Getting SMART about Adaptive Interventions: A Conceptual Introduction Rona L. Levy With many thanks to Daniel Almirall

Transcription

1 Getting MART about Adaptive Interventions: A Conceptual Introduction Rona L. Levy rlevy@uw.edu With many thanks to Daniel Almirall Institute for ocial Research, University of Michigan

2 Adaptive Interventions Definition: An Adaptive Intervention is a sequence of individually tailored decision rules that specify whether, how, or when and based on which measures to alter the dosage (duration, frequency or amount), type, or delivery of treatment(s) at critical decision points. Adaptive Interventions (AIs) help guide the type of sequential treatment decision making that is typical of (and often needed in!) educational or clinical practice.

3 Concrete Example of an Adaptive Intervention Child ADHD in chools, Ages 6-12 First-line Txt Tailoring Variable econd-line Txt Responder Continue: MED MED Non-Responders Augment: MED + BM Responder status measured by school-teacher. Goal is to min. symptoms / max. school performance.

4 What makes up an Adaptive Intervention? 1. Critical decision points: based on time or other measures 2. Treatment options at each stage 3. Tailoring variables: to decide how to adapt treatment 4. Decision rules: inputs tailoring variable, outputs treatments aka: dynamic treatment regiment, adaptive txt strategies, treatment algorithms, medication algorithms, stepped care, txt policies, multi-stage strategies

5 Why are Adaptive Interventions Necessary? In Clinical Practice... Nature of chronic disorders/phenomena (substance use, obesity, mental health, autism, diabetes, cancer, HIV/AID, pain) Waxing and waning course (multiple relapse, recurrence) Life events, comorbidities, non-adherence may arise Disorders for which there is no widely effective treatment. Disorders for which there are widely effective treatments, but they are costly or burdensome. Bottom line: High heterogeneity in response to treatment Within person (over time) and between person All require sequences of treatment decisions!

6 GENERATING HYPOTHEE vs BUILDING vs EVALUATING ADAPTIVE INTERVENTION

7 3 Different Research Questions/Aims = 3 Different Research Designs Aim 1: When generating hypotheses about an Adaptive Intervention: e.g., Does augmenting txt (as observed in a previous trial) for non-responders correlate with better outcomes? Aim 2: When building an Adaptive Intervention: e.g, What are the best tailoring variables and/or decision rules? Aim 3: When evaluating a particular Adaptive Intervention: e.g. Does the AI have a (statistically powered) clinically significant effect compared to suitable control?

8 3 Different Research Questions/Aims = 3 Different Research Designs Ex. Q1: Does augmenting txt for non-responders (as observed in a previous trial) correlate with better outcomes? Ex. Q2: What are the best tailoring variables or decision rules? Ex. Q3: Does an already-developed adaptive intervention have a statistically and clinically signif. effect as compared to control intervention? Observational tudies Experimental tudies Question Aim e.g., analysis of previous RCT e.g., MART e.g., RCT 1 Hypothesis Gen. YE 2 Building YE 3 Evaluating YE

9 EQUENTIAL MULTIPLE AIGNMENT RANDOMIZED TRIAL (MARTs) Multi-stage trials; same participants throughout Each stage is pre-fixed and corresponds to a critical decision point The goal of a MART is to inform the development of adaptive interventions.

10 Background for ADHD Treatment Example MART: Children Ages 6-12 Both medication (MED) and behavioral modification (BM) have been shown to be efficacious However, there is much debate on whether first-line intervention should be pharmacological or behavioral, especially in younger children Further, there is a need for a rescue treatment if the first treatment does not go well because 20-50% of children do not substantially improve on BM or MED o important questions for clinical practice include What treatment do we begin with: BM or MED? Among non-responders, what second treatment is best?

