Enhancing the Representational Power of i2b2 through Referent Tracking

Area of Science:

• Biomedical Informatics
• Clinical Data Management
• Health Data Analytics

Background:

• The Informatics for Integrating Biology and the Bedside (i2b2) platform effectively identifies patient cohorts using demographic, phenotypic, and genetic data.
• A limitation of current i2b2 applications is the inability to search based on characteristics of clinical assertions (e.g., diagnoses) themselves, rather than just the conditions observed.

Purpose of the Study:

• To explore the feasibility of using i2b2's modifier system to represent relationships between observations and their referents.
• To enable cohort identification based on disagreements among healthcare providers regarding patient diagnoses.

Main Methods:

• Utilized i2b2's modifier system to explicitly represent relationships between clinical observations and their referents (explicit or implied).
• Stored unique instance identifiers for observations and referents within i2b2's fact table.
• Leveraged i2b2's base functionality and implementation specifications.

Main Results:

• Demonstrated that i2b2's modifier system can represent complex relationships between observations and referents.
• Successfully stored unique identifiers for observations and referents, facilitating detailed data analysis.
• The proposed approach makes explicit previously undetected ambiguities and confusions in clinical assertions.

Conclusions:

• The i2b2 modifier system can be extended to capture nuanced information about clinical assertions, including provider disagreements.
• This method enhances the ability to identify specific patient cohorts by analyzing the characteristics of diagnostic assertions.
• The approach improves the transparency of clinical data by revealing underlying ambiguities.