SciKit Digital Health: Python Package for Streamlined Wearable Inertial Sensor Data Processing

Area of Science:

• Digital Health
• Biomedical Engineering
• Data Science

Background:

• Wearable inertial sensors offer valuable patient mobility and health data.
• Limited availability of open-source software tools for processing daily living activities from sensor data.
• Lack of readily available code for research replication and off-the-shelf software packages.

Purpose of the Study:

• Introduce SciKit Digital Health (SKDH), an open-source Python package.
• Provide algorithms for deriving clinical features from wearable sensor data.
• Streamline digital endpoint generation for research and clinical applications.

Main Methods:

• Developed SKDH as a Python software package.
• Integrated algorithms for gait, sit-to-stand, physical activity, and sleep analysis.
• Designed an extensible framework for data ingestion, preprocessing, and analysis.

Main Results:

• SKDH offers a unified pipeline for digital health data processing.
• The package promotes reproducibility through a convention-over-configuration approach.
• Standardized settings are provided for healthy and mildly impaired adult populations.

Conclusions:

• SKDH simplifies the creation of digital health data processing pipelines.
• The open-source package facilitates reproducible research in digital health.
• SKDH is freely available under an MIT license for use and extension.