Effectiveness of a Web-based and Mobile Therapy Chatbot on Anxiety and Depressive Symptoms in Subclinical Young Adults: A Randomized Controlled Trial (Preprint)

Abstract

Recent years have brought forth an increase in the need for providing specialized help for people with depressive and anxiety symptoms, particularly for teenagers and young adults. There is evidence that chatbots are effective in reducing depression and anxiety in a two-week intervention, an effect that was not detected in the control group that was provided with self-help materials. Although it seems that chatbots are a promising solution, there is limited scientific evidence for the efficacy of agent-guided CBT outside the English language, especially for highly inflected languages.

This study aimed to measure the efficacy of Fido, a therapy chatbot that uses the Polish language. It targets depressive and anxiety symptoms using Cognitive-Behavioral Therapy techniques.

We conducted a two-arm, open-label RCT on 81 participants with subclinical depression or anxiety recruited via social media. Participants were divided into two groups: experimental (that received a fully-automated Fido chatbot) and control (that received a self-help book). Both intervention methods addressed topics such as general psychoeducation and cognitive distortions identification and modification via socratic questioning. The chatbot also featured suicidal ideation identification and redirection to suicide hotlines. We used self-assessment scales to measure depression, anxiety, worry tendencies, satisfaction with life and loneliness at baseline, after 2 weeks intervention and at 1-month follow-up. We also controlled engagement and frequency of usage.

There was no difference in anxiety and depressive symptoms at the enrollment and the baseline measurement. After intervention, depressive and anxiety symptoms were reduced in both groups (chatbot: N=36, control: N=38), which remained stable at 1-month follow-up. There was no loneliness difference after intervention in both groups, however in-depth exploratory analysis provided evidence that there was a decline among participants that had been using Fido more frequently. Both groups used their intervention technique with similar frequency, however the control group (mean 117.57 minutes, SD 72.40) spent more time on the intervention than the Fido group (mean 79.44 minutes, SD 42.96).

We didn’t manage to replicate previous findings from Woebot studies, as both arms yielded therapeutic effects. However, such results stand in line with other Internet interventions research. Fido provides sufficient help in reducing anxiety and depressive symptoms. It also managed to decrease perceived loneliness among high-frequency users, which is one of the first pieces of evidence of chatbot efficacy with agents that use a highly inflected language. Further research is needed to determine the long-term, real-world effectiveness of Fido and its efficacy in a clinical sample.

ClinicalTrials.gov Registry: NCT05762939 (retrospective). Open Science Foundation Registry: 2cqt3 (prospective).