Related Experiment Video
Updated: Apr 7, 2026
08:28
Development of an In Vitro Ocular Platform to Test Contact Lenses
Published on: April 6, 2016
11.4K
Large Language Models for Rapid Instrument Prototyping: Design and Structural Optimization of the Dry Eye Disease in
Marta Jaruchowska1, Musa Aamir Qazi1, Muhammad Jalal Haidar1
1Department of Experimental Physiology and Pathophysiology, Medical University of Warsaw, Warsaw, Poland.
Clinical Ophthalmology (Auckland, N.Z.)
|April 6, 2026
Summary
This study developed the Dry Eye Disease in Pregnancy Questionnaire (DED-PREG) using AI simulation. This novel tool captures pregnancy-specific symptoms, aiding in better diagnosis and management of gestational dry eye disease.
Area of Science:
- Ophthalmology
- Reproductive Medicine
- Computational Biology
Background:
- Gestational Dry Eye Disease (DED) affects up to 50% of pregnant individuals.
- Existing diagnostic tools are not tailored for pregnancy-specific DED symptoms.
- Developing new instruments for pregnant populations faces logistical and ethical challenges.
Purpose of the Study:
- To describe the development and computational prototyping of a pregnancy-specific DED instrument.
- To utilize a Generative Artificial Intelligence (GenAI) framework for instrument design.
- To create the Dry Eye Disease in Pregnancy Questionnaire (DED-PREG).
Main Methods:
- Employed a multi-stage in silico framework with synthetic cohorts.
- Generated qualitative focus group simulations for content derivation.
- Utilized semantic vectorization for item reduction and validated with 500 synthetic pregnant personas.
- Evaluated the 20-item instrument for internal consistency, structural validity, and test-retest reliability using a longitudinal simulation engine across five gestational timepoints.
Main Results:
- The DED-PREG instrument comprises three domains: Ocular Symptoms, Functional Impact, and Lifestyle & Environmental Modulators.
- Demonstrated satisfactory internal consistency (Cronbach's alpha = 0.89) and excellent temporal stability (ICC = 0.99).
- Confirmatory Factor Analysis showed acceptable model fit (CFI = 0.82; RMSEA = 0.11).
- Longitudinal analysis confirmed responsiveness to gestational changes (Global Cohen's d = 0.44) and identified a significant interaction between low socioeconomic status and symptom exacerbation (p < 0.001).
Conclusions:
- Presents the first pregnancy-specific DED instrument optimized via AI simulation.
- Demonstrates GenAI's capability for rigorous instrument design and rapid prototyping prior to clinical deployment.
- Highlights the potential of computational approaches to overcome challenges in developing clinical tools for specific populations.
