1 / 30

Designing An Adaptive Treatment

Designing An Adaptive Treatment . Susan A. Murphy Univ. of Michigan Joint with Linda Collins & Karen Bierman Pennsylvania State Univ. Outline. Adaptive Treatments Why Use? When to Use? Design Goals What Does the Treatment Include? Summary & Discussion. Adaptive Treatments.

Download Presentation

Designing An Adaptive Treatment

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.


Presentation Transcript

  1. Designing An Adaptive Treatment Susan A. Murphy Univ. of Michigan Joint with Linda Collins & Karen Bierman Pennsylvania State Univ.

  2. Outline • Adaptive Treatments • Why Use? • When to Use? • Design Goals • What Does the Treatment Include? • Summary & Discussion

  3. Adaptive Treatments • Individualized tailoring of dosage type and amount to the subject across time. • dosage moderators: variables expected to moderate the effect of treatment component • link values on the dosage moderator with specific dosage via a priori rules

  4. Example : Aftercare for Alcohol Dependency • Overall Goal: deter heavy drinking • Adaptive Treatment Condition: Naltrexone, and CBI • Dosage Moderator: # days heavy drinking • Frequency of Decisions: weekly

  5. Why Use? • Subjects are heterogeneous in their need for treatment • Increase salience • To devote additional resources to higher-risk individuals

  6. Why Use? • Variations of treatment may enhance compliance • Excessive treatment leads to non-compliance or other side effects • Treatment is costly

  7. When to Use? • Use if you expect that there will be significant variation in treatment effects across subjects in comparisons of fixed treatments.

  8. Design Goals: • Maximize strength of treatment By well chosen moderators, well measured moderators, & well conceived dosage assignment rules

  9. Design Goals: • Maximize replicability in future experimental and real-world implementation conditions By fidelity of implementation & by clearly defining the treatment.

  10. Parts of the Treatment: • Choice of dosage moderator • Measurement of dosage moderator • Rules linking dosage moderator to dosage assignment • Implementation of the rules

  11. Dosage Moderators: • Significant differences in effect sizes in a comparison of fixed treatments as a function of characteristics. • Dosage moderator=individual, family, contextual characteristics.

  12. Aftercare for Alcohol Dependency • Individuals who return to heavy drinking while on Naltrexone need additional help to maintain a non-drinking lifestyle. • Dosage Moderator is heavy drinking • Providing CBI to individuals who are maintaining a non-drinking lifestyle is costly.

  13. Technical Interlude! s=dosage moderator t=treatment type (0 or 1) Y=response Y = 0 + 1s + 2t + 3st + error = 0 + 1s + (2 + 3s)t + error If (2 + 3s) is zero or negative for some s and positive for others then s is a dosage moderator.

  14. Measurement of Dosage Moderators • Reliability -- high signal to noise ratio • Validity -- unbiased

  15. Derivation of Rules • Articulate a theoretical model for how treatment effect on key outcomes should differ across values of the moderator. • Use scientific theory and prior clinical experience. • Use prior experimental and observational studies. • Discuss with research team and clinical staff, “What dosage would be best for people with this value on the moderator?”

  16. Derivation of Rules • Good dosage assignment rules are objective, are operationalized. • Strive for comprehensive rules (this is hard!) –cover situations that can occur in practice, including when the dosage moderator is unavailable.

  17. Implementation • Try to implement rules universally, applying them consistently across subjects, time, site & staff members. • Document values of dosage moderator!

  18. Implementation • Exceptions to the rules should be made only after group discussions and with group agreement. • If it is necessary to make an exception, document this so you can describe the implemented treatment.

  19. Summary & Discussion • Research is needed to build a theoretical literature that can provide guidance: • in identifying dosage moderators, • in the development of reliable and valid indices of the moderators that can be used in the course of repeated clinical assessments

  20. Summary & Discussion • Research is needed on how we might use existing experimental and observational studies to • identify useful dosage moderators • Formulate best rules. • Research is needed on how we might design experiments that find good moderators and rules.

  21. Summary & Discussion • In comparison to fixed treatments, adaptive treatments hold much promise in terms of increasing potency, improving compliance, reducing side effects and reducing waste. • As treatment and prevention programs move in the direction of more comprehensive, multi-layered systems, adaptive components should become more common, particularly for chronic problems.

  22. Extra slides follow

  23. Example : Fast Track • Overall Goal: reduce incidence of conduct disorder • Adaptive Treatment Component: family counseling via home visits • Dosage Moderator: level of family functioning • Frequency of Decisions: 3 x year for 10 years

  24. Fast Track: • Multiple pathways to conduct disorder • Use family functioning as a dosage moderator • It was expected that less frequent home visits would be sufficient to promote positive child behavior in families with few family functioning problems.

  25. Fast Track: • Researchers anticipated that higher levels of home visiting to families with few family functioning problems would have a negative impact on child behavior • By stigmatizing the family • By burdening the family and inducing noncompliance

  26. Fast Track • Family Functioning is a latent construct and can not be measured with the same precision as a biological measurement. • Frequent assessment via standardized interviews, family observations, teacher and staff ratings was untenable. • Fast Track used a rating of family functioning completed by home visiting staff.

  27. Fast Track & Home Visiting • Formulated a 6 item assessment, each item results in a 0,1,2,3,4. They summed the items. • Assign weekly home visits if sum is less than 9. • Assign biweekly home visits if sum is between 9 and 16. • Assign monthly home visits if sum is greater than 16. • Deviations were permitted in exceptional(!!!!) circumstances.

  28. Statistical Evaluation • Standard comparisons between adaptive treatments proceed as in fixed treatments comparison. • No dose response analyses unless dose is assigned by randomization. • Assessing planned treatment effect rather than the intention to treat effect when the rules are not followed is an immature area of statistics!

  29. Summary & Discussion • In order to develop innovative statistical analyses examining how adaptive treatments work, we need to think about the dosage moderator – outcome relationship within an effective treatment condition.

More Related