Protocol for an Observational Study of the Effects of High School Football on Cognition Late in Life Using a New Matching Method

Abstract

A study in JAMA Neurology of the effect of high school football exposure on cognition at age 72 did not detect a significant effect. Using newly available data at the later age of 81, we look for such an effect at an age when cognitive decline is substantially more common. We propose a longitudinal observational study with 1155 high school graduates after exclusion criteria. 426 of these graduates played football in high school and 729 did not. The primary outcome is a combination of two measures collected in the 2020 wave of the Wisconsin Longitudinal Study (WLS). One component is a recently developed measure of cognitive functioning, the “Telephone Interview for Cognitive Status-modified” (TICSm) score. The second component is a consensus diagnosis made after a longer interview of participants who performed poorly in terms of the TICSm score. A new matching method is created in order to construct triples of either one football player and two controls or two football players and one control such that the measured covariates are balanced within triples of each type. Such a design has greater design sensitivity than one which also allows matched pairs. We will use the “aberrant rank statistic” to evaluate cognitive scores outside the normal range. Additionally, we will conduct a secondary analysis on the TICSm score alone. The sensitivity of these analyses to unmeasured biases will be assessed. Finally, we will consider two unaffected outcomes, periodontitis and cancer, in an effort to detect unmeasured biases. This study, to the best of our knowledge, is the first prospective longitudinal study to look at the effects of high school football at such a late age of 81.