TY - GEN
T1 - Using Wearable Photoplethysmography for Detecting Atrial Fibrillation in Ambulatory Conditions
AU - Halkola, T.
AU - Yli-Mäyry, S.
AU - Nikus, K.
AU - Vehkaoja, A.
PY - 2023
Y1 - 2023
N2 - Atrial fibrillation is the most common sustained cardiac arrhythmia, and its occurrence will increase in the following decades due to aging of the population. As more resources will be needed to treat atrial fibrillation, more efficient detection methods than the current obtrusive 1–3-day Holter electrocardiogram are needed. A promising option is to use a photoplethysmography device worn on the wrist for an extended period to have an unobtrusive patient experience. In this study, an optical method for atrial fibrillation detection is presented. Thirty patients suspected of having atrial fibrillation wore an optical device on their wrist for up to 48 hours. Proprietary algorithms were used to detect atrial fibrillation from the recorded photoplethysmography data. This method was then compared to the gold standard electrocardiogram recorded at the same time. Post-processing was done in 5-minute and 1-minute segments. A 97.9% accuracy, 94.1% sensitivity, and 98.1% specificity in detecting atrial fibrillation were obtained for the 5-minute segment length with 39.3% of the segments marked as undetermined and discarded from the analysis. The results show that the performance of wrist photoplethysmography is adequate for the screening of atrial fibrillation in ambulatory conditions.
AB - Atrial fibrillation is the most common sustained cardiac arrhythmia, and its occurrence will increase in the following decades due to aging of the population. As more resources will be needed to treat atrial fibrillation, more efficient detection methods than the current obtrusive 1–3-day Holter electrocardiogram are needed. A promising option is to use a photoplethysmography device worn on the wrist for an extended period to have an unobtrusive patient experience. In this study, an optical method for atrial fibrillation detection is presented. Thirty patients suspected of having atrial fibrillation wore an optical device on their wrist for up to 48 hours. Proprietary algorithms were used to detect atrial fibrillation from the recorded photoplethysmography data. This method was then compared to the gold standard electrocardiogram recorded at the same time. Post-processing was done in 5-minute and 1-minute segments. A 97.9% accuracy, 94.1% sensitivity, and 98.1% specificity in detecting atrial fibrillation were obtained for the 5-minute segment length with 39.3% of the segments marked as undetermined and discarded from the analysis. The results show that the performance of wrist photoplethysmography is adequate for the screening of atrial fibrillation in ambulatory conditions.
U2 - 10.22489/CinC.2023.384
DO - 10.22489/CinC.2023.384
M3 - Conference contribution
T3 - Computing in Cardiology
BT - 2023 Computing in Cardiology (CinC)
PB - Computing in Cardiology
T2 - Computing in Cardiology
Y2 - 1 October 2023 through 4 October 2023
ER -