OBJECTIVES

Areas of increased school-entry vaccination exemptions play a key role in epidemics of vaccine-preventable diseases in the United States. California eliminated nonmedical exemptions in 2016, which increased overall vaccine coverage but also rates of medical exemptions. We examine how spatial clustering of exemptions contributed to measles outbreak potential pre- and postpolicy change.

METHODS

We modeled measles transmission in an empirically calibrated hypothetical population of youth aged 0 to 17 years in California and compared outbreak sizes under the observed spatial clustering of exemptions in schools pre- and postpolicy change with counterfactual scenarios of no postpolicy change increase in medical exemptions, no clustering of exemptions, and lower population immunization levels.

RESULTS

The elimination of nonmedical exemptions significantly reduced both average and maximal outbreak sizes, although increases in medical exemptions resulted in more than twice as many infections, on average, than if medical exemptions were maintained at prepolicy change levels. Spatial clustering of nonmedical exemptions provided some initial protection against random introduction of measles infections; however, it ultimately allowed outbreaks with thousands more infections than when exemptions were randomly distributed. The large-scale outbreaks produced by exemption clusters could not be reproduced when exemptions were distributed randomly until population vaccination was lowered by >6 percentage points.

CONCLUSIONS

Despite the high overall vaccinate rate, the spatial clustering of exemptions in schools was sufficient to threaten local herd immunity and reduce protection from measles outbreaks. Policies strengthening vaccine requirements may be less effective if alternative forms of exemptions (eg, medical) are concentrated in existing low-immunization areas.

What’s Known on This Subject:

Nonmedical exemptions to required school vaccinations cluster geographically in California. The elimination of nonmedical exemptions resulted in an increase in physician-granted medical exemption in similar geographic areas. Clustering of susceptible individuals can compromise local herd immunity.

What This Study Adds:

Our realistic simulation model demonstrated that school-level clustering of vaccine exemptions before California’s elimination of nonmedical exemptions was sufficient to threaten local herd immunity despite high overall immunization coverage; increases in medical exemptions postpolicy change reduced its effectiveness substantially.

Substantial declines in pediatric vaccination rates due to the coronavirus disease 2019 (COVID-19) pandemic pose a long-term public health threat.1,2  Before the pandemic, >1200 measles infections were reported in the United States in 2019,3  despite high population measles, mumps, and rubella (MMR) vaccination coverage among children nationally.4  Geographic areas with low vaccination coverage have played a significant role in recent measles outbreaks both in the United States and abroad.3,58  Nonmedical exemptions to school immunization requirements are an important measure of the prevalence of intentionally unvaccinated children in the United States9  and have been linked to increased risk of measles infection and transmission.10,11 

In California, the percentage of children enrolling in schools with nonmedical personal beliefs exemptions (PBEs) more than doubled between 1992 and 2014.12  Children with these exemptions tended to be spatially clustered within and across schools.13,14  Increased prevalence of vaccine-preventable disease, including measles,15  prompted California to pass Senate Bill 277 (SB277) in 2015, eliminating access to new PBE petitions (but not canceling previously granted PBEs). PBE rates dropped sharply after the implementation of SB277 in 2016; however, rates of students entering schools with permanent medical exemptions (PMEs) granted by physicians have steadily increased in the years since,16,17  suggesting a replacement effect.18  Spatial clusters of PMEs have emerged in areas with previously high rates of PBEs,19  potentially contributing to the continuation of clusters of low vaccine coverage.20 

The primary objective of this study was to examine how the spatial clustering of vaccine exemptions before and after SB277 affected measles outbreak potential. Measles is a highly contagious viral illness, with an estimated basic reproduction number of 12 to 18.21,22  The MMR vaccine is effective at preventing infection (96.9% efficacy after 2 appropriately spaced doses)23 ; however, because of the high level of contagion, high rates of population vaccination (92% to 96%) are required to achieve herd immunity.24  Clustering of susceptible individuals can weaken local herd immunity, allowing outbreaks even with high rates of population vaccination.25  We simulated measles transmission in an empirically calibrated hypothetical population of all youth in California with distributions of PBEs and PMEs observed pre- and post-SB277. We created 3 additional hypothetical distributions of exemptions to further isolate the effects of spatial clustering from changes in the prevalence of PBEs and PMEs over time.

Our hypothetical population represented all children ages 0 to 17 years in California in 2014 (N = 9 059 020). Including children younger than the Centers for Disease Control and Prevention recommended age for MMR vaccination and older children with higher overall vaccination rates was important for estimating the community-level risk of measles transmission. We used data from the 2010 US Census, 2014 American Community Survey (ACS), 2015 Environmental Systems Research Institute (ESRI) Business Analyst, and 2000 California Birth Master File to create a population representative of the age, race and ethnicity, socioeconomic, parental education, and family size distributions of the Californian population at the Census block group level. Children were given point-level geographical locations on the basis of block-level population density data from the 2010 Census.

