When the next outbreaks hit, they’ll likely be in Texas (2019 edition)

I’ve been writing a long time about how Texas is overdue for massive measles outbreaks on par with what we are observing in Brooklyn and and Rockland Counties in New York. The reason is that the antivaccine movement in Texas has been particularly influential, fusing as it has antivaccine conspiracy mongering with conservative, anti-government, -antiregulation beliefs, making appeals to “parental rights” and “health freedom,” particularly effective there. As I’ve said before, “parental rights,”
“health freedom,” and anti-government beliefs combine to produce a highly effective gateway drug to antivaccine pseudoscience, quackery, and conspiracy theories. Indeed, the number of children whose parents claim “personal belief exemptions” (or, as I like to call them, “I don’t wanna” exemptions) to school vaccine mandates has skyrocketed since 2003.

That’s why Dr. Peter Hotez and others have been warning of potential disaster in Texas for a while now. Now, a new study from the University of Pittsburgh and Texas Pediatric Society, the Texas Chapter of the American Academy of Pediatrics, Austin, TX: Forecasted Size of Measles Outbreaks Associated With Vaccination Exemptions for Schoolchildren. The study, whose lead author is David Sinclair, is hot off the presses last week in JAMA Network Open. The key question asked was: What is the expected size of measles outbreaks in Texas at current (i.e., 2018) and decreased vaccination rates? Mathematical modeling was used to estimate the expected size of measles outbreaks in Texas under various scenarios, taking into account how individual cases could spread based on the activities and personal contacts of the infected individuals, noting that the number of personal belief exemptions in Texas has increased 28-fold, from 2,300 in 2003 to 64,000 in the 2017-2018. I note that this means that the personal belief exemption rate has climbed from 45,000 since the 2015-2016 school year.

The authors note in the introduction:

Efforts to achieve or maintain herd immunity have been hampered by a small segment of the population declining vaccinations for their children for various reasons, including concerns regarding adverse effects of vaccination, lack of knowledge of the vaccine, and social influences.13,14 These concerns are not homogeneously spread in the population, creating spatial foci with greater risk of measles outbreaks.15,16 Because of these fears increasing in the past 2 decades, vaccine declination disproportionately affects children. A 5% decrease in vaccination rates has been estimated to cause a tripling of measles cases in children aged 2 to 11 years.17

Schools with a large number of students with vaccine exemptions provide environments in which measles can spread among susceptible students and to the wider population.18 The merits of requiring children to be vaccinated to attend schools have been debated extensively,19 with requirements and incentives varying around the world.20 In the United States, states individually choose whether to grant vaccination exemptions to allow unvaccinated children to attend school. As of June 2019, exemptions for religious or personal reasons are permissible in 45 states.21

The authors note that Texas is the second largest state in the US by population, with metropolitan statistical areas (MSAs) encompassing a range of sizes representing 92% of the MSAs in the US, before proposing to do the following:

We use agent-based simulations of the Texas population, including the size, location, and vaccination rate of each school, to evaluate the risk of widespread measles outbreaks occurring within and beyond Texas schools. We analyze the risk of outbreaks at 2018 vaccination rates (from the 2017-2018 school year) and the risk if vaccination rates continue to decrease. Advocacy groups in Texas have argued both in favor of and against changes to vaccine exemption regulations in Texas31-33; this investigation aims to help inform such discussions. Early (unpublished) versions of these simulations directly influenced legislative votes in favor of the elimination of conscientious exemptions to vaccination in California.34 Simulations were run for each Texas MSA (except Texarkana). Texas counties not located in an MSA are not discussed in this article; however, the results of similar simulations are provided elsewhere.35

The authors developed an agent-based decision model using a tool called Framework for Reconstructing Epidemiological Dynamics (FRED). (I kind of like that the name of the tool is Fred. I don’t know why.) Agent-based models can simulate the actions and interaction of a collection of individual agents that follow a set of rules and thus allow the modeling of the behavior of agents in a complex, interconnected system when various parameters and rules are changed. The authors note that the FRED tool can model the spread of infectious diseases through a population, based on the daily interactions of agents in their households, neighborhoods, and schools and workplaces.

The agents in this particular set of simulations were made up of a synthetic population of Texas generated from 2010 US Census data, including age, gender, race, household size, and household income. Each member of the synthetic population was assigned a household location according to US Census tract populations and demographic distributions, and agents were assigned workplaces corresponding to commuting patterns, business sizes, and employment rates. Finally, each school-aged agent was assigned a school type (public or private) and school location on the basis of their demographic characteristics. The schools in the synthetic population were created using real school locations and sizes. The agent-based model assumes that, during each work day, agents interact with the other agents in their household, neighborhood, and workplace or school.

