Evaluation and Development of a Safety-Oriented Chatbot for Women's Sexual Well-Being: A Feasibility Study Focused on Methodology - Scorecard - MDSpire

Evaluation and Development of a Safety-Oriented Chatbot for Women's Sexual Well-Being: A Feasibility Study Focused on Methodology

  • By

  • Alice E McGee

  • Guy Parsons

  • Liudmila Zhaunova

  • Alison Paul

  • Aliaksandr Kazlou

  • Yihan Xu

  • Heorhi Stsefanovich

  • Anna Klepchukova

  • András Meczner

  • April 1, 2026

  • 0 min

Share

Clinical Scorecard: Evaluation and Development of a Safety-Oriented Chatbot for Women's Sexual Well-Being: A Feasibility Study Focused on Methodology

At a Glance

CategoryDetail
ConditionWomen's Sexual Well-Being
Key MechanismsUtilizes LLMs for personalized health education while addressing safety concerns through structured governance.
Target PopulationIndividuals seeking information on sexual health and well-being.
Care SettingCommercial mobile application integrated with a sexual wellness program.

Key Highlights

  • Developed an LLM-based educational chatbot named 'Expert' focused on sexual well-being.
  • Incorporated a secondary LLM, 'Critic', for semi-automated evaluation of chatbot responses.
  • Emphasized safety and risk governance as central design outcomes.
  • Conducted user journey mapping to identify potential hazards and inform design decisions.
  • Implemented structured clinical risk assessments throughout the development process.

Guideline-Based Recommendations

Diagnosis

  • The chatbot is not designed for diagnostic or triage purposes.

Management

  • Focus on educational content and user reflection rather than clinical advice.

Monitoring & Follow-up

  • Utilize a hazard log and risk matrix for ongoing assessment of chatbot interactions.

Risks

  • Address potential misinformation and inappropriate responses through proactive design and review.

Patient & Prescribing Data

Users of the sexual wellness mobile application.

Chatbot provides educational information and resources post-engagement with in-app content.

Clinical Best Practices

  • Assemble cross-functional teams for hazard identification and risk mitigation.
  • Conduct proof-of-concept studies to evaluate user interactions and safety.
  • Iteratively update risk assessments based on user feedback and clinical findings.

References

Original Source(s)

Related Content