We assigned children to schools and child care centers. School and child care assignment were important for 2 reasons. First, vaccine exemptions were measured at the school and child care center level. Individual-level vaccination data are not available for privacy reasons. Second, contact in schools and child care centers constitutes a substantial proportion of both social interaction and disease transmission among children.26  We predicted which type of school (public, private, charter, or home-based) or child care center (day care or Head Start) children were likely to attend by race and ethnicity, maternal education level, maternal employment status, and household income. We assigned children to schools on the basis of predicted school type, geographic distance to school locations, and enrollment records reported by the California Department of Education in 2014. All school-aged children (5 to 17 years) attended 1 of the 4 school types. Grade levels were assigned on the basis of age. We used this same approach to assign children to child care centers using 2014 enrollment data from the California Department of Social Services. Consistent with 2014 data from the California Department of Public Health (CDPH), 32.3% of children ages 2 to 4 years were assigned to child care.27 

To examine the effects of spatial clustering of PBEs and PMEs before and after SB277, we used the hypothetical population to create 5 separate scenarios which varied only by the assignment of exemptions. We downloaded annual CDPH data on exemptions at the school and child care center levels. We recreated the prevalence and spatial distribution of exemptions for 2 school years: 2014–2015 and 2018–2019. We refer to these school years by their fall semesters: 2014 and 2018, respectively. We chose 2014 because it represented the last school year before SB277 was passed in the California Legislature (2015); 2018 was the most recent year of postimplementation data available. Distinctions in the prevalence, spatial clustering, and age distributions of exemptions between the 5 scenarios are shown in Fig 1.

FIGURE 1

The prevalence, spatial clustering, and age distributions of nonmedical PBEs and PMEs, by exemption scenario. A, Pre-SB277. B, Post-SB277. C, Post-SB277, no PME increase. D, Pre-SB277, PBEs random by space. E, Pre-SB277, PBEs random by space and age. Note: Shaded areas on maps show clusters of schools and child care centers with statistically significant higher relative risk of exemptions compared with the statewide prevalence of exemptions in the hypothetical population. These clusters were detected by using Kulldorff spatial scan statistics,50  which compares the observed distribution of exemptions across schools to randomly generated data sets of exemptions after a Poisson distribution using the observed school enrollment totals. The program then iteratively draws circular windows of varying sizes around schools in both the simulated and observed data sets to maximize the relative risk of exemptions in schools inside (as compared with schools outside) the windows. Statistical significance of clusters is calculated by comparing the likelihood ratios for exemptions in observed data to the distribution of exemption rates in the simulated data sets (N = 999 simulated data sets). The clusters shown in the figure are significant at α < 0.05.

FIGURE 1

The prevalence, spatial clustering, and age distributions of nonmedical PBEs and PMEs, by exemption scenario. A, Pre-SB277. B, Post-SB277. C, Post-SB277, no PME increase. D, Pre-SB277, PBEs random by space. E, Pre-SB277, PBEs random by space and age. Note: Shaded areas on maps show clusters of schools and child care centers with statistically significant higher relative risk of exemptions compared with the statewide prevalence of exemptions in the hypothetical population. These clusters were detected by using Kulldorff spatial scan statistics,50  which compares the observed distribution of exemptions across schools to randomly generated data sets of exemptions after a Poisson distribution using the observed school enrollment totals. The program then iteratively draws circular windows of varying sizes around schools in both the simulated and observed data sets to maximize the relative risk of exemptions in schools inside (as compared with schools outside) the windows. Statistical significance of clusters is calculated by comparing the likelihood ratios for exemptions in observed data to the distribution of exemption rates in the simulated data sets (N = 999 simulated data sets). The clusters shown in the figure are significant at α < 0.05.

Close modal

In the first exemption scenario (Fig 1A), we assigned PBEs and PMEs by grade within schools and child care centers consistent with the pre-SB277 (2014) CDPH data. Children ages 1 to 4 years not enrolled in child care were assigned exemption rates consistent with child care centers located in the same county. Infants <1 year of age were considered unvaccinated as the first MMR dose is typically not recommended until ages 12 to 15 months.28  Siblings were given a tendency to share PBE status to represent consistency in parental vaccination decisions. In this pre-SB277 scenario, siblings did not tend to share PME status because these exemptions would have been expected to reflect medical rather than parental decisions because of relatively easy access to PBEs.29  Across all children in the population (ages 0 to 17 years), the population vaccination rate was 92.2%.

In the second scenario (Fig 1B), we assigned PBEs and PMEs consistent with the post-SB277 (2018) CDPH data. Comparing the first and second scenarios allowed us to examine changes in measles outbreak potential due to the observed changes in PBE and PME rates before and after SB277. The population vaccination rate in the second scenario was 93.0% across all children. Exemptions were assigned in schools and child care centers according the same methods in the first scenario, except that siblings tended to share exemption status regardless of type because new PBEs were not permitted (eg, if an older sibling had a pre-SB277 PBE, the younger sibling had a higher likelihood of receiving a post-SB277 PME).18 

We created 3 hypothetical exemption scenarios to better understand how measles outbreak potential may have differed under alternative (unobserved) trends in distributions of PBEs and PMEs. In the third scenario (Fig 1C), we paired post-SB277 (2018) PBE prevalence with pre-SB277 (2014) PME prevalence. This hybrid scenario allowed us to evaluate the effect of SB277 under the counterfactual condition that there was no increase in PMEs after eliminating PBE petitions. The population vaccination rate across all children in this third scenario was 93.3%.

Finally, the fourth and fifth scenarios disentangled the effect of spatial clustering of PBEs from the temporal trend. In the fourth scenario (Fig 1D), we maintained the same number of PBEs as observed within grade cohorts in 2014 (or age cohorts for non–school-aged children) but distributed them spatially at random. This preserved differences in PBE prevalence across time (ie younger children still had higher PBE rates than older children) but eliminated clustering both within schools and across geographically-near schools. To assess the impact of PBE clustering on herd immunity, we created supplementary versions of this fourth scenario with population vaccination rates decreased incrementally (by 1%) from 92% to 85%. Finally, in the fifth scenario (Fig 1E), we randomly shuffled the same number of pre-SB277 PBEs throughout the entire population. Infants <1 year of age were still considered to be unvaccinated and ineligible for PBEs. This removed both the spatial clustering and differences in prevalence across grade and age. We did not randomly reassign PMEs in either of these scenarios; children retained their pre-SB277 PME assignments.

We simulated 1000 trials of measles transmission in each of the 5 exemption scenarios using a stochastic, agent-based Susceptible-Exposed-Infected-Recovered (SEIR) model. This model separates agents into 4 stages: susceptible to infection (S), incubating after exposure (E), infectious (I), and recovered and immune (R); this model is often used for measles transmission because of the extended incubation period after exposure and immunity after infection.3032  Children could be exposed to the virus during daily interaction in homes, child care centers, schools, neighborhoods, larger community, and physician offices. Infections occurred probabilistically after exposure to an infectious child. Children ages 1 to 17 years without a PBE or PME were assumed to be properly vaccinated, with vaccine efficacy set at 96.9%. Each trial was initiated with 1 seeded infection.25,30,32  We measured outbreak size as the number of nonseeded infections that occurred over a 36-week period (the typical length of a school year). Full details of the hypothetical population, contact network, transmission parameters, and simulation model are included in the Supplemental Information.

Most trials generated no or few secondary infections across exemption scenarios (Fig 2). When exemptions were assigned as observed in schools and child care centers pre-SB277, the initial seeded infections generated no additional cases in 62.0% of trials. This proportion was slightly higher post-SB277, with 66.1% of trials generating no secondary infections, although this difference was only marginally statistically significant (P = .056, Supplemental Table 1). The hypothetical scenario in which there was no post-SB277 increase in PMEs generated the highest share of trials with no secondary infections (67.4%); this was a significant increase in the share of trials with no infections as compared with pre-SB277 (P = .012, Supplemental Table 1). The hypothetical scenarios in which pre-SB277 PBEs were distributed randomly across space and space and age had fewer trials with no secondary infections (58.1% and 57.1%, respectively) but higher shares of trials with 1 to 5 secondary infections. This finding was not sensitive to the number of seeded infections introduced in the population (Supplemental Fig 5).

FIGURE 2

Proportion of simulation trials across measles outbreak sizes, by exemption scenario Note: N = 1000 simulation trials in each scenario.

FIGURE 2

Proportion of simulation trials across measles outbreak sizes, by exemption scenario Note: N = 1000 simulation trials in each scenario.