Vaccination rates were assigned on the basis of the reported vaccination rates of the schools attended by agents in the model. Because the two-dose series of MMR is 97% effective in preventing infection, a random 3% of each school’s population was assigned to be susceptible to measles. Approximately 0.2% of students in Texas are ineligible to be vaccinated because of contraindications46; medical exemptions to vaccination were assigned randomly to agents in schools.

Finally:

People born in the United States before 1957 are assumed to be immune to measles because they were exposed before mass vaccination, when measles was widespread.47 Agents representing population members aged 62 years or older (the age of someone born before 1957 on January 1, 2019) were, thus, immune to measles in these simulations. The vaccination rate of the remaining population was assumed to be 94.8%, on the basis of antibody seroprevalence analysis of the US population48 and assuming 97% effectiveness of administered vaccinations.

Vaccination rates for each school were obtained from the Texas Department of State Health Services.23 The most recent available vaccination rate for each school was used (from the 2017-2018 to 2015-2016 school years). Vaccination rates for schools with no data available were estimated using the vaccination rates of similar, nearby schools (eAppendix 2 and eFigure 1 in the Supplement). Private schools in Texas report vaccination rates on a per-school basis; however, public schools report by school district, so the vaccination rate of a school district was applied to each public school in the district. Seventh-grade (ages 12-13 years) vaccination rates were used (eAppendix 3 in the Supplement).

And…enter a single case of measles:

In each simulation run, a single case of measles was introduced via a randomly selected student (aged 5-15 years) for whom a vaccine had been refused. Measles outbreaks were simulated for the 2018 vaccination rates at each school, and further scenarios in which the vaccination rate of schools with vaccine refusers was decreased by 1% to 10%. Approximately 0.2% of students in Texas are reported as being medically exempt from vaccination. We assume an uncertainty of 0.2% on this value; therefore, any school with less than 99.6% of students vaccinated was considered to have vaccine refusers when deciding which schools’ vaccination rates to decrease.

This leads to the money figure, which shows an exponential increase in measles cases in six MSAs, including the four largest MSAs in Texas (Austin–Round Rock, Dallas–Fort Worth–Arlington, Houston–The Woodlands–Sugar Land, and San Antonio) and two with more typical population sizes (Lubbock and Tyler):

Note that this is a logarithmic scale. Each major tick mark represents an increase by a factor of ten in the number of measles cases based on percent decline in vaccination coverage. Also note that it’s not just the unvaccinated who are at risk. Bystanders (those with medical exemptions and the 3% of the vaccinated who remain susceptible to measles) are also at risk. Even though this modeling predicts that 64% of the measles cases would occur among children for whom vaccination was refuse, that still means 36% occur bystanders who did not refuse vaccination or the small number of children with legitimate medical exemptions to school vaccine requirements.

The authors note in the conclusion:

These simulations estimate that large measles outbreaks of more than 400 cases could occur in Austin–Round Rock and Dallas–Fort Worth–Arlington, with their 2018 distribution of vaccination rates. This finding is consistent with the largest measles outbreaks that have occurred since measles was eliminated in the United States9,10 and suggests that the vaccination rate of these areas should be increased to reduce the chance of a large outbreak.

These simulations estimate that large measles outbreaks of more than 400 cases could occur in Austin–Round Rock and Dallas–Fort Worth–Arlington, with their 2018 distribution of vaccination rates. This finding is consistent with the largest measles outbreaks that have occurred since measles was eliminated in the United States9,10 and suggests that the vaccination rate of these areas should be increased to reduce the chance of a large outbreak.

The Tyler MSA simulations suggest a significant chance of large measles outbreaks associated with students attending 2 schools with vaccination rates of 70% and 85% (eAppendix 6 in the Supplement). This highlights that a small number of significantly undervaccinated networks could be associated with measles spreading widely in a population. In the event of an outbreak in schools with undervaccinated populations, interventions targeted at these schools may be especially effective. Such interventions may involve isolation measures or mandatory vaccinations, as happened in New York State in 2019,12 or voluntary vaccination programs,18 if vaccination acceptance increases during an outbreak. Undervaccinated close-knit communities present an increased risk of outbreaks; mandating that schools with low vaccination rates plan for outbreak scenarios may help reduce outbreak sizes.

Also, the lower the vaccination rate, the larger the chance of outbreaks and the larger the size of outbreaks. The authors also found that decreases in vaccination rates in schools with undervaccinated populations in 2018 were associated with exponential increases in the potential size of outbreaks. Indeed, even a 5% decrease in vaccination rate was associated with a 40% to 4000% increase in potential outbreak size, depending on the metropolitan area.

As I’ve said many times before, when the next outbreaks occur, they’ll probably happen in Texas. In fact, I’m rather surprised that Texas has managed to get away with this situation for as long as it has without suffering large outbreaks. Moreover, by refusing to vaccinate their children based on fears based on pseudoscience and conspiracy theories, antivaxers endanger not just their own children, but all children.