542-10-#1

Statistics 542Statistics 542

Introduction to Clinical TrialsIntroduction to Clinical Trials

Issues in Analysis of Issues in Analysis of Randomized Clinical TrialsRandomized Clinical Trials

542-10-#2

Issues in Analysis of Issues in Analysis of Randomized Clinical TrialsRandomized Clinical Trials

• Reference:

May, DeMets et al (1981)

Circulation 64:669-673

Peto et al (1976)

British Journal of Cancer

542-10-#3

Sources of BiasSources of Bias1. Patient selection

2. Treatment assignment

3. Patient Evaluation

4. Data Analysis

Methods to Minimize Bias1. Randomized Controls

2. Double blind (masked)

3. Analyze what is randomized

542-10-#4

What Data Should Be Analyzed?What Data Should Be Analyzed?• Basic Intention-to-Treat Principle

– Analyze what is randomized!– All subjects randomized, all events during

follow-up• Randomized control trial is the “gold”

standard”

• Definitions

Exclusions– Screened but not randomized– Affects generalizability but validity OK

Withdrawals from Analysis– Randomized, but not included in data analysis– Possible to introduce bias!

542-10-#5

Patient CloseoutPatient Closeout

• ICH E9 Glossary– “Intention-to-treat principle - …It has the

consequence that subjects allocated to a treatment group should be followed up, assessed, and analyzed as members of that group irrespective of their compliance with the planned course of treatment.”

542-10-#6

Intention To Treat (ITT) Intention To Treat (ITT) PrinciplePrinciple

• Analyze all subjects randomized & all events

• Beware of “look alikes”– Modified ITT: Analyze subjects who get

some intervention– Per Protocol: Analyze subjects who comply

according to the protocol

542-10-#7

Patient Withdrawn in Analysis (1)Patient Withdrawn in Analysis (1)

• Common Practice - 1980s– Over 3 years, 37/109 trials in New England Journal of Medicine

published papers with some patient data not included

• Typical Reasons Given

a. Patient ineligible (in retrospect)

b. Noncompliance

c. Competing events

d. Missing data

542-10-#8

Patient Withdrawn in Analysis (2)Patient Withdrawn in Analysis (2)

A. Patient INELIGIBLE

– After randomization, discover some patients did not in fact meet entry criteria

– Concern ineligible patients may dilute treatment effect

– Temptation to withdraw ineligibles

– Withdrawl of ineligible patients, post hoc, may introduce bias

542-10-#9

Betablocker Heart Attack TrialBetablocker Heart Attack Trial(JAMA, 1982)(JAMA, 1982)

• 3837 post MI patients randomized• 341 patients found by Central Review to be ineligible• Results

% Mortality

Propranolol Placebo

Eligible 7.3 9.6

Ineligible 6.7 11.3 Best

Total 7.2 9.8

In the ineligible patients, treatment works best

542-10-#10

Anturane Reinfarction Trial (1980) Anturane Reinfarction Trial (1980) NEJMNEJM

• Randomized, double blind, placebo controlled

Anturane Placebo Total

Randomized 813 816 1629

Ineligible 38 33 71

• Reasons for ineligible1/3 - time since MI: < 25 days or > 35 days

1/3 - enzymes not elevated

1/3 - other: age, enlarged heart, prolonged hospitalization, .…

• Number ineligible about the same in each treatment group

BUT

542-10-#11

Anturane Reinfarction Trial (1980)Anturane Reinfarction Trial (1980)

• 1629 patients randomized– 1631 entered, but two patients randomized twice

– Need to delete 03013, 17008

– Use first randomization!

• Declared post hoc 71 “ineligible” patients

542-10-#12

Anturane Reinfarction Trial Anturane Reinfarction Trial (1980)(1980)

Placebo Anturane Total

All 817 812 1629

Ineligible 33 38 71

Eligible 784 774 1558

< 7 day rule 15 15 30

Analyzable subjects 769 759 1528

(Table 3)

