1 September 2016

I was glued to my television this August watching the Rio Olympics.  Beyond the medal counts, I enjoyed the personal vignettes about the athletes.  Equally compelling to me were stories about the contribution of science to sports.  I was amazed that experts in physics and computer science can evaluate a swimmer’s stroke and determine how long the athlete should glide underwater.  Who knew that swimsuit design could unduly influence race times and as such, the 2016 Olympian men could not wear full bodysuits?  By contrast, I recall watching the mustachioed swimmer Mark Spitz win gold medal after gold medal in the 1972 Olympics, with not even a swim cap to reduce the drag.

It is clear that simple athletic talent and hard work are no longer sufficient to be competitive at a global level.  Innovation and integration of multiple scientific disciplines has made athletes better, faster and stronger.  Continuous and subtle scientific advancements are all around us, not just in sports.  Easily obvious is our telephone, going from clunky land lines and operator assistance to making international calls from our cell phones.  Agriculture has also changed.  Farmers have gone from high-energy practices to no-till farming and have adopted the mantra, “every drop on the crop.”  Sadly, I have seen less innovation in my own discipline, epidemiology.

Epidemiologists have been looking at correlations between agricultural activities and disease for many years.  In the early 1970s and 1980s, occupational groups such as crop protection product applicators and manufacturers were studied with a focus on cancer.  Other approaches relied upon study participants to self-report products used and jobs held.  The findings of these studies were marked by inconsistencies.  In other words, the link between disease and a crop protection product identified by one investigation would not be confirmed in another.

Scientists asked themselves why these studies didn’t agree.  One important limitation was the exposure assessment.  Are all sprayers equally exposed when applying an insecticide?  To what degree do boots, gloves and coveralls offer protection?  How much does the application method affect exposure, be it from a back pack or an enclosed cab tractor?  The U.S. Agricultural Health Study (AHS) was launched in the mid-1990s to study more than 50,000 crop protection product applicators and ask them these questions.  The AHS was a step in the right direction, but it still relied upon questionnaires and the investigator still depended upon the memory of each participant.  Moreover, interviewing individuals is expensive.

What has been the solution?  Agricultural records.  In France, investigators have introduced a Crop Exposure Matrix (CEM) to create an estimate of which crop protection products might be used on a given crop in a given month.  In the U.S., the California Pesticide Use Report (PUR) has been used to match product applications and residence.  As an example, the recent analysis of autism spectrum disorder by Shelton et al.[1] used the PUR to estimate maternal residential exposures.  However, the CEM and PUR approaches fail to incorporate crop protection product information (formulation, drift potential, application method) and weather conditions.  In short, validation of actual internal dose is absent.

Most recently, epidemiological studies have been considered in the risk assessment of crop protection products for potential endocrine disruption.  With increases in the diagnosis of endocrine-related diseases, multiple risk factors are being examined. Well designed and interpreted studies can be important to narrowing down these risk factors and resulting in sound regulatory decisions.  But poorly done studies can jeopardize product registrations, lead to poor policy-making and potentially cause unjustified public concerns.

Epidemiology study findings are being debated and discussed.  These “new” studies and approaches aren’t characterizing exposure any better than their counterparts 40 years ago.  In fact, the assumptions about exposure are probably poorer because they ignore societal and technological changes.  Instead of providing important contributions to public health decisions, epidemiology studies with inadequate exposure measures are introducing controversy related to lack of biological concordance with other studies and poor transparency of results.

To this end, epidemiology research is due for innovation and an evaluation of which results are considered noteworthy, not merely the 5 percent that attain statistical significance.  Epidemiology has leaned heavily on advancements in statistical methods but has not spent the same effort on exposure assessment.  If the exposure data are poor, how can we rely upon the statistical testing?  How can we link exposure and disease?  Epidemiologists need to work more closely with exposure specialists, agronomists, farmers, toxicologists and other experts.

The Olympians at Rio knew that to be competitive they must train hard and make use of all the scientific innovation in their sport.  Epidemiology is not advancing at the same pace.  As a result, this discipline has been called out to innovate for the 21^st century.  Exposure assessment is at the core for that change.  If we want to “swim” competently and be a “team player” on policy decisions, epidemiologists must innovate and collaborate with other scientists.

Carol J. Burns, M.P.H., Ph.D., F.A.C.E., is president of Burns Epidemiology Consulting, LLC in Sanford, Mich. She is a Fellow of the American College of Epidemiology.

[1] Shelton JF, Geraghty EM, Tancredi DJ, Delwiche LD, Schmidt RJ, Ritz B, et al. 2014. Neurodevelopmental disorders and prenatal residential proximity to agricultural pesticides: the CHARGE study. Environ Health Perspect 122(10):1103–1109; doi:10.1289/ehp.1307044.