18
Inference for Clinical Decision Making Policies D. Lizotte, L. Gunter, S. Murphy INFORMS October 2008

Inference for Clinical Decision Making Policies

Embed Size (px)

DESCRIPTION

Inference for Clinical Decision Making Policies. D. Lizotte, L. Gunter, S. Murphy INFORMS October 2008. Sequenced Treatment Alternatives to Relieve Depression STAR*D. - PowerPoint PPT Presentation

Citation preview

Page 1: Inference for Clinical Decision Making Policies

Inference for Clinical Decision Making Policies

D. Lizotte, L. Gunter, S. Murphy

INFORMS

October 2008

Page 2: Inference for Clinical Decision Making Policies

2

Sequenced Treatment Alternatives to Relieve Depression

STAR*D

• A goal of the clinical trial is construct good treatment sequences for patients suffering from treatment resistant depression.

• The goal is to achieve remission.

www.star-d.org

Page 3: Inference for Clinical Decision Making Policies

3

SER, BUP, VEN CIT+BUS, CIT+BUP

MIRT, NTP L2+Li, L2+THY

TCP, MIRT+VEN

Level 2Max 12 Weeks

Level 3Max 12 Weeks

Level 4 12 Weeks

QIDS > 5

QIDS > 5

CITLevel 1Max 12 Weeks

QIDS > 5

QIDS ≤ 5

Follow-up

QIDS ≤ 5

Follow-up

QIDS ≤ 5

Follow-up

QIDS ≤ 5

Follow-up

Preference to Switch

Preference to Switch

Preference to Augment

Preference to Augment

Page 4: Inference for Clinical Decision Making Policies

4

STAR*D • Level 1 Observation:

• QIDS: Quick Inventory of Depressive Symptoms• 16 Items. Score range: 0-27. Self-reported.

• Preference for type of Level 2 treatment: Switch or Augment

• Level 2 Treatment Action: If Level 1 preference is Switch then switch to either Ser, Bup or Ven; if Level 1 preference is Augment then augment with Bup or Bus.

• Level 2 Observation:• QIDS• Preference for type of Level 3 treatment: Switch or Augment

• Level 3 Treatment Action: If Level 2 preference is Switch then switch to either Mirt or Ntp: if Level 2 preference is Augment then augment with Li or Thy

• Level 3 Observation:• QIDS

• Patients exit to follow-up if remission is achieved (QIDS ≤ 5).

Page 5: Inference for Clinical Decision Making Policies

5

Construct the policy to maximize average sum of rewards

•Reward: Convert Level 2 and Level 3 QIDS scores to standardized percentiles → %QIDS

•Reward: Rj=1-(%QIDSj-%5)/100 for j=2,3

•If a patient remits in Level 2, R2=1+1, R3=0.

•Construct policy so as to maximize E[R2 +R3]

Page 6: Inference for Clinical Decision Making Policies

6

Batch version of Q-learning for finite horizon problems

•Approximate Q3 by regressing R3 on Levels 1 & 2 QIDS within each (present action, preference, past action category.

•The best level 3 action is and the value is

Page 7: Inference for Clinical Decision Making Policies

7

Batch version of Q-learning for finite horizon problems:

•Approximate Q2 by regressing R2 + V3 on Level 1 QIDS within each present action x preference category.

•The best level 2 action is and the value is

Page 8: Inference for Clinical Decision Making Policies

8

Use voting across bootstrap samples to assess confidence

• 100 bootstrap samples

• Each sample produces a Q2; for each level 1 QIDS score we calculate the level 2 action that maximizes Q2(o,a). This is a vote by this bootstrap sample for the action.

Page 9: Inference for Clinical Decision Making Policies

9

Page 10: Inference for Clinical Decision Making Policies

10

Conclusion

• If level 1 QIDS is >12 then Ven is best treatment action at level 2

• If level 1 QIDS is <11 then Ser is best treatment action at level 2

• If level 1 QIDS is around 11 or 12 then Ven and Ser are best treatment actions at level 2.

Page 11: Inference for Clinical Decision Making Policies

11

The Problem

• Many patients dropout of the study.

Level 2 Level 3

Remit 383 36

Move to next level

456 260

Dropout 362 160

Sum 1201 456

Page 12: Inference for Clinical Decision Making Policies

12

Two Approaches to Study Dropout

• Complete Case Analysis (Remove all patients with incomplete data from the analysis)--- gross assumptions on why people do or do not dropout. N=1201→N=679.

• Use a Bayesian method: Multiple Imputation.

•This method multiply imputes the missing data. Intuitively, an imputation model is used to group similar patients. Data from similar patients who remain in the study is used to construct the imputations for the missing data of dropouts.

Page 13: Inference for Clinical Decision Making Policies

13

Page 14: Inference for Clinical Decision Making Policies

14

Page 15: Inference for Clinical Decision Making Policies

15

Conclusion

• If level 1 QIDS is > 20 then Ven and Bup are best treatment actions at level 2

• If level 1 QIDS is <12 then Ven and Ser are best treatment actions at level 2

• If level 1 QIDS is around 12 to 20 then Ven is best treatment action at level 2.

Page 16: Inference for Clinical Decision Making Policies

16

Discussion

• If reinforcement learning and modern day control methods are to be used with clinical trial data then these methods must be combined with modern missing data methods and methods for assessing confidence.

•The multiple imputation + bootstrap we used is likely conservative in terms of the assessment of confidence.

• We are developing more principled methods of assessing confidence.

Page 17: Inference for Clinical Decision Making Policies

17

This seminar can be found at:http://www.stat.lsa.umich.edu/~samurphy/

seminars/INFORMS10.08.ppt

Email me with questions or if you would like a copy!

[email protected]

Page 18: Inference for Clinical Decision Making Policies

18