Activity Number:
|
401
- Lead with Statistics: Case Studies and Methods for Learning and Improving Healthcare Through EHRs
|
Type:
|
Topic Contributed
|
Date/Time:
|
Tuesday, July 31, 2018 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Biometrics Section
|
Abstract #329871
|
|
Title:
|
Predicting Suicide Risk: Statistical Methods for Using EHR Data to Inform Mental Health Care
|
Author(s):
|
Rebecca Coley* and Susan Shortreed and Rod Walker and Eric Johnson
|
Companies:
|
Kaiser Permanente Washington Health Research Institute and Kaiser Permanente Washington Health Research Institute and Kaiser Permanente Washington Health Research Institute and Kaiser Permanente Washington Health Research Institute
|
Keywords:
|
risk stratification;
prediction;
precision medicine;
mental health;
suicide;
longitudinal data analysis
|
Abstract:
|
Prediction and prevention of suicide attempts and deaths is an urgent priority for health care systems and mental health providers. Our study uses patient data available in the EHR, including longitudinal information on diagnoses, prescriptions, and patient-reported mental health symptoms, to predict suicide risk for a patient at the time of a mental health visit. This talk will discuss statistical considerations when designing a tool for clinical use including: how should we handle multiple mental health visits per patient? what statistical methods are most accurate given our large sample size and low event rate? how do we split our training and validation sets given our sampling scheme and the target of prediction? Conclusions drawn are applicable to a range of prediction and quality improvement projects using EHR data.
|
Authors who are presenting talks have a * after their name.