Feasibility of a Machine Learning-Based Smartphone Application in Detecting Depression and Anxiety in a Generally Senior Population.
NLP
artificial intelligence
behavioral health monitoring
biomarkers
machine learning
mental health screening
smartphone
speech
Journal
Frontiers in psychology
ISSN: 1664-1078
Titre abrégé: Front Psychol
Pays: Switzerland
ID NLM: 101550902
Informations de publication
Date de publication:
2022
2022
Historique:
received:
08
11
2021
accepted:
01
03
2022
entrez:
28
4
2022
pubmed:
29
4
2022
medline:
29
4
2022
Statut:
epublish
Résumé
Depression and anxiety create a large health burden and increase the risk of premature mortality. Mental health screening is vital, but more sophisticated screening and monitoring methods are needed. The Ellipsis Health App addresses this need by using semantic information from recorded speech to screen for depression and anxiety. The primary aim of this study is to determine the feasibility of collecting weekly voice samples for mental health screening. Additionally, we aim to demonstrate portability and improved performance of Ellipsis' machine learning models for patients of various ages. Study participants were current patients at Desert Oasis Healthcare, mean age 63 years (SD = 10.3). Two non-randomized cohorts participated: one with a documented history of depression within 24 months prior to the study (Group Positive), and the other without depression (Group Negative). Participants recorded 5-min voice samples weekly for 6 weeks Protocol completion rate was 61% for both groups. Use beyond protocol was 27% for Group Positive and 9% for Group Negative. The Ellipsis Health App showed an AUC of 0.82 for the combined groups when compared to the PHQ-8 and GAD-7 with a threshold score of 10. Performance was high for senior participants as well as younger age ranges. Additionally, many participants spoke longer than the required 5 min. The Ellipsis Health App demonstrated feasibility in using voice recordings to screen for depression and anxiety among various age groups and the machine learning models using Transformer methodology maintain performance and improve over LSTM methodology when applied to the study population.
Sections du résumé
Background
UNASSIGNED
Depression and anxiety create a large health burden and increase the risk of premature mortality. Mental health screening is vital, but more sophisticated screening and monitoring methods are needed. The Ellipsis Health App addresses this need by using semantic information from recorded speech to screen for depression and anxiety.
Objectives
UNASSIGNED
The primary aim of this study is to determine the feasibility of collecting weekly voice samples for mental health screening. Additionally, we aim to demonstrate portability and improved performance of Ellipsis' machine learning models for patients of various ages.
Methods
UNASSIGNED
Study participants were current patients at Desert Oasis Healthcare, mean age 63 years (SD = 10.3). Two non-randomized cohorts participated: one with a documented history of depression within 24 months prior to the study (Group Positive), and the other without depression (Group Negative). Participants recorded 5-min voice samples weekly for 6 weeks
Results
UNASSIGNED
Protocol completion rate was 61% for both groups. Use beyond protocol was 27% for Group Positive and 9% for Group Negative. The Ellipsis Health App showed an AUC of 0.82 for the combined groups when compared to the PHQ-8 and GAD-7 with a threshold score of 10. Performance was high for senior participants as well as younger age ranges. Additionally, many participants spoke longer than the required 5 min.
Conclusion
UNASSIGNED
The Ellipsis Health App demonstrated feasibility in using voice recordings to screen for depression and anxiety among various age groups and the machine learning models using Transformer methodology maintain performance and improve over LSTM methodology when applied to the study population.
Identifiants
pubmed: 35478769
doi: 10.3389/fpsyg.2022.811517
pmc: PMC9037748
doi:
Types de publication
Journal Article
Langues
eng
Pagination
811517Informations de copyright
Copyright © 2022 Lin, Nazreen, Rutowski, Lu, Harati, Shriberg, Chlebek and Aratow.
Déclaration de conflit d'intérêts
All authors were employed by the company Ellipsis Health Inc. The authors declare that this study received funding from Ellipsis Health Inc. Ellipsis Health Inc. had the following involvement in the study: study design, collection, analysis, interpretation of data, the writing of this article and the decision to submit it for publication.
Références
Psychiatry Investig. 2019 Apr;16(4):262-269
pubmed: 30947496
JMIR Mhealth Uhealth. 2018 Aug 22;6(8):e10748
pubmed: 30135050
J Affect Disord. 2018 Jul;234:117-123
pubmed: 29525352
J Consult Clin Psychol. 2020 Jan;88(1):1-13
pubmed: 31697093
CMAJ. 2012 Feb 21;184(3):E191-6
pubmed: 22184363
BMC Public Health. 2019 Dec 16;19(1):1649
pubmed: 31839006
BMC Public Health. 2019 Aug 07;19(1):1059
pubmed: 31391010
NCHS Data Brief. 2018 Feb;(303):1-8
pubmed: 29638213
JAMA. 2016 Jan 26;315(4):380-7
pubmed: 26813211
JMIR Mhealth Uhealth. 2018 Jan 23;6(1):e26
pubmed: 29362211
Int J Bipolar Disord. 2020 Jan 10;8(1):2
pubmed: 31919635
Geriatrics (Basel). 2017 Nov 15;2(4):
pubmed: 31011045
Arch Intern Med. 2006 May 22;166(10):1092-7
pubmed: 16717171
N Engl J Med. 2015 Nov 19;373(21):2059-68
pubmed: 26580998
JMIR Mhealth Uhealth. 2015 Nov 04;3(4):e101
pubmed: 26537656
J Med Internet Res. 2019 Sep 25;21(9):e14567
pubmed: 31573916
World Psychiatry. 2017 Oct;16(3):287-298
pubmed: 28941113
Psychiatr Serv. 2019 Jul 1;70(7):538-544
pubmed: 30914003
Psychiatr Serv. 2011 Oct;62(10):1167-79
pubmed: 21969643
Psychiatr Serv. 2017 Jul 1;68(7):660-666
pubmed: 28196461
Clin Ophthalmol. 2020 Dec 03;14:4229-4251
pubmed: 33299297
Curr Pharm Des. 2012;18(36):5936-57
pubmed: 22681171
Med Care. 2004 Dec;42(12):1194-201
pubmed: 15550799
Laryngoscope Investig Otolaryngol. 2020 Jan 31;5(1):96-116
pubmed: 32128436
J Affect Disord. 2021 Jan 1;278:122-130
pubmed: 32956961
Biometrics. 1988 Sep;44(3):837-45
pubmed: 3203132
Gen Hosp Psychiatry. 2019 Jan - Feb;56:13-18
pubmed: 30508772