Use of Device Data to Predict Frailty in Individuals

Technology is providing new opportunities to gain insight into patient health in a manner far superior to what can be obtained from routine office visits. For geriatric patients, frailty is one such measure, which has been shown to have a negative correlation with outcomes, including hospitalizations, complications following invasive procedures, and overall mortality. Methods to quantify frailty have been highly limited and somewhat controversial, largely because of their cross-sectional nature. Efforts to obtain longitudinal measures of frailty using activity monitors and other wearable devices have been promising, although they are limited by patient cooperation and cost. In addition to their therapeutic functions, cardiovascular implantable electrical devices (CIEDs) collect large amounts of data on individuals in whom they are implanted, which can be accessed during office visits as well as remotely. Importantly, the data is obtained passively and routinely from all patients, requiring little to no additional effort or expense. The possibility that activity measures from CIEDs could be used to evaluate frailty, as well as other outcomes, would provide a powerful tool for clinicians to apply data science directly to patients. This pilot investigation will assess the feasibility of using CIED-derived activity data to predict frailty, and other outcomes, in older populations. The investigators plan to enroll subjects in whom CIEDs are implanted from the University of Colorado Hospital Device clinic, and analyze their data to assess the correlation of daily activity with frailty measures, as well as with cardiac and clinical outcomes.