Construction of a chatbot to assess depressivity and suicidal tendencies of children and adolescents: a pilot study
- Conditions
- mild, moderate and severe depressive episodesuicidal tendency
- Registration Number
- DRKS00023733
- Lead Sponsor
- Klinik und Poliklinik für Kinder- und Jugendpsychiatrie, Psychosomatik und Psychotherapie der Universität Regensburg am Bezirksklinikum Regensburg
- Brief Summary
8 transcripts (depression group: n = 5, control group: n = 3) were analysed with a BERT network (Bidirectional Encoder Representations from Transformers). Based on the transformer, only one additional layer was used to avoid overfitting. A group classification accuracy of over 80 % was achieved.
- Detailed Description
Not available
Recruitment & Eligibility
- Status
- Complete
- Sex
- All
- Target Recruitment
- 11
Inclusion Criteria
Three out of nine participants have been diagnosed with mild, moderate or severe depression. A further inclusion criterion is a sufficient understanding of the German language.
Exclusion Criteria
Persons with other neurological or mental (pre-)illness cannot participate in the study. A further exclusion criterion is intellectual impairment with an IQ < 85 and insufficient understanding of the German language.
Study & Design
- Study Type
- observational
- Study Design
- Not specified
- Primary Outcome Measures
Name Time Method The presence or absence of a depressive disorder in the participants is evident in both the BDI-II and the MINI-Kid.
- Secondary Outcome Measures
Name Time Method - This pilot study will explore whether the answers in the MINI-Kid on depression can be used to determine the severity of the depressive disorder using the BDI-II score.<br><br>- This pilot study will explore whether the answers in the MINI-Kid on suicidal tendencies can be used to draw conclusions about the severity of suicidal tendencies using the BSS score.