Effectiveness of a Web-based and Mobile Therapy Chatbot on Anxiety and Depressive Symptoms in Subclinical Young Adults: Randomized Controlled Trial
Bibliographic record
Abstract
BACKGROUND: There has been an increased need to provide specialized help for people with depressive and anxiety symptoms, particularly teenagers and young adults. There is evidence from a 2-week intervention that chatbots (eg, Woebot) are effective in reducing depression and anxiety, an effect that was not detected in the control group that was provided self-help materials. Although chatbots are a promising solution, there is limited scientific evidence for the efficacy of agent-guided cognitive behavioral therapy (CBT) outside the English language, especially for highly inflected languages. OBJECTIVE: This study aimed to measure the efficacy of Fido, a therapy chatbot that uses the Polish language. It targets depressive and anxiety symptoms using CBT techniques. We hypothesized that participants using Fido would show a greater reduction in anxiety and depressive symptoms than the control group. METHODS: We conducted a 2-arm, open-label, randomized controlled trial with 81 participants with subclinical depression or anxiety who were recruited via social media. Participants were divided into experimental (interacted with a fully automated Fido chatbot) and control (received a self-help book) groups. Both intervention methods addressed topics such as general psychoeducation and cognitive distortion 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 primary outcomes, including the levels of depression, anxiety, worry tendencies, satisfaction with life, and loneliness at baseline, after the 2-week intervention and at the 1-month follow-up. We also controlled for secondary outcomes, including engagement and frequency of use. RESULTS: There were no differences in anxiety and depressive symptoms between the groups at enrollment and baseline. After the intervention, depressive and anxiety symptoms were reduced in both groups (chatbot: n=36; control: n=38), which remained stable at the 1-month follow-up. Loneliness was not significantly different between the groups after the intervention, but an exploratory analysis showed a decline in loneliness among participants who used Fido more frequently. Both groups used their intervention technique with similar frequency; however, the control group spent more time (mean 117.57, SD 72.40 minutes) on the intervention than the Fido group (mean 79.44, SD 42.96 minutes). CONCLUSIONS: We did not replicate the findings from previous (eg, Woebot) studies, as both arms yielded therapeutic effects. However, such results are in line with other research of Internet interventions. Nevertheless, Fido provided sufficient help to reduce anxiety and depressive symptoms and decreased 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. TRIAL REGISTRATION: ClinicalTrials.gov NCT05762939; https://clinicaltrials.gov/study/NCT05762939; Open Science Foundation Registry 2cqt3; https://osf.io/2cqt3.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.005 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.005 | 0.003 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.011 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".