Using Large Language Models for In Silico Development and Simulation of a Patient-Reported Outcome Questionnaire for Cataract Surgery with Various Intraocular Lenses - A Pre-Validation Study

Sep 30, 2026

Uploaded by Cart

Development of patient-reported outcome measures (PROMs) in ophthalmology faces high patient burden during early validation. An In Silico Pre-validation Framework is proposed using large language models (LLMs) to stress-test instruments before clinical deployment. Methods generate a PROM questionnaire and a synthetic cohort of 500 patient profiles instantiated as structured JSON, including IOL type and dysphotopsia severity. Stateless simulation supports test–retest reliability. Results show strong psychometric fit via CFA, no significant DIF by age, sex, or IOL type, robust internal consistency (Cronbach’s alpha >0.80), high ICC (>0.90), and convergent validity versus NEI-VFQ-25. Conclusions emphasize algorithmic fidelity to identify structural failures in silico, reducing ethical and logistical burden.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Background/Objectives

Methods

Results

Conclusions

Introduction

Patient-centered practice and PROMs

Need for tailored tools in cataract surgery

Role of large language models in medical research

Article Using Large Language Models for In Silico Development and Simulation of a Patient-Reported Outcome Questionnaire for Cataract Surgery with Various Intraocular Lenses: A Pre-Validation Study Ewelina Trojacka 1, Joanna Przybek-Skrzypecka 2,3, Justyna Izdebska 1,2,3, Jacek P. Szaflik 1,2,3, Musa Aamir Qazi 4, Abdullah Azhar 4 and Janusz Skrzypecki 4, * Academic Editors: Kyung Chul Yoon and Yoshihiro Takamura Received: 2 December 2025 Revised: 20 December 2025 Accepted: 28 December 2025 Published: 30 December 2025 Copyright: © 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license. 1 Center of Ocular Microsurgery, Professor Jerzy Szaflik’s Clinic in Warsaw, 00-215 Warszawa, Poland 2 SPKSO Ophthalmic University Hospital in Warsaw, 03-709 Warszawa, Poland 3 Department of Ophthalmology, Medical University of Warsaw, 02-091 Warszawa, Poland 4 Department of Experimental Physiology and Pathophysiology, Medical University of Warsaw, 02-091 Warszawa, Poland * [Correspondence: janusz.skrzypecki@gmail.com](Correspondence: janusz.skrzypecki@gmail.com) Abstract Background/Objectives: Development of Patient-Reported Outcome Measures (PROMs) in ophthalmology is limited by high patient burden during early validation. We propose an In Silico Pre-validation Framework using Large Language Models (LLMs) to stresstest instruments before clinical deployment. Methods: The LLM generated a PROM questionnaire and a synthetic cohort of 500 distinct patient profiles via a Python-based pipeline. Profiles were instantiated as structured JSON objects with detailed attributes for demographics, lifestyle, and health background, including specific clinical parameters like IOL type (Monofocal, Multifocal, EDOF) and dysphotopsia severity. To eliminate memory bias, a stateless simulation approach was used for test–retest reliability; AI agents were re-instantiated without access to prior conversation history. Psychometric validation included Confirmatory Factor Analysis (CFA) using WLSMV estimation and Differential Item Functioning (DIF) . Results: The model demonstrated excellent fit (CFI = 0.962, TLI = 0.951, RMSEA = 0.048, SRMR = 0.063), confirming structural validity. DIF analysis detected no significant bias based on age, sex, or IOL type (0/20 items flagged) . Internal consistency was robust (Cronbach’s alpha > 0.80) and stateless test–retest reliability was high (ICC > 0.90), indicating stability independent of algorithmic memory. Convergent validity was established via significant correlations with NEI-VFQ-25 scores (Spearman’s:−0.425 to −0.652,) . While responsive to change, known-groups validity reflected realistic clinical overlap. Conclusions: LLM-based pre-validation effectively mirrors complex human response patterns through “algorithmic fidelity.” By identifying structural failure points in silico, this framework ensures PROMs are robust and unbiased before clinical trials, reducing the ethical and logistical burden on real-world populations. Keywords: visual symptoms questionnaire; PROM; cataract surgery; GenAI; LLMsassisted PROM 1. Introduction Modern clinical practice is increasingly centered on improving the patient experience and overall quality of life, marking a paradigm shift from purely objective clinical metrics to subjective well-being [1–5] . A key component of this patient-centered approach is the use of patient-reported outcome measures (PROMs), which offer indispensable insights into patients’ subjective experiences, symptom burden and treatment satisfaction that cannot be captured by clinical examination alone [6–14] . However, the development of high-quality, scientifically robust PROMs is notoriously resource-intensive and time-consuming. It requires extensive input from diverse patient populations through iterative focus groups and cognitive debriefing.

FAQs

What problem does the study address in PROM development for ophthalmology?

Early PROM validation is limited by high patient burden, plus delays caused by ambiguities and omissions discovered only after interviews and statistical validation.

How does the proposed framework generate and validate the questionnaire?

It uses LLMs to generate a PROM questionnaire and a synthetic cohort of 500 structured patient profiles, then applies stateless simulation to support test–retest reliability and conducts psychometric validation via CFA and DIF.

What psychometric evidence supports the framework’s effectiveness?

The results report excellent model fit (CFI, TLI, RMSEA, SRMR), no significant DIF across age, sex, or IOL type (0/20 flagged items), strong internal consistency (Cronbach’s alpha >0.80), and high stateless test–retest reliability (ICC >0.90), with convergent validity against NEI-VFQ-25.

Sep 30, 2026

Uploaded by Cart