1. Development and validation of a nomogram prediction model for ADHD in children based on individual, family, and social factors.
- Author
-
Gao, Ting, Yang, Lan, Zhou, Jiayu, Zhang, Yu, Wang, Laishuan, Wang, Yan, and Wang, Tianwei
- Subjects
- *
CHILDREN with intellectual disabilities , *NOMOGRAPHY (Mathematics) , *SOCIAL factors , *MACHINE learning , *PREDICTION models , *ATTENTION-deficit hyperactivity disorder , *INTELLECTUAL disabilities - Abstract
A reliable, user-friendly, and multidimensional prediction tool can help to identify children at high risk for ADHD and facilitate early recognition and family management of ADHD. We aimed to develop and validate a risk nomogram for ADHD in children aged 3–17 years in the United States based on clinical manifestations and complex environments. A total of 141,356 cases were collected for the prediction model. Another 54,444 cases from a new data set were utilized for performing independent external validation. The LASSO regression was used to control possible variables. A final risk nomogram for ADHD was established based on logistic regression, and the discrimination and calibration of the established nomogram were evaluated by bootstrapping with 1000 resamples. A final risk nomogram for ADHD was established based on 13 independent predictors, including behavioral problems, learning disabilities, age, intellectual disabilities, anxiety symptoms, gender, premature birth, maternal age at childbirth, parent-child interaction patterns, etc. The C-index of this model was 0.887 in the training set, and 0.862 in the validation set. Internal and external validation proved that the model was reliable. A nomogram, a statistical prediction tool that assesses individualized ADHD risk for children is helpful for the early identification of children at high risk for ADHD and the construction of a conceptual model of society-family-school collaborative diagnosis, treatment, and management of ADHD. To construct a conceptual model of society-family-school collaborative diagnosis, treatment, and management of ADHD. We employed machine learning algorithms to identify 13 independent prognostic factors of ADHD from multi-dimensional predictors. We successfully developed a highly accurate early ADHD prediction model with a C-index of 0.887. This reliable and user-friendly tool incorporates multiple dimensions of prediction, enabling g parents, teachers, and physicians to identify children at a high risk of ADHD. By empowering parents, teachers, and physicians with the ability to identify children at high risk, we can facilitate early recognition and family management, leading to improved outcomes for children with ADHD. [Display omitted] • We developed a nomogram for ADHD in children aged 3-17, which can aid in early recognition and family management of ADHD. • The nomogram was validated externally by using another database, proving its reliability in accurately predicting ADHD risk. • This nomogram forms a collaborative diagnostic, therapeutic, and management model involving society, family, and school. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF