Clinical report writing – Analysis of uncertainty terms and usage of medical ontologies

From REU@MU
Revision as of 23:45, 12 March 2023 by Praveen (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Title : Clinical report writing – Analysis of uncertainty terms and usage of medical ontologies

Uncertainty in clinical reports has always been a challenge in the diagnosis process. Uses of uncertainty terms can mislead diagnosis and develop a miscommunication among different stakeholders at clinical settings. Standardized clinical report formats are recommended, but in practice limited usage is noted. Medical ontologies play a key role in standardized clinical terms, but not every clinical expert uses those ontologies. There is a need for computer-aided systems that can classify these uncertainty terms in clinical reports and notify users about that. Clinical reports quality can be improved using classification models and identify uncertainty in those reports. Qualifying uncertainty in these reports is a challenging problem due to the unstructured data format supported by these reports. Limited research has been done to identify uncertainty in clinical report writings and limited availability of standardized vocabulary makes this task more challenging. In this proposed research work, we will be using publicly available and in-house (from hospitals) clinical reports and will identify uncertainty levels using machine learning and natural language processing algorithms. We will identify measures to identify the quality of clinical reports and predict uncertainty levels in existing reports. We will use semi-supervised machine learning techniques and natural language processing techniques.

Skills: Python, Machine learning, Natural language Processing, Transformer-based language models (if any)

Faculty details: Priya Deshpande, PhD Email: Priya.deshpande@marquette.edu