Objectives Discuss the role of randomization in controlled trials Discuss the role of blinding in controlled trials Identify the general strengths and weaknesses of controlled trials Identify the advantages to using a runin design a factorial design a randomized matched pair design or ID: 912774
Download Presentation The PPT/PDF document "Chapter 8 Experimental Studies in Epidem..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Chapter 8
Experimental Studies in Epidemiology
Slide2Objectives
Discuss the role of randomization in controlled
trials
Discuss the role of blinding in controlled
trials
Identify the general strengths and weaknesses of controlled
trials
Identify the advantages to using a run-in design, a factorial design, a randomized matched pair design, or a group-randomized
design
Discuss some of the ethical issues associated with experimental
studies
Slide3Epidemiologic
Experimental Studies
Resemble controlled experiments performed in scientific
research
Best reserved for relatively mature questions
Best for establishing cause-effect relationships and for evaluating the efficacy of prevention and therapeutic
interventions
Slide4Experimental
Studies
Also called intervention studiesInvestigators
influence
the exposure of the study subjects
Two types of
experimental
trials
Controlled trials
Community trials
Slide5Between-Group Design
The
strongest methodological design
is
a
between-group
design
, where
outcomes are compared between two or more groups of people receiving different levels of
the
intervention
Slide6Within-Group Design
May be used where the outcome in a single group is compared before and after the assignment of an intervention
Strength – Individual
characteristics that might confound an association (e.g., gender, race, genetic susceptibility) are
controlled
Weakness – Susceptible
to confounding from time-related factors such as the media but may be adjusted for in
the
analysis
Slide7Controlled
Trial
The unit of analysis is the individual A randomized controlled trial in a clinical setting is referred to as a clinical trial
Slide8Community
Trial
The unit of analysis is the group or community An experimental epidemiologic study where one group
of
people,
or
one community,
receives an intervention and
another
does
not is a community trial
Slide9Natural
Experiment
In some rare situations in nature, unplanned events produce a natural experiment
Levels of exposure to a presumed cause differ among a population in a way that is relatively unaffected by extraneous factors so that the situation resembles a planned experiment
Slide10Example of Natural Experiment
Screening and treatment for prostate cancer in the Seattle-Puget Sound area differed considerably from that in Connecticut during 1987-90
Specifically, prostate-specific antigen testing was 5.39 (95% confidence interval: 4.76 to 6.11) times higher in Seattle than
in
Connecticut,
and the prostate biopsy rate was 2.20 (1.81 to 2.68) times higher than in Connecticut
Slide11Example
of
Natural ExperimentTen-year cumulative incidences of radical prostatectomy and external beam radiation up to
1996
were
2.7% and 3.9% for those in Seattle
0.5% and 3.1% for those in Connecticut
Slide12Example
of
Natural ExperimentDid mortality from prostate cancer from 1987 to 1997
differ
between Seattle and
Connecticut?
The adjusted rate ratio of prostate cancer mortality during the study period for
Seattle
vs.
Connecticut was 1.03 (0.95 to
1.11
)
In other words,
the
11-year
follow-up data showed no difference in prostate cancer mortality between the two areas, despite much more intensive screening and treatment
in
Seattle
Slide13Experimental
Studies
Experimental studies may beBetween-group designs
Within-group designs
Slide14Between
Group Design
When a comparison is made between outcomes observed in two or more groups of subjects that receive different levels of the intervention, we call this a between-group design
Slide15Within
Group Design
When we compare the outcomes observed in a single group before and after the
intervention,
it is called a within-group design
Slide16Random
Assignment
Random assignment makes intervention and control groups
look as similar as possible
Chance
is the only factor that determines group assignment
Neither the patient or the physician know in advance which prevention program or therapy will be assigned
Confounding and sample size
Slide17Non-Randomized Study
Also called convenience sample
Concurrent
comparison group is allocated by a
non-random process
Assignment
Problems
Not effective at controlling unmeasured
confounding
variables
Measured
confounding
variables;
however, may be adjusted through analytic methods
Slide18Advantages and disadvantages of randomized controlled clinical trials
Advantages
of randomization
Eliminates conscious bias due to physician or patient selection
Averages out unconscious bias due to unknown factors
Groups are “alike on average”
Disadvantages
of
randomization
Ethical issues
Interferes with the doctor-patient relationship
Slide19Blinding
Three levels of blinding
Single blind – Subjects
Double blind – Investigators
Triple blind – Analyses
Slide20Single-Blinded Study
In a
single-blinded
placebo-controlled
study, the subjects are blinded but investigators are aware of who is receiving the active treatment
Slide21Double-Blinded Study
In a
double-blind
study
, neither
the subjects nor the investigators know who is receiving the active treatment
Slide22Triple-Blinded Study
In a
triple-blind
study
, not
only are the treatment and research approaches kept a secret from the subjects and investigators, but the analyses are completed in a manner that is removed from the investigators
Slide23Blinding
Patients
Why blind patients?
Patients try to get well/please physicians
Minimize potential bias from a
placebo
effect
A
placebo effect
is defined as the effect on patient outcomes (improved or worsened) that may occur due to the expectation by a patient (or provider) that a particular intervention will have
an
effect
Slide24Chronic
Severe Itching Study
Forty-six patients randomly assigned to one of four groups
Treatment
Itching Score
Cyproheptadine HCI 27.6
Trimeprazine tartrate 34.6
Placebo 30.4
Nothing 49.6
Slide25Blinding
Patients
To blind
patients,
use
a
placebo; for example,
Pill
of same size, color, shape as treatment
Sham
operation (anesthesia and incision) for angina relief (unethical)
Problems
In non-drug studies may be impossible/unethical
In drug studies if treatment has characteristic side effects
Slide26Blinding
Patients
More subjective outcomes call for more serious consideration of placebo
For example, time
to death vs. pain relief
Placebos improve comparability of treatment groups in terms of compliance and follow-up
For example,
if patient perceives improvement
because of medication, more likely to remain in study
Slide27Blinding physician or outcome assessing investigator
Best way to avoid unconscious bias is to blind
Physicians –
don’t know which patient is taking the placebo and which patient is taking the drug
Assessors –
of the outcome;
are not the treating doctors, and are not told which treatment was used
What if
physician
blinding is
not possible (e.g., surgery or radiation trial)?
Slide28Problems w
ith Blinding
For non-drug
studies,
such as those involving behavior changes or surgery, it may be impossible or unethical to blind
It may also be problematic to blind in drug studies where a treatment has characteristic side effects
Slide29Purpose o
f Experimental Studies
To identify clinical and public health approaches to solving public health problems (how to prevent or treat)
Slide30Strengths of blinded randomized controlled clinical trials
Demonstrate cause-effect relationships
May produce a faster and cheaper answer than observational studies
Only appropriate approach for some research questions
Allow investigators to control the exposure levels as needed
Slide31Weaknesses of blinded randomized controlled clinical trials
Often more costly in time and money
Many research questions are not suitable for experimental designs because of ethical barriers and because of rare outcomes
Many research questions are not suitable for blinding
Standardized interventions may be different from common practice
May have limited generalizability due to the use of volunteers, eligibility criteria, and loss to follow-up
Slide32Early
History of Clinical Trials
1600 – East India Company
1747 – James Lind
1835 – P.C.A. Louis, Charite Hospital, Paris
Day bled after onset Died Lived % Surviving
1-3 12 12 50
4-6 12 22 65
7-9 3 16 84 .
Note: 1827 33,000,000 leeches imported to Paris
1837 7,000 leeches imported to Paris
Designing
a Clinical Trial
Assembling study
cohort
Measuring baseline variables
Choosing
comparison group
Assuring compliance
Selecting
treatment
Selecting
patient
population
Selecting
outcome
(endpoint)
Ethical considerations
Slide34Assembling
Study Cohort
Inclusion criteria
Broad vs.
specific – related to the extent
of generalization
Is
the outcome rare (e.g., CHD incidence)?
Then
recruit from populations at high risk
such
as males.
Slide35Assembling
Study Cohort
Exclusion criteria
Define
exclusion criteria that will help
control
error.
Example
: An advanced cancer that may be fatal
before
the end of the follow-up period in
a
subject
entering
a
CHD-prevention
study
Exclude
those with difficulty
in
complying
Examples
: Alcoholics, psychotic patients, individuals planning to move out of state
Sample size
Slide36Measuring
Baseline Variables
Characterize the
study cohort
Identifying information (name, address, ID#)
Demographics (age, race, gender, etc.)
Clinical
factors
The
first table of a final report of any randomized blinded trial typically compares the level of baseline characteristics in the two study groups
Slide37Measuring
Baseline Variables
Consider measuring the outcome variable
Change (appropriate for within-group design)
To assure disease is or is not present at baseline (appropriate for
between-group
design)
Measure various predictors of outcomes (e.g., smoking habits) to allow for statistical adjustment
Be parsimonious (i.e., keep simple)
Slide38Choosing Comparison Group
Not
contaminated by
treatment
Ideal – possible to blind, usually meaning placebo
used
Status
quo
vs.
new treatment
Slide39Assuring
Compliance
Calling the day before clinical visitProviding reimbursement
Adhering to the intervention protocol
Drug or behavioral intervention should be well tolerated
Taking once a day vs. complex schedule
Measuring compliance (self-report, pill counts, urinary metabolite levels)
Slide40Selecting
Treatment
What is the research objective?Are the therapies safe and active against the disease?
Is there evidence that one therapy is better than another?
Is the intervention likely to be implemented in a clinical practice?
Is the intervention “strong” enough to have a chance of producing a detectable effect?
Slide41Selecting
Patient Population
Often a compromise between
The population most efficient for answering the clinical question
The population best for generalizing the study findings
For example, many CHD-prevention
trials do not include subjects over age 60 because such elderly subjects might already have extensive atheriosclerosis of their coronary arteries that would no longer be responsive to preventive efforts.
Slide42Selecting
Outcomes (Endpoints
)Primary endpointFor example,
the primary endpoint for
most
phase
III clinical trials in HIV disease is
an
AIDS-defining
event
or
death
Major
AIDS-defining
events
are:
parasitic infections; fungal infections; bacterial infections; viral infections; neoplastic disease; HIV dementia; HIV wasting syndrome
Selection of the “best” endpoint is often complicated
Surrogate endpoint
Slide43Phase
I
TrialUnblinded
, uncontrolled study with
typically
fewer than
30
patients
The
purpose of phase I trials is to determine the safety of a test
in
humans
Patients
in phase I trials often have advanced disease and have already tried
other
options
They
often undergo
intense
monitoring
Slide44Phase
II
TrialRelatively
small (up to 50 people) randomized blinded trials that test
Tolerability
Safe dosage
Side effects
How the
body copes with the
drug
Also
evaluate which types of disease a treatment is effective against, further assess side effects and how they can be managed, and reveal the most effective
dosage
level
Slide45Phase
III
TrialTypically
much larger and may involve thousands
of
patients
These
trials typically involve random assignment and are used to evaluate the efficacy of a
new
treatment
Different
dosages or methods of administration of the treatment are often part of
the
evaluation
Slide46Phase
IV Trial
A large study conducted after the therapy has been approved by the FDA to assess the rate of serious side effects,
and explore further
therapeutic uses
Slide47Selected special
types
ofrandomized designs
Basic randomized designs
Run-in designs
Factorial design
Randomization of matched pairs
Group randomization
Slide48Run-In Design
All subjects in the cohort are placed on placebo (or treatment), followed for some period of time (usually a week or two), and then those who have remained in the study are randomly assigned to either the treatment or placebo arms of the study
A limitation of this study design is that the subjects in the cohort at the time of randomization may no longer reflect the population of interest
Slide49Factorial
Design
Subjects are randomly assigned to one of four groups. The groups represent the different combinations of the two
interventions.
Slide50Randomization o
f Matched Pairs
Improves covariate balance on potential confounding variables
Matched randomization provides more accurate estimates than
unmatched
randomization,
and may involve matching on several potential confounders
Slide51Randomization o
f Matched Pairs
Subjects are matched in pairs according to some confounding factor (e.g., age, sex, race/ethnicity)
One subject is then randomly assigned the study group (e.g., a dietary program, a drug) and the other is assigned to the comparison control group
Slide52Group
Randomization
Groups or naturally forming clusters are randomly assigned the intervention
Groups may involve
Practices
Schools
Hospitals
Communities
Individuals or patients within a cluster are likely to be more similar to each other compared to those in other clusters
Slide53Summary o
f Ethical Principles
Benefits –
Maximize
good
Risks – Avoid doing harm
Subject – Respect for all persons
Society – Fairness to all
Is
consent necessary?
Is it necessary to disclose to the subject the fact that they will be determined by chance?
Should subjects be compensated for injury?
Who should be permitted (or encouraged) to participate in clinical research?
When and how should a clinical trial be
stopped
?
Slide54Tuskegee
Syphilis Study
The Tuskegee study had nothing to do with medical experiments
No treatment was offered for syphilis
No new drugs were tested, nor
were
any efforts
made to establish the efficacy of older chemical treatments
Slide55Summary o
f Ethical Principles
Competent investigators and good research design lead to a greater likelihood of benefits, protect subjects from harm, ensure that peoples’ time is not
wasted,
and their desire to participate in a meaningful activity not frustrated.
Source:
Levine and Labacqz