Using the Fitbit for early detection of Infection and reduction of healthcare utilization after Discharge in Pediatric Surgical Patients

PROJECT SUMMARY
Pediatric appendectomy, the most prevalent inpatient procedure in children, is associated with significant
burden to the patient, their parents, healthcare systems and third party payors. After discharge, monitoring by
parents consists only of such “proxy” subjective assessments, which have been reported as inaccurate, and
resulted in both increased complications (e.g., readmissions), and wasted healthcare resources (e.g.,
potentially avoidable emergency department (ED) visits after surgery). Advances in consumer wearable
devices (“CWDs”) that passively and non-invasively monitor physical activity (PA), heart rate (HR), and sleep
are ushering in a new era of symptoms science, particularly after surgery. Their expanding capability to
generate continuous, valid, objective, and actionable measures in near-real time in children, provide
opportunities to detect altered post-operative recovery patterns early, and therefore improve the precision and
timeliness of any necessary clinical interventions. The proposed study will use a CWD, the Fitbit Inspire 2, and
will apply machine learning methods to the Fitbit data (physical activity, HR, and sleep) to create clinically
meaningful alerts for early detection of postoperative infection. During hospitalization and continuing after
discharge, a Fitbit Inspire 2, a widely-used, commercially wearable device well-tolerated by young children (3-
18 years old) will be used to measure step counts, sleep, and HR. The proposal has 2 aims. Aim 1 develops
and validates machine learning algorithm for infection using the Fitbit. Aim 2 prospectively feeds near-real time
Fitbit data on postoperative appendectomy patients to clinicians, and examines their effect on clinical decision
making, time to first contact with the healthcare system, and on overall healthcare use patterns. The proposal
is aligned with NINR’s research priorities. Methods developed from this work will pave the way to develop
similar algorithms for other patient populations needing a proxy, as well as to characterize other surgeries and,
should improve overall postoperative management for all surgical patients.

Grant Number: 7R01NR020918-03
NIH Institute/Center: NIH
Principal Investigator: FIZAN ABDULLAH