11 Concrete Example of a MART: Child ADHD PI: William Pelham, PhD, Florida International University, IE-Funded Grant N = 153, 8 month study, Monthly non-response (ITB < 75% and IR > 1 domain) R MED Responders Non-Responders R Continue: MED Intensify: Higher Dose MED Augment: MED + BM BM Responders Non-Responders R Continue: BM Intensify: Increase BM Augment: MED + BM

12 One of Four Adaptive Interventions Within the MART R MED Responders Non-Responders R Continue: MED Intensify: Higher Dose MED Augment: MED + BM BM Responders Non-Responders R Continue: BM Intensify: Increase BM Augment: MED + BM

13 4 Embedded Adaptive Interventions in this MART AI 2 AI 3 AI 4 AI 1 MED Non-Responders Intensify: Higher Dose MED Responders Continue MED MED Non-Responders Responders Augment: MED+BM Continue MED BM Non-Responders Responders Intensify: Increase BM Continue BM BM Non-Responders Augment: MED+BM Responders Continue BM

14 Background to adult obesity study with MART Design PI: Nancy herwood, PhD, HealthPartners and Univ. of MN, Minneapolis MN Rationale: (1) Response heterogeneity is observed in weight loss programs, with half of participants unable to achieve desired goals, a one size fits all approach is the norm, (2) non-response can be detected early. The primary aim is to evaluate the benefit of augmenting treatment with Meal Replacements (MR) versus switching to Acceptance Commitment Therapy (ACT) for non-responders. The secondary aim is to evaluate the optimal time to intervene with non-responders

15 Ongoing Adult Obesity tudy R * IBT for 3 weeks (N=250) A E R E P O N E IBT for 7 weeks (N=250) Responders Non-Responders A E R E P O N E R Responders Non-Responders Continue IBT Augment IBT: IBT + Meal Replacements witch from IBT: ACT R Continue IBT Augment IBT: IBT + Meal Replacements witch from IBT: ACT A B C D E F Baseline 5 Weeks 10 Weeks 20 Weeks * Participants and coaches are blind to group assignment until their randomly assigned response assessment time to reduce the impact of treatment assignment knowledge on early treatment performance

16 MART Designs MARTs are NOT Adaptive Experimental Designs (where, for example, a study is stopped if a treatment is ineffective or harmful) MART are fixed from the start Again, the goal of a MART is to inform the development of Adaptive Interventions

17 MART Design Principles 1. KI Principle: Keep It imple, traightforward 2. Power for simple important primary hypotheses 3. Take Appropriate steps to develop a more deeplyindividualized (optimized) Adaptive Intervention

18 KI Principle: Keep It imple, traightforward (Overarching Principle) At each stage, or critical decision point,... Restrict class of treatment options only by ethical, feasibility, or strong scientific considerations If you do restrict randomizations, use low dimensional summary to restrict subsequent treatments Use binary responder status hould be easy to use in actual clinical practice Collect additional, auxiliary time-varying measures To develop a more deeply-tailored Adaptive Intervention Think time-varying effect moderators

19 MART Design: Primary Aims Choose a simple primary aim/question that aids development of an adaptive intervention. tatistical methods used here aim to reduce uncertainty so the investigator can come away with a solid answer. ample size for the MART chosen based on the hypothesis test associated with this aim (e.g., use standard α = 5%).

20 MART Design: econdary Aims Choose secondary aims/questions that further develop the Adaptive Intervention and take advantage of sequential randomization to eliminate confounding. tatistical methods used here aim to generate hypotheses, e.g., generate good hypotheses about additional tailoring variables or moderators. Here, investigators will tolerate hypothesis tests with higher Type-I error, e.g., α = 10%.

21 Final Comments Adaptive Interventions individualize treatment up-front and throughout; they are guides for clinical practice MARTs are not Adaptive Trial Designs, but are used to build better Adaptive Interventions (ubsequent study: RCT of MART-optimized Adaptive Intervention vs control) MARTs do not have to be complicated; MARTs do not necessarily require larger sample sizes

22 Additional recommended reference H. Lei, I. Nahum-hani, K. Lynch, D. Oslin, 5and.A. Murphy A MART Design for Building Individualized Treatment equences, Annu. Rev. Clin. Psychol :