Capturing the curve: Functional data analysis for validated digital outcome measures
Discussing summary metrics derived from Multilevel Functional Principal Component Analysis using smartwatch ECG data as a motivating example.
The use of digital health technologies to measure outcomes in clinical trials opens new opportunities as well as methodological challenges. Digital outcome measures can provide more convenient data collection via remote assessment and may potentially detect clinical change earlier. However, there are unresolved methodological challenges around how digital outcomes should be defined and validated, and very few digital outcomes are currently accepted as primary outcomes in late-phase trials.
A major difficulty lies in summarising the volume of data collected by digital devices. Many physiological signals, such as electrocardiogram (ECG) traces or measurements of gait, have a hierarchical functional structure: repeated physiological functions are observed over time and on multiple time scales (seconds, days, weeks). While many summaries can be derived from digital data, typically, only a small subset of pre-defined, clinically interpretable scalars are validated as clinical outcome measures.
As an alternative approach, we evaluate data-driven summaries or scores from Multilevel Functional Principal Component Analysis (MFPCA). These scores are derived from a formal method for obtaining meaningful low-dimensional representations of functional data with robust statistical properties. Through a simulation study motivated by smartwatch electrocardiogram (ECG) signals, we compare the performance of MFPCA scores with pre-specified summaries on validation criteria for digital outcome measures. We discuss the potential benefits and challenges for this approach in the context of trials.
Speaker
- Dr Mia Tackney - NIHR Advanced Fellow, MRC Biostatistics Unit
Event notices
- Please note that you can join this event in person or you can join the session remotely
- Please note that the recording link will be listed on this page when available
Admission
Contact
