Binge-Eating Genetics Initiative (BEGIN) - Dynamical Systems Approach (BEGIN)

Predicting Binge and Purge Episodes From Passive and Active Apple Watch Data Using a Dynamical Systems Approach

The Binge Eating Genetics Initiative (BEGIN) is an observational study where individuals with binge-eating disorder (BED) or bulimia nervosa (BN) complete assessments about eating disorder history, current disordered eating behavior, and mood. Participants also provide active data on binge eating, purging, nutrition, and cognitions using Recovery Record on the Apple Watch. Passive sensor data are collected via native applications over a 30-day period in 1000 individuals with BED or BN. Investigators will combine longitudinal passive (Apple Watch) and active (Recovery Record) data to predict when patients are at high risk of binge eating or purging. Results will enable the deployment of real-time, in-the-moment, personalized signaling of impending binge or purge episodes that will interrupt automatic behaviors and empower patients to exert control over binge eating and purging by engaging in therapeutic alternatives.

Study Overview

• Status: Completed
• Conditions: Binge-Eating Disorder; Bulimia Nervosa
• Intervention / Treatment: Other: Recovery Record

Detailed Description

The parent study, BEGIN, has full ethical approval. From within Recovery Record, individuals are directed to a link to the BEGIN study. From outside Recovery Record, interested individuals complete a brief screen. If they are potentially eligible, they are directed to Recovery Record. Informed consent for phenotyping is obtained digitally via the Recovery Record app. After consent, interested participants complete a validated eating disorders diagnostic questionnaire based on Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria adapted from the Structured Clinical Interview for DSM-5 (ED100K). This questionnaire was developed for use in large genetic studies as a means to recruit large samples efficiently with high diagnostic accuracy. Those who screen case positive are offered the opportunity to participate in the full study (with a second informed consent also delivered from within Recovery Record). Eligible participants are mailed a package containing a description of the study, saliva collection kit, stool collection kit, and an Apple Watch. Participants complete the Eating Disorder Examination Questionnaire (EDE-Q), Generalized Anxiety Disorder 7-item scale (GAD-7), and Patient Health Questionnaire-9 (PHQ-9) at enrollment, midpoint, and endpoint. Participants are asked to log mood and food in the Recovery Record app on their Apple Watch for 30 days, the duration of the study. During those 30-days, participants are also asked to submit their saliva and stool samples. Saliva kits are returned directly to Rutgers University Cell and DNA Repository (RUCDR).

The questionnaire, active, and passive data will all be used to characterize patterns of when and where individuals are more/less likely to binge and/or purge in their daily lives. The investigators will identify low-risk and high-risk passive data patterns that will facilitate the prediction of transitions to high risk states signaling impending binge or purge episodes.

• Study Type: Observational
• Enrollment (Actual): 1166

Contacts and Locations

Study Locations

• United States
  • Virginia
    • Chapel Hill, Virginia, United States, 27599-7160
      • University of North Carolina

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 16 years to 43 years (Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

Individuals between the ages of 18 and 45, who have an iPhone 5 or later, and who are currently experiencing episodes of binge eating can join the study.

Description

Inclusion Criteria:

• Resident of US;
• a current BN or BED diagnosis (confirmed via questionnaire in screening instrument);
• age 18-45 years, inclusive;
• reads, speaks English;
• existing iPhone user with iPhone 5 or later;
• provides informed consent;
• ambulatory

Exclusion Criteria:

• Current use of hormone therapy;
• bariatric surgery, including the following: Roux-en-Y, gastric bypass, laparoscopic adjustable gastric banding, sleeve gastrectomy, duodenal switch with biliopancreatic diversion;
• currently pregnant or breastfeeding;
• thoughts of suicide with a plan in the 2 weeks prior to study enrollment;
• hospitalization due to ED in the 2 weeks prior to study enrollment;
• antibiotic use in the past 30 days (renders temporary ineligibility, but can be - eligible to participate again 30 days after the last dose);
• probiotic use in the past 30 days

Study Plan

How is the study designed?

