Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Wednesday, July 01, 2020 at 8:00 PM to 10:00 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

Accepted for/Published in: JMIR Formative Research

Date Submitted: Feb 10, 2021
Date Accepted: Apr 4, 2021

The final, peer-reviewed published version of this preprint can be found here:

Evidence of Human-Level Bonds Established With a Digital Conversational Agent: Cross-sectional, Retrospective Observational Study

Darcy A, Daniels J, Salinger D, Wicks P, Robinson A

Evidence of Human-Level Bonds Established With a Digital Conversational Agent: Cross-sectional, Retrospective Observational Study

JMIR Form Res 2021;5(5):e27868

DOI: 10.2196/27868

PMID: 33973854

PMCID: 8150389

Evidence of human-level bond established with a digital conversational agent: An observational study

  • Alison Darcy; 
  • Jade Daniels; 
  • David Salinger; 
  • Paul Wicks; 
  • Athena Robinson

ABSTRACT

Background:

There are more patients in mental distress than healthcare professional hours to support them. While digital tools may help, critics suggest that technology without human empathy prohibits a bond or therapeutic alliance from being formed, limiting their applicability.

Objective:

To investigate whether users of a cognitive behavioral therapy (CBT) based conversational agent report a similar level of therapeutic “bond” to literature from other CBT modalities.

Methods:

Cross sectional, retrospective analysis of de-identified aggregate data from a fully automated conversational agent delivering CBT (“Woebot”) available to smartphone users as “your self-care expert” via app stores in English-speaking countries between November 2019 and August 2020. Working alliance was measured by the Working Alliance Inventory-Short Revised (WAI-SR), and the 2-item Patient Health Questionnaire (PHQ-2) for depression symptom status, with WAI-SR data compared to data abstracted from recent reviews of the scientific literature.

Results:

The study analyzed aggregate, de-identified data from 36,070 adult users self-referred to a self-directed cognitive behavior therapy app for symptoms of anxiety and depression, guided by a fully automated conversational agent. Participants ranged in age from 18 to 78 years and 57% reported female gender. Mean PHQ-2 score was 3.03 (SD 1.79) with 55% scoring over the cutoff score of 3 for depression screening. Within 5 days of initial app use, mean WAI-SR score was 3.36 (SD 0.8) with a mean Bond subscale score of 3.8 (SD 1.0) comparing favorably with recent studies from the literature of traditional outpatient individual CBT and group CBT (mean Bond subscale 4 and 3.8, respectively). Bond did not differ meaningfully among Woebot users by age or gender.

Conclusions:

Woebot users spontaneously report that sharing their feelings and engaging in cognitive restructuring techniques feels easier with a non-human informant, with less potential to feel stigma or shame. Future research might consider further development of these “relational agents” and investigate the role of bond as a mediator of clinical outcomes. Clinical Trial: Not applicable


 Citation

Please cite as:

Darcy A, Daniels J, Salinger D, Wicks P, Robinson A

Evidence of Human-Level Bonds Established With a Digital Conversational Agent: Cross-sectional, Retrospective Observational Study

JMIR Form Res 2021;5(5):e27868

DOI: 10.2196/27868

PMID: 33973854

PMCID: 8150389

Download PDF


Request queued. Please wait while the file is being generated. It may take some time.

© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.