Classification of Free-Living Body Posture with ECG Patch Accelerometers: Application to the Multicenter AIDS Cohort Study

• Published in: Statistics in biosciences Vol. 16; no. 1; pp. 25 - 44
• Main Authors: Etzkorn, Lacey H., Heravi, Amir S., Knuth, Nicolas D., Wu, Katherine C., Post, Wendy S., Urbanek, Jacek K., Crainiceanu, Ciprian M.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.04.2024; Springer Nature B.V
• Subjects: Accelerometers; Acquired immune deficiency syndrome; AIDS; Algorithms; Biostatistics; Classification; Clustering; Cohort analysis; EKG; Health Sciences; HIV; Human immunodeficiency virus; Labels; Mathematics and Statistics; Medicine; Physical activity; Position measurement; Posture; Sensitivity analysis; Statistics; Statistics for Life Sciences; Theoretical Ecology/Statistics; Actigraphy; Sedentary behavior; Change point; Static activity; Clustering; Electrocardiogram; Changepoint; Static Activity; Sedentary Behavior
• ISSN: 1867-1764; 1867-1772
• DOI: 10.1007/s12561-023-09377-7

As health studies increasingly monitor free-living heart performance via ECG patches with accelerometers, researchers will seek to investigate cardio-electrical responses to physical activity and sedentary behavior, increasing demand for fast, scalable methods to process accelerometer data. We extend a posture classification algorithm for accelerometers in ECG patches when researchers do not have ground-truth labels or other reference measurements (i.e., upright measurement). Men living with and without HIV in the Multicenter AIDS Cohort study wore the Zio XT® for up to 2 weeks ( n  = 1250). Our novel extensions for posture classification include (1) estimation of an upright posture for each individual without a reference upright measurement; (2) correction of the upright estimate for device removal and re-positioning using novel spherical change point detection; and (3) classification of upright and recumbent periods using a clustering and voting process rather than a simple inclination threshold used in other algorithms. As no posture labels exist in the free-living environment, we perform numerous sensitivity analyses and evaluate the algorithm against labeled data from the Towson Accelerometer Study, where participants wore accelerometers at the waist. On average, 87.1% of participants were recumbent at 4 a.m. and 15.5% were recumbent at 1 p.m. Participants were recumbent 54 min longer on weekends compared to weekdays. Performance was good in comparison to labeled data in a separate, controlled setting (accuracy = 96.0%, sensitivity = 97.5%, specificity = 95.9%). Posture may be classified in the free-living environment from accelerometers in ECG patches even without measuring a standard upright position. Furthermore, algorithms that fail to account for individuals who rotate and re-attach the accelerometer may fail in the free-living environment.