Statistical distance: A promising measure of physiological dysregulation in biodemographic studies

Many biodemographic studies incorporate biomarkers of physiological dysregulation and/or allostatic load. However, it is not clear how well these measures represent a global process of dysregulation, nor whether the various biomarkers incorporated into summary measures have been appropriately weighted or considered. Here we introduce a measure based on statistical distance – how common or rare a combination of variables is, according to its multivariate distribution – as an alternative approach to measuring physiological dysregulation. The measure is based on the principle that the more “strange” an individual’s biomarker profile is relative to a baseline population (used to define the distribution, ideally composed of young, healthy individuals), the more dysregulation is occurring. We use three examples to show strengths and limitations of the method. The biomarkers included must be chosen carefully (based on statistical or biological criteria) and the interpretation should be made in light of the variables included. Suites of appropriately chosen variables can help illuminate underlying physiology as well as serve as useful proxies for both mortality risk and biological age. Individual trajectories of statistical distance and in multivariate space are extensions that may help in understanding mortality and disease risk.