Remote monitoring of heart failure exacerbations using a smartwatch

Figures

• 

  Fig. 1. Schematic representation of the TRUE-HF observational cohort study design. a, The TRUE-HF study protocol incorporating study-entry clinic visits and 3-month end-of-study visits to assess patient status and cardiopulmonary fitness. Wearable data and daily surveys were collected during the free-living observation period. b, Distribution of key HealthKit variables, illustrating the number of HealthKit recordings collected during the study. c, The TRUE-HF model processing a 30-d sliding window of wearable data for outcome prediction (pVO2), which was tested for concordance with CPET-derived pVO2 and clinical events including declines in CPET pVO2, hospitalization, emergency visits and intravenous furosemide administration.

• 

  Fig. 2. Consort diagram for the TRUE-HF study. Reasons for patient ineligibility or nonparticipation are included. aPatients were deemed ineligible or not approached for the following reasons: unable to give consent, had documentation requesting not to be approached for research, on dialysis, had documented noncompliance, had active psychiatric problems, new patients to the clinic or purposefully selected to ensure demographic diversity in the study cohort. In addition, the final 50 patients (held-out set) were recruited to match the CPET pVO2 distribution of the training cohort. bAmong the group of individuals who were excluded for having insufficient data, one patient had received a left ventricular assist device (LVAD) and one patient had died. The 15 participants who did not complete the end-of-study assessments nevertheless had sufficient follow-up data to be included in the unplanned healthcare utilization analysis.

• 

  Fig. 3. Biometric data from Apple Watch can be used to estimate cardiorespiratory fitness. a, Scatterplot representation of predicted pVO2 compared to the measured ground truth of CPET pVO2 at the end-of-study clinic visit (median 94.5 d post study-entry). Points represent individual participants. The solid line indicates the fitted linear relationship between predicted and measured pVO2. Shaded bands denote the 95% confidence band around the fitted linear relationship. b, Spearman’s correlation of the TRUE-HF model (n = 50) and Apple model (n = 35) to the CPET pVO2 measured at the end-of-study clinic visits. The pVO2 according to the Apple model was not measured for 15 participants. c, AUROC curve in clinical pVO2 measurement from study-entry to end-of-study clinic visits. AUROCs were calculated by comparing each of the model’s prediction of pVO2 decline (based on comparing the last, temporally closest to the end-of-study clinic visit date, prediction to the study-entry CPET pVO2) against the actual observed declines in CPET pVO2. Solid lines represent the mean ROC curve and shaded bands indicate the 95% confidence interval (CI), estimated by bootstrap resampling. A statistically…

• 

  Fig. 4. The TRUE-HF model and its reduced-sensor variant (TRUE-HF-RS) predict declines in daily pVO2 before unplanned healthcare use. a–c, Performance of TRUE-HF model on the held-out TRUE-HF test set (n = 63), of which 58 patients had sufficient data for analysis (>10 d, 58 of 63). a, AUROC of unplanned healthcare utilization prediction of the TRUE-HF model against study-entry CPET pVO2, NT-proBNP, 6MWTD and NYHA functional class. AUROC was calculated using the maximum decrease in TRUE-HF daily pVO2 between the first model prediction and drops before the event for TRUE-HF. For NT-proBNP, 6MWT and NYHA, AUROC was calculated using the corresponding study-entry measures. The circled point on the AUROC curve represents the sensitivity and specificity of a 10% drop in TRUE-HF daily pVO2. b, Time-varying, extended, Simon and Makuch’s Kaplan–Meier cumulative risk curve for the TRUE-HF model of a 10% drop detected for event-free survival during the study. The shadowed areas indicate 95% CIs. The TRUE-HF drop group is a time-dependent covariate, which more accurately reflects the impact of time-dependent predictions on survival. c, Longitudinal TRUE-HF daily pVO2 model predictions in patients…

• 

  Extended Data Fig. 1. TRUE-HF transformer model framework. TRUE-HF model processes 30-days of data comprising HealthKit features from nine prevalent HealthKit data types in 90-min summaries to predict outcomes. First, the TRUE-HF model uses a TRUE-HF Transformer block to learn temporal relationships in the data sequence and reduces the representations and temporal resolutions (more oversized windows). The length of time is reduced deeper in the model through pooling. TRUE-HF integrates demographic and drug (clinical features) data with time-series through affine transformation, where these clinical details can affect the interpretation of the HealthKit variables. A linear layer generates the final prediction of outcomes (that is, daily pVO2 regression).

• 

  Extended Data Fig. 2. Consort diagram for All of Us external validation set.

• 

  Extended Data Fig. 3. Landmark analysis of time-dependent performance of TRUE-HF-RS for predicting unplanned healthcare utilization events in All of Us. Solid lines indicate mean estimates of the AUROC, while dashed lines represent Specificity and dot–dash lines represent Sensitivity. Shaded regions denote 95% confidence intervals derived from bootstrapping.

• 

  Extended Data Fig. 4. Biometric data from Apple Watch can be used to estimate 6MWTD. (A) Scatterplot representation of predicted 6MWTD compared to the measured ground truth of clinical 6MWTD at the end-of-study clinical visit (median 94.5 d post study-entry). Points represent individual participants (n = 49). The solid line indicates the fitted linear relationship between predicted and measured 6MWTD. Shaded bands denote the 95% confidence band around the fitted linear relationship (B) Spearman’s correlation of the TRUE-HF 6MWTD model (n = 49) and Apple’s SixMinuteWalkDistance model (n = 44) to the clinical 6MWTD measured at the end-of-study clinical visit. 6MWTD according to Apple’s model was not measured for 5 participants. (C) AUROC curve in clinical 6MWTD measurements from study-entry to end-of-study clinical visit. AUROCs were calculated by comparing each of the model’s prediction of 6MWTD decline (based on comparing the last, temporally closest to the end-of-study clinical visit date, prediction to the study-entry clinical 6MWTD) against the actual observed declines in clinical 6MWTD. Solid lines represent the mean ROC curve, and shaded bands indicate the 95% confidence interval, estimated by…

• 

  Extended Data Fig. 5. TRUE-HF 6MWTD and Apple Algorithm 6MWTD predictions of declines in daily 6MWTD prior to unplanned healthcare use. (A–B) Experiments were conducted in the full TRUE-HF held-out test set (n = 63). (A) AUROC of unplanned healthcare utilization prediction of the TRUE-HF 6MWTD model against the Apple 6MWTD model. AUROC was calculated using the maximum decrease in model 6MWTD between the first model prediction and drops before the event. Solid lines represent mean estimates, and shaded regions indicate 95% confidence intervals. (B) Forest plot of hazard ratios from Cox proportional hazards models evaluating the association between declines in daily 6MWTD and time to unplanned healthcare utilization for TRUE-HF and Apple estimates. Points denote HR estimates, horizontal lines indicate 95% confidence intervals, and p-values correspond to two-sided Wald tests. The vertical dashed line denotes HR = 1.