1. Optimizing weight loss treatment: Design and implementation of the Opt-IN study
Bonnie Spring, PhD
Northwestern University
Director, Center for Behavior and Health
OBSSR & NIDA: Optimization of Behavioral Interventions, 5/16/16, Bethesda, Maryland

2. Overview Obesity as public health problem
Gold standard obesity treatments: challenges to scalability
Are new treatment packages tested by RCT the best way forward?
Treatment optimization by factorial experiment:
Conceptual model, intervention components, optimization criteria
The Opt-IN study
One optimization experiment: Tips, tricks, and a workaround

3. Obesity – The Public Health Problem
Prevalence:
U.S: 30% obese; 69% overweight/obese (CDC, 2008)
World: 1.4 bill overweight/obese (WHO, 2013)
U.S. employer annual cost of obesity-related medical expense, absenteeism, and presenteeism for full-time workers
$73.1 billion (Finkelstein et al, 2010)

4. The Intervention Challenge
Effective treatment (intensive lifestyle treatment – e.g., DPP, Look AHEAD) produces 7% sustained weight loss and metabolic improvement but burdensome (16-36 sessions) and costly ($1800/patient)
Diabetes Prevention Program, Diabetes Care, 2012

5. real-time feedback and decision support on diet and PA
Treatment Package Approach: The ENGAGED Trial Reduce DPP treatment intensity by half; reconfigure components to increase efficiency
8 in-person treatment sessions
Telephone coaching
Peer support (groups, message board,
adherence info)
4. Incentives
5. Technology (app, accelerometer, texts)
“Happy Birthday,Philly!
May you eat well, but
stay within the Fan
Meter safe zone…”
“Love Shimmer –
so reinforcing!”
real-time feedback and decision support on diet and PA
supportive accountability via coach monitoring and remote counseling
supportive accountability via group monitoring and online
social networking
RC1DK087126

6. What don’t we know at end of RCT?
If Treatment > Control
If Treatment < Control
Which components are making positive contribution
Whether all components are needed
Which components’ contribution to effect offsets its cost
How to make intervention more effective
Whether any components are worth retaining
Whether one component had a negative effect that offset a positive effect
Explicitly what went wrong and how to do it better next time
Courtesy of Kari Kugler and Linda Collins

7. Opt-In Study
Optimization of Remotely Delivered Intensive Lifestyle Treatment for Obesity
Principal Investigators
Bonnie Spring, Ph.D. (Northwestern University)
Linda Collins, Ph.D (Pennsylvania State University)
Funded by the National Institute of Diabetes and
Digestive and Kidney Diseases (NIDDK R01 DK097364)

8. Who We Are Shirlene Wang Coach Gwen Ledford Coach
Christine Pellegrini, Ph.D. Project Coordinator
Angela Pfammatter, Ph.D.
Project Coordinator
Elyse Daly
Assessment Coordinator
Amanda Paluch, PhD
Postdoctoral Fellow
Sara Hoffman
Coach/Coordinator
Kara DeWalt
Coach
Gwen Ledford
Coach
Shirlene Wang
Coach
Garseng .Wong
Coach

9. Five components Primary Aim 1 Coaching Intensity Text Messaging
Identify which components/component levels, contribute most to
average weight loss
percent achieving ≥ 7 % weight loss
among overweight and obese adults over a 6-month period.
Coaching Intensity
12 v. 24 phone sessions
Text Messaging
No v. Yes
Progress Reports for PCP
Recommendations to use meal replacements
Training participants’ self selected buddies to be supportive
Five components

10. Primary Aim 2 Apply these results to build an intervention strategy
Made up of only active components >ES .25
Intervention package < $500 dollars (scalable)

11. Step 1: Opt-IN Conceptual Model
Intra-Individual
The two intra-individual mediators are: self-regulation (skill at controlling oneself through self-monitoring, goal-setting, and use of feedback) and self-efficacy (confidence in one’s personal ability to behave in ways that bring about desired. Facilitation entails reliable access to tools (e.g., Opt-IN Smartphone app) and environmental resources (e.g., meal replacements) that make healthy behavior changes easier to perform. Supportive accountability is the manner in which a coaching relationship conveys positive expectations, support, and a sense of holding the participant responsible to attain behavioral goals17,49,61.
Environmental

12. Candidate Intervention Components
For example, text messaging for obesity treatment has been bundled with telephone coaching.54,59 Informing the primary care provider (PCP) about the patient’s weight loss progress has been confounded with having the physician provide behavioral weight loss treatment or having allied health professionals deliver treatment in high-cost medical space, neither of which is cost-efficient.71,37 Other components, e.g., meal replacements, have been tested in such expensive formats (e.g., free food provision) as to preclude real world implementation.41,71 Another example involves the use of peers to provide social support who are given so much training (e.g., 36 hours) that they can more aptly be described as semiprofessionals38.

