Machine learning and natural language processing for the identification of potential mental disorders among
Shanquan Chen1, Ting Dang2, Mengjie Qian3
1School of Public Health, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China. shanquan.chen@hku.hk.
Insights
Combining natural language processing (NLP) of children's essays with key risk factors significantly improves early detection of mental health disorders. This approach offers a scalable and objective method for identifying potential issues in school-aged children.
Area of Science:
- Child and Adolescent Psychiatry
- Computational Linguistics
- Public Health Screening
Background:
- Early identification of childhood mental health disorders is crucial for public health.
- Current screening methods rely on resource-intensive observer reports and may miss subtle symptoms.
- Analyzing children's linguistic expression offers a scalable and objective screening alternative.
Purpose of the Study:
- To evaluate if combining natural language processing (NLP) of children's essays with traditional risk factors enhances mental health disorder detection.
- To compare the predictive performance of integrated models against models using single data sources.
Main Methods:
- Prospective analysis of the National Child Development Study (NCDS) cohort (N=8,981).
- Used creative writing essays from age 11, alongside birth, age 7, and age 11 assessment data.
- Machine learning models integrated traditional risk factors and NLP-derived linguistic features to predict mental health disorders (BSAG/Rutter A Scale).
Main Results:
- Integrated models combining NLP essay features and top risk factors showed significantly higher predictive capability (AUROC: 0.77 at 95th percentile BSAG).
- This approach outperformed models using all variables (AUROC: 0.70) or essay features alone (AUROC: 0.67).
- Key predictors included perinatal, socioeconomic, parental engagement, and child BMI variables.
Conclusions:
- Integrating NLP analysis of children's essays with key risk factors substantially improves the identification of potential mental health disorders.
- This novel approach shows promise for developing scalable, objective screening tools.
- Further validation in diverse pediatric populations is needed before clinical implementation.
Background:
Early identification of childhood mental health disorders is a critical public health objective. Existing screening approaches, largely dependent on observer reports, are resource-intensive and may overlook subtle internalized symptoms. The analysis of children's linguistic expression presents a scalable and potentially more objective alternative. This study evaluates whether combining natural language processing (NLP) of children's essays with conventional risk factors improves the detection of mental health difficulties in school-age populations, relative to models based on a single data source.
Methods:
We conducted a prospective analysis using data from the UK-based National Child Development Study (NCDS), a national birth cohort initiated in 1958. Data from birth, age 7, and age 11 assessments were analyzed. The final sample included 8,981 children (4,428 [49.3%] female) who completed a creative writing essay at age 11 describing their imagined life at age 25. Predictors comprised traditional risk factors (perinatal, socioeconomic, and parental engagement variables) and linguistic features computationally extracted from the essays. The primary outcome was potential mental health disorder at age 11, defined as scoring above the 95th or 90th percentile on the teacher-completed Bristol Social Adjustment Guide (BSAG). The mother-completed Rutter A Scale was used for sensitivity analysis. Machine learning models incorporating various predictor combinations were developed, and their predictive performance was evaluated using area under the receiver operating characteristic (AUROC) values.
Results:
Using BSAG 95th percentile threshold, models combining top five selected variables with essay features achieved significantly higher predictive capability (AUROC:0.77, 95%CI:0.71-0.83) compared to models using all variables (AUROC:0.70, 95%CI:0.63-0.76) or essay features alone (AUROC:0.67, 95%CI:0.60-0.74). At 90th percentile threshold, this integrated approach showed similar improvement (AUROC:0.81, 95%CI:0.78-0.85). Key predictors included gestational length, maternal parity, parental age, residential characteristics, parental engagement metrics, and children's body mass index. Sensitivity analyses using Rutter A Scale confirmed these findings.
Conclusions:
In this prospective birth cohort study, integrating NLP analysis of children's essays with a small set of key risk factors substantially improved the identification of potential mental health disorders. This integrated approach represents a potential paradigm for developing scalable, objective screening tools, but requires validation in contemporary, diverse pediatric populations before clinical consideration.
