Related Experiment Video
Updated: Feb 12, 2026
Involving Individuals with Developmental Language Disorder and Their Parents/Carers in Research Priority Setting
Published on: June 6, 2020
A Large Language Model Approach to Functional Status Scale Assessment
Blake Martin1,2,3, Anna M Janas1,3, Kristen R Miller1,3
1Section of Critical Care Medicine, Department of Pediatrics, University of Colorado School of Medicine, Aurora, CO.
Insights
A fine-tuned artificial intelligence (AI) model, FSS-AI, can estimate Functional Status Scale (FSS) scores in critically ill children. The AI showed moderate agreement with manual scores and could identify normal versus abnormal FSS.
Area of Science:
- Artificial Intelligence in Medicine
- Pediatric Critical Care
- Health Informatics
Background:
- Estimating functional status in critically ill children is crucial for care planning and outcomes assessment.
- The Functional Status Scale (FSS) is a common tool, but manual scoring can be time-consuming and subjective.
- Leveraging advanced AI like GPT-4o offers potential for automated and efficient FSS scoring.
Purpose of the Study:
- To develop and evaluate a fine-tuned generative pretrained transformer (GPT)-4o artificial intelligence (AI) model, named FSS-AI, for estimating FSS scores in critically ill children.
- To assess the accuracy and agreement of AI-generated FSS scores compared to manually assigned scores across different hospitalization timepoints.
Main Methods:
- A secondary analysis of a prospective cohort of mechanically ventilated children (1 month to 18 years) was performed.
- Patient notes from baseline, PICU transfer, and hospital discharge were used to train and test the custom GPT-4o model (FSS-AI).
- FSS-AI performance was evaluated by comparing its generated scores against prospectively determined manual FSS scores using weighted Cohen's Kappa and accuracy metrics.
Main Results:
- FSS-AI analyzed 428 patient notes, demonstrating moderate agreement with manual FSS scores at baseline (Kappa=0.59) and discharge (Kappa=0.51).
- The model showed good discrimination between normal and abnormal FSS scores, with highest accuracy (0.90) and PPV (0.95) at baseline.
- FSS-AI identified new morbidities at discharge with 75% accuracy and 56% sensitivity.
Conclusions:
- A custom GPT-4o model (FSS-AI) can effectively estimate FSS scores in critically ill children at various hospitalization stages.
- The AI tool shows moderate agreement with manual scoring and can differentiate between normal and abnormal functional status.
- FSS-AI demonstrates potential as an efficient tool for assessing functional status and identifying new morbidities in pediatric critical care settings.
Objectives:
To develop a fine-tuned version of the generative pretrained transformer (GPT)-4o artificial intelligence (AI) model able to estimate Functional Status Scale (FSS) scores among critically ill children.
Design:
Secondary analysis of a prospective, observational cohort of critically ill children 1 month to 18 years old who required invasive mechanical ventilation for greater than or equal to 3 days. Four patient notes from each of three hospitalization timepoints-baseline (history & physical), PICU transfer, and hospital discharge-along with retrospectively assigned FSS scores were used to train the model. The resulting custom GPT (hereafter, FSS-AI) was then applied to the remaining 428 notes. The GPT-generated FSS scores were then compared with the manually assigned scores determined prospectively during the original study.
Setting:
A single, quaternary-care academic pediatric hospital.
Patients:
Children who completed the original study, survived to discharge, and had FSS scores documented.
Interventions:
None.
Measurements And Main Results:
FSS-AI analyzed 428 notes from 147 patients over 255 minutes (averaging 35.7 s/note). FSS-AI demonstrated moderate agreement with manually determined total FSS scores at the pre-illness baseline (weighted Cohen's Kappa, 0.59; 95% CI, 0.49-0.70) and hospital discharge (0.51; 95% CI, 0.43-0.58) timepoints, with slightly lower agreement at PICU transfer (0.45; 95% CI, 0.37-0.54). For discrimination of normal total FSS scores (6-7) from abnormal scores (≥ 8), FSS-AI accuracy and positive predictive value were highest at the pre-illness baseline (0.90 and 0.95, respectively) and hospital discharge (0.81-0.75) timepoints. FSS-AI identified children with a new morbidity at hospital discharge (total FSS increase ≥ 3 or domain FSS increase ≥ 2) with accuracy and sensitivity of 0.75 and 0.56.
Conclusions:
A custom version of GPT-4o was able to estimate FSS scores at multiple hospitalization timepoints. The tool demonstrated moderate agreement with manually determined scores, could discriminate children with normal vs. abnormal FSS (best performance at baseline and hospital discharge timepoints), and had fair accuracy for detecting new morbidities present at hospital discharge.
More Related Videos
Related Concept Videos
01:10Higher Mental Functions of the Brain: Language
Language formation and comprehension take place in the dominant hemisphere. The dominant hemisphere is responsible for understanding the meaning of spoken, written, or sign language, as well as the ability to communicate. For most people, the left hemisphere is the dominant one. The right hemisphere, then, gives tone and emotional context to the...
01:16Language
Corballis and Suddendorf (2007) and Tomasello and Rakoczy (2003) highlight the role of language in...
02:41pH Scale
01:24Components of Language
01:22Language Development
The critical period for language acquisition suggests that the ability to acquire language is at its peak early in life. As people age, this proficiency decreases. Language development begins very...
01:27Language and Cognition