A mixed-methods feasibility study of a novel AI-enabled, web-based, clinical decision support system for the treatment of major depression in adults

Sabrina Qassim, Grace Golden, Dominique Slowey, Mary Sarfas, Kate Whitmore, Tamara Perez, Elizabeth Strong, Eryn Lundrigan, Marie Jeanne Fradette, Jacob Baxter, Bennet Desormeau, Myriam Tanguay-Sela, Christina Popescu, Sonia Israel, Kelly Perlman, Caitrin Armstrong, Robert Fratila, Joseph Mehltretter, Karl Looper, Warren SteinerSoham Rej, Jordan F. Karp, Katherine Heller, Sagar V. Parikh, Rebecca McGuire-Snieckus, Manuela Ferrari, Howard Margolese, David Benrimoh

Research output: Contribution to journalArticlepeer-review

Abstract

Background: The objective of this paper is to discuss perceived clinical utility and impact on physician-patient relationship of a novel, artificial-intelligence (AI) enabled clinical decision support system (CDSS) for use in treating adults with major depression. Methods: A single arm, naturalistic follow-up study aimed at assessing the acceptability and useability of the software. Patients had a baseline appointment, followed by a minimum of two appointments with the CDSS. Study exit questionnaires and interviews were conducted to assess perceived clinical utility, impact on patient-physician relationship, and understanding and trust. 7 physicians and 17 patients, of which 14 completed, consented to participate. Results: 86 % of physicians (6/7) felt the information provided by the CDSS provided more comprehensive understanding of patient situations. 71 % (5/7) felt the information was helpful. 86 % of physicians (6/7) reported the AI/predictive model was useful when deciding treatment. 62 % of patients (8/13) reported improved care due to the tool, and 46 %(6/13) reported a significantly or somewhat improved physician-patient relationship 54 % reported no change. 71 % of physicians (5/7) and 62 % of patients (8/13) rated trusting the tool. Limitations: Small sample size and treatment changes prior to CDSS introduction limits ability to verify impact on outcomes. Conclusions: Qualitative results from 12 patient exit interviews are analyzed and presented. Findings suggest physicians perceived the tool as useful in conducting appointments and used it while deciding treatment. Physicians and patients generally found the tool trustworthy, and it may have positive effects on physician-patient relationships. (Study identifier: NCT04061642).

Original languageEnglish (US)
Article number100677
JournalJournal of Affective Disorders Reports
Volume14
DOIs
StatePublished - Dec 2023

Keywords

  • Artificial intelligence
  • Clinical decision support system
  • Feasibility
  • Major depressive disorder
  • Physician-patient relationship
  • Trust

ASJC Scopus subject areas

  • Clinical Psychology
  • Psychiatry and Mental health

Fingerprint

Dive into the research topics of 'A mixed-methods feasibility study of a novel AI-enabled, web-based, clinical decision support system for the treatment of major depression in adults'. Together they form a unique fingerprint.

Cite this