Evaluating Diversity in Open Photoplethysmography Datasets: Protocol for a Systematic Review

Area of Science:

• Biomedical Engineering
• Data Science
• Medical Informatics

Background:

• Photoplethysmography (PPG) is a noninvasive optical technique measuring blood volume changes, crucial for pulse oximeters and wearables.
• Technical limitations of PPG can introduce bias, potentially impacting specific patient and consumer groups.
• Openly accessible datasets are increasingly driving PPG technology development, but may underrepresent certain populations.

Purpose of the Study:

• To systematically review open-access PPG datasets, evaluating their reporting patterns and structure.
• To identify included biosignals and demographic variables to understand current medical device development parameters.
• To pinpoint gaps and areas for improvement in PPG datasets to mitigate bias in medical device development.

Main Methods:

• Systematic review following PRISMA guidelines, including studies with PPG and open-access datasets since 2000.
• Searches across MedLine OVID, IEEE Xplore, Scopus, and PhysioNet for datasets with physiological parameters and PPG waveforms.
• Evaluation based on Standing Together Initiative recommendations, with statistical analysis of demographic and biosignal variables.

Main Results:

• Analysis of dataset diversity and structural characteristics, including statistical comparison of nominal variables.
• Evaluation of frequencies for devices, biosignals, clinical parameters, demographics, and geographic information.
• The review is ongoing, with expected completion by September 2025.

Conclusions:

• The review will highlight potential gaps in current open-access PPG datasets.
• Findings will inform future data collection and design of PPG datasets to ensure diverse representation.
• This aims to prevent bias perpetuation in medical devices and enable broader clinical application.