Methods to model and characterize person-level features of the salivary cortisol profile in population research—The Multi Ethnic Study of Atherosclerosis

Many studies have hypothesized that stress and its biological consequences may contribute to disparities in cardiovascular disease. However, our understanding of the most appropriate statistical methods useful to analyze biological markers of stress such as salivary cortisol remains limited. We explore the utility of various statistical methods in modeling the stress response curve, as well as characterizing features of stress response in population based studies. In addition to commonly used summaries such as the cortisol awakening response and the area under the curve, we also propose alternative curve features. We compare the inference and interpretations drawn from these methods, and use data collected as part of the Multi-Ethnic Study of Artherosclerosis (MESA) to illustrate them.