Analyzable Deaths - Within 7 days of being off drug

542-10-#13

1980 Anturane Mortality 1980 Anturane Mortality ResultsResults

Anturane Placebo P-Value

Randomized 74/813 (9.1%) 89/816 (10.9%) 0.20

“Eligible” 64/775 (8.3%) 85/783 (10.9%) 0.07

“Ineligible” 10/38 (26.3%) 4/33 (12.1%) 0.12

P-Values for 0.0001 0.92 eligible vs ineligible

Reference: Temple & Pledger (1980) NEJM, p. 1488

542-10-#14

1980 Anturane Mortality 1980 Anturane Mortality ResultsResults

Anturane Placebo

Withdrawn 10 4

• “Early” discontinuation 4 3

• “Late” discontinuation 6 1

542-10-#15

Total MortalityTotal MortalityAnturane Reinfarction Trial (1980)Anturane Reinfarction Trial (1980)

I. All Pts - 1629All Deaths Random Deaths * NEJM accounts for onlyAnturane 812 74 106Placebo 817 89 43Total 1629 *163 149

P = 0.20 (Table 3+6)

II. 1558 Subjects - Exclude 71 Non-eligiblesAll Deaths Random Deaths “71"Anturane 774 64 (10)Placebo 784 85 (4)Total 1558 149 P = 0.07

542-10-#16

Total MortalityTotal MortalityAnturane Reinfarction Trial (1980)Anturane Reinfarction Trial (1980)

III. 1528 Subjects - Exclude 71 Non-eligibles + 30<7 days

All Deaths Random DeathsAnturane 759 60Placebo 769 80Total 1528 140

IV. 1528 SubjectsAnalyzable Deaths Random DeathsAnturane 759 44Placebo 769 62Total 1528 106

P = 0.076

542-10-#17

Total MortalityTotal Mortality Anturane Reinfarction Trial (1980)Anturane Reinfarction Trial (1980)

Consider Patients Excluded

I. < 7 day rule - 30 pts

Alive Dead TotalPlacebo 10 5 15Anturane 11 4 15Total 21 9 30

542-10-#18

Total MortalityTotal Mortality Anturane Reinfarction Trial (1980)Anturane Reinfarction Trial (1980)

II. 71 "ineligibles"

Alive Dead TotalPlacebo 29 4 33Anturane 28 10 38Total 57 14 71

542-10-#19

Anturane Sudden Death (SD)Anturane Sudden Death (SD)I. All Patients (N = 1629)

Randomized NDA-SDs NEJM-SDsPlacebo 816 41 49 8Anturane 813 27 30 3Total 1629 68 79P-value 0.08 0.0311 additional SD's were defined from submission of NDs to publications

II. Exclude 71 Protocol Violators (N = 1558)Randomized NDA-SDs NEJM-SDs

Placebo 783 40 46Anturane 775 26 28Total 1558 66 64P-value 0.08 0.03Difference of 8 SD's

542-10-#20

Anturane Sudden Death (SD)Anturane Sudden Death (SD)for Total Follow-upfor Total Follow-up

III. Exclude 71 Protocol Violators & 30 7 Day Rule Violators (N = 1528)

Randomized NDA-SDs NEJM-SDsPlacebo 768 37 37Anturane 760 24 22Total 1528 61 59P-value 0.09 0.04

* Information not necessarily given in NEJM articlebut used to prepare tables presented

542-10-#21

Anturane AnalysisAnturane Analysis

Percent Mortality

Patient Group Anturane Placebo P-value

1620 5.6 (45/806)+ 7.5 (61/814) 0.10

1567 5.2 (41/781) 7.4 (58/786) 0.07

1547 4.6 (35/768) 7.4 (58/779) 0.01

1475 4.4 (32/733) 7.1 (53/742) 0.02

Article '78 3.4 (25/733) 5.9 (44/742) 0.016+ Number of deaths/number at risk

Table D - 1978 ArticleComparison of the Mortality Experience for

the 4 Patient Groups

542-10-#22

ART (NEJM, 1978)ART (NEJM, 1978)Comparisons of the Mortality Experience for the 73 Patients with

"Objective" and "Subjective" Baseline Exclusions

Groups Compared % Mortality

Placebo vs. Anturane in the 73** 8.6 (3/35)* 26.3 (10/38)

73 vs. 1547*** 17.8 (13/73) 6.0 (93/1547)

(Both Treatment Groups)

73 vs. 1547 (Anturane Group) 26.3 (10/38) 4.6 (35/768)

73 vs. 1547 (Placebo Group) 8.6 (3/35) 7.5 (58/779)* Number of deaths/number at risk** 73 refers to the group of 73 patients with "objective" or

"subjective" reasons at baseline for exclusion*** 1547 refers to the total group of randomized patients with the 73 patients

with objective and subjective baseline exclusions removed

542-10-#23

ART (NEJM, 1978)ART (NEJM, 1978)P-Values Using Two Techniques for

Survival Curve Comparisons of the Groups

P-Values Groups Compared Mantel-Haenszel Gehan

Method Method

Placebo vs. Anturane in the 73 0.045 0.052

73 vs. 1547

(Both Treatment Groups) 0.0009 0.0003

73 vs. 1547 (Anturane Only) < 0.0001 < 0.0001

73 vs. 1547 (Placebo Only) 0.91 0.98

542-10-#24

Acceptable PoliciesAcceptable PoliciesFor Ineligible SubjectsFor Ineligible Subjects

1. Delay randomization, confirm eligibility and allow no withdrawals (e.g. AMIS) (Chronic Studies)

2. Accept ineligibles, allow no withdrawals

(e.g. BHAT, MILIS) (Acute Studies)

3. Allow withdrawals if:

a. Procedures defined in advance

b. Decision made early (before event)

c. Decision independent and blinded

d. Use baseline covariates only (two subgroups)

e. Analysis done with and without

542-10-#25

B. WITHDRAWL FOR NON-COMPLIANCE

References: Sackett & Gent (1979) NEJM, p. 1410Coronary Drug Project (1980) NEJM, p. 1038

• Two Types of Trials

1. Management

- "Intent to Treat" Principle

- Compare all subjects, regardless of compliance

2. Explanatory

- Estimate optimum effect, understand mechanism

- Analyze subjects who fully comply

WITHDRAWALS FOR NON-COMPLIANCE MAY LEAD TO BIAS!

542-10-#26

Breast Cancer Adjuvant TherapyBreast Cancer Adjuvant TherapyProbability of Disease Free Survival for Probability of Disease Free Survival for

Years Post Mastectomy (Method I)Years Post Mastectomy (Method I)

Redmond et al (1983) Cancer Treatment Report

dose protocol total

received dose I Method

542-10-#27

Breast Cancer Adjuvant TherapyBreast Cancer Adjuvant TherapyProbability of Disease Free Survival for Probability of Disease Free Survival for

Years Post Mastectomy (Method II)Years Post Mastectomy (Method II)

(possible)

study on while dose

received dose II Method

Redmond et al (1983) Cancer Treatment Report

542-10-#28

542-10-#29

Breast Cancer Adjuvant TrialBreast Cancer Adjuvant Trial

• Results using stratification by compliance analysis can be re-ordered according to definition

• Both previous graphs are for the placebo arm

• Lesson: Compliance is an outcome & analysis of one outcome, stratified by another, is highly vulnerable to bias

542-10-#30

Cancer Trial (5-FU & Radiation)Cancer Trial (5-FU & Radiation)Gastric CarcinomaGastric Carcinoma

• Reference: Moertel et al. (Journal of Clinical Oncology, 1984)

• 62 patients randomized– No surgical adjuvant therapy

vs.

– 5-FU and radiation

• 5 year survival results

Randomized Percent (%)

Treatment 23% P < 0.05

No Treatment 4%

542-10-#31

Cancer Trial (5-FU & Radiation)Cancer Trial (5-FU & Radiation)Gastric CarcinomaGastric Carcinoma

• According to treatment received 5 year survival

Received % Survival

Treatment 20%

Refused Treatment 30% NS

Control 4%

542-10-#32

Example: Coronary Drug ProjectExample: Coronary Drug Project5-Year Mortality5-Year Mortality

Clofibrate Placebo

N % Deaths N % Deaths

Total (as reported) 1103 20.0 2782 20.9

By Compliance 1065 18.2 2695 19.4

< 80% 357 24.6 882 28.2

> 80% 708 15.0 1813 15.1

• Adjusting for 40 covariates had little impact

• Compliance is an outcome

Compliers do better, regardless of treatment

542-10-#33

Example: Coronary Drug ProjectExample: Coronary Drug Project2-Year Mortality2-Year Mortality