Close modal

Although the pre- and post-SB277 scenarios in which exemptions were allowed to cluster as observed in schools and child care centers had higher shares of trials with no secondary infections, they ultimately permitted outbreaks with hundreds of cases (Fig 3). The largest outbreaks were observed pre-SB277, with maximal outbreaks containing >4000 infections. The elimination of new PBEs by SB277 significantly reduced both the mean (86.5% reduction) and maximal outbreak sizes (70.8% reduction) (Supplemental Table 2), although the largest outbreaks still generated >1000 cases. SB277 reduced PBE prevalence in the population by 51.7%; however, PME prevalence increased by 121.4% (Fig 1), offsetting some of the gains in vaccine coverage. Although increases in PMEs did not fully replace the volume of eliminated PBEs, they increased potential for larger measles outbreaks. The post-SB277 increase in PME prevalence resulted in a 115% increase in infections (7.0 compared with 3.2), on average, and 47.4% more cases in maximal outbreaks (1269 compared with 861) than would be expected if PMEs were maintained at pre-SB277 levels observed in 2014.

FIGURE 3

Distributions of measles outbreak sizes, by exemption scenario. Note: N = 1000 simulation trials in each scenario; lower bound of box shows the 25th percentile (same as minimum for all exemption scenarios), middle line in box shows the 50th percentile (median), upper bound of box shows the 75th percentile, upper whisker shows 1.5 times the interquartile range, points show outbreak sizes above this range. The x-axis (number of infections) is shown on log10 scale to increase figure legibility. Descriptive statistics for outbreak distributions by scenario are shown in Supplemental Table 2. For the hypothetical post-SB277 scenario in which PMEs did not increase after the elimination of PBEs, the values for the 25th and 50th percentiles were the same (1 infection), which is reflected in the seemingly “empty” box.

FIGURE 3

Distributions of measles outbreak sizes, by exemption scenario. Note: N = 1000 simulation trials in each scenario; lower bound of box shows the 25th percentile (same as minimum for all exemption scenarios), middle line in box shows the 50th percentile (median), upper bound of box shows the 75th percentile, upper whisker shows 1.5 times the interquartile range, points show outbreak sizes above this range. The x-axis (number of infections) is shown on log10 scale to increase figure legibility. Descriptive statistics for outbreak distributions by scenario are shown in Supplemental Table 2. For the hypothetical post-SB277 scenario in which PMEs did not increase after the elimination of PBEs, the values for the 25th and 50th percentiles were the same (1 infection), which is reflected in the seemingly “empty” box.

Close modal

Still, maximal outbreak sizes in the first 3 scenarios were orders of magnitude larger than those in which pre-SB277 PBEs were randomly distributed across space. The distributions of outbreak sizes in these scenarios were similar. Both had <2 secondary infections, on average, and maximal outbreaks of no more than 26 cases, only 3% of the size of those observed even if there had been no post-SB277 increase in PMEs. The observed spatial clustering of exemptions in schools and child care centers reduced population protection against outbreaks and increased transmission of the virus locally. Figure 4 shows that the largest outbreaks observed in the pre-SB277 scenario (solid black line) were only generated when 86% of the population was vaccinated but the PBEs were spatially random (teal dashed line).

FIGURE 4

Distributions of measles outbreak sizes pre-SB277, by level of population vaccination. N = 1000 simulation trials in each scenario. The y-axis (number of infections) shown on log10 scale to increase figure legibility. Descriptive statistics for outbreak distributions by scenario are shown in Supplemental Table 2.

FIGURE 4

Distributions of measles outbreak sizes pre-SB277, by level of population vaccination. N = 1000 simulation trials in each scenario. The y-axis (number of infections) shown on log10 scale to increase figure legibility. Descriptive statistics for outbreak distributions by scenario are shown in Supplemental Table 2.

Close modal

We demonstrated that the clustering of vaccine exemptions in schools and child care centers increased measles outbreak potential. The use of empirical data on vaccine exemptions and local interaction points answers previous calls for research to better understand how the clustering of susceptibility in schools and day care centers, specifically, affects vaccination coverage required to prevent outbreaks.30  These findings illustrate why measles outbreaks continue to occur in places with high overall population vaccination rates.

Our simulation results showed that spatial clustering of exemptions is associated with fewer small outbreaks but increased risk of large outbreaks. Although surprising, there is a logical explanation for lower risk of small outbreaks. When infections are randomly introduced into a population, randomly distributed exemptions make it easier for the virus to initially reach an unvaccinated child. The lack of other unvaccinated children clustered in the local area, however, makes it difficult to reach additional susceptible individuals and transmission dies out. Clustering of exemptions has the opposite effect; although it may be more challenging for a randomly introduced infection to reach unvaccinated children initially, once the virus encounters an exemption cluster, the presence of other susceptible children locally accelerates transmission and produces the potential for large outbreaks.

In practice, the initial protective effects of clustered exemptions are likely compromised for 2 reasons. First, introduction of measles often results from an unvaccinated individual infected abroad who returns to a local community with low vaccine coverage.6  Second, the virus does not need to be introduced directly into an exemption cluster, only sustain minimal transmission until a cluster is reached. In the measles outbreak originating at Disneyland in December 2014, an infected visitor carried the virus back to a nonimmunizing religious community near Quebec, Canada, resulting in >150 infections.33 

Our results also showed that the increased potential for large-scale outbreaks remained even after SB277 eliminated access to new nonmedical exemptions. The scale of outbreaks observed in recent years in the United States, including those in New York City and New York State in 2018–2019, were only observed in the scenarios with spatially clustered exemptions; sustained transmission during large-scale outbreaks threatens the nonendemic status of measles.3  Measles outbreaks continue to occur through contact in schools and child care centers in the postelimination era in the United States,6,34  and significant action has been taken to limit transmission in these settings during recent outbreaks, including excluding children without proof of vaccination from schools and community-wide vaccination events.35 

Spatial clustering of exemptions is not easily addressed with policy changes targeted to increasing population vaccine coverage. Increases in PMEs in schools with previously high rates of PBEs reinforces existing exemption clusters,19  which can increase local vulnerability to vaccine-preventable disease outbreaks. In California, lack of clear guidance on qualifying medical conditions, authority to grant medical exemptions, and which institutions, if any, bear responsibility for monitoring medical exemptions has likely contributed to the uneven increases in these exemptions across the state.36,37  It is essential that those with the authority to grant or review exemptions be aware of the danger that spatial clustering of vaccine refusals presents for both infections in local communities and sustained transmission in the population.38 

Our results confirmed that spatial clustering of exemptions lead to underestimations of outbreak potential. To reproduce the large-scale outbreaks observed when exemptions were spatially clustered, population vaccination rates had to be reduced by ∼6% (from 92.2% to 86%) when exemptions were spatially random. Although an absolute difference of 6% may seem small, spatially clustered exemptions can generate large differences in outbreak potential when population vaccination is near levels required for herd immunity.24,39,40  Calculations used to estimate vaccine coverage necessary for herd immunity (eg, subtracting the inverse of the basic reproduction number from 1 (1 – 1/R 0)) usually assume homogenous vaccination and random mixing between contacts.40  These estimates do not capture the uneven risk for disease transmission created by the clustering of susceptible individuals.

Spatial clustering of susceptible individuals has implications for estimating vaccine coverage needed to prevent other viral outbreaks, including severe acute respiratory syndrome coronavirus 2. Current COVID-19 vaccination rates show stark, county-level differences in the United States, with lower coverage in many Southern states and rural areas.41,42  Lower vaccination rates in counties with greater socioeconomic vulnerability43  or communities that face other social, economic, or structural barriers to vaccination can increase risk of COVID-19 infections locally. The results presented here indicate that determining the level of population vaccination required to achieve herd immunity can be challenging when spatial disparities are present in vaccine coverage.

This work has several limitations. First, the simulation results presented here provide a general illustration of consequences of clustering of vaccine exemptions rather than a definitive statement on measles outbreak risk. Simulation models are stylized representations of populations with built-in behavioral and biological assumptions which cannot capture the real-world complexity of social interaction. Yet, these studies can help model dynamic trajectories of disease transmission and isolate effects using counterfactual scenarios unable to be observed in the real world.44  Second, our data did not contain information on MMR vaccination status; children with exemptions may have been unvaccinated, partially vaccinated, or fully vaccinated against measles. Previous research has shown, however, that MMR coverage was relatively low among children with PBEs in California in 2009, particularly in schools with high PBE rates.45  Third, because our goal was to estimate potential for measles outbreaks rather than examine effects of intervention strategies,30  we allowed transmission to spread through the population without interventions. Mitigation strategies would have been deployed in real-world outbreaks.

Recent increases in measles incidence in the United States and abroad suggest that focusing efforts on maintaining high population vaccine coverage is not sufficient to prevent outbreaks. Population-level vaccination programs allow for the creation of pockets of partially vaccinated or unvaccinated individuals that increase the risk of network-wide epidemics.46  Although the majority of infections in postelimination settings occur in unvaccinated individuals,34  herd immunity is a public benefit that provides protection to those too young or medically ineligible to be vaccinated. Ongoing debates over school-vaccine policies15,47  should consider the effects of even low rates of vaccine refusals when clustered in the same schools or communities. Interventions targeting parents in schools and communities with high rates of vaccine refusals, including increasing the visibility of decisions to vaccinate among close social contacts, may also be effective at leveraging social influence to change local vaccination norms.48,49  Vaccine programs and policies must recognize that population-wide protection against disease outbreaks depends not only on overall vaccination coverage but on locations of nonvaccination.

