Two-Minute Home Wearable Vital-Sign Monitoring Around Thoracic Surgery: A Single-Participant Technical Feasibility and Data Quality Analysis

Background: High-frequency home wearable monitoring may offer longitudinal physiological recordings, but timestamp availability, signal-specific technical usability, and the impact of temporal subsampling require characterization before clinical interpretation.Methodology: This single-participant methodological study examined a man in his 70s who underwent thoracoscopic right upper-lobe wedge resection.A nonmedical wrist-worn monitoring system (Vital Voice VX-2; Vital Voice Co., Ltd., Okayama, Japan) recorded heart rate (HR), peripheral oxygen saturation (SpO₂), systolic blood pressure (SBP), and diastolic blood pressure (DBP) during seven preoperative and 28 postoperative home-monitoring days, separated by hospitalization.Wearing was permitted as tolerated, without a fixed daily target.The primary objective was to quantify timestamp availability and signal-specific technically usable data yield using a hierarchical two-minute epoch framework.A retrospective exploratory secondary analysis examined the retention of operationally defined sustained episodes after four-and tenminute subsampling.Results: Among 25,200 planned epochs, 22,226 were timestamped (88.2% timestamp availability).Technically valid data yield was 96.7% for HR and the BP pair, but 46.7% for SpO₂; concurrent validity across all four vital signs was 46.6% of timestamped epochs.Zero-value outputs accounted for most technically invalid observations, particularly for SpO₂.There were 166 threshold-defined abnormal vital-direction epochs (four HR-high, eight HR-low, and 154 SpO₂-low) and 12 sustained episodes (one HR-high and 11 SpO₂-low).Under the specified retention criterion, four-minute subsampling retained all reference episodes, whereas ten-minute subsampling retained a median of nine of 11 SpO₂-low episodes across offsets (range, 8-11).Conclusions: A hierarchical framework characterized recording availability and signal-specific data usability in this single-participant home-monitoring dataset.The exploratory findings on episode retention depended on operational definitions and sampling phase.The results do not establish device accuracy, patient adherence, or clinical event-detection performance.