Outcome measures based on digital health technology sensor data: data- and patient-centric approaches

Area of Science:

• Digital Health
• Biomedical Engineering
• Clinical Research Methodology

Background:

• Digital health technology tools (DHTTs) like apps and wearables collect remote health data.
• DHTTs offer objective, sensitive, and high-frequency data collection, enabling detailed patient phenotyping.
• A gap exists in developing robust outcome measures from DHTT sensor data.

Purpose of the Study:

• To outline two roadmaps for generating outcome measures from DHTT data.
• To address the need for meaningful digital outcome measures in clinical research.
• To facilitate the use of DHTTs in drug development and patient monitoring.

Main Methods:

• A data-centric approach summarizing sensor features sensitive to specific concepts (e.g., disease progression).
• A patient-centric approach summarizing sensor features relevant to patients' daily functioning.
• Exemplification using Parkinson's disease motor symptom and experience tracking.

Main Results:

• Two distinct roadmaps (data-centric and patient-centric) were developed for creating digital outcome measures.
• These roadmaps are applicable to observational and interventional studies.
• The methods are demonstrated for Parkinson's disease research.

Conclusions:

• Developing methods to summarize sensor data into meaningful outcome measures is crucial for DHTT research.
• The proposed roadmaps aim to foster scientific discourse and consensus on digital outcome measures.
• This work supports the creation of novel digital outcome measures for clinical research and drug development.