by Patricia Salber

“Tax payments influence the long-term health of the economy, but no study has tested how tax deadlines might affect the immediate health of individuals.”  DA Redelmeier

I think I will do the caveats first.  Although the study I am about to describe was published in JAMA (4/11/2012. p. 487), a pretty good peer-reviewed journal, it was published as a “Research Letter” as opposed to a full-blown article.  I don’t really know if the same peer-review process was applied to this letter as are applied to articles.  The lead author, Donald A. Redelmeier, the author with the the most letters after his name (MD, FRCPC, MS(HSR)) seems to like to write about this type of stuff–one of his listed references is titled “Road crashes and the next US presidential election Chance.” A quick perusal of his publications on PubMed, however, show him to be a prolific author and his articles are usually in prestigious journals, such as JAMA, NEJM, and the Journal of Clinical Epidemiology.

So, with that out of the way, what did and his co-author, Christopher J. Yarnell, AB do and what did they find?  First, they got a hold of Internal Revenue Service tax deadline data to determine actual tax days.  Although, most of us think that Tax Day = April 15, in fact the date varies “to accommodate holidays or extreme weather conditions.”  This year, 2012, Tax Day is April 17.

Daniel Redelmeier, MD, Dept of Medicine, Toronto (photo from NY Times)

Next, they examined fatal road crash data from the  National Highway Traffic Safety Administration for the past three decades using the Fatality Analysis Reporting (FAR) System.  The FAR system includes all car crashes  in which someone died within 30 days of the event.  They compared the number of fatalities on Tax Day to the number per day that occurred in either the week before or the week after the dreaded deadline.  Although we don’t know much about the details, they note that  they “controlled for year, month, and weekday as well as minimized bias from differences in vehicle technology, gasoline prices, health care access, driver training, and other confounders.”  The also say that “stratified analyses assessed robustness based on age, sex, religion [religion??], decade, time of day, position, and initial outcome.”

Here is what they found:  There were 6,783 fatalities or 226 per day on Tax Days and 12,758 or 213 per day on non-tax days, an increase of about 13 deaths on an average tax day.  The magnitude of the increase was about 6% – not a huge increase – however 6% of a lot of people is, well, a lot of people. The confidence interval (1.03-1.10) was broad, however the P value for the study was <.001 which is statistically significant.

No matter how the researchers sliced and diced the data (e.g., by region of the country, decade, time of day, age or sex of deceased, position of the deceased (driver, passenger, passenger) or initial outcome (dead or alive), the risk of dying on Tax Day was higher compared to the control days (described above) in all but two parameters (the decade of the 80s and age >65). In those two parameters,  the odds of dying in a car crash on tax day was more or less equal to a non-tax day. The increase in relative risk was highest in the last decade (2000-2009) and in people less than 65.

The authors did not examine why there are more fatal crashes on Tax Day compared to the days immediately preceeding or following it because they did not have data on alcohol, stress, or driving patterns. The study notes that a Gallup Poll showed that Tax Day was the second most stressful day of 2011 – the most stressful day being April 27, “when one of the largest tornado outbreaks ever recorded tore through the southeastern U.S., killing more than 300 people and decimating entire communities.”

Overall, I am not really sure what to make out of this little research paper, but I was pretty sure that you might want to read it before April 17.  Happy Tax Day, everyone.