Heart rate recovery (HRR) after physical exercise is a convenient method to assess cardiovascular autonomic function. Since stair climbing is a common daily activity, usually followed by a slow walking or rest, this type of activity can be considered as an alternative HRR test. The present study explores the feasibility to estimate HRR parameters after stair climbing using a wrist-worn device with embedded photoplethysmography and barometric pressure sensors. A custom-made wrist-worn device, capable of acquiring heart rate and altitude, was used to estimate the time-constant of exponential decay $\tau$ , the short-term time constant $\mathcal{S}$ , and the decay of heart rate in 1 min $\mathcal{D}$ . Fifty-four healthy volunteers were instructed to climb the stairs at three different climbing rates. When compared to the reference electrocardiogram, the absolute and percentage errors were found to be ≤ 21.0 s (≤ 52.7%) for $\tau$ , ≤ 0.14 (≤ 19.2%) for $\mathcal{S}$ , and ≤ 7.16 bpm (≤ 20.7%) for $\mathcal{D}$ in 75% of recovery phases available for analysis. The proposed approach to monitoring HRR parameters in an unobtrusive way may complement information provided by personal health monitoring devices (e.g., weight loss, physical activity), as well as have clinical relevance when evaluating the efficiency of cardiac rehabilitation program outside the clinical setting.

Exercise testing to assess the response to physical rehabilitation or lifestyle interventions is administered in clinics thus at best can be repeated only few times a year. This study explores a novel approach to collecting information on functional performance through walk tests, e.g., a 6-min walk test (6MWT), unintentionally performed in free-living activities. Walk tests are detected in step data provided by a wrist-worn device. Only those events of minute-to-minute variation in walking cadence, which is equal or lower than the empirically determined maximal SD (e.g., 5-steps), are considered as walk test candidates. Out of detected walk tests within the non-overlapping sliding time interval (e.g., 1-week), the one with the largest number of steps is chosen as the most representative. This approach is studied on a cohort of 99 subjects, assigned to the groups of patients with cardiovascular disease (CVD) and healthy subjects below and over 40-years-old, who were asked to wear the device while maintaining their usual physical activity regimen. The total wear time was 8,864 subject-days after excluding the intervals of occasionally discontinued monitoring. About 82% (23/28) of patients with CVD and 88% (21/24) of healthy subjects over 40-years-old had at least a single 6MWT over the 1st month of monitoring. About 52% of patients with CVD (12/23) and 91% (19/21) of healthy subjects over 40-years-old exceeded 500 m. Patients with CVD, on average, walked 46 m shorter 6MWT distance (p = 0.04) compared to healthy subjects. Unintentional walk testing is feasible and could be valuable for repeated assessment of functional performance outside the clinical setting.

Due to frailty, cardiac rehabilitation in older patients after open-heart surgery must be carefully tailored, thus calling for informative and convenient tools to assess the effectiveness of exercise training programs. The study investigates whether heart rate (HR) response to daily physical stressors can provide useful information when parameters are estimated using a wearable device. The study included 100 patients after open-heart surgery with frailty who were assigned to intervention and control groups. Both groups attended inpatient cardiac rehabilitation however only the patients of the intervention group performed exercises at home according to the tailored exercise training program. While performing maximal veloergometry test and submaximal tests, i.e., walking, stair-climbing, and stand up and go, HR response parameters were derived from a wearable-based electrocardiogram. All submaximal tests showed moderate to high correlation ( $r$ = 0.59–0.72) with veloergometry for HR recovery and HR reserve parameters. While the effect of inpatient rehabilitation was only reflected by HR response to veloergometry, parameter trends over the entire exercise training program were also well followed during stair-climbing and walking. Based on study findings, HR response to walking should be considered for assessing the effectiveness of home-based exercise training programs in patients with frailty.

Frailty in patients after open-heart surgery influences the type and intensity of a cardiac rehabilitation program. The response to tailored exercise training can be different, requiring convenient tools to assess the effectiveness of a training program routinely. The study aims to investigate whether kinematic measures extracted from the acceleration signals can provide information about frailty trajectories during rehabilitation. One hundred patients after open-heart surgery, assigned to the equal-sized intervention and control groups, participated in exercise training during inpatient rehabilitation. After rehabilitation, the intervention group continued exercise training at home, whereas the control group was asked to maintain the usual physical activity regimen. Stride time, cadence, movement vigor, gait asymmetry, Lissajous index, and postural sway were estimated during the clinical walk and stair-climbing tests before and after inpatient rehabilitation as well as after home-based exercise training. Frailty was assessed using the Edmonton frail scale. Most kinematic measures estimated during walking improved after rehabilitation along with the improvement in frailty status, i.e., stride time, cadence, postural sway, and movement vigor improved in 71%, 77%, 81%, and 83% of patients, respectively. Meanwhile, kinematic measures during stair-climbing improved to a lesser extent compared to walking. Home-based exercise training did not result in a notable change in kinematic measures which agrees well with only a negligible deterioration in frailty status. The study demonstrates the feasibility to follow frailty trajectories during inpatient rehabilitation after open-heart surgery based on kinematic measures extracted using a single wearable sensor.

Background: A wrist-worn wearable device for acquiring limb and chest ECG leads (wECG) may constitute a promising approach to detection of acute myocardial infarction (AMI). However, it remains to be demonstrated whether the information conveyed by the wECG is sufficient for AMI detection. Objective: To explore explainable machine learning models for detecting AMI using the wECG. Methods: Two types of machine learning models are explored: a convolutional neural network (CNN) using the raw ECG as input and a gradient-boosting decision tree (GBDT) using clinically informative features. 123 participants were included, divided into patients with AMI, patients with other cardiovascular diseases, and healthy individuals. A wrist-worn device equipped with three biopotential electrodes was used to acquire two ECG leads with a single touch: limb lead I and another lead involving a specific body site, i.e., either the V3 or V5 electrode positions, or the abdomen. Results: The best performance on the test dataset is obtained using models that incorporate all four leads. The CNN model performs slightly better than the GBDT model, with a sensitivity of 0.77 and specificity of 0.75 compared to 0.77 and 0.72, respectively. When distinguishing between AMI and healthy participants, the specificity increases to 0.94 for the CNN model and 0.90 for the GBDT model. Feature importance analysis shows that the GBDT model primarily relies on the J point, while the CNN model primarily relies on the QRS complex. Conclusions: wECG-based AMI detection shows considerable promise in out-of-hospital settings. However, caution is needed as CNN explanations rarely agree with the ECG intervals typically analyzed in clinical practice.