EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Machine learning for prediction of childhood mental health problems in social care

Ryan Crowley iD, Katherine Parkin iD, Emma Rocheteau iD, Efthalia Massou, Yasmin Friedmann, Ann John iD, Rachel Sippy, Pietro Liò, Anna Moore iD

BJPsych Open · 2025-05 · 11 (3)

DOI: 10.1192/bjo.2025.32 ↗

Kurz gefasst

Fachartikel in der Zeitschrift BJPsych Open, online erschienen am 11.04.2025. Verfasst von Ryan Crowley und 8 weiteren Autor:innen. Beteiligte Einrichtung: New York University Grossman School of Medicine, New York, US. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Rates of childhood mental health problems are increasing in the UK. Early identification of childhood mental health problems is challenging but critical to children’s future psychosocial development. This is particularly important for children with social care contact because earlier identification can facilitate earlier intervention. Clinical prediction tools could improve these early intervention efforts.

Aims

Characterise a novel cohort consisting of children in social care and develop effective machine learning models for prediction of childhood mental health problems.

Method

We used linked, de-identified data from the Secure Anonymised Information Linkage Databank to create a cohort of 26 820 children in Wales, UK, receiving social care services. Integrating health, social care and education data, we developed several machine learning models aimed at predicting childhood mental health problems. We assessed the performance, interpretability and fairness of these models.

Results

Risk factors strongly associated with childhood mental health problems included age, substance misuse and being a looked after child. The best-performing model, a gradient boosting classifier, achieved an area under the receiver operating characteristic curve of 0.75 (95% CI 0.73–0.78). Assessments of algorithmic fairness showed potential biases within these models.

Conclusions

Machine learning performance on this prediction task was promising. Predictive performance in social care settings can be bolstered by linking diverse routinely collected data-sets, making available a range of heterogenous risk factors relating to clinical, social and environmental exposures.

Schlagwörter

Stichwörter der Autor:innen

Mental health servicescommunity mental health teamsmachine learning methodsmedical technologyprecision medicine

Autor:innen und Institutionen

  • New York University Grossman School of Medicine, New York, US.
  • Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK.
  • Department of Psychiatry, University of Cambridge, Cambridge, UK.
  • Cambridge Public Health, University of Cambridge, Cambridge, UK.
  • Department of Computer Science, University of Cambridge, Cambridge, UK.
  • Neath Port Talbot County Borough Council, Port Talbot, UK.

ORCID-Profile: Ryan Crowley, Katherine Parkin, Emma Rocheteau, Ann John, Anna Moore

Zitieren

APA

Crowley, R., Parkin, K., Rocheteau, E., Massou, E., Friedmann, Y., John, A., Sippy, R., Liò, P., & Moore, A. (2025). Machine learning for prediction of childhood mental health problems in social care. BJPsych Open, 11(3). https://doi.org/10.1192/bjo.2025.32

Vancouver

Crowley R, Parkin K, Rocheteau E, Massou E, Friedmann Y, John A, et al. Machine learning for prediction of childhood mental health problems in social care. BJPsych Open. 2025;11(3). doi:10.1192/bjo.2025.32

BibTeX
@article{Crowley2025,
  author = {Crowley, Ryan and Parkin, Katherine and Rocheteau, Emma and Massou, Efthalia and Friedmann, Yasmin and John, Ann and Sippy, Rachel and Liò, Pietro and Moore, Anna},
  title = {Machine learning for prediction of childhood mental health problems in social care},
  journal = {BJPsych Open},
  year = {2025},
  volume = {11},
  number = {3},
  doi = {10.1192/bjo.2025.32}
}
Kostenlos starten →