Smart Technology for Anorexia Nervosa Recovery: A Pilot Intervention for the Post-Acute Treatment of Anorexia Nervosa

PROJECT SUMMARY/ABSTRACT Anorexia nervosa (AN) has the highest mortality rate of any mental illness, with a typical onset in adolescence. Although family-based interventions are efficacious for up to 75% of adolescents with AN, approximately 30% will relapse after recovery. There is a critical need to optimize treatments and prevent post-discharge relapse

following acute treatment to improve outcomes for adolescents with AN. To address this critical need, our team developed a suite of digital tools that advance the science of assessment, risk prediction, and clinical-decision support for use in the post-acute treatment window, called “Smart Treatment for Anorexia Recovery (STAR).”

STAR uses cutting-edge assessment technology to shorten test administration and machine-learning to predict likelihood of recovery. This information is then provided back to the clinician via an easy-to-use clinical- decision support tool to alert the clinician when user-entered data suggests the patient is not progressing. In

the current application, we propose to expand STAR to test an adaptive mHealth intervention delivered in the post-discharge window. Our scientific premise is that a transdiagnostic assessment and clinical-decision support tool delivered within the STAR suite will optimize face-to-face clinical service and the addition of an

adaptive mHealth intervention will improve outpatient treatment response and reduce relapse in adolescents discharged from intensive treatment for AN. Our previous work supports our scientific premise. Specifically, our studies provide robust support for the predictive validity and clinical utility of our assessment tool for predicting

ED-related psychiatric impairment and recovery. However, the number of items across our paper-based assessment tool is 144, which is overly long for routine use. To overcome this challenge, we developed a mobile phone app that uses computerized adaptive testing to reduce assessment length by up to 50% while

retaining the reliability and validity of the original paper-and-pencil measure. We propose to leverage this innovation to optimize both face-to-face and mHealth treatment for AN. Our objectives are to: 1) develop the mHealth intervention (with clinician and stakeholder input) and 2) establish feasibility, acceptability, and

preliminary effect size of our mHealth intervention using both clinician and patient data. To accomplish our objectives, we will employ a computerized adaptive test coupled with machine learning algorithms, delivered within our app to signal clinicians when their clients are at-risk for poor outcomes and relapse. Specific aims

include: 1) adapt our existing clinical tool to provide therapist support modules and patient mHealth messages; 2) conduct a preliminary randomized controlled trial (RCT) of our integrated assessment and mHealth intervention tool ; 3) test preliminary mechanisms that lead to changes in AN symptoms. Given there is a

scarcity of specialty care for AN following acute treatment, yet 95% of adolescents have smart phones, the proposed research is innovative and significant because it has the future potential to reduce relapse and optimize existing community-delivered interventions for AN over the post-acute treatment window.