Study Rationale: The combination of digital measures and machine learning (ML) holds great promise for improving critical aspects of Parkinson’s disease (PD) care, practice and research. Earlier PD diagnoses may yield better treatment regimens that improve patients’ quality of life. Remote PD monitoring could help reduce in-person clinic visit burdens. However, additional research is needed to determine how well digital measures can classify and monitor PD in practice and at scale. This study will be completed with a retrospective dataset of digital measures using the Verily Study Watch (Verily Life Sciences, Inc.) in the Parkinson’s Progression Markers Initiative (PPMI).
Hypothesis: We hypothesize that digital measures can accurately categorize subjects into known groups of healthy, prodromal PD and severity of diagnosed PD (mild to severe) and can detect empirically established minimal clinical important differences in PD disease progression.
Study Design: This is a retrospective study to analyze the dataset obtained from the Verily Study Watch (PPMI). The dataset consists of 341 enrolled subjects with varying PD diagnosis and healthy controls and includes sensor data for physical activity tracking, as well as photoplethysmography (PPG) sensors for monitoring cardiac activity. Analysis for quality and wear compliance will include descriptive statistics for each measure, ceiling and floor effects and outlier analysis. Derived measures with functional significance to PD patients will be grouped by physical activity, gait, sleep, and cardiac activity. ML models will be used to analyze the dataset and distinguish between cohorts.
Impact on Diagnosis/Treatment of Parkinson’s Disease: Clinicians may have better PD diagnosis and disease progression monitoring tools which provide more fine-grained clinical signals. Patients diagnosed earlier may immediately benefit from better prognostic counseling, earlier initiation of disease-modifying or neuroprotective therapies at a stage when such therapies would likely be most effective.
Next Steps for Development: The next step will be clinical validation of the digital measures to determine the ability to measure change in PD disease symptoms through disease progression or treatment effect.