Epidemiology and Biostatistics Clinical Trials

1
Epidemiology and BiostatisticsClinical Trials
Christine E. McLaren, Ph. D. Epidemiology
Division
2
Topics

• Clinical trials
• Selection of subjects
• Controls
• Randomization
• Data collection
• Crossover
• Noncompliance
• Generalizability of results
• Phase I, II, III, and IV trials
• Ethical concerns

3
Clinical Trial

• Intervention study in medicine that in
• involves humans as research subjects
• Controlled clinical trial clinical trial
• that compares treatment(s) to a control
• Control seen as very important. Many
• examples in literature in which
• uncontrolled studies or use of historical
• controls (results taken from literature
• or records ( yields misleading results).

4
Controls

• Passive control
• No treatment placebo
• Active control
• Standard treatment often gold standard
• or standard of care

5
Import Ethical Considerations

• Nuremberg Code
• Belmont Report
• Informed consent of human subject or
• VOLUNTEER
• Signed document on file
• Patient may withdraw at any time
• Confidentiality of patient guaranteed
• Good Clinical Practice GCP.
• Only use active control if current treatment
• standard exists

6
Treatment Arms

• Each treatment modality referred to as
• arm, e.g. Control arm,
• Treatment arm
• Multi-arm Trial
• More than one control
• More difficult to analyze
• Often not controls
• e.g., comparing 3 drugs in a
• three-arm trial

7
Randomization

• Subject assignment to an arm of a clinical
• trial by randomization.
• Patients recruited to a clinical trial after
• they report for treatment to a medical
• facility belong to what is called a
• convenience sample.

8
Randomized Controlled Clinical Trial

• Gold standard of medical research
• Intervention is assigned to patient in an
• unbiased, random manner
• Randomization method may be complicated
• Simple or sophisticated strategies for
• statistical analysis

9
Randomization

• Objective treatment assignmentrule out
• Conscious or unconscious bias of
• Investigators
• Steering of promising patients to
• treatment group is to be avoided.
• Also avoid steering poor prognosis
• patients
• Balances effects of unknown factors among
• treatment groups
• Want treatment groups to be comparable
• Does not solve all bias problems

10
Randomization list
11
Why Gold Standard?

• Timelyprospective, dealing with a current
• issue
• Prospective
• Investigator can control bias and
• variability with careful planning and
• sample size calculations
• Randomization
• Balances groups to make them
• comparable
• Balances adjuvant care over time, as
• opposed historical controls

12
Bias vs. Confounding

• Bias systematic error in study results
• Confounding confusion or error in study
• results due to hidden or mixed-up factors
• e.g., Clinical trial in which males make up
• the treatment group, females make up
• control groupan extreme but
• illuminating example. Confounding cannot
• be reliably corrected by statistical
• adjustment (the information is not there).

13
In a study

• Internal validity
• Are the results correct for the study itself?

• External validity
• Are the results correct for the target
• population?
• A significant p-value is not enough
• Must show sample is representative of
• target population

14
Special Trials Cross-over Trial

• Usually for chronic disease (why?)
• Randomize to whether receive
• treatment first or control first
• Wash-out period for each subject
• between periods
• Dropouts a problem

15
Phases of Clinical Trials

• Phase IEarly stage of clinical pharmacology
• Phase II IIILater stages of clinical
  development
• Phase IV Post-marketing surveillance

16
Cancer Trials
Phase Goal
Endpoint
I Establish Maximum Toxicity tolerated dose
(MTD) II Establish activity Tumor Shrink
age III Establish quantitative Disease-free ef
fect survival/ overall survival
17
Phase I Study of ILX23-7553 in Patients with
Advanced Malignancies
Treatment dosages and cycles  
Treatment dosages and cycles
Wider et al.. Investigational New Drugs 212003.
18
Phase I Study of ILX23-7553 in Patients with
Advanced Malignancies
Distribution of patients with grade 1 or 2
drug-related toxicities among dose levels
Other Toxicity
 
G1
G2
G1
G2
G1
G2
 
Wider et al.. Investigational New Drugs 212003.
19
Phase II

• Establish efficacy, while monitoring safety
• Endpoint often dichotomous response
• Stop early if evidence of low activity and
• reject treatment for further study
• Back to Phase I if no activity and can
• alleviate toxicity with other drugs.

20
A Phase II Trial of Gemcitabine and Docetaxel in
Patients with Chemotherapy-Naïve, Advanced
Nonsmall Cell Lung Carcinoma
Patient Characteristics (N32)
Popa et al., CANCER 95 2002
21
A Phase II Trial of Gemcitabine and Docetaxel in
Patients with Chemotherapy-Naïve, Advanced
Nonsmall Cell Lung Carcinoma
Popa et al., CANCER 95 2002
22
Phase III

• Comparative trials
• Larger trials
• In cancer trials, endpoints are usually
• survival, disease-free survival, or event-free
• survival

23
Randomized Phase III Trial of Paclitaxel,
Etoposide, and Carboplatin versus Carboplatin,
Etoposide, and Vincristine in Patients with
Small-Cell Lung Cancer
Reck, et al. JNCI 95, 2003
24
Randomized Phase III Trial of Paclitaxel,
Etoposide, and Carboplatin versus Carboplatin,
Etoposide, and Vincristine in Patients with
Small-Cell Lung Cancer
Results of he Cox model (univariate and
multivariate analyses) to adjust the risk
associated with therapy for various prognostic
factors
HR hazard ratioT Tumor N Node M
metastasis
Reck, et al. JNCI 95, 2003
25
Internal and External Validity

• Must have absence of bias and confounding
• for internal validity
• In addition, must have representation
• of general or target population for
• external validity

26
Important Documents in Clinical Trials

• Protocolthe plan for the trial
• Informed Consent Document
• Patient assignment sheet
• Data collection sheets (Case Report forms)
• Statistical analysis plan
• Final report or article

27
Data Collection

• Primary outcome The measurement that will
  give the most important answer for this trial
• Often efficacy
• How well does the treatment work?
• Secondary outcomes
• Side effects, toxicities
• Patient satisfaction, Quality of Life
• Cost

28
Data Types Used in Clinical Trials
29
Safeguards for Ethical Concerns

Alphabet Soup The ABCs of Clinical Research
IRB - Institutional Review Board HIPAA -
Health Insurance Portability
and Accountability Act -
Protected Health Information FDA - Food and
Drug Administration NIH - National
Institutes of Health ICH -International
Conference on Harmonisation