Getting SMART about Developing Individually-Tailored, Adaptive - PDF document

Getting SMART about Developing Individually-Tailored, Adaptive Health Interventions Addiction Health Services Research - Monday, October 3, 1PM-5PM Instructors: Daniel Almirall, PhD & Susan A. Murphy, PhD (University of Michigan) Module Module Description INTRODUCTION Course Outline, Structure, and Introductions Exercises: Identify working groups of 2-3 investigators by scientific interests/discipline. 1:00-1:10PM (10 min) MODULE 1 What are Adaptive Treatment Strategies?  1:10-1:40PM (30 min) What are adaptive treatment strategies (ATS)? Give examples of ATSs.  Discuss why ATSs are needed and how they inform clinical practice.  Compare simple ATSs versus more deeply tailored ATSs. Q&A Question, Answer, Discussion & Practice Exercise Exercise: Write down 2-3 simple ATSs to address a condition in your research. 1:40-2:00PM (20 min) MODULE 2 What are Sequential Multiple Assignment Randomized Trials (SMARTs)?  2:00-2:40PM (40 min) What are SMARTs? Why do we need SMARTs?  Compare SMARTs to using a multiple-RCT approach for building ATSs.  Discuss SMART design principles. What are typical primary and secondary aims?  Address misconception that SMARTs necessarily require large sample sizes. Q&A Question, Answer, Discussion & Practice Exercise Exercise: Using the 2-3 simple ATSs written above, (a) construct a draft SMART design and 2:40-3:10PM (30 min) (b) identify your primary scientific aim. BREAK 3:10-3:30PM (20 min) Break for water, snacks, and restrooms. MODULE 3 Primary Data Analytic Methods using Data Arising from a SMART  3:30-3:55PM (25 min) Discuss common primary research questions in a SMART.  Present SAS code and worked examples using simulated/fake data. Q&A Question, Answer, Discussion & Practice Exercise 3:55-4:10PM (15 min) Exercise: Write down a primary research question of interest to you. What data analysis approach would you use to address this primary question? MODULE 4 Secondary Data Analytic Methods using Data Arising from a SMART  4:10-4:35PM (25 min) Discuss common secondary research questions in a SMART.  Present SAS code and worked examples using simulated/fake data. Q&A Question, Answer, Discussion & Practice Exercise 4:35-4:50PM (15 min) Exercise: Write down a secondary research question of interest to you. What data analysis approach would you use to address this question? WRAP-UP 4:50-5:00PM (10 min) Wrap-up early to address final questions & to share contact information, etc.

ADAPTIVE INTERVENTION Baseline Risk High Risk Low Risk Assessment BI-WEEKLY COURT AS-NEEDED COURT HEARINGS HEARINGS Monthly Progress Assessments NON-COMPLIANT RESPONSIVE & NON-RESPONSIVE COMPLIANT • ≥ 2 missed counseling sessions • ≥ 2 drug-positive urines and/or • ≥ 2 unexcused failures to provide a urine specimen If W a s As-Needed : If W a s Bi-Wee k l y: INTENSIVE CONTINUE AS BI-WEEKLY JEOPARDY CLINICAL CASE PREVIOUSLY COURT CONTRACT MANAGEMENT ASSIGNED HEARINGS

TCA + Typical Antipsychotic Stage 1 SSRI + Typical Antipsychotic Amoxapine Response TCA + Olanzapine/Risperidone Continuation SSRI + Olanzapine/Risperidone VLF + Typical Antipsychotic VLF + Olanzapine/Risperidone Partial Response or Nonresponse Efficacy Failure: Stage 2 (1) If Non-TCA, Go to TCA (2) If TCA, Go to Non-TCA Response or to Stage 3 Continuation Side Effect Failure: Different Drug Partial Response or Nonresponse Response Stage 3 ECT Continuation Partial Response or Nonresponse Previously Untried Stage 4 Response Stage 1 Agent + Lithium Continuation Augmenting Agent Partial Response or Nonresponse Other Stage 5 Response Agent Not Used in Continuation Stages 1 or 2 Go to Maintenance Phase When Indicated Figure 2. Strategies for the treatment of psychotic major depressive disorder. ECT indicates electroconvulsive therapy; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant; VLF, venlafaxine. This figure is published with permission from the Texas Department of Mental Health and Mental Retardation and is part of a state-funded project.

ADHD in Children SMART Design Principal Investigator: W. Pelham Continue Responders Medication Medication Increase Medication Dose Non-Responders R Add Behavioral Intervention R Continue Responders Behavioral Intervention Behavioral Intervention Increase Behavioral Non-Responders R Intervention Add Medication O1 A1 O2 / R Status A2 Y