Indices of Cumulative Deficits

Despite broad interest in the mechanisms responsible for human aging and numerous efforts to identify factors contributing to morbidity, biological senescence, and longevity, these processes still remain elusive. This makes the systemic description of aging-related changes embedded in data from different studies a difficult task. Indeed, observational studies typically measure not only major changes in health and well-being captured by well-defined risk factors (e.g., physiological measurements), but also various aging-related changes spread throughout hundreds of distinct variables. The connection between such variables as well as between each of these variables and health or survival outcomes is unclear and often cannot be evaluated statistically with acceptable accuracy. This is due to the fact that the number of these variables is typically large, while the effect of each on health and survival is small, so most estimates of effect parameters in corresponding statistical models are statistically non-significant. This chapter describes a line of analysis that is based on the premise that, by taking such “mild-effect” variables into account, the description of aging-related deterioration in health and well-being in humans can be substantially improved without costly investments in collecting new data. To realize this potential, new statistical methods are required.