EUVIMEDEUVIMED entdecken
Systematische ÜbersichtMetaanalyseOpen Access · CC BY 4.0

Machine learning algorithms and their predictive accuracy for suicide and self-harm: Systematic review and meta-analysis

Machine-Learning-Algorithmen und ihre Vorhersagegenauigkeit für Suizid und Selbstverletzung: systematische Übersicht und Metaanalyse

Matthew J Spittal iD, Xianglin Aneta Guo, Laurant Kang iD, Olivia J Kirtley iD, Angela Clapperton iD, Keith Hawton iD, Nav Kapur, Jane Pirkis, Greg Carter iD

PLOS Medicine · 11.09.2025 · 22 (9) · e1004581

DOI: 10.1371/journal.pmed.1004581 ↗

Auf einen Blick

53 Studien untersuchten Machine-Learning-Algorithmen zur Vorhersage von Suizid und hospitalbehandelter Selbstverletzung. Die Vorhersagegenauigkeit war für Screening oder Risikoselektion zu niedrig.

Fragestellung

Wie genau sagen Machine-Learning-Algorithmen Suizid und hospitalbehandelte Selbstverletzung voraus?

Methode

Systematische Übersicht und Metaanalyse; Suche in PubMed, PsycINFO, Scopus, EMBASE, IEEE, Medline, CINALH und Web of Science bis 30. April 2025.

Ergebnisse

53 Studien. AUC 0,69–0,93; Sensitivität 45 %–82 %; Spezifität 91 %–95 %; LR+ 6,5–9,9; LR− 0,2–0,6; PPV 6 %–17 %.

Einschränkungen

Ausgeschlossene relevante Studien ohne ausreichende Daten; unterschiedliche Nachbeobachtungszeiten zwischen den Studien.

Schlussfolgerung der Autor:innen

Die Genauigkeit war zu niedrig für Screening oder zur Priorisierung Hochrisikopersonen für Interventionen.

Kernaussagen

  • 53 Studien zur Vorhersage von Suizid und hospitalbehandelter Selbstverletzung
  • AUC 0,69–0,93; Sensitivität 45 %–82 %; Spezifität 91 %–95 %
  • Laut Autor:innen zu ungenau für Screening und Risikoselektion

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

Kurz gefasst

Systematische Übersicht in der Zeitschrift PLOS Medicine, online erschienen am 11.09.2025. Verfasst von Matthew J Spittal und 8 weiteren Autor:innen. Beteiligte Einrichtung: Centre for Mental Health and Community Wellbeing, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, Australia. Hauptthemen laut MEDLINE: Self-Injurious Behavior, Machine Learning, Suicide. Die Arbeit wurde bisher 42-mal zitiert (Crossref) und nennt 98 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Health and Medical Research Council, KU Leuven, Suicide Prevention Australia.

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

There has been rapid expansion in the development of machine learning algorithms to predict suicidal behaviours. To test the accuracy of these algorithms for predicting suicide and hospital-treated self-harm, we undertook a systematic review and meta-analysis. The study was registered (PROSPERO CRD42024523074). Methods and findings We searched PubMed, PsycINFO, Scopus, EMBASE, IEEE, Medline, CINALH and Web of Science from database inception until 30 April 2025 to identify studies using machine learning algorithms to predict suicide, self-harm and a combined suicide/self-harm outcome. Studies were included if they examined suicide or hospital-treated self-harm outcomes using a case-control, case-cohort or cohort study design. Studies were excluded if they used self-reported outcomes or examined outcomes using other study designs. Accuracy was assessed using statistical methods appropriate for diagnostic accuracy studies. Fifty-three studies met the inclusion criteria. The area under the receiver operating characteristic curves ranged from 0.69 to 0.93. Sensitivity was 45%–82% and specificity was 91%–95%. Positive likelihood ratios were 6.5–9.9 and negative likelihood values were 0.2–0.6. Using in-sample prevalence values, the positive predictive values ranged from 6% to 17%. Using out-of-sample prevalence values at an LR+ value of 10, the positive predictive value was 0.1% in low prevalence populations, 17% in medium prevalence populations and 66% in high prevalence populations. The main study limitations were the exclusion of relevant studies where we could not extract sufficient information to calculate accuracy statistics and between-study differences in the follow-up time over which the outcomes were observed.

Conclusions

The accuracy of machine learning algorithms for predicting suicidal behaviour is too low to be useful for screening (case finding) or for prioritising high-risk individuals for interventions (treatment allocation). For hospital-treated self-harm populations, management should instead include three components for all patients: a needs-based assessment and response, identification of modifiable risk factors with treatment intended to reduce those exposures, and implementation of demonstrated effective aftercare interventions.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Self-Injurious BehaviorMachine LearningSuicideHumansRisk AssessmentAlgorithms

Autor:innen und Institutionen

  • Centre for Mental Health and Community Wellbeing, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, Australia.
  • Hunter New England Local Health District, Waratah, Australia.
  • Center for Contextual Psychiatry, KU Leuven, Leuven, Belgium.
  • Centre for Suicide Research, Department of Psychiatry, University of Oxford, Warneford Hospital, Oxford, United Kingdom.
  • National Confidential Inquiry into Suicide and Safety in Mental Health (NCISH), Centre for Mental Health and Safety, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, United Kingdom.
  • NIHR Greater Manchester Patient Safety Research Collaboration, University of Manchester, Manchester, United Kingdom.

ORCID-Profile: Matthew J Spittal, Laurant Kang, Olivia J Kirtley, Angela Clapperton, Keith Hawton, Greg Carter

Förderung

  • National Health and Medical Research Council
  • KU Leuven
  • Suicide Prevention Australia
  • National Institute for Health Research Greater Manchester Patient Safety Research Collaboration
  • Mersey Care NHS Foundation

Zitieren

APA

Spittal, M. J., Guo, X. A., Kang, L., Kirtley, O. J., Clapperton, A., Hawton, K., Kapur, N., Pirkis, J., & Carter, G. (2025). Machine learning algorithms and their predictive accuracy for suicide and self-harm: Systematic review and meta-analysis. PLOS Medicine, 22(9), e1004581. https://doi.org/10.1371/journal.pmed.1004581

Vancouver

Spittal MJ, Guo XA, Kang L, Kirtley OJ, Clapperton A, Hawton K, et al. Machine learning algorithms and their predictive accuracy for suicide and self-harm: Systematic review and meta-analysis. PLOS Medicine. 2025;22(9):e1004581. doi:10.1371/journal.pmed.1004581

BibTeX
@article{Spittal2025,
  author = {Spittal, Matthew J and Guo, Xianglin Aneta and Kang, Laurant and Kirtley, Olivia J and Clapperton, Angela and Hawton, Keith and Kapur, Nav and Pirkis, Jane and Carter, Greg},
  title = {Machine learning algorithms and their predictive accuracy for suicide and self-harm: Systematic review and meta-analysis},
  journal = {PLOS Medicine},
  year = {2025},
  volume = {22},
  number = {9},
  pages = {e1004581},
  doi = {10.1371/journal.pmed.1004581}
}
Kostenlos starten →