13. Implemented Design

14. Real World Constraints Impeding Implementation of EBPs
Optimization
Goals
Brevity
Economy
Effect size
Efficiency
Reach
Time
Money
Small benefit
Workflow
Low uptake

16. Opt-IN Goals
Design maximally effective and efficient weight loss intervention such that all components are
active
feasible for real world implementation
make lowest possible resource demands
2. Components meet 2 criteria:
relative cost/burden is known
relative impact on weight loss is unknown
3. Build an intervention:
made up of active components [ES > .25]
yielding largest treatment effect obtainable for <$500

17. Opt-IN Study Protocol
11/11/2018

18. Enrollment Criteria Exclusion Inclusion Unstable medical conditions
History of diabetes w/insulin, CVD, Crohn’s, sleep apnea (CPAP)
Previous hospitalization for psychiatric reasons in past 5 yrs.
At risk for cardiovascular events with MVPA or use assistive device for mobility
Taking weight loss/gain meds
Meet criteria for eating disorders or substance abuse
Pregnant, trying to get pregnant, or lactating
Living with another Opt-IN participant**
Plantar fasciitis diagnosed by physician or podiatrist***
18-60 years (at least half ≥ 45)
BMI kg/m2*
Weight stable
Must have Android or iPhone smartphone
No other major medical concerns
Willing to use smartphone app
*Extended BMI range from to on 8/26/14
**Approved 9/21/15
***Approved 11/16/15

19. Study Sequence Web Screen Telephone Screen Orientation Session
Baseline Assessment
Randomization
Study Sequence
3 Month Assessment
6 Month Assessment

20. Core Treatment (All receive)
Behavioral Coaching and Curriculum
Online lessons
Phone coaching (12 vs. 24)
Self-Monitoring Tool
Smartphone application
Dietary Intervention
Calorie/fat goals, feedback
Physical Activity Intervention
Moderate intensity (brisk walking) physical activity goal, increasing duration over time
Pellegrini CA, Hoffman SA, Collins LM, Spring B. Optimization of remotely delivered intensive lifestyle treatment for obesity using the multiphase optimization strategy: Opt-in study protocol. Contemporary clinical trials. 2014;38:

21. Intervention Components
Coaching Calls over 6 Months
12 calls
24 calls
Progress Report Mailed to Primary Care Physician
Report mailed at 3- and 6-months to participant and Provider
Outlines weight loss progress, behavioral recommendations
Text Messages
Encouraging words, personalized feedback based on self-monitoring
Automated messages in real-time

22. Intervention Components (continued)
Meal Replacement Recommendations
Provided with one week’s supply at randomization session
Recommendations to continue use made by coach throughout intervention
Buddy Training
One telephone training session
Four minute online webinar sessions
Buddies receive $5 for each session complete and additional $20 for completing 3 out of 4 webinars
2 Core: Introduction to Healthy Weight Loss for Buddies, How to Be a Supportive Buddy
4 Additional (choose 2): Support During Setbacks, Building Motivation, Problem Solving, Healthy Eating Out

23. Behind the Scenes: Implementation

24. How to Explain this to Participants…
Ensure they know the 5 components
Show them the 16 different conditions
Highlight:
No control group
Everyone gets something
Study staff & coaches have no idea which conditions/components yield greatest weight loss
In some ways easy because coaches have to follow protocol – can’t establish 32 difft treatment biases.

25. How to Implement a Study with 16 Conditions?
Checks, double checks, and triple checks!

26. Sample visual aids in coaching folder

27. 12 24

28. And how to course correct..
What can go wrong
And how to course correct..
11/11/2018
Presentation or Section Title

29. The Corrigendum
Copy Paste Error – implemented design (proposed in R01) did not match intended fractional factorial Resolution V design
Error discovered when enrollment almost halfway completed
Solution: expand the design to a complete 25 factorial with 32 experimental conditions
Eliminates aliasing of effects among experimental factors and enables us to examine all five components
Does not require randomization of additional subjects
Res 5: each main effect is aliased with one four-way interaction or the five-way interaction, and each two way interaction is aliased with one three-way or higher-order interaction.
Pellegrini CA, Hoffman SA, Collins LM, Spring B. Corrigendum to optimization of remotely delivered intensive lifestyle treatment for obesity using the multiphase optimization strategy: Opt-in study protocol; Contemporary clinical trials. 2015;45:

30. Project Recruitment Timeline
CTAs
Projected
Actual
Average carried forward
CTA
CTA
CTA
CTA
Randomization Schedule ( )
Total # to Randomize: 560 (current: 425)
Time Remaining: 7 Months (June 2016 – Dec 2016)
Monthly Randomization Rate: 17 (need 19.3/month)
May Randomization = 5 so far, 3 scheduled

31. Current Recruitment Status
Web Screen
N = 6175
Common IE Reasons
BMI too high = 1214
BMI too low = 440
Weight unstable = 712
Common IE Reasons
Unable to contact = 608
Didn’t attend session = 628
Not interested = 256
Phone Screen
N = 2949
Eligible
N = 1649
Randomization
N = 425
Baseline
N = 848
560
Gender: Female = 266 / Male = 52
Cond 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 n 18 19
9/13-7/15
Cond 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 n 7 8 6
7/15-12/16

32. Approaching the Finish Line
Current Progress:
Randomized 405 participants since October 2013
Goal = 560 participants by December 2016 (March 2017?)
Evidence-based intervention design: genuine equipoise
Appealing to ppts: no inert control
Some implementation challenges, but manageable
Error highlights design’s efficiency (no additional ppts needed), flexibility, ability to course correct

33. Thank you!