Related Experiment Video
Updated: Jul 20, 2025
05:40
Using an Automated Hirschberg Test App to Evaluate Ocular Alignment
Published on: March 24, 2020
12.3K
A case study in applying artificial intelligence-based named entity recognition to develop an automated ophthalmic
Carmelo Z Macri1,2, Sheng Chieh Teoh3, Stephen Bacchi4,3
1Discipline of Ophthalmology and Visual Sciences, The University of Adelaide, Adelaide, South Australia, Australia. carmelo.macri@adelaide.edu.au.
Summary
We developed an automated ophthalmic disease registry using artificial intelligence (AI) and a low-code tool. This system extracts diagnoses from electronic health records, aiding clinicians in disease case finding.
Area of Science:
- Ophthalmology
- Medical Informatics
- Artificial Intelligence
Background:
- AI-based named entity extraction (NER) enhances diagnostic entity extraction from electronic health records (EHRs).
- Clinicians often lack AI expertise, limiting the adoption of advanced NLP tools.
- A need exists for user-friendly tools to leverage AI for clinical data analysis.
Purpose of the Study:
- To demonstrate a case study for developing an automated ophthalmic disease registry.
- To create a ready-to-use, low-code tool for clinicians to facilitate AI adoption.
- To improve the extraction of diagnostic entities from unstructured EHR data.
Main Methods:
- Extracted deidentified clinical records from an adult outpatient ophthalmology clinic (Nov 2019 - May 2022).
- Utilized low-code annotation software (Prodigy) for diagnosis annotation.
- Trained a custom spaCy NER model to extract diagnoses and build the registry.
Main Results:
- Extracted 123,194 diagnostic entities from 33,455 clinical records.
- Identified 5070 distinct diagnostic entities after data cleaning.
- The NER model achieved a precision of 0.8157, recall of 0.8099, and F-score of 0.8128.
Conclusions:
- A case study demonstrated the creation of an automated ophthalmic disease registry using low-code AI and NLP tools.
- The developed NER model showed moderate capability in extracting diagnoses from clinical text.
- A ready-to-use tool was created to encourage clinician adoption of AI for EHR case finding.
