Credit…Charlie Riedel/Associated Press
Does voting spread disease?
Common sense, and a basic understanding of how respiratory viruses spread, make it a reasonable question.
We’ve conducted research on the subject — looking at the past five federal elections (presidential and midterm) — and concluded that past elections have not led to national surges of viral disease.
Our findings shed light on the behavior of respiratory viruses in general, but it’s important to note that they should not be applied to the current coronavirus, a new and unpredictable virus that health officials say is more lethal.
How a virus could spread
Although some sick voters may stay home on Election Day, others will arrive coughing, sneezing and sniffling at their polling location. There, they will write with the same pens and touch the same voting machines as their fellow citizens.
Droplets from coughs or sneezes from an infected person can be left on surfaces and then touched by uninfected people. If those people do not wash their hands thoroughly before touching their eyes, nose or mouth, they can become infected with a virus and then go on to infect others.
Respiratory viruses include influenza as well as those that can cause influenza-like illnesses, such as upper respiratory infections (the common cold), viral sinusitis, viral bronchitis and viral ear infections. Every year, they infect millions and kill thousands of Americans; health officials closely track cases by the week.
How we conducted the experiment
To test the hypothesis that these elections could lead to more viral infections, we used a large national insurance database focusing on the years 2007 to 2016. This database included more than 14.5 million patients over 40, an age group we chose to study because they are more likely to vote.
We calculated rates of influenza infection or influenza-like illness for the 21 days after federal elections, which fall on the first Tuesday of November. We chose a 21-day window to allow enough time for any newly infected patients to fall ill and seek medical care. In years without an election, we calculated infection rates for the 21 days after the first Tuesday of November — the day that would have been Election Day had there been a federal election.
We then grouped the data into years with a presidential election, years with a midterm election, and years with no election. Although state or local elections might have taken place in the years without a federal election, they would have drawn far fewer voters, and so these years made a good group for comparison.
We found that among adults over age 40, rates of doctor visits for viral illnesses gradually increased in the weeks before and after the first Tuesday in November — as we would expect in every cold and flu season — but election years were no different than years without an election.
This suggests that there was no major difference in outpatient visits for viral infections in years with elections years versus those without.
While we observed minor differences between presidential, midterm and nonelection years, they were within the margin of error of our model and are more likely related to varying severity of the overall flu season for the years included in each group.
A number of factors could explain these findings. First, polling locations bring in voters only from a local community; if a virus is not circulating in that community, it cannot be spread at a polling location. Second, voting is a brief individual activity in which the main potential for viral spread is through touching common surfaces. If surfaces are disinfected periodically and voters wash their hands or wear gloves, viral spread could be easily limited.
In contrast, a music festival or international tennis tournament attracts crowds from afar who pack into a crowd, spreading their respiratory droplets as they talk, sing, scream and cheer before returning home.
Up to a third of voters might not have cast their ballots in person on Election Day, instead using absentee balloting or mail-in votes. While respiratory viruses can theoretically survive for days outside the body — such as on a licked mail-in ballot envelope — it stands to reason that in-person voting carries greater potential for person-to-person viral spread than mail-in voting. The election expert Rick Hasen, a professor of law and political science at U.C. Irvine, has recently called on Congress to pass a law making it easy for Americans in all states to cast absentee ballots in November.
Decisions by election officials
Our analysis was designed to detect a large-scale, national effect of voting on spread of viral disease. Person-to-person viral transmission certainly occurs at polling locations. Small but difficult-to-measure outbreaks may also occur, but our statistical model was not designed to detect these, nor can it be extrapolated to coronavirus.
Likewise, our findings in no way refute decisions by public health officials who have recently felt that it is in the public’s best interest — with social distancing crucial at this time — to delay or alter voting procedures during this health crisis. Louisiana, Georgia and Kentucky have announced plans to postpone their primary elections, and other states are considering doing the same. Florida, Ohio, Arizona and Illinois were scheduled to vote today, but late Monday Ohio’s governor succeeded in postponing the primary. Other states have taken steps to make voting safer, including making sure hand sanitizer is available, moving polling places out of assisted-living facilities and making it easier to drop off a vote-by-mail ballot.
Last week, states like Washington and Missouri held elections despite confirmed cases of the coronavirus in their communities, so researchers will have the opportunity to assess any possible impact of elections on its spread.
Christopher M. Worsham, M.D., is a pulmonologist and critical care physician at Harvard Medical School. Follow him on Twitter at @ChrisWorsham. Jaemin Woo is a research analyst at Harvard Medical School. Follow him on Twitter at @_JaeminWoo. Anupam B. Jena, M.D., Ph.D., is an economist, a physician and the Ruth L. Newhouse Associate Professor at Harvard Medical School. Follow him on Twitter at @AnupamBJena.