Design Details

• Observational Models: Case-Only
• Time Perspectives: Prospective

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
BEGIN Case; Enrolled participants with BN/BED will participate in a 30-day observational study.	Other: Recovery Record; Observational data will be obtained through participants entering data on disordered eating episodes, mood, and food in Recovery Record, an app to assist people in managing their eating disorder during the 30 days of the trial. In addition, participants wear an Apple Watch from which passive data (e.g., heart rate, activity, and geolocation) are collected.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Prediction models for the occurrence of disordered eating episodes in individuals with BN/BED using passive sensor data	Dynamic systems approach will use passive and active data to illustrate the dynamics that lead up to a binge or purge event. Using passive data up to two hours prior to a binge or purge event as recorded within the Recovery Record app by the participant (but not within an hour of a previous event), changes in heart rate and actigraphy will be modeled as a function of current heart rate and actigraphy. Time leading up to the binge or purge event, type of event (binge or purge), and an interaction with the type of event will be included as moderators of the relationship of heart and steps in predicting their changes. As one approaches a binge or purge, this will capture a reduction in pattern stability as the strength of association between current heart rate and steps and changes in these same variables should become weaker as one approaches the binge or purge event. This analysis will be able to identify how early signs of instability are detectable before binge and purge events.	all reported disordered eating events during the 30 day observation period
Models of the dynamics in the 2 hours after a disordered eating event occurs	Using dynamical system approaches, changes in heart rate and steps will be modeled as a function of current levels in heart rate and steps. Time since the binge/purge, type of event, and BN/BED diagnosis (along with interactions between them) will be used as moderators of the relationships of heart rate and actigraphy in predicting the changes.	all reported disordered eating events during the 30 day observation period
Prediction models with contextual factors of disordered eating episodes	Contextual factors such as positive and negative mood, time of day, prior same day binge eating and/or purging, location, and use of coping skills will be used to inform the models from OUTCOME 1 to see how these factors moderate likelihood of disordered eating episodes. Current mood is assessed within Recovery Record. When participants log a meal, or a binge or purge, they are asked "Which feelings are you experiencing?" There are 28 response options. These include, but are not limited to Happy, Anxious, Lonely, Guilt, Angry, Depressed. Participants can select as many as they like. For each feeling they endorse, they can indicate the level to which they are experiencing these feeling on a 5-point Likert scale, ranging from options such as 'A little' to 'Extreme', where higher values indicate greater intensity of the feeling endorsed.	all reported disordered eating events during the 30 day observation period

Collaborators and Investigators

• Sponsor: University of North Carolina, Chapel Hill
• Collaborators: National Institute of Mental Health (NIMH); University of Utah; uBiome; Recovery Record
• Investigators: Principal Investigator: Cynthia M Bulik, UNC at Chapel Hill

Publications and helpful links

General Publications

• Butner JE, Deits-Lebehn C, Crenshaw AO, Wiltshire TJ, Perry NS, Kent de Grey RG, Hogan JN, Smith TW, Baucom KJW, Baucom BRW. A multivariate dynamic systems model for psychotherapy with more than one client. J Couns Psychol. 2017 Nov;64(6):616-625. doi: 10.1037/cou0000238.
• Beltz AM, Wright AG, Sprague BN, Molenaar PC. Bridging the Nomothetic and Idiographic Approaches to the Analysis of Clinical Data. Assessment. 2016 Aug;23(4):447-458. doi: 10.1177/1073191116648209.
• Butner JE, Wiltshire TJ, Munion AK. Modeling Multi-Agent Self-Organization through the Lens of Higher Order Attractor Dynamics. Front Psychol. 2017 Mar 20;8:380. doi: 10.3389/fpsyg.2017.00380. eCollection 2017.
• Butner JE, Gagnon KT, Geuss MN, Lessard DA, Story TN. Utilizing topology to generate and test theories of change. Psychol Methods. 2015 Mar;20(1):1-25. doi: 10.1037/a0037802. Epub 2014 Nov 3.
• Kilshaw RE, Adamo C, Butner JE, Deboeck PR, Shi Q, Bulik CM, Flatt RE, Thornton LM, Argue S, Tregarthen J, Baucom BRW. Passive Sensor Data for Characterizing States of Increased Risk for Eating Disorder Behaviors in the Digital Phenotyping Arm of the Binge Eating Genetics Initiative: Protocol for an Observational Study. JMIR Res Protoc. 2022 Jun 2;11(6):e38294. doi: 10.2196/38294.
• Bulik CM, Butner JE, Tregarthen J, Thornton LM, Flatt RE, Smith T, Carroll IM, Baucom BRW, Deboeck PR. The Binge Eating Genetics Initiative (BEGIN): study protocol. BMC Psychiatry. 2020 Jun 16;20(1):307. doi: 10.1186/s12888-020-02698-7.

Study record dates

Study Major Dates

• Study Start (Actual): August 24, 2017
• Primary Completion (Actual): August 31, 2021
• Study Completion (Actual): August 31, 2021

Study Registration Dates

• First Submitted: November 11, 2019
• First Submitted That Met QC Criteria: November 11, 2019
• First Posted (Actual): November 14, 2019

Study Record Updates

• Last Update Posted (Actual): September 29, 2021
• Last Update Submitted That Met QC Criteria: September 27, 2021
• Last Verified: September 1, 2021

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Mental Disorders; Signs and Symptoms, Digestive; Hyperphagia; Bulimia; Feeding and Eating Disorders; Binge-Eating Disorder; Bulimia Nervosa
• Other Study ID Numbers: 17-0242; R01MH119084 (U.S. NIH Grant/Contract)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: Yes
• IPD Plan Description: Anonymized data and scripts will be made available to the general scientific community at the conclusion of the investigation. Specifically, phenotype data will be submitted to the database of Genotypes and Phenotypes (dbGaP) and National Institute of Mental Health Repository & Genetics Resource (NRGR) and genotype data to dbGaP.
• IPD Sharing Time Frame: Assessment phenotype data will be submitted within 6 months after the end of the study. Genotype data will be submitted after the first set of analyses with these data are complete. All data will be available as long as the repositories maintain the datasets (indefinitely).
• IPD Sharing Access Criteria: The databases are controlled-access meaning that individuals who wish to have access to the data must apply to the respective repository and meet all of their criteria.
• IPD Sharing Supporting Information Type: Study Protocol

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No