Compliance Assessed Estrogen Placebo

N % Deaths N % Deaths

Total 903 6.2 2361 5.7

< 80% 488 6.1 436 9.9

> 80% 415 6.3 1925 4.8

Comments• Higher % of estrogens patients did not comply• Beneficial to be randomized to estrogen & not take it • (6.1% vs. 9.9%)• Best to be randomized to placebo & comply (4.8%)

542-10-#34

Example: Wilcox et al (1980) Trial, Example: Wilcox et al (1980) Trial, BMJBMJ6-Week Mortality6-Week Mortality

Propranolol Atenolol Placebo

N % Deaths N % Deaths N % Deaths

Total 132 7.6 127 8.7 129 11.6

Compliers 88 3.4 76 2.6 89 11.2

Non-compliers 44 15.9 51 17.6 40 12.5

Comments• Compliers did better than placebo• Treatment non-compliers did worse than placebo• Placebo non-compliers only slightly worse than compliers• Analysis by compliers overestimates benefit

542-10-#35

Aspirin Myocardial Infarction Aspirin Myocardial Infarction Study (AMIS)Study (AMIS)

% Mortality

Compliance Aspirin Placebo

Good 6.1 5.1

Poor 21.9 22.0

Total 10.9 9.7

542-10-#36

Summary of ComplianceSummary of Compliance

• No consistent pattern

Example Non-compliance Did Worse

CDP Clofibrate, AMIS Both Treatment & Control

CDP Estrogen Control OnlyBeta-blocker, Wilcox Two Treatments, Not

Control• Compliance an outcome, not always independent

of treatment• Withdrawal of non-compliers can lead to bias• Non-compliers dilute treatment• Try hard not to randomize non-compliers

542-10-#37

II. Competing EventsII. Competing Events

• Subject may be censored from primary event by some other event (e.g. cancer vs. heart disease)

• Must assume independence

• If cause specific mortality used, should also look at total death

• If non-fatal event is primary, should also look at total death and non-fatal event

• Problem for some response measures

542-10-#38

COMPANIONCOMPANIONExample of Competing EventsExample of Competing Events

• COMPANION trial was a device trial in CHF patients

• Best care vs pacemaker vs pacemaker+defibrillator

• Another device approved during trial• Patients in best care arm withdrew

consent; censored follow-up• Differential censoring biases analysis

542-10-#39

Optimal PharmacologicalTherapy (OPT)

(OPT) + CRT(CONTAK TR/EASYTRAK)

(OPT) + CRT + ICD(CONTAK CD/EASYTRAK)

COMPANION (COMPANION (COCOmparison of mparison of MMedical edical Therapy, Therapy, PPacing, acing, ANANd Defibrillatd DefibrillatIONION in in

Heart Failure): Study DesignHeart Failure): Study Design

PatientEnrollment

OPTAlone

RandomizeBaselineTesting

OPT +CRT

OPT +CRT-D

Target Time to Implant ≤ 2 days from randomization

Randomization stratifications:by site, +/- -blocker therapy

Patients randomized 1:2:2 to the following three arms:

1

2

2

HFSA Late-Breaker Sept 24, 2003

542-10-#40

COMPANION: Endpoints (1)COMPANION: Endpoints (1)

• Primary Endpoint:

– Composite of time to first all-cause mortality or all-cause hospitalization analyzed from randomization

• Hospital emergency or outpatient (unscheduled) administration of IV inotropes or vasoactive drugs for more than 4 hours were considered a hospitalization primary event

HFSA Late-Breaker Sept 24, 2003

542-10-#41

COMPANION: Endpoints (2)COMPANION: Endpoints (2)

• Highest order secondary endpoint:– All-Cause Mortality

• Other outcomes analyzed: – Combined mortality or CV, heart failure

hospitalizations

HFSA Late-Breaker Sept 24, 2003

542-10-#42

COMPANION: Statistical PlanCOMPANION: Statistical Plan• Intention to treat, endpoint data collection begins with

randomization; open-label (ethical reasons)• Steering and Endpoints Committees, Exercise Core

Laboratory, Sponsor were blinded• Alpha allocation: OPT vs. CRT = 0.02; OPT vs. CRT-D

= 0.03• Sample size assumptions and calculations:

– Primary endpoint: 12 month event rate of 40% in OPT arm, 25% reduction in either device arm would require 2200 patients followed for ≥12 month (would translate to 1000 primary events), power ≥90%

– Mortality (secondary endpoint): 12 month event rate of 24% in the OPT arm, 25% reduction in either device arm;power = 80%

HFSA Late-Breaker Sept 24, 2003

542-10-#43

COMPANION: Data UpdateCOMPANION: Data Update• ACC March 2003 (Preliminary Data)

– Data indicated a disproportionate withdrawl rate among OPT, CRT and CRT-D (13%, 2%,2% w/o prior PEPs)

– After deliberations with the independent SDAC and DSMB, a decision was made by the Steering Committee to:

• Re-consent withdrawn patients to collect endpoint data and vital status• Not count elective device admissions as hospitalization EPs

• HFS 2003 (Final Data)The process of collecting endpoint data and vital status on patients that withdrew prior to 12/01/02 is complete:– OPT = 95%, CRT = 99%, and CRT-D = 99%– Median follow-up times (days) are 442 for OPT, 495 for CRT

(p = .03), and 479 for CRT-D (p = .13)

HFSA Late-Breaker Sept 24, 2003

542-10-#44

COMPANION: Primary EndpointCOMPANION: Primary Endpoint

HFSA Late-Breaker Sept 24, 2003

542-10-#45

COMPANION: Secondary Endpoint COMPANION: Secondary Endpoint of All-Cause Mortalityof All-Cause Mortality

HFSA Late-Breaker Sept 24, 2003

542-10-#46

CONCLUSIONSCONCLUSIONS

When added to optimal pharmacological therapy in patients with modern-severe LV dysfunction, NYHA class III or IV symptoms and QRS lengthening:

• CRT or CRT-D reduces mortality + hospitalization• CRT-D reduces mortality

– 2/3 of the effect size can be attributed to CRT

HFSA Late-Breaker Sept 24, 2003

542-10-#47

III. Problem of DefinitionsIII. Problem of Definitions

Classification Anturane Placebo P-value

ART 30/812 48/817 0.03

Another Committee 28/812 39/817 0.17

• Cause specific definitions hard to apply

• Example: Anturane Reinfarction Trail (ART)(NEJM, 1980)

Sudden Death

542-10-#48

Anturane Reinfarction TrialAnturane Reinfarction TrialSudden DeathSudden Death

Category Source Placebo Anturane P-value

All patients & all NEJM 48/817 30/812 0.03

sudden deaths AC 39/817 28/812 0.17

"Eligible" patients & NEJM 46/785 28/775 0.03

all sudden deaths AC 37/782 25/773 0.12

• Problem of cause specific definitions• AC = Another review committee

542-10-#49

IV. "Wrong", Inconsistent, IV. "Wrong", Inconsistent, Outlying DataOutlying Data

• "Wrong" or "outlying" data may in fact be real

• Decisions must be made blind of group assignment

• All modifications or withdrawals must be documented

542-10-#50

V. Missing Outcome DataV. Missing Outcome Data

• Design with zero– missingness may be associated with treatment

• for analysis, data are not missing at random• even if same number missing, missing may be for different

reason in each treatment group

• Implement with minimum possible• Missing data happens!• Analyze exploring different approaches

– If all, or most, agree, then more persuasive– Sensitivity / robustness analysis

542-10-#51

Common SolutionsCommon Solutions

• Last observation carried forward (LOCF)

• Best Case – Worst Case

• Multiple Imputation Methods

542-10-#52

““Best” and “Worst” Best” and “Worst” Case AnalysesCase Analyses

Treatment Control

Total Events 170 220

Lost to Follow-up 30 10

"Best" Case 170 230

"Worst" Case 200 220

542-10-#53

Multiple Imputation (Rubin, 2006)Multiple Imputation (Rubin, 2006)

MI is well established as a valid method of dealing with missing data when realistic MI model is used

• Each missing datum is replaced my multiple values• Reflects uncertainty about the correct value to impute• Allows standard complete-data methods of analysis

• MI proposed by Rubin in the 1970s and has a very large number of evaluations supporting its validity and robustness to modeling assumptions

• Standard MI assumes “ignorable” missingness• “ignorable” in a particular technical sense

542-10-#54

Multiple Imputation (Rubin, 2006)Multiple Imputation (Rubin, 2006)

542-10-#55

VI: Poor Quality DataVI: Poor Quality Data

• Not all data need to be of same quality• Focus on key data (eligibility, primary

outcome, SAEs• Poor quality data, like missing data, may be

treatment related, not at random• Best solution is prevention, minimize

occurrence of poor quality data for key outcomes

• Conduct sensitivity analysis to see possible impact

• Use rank statistics

542-10-#56

VII. Poor Clinic Performance in VII. Poor Clinic Performance in a Multicenter Studya Multicenter Study

• If randomization was stratified by clinic, then withdrawal of a clinic is theoretically valid

• Withdrawal must be done independent of the outcome at that clinic

542-10-#57

Mortality in Aspirin Myocardial Mortality in Aspirin Myocardial Infarction Study (AMIS)Infarction Study (AMIS)

Aspirin Placebo P-value

All 30 Centers 246/2267 219/2257 0.997 “Selected” Centers 39 66 < 0.01

• In “selected” centers, aspirin showed superiority

542-10-#58

Mortality in Beta-Blocker Mortality in Beta-Blocker Heart Attack Trial (BHAT)Heart Attack Trial (BHAT)

Propranolol Placebo P-value

All 32 Centers 138/1916 188/1921 < 0.01

Cox adjusted Z = 3.05

6 “Selected” Centers 43 26 < 0.05

• In “selected” centers, propranolol worse

542-10-#59

VIII. Special Counting RulesVIII. Special Counting Rules• Events beyond a specified number of days after

treatment stopped not counted "non-analyzable"

• Examples

1. "7 Day Rule" Anturane (1978) NEJM2. "28 Day Rule" Timolol (1981) NEJM

3. “14 Day Rule” Approve (2005) NEJM

• If used, must– Specify in advance– Be a long period to insure termination not related to

outcome– Must analyze results both ways

542-10-#60

APPROVE TrialAPPROVE TrialNEJM (2005)NEJM (2005)

• Trial of COX II inhibitor (Vioxx) in patients with colon polyps

• Reduce risk of recurrent polyps

• Randomized, multicenter, placebo controlled trial

• 2586 patients randomized

• Follow-up terminated 14 days after patient going off study medication

542-10-#61

APPROVE TrialAPPROVE TrialNEJM (2005)NEJM (2005)

• DMC terminated trial early for excess thrombotic events (46 vs. 26)

• Separation of KM curves after 18 months

• Initial paper claimed risk the same for 18 months

• Controversy followed

542-10-#62

APPROVE TrialAPPROVE TrialNEJM (2005)NEJM (2005)

Kaplan-Meier Estimates of the Cumulative Incidence of Confirmed Serious Thrombotic

Events

542-10-#63

APPROVE TrialAPPROVE Trialwith additional follow-upwith additional follow-up

NEJM, Nissen (2006)NEJM, Nissen (2006)

Kaplan-Meier Estimates of the Cumulative Incidence of Confirmed Thrombotic Cardiovascular Events in the Rofecoxib and Placebo Groups, According to the

Intention-to-Treat Principle

542-10-#64

APPROVE LessonsAPPROVE Lessons

• Use of “14 Day” rule was not justified

• Likely that going off study medication was related to patient risk

• Caused a bias in assessment of risk

• In general, censoring rules such as this should be avoided

542-10-#65

IX. Fishing orIX. Fishing orDichotomizing OutcomesDichotomizing Outcomes

• Common practice to define a response (S,F) from a non-dichotomous variable

• By changing our definition, we can alter results

• Thus, definitions stated in advance

• Definitions should be based on external data

542-10-#66

Dichotomizing OutcomesDichotomizing Outcomes

Heart Rate

Trt A Trt B

Subject Pre Post Pre Post

1 72 72 0 72 702

2 74 73 1 71 68 3

...

25 73 73 0 79 79 0

Mean 74.0 73.2 0.8 74.4 74.0 0.4

Example

542-10-#67

Three Possible Analyses (1)Three Possible Analyses (1)

Change Treatment A Treatment B P-Value

1.F = < 7 23 25 0.49

S = > 7 2 0

542-10-#68

Three Possible Analyses (2)Three Possible Analyses (2)

Change Treatment A Treatment B P-Value

1.F = < 7 23 25 0.49

S = > 7 2 0

2.F = < 5 19 25 0.02

S = > 5 6 0

542-10-#69

Three Possible Analyses (3)Three Possible Analyses (3)

Change Treatment A Treatment B P-Value

1.F = < 7 23 25 0.49

S = > 7 2 0

2.F = < 5 19 25 0.02

S = > 5 6 0

3.F = < 3 17 18 0.99

S = > 3 8 7

542-10-#70

X. Time Dependent Covariate X. Time Dependent Covariate AdjustmentAdjustment

• Classic covariate adjustment uses baseline prognostic factors only– Adjust for Imbalance– Gain Efficiency

• Adjustment by time dependent variates not recommended in clinical trials (despite Cox time dependent regression model)

• Habit from epidemiology studies

542-10-#71

Coronary Drug ProjectCoronary Drug Project5-Year Mortality5-Year Mortality

Baseline Cholesterol % DeathsCholesterol Change Clofibrate Placebo

< 250mg%* Fall 16.0 21.2

< 250 Rise 25.5 18.7

> 250 mg% Fall 18.1 20.2

> 250 ** Rise 15.5 21.3

• Little change in placebo group• Best to have

a. Low cholesterol getting lower *b. High cholesterol getting higher **

Example

542-10-#72

Example: Cancer TrialsExample: Cancer Trials

• A common practice to compare survival on patients with a tumor response

• Problem is that patient must first survive to be a responder

length - bias sampling

542-10-#73

Cancer Trials (1)Cancer Trials (1)

Advanced Breast Cancer: Surgery vs. Medicine

Santen et al. (1981) NEJM

(Letter to editor, Paul Meier, U of Chicago)

• A randomized clinical trial comparing surgical adrenalectomy vs. drug therapy in women with advanced breast cancer

• 17 pts withdrawn from surgery group

10 pts withdrawn from medical group

542-10-#74

Cancer Trials (2)Cancer Trials (2)• Reasons

– Medical group (10 pts)

2 stopped taking their drugs

5 drug toxicity

– Surgical group (17 pts)

7 later refused surgery

8 rapid progression precluding surgery

• No follow-up data on these 27 pts presented

542-10-#75

XI. Subgroup AnalysesXI. Subgroup Analyses

542-10-#76

False Positive RatesFalse Positive RatesThe greater the number of subgroups analyzed separately, the larger the probability of making false positive conclusions.

No. of Subgroups False Positive Rate

1 .05

2 .08

3 .11

4 .13

5 .14

10 .19

542-10-#77

Subgroup AnalysesSubgroup Analyses

• Focusing on a particular “significant” subgroup can be risky– Due to chance– Results not consistent

• Estimates not precise due to small sample size

542-10-#78

MERIT Total MortalityMERIT Total Mortality

542-10-#79

MERITMERIT

542-10-#80

MERITMERIT(AHJ, 2001)

542-10-#81

Praise IPraise IRef: NEJM, 1996

• Amlodipine vs. placebo• NYHA class II-III• Randomized double-blind• Mortality/hospitalization outcomes• Stratified by etiology (ischemic/non-ischemic)• 1153 patients

542-10-#82

PRAISE IPRAISE I

542-10-#83

PRAISE I - InteractionPRAISE I - Interaction

• Overall P = 0.07

• Etiology by Trt InteractionP = 0.004

• Ischemic P = NS

• Non-Ischemic P < 0.001

542-10-#84

PRAISE I - IschemicPRAISE I - Ischemic

542-10-#85

PRAISE I – Non- IschemicPRAISE I – Non- Ischemic

542-10-#86

PRAISE IIPRAISE II

• Repeated non-ischemic strata

• Amlodipine vs. placebo

• Randomized double-blind

• 1653 patients

• Mortality outcome

• RR 1.0

542-10-#87

PRAISE I vs PRAISE IIPRAISE I vs PRAISE IIPlacebo armsPlacebo arms

542-10-#88

Three ViewsThree Views

• Ignore subgroups and analyze only by treatment groups.

• Plan for subgroup analyses in advance. Do not “mine” data.

• Do subgroup analyses

However view all results with caution.

542-10-#89

Analysis Issues SummaryAnalysis Issues Summary

• Important not to introduce bias into the analysis

• ITT principle is critical

• Important to have “complete” follow-up

• Off treatment is not off study

542-10-#90

Henry MallHenry Mall