Distributed Data Processing for Public Health Surveillance
Author Information
Author(s): Lazarus Ross, Yih Katherine, Platt Richard
Primary Institution: Channing Laboratory, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School
Hypothesis
Can a distributed data processing model effectively support public health surveillance while minimizing the risk of disclosing personal health information?
Conclusion
The distributed processing model can be successfully deployed with very low risk of inadvertent disclosure of personal health information.
Supporting Evidence
- The distributed model minimizes the risk of disclosing personal health information.
- Only aggregate data is transferred to the central datacenter for analysis.
- The system supports rapid and efficient syndromic surveillance.
Takeaway
This study shows that we can collect health data without sharing personal information, making it safer for everyone involved.
Methodology
The study describes a distributed data processing model where personal health information is processed locally and only aggregate data is sent for analysis.
Potential Biases
There is a risk of bias if the model cannot distinguish multiple encounters from the same individual.
Limitations
The model requires a clinical responder to provide case-level information when clusters are detected, which may delay responses.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website