Clinical Scorecard: A chatbot-facilitated psychological intervention for older adolescents and young adults experiencing psychological distress: results from a randomized clinical trial conducted in Jordan

At a Glance

Key Highlights

• STARS chatbot intervention significantly reduced anxiety and depression compared to enhanced usual care (EUC) at 3-month follow-up with moderate to large effect sizes.
• High engagement was observed with 66.7% completing at least seven chatbot lessons and 61.4% attending at least four e-helper sessions.
• The intervention addresses treatment gaps in LMICs by providing accessible, stigma-reducing, evidence-based mental health care through a decision-tree logic chatbot.

Guideline-Based Recommendations

Diagnosis

• Use validated symptom checklists such as the Hopkins Symptom Checklist (HSCL) to assess anxiety and depression severity in young adults.

Management

• Implement rule-based chatbot interventions like STARS to deliver stress coping strategies with multimedia and interactive storylines.
• Provide brief weekly support sessions from trained non-specialist helpers (e-helpers) to enhance engagement and adherence.

Monitoring & Follow-up

• Monitor symptom changes using standardized scales (e.g., HSCL, K10) at baseline, post-treatment, and follow-up (e.g., 3 months).
• Track intervention adherence through lesson completion and e-helper session attendance.

Risks

• Be aware of potential high dropout rates in digital interventions; supplement chatbot use with human support to mitigate this.
• Avoid generative AI chatbots for mental health interventions in LMICs due to risks of inaccurate or improper responses; prefer rule-based decision-tree chatbots.

Patient & Prescribing Data

Psychologically distressed young adults in Jordan, representative of LMIC youth populations with limited mental health care access.

STARS chatbot plus e-helper support led to significantly greater reductions in anxiety and depression symptoms compared to enhanced usual care, with number needed to treat around 2.7–3.1 for clinically meaningful improvement.

Clinical Best Practices

• Use human-centred design to develop culturally appropriate digital mental health interventions for LMIC youth.
• Combine digital chatbot interventions with brief, structured human support to improve engagement and fidelity.
• Employ decision-tree logic chatbots to ensure consistent delivery of evidence-based content and reduce risks associated with generative AI.
• Regularly assess symptom severity and treatment response using validated scales to guide ongoing care.

References

• Global prevalence of anxiety and depression in young people
• WHO Sustainable Technology for Adolescents and Youth to Reduce Stress (STARS) intervention
• Effectiveness of digital mental health interventions in LMICs