Screening for Depression and Anxiety Using a Nonverbal Working Memory Task in a Sample of Older Brazilians: Observational Study of Preliminary Artificial Intelligence Model Transferability
2024

Screening for Depression and Anxiety in Older Brazilians Using AI

Sample size: 69 publication 10 minutes Evidence: moderate

Author Information

Author(s): Georgescu Alexandra Livia, Cummins Nicholas, Molimpakis Emilia, Giacomazzi Eduardo, Rodrigues Marczyk Joana, Goria Stefano

Primary Institution: thymia, International House, London, United Kingdom

Hypothesis

The study aims to illustrate the preliminary transferability of AI models designed to detect high depression and anxiety symptom scores in older Brazilian adults.

Conclusion

The study demonstrates that AI models for depression and anxiety classification can be effectively transferred to a new demographic of older Brazilian adults.

Supporting Evidence

  • The depression classification model achieved an AUC of 0.78, indicating good performance.
  • The anxiety classification model showed an AUC of 0.63, which improved to 0.76 when analyzing first-time exposures.
  • Participants completed activities on a dashboard that included a 5-minute n-back game.
  • Models were trained on data from a diverse population and tested on older Brazilian adults.

Takeaway

Researchers used a game to help find out if older people in Brazil are feeling depressed or anxious, and it worked pretty well.

Methodology

The study involved 69 Brazilian participants aged 51-92 who completed n-back tasks and questionnaires over 6 months to evaluate AI models for depression and anxiety.

Potential Biases

Potential conflict of interest due to the authors' affiliations with thymia Limited and Laços Saúde.

Limitations

The sample size was small, and the self-report instruments may lack sensitivity for older adults.

Participant Demographics

Participants were older Brazilian adults aged 51-92, with a mean age of 71 years.

Statistical Information

P-Value

<0.001

Confidence Interval

95% CI 0.650, 0.871

Statistical Significance

p<0.001

Digital Object Identifier (DOI)

10.2196/55856

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication