Related Experiment Video
Updated: Feb 22, 2026

An R-Based Landscape Validation of a Competing Risk Model
Published on: September 16, 2022
Safety and Efficiency Reducing Retinopathy of Prematurity Guideline Sensitivity: An External Validation Using a Large
Robert W Arnold1, Jack Jacob2, Lance Siegel3
1From the Alaska Blind Child Discovery (R.W.A.), Alaska Children's EYE & Strabismus, Anchorage, Alaska, USA.
Purpose:
Recent publications have proposed alternative retinopathy of prematurity (ROP) screening guidelines balancing efficiency with safety. We used the 28-hospital ROP CheckⓇ higher-risk, racially diverse database to perform external validation. ROP Check is neonatology/ophthalmology software that combines customizable, guideline-based screening and ophthalmology monitoring of staging ROP examinations and treatment.
Design:
Retrospective analysis of screening guideline sensitivity.
Methods:
We reviewed encounter volumes from 2010 to 2022 of infants whose entire hospital stay was monitored with ROP Check. Patients with primary laser or anti-vascular endothelial growth factor (anti-VEGF) therapy were compared with recent-encounter volume-reducing (2%-8.6%) proposals.
Results:
A total of 5058 infants had 23,557 physician-patient encounters. Of these, 388 infants required ROP treatment, including 3 patients (1 of Pacific and 2 of non-Pacific race/ethnicity) with gestational age (GA) of 30.0 to 30.9 weeks. The 174 patients treated initially with laser had larger median bodyweight (BW) (706 g) than the 214 initially treated with anti-VEGF (620 g), older GA (25.1 vs 24.4 weeks), older GA at initial treatment (36.6 vs 35 weeks), fewer encounters (10 vs 13), and fewer examinations after treatment (3 vs 7). Six cases reached stage 4. We raise concerns that some of our treatment-warranted infants reached the threshold for treatment before the GA recommended by some recently proposed screening guideline modifications. However, cutting the last half-week GA screening guideline reduced encounters by 2.5%, whereas using ROP Check "if necessary" for BW <1400 and GA <30 weeks for infants of non-Pacific races/ethnicities cut encounters 7.8% without missing treatable infants.
Conclusions:
The proposed modifications of currently accepted screening guidelines for premature infants in the United States could reduce the volume of encounters, but must account for high-risk Pacific race/ethnicity infants, who require more sensitive screening.
More Related Videos
05:47Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
Published on: June 13, 2025
07:15Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Related Concept Videos
Data Validation
Key parameters for method validation include:
Improving Translational Accuracy
Improving Translational Accuracy