Automatic analysis of natural language in online therapy transcripts.
Research type
Research Study
Full title
What can automatic analysis of natural language in online mental health records and therapy transcripts tell us about treatment process and outcome?
IRAS ID
141708
Contact name
Michael King
Contact email
Sponsor organisation
University College London
Research summary
This project aims to harness the potential of vast amounts of therapeutic data through the use of advanced linguistic analysis software. Developing patterns of analysis such as these allows for the processing of large volumes of data that could provide valuable insight into the process and outcome of online therapy. The linguistic analysis process will require the development of new query patterns adapted to this form of data and will provide structured results with which the research team will build predictive models of clinical outcomes to be tested and adjusted.
Detailed analysis of the data will allow the researchers to develop models of textual patterns that predict 1) severity from questionnaire scores; 2) comorbidity with other disorders; 3) relationship with the therapist; 4) levels of risk (e.g. self-harm); 5) early termination of therapy and poor adherence.
Refining and adapting an analysis method for this type of data has vast implications for the patient population involved. It would allow for real-time feedback on the progress of the individual and consequently transform the current approach to monitoring outcomes of treatment. It would allow a shift from questionnaire-based and therapist-led therapeutic decisions to patient-guided treatment.
The study is secondary research and involves the use of two data sets. The first will be therapy transcripts collected during a clinical trial of online therapy and the second will be a set of therapy transcripts and other anonymised clinical data that are routinely collected by the therapy provider over the course of treatment. Each patient follows a course of approximately 10 sessions.REC name
London - Riverside Research Ethics Committee
REC reference
13/LO/1929
Date of REC Opinion
10 Dec 2013
REC opinion
Favourable Opinion