Transcript Slide 1
Methodology for Adaptive Treatment Strategies R21 DA019800 The Goal Susan Murphy, PI To facilitate methodological collaborations necessary for producing evidence-based adaptive treatment strategies. This network involves engineers, computer scientists, psychologists, physicians and statisticians interested in forming collaborate groups which will conduct the necessary research and pilot studies so as to apply for R03 or R01 funding. Our long term goal is to enhance the adaptive, sequential, clinical decision making necessary in treating/managing chronic, relapsing disorders. University of Michigan What are Adaptive Treatment Strategies? Adaptive Treatment Strategies are individually tailored sequences of treatments, with treatment type and dosage adapted and readapted to the individual. • Generalization from a one-time decision to a sequence of decisions concerning treatment • Operationalize clinical practice Example of an Adaptive Treatment Strategy If high risk, then provide biweekly court hearings. Innovations Intervention Trial Design Sequential, Multiple Assignment, Randomized Trials Else maintain on current treatment. 8 wks Responder 2nd Tx R rRBT Generalization of regression from one stage of treatment to multiple stages of treatment. aRBT rRBT rRBT tRBT CATIE (simplified version) R Responder R 1st Tx rRBT SMART Precursors are STAR*D (depression) and CATIE (schizophrenia) (NIMH funded) Identify a few critical decisions, randomize individuals at each critical decision among treatment options. Critical decisions include: Which treatment to use first?, How long to wait for the treatment to work?, Which treatment to provide next? Olan Timing R Zip Other SMART Studies Sequencing Pharmacology and Behavioral Interventions Risp in ADHD. Pelham, PI; IES Perp Early Trigger NonR Responder NTX NTX+TDM CBI NTX+CBI Children. Kasari, PI; Autism Speaks Analyses of SMART Studies Compare two adaptive treatment strategies Compare outcomes from blue and green groups! NTX R R Late Trigger NTX NTX+TCM CBI NonR Responder NTX+CBI Cloz NonResp. Efficacy R Other Tolerance R Olan Quet Zip Risp Sequencing Behavioral Treatments for Nonverbal Autistic 2nd Tx R Quet tRBT Adaptive Aftercare for Alcoholism. McKay PI, NIAAA SMART Study (ExTENd; Oslin, PI: NIAAA) Responder Same Tx Resp. Sequencing Treatments for Prostate Cancer. Millikan, PI Responder 2nd Tx tRBT NonR Responder 1st Tx Preference tRBT eRBT Monitor compliance and symptoms monthly. If offender becomes non-responsive, then provide intensive case management along with assessment and referral for adjunctive services. 1st Tx NonR Responder (SMART) Intervention Trial Data Analysis Q-Learning Exploratory Tailoring Analyses Else provide “as-needed court hearings.” If the offender becomes noncompliant, then the offender is subject to a court determined disposition. strategies that use more than simply response/non-response to adapt treatment to the individual. SMART Study II (RBT; Jones, PI; NIDA) (Drug Court Program for drug abusing offenders.) Provide standard drug counseling Q-Learning: A method from Computer Science Can be used to construct proposals for adaptive treatment Construct more deeply tailored strategies: Should the level of adherence during 1st treatment 2nd be used to select treatment? Should severity of disorder influence which treatment should be 1st? Q-Learning--Example of Results using CATIE Begin with Olanzapine If non-responder then If preference is try for efficacy then If PANSS > 94 then switch to Clozapine Else switch to either Quetiapine or Risperidone If preference is to try for tolerability then If Olanzapine was not tolerable then switch to Risperidone If Olanzapine was not efficacious then switch to Quetiapine PANSS: Positive and Negative Syndrome Scale Acknowledgements: MCAT network and NIH