JMIR Publications

ABSTRACT

Background:

Biological sex is an important consideration in biomedical research, yet females are still underrepresented in both human and animal biomedical research. Hesitancy to include female subjects is partially due to the hypothesis that biological rhythms driven by menstrual cycles, and occurring on the timescale of roughly 28 days, increase biological variability and weaken statistical power.

Objective:

We aimed to determine if variability of physical activity (PA) is affected by biological sex, and if so, whether having menstrual cycles (as indicated by temperature rhythms) contributes to increased female PA variability. We then sought to compare the effect of sex and menstrual cycles on PA variability to the effect of PA rhythms on the timescales of days and weeks and to the effect of non-rhythmic temporal structure in PA on the timescale of decades of life (age).

Methods:

We used minute-level metabolic equivalent task (MET) data collected using a wearable device across a 206-day study period for each of 596 individuals as an index of physical activity (PA) to assess the magnitudes of variability in PA accounted for by biological sex and temporal structure on different timescales. We represented intraindividual variability in PA with consecutive disparity index (CDI).

Results:

Females (regardless of whether they had menstrual cycles) demonstrated lower intraindividual variability in PA than males (Kruskal-Wallis, H=29.51, P<.001). Furthermore, people with menstrual cycles did not have greater intraindividual variability than people without menstrual cycles (Kruskal-Wallis, H=0.54, P=.46). PA rhythms differed at the weekly timescale: individuals with increased or decreased PA on weekends had larger intraindividual variability (Kruskal-Wallis, H=10.13, P=.001). Additionally, intraindividual variability differed by decade of life, with older age groups tending to have less variability in PA (Kruskal-Wallis, H=40.55, P=1x10-7, Bonferroni corrected significance threshold for 15 comparisons: P=3x10-3). A generalized additive model (GAM) predicting CDI of 24-hour MET sums (variability of PA) showed that sex, age, and weekly rhythm accounted for only 11% of PA variability.

Conclusions:

The exclusion of people from biomedical research based on their biological sex or the presence of menstrual cycles is not supported by our analysis. Menstrual cycles did not significantly affect female PA variability. Temporal structures in PA on other timescales had significant effects on both female and male PA. Our findings highlight the potential for emerging longitudinal data sources to allow for phenotyping of individuals by their temporal structure on relevant timescales. This may improve precision in statistical and machine learning models as an alternative to excluding any groups.