IEPA 11 has ended
Back To Schedule
Wednesday, October 10 • 1:35pm - 1:45pm
Oral 14, Talk 4. "The Early Psychosis Screener (EPS-SVM): a Practical Self-Report Tool using Machine Learning to Predict Psychosis"

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Feedback form is now closed.
Benjamin Brodey1, Ragy Girgis2, Oleg Favorov3, Carrie Bearden4, Scott Woods5, Jean Addington6, Diana Perkins3, Elaine Walker7, Barbara Cornblatt8, Gary Brucato2, Susan Purcell1, Kristin Cadenhead9; 1TeleSage, Inc., 2New York State Psychiatric Institute, 3University of North Carolina at Chapel Hill, 4University of California Los Angeles, 5PRIME Psychosis Prodrome Research Clinic, Connecticut Mental Health Center, 6University of Calgary, 7Emory University, 8The Zucker Hillside Hospital, 9University of California, San Diego
Introduction: A more accurate self-report screener for early psychosis would help to promote early identification and intervention. Methods: Self-report Likert-scale survey items were administered to individuals being screened with the gold standard Structured Interview for Psychosis-risk Syndromes (SIPS) at eight specialty early psychosis clinics. An a priori analytic plan included Spectral Clustering Analysis (SCA) to reduce the item pool followed by development of Support Vector Machine (EPS-SVM) identification algorithms. Results: The cross-validated positive predictive value (PPV) of the EPS-SVM at the optimal cut-off (76.5%) exceeded that of the clinician administered SIPS (68.5%) at differentiating clients in specialty early psychosis clinics who would not convert to psychosis within 12 months from those who either would convert within 12 months or who had already converted. When used in tandem with the SIPS with CHR participants, the EPS-SVM increased the combined PPV to 86.6%. The SVM classified as FEP/converters only 1% of individuals in non-clinical and 4% of clinical control populations. Sensitivity of the EPS-SVM, however, was approximately 50%. Discussion: The EPS-SVM can identify, comparably to the SIPS but in far less time and with fewer resources, a group of individuals who are either at very high risk to develop a psychotic disorder within 12 months or who are already experiencing a FEP. Compared to the SIPS, however at the optimal cut-off, current or future psychotic cases are missed by EPS-SVM approximately 50% of the time.  The cut-off can be selected based on purpose.  A free on-line screening system, eps.telesage.org, is under development.


Wednesday October 10, 2018 1:35pm - 1:45pm EDT
St. George AB Westin Copley Place, third floor