Trebuchet 48pt

1
CLASS SESSION 11
RANDOMIZED CONTROLLED TRIALS
Epidemiology 503, Section 2
Announcements

Exams expected back Wednesday

New groups!
Why Perform Studies?





Etiology or cause of disease
Extent of disease in the community
Natural history and prognosis
Evaluate interventions
Provide foundation for developing public policy
Objective of Epidemiologic Studies
Causal?
Factor (“exposure”)
(social, behavioral,
genetic, drug
treatment, policy, etc.)
Does exposure CAUSE outcome?
Health Outcome
(health event,
disease occurrence,
death, relapse)
Hierarchy of Study Designs
Observational Studies
Ecologic Studies
Generate
Hypotheses
Cross Sectional
Studies
Good place to
start investigating
a research
question
Case Control Studies
Cohort Studies
Experimental
Studies
Randomized Trials /
Intervention Studies
Establish
Causality
Gold standard for
establishing
causality
The Gold Standard of Study Designs

Randomized-Controlled Trials (RCTs)
 Randomly
assign persons or groups to either receive or
not receive the exposure

Considered the ideal design for evaluating
effectiveness and side effects of new interventions
or treatments
Source Population
Study Population / Defined Population
RANDOMIZATION
Exposed to
factor
Event
No event
Not exposed to
factor
Event
No event
Source Population
Study Population / Defined Population
RANDOMIZATION
New treