Data Integration for Health and Stress Monitoring: Biological Metabolites, Wearables Data, and Self-Reporting
Abstract
Integrative and unobtrusive approaches to monitoring health and stress can assist in preventative medicine and disease management, and provide capabilities for complex work environments, such as military deployments and long-duration human space exploration missions. With many data streams that could potentially provide critical information about the health, behavior, and psychosocial states of individuals or small groups, the central question of this research is how to reliably measure health and stress states over time. This integrative approach to health and stress monitoring has implemented biological metabolite profiling, wearables data analysis, and survey assessment for comparing biological, behavioral, and psychological perspectives. Health monitoring technologies aim to provide objective data about health status. Providing objective information can help mitigate biases or blind spots in an individual’s perception. Consider an individual who is unwilling to openly admit to psychosocial distress and unhealthy habits, or an individual who has habituated to long-term stressors and is unable to recognize a chronic state of high stress. Both honesty and self-awareness are required for accurate self-reporting. Digital health technologies, such as wearable devices, provide objective data for health monitoring. Compared to surveys, wearables are less influenced by participant openness, and compared to biological samples, wearables require less equipment and less labor for analysis. However, inherent to every data stream are limitations due to uncertainty and sensitivity. This research has been conducted in collaboration with Hawaii Space Exploration Analog and Simulation (HI-SEAS), which is a Mars analog research site on the slopes on Mauna Loa volcano in Hawaii. During 8-month and 12-month HI-SEAS missions in the 2014-2016 timeframe, twelve individuals provided hair and urine samples for metabolite profiling, utilized consumer-grade wearables to monitor sleep and activity behaviors, and responded to surveys for recording perceived health and stress levels. This work has developed a self-report instrument for stress characterization, efficient protocols for metabolite profiling, novel measures of sleep quality and activity levels, and has evaluated performance differences of Jawbone® and Fitbit® wearable devices that were worn in tandem. There is considerable debate about the accuracy of data collected from wearable devices. Therefore, the success of next-generation wearable devices is hinging on the ability to reliably process wearables data into meaningful health information. By simultaneously quantifying biological metabolites, sleep and activity behaviors, and psychological perceptions of health, this research is evaluating possible predictors of health and stress, such as evaluating if activity and sleep behaviors recorded by wearables can be predictive of biological metabolites and perceived health. This research has developed data-driven insights for advancing the next-generation of biological, behavioral, and psychological health monitoring capabilities.
Degree
Ph.D.
Advisors
Landry, Purdue University.
Subject Area
Aerospace engineering|Health sciences|Behavioral Sciences
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.