This technology is a modeling approach that uses Electronic Health Record (EHR) data to characterize clinical behaviors that can be used to prioritize clinician concerns and identify high-risk patients.
Unmet Need: Method for analysis of EHR data to inform clinical practice
Electronic Health Records (EHRs) contain valuable patient information, including demographics, medications, diagnoses, and the recommendations and concerns of clinicians and nurses. Streamlining the use of EHR data to inform clinical practice has the potential to improve hospital efficiency and patient health management. However, current methods to analyze EHR data that utilize machine learning and predictive algorithms have struggled to account for noise, interdependent variables, dataset integration, and subjective variables.
The Technology: Predictive modeling framework to improve clinical practice
This technology describes a modeling approach that uses EHR data to characterize clinician behaviors in order to predict and/or determine patient or clinician attributes that are not directly measurable. Using multiple natural language processing methods, this technology incorporates subjective notes and identifies quality phrases as predictors of significant risk factors for negative patient outcomes. Overall, this technology enables prioritization of clinician concerns and identification of high-risk patients, potentially improving patient outcomes and guiding future clinical practices.
Applications:
- Prediction and learning of clinician concern and decision-making
- Patient acuity scoring
- Resource tracking
- Clinician work burden assessment
Advantages:
- Sensitive to changes in clinician behavior
- Quantifies patient and clinician properties that are not directly measurable
- Compatible with existing EHR data
- Increased usability via integrated clinician expertise
- Tunable to range of clinical concepts
Lead Inventor:
Kenrick Cato, Ph.D.
Sarah C. Rossetti, RN, Ph.D.
Patent Information:
Patent Pending
Related Publications:
Rossetti SC, Knaplund C, Albers D, Tariq A, Tang K, Vawdrey D, Yip NH, Dykes PC, Klann JG, Kang MJ, Garcia J, Fu LH, Schnock K, Cato K. “Leveraging clinical expertise as a feature - not an cutcome - of predictive models: evaluation of an early warning system use case” AMIA Annu Symp Proc. 2020 Mar 4; 2019:323-332.
Korach ZT, Yang J, Rossetti SC, Cato KD, Kang MJ, Knaplund C, Schnock KO, Garcia JP, Jia H, Schwartz JM, Zhou L. “Mining clinical phrases from nursing notes to discover risk factors of patient deterioration” Int J Med Inform. 2020 Mar; 135:104053.
Kang MJ, Dykes PC, Korach TZ, Zhou L, Schnock KO, Thate J, Whalen K, Jia H, Schwartz J, Garcia JP, Knaplund C, Cato KD, Rossetti SC. “Identifying nurses’ concern concepts about patient deterioration using a standard nursing terminology” Int J Med Inform. 2020 Jan; 133:104016.
Tech Ventures Reference:
