中国全科医学 ›› 2026, Vol. 29 ›› Issue (08): 1029-1036.DOI: 10.12114/j.issn.1007-9572.2024.0473

• 论著 • 上一篇    

基于多目标规划的成年炎症性肠病患者健康膳食推荐模型研究

尹婷婷, 徒文静, 柏亚妹, 黄丽娜, 李伊婷, 徐桂华*()   

  1. 210023江苏省南京市,南京中医药大学护理学院
  • 收稿日期:2024-08-13 修回日期:2024-12-15 出版日期:2026-03-15 发布日期:2026-02-03
  • 通讯作者: 徐桂华

  • 作者贡献:

    尹婷婷、徒文静、黄丽娜负责研究设计、课题实施、资料分析、论文撰写;李伊婷负责论文修改;柏亚妹、徐桂华负责研究指导、论文审阅。

  • 基金资助:
    国家自然科学基金青年基金资助项目(72204124); 南京中医药大学国自然青年基金经费配套项目(XPT72204124)

Study on Healthy Dietary Recommendation Model for Inflammatory Bowel Disease Patients Based on Multi-objective Planning

YIN Tingting, TU Wenjing, BAI Yamei, HUANG Lina, LI Yiting, XU Guihua*()   

  1. School of Nursing, Nanjing University of Chinese Medicine, Nanjing 210023, China
  • Received:2024-08-13 Revised:2024-12-15 Published:2026-03-15 Online:2026-02-03
  • Contact: XU Guihua

摘要: 背景 膳食是炎症性肠病(IBD)患者健康管理的关键部分,构建健康膳食推荐模型有助于为IBD患者提供工具,辅助膳食管理和疾病康复。 目的 构建成年IBD患者健康膳食推荐模型并初步验证。 方法 2023年9月成立研究小组,采用文献研究和德尔菲专家咨询法明确IBD患者推荐营养素种类及摄入量,筛选具有抗炎或促炎特性的营养素,采用食物-频次矩阵方法识别患者膳食偏好,在此基础上利用多目标优化算法和协同过滤算法建立成年IBD患者健康膳食推荐模型。采用目的性抽样和最大差异抽样,于2023年12月选取在南京中医药大学附属南京中医院IBD中心就诊的20例成年IBD患者,采用多目标粒子群算法和协同过滤算法为IBD患者推荐个性化膳食种类和摄入量,据此验证模型可行性和科学性。 结果 成年IBD患者健康膳食推荐模型需同时满足营养需求、辅助治疗和膳食偏好三大目标。营养需求主要考虑能量、蛋白质、膳食纤维、维生素D、钙和铁6个指标,辅助治疗从膳食纤维、维生素A、维生素C、维生素E、硒、镁和锌等7类抗炎营养素以及能量、脂肪、蛋白质和铁等4大促炎营养进行综合考量,算法求解得出符合20例IBD患者膳食偏好的个性化膳食推荐方案。验证结果显示该模型食物种类推荐平均准确率为95.5%,营养素平均误差为12.60%。 结论 该研究构建的成年IBD患者健康膳食推荐模型准确有效,有助于提升IBD膳食管理精细化。

关键词: 炎症性肠病, 健康膳食, 膳食推荐, 模型构建, 多目标规划

Abstract:

Background

Diet is a critical component of health management for patients with inflammatory bowel disease (IBD), and modeling healthy dietary recommendations can help provide patients with tools to aid in dietary management and disease recovery.

Objective

To construct and validate a healthy dietary recommendation model for adult patients with IBD.

Methods

A research group was set up in September 2023 to clarify the recommended nutrient groups and intakes for IBD patients using literature research and Delphi expert consultation method, then screen nutrients with anti-inflammatory or pro-inflammatory properties, and identify patients' dietary preferences using food-frequency matrix method, finally establish a model for recommending healthy diets to patients with IBD using multi-objective optimization algorithms and collaborative filtering algorithm. Using purposive sampling and maximum variation sampling, 20 IBD adult patients were selected from a tertiary hospital in Nanjing in December 2023, and the model was applied to recommend personalized dietary types and intake for IBD patients using a multi-objective particle swarm algorithm and collaborative filtering algorithm to validate the model's feasibility and scientificity.

Results

The model of healthy dietary recommendations for IBD adult patients needs to meet the three main goals of nutritional needs, complementary therapies and dietary preferences at the same time. Nutritional requirements are mainly considered in six indicators: energy, protein, dietary fiber, vitamin D, calcium and iron; adjuvant therapy is based on the comprehensive consideration of seven anti-inflammatory nutrients such as dietary fiber, vitamin A, vitamin C, vitamin E, selenium, magnesium and zinc, as well as the four major pro-inflammatory nutrients such as energy, fat, protein and iron; and algorithmic solving results in a personalized dietary recommendation plan in line with the dietary preferences of the 20 patients with IBD. The validation results showed that the average accuracy of the model's food group recommendations was 95.5%, and the average nutrient error was 12.60%.

Conclusion

The model of healthy dietary recommendations for IBD adult patients constructed in this study is accurate and effective and helps to improve the refinement of IBD dietary management.

Key words: Inflammatory bowel diseases, Healthy diet, Dietary recommendations, Model construction, Multi-objective planning

中图分类号: