Approaches to Immunization Across the European Region: December 2018. https://www.sabin.org/sites/sabin.org/files/sabin_european_immunization_.... Published 2018. Accessed February 17, 2020.
5. Brewer NT, Chapman GB, Rothman AJ, Leask J, Kempe A. Increasing Vaccination: Putting Psychological Science Into Action. Psychological Science in the Public Interest. 2017;18(3):149-207.
RE: Let’s not make conclusions about mandatory vaccination from imperfect regression analysis
Vaz and colleagues (1) used regression analysis to conclude that mandatory vaccination led to increased vaccination coverage and reduced measles incidence in Europe. They compared 21 countries classified without mandatory vaccination and seven countries classified with: Bulgaria, Czech Republic, Hungary, Latvia, Poland, Slovenia and Slovakia. The authors found positive association between vaccination coverage and size of monetary fines imposed for not taking children for vaccination.
We have several reservations on the methodology and analysis. While the seven countries have indeed reported high coverage rates for several decades, we do not believe that mandatory vaccination is the primary reason. Instead, strong immunization programmes is the major factor, with fines rarely levied in the countries with mandates. In Hungary, for example, an important reason for high coverage is that pediatric nurses make home visits to new parents, keep children on record, and follow up with defaulters (2).
Regression analysis only shows correlations and not causal effects. Rutter warns against papers like the one by Vaz et al: “From an early point in their training, all behavioral scientists are taught that statistically significant correlations do not necessarily mean any kind of causative effect. Nevertheless, the literature is full of studies [containing direct or implied causal conclusions] that are exclusively based on correlational evidence.” (p. 377) (3).
Regression results are only valid if the model holds. Unfortunately, the authors do not show any metrics to validate their confounders. This is particularly concerning given the small sample size of seven countries with mandatory vaccination and a relatively large number of controlled variables. Moreover, Table 2 shows considerable outliers in the non-mandatory group, but no adjustments are made.
Differences in coverage between the two groups seemed only to be significant before 2013. Indeed, when we ran the analysis with the addition of 2017 and 2018 data, the result does seem not hold. Measles coverage decreased to 93% in Bulgaria, Poland and Slovenia in 2018, compared to 96%-98% in earlier years (4).
In addition, we are concerned about the country categorization. The authors classified countries with mandatory vaccination if measles and pertussis vaccines were mandated, but Belgium, Cyprus, France, Greece, Italy and Malta mandate other vaccines, particularly polio (5). As polio vaccine is combined with pertussis in most European countries, pertussis vaccine is effectively also mandated. Hence, those six countries should have been classified as having mandatory vaccination, which would, of course, have led to different results.
We question the authors’ conclusion that mandatory vaccination has the potential to decrease negative impacts of vaccine-preventable diseases. In our opinion, mandatory vaccination is primarily a political intervention put forward as a quick fix without appreciating and seeking to relieve the root causes of low vaccination coverage. Several of the countries with the best measles control in Europe do not have mandatory vaccination (Netherlands, Norway, Sweden) (4,5). They have inclusive vaccination programmes with follow-up systems for parents who fall outside of the normative behavior. We believe that a public health system should be built on trust and quality services. Threats of penalties do not foster trust.
1 Vaz OM, Ellingson MK, Weiss P, et al. Mandatory vaccination in Europe. Pediatrics. January 2020; e20190620.
2 European Observatory on Health Systems and Policies and the World Health Organisation. The organization and delivery of vaccination services in the European Union. London, UK: World Health Organisation; 2018.
3 Rutter M. Proceeding from observed correlation to causal inference: the use of natural experiments, Perspectives on Psychological Science. 2007;2:377-395.
4 UNICEF, Immunization coverage by antigen (country, regional, and global trends), 2019WHO European Region, Routine immunization profile, WHO Regional Office for Europe, 2019
5 Sabin Vaccine Institute, Legislative Landscape Review: Legislative Approaches to Immunization Across the European Region, December 2018
RE: Caution when concluding on mandatory vaccination using European surveillance data
In the paper ‘Mandatory Vaccination in Europe’, Vaz and colleagues conclude that mandatory vaccination and financial penalties are associated with higher vaccination coverage in European countries. They also conclude that there is an association between mandatory vaccination and incidence rate of pertussis and measles. A pivotal premise for this association study is the assumption that data on disease incidence and vaccination coverage are comparable across European surveillance sites. We find that this assumption is not valid, and we argue that the conclusions should be reconsidered.
The WHO data on vaccination coverage are obtained from different sources, such as administrative systems, surveys or registries. Furthermore, data from many sites are adjusted to account for underreporting. In Norway, data on vaccination coverage is based on entry of all administered vaccinations in a national electronic immunization registry. The data are not adjusted for the WHO report, and are likely to slightly underreport the true vaccine uptake. The differences in data on vaccination coverage that are reported to WHO should be sufficiently acknowledged and discussed when compared across sites.
The incidence of pertussis differs greatly in European countries. Different diagnostic practices and diagnostic availability can partly explain these differences in notification rate, and underscore that data on infectious disease incidence are not directly comparable across European sites. The authors briefly discuss that differences in medical attention seeking behavior and in notification of cases to national surveillance systems can lead to underreporting in sites. However, they assume that these differences are evenly distributed in sites with and without mandatory vaccination. This might be true, but it is also possible that differences in health systems and diagnostic practices leading to under-reporting are associated with sites with a mandatory vaccination policy.
Awareness of pertussis among clinicians in Norway is high and diagnostic testing is easily available for all age groups. The overall notification rate for pertussis has been higher than in all other European countries for the past 20 years. In the study, the average notification rate was estimated to be 73.6 cases per 100 000 individuals per year. This rate reflects notification of cases in all age groups. The main objective for the pertussis immunization programme is to prevent severe disease in infants and young children. Immunity conferred by acellular pertussis vaccines is short-lived, and booster doses are needed to maintain protection. Indeed, in Norway, the notification rate of pertussis declined in 7-10 year olds and in adolescents after introduction of school-entry and adolescent booster doses. Furthermore, Norway has not experienced an increase of cases in infants, as has been seen in other European countries in recent years.
We argue that the authors cannot compare vaccination coverage and notification rates of pertussis and measles across European surveillance sites without acknowledging and discussing the strengths and weaknesses in the data. Furthermore, the overall notification rate of pertussis should not be used to evaluate the impact of immunization or the need for mandatory vaccination of infants and toddlers. The conclusions should be reconsidered.
RE: East vs West Europe
A comparison of vaccination coverage and outcome with respect to mandatory vaccinations in Europe is basically a comparison between East and West Europe, unfortunately still two completely different realities. Moreover, there are differences between North (high coverage, no mandatory) and South-Europe (low coverage, more states with mandatory vaccinations).
Furthermore, some vaccinations are indirectly mandatory as monovalent vaccines are missing. For instance, pertussis vaccination was not mandatory in Italy before 2017, but tetanus, diphtheria, hepatitis B and polio was. Therefore, pertussis is “mandatory” in Italy. The same holds true for France, Malta and Romania. Finally, ECDC vaccine scheduler has only valid data for mandatory vaccinations in the most recent dataset, but not for older versions.