Current status of external validation of risk prediction models in the nursing field in China: a scoping review
10.3760/cma.j.cn115682-20240417-02109
- VernacularTitle:国内护理领域风险预测模型开展外部验证现状的范围综述
- Author:
Xinyi ZHOU
1
;
Liyun ZHU
;
Yuan XU
;
Jianhua SUN
;
Haibo DENG
;
Lei WANG
;
Xiaojie WANG
;
Ni YANG
;
Manna SHAO
;
Yufen MA
Author Information
1. 中国医学科学院北京协和医学院护理学院，北京 100144
- Publication Type:Journal Article
- Keywords:
Review;
Predictive model;
External validation;
Nursing research;
Scoping review
- From:
Chinese Journal of Modern Nursing
2025;31(1):100-105
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To describe the current status and research methods of external validation studies of risk prediction models conducted by nursing scholars in China.Methods:Using the search terms "predictive model, external validation, nursing, risk prediction, external validation, nursing" this study searched eight databases (China Biology Medicine disc, China National Knowledge Infrastructure, Wanfang Data, VIP, Embase, PubMed, CINAHL, and Web of Science) for studies published up to June 2023. Literature was imported into EndNote software for organization and deduplication. Two researchers independently conducted initial and secondary screenings by reading the titles, abstracts, and full texts of studies according to inclusion and exclusion criteria. Descriptive analysis was conducted on the included studies under the framework of the Joanna Briggs Institute (JBI) scoping review methodology.Results:A total of 70 studies (15 dissertations and 55 journal articles), primarily published from 2021 to 2023, were included. Among 246 risk prediction models constructed by nurses, 28.5% (70/246) underwent external validation. The models focused on issues such as infection, cognitive impairment, skin injury, thrombosis, and malnutrition. Most studies were conducted in single-center settings (71.4%, 50/70), with temporal validation being the most common type (67.1%, 47/70). The majority of models were presented as nomograms, though some studies had methodological issues in validation.Conclusions:In recent years, new prediction models for specific diseases or endpoints have continued to emerge from nursing research in China, yet few have undergone external validation for clinical application, and the quality and methods of validation require improvement. Future researchers should standardize and strengthen the external validation and optimization of risk prediction models, focusing on clinical applicability to enhance the practical value of these models.
