SHARPS DS2 is a technical architecture and open source prototype that leverages OpenCDS (http://www.opencds.org/), a popular Clinical Decision Support (CDS) framework, for the identification and sequestration of certain types of sensitive information from patient records flowing through a Health Information Exchange (HIE). The project adopted the name “DS2” – Decision Support for Data Segmentation – because of its unique focus on the ability to detect clinical facts that may imply a sensitive condition, in addition to detecting clinical facts that are directly related to the condition.
The project demonstrated that the redaction of a condition and its related clinical facts sometimes leaves residual facts that, through clinical inference, can still reveal the redacted condition. To address this challenge, OpenCDS deterministic rules were combined with Bayesian and other machine learning classifiers to redact targeted conditions along with certain co-occurrences and co-morbidities.
It is important to note that DS2 is prototype software, intended to demonstrate the DS2 architecture, and has not been thoroughly tested nor is intended for use in a production environment. The rules and classifiers that have been developed were created to demonstrate the methodology and the relative effectiveness of different approaches for a small subset of conditions, and should not should not be considered complete for production use. Visitors who have arrived at this site looking for pilot or production software related to Data Segmentation for Privacy should see the DS4P Pilots at http://wiki.siframework.org/Data+Segmentation+for+Privacy+RI+and+Pilots+Sub-Workgroup. DS4P is a very broad topic that encompasses consent, policy, privacy law, health data standards, clinical vocabularies, authentication and access control. By contrast, DS2 focuses on a few niche areas that we hope will contribute to existing and future DS4P research and development, as well as other work related to HIE and CDS:
- Clinical inferencing: Segmentation utilizing co-occurring concepts in addition to child and synonymous concepts
- Utilizing the Virtual Medical Record (vMR) and OpenCDS to make clinical inferences (and translating CCD to vMR)
- Predicate/Reducer approach/dichotomy and utilization of both knowledge-based and machine learning CDS models
- OpenCDS tools, including template-driven CDA Editor/Test Manager
- Safety: vMR-based drug interaction checking using non-redacted content
- Unique HIE proxy-based architecture
To learn more about the project and to view documentation on the prototype software, see our white papers in Publications, view the slide set at the bottom of this page, or click on any of the sub-topics listed below:
All of the source code for the project is available under an open source license and hosted at the SHARPS DS2 Bitbucket Team repository at https://bitbucket.org/sharps-ds2.
For more information, please contact: firstname.lastname@example.org.
This work was supported by HHS 90TR0003-01 (SHARPS) and the Illinois Office of Health Information Technology (OHIT). The views expressed are those of the authors only.