We thank William Roy, Edward Walker, Gabriel Rossman, Peter Bearman, and the anonymous reviewers for insightful feedback on earlier versions of this work. We thank Thomas Rice and participants of the University of California, Los Angeles Department of Health Policy and Management Doctoral Studies Discussion Forum for helpful comments while revising this draft.

FUNDING: Ashley Gromis’ work is supported by grant T32HS000046 from the Agency for Healthcare Research and Quality. Ka-Yuet Liu benefited from facilities and resources provided by the California Center for Population Research at the University of California, Los Angeles, which receives core support from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (grant P2CHD041022). Her work was partly supported by a Hellman Foundation Fellowship. The funders of the study had no role in design and conduct of the study; collection, management, analysis, or interpretation of the data; and preparation, review, or approval of the manuscript or the decision to submit for publication.

Dr Gromis conceptualized and designed the study, conducted the data analysis, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Liu conceptualized and designed the study, and reviewed and revised the manuscript; and both authors approved the final manuscript as submitted.

COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2021-054369.

     
  • ACS

    American Community Survey

  •  
  • CDPH

    California Department of Public Health

  •  
  • COVID-19

    coronavirus disease 2019

  •  
  • ESRI

    Environmental Systems Research Institute

  •  
  • MMR

    measles, mumps, and rubella

  •  
  • PBE

    personal beliefs exemption

  •  
  • PME

    permanent medical exemption

  •  
  • SB277

    Senate Bill 277

  •  
  • SEIR

    Susceptible-Exposed-Infected-Recovered

1
Santoli
JM
,
Lindley
MC
,
DeSilva
MB
, et al
.
Effects of the COVID-19 pandemic on routine pediatric vaccine ordering and administration - United States, 2020
.
MMWR Morb Mortal Wkly Rep.
2020
;
69
(
19
):
591
593
2
Bode
SM
,
Gowda
C
,
Mangini
M
,
Kemper
AR
.
COVID-19 and primary measles vaccination rates in a large primary care network
.
Pediatrics.
2021
;
147
(
1
):
e2020035576
3
Patel
M
,
Lee
AD
,
Clemmons
NS
, et al
.
National update on measles cases and outbreaks - United States, January 1-October 1, 2019
.
MMWR Morb Mortal Wkly Rep.
2019
;
68
(
40
):
893
896
4
Hill
HA
,
Elam-Evans
LD
,
Yankey
D
,
Singleton
JA
,
Kang
Y
.
Vaccination coverage among children aged 19-35 months - United States, 2017
.
MMWR Morb Mortal Wkly Rep.
2018
;
67
(
40
):
1123
1128
5
Majumder
MS
,
Cohn
EL
,
Mekaru
SR
,
Huston
JE
,
Brownstein
JS
.
Substandard vaccination compliance and the 2015 measles outbreak
.
JAMA Pediatr.
2015
;
169
(
5
):
494
495
6
Sugerman
DE
,
Barskey
AE
,
Delea
MG
, et al
.
Measles outbreak in a highly vaccinated population, San Diego, 2008: role of the intentionally undervaccinated
.
Pediatrics.
2010
;
125
(
4
):
747
755
7
Gastañaduy
PA
,
Funk
S
,
Paul
P
, et al
.
Impact of public health responses during a measles outbreak in an Amish community in Ohio: modeling the dynamics of transmission
.
Am J Epidemiol.
2018
;
187
(
9
):
2002
2010
8
Patel
MK
,
Dumolard
L
,
Nedelec
Y
, et al
.
Progress toward regional measles elimination - worldwide, 2000-2018
.
MMWR Morb Mortal Wkly Rep.
2019
;
68
(
48
):
1105
1111
9
Omer
SB
,
Salmon
DA
,
Orenstein
WA
,
deHart
MP
,
Halsey
N
.
Vaccine refusal, mandatory immunization, and the risks of vaccine-preventable diseases
.
N Engl J Med.
2009
;
360
(
19
):
1981
1988
10
Feikin
DR
,
Lezotte
DC
,
Hamman
RF
,
Salmon
DA
,
Chen
RT
,
Hoffman
RE
.
