# Development and validation of a nutritional risk prediction model for patients with ulcerative colitis: a single-center retrospective study

## TL;DR

This study created a model to predict nutritional risk in ulcerative colitis patients using clinical data, offering a tool for early intervention.

## Contribution

A new nomogram model was developed and validated for identifying nutritional risk in UC patients using routine clinical parameters.

## Key findings

- The model achieved an AUC of 0.914 in internal validation and 0.875 in temporal validation.
- Five independent predictors were identified: diet, gender, disease activity, albumin, and hemoglobin.
- Calibration curves and DCA confirmed the model's clinical utility and accuracy.

## Abstract

Nutritional risk is commonly overlooked in patients with ulcerative colitis (UC) due to a lack of specific assessment tools. This study developed and validated a prediction model to identify nutritional risk using routine clinical parameters. Conducted at a tertiary hospital in China, this retrospective analysis utilized data from 451 UC patients between January 2016 and October 2023, with temporal validation in a cohort of 115 patients from November 2023 to September 2024. Through multivariable logistic regression, we identified five independent predictors: diet, gender, disease activity, albumin levels, and hemoglobin levels, and created a nomogram. The model demonstrated strong performance, achieving an area under the ROC curve (AUC) of 0.914 in internal validation and 0.875 in temporal validation. Calibration curves showed good agreement between predicted and observed outcomes, and DCA confirmed clinical utility across different risk thresholds. In conclusion, our nomogram offers clinicians a practical tool for early nutritional intervention in UC, potentially reducing long-term complications. However, as a preliminary model derived from a single center, it requires further multi-center external validation before widespread clinical application.

The online version contains supplementary material available at 10.1038/s41598-025-33379-8.