Oura Ring Gen 4 sensor data, not clinical measurementsN=1 case study, not validated for clinical decisionsHEV diagnosed Mar 18; Day 109 post-ruxolitinibMore
Consumer wearable data can support exploratory review only. The HEV diagnosis, temporally confounded with treatment start, remains a material confounder.

Oura Digital Twin Dashboard

Oura Gen 4: 2026-01-08 to 2026-07-02 (176 unique days, 202 sleep periods) · 10/10 analysis modules loaded · Ruxolitinib 10 mg BID from 2026-03-16
POST-HSCT BIOMETRIC MONITORING

Consumer wearable suggests a physiological shift after JAK inhibitor start
exploratory N=1 monitoring in a confounded window

+15.5
HRV shift
ms (pre to post)
-12.6
Resting HR shift
bpm (pre to post)
176
Days
modelled
109
Post-drug
days observed
Oura Ring Gen 4 · 176 days · 10/10 analysis modules · Ruxolitinib 10 mg BID from 2026-03-16
Pre-Ruxolitinib to Post-Ruxolitinib
67 days (73 sleep periods) vs 109 days (129 sleep periods)
Key Finding
Analysis modules detected multiple biosignal shifts after ruxolitinib initiation, but the 109-day post-intervention window remains too short for strong causal claims. 0/0 CausalImpact streams reach raw p<0.05; 0/0 survive FDR correction. Separate UKF comparisons: Cardiac Reserve shift p<0.001, Sleep Quality shift p<0.001.
Temperature
N/A
No data
HRV (RMSSD)
18.9ms
vs healthy 42-49 ms | Kleiger 1987 / Bigger 1992
Resting HR
77bpm
IST threshold: 90 bpm
SpO2
96.0%
BOS risk: MODERATE (22/100)
Sleep Efficiency
74%
Below 85% threshold
Autonomic Tone
0.00SD
UKF state | improved | p<0.001
GVHD Score
17/100
7d avg: 26
Anomaly Days
18/202
Feb 9 detected (18 anomaly days, N=1)

Interrupted Time Series: Temperature Response

CausalImpact BSTS

Treatment Effect Sizes

CausalImpact + UKF

Disease State Trajectory

rSLDS 4-State Model

Anomaly Detection

5-Algorithm Ensemble

Composite Biomarker Trends

6 Biomarkers

SpO2 & BOS Risk

Pulmonary Screening

Combined Biosignal Timeline

5-Channel View

Heart Rate Terrain Map

3D Surface

Phase Space Reconstruction

Nonlinear Dynamics

Sleep Architecture Heatmap

Epoch-Level

Circadian Rhythm Analysis

Cosinor Model
Methodology & Data Provenance
Data Source
Oura Ring Gen 4 (continuous PPG, accelerometry, thermometry)
Volume
104,855 HR readings · 14,150 HRV samples · 16,423 sleep epochs
Analysis Pipeline
CausalImpact (BSTS) · Unscented Kalman Filter (5 latent states) · rSLDS (4 disease states) · 5-algorithm anomaly ensemble · 6 composite biomarkers · Chronos-2 / ARIMA forecasting
Sleep Architecture
Deep: 14.8% Light: 52.4% REM: 12.2% Awake: 20.6%
For research and clinical discussion only. Not a diagnostic tool. All p-values are exploratory.