Individual and community risks of measles and pertussis associated with personal exemptions to immunization
.
JAMA.
2000
;
284
(
24
):
3145
3150
11
Salmon
DA
,
Haber
M
,
Gangarosa
EJ
,
Phillips
L
,
Smith
NJ
,
Chen
RT
.
Health consequences of religious and philosophical exemptions from immunization laws: individual and societal risk of measles
.
JAMA.
1999
;
282
(
1
):
47
53
12
California Department of Public Health
.
2015–2016 Kindergarten Immunization Assessment
.
2016
.
13
Carrel
M
,
Bitterman
P
.
Personal belief exemptions to vaccination in California: a spatial analysis
.
Pediatrics.
2015
;
136
(
1
):
80
88
14
Richards
JL
,
Wagenaar
BH
,
Van Otterloo
J
, et al
.
Nonmedical exemptions to immunization requirements in California: a 16-year longitudinal analysis of trends and associated community factors
.
Vaccine.
2013
;
31
(
29
):
3009
3013
15
Yang
YT
,
Barraza
L
,
Weidenaar
K
.
Measles outbreak as a catalyst for stricter vaccine exemption legislation
.
JAMA.
2015
;
314
(
12
):
1229
1230
16
Buttenheim
AM
,
Jones
M
,
Mckown
C
,
Salmon
D
,
Omer
SB
.
Conditional admission, religious exemption type, and nonmedical vaccine exemptions in California before and after a state policy change
.
Vaccine.
2018
;
36
(
26
):
3789
3793
17
Delamater
PL
,
Leslie
TF
,
Yang
YT
.
Change in medical exemptions from immunization in California after elimination of personal belief exemptions
.
JAMA.
2017
;
318
(
9
):
863
864
18
Delamater
PL
,
Pingali
SC
,
Buttenheim
AM
,
Salmon
DA
,
Klein
NP
,
Omer
SB
.
Elimination of nonmedical immunization exemptions in California and school-entry vaccine status
.
Pediatrics.
2019
;
143
(
6
):
e20183301
19
Gromis
A
,
Liu
KY
.
The emergence of spatial clustering in medical vaccine exemptions following California Senate Bill 277, 2015-2018
.
Am J Public Health.
2020
;
110
(
7
):
1084
1091
20
Delamater
PL
,
Leslie
TF
,
Yang
YT
.
Examining the spatiotemporal evolution of vaccine refusal: nonmedical exemptions from vaccination in California, 2000-2013
.
BMC Public Health.
2018
;
18
(
1
):
458
21
Fine
PE
.
Herd immunity: history, theory, practice
.
Epidemiol Rev.
1993
;
15
(
2
):
265
302
22
Anderson
RM
,
May
RM
,
Anderson
B
.
Infectious Diseases of Humans: Dynamics and Control.
Oxford, UK
:
Oxford University Press
;
1992
.
23
Davis
RM
,
Whitman
ED
,
Orenstein
WA
,
Preblud
SR
,
Markowitz
LE
Hinman
AR
.
A persistent outbreak of measles despite appropriate prevention and control measures
.
Am J Epidemiol.
1987
;
126
(
3
):
438
449
24
Anderson
RM
,
May
RM
.
Vaccination and herd immunity to infectious diseases
.
Nature.
1985
;
318
(
6044
):
323
329
25
Masters
NB
,
Eisenberg
MC
,
Delamater
PL
,
Kay
M
,
Boulton
ML
,
Zelner
J
.
Fine-scale spatial clustering of measles nonvaccination that increases outbreak potential is obscured by aggregated reporting data
.
Proc Natl Acad Sci USA.
2020
;
117
(
45
):
28506
28514
26
Viboud
C
,
Boëlle
P-Y
,
Cauchemez
S
, et al
.
Risk factors of influenza transmission in households
.
Br J Gen Pract.
2004
;
54
(
506
):
684
689
27
California Department of Health Services
.
2014–2015 Childcare immunization assessment results
.
2015
.
28
Centers for Disease Control and Prevention
.
Recommended Immunization Schedule for Children and Adolescents Aged 18 Years or Younger, United States, 2017
.
2017
.
29
Omer
SB
,
Pan
WKY
,
Halsey
NA
, et al
.
Nonmedical exemptions to school immunization requirements: secular trends and association of state policies with pertussis incidence
.
JAMA.
2006
;
296
(
14
):
1757
1763
30
Liu
F
,
Enanoria
WTA
,
Zipprich
J
, et al
.
The role of vaccination coverage, individual behaviors, and the public health response in the control of measles epidemics: an agent-based simulation for California
.
BMC Public Health.
2015
;
15
:
447
31
Hethcote
HW
.
The mathematics of infectious diseases
.
SIAM Rev.
2000
;
42
(
4
):
599
653
32
Sinclair
DR
,
Grefenstette
JJ
,
Krauland
MG
, et al
.
Forecasted size of measles outbreaks associated with vaccination exemptions for schoolchildren
.
JAMA Netw Open.
2019
;
2
(
8
):
e199768
33
Blumberg
S
,
Worden
L
,
Enanoria
W
, et al
.
Assessing measles transmission in the United States following a large outbreak in California
.
PLoS Curr.
2015
;
7
:
ecurrents.outbreaks.b497624d7043b1aecfbfd3dfda3e344a
34
Gastañaduy
PA
,
Funk
S
,
Lopman
BA
, et al
.
Factors associated With measles transmission in the United States during the postelimination era
.
JAMA Pediatr.
2020
;
174
(
1
):
56
62
35
McDonald
R
,
Ruppert
PS
,
Souto
M
, et al
.
Notes from the field: measles outbreaks from imported cases in Orthodox Jewish communities - New York and New Jersey, 2018-2019
.
MMWR Morb Mortal Wkly Rep.
2019
;
68
(
19
):
444
445
36
Mohanty
S
,
Buttenheim
AM
,
Joyce
CM
,
Howa
AC
,
Salmon
D
,
Omer
SB
.
California’s Senate Bill 277: local health jurisdictions’ experiences with the elimination of nonmedical vaccine exemptions
.
Am J Public Health.
2019
;
109
(
1
):
96
101
37
Mohanty
S
,
Buttenheim
AM
,
Joyce
CM
,
Howa
AC
,
Salmon
D
,
Omer
SB
.
Experiences with medical exemptions after a change in vaccine exemption policy in California
.
Pediatrics.
2018
;
142
(
5
):
e20181051
38
May
T
,
Silverman
RD
.
‘Clustering of exemptions’ as a collective action threat to herd immunity
.
Vaccine.
2003
;
21
(
11-12
):
1048
1051
39
Salathé
M
,
Bonhoeffer
S
.
The effect of opinion clustering on disease outbreaks
.
J R Soc Interface.
2008
;
5
(
29
):
1505
1508
40
Truelove
SA
,
Graham
M
,
Moss
WJ
,
Metcalf
CJE
,
Ferrari
MJ
,
Lessler
J
.
Characterizing the impact of spatial clustering of susceptibility for measles elimination
.
Vaccine.
2019
;
37
(
5
):
732
741
41
Centers for Disease Control and Prevention
.
COVID data tracker: COVID-19 integrated county view
.
2021
.
Available at: https://covid.cdc.gov/covid-data-tracker/#county-view. Accessed September 1, 2021
42
Murthy
BP
,
Sterrett
N
,
Weller
D
, et al
.
Disparities in COVID-19 vaccination coverage between urban and rural counties - United States, December 14, 2020-April 10, 2021
.
MMWR Morb Mortal Wkly Rep.
2021
;
70
(
20
):
759
764
43
Hughes
MM
,
Wang
A
,
Grossman
MK
, et al
.
County-level COVID-19 vaccination coverage and social vulnerability - United States, December 14, 2020-March 1, 2021
.
MMWR Morb Mortal Wkly Rep.
2021
;
70
(
12
):
431
436
44
Yang
W
.
Transmission dynamics of and insights from the 2018-2019 measles outbreak in New York City: a modeling study
.
Sci Adv.
2020
;
6
(
22
):
eaaz4037
45
Buttenheim
AM
,
Sethuraman
K
,
Omer
SB
,
Hanlon
AL
,
Levy
MZ
,
Salmon
D
.
MMR vaccination status of children exempted from school-entry immunization mandates
.
Vaccine.
2015
;
33
(
46
):
6250
6256
46
Perisic
A
,
Bauch
CT
.
Social contact networks and disease eradicability under voluntary vaccination
.
PLOS Comput Biol.
2009
;
5
(
2
):
e1000280
47
Yang
YT
,
Silverman
RD
.
Legislative prescriptions for controlling nonmedical vaccine exemptions
.
JAMA.
2015
;
313
(
3
):
247
248
48
Valente
TW
.
Network interventions
.
Science.
2012
;
337
(
6090
):
49
53
49
Bond
RM
,
Fariss
CJ
,
Jones
JJ
, et al
.
A 61-million-person experiment in social influence and political mobilization
.
Nature.
2012
;
489
(
7415
):
295
298
50
Kulldorff
M
.
Spatial scan statistics: models, calculations, and applications
. In:
Glaz
J
,
Balakrishnan
N
, eds.
Scan Statistics and Applications.
Statistics for Industry and Technology
.
Birkhäuser, Boston, MA
;
1999
:
303
322

Competing Interests

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

Supplementary data