2019 HSR&D/QUERI National Conference
4107 — Eating Disorder Screening and Diagnostic Tools for the Veteran Healthcare System
Lead/Presenter: Shira Maguen, MIRECC
All Authors: Maguen S (San Francisco VA Health Care System, and UCSF School of Medicine) (multi-PI), Masheb RM (VA Connecticut Healthcare System, Yale School of Medicine), (multi-PI) Marsh AG (VA Connecticut Healthcare System) Huggins J (San Francisco VA Health Care System) Hebenstreit C (VA Palo Alto Health Care System) Li Y (San Francisco VA Health Care System)
To develop an evidence-based clinical pathway that will improve access and outcomes for Veterans with eating disorders. We describe a HSRD funded study that began December 2018 that will enact this objective, and we will present data from our pilot study.
Existing measures for eating disorders fail to identify the full range of DSM-5 eating disorders, and screeners perform poorly in samples representative of Veterans. In our pilot study we surveyed 407 female Veterans with the most widely used screen, the SCOFF (not an acronym), and additional eating questions. Receiver operating characteristic curves were analyzed to create an adapted measure, the screen for disordered eating (SDE), and to compare this new screen to the SCOFF. The HSRD funded project extends this work to include male and female Veterans, and to develop a full clinical pathway toward eating disorder treatment. In our quantitative study, Veterans (N = 400) from two sites will complete surveys of eating behavior. A random subset of those with and without probable eating disorders (n = 156) will participate in gold-standard diagnostic interviews to validate the proposed screen and self-report measure for DSM-5 eating disorder diagnoses. Measures will then be revalidated in a national sample (N = 400). Our qualitative study will involve stakeholder interviews (N = 38) on how to best implement the tools across VHA.
The SDE correctly classified 87.2% (CI: 74.3%-95.2%) of Binge Eating Disorder (BED) cases, all cases of bulimia and anorexia, and 90.5% (CI: 80.4%-96.4%) of any eating disorder (AED) cases. Sensitivity estimates were higher than the SCOFF, which correctly identified 69.6% (CI: 54.2%-82.3%) of BED, 77.8% (CI: 40.0%-97.2%) of bulimia, 37.5% (CI: 8.52%-75.5%) of anorexia, and 66.1%(CI: 53%-77.7%) of AED.
We have demonstrated that our initial screen outperforms the current screen that is most commonly used in the VHA. Further development of this screen, validation of a self-report diagnostic tool, and development of methods for implementation are underway.
Eating disorders are serious psychiatric conditions with high prevalence rates among Veterans. An evidence-based pathway will improve identification, triage, and treatment of eating disorders across VHA.