Overview of Adaptive Treatment Regimes

1
Overview of Adaptive Treatment Regimes
Sachiko MiyaharaDr. Abdus Wahed
2
Before starting the presentation
Adaptive Experimental Design
Adaptive Treatment Regimes
?
3
Adaptive Treatment Regimes vs. Adaptive
Experimental Design

• Adaptive Treatment Regimes
• adaptive as used here refers to a time-
  varying therapy for managing a chronic illness
  (Murphy,2005)
• Adaptive Experimental Design
• such as designs in which treatment allocation
• probabilities for the present patients depend
  on
• the responses of past patients (Murphy,2005)

4
Outline

• 1. What is Adaptive Treatment Regime?
• -Definition
• -Example
• -Objective
• 2. How to decide the best regime?
• - 3 different study designs
• - Comparison of 3 designs
• 3. Trial Example (STARD)
• 4. Inference on Adaptive Treatment Regimes

5
What is Adaptive Treatment Regime?

• Definition
• a set of rule which select the best treatment
  option, which are made based on subjects
  condition up to
• that point.

6
What is Adaptive Treatment Regime?
B1
Responder
B1
A
B2
Non Responder
B2
Patient
B1
Responder
A
B1
B2
Non Responder
B2
7
8 Possible Policies

• (1) Trt A followed by B1 if response, else B2
  (AB1B2)
• (2) Trt A followed by B1 if response, else B2
  (AB1B2)
• (3) Trt A followed by B1 if response, else B2
  (AB1B2)
• (4) Trt A followed by B1 if response, else B2
  (AB1B2)
• (5) Trt A followed by B1 if response, else B2
  (AB1B2)
• (6) Trt A followed by B1 if response, else B2
  (AB1B2)
• (7) Trt A followed by B1 if response, else B2
  (AB1B2)
• (8) Trt A followed by B1 if response, else B2
  (AB1B2)

8
What is the objective of the Adaptive Treatment
Regimes?

• Objective
• To know which treatment strategy works the best,
  given a patients history.

9
What is the objective of the Adaptive Treatment
Regime?

• A treatment naïve patient comes to a physicians
  office.
• Questions
• 1. What treatment strategy should the
• physician follow for that patient?
• 2. How should it be decided?

10
If one knew

• (T be the outcome measurement)
• 1. E(T AB1B2) 15
• 2. E(T AB1B2) 14
• 3. E(T AB1B2) 18
• 4. E(T AB1B2) 17
• 5. E(T AB1B2) 20
• 6. E(T AB1B2) 19
• 7. E(T AB1B2) 13
• 8. E(T AB1B2) 12

Best Regime for the patient
11
In Reality

• Problems
• 1. E(T . ) are not known (need to
• estimate)
• 2. How can one accurately and
• efficiently estimate E(T . )?

12
How to estimate the expected outcome?

• Three study designs
• 1. A clinical trial with 8 treatments
• 2. Combine existing trials
• 3. SMART (Sequential Multiple
• Assignment Randomized Trials)

13
Design 1 A clinical trial with 8 Treatment
Policies
AB1B2
AB1B2
AB1B2
AB1B2
Sample
AB1B2
AB1B2
AB1B2
AB1B2
Randomization
14
Design 2 Combining Existing Trials
Trial 1
Trial 5
Trial 3
Trial 2
Trial 4
B1
B1
B2
B2
A




B1
B1
B2
B2
A
Responder to A only
Responder to A only
Non Responder to A only
Non Responder to A only
15
Design 3 SMART

• Sequential Multiple Assignment Randomized
• Trials (SMART) proposed by Dr. Murphy
• The SMART designs were adapted to
• - Cancer (Thall 2000)
• - CATIE (Schneider 2001) Alzheimer's Disease
• - STARD (Rush 2003) Depression

16
Design 3 SMART
B1
Responder
B1
A
B2
Non Responder
B2
Sample
B1
Responder
A
B1
B2
Non Responder
B2
Randomization
17
Comparisons of 3 Study Designs
Question A Trial with 8 Trts Combined Trial SMART
1. Does it serve the purpose of finding the best strategy?
2. Is it feasible?
3.Can we assess the trt effects using a standard statistical method?
Yes
Yes
Maybe
Yes
No
No
Maybe
Yes
No
18
Sequenced Treatment Alternatives To Relieve
Depression (STARD)

• 1.What is STARD?
• 2.The Study Design

19
What is STARD?

• Multi-center clinical trial for depression
• Largest and longest study to evaluate depression
• N4,041
• 7 years study period
• Age between 18-75
• Referred by their doctors
• 4 stages (3 randomizations)

20
STARD Study Design Level 1
CIT
Responder
Eligible Subjects
CIT
Non Responder
Go to Level 2
21
STARD Study Design Level 2
CITBUP
Add on
CITBUS
CITCT
Lev 1 Non Responder
BUP
SER
Switch
VEN
Subjects Choice
CT
Randomization
22
STARD Study Design Level 3
Lev3 MedLi
Add on
Lev3 MedLi
Lev 2 Non Responder
MIRT
Switch
NTP
Subjects Choice
Randomization
23
STARD Study Design Level 4
TCP
Lev 3 Non Responder
VENMIRT
Randomization
24
Details on Inference from SMART

• Remember the goal is to estimate E(TAB1B2)
• First, how can we construct an unbiased estimator
  for
• E(TAB1B2)?

25
Details on Inference from SMART

• Let us ask ourselves, what would we have done if
  everyone in the sample were treated according to
  the strategy AB1B2 ?

Responder
B1
A
Patient
Non Responder
B2
26
Details on Inference from SMART

• What would we have done if everyone in the sample
  were treated according to the strategy AB1B2 ?
• Answer
• E(TAB1B2) STi/n

Applies to 8-arm randomization trial
27
Details on Inference from SMART

• But in SMART, we have not treated everyone with
  AB1B2

28
Details on Inference from SMART

• Let C(AB1B2) be the set of patients who are
  treated according to the policy AB1B2

29
Details on Inference from SMART

• We define
• R Response indicator (1/0)
• Z1 Treatment B1 indicator (1/0)
• Z2 Treatment B2 indicator (1/0)
• Then
• C(AB1B2) i RiZ1i (1-Ri)Z2i1

30
Details on Inference from SMART

• One would define
• E(TAB1B2) SRiZ1i (1-Ri)Z2iTi/n
• Where n is the number of patients in C(AB1B2).

This estimator would be biased as it ignores the
second randomization.
31
Details on Inference from SMART

• There are two types of patients in the set
  C(AB1B2) who were treated according to the policy
  AB1B2
• A responder who received B1
• and
• A nonresponder who received B2

32
Details on Inference from SMART
33
Details on Inference from SMART

• Assuming equal randomization,
• A responder who received B1 was equally eligible
  to receive B1
• A responder who received B2 was equally
  eligible to receive B2

34
Details on Inference from SMART

• Thus
• A responder who received B1 in C(AB1B2) is
  representative of another patient who received
  B1
• and
• A non-responder who received B2 in C(AB1B2) is
  representative of another patient who received
  B2

35
Details on Inference from SMART

• We define weights as follows
• A responder who received B1 in C(AB1B2) receives
  a weight of 2 1/(1/2), also
• A non-responder who received B2 in C(AB1B2)
  receives a weight of 2 1/(1/2)
• While everyone else receives a weight of zero.

36
Details on Inference from SMART

• Unbiased estimator
• E(TAB1B2) SRiZ1i (1-Ri)Z2iTi/(n/2)
• And, in general,
• E(TAB1B2) SRiZ1i /p1 (1-Ri)Z2i /p2Ti/n

This estimator is unbiased under certain
assumptions
37
Issues

• Compare treatment strategies
• Wald test possible but needs to derive covariance
  between estimators (which may not be independent
  of each other)
• In survival analysis setting, how to derive
  formal tests to compare survival curves under
  different strategies
• Is log-rank test applicable?
• Can the proportional hazard model be applied here?

38
Issues

• Efficiency issues
• How can one improve efficiency of the proposed
  estimator
• How to handle missing data (missing response
  information, censoring, etc.)
• How to adjust for covariates when comparing
  treatment strategies
• And most importantly,

39
Issues

• Is it possible to tailor the best treatment
  strategy decisions to individual characteristics?
• For instance, could we one day hand over an
  algorithm to a nurse (not physician) which would
  provide decisions like If the patient is a
  caucacian female, age 50 or over, have normal HGB
  levels, bla bla blathe best strategy for
  maintaining her chronic disease would be..

40
ATSRG link

• http//www.pitt.edu/wahed/ATSRG/main.htm