Strados Labs Whitepaper
Data Processing Architectures for Cough Monitoring: The Case for Cloud Processing
By Jason Kroh, Chief Technology Officer and Tom deLaubenfels, PhD, Director of Data Science at Strados Labs, 2025
What’s Inside?
This whitepaper explores two approaches to cough monitoring data architectures, edge processing and cloud processing, and their implications for clinical trials. Edge processing analyzes cough events directly on the device, offering real-time results and reduced data transmission, but limits transparency by discarding raw audio and preventing retrospective review.
Cloud processing, by contrast, preserves raw recordings for centralized analysis, enabling human annotation (the FDA gold standard), hybrid machine learning review, and complete audit trails for regulatory submissions. This architecture also supports algorithm development, multi-annotator quality control, and retrospective analyses, while maintaining subject privacy through encryption and speech obfuscation.
These findings make the case that cloud-based processing provides the auditability, flexibility, and scientific rigor needed for cough monitoring as a defensible digital endpoint in clinical research and drug development.
