Your browser doesn't support javascript.
loading
Measuring algorithmic bias to analyze the reliability of AI tools that predict depression risk using smartphone sensed-behavioral data.
Adler, Daniel A; Stamatis, Caitlin A; Meyerhoff, Jonah; Mohr, David C; Wang, Fei; Aranovich, Gabriel J; Sen, Srijan; Choudhury, Tanzeem.
Afiliación
  • Adler DA; Cornell Tech, Information Science, 2 W Loop Rd, New York, NY, 10044, USA. daa243@cornell.edu.
  • Stamatis CA; Northwestern University Feinberg School of Medicine, Center for Behavioral Intervention Technologies, Chicago, IL, 60611, USA.
  • Meyerhoff J; Northwestern University Feinberg School of Medicine, Center for Behavioral Intervention Technologies, Chicago, IL, 60611, USA.
  • Mohr DC; Northwestern University Feinberg School of Medicine, Center for Behavioral Intervention Technologies, Chicago, IL, 60611, USA.
  • Wang F; Weill Cornell Medicine, Population Health Sciences, New York, NY, 10065, USA.
  • Aranovich GJ; Cornell Tech, Information Science, 2 W Loop Rd, New York, NY, 10044, USA.
  • Sen S; Michigan Medicine, Department of Psychiatry, Ann Arbor, MI, 48109, USA.
  • Choudhury T; Cornell Tech, Information Science, 2 W Loop Rd, New York, NY, 10044, USA.
Npj Ment Health Res ; 3(1): 17, 2024 Apr 22.
Article en En | MEDLINE | ID: mdl-38649446
ABSTRACT
AI tools intend to transform mental healthcare by providing remote estimates of depression risk using behavioral data collected by sensors embedded in smartphones. While these tools accurately predict elevated depression symptoms in small, homogenous populations, recent studies show that these tools are less accurate in larger, more diverse populations. In this work, we show that accuracy is reduced because sensed-behaviors are unreliable predictors of depression across individuals sensed-behaviors that predict depression risk are inconsistent across demographic and socioeconomic subgroups. We first identified subgroups where a developed AI tool underperformed by measuring algorithmic bias, where subgroups with depression were incorrectly predicted to be at lower risk than healthier subgroups. We then found inconsistencies between sensed-behaviors predictive of depression across these subgroups. Our findings suggest that researchers developing AI tools predicting mental health from sensed-behaviors should think critically about the generalizability of these tools, and consider tailored solutions for targeted populations.

Texto completo: 1 Colección: 01-internacional Base de datos: MEDLINE Idioma: En Revista: Npj Ment Health Res Año: 2024 Tipo del documento: Article País de afiliación: Estados Unidos Pais de publicación: Reino Unido

Texto completo: 1 Colección: 01-internacional Base de datos: MEDLINE Idioma: En Revista: Npj Ment Health Res Año: 2024 Tipo del documento: Article País de afiliación: Estados Unidos Pais de publicación: Reino Unido