Experimental Studies Developed through the APTR Initiative to Enhance Prevention and Population Health Education in collaboration with the Brody School of Medicine.
Download
Report
Transcript Experimental Studies Developed through the APTR Initiative to Enhance Prevention and Population Health Education in collaboration with the Brody School of Medicine.
Experimental Studies
Developed through the APTR Initiative to Enhance Prevention and Population
Health Education in collaboration with the Brody School of Medicine at East
Carolina University with funding from the Centers for Disease Control and
Prevention
APTR wishes to acknowledge the following individual that
developed this module:
Jeffrey Bethel, PhD
Department of Public Health
Brody School of Medicine at East Carolina University
This education module is made possible through the Centers for Disease Control and Prevention (CDC) and the
Association for Prevention Teaching and Research (APTR) Cooperative Agreement, No. 5U50CD300860. The module
represents the opinions of the author(s) and does not necessarily represent the views of the Centers for Disease
Control and Prevention or the Association for Prevention Teaching and Research.
1.
Recognize use of experimental studies as an
epidemiologic study design
2.
Distinguish between types of experimental studies
3.
Describe key features of conducting experimental
studies
4.
Recognize special considerations of experimental
studies
Experimental studies (experimental)
Researcher determines who is exposed
(treatments received)
Cohort studies (observational)
Case-control studies (observational)
Cross-sectional studies (observational)
Goal of public health and clinical medicine is to
modify natural history of disease and improve
morbidity and mortality
How do we select the best preventive and
therapeutic measures?
Carry out studies to determine value of various
measures
Smith, AH. The Epidemiologic Research Sequence. 1984
Most closely resemble controlled laboratory
experiments
Gold standard of epidemiological research
High status and validity and can pick up small and
modest effects
James Lind identified symptoms of scurvy among
sailors at sea after as little as a month
Conducted early experimental study on treatment of
scurvy in mid-1700’s among British sailors
Small sample size (6 groups of 2 ill sailors)
Group eating oranges and lemons were fit for duty in
6 days
Evaluate new drugs and other treatments for
diseases
Evaluate new medical and health care technology
Evaluate new screening programs or techniques
Evaluate new ways of organizing or delivering health
services (e.g. home v. hospital care following
myocardial infarction)
Preventive
Does prophylactic agent given to healthy or high-risk
individual to prevent disease?
Therapeutic
Does treatment given to diseased individual reduce risk of
recurrence, improve survival, quality of life?
Individual
Do women with stage I breast cancer given a lumpectomy
alone survive as long without recurrence of disease as
women given a lumpectomy plus radiation?
Community
Does fluoride in the water supply decrease the frequency
of dental caries in a community compared to a similar
community without such water treatment?
STUDY
POPULATION
RANDOM ASSIGNMENT
CURRENT
TREATMENT
IMPROVE
DO NOT
IMPROVE
NEW
TREATMENT
IMPROVE
DO NOT
IMPROVE
Hypothesis formed
Participants recruited based on specific criteria and their
informed consent is sought
Eligible and willing subjects randomly allocated to receive
one of the two or more interventions being compared
Study groups are monitored for outcome under study
(recurrence of disease, first occurrence of disease, getting
better, side effects)
Rates of the outcome in the various groups are compared
Women with stage I breast cancer given a
lumpectomy alone will survive as long without
recurrence of disease as women given a lumpectomy
plus radiation
Water supply with fluoride will decrease the
frequency of dental caries in a community compared
to a similar community without water treated with
fluoride
Who will be in the study?
Must be defined specifically before study
begins
Remove subjectivity
Reproducibility
Women’s Health Study
≥ 45 years
No history of coronary heart disease, cerebrovascular
disease, cancer, or other major chronic illness
No history of side effects to any of study medications
Were not taking any of following meds more than once per
week: aspirin, NSAIDs, supplements of vitamin A, E, or
beta-carotene
Were not taking anticoagulants or corticosteroids
NEJM 352;13:1293-1303
How many participants do we need to enroll
in the study?
Programs and tables exist to calculate sample
size based on various parameters
TRUTH IN THE POPULATION
CONCLUSION
FROM SAMPLE
Fail to reject Ho
(no difference)
Ho
(no difference)
H1
(there is a
difference)
Correct decision
Type II error
(Probability = b)
False negative
Reject Ho
(there is a
difference)
Type I error
(Probability = a)
Correct decision
(Probability = 1- b)
False positive
Type I and II errors can be reduced by increasing sample size
The difference in effect to be detected
Estimate of effect in one group
Level of significance (a)
Probability of concluding treatments differ when
they do not differ
Level of power desired (1 - β)
Probability of concluding treatments differ when
they do differ
1-sided or 2-sided test
Compare the outcome among “exposed” to what the
outcome would have been if unexposed
This comparison is counterfactual
Instead, compare the outcome among “exposed”
group to the outcome in a “substitute” population
Validity of inference depends on finding a valid
substitute population
Need to randomly assign participants to one of the
intervention groups (test or control)
Randomization
Next assignment is unpredictable
Coin toss to determine group allocation
Random number table, opaque envelopes
Computer
Main purpose
Reduces selection bias in the allocation of treatment
Each participant has an equal chance of being in test or
control group
Secondary purpose
If large enough sample size, produce treatment and control
groups with similar baseline characteristics
Control for known and unknown factors
Baseline Characteristics in a study of heart disease patients
Test Group
(n = 9,599)
Control Group
(n = 9,586)
Male (%)
72
72
White (%)
95
95
Current smoker (%)
29
30
Hypertension (%)
52
51
Stable angina (%)
22
22
High cholesterol (%)
41
41
Characteristic
Patients with a history of:
Baseline Characteristics in a study Maternal-Infant HIV Transmission
Test Group
(n = 239)
Control Group
(n = 238)
Median age at entry (yrs)
24
25
White (%)
48
38
Gestational age at entry
29
30
Median (weeks)
26
27
14-26 weeks (%)
52
50
> 26 weeks (%)
48
50
41
41
Characteristic
Median CD4 county at entry
Treatment
Keep track of which treatment group the participant was
assigned
Keep track of which therapy received
Baseline data
Collect baseline demographic and other risk factor data
Compare treatment groups
Measuring outcome
Must be conducted in same fashion for all treatment
groups
Preventive studies
▪ Precursors of disease or first occurrence of disease
Therapeutic studies
▪ Symptom improvement
▪ Length of survival
▪ Disease recurrence
Myocardial infarction
Symptoms met WHO criteria
Abnormal levels of cardiac enzymes or diagnostic
electrocardiograms
Stroke
New neurologic deficit of sudden onset that persisted for
at least 24 hours
Death from cardiovascular disease
Examination of autopsy reports, death certificates, medical
records, and information obtained from the next of kin or
other family members
Masking (Blinding)
Prevents conscious and subconscious bias in research
Use placebo to mask
Single blind: participants do not know which treatment
they are receiving
Double blind: participants and observers (data collectors)
do not know participant treatment status
Parallel
Participants in each group simultaneously receive one
study treatment
Treatment and comparison groups consist of different
participants
Crossover
Planned reversal of intervention and control groups
Each participant can serve as his/her own control
STUDY
POPULATION
RANDOMLY ASSIGNED
NEW TREATMENT
CURRENT TREATMENT
Group 1
Group 2
Observe and
Measure Effects
Group 1
Group 2
Group 2
Group 1
Observe and
Measure Effects
Group 2
Group 1
Simple
Each group receives a treatment consisting of one
component (e.g. one drug)
Factorial
Use same study population to compare 2 or more
treatments
2 x 2 factorial design
Similar to 3 arms (drug A, drug B, and placebo) with fewer
participants
Drug A
Yes
No
Yes
Both
A and B
(cell a)
B only
(cell b)
No
A only
(cell c)
Neither
(cell d)
Drug B
Efficacy
of A
a+c v. b+d
Efficacy
of B
a+b
v.
c+d
Aspirin
Betacarotene
Yes
No
Efficacy
of Aspirin
Yes
No
Aspirin and Beta-carotene
Beta-carotene
only
(cell a)
(cell b)
Aspirin only
(cell c)
Neither
(cell d)
a+c v. b+d
Efficacy of
BetaCarotene
a+b
v.
c+d
Overt
Notify investigators he/she is dropping out of study
Drop outs
Covert
Stop taking assigned treatment without telling investigators
Need to build compliance checks in to the study (e.g. test
urine, count pills, etc.)
Efficacy
Reduction in risk
Calculate risk of death, developing disease,
complications in each group
=
Vaccine example
(Rate in placebo group) – (Rate in vaccine group)
Rate in placebo group
Relative risk
Kaplan-Meier plot
Hazard ratio
Number of patients who would need to be treated
(NNT) to prevent 1 adverse event
Number needed to harm (NNH) indicates number
patients treated to cause harm in 1 patient who
would not otherwise have been harmed
Internal validity
Extent to which the study groups are comparable
Comparability
Reflected by selection/randomization
External validity
Extent to which the results of a study can be applied to
people not in it
Generalizability
Representativeness
REFERENCE
POPULATION
External
Validity
STUDY
POPULATION
RANDOMLY ASSIGNED
CURRENT
TREATMENT
NEW
TREATMENT
Internal Validity
Items affecting internal validity
Loss to follow-up
Lack of randomization
Items affecting external validity
Loss to follow-up
Low response rate
Narrow inclusion criteria
Randomization
There must be genuine uncertainty about which treatment
is better
Informed consent
Some trials enroll participants immediately after diagnosis
When to stop the study?
Harmful or beneficial effects of one treatment arm
Outside board monitors study
Expensive and time-consuming
Ethical concerns may arise
A large number of participants may be required
Participant exclusion may limit generalizability
Compliance may be an issue
Influence of sponsorship
Randomization tends to balance risk factors across
study groups
Blinding of participants can reduce bias in
assessment of outcomes
Prospective design
Eliminate bias by comparing two otherwise identical
groups
Detailed information collected at baseline and
throughout study period
Experimental studies top epidemiologic study design
hierarchy in terms of validity
Investigators assign treatment to participants
(experimental)
Randomization reduces selection bias in treatment
allocation
Data collection must be conducted systematically
Noncompliance and drop-outs must be minimized to
increase validity of results
Center for Public Health Continuing Education
University at Albany School of Public Health
Department of Community & Family Medicine
Duke University School of Medicine
Mike Barry, CAE
Lorrie Basnight, MD
Nancy Bennett, MD, MS
Ruth Gaare Bernheim, JD, MPH
Amber Berrian, MPH
James Cawley, MPH, PA-C
Jack Dillenberg, DDS, MPH
Kristine Gebbie, RN, DrPH
Asim Jani, MD, MPH, FACP
Denise Koo, MD, MPH
Suzanne Lazorick, MD, MPH
Rika Maeshiro, MD, MPH
Dan Mareck, MD
Steve McCurdy, MD, MPH
Susan M. Meyer, PhD
Sallie Rixey, MD, MEd
Nawraz Shawir, MBBS
Sharon Hull, MD, MPH
President
Allison L. Lewis
Executive Director
O. Kent Nordvig, MEd
Project Representative