Development and validation of a predictive model for healthcare-seeking time in patients with diabetic foot
10.3760/cma.j.cn115682-20240531-03060
- VernacularTitle:糖尿病足患者就医时间预测模型的构建及验证
- Author:
Shuqing ZHU
1
;
Xueke LI
;
Zichen JIN
;
Gang CHEN
;
Youyou ZHAI
;
Yawei ZHANG
;
Teng LI
Author Information
1. 郑州大学第一附属医院内分泌与代谢科，郑州 450052
- Publication Type:Journal Article
- Keywords:
Diabetic foot;
Healthcare-seeking time;
Predictive model
- From:
Chinese Journal of Modern Nursing
2025;31(7):926-932
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the factors influencing healthcare-seeking time in diabetic foot patients and to develop and validate a predictive model for healthcare-seeking time.Methods:A total of 299 diabetic foot patients hospitalized in the Department of Endocrinology and Metabolism at the First Affiliated Hospital of Zhengzhou University from March 2023 to January 2024 were recruited for model development and internal validation. Sixty additional patients from the Second Affiliated Hospital of Zhengzhou University from September 2023 to January 2024 were used for external validation. Kaplan-Meier survival curves were used to estimate healthcare-seeking times. Cox regression analysis identified influencing factors and constructed the model. Random Survival Forest (RSF) was employed for variable selection and model construction. Internal validation was conducted using 10-fold cross-validation, and model evaluation utilized the integrated Brier score, C-index, and prediction error curve. Results:Kaplan-Meier analysis revealed that education level, foot self-care ability, lower extremity vascular disease, and disease perception significantly influenced healthcare-seeking time ( P<0.05). Cox regression identified gender, income level, medical payment method, living situation, marital status, ulcer history, social support, disease perception, and healthcare behavior perception as significant influencing factors ( P<0.05). RSF variable selection indicated that social support, disease perception, e-health literacy, healthcare behavior perception, and age were the most valuable factors for model construction. In external validation, the Brier scores for the Cox regression and RSF models were 0.059 and 0.088, respectively, while the C-indices were 0.862 and 0.683. Prediction error curves showed that the Cox regression model had lower prediction errors and higher predictive performance. Conclusions:The Cox regression model demonstrated superior performance and can assist nurses in effectively identifying high-risk populations for delayed healthcare-seeking in diabetic foot patients. This allows for timely interventions to improve healthcare behavior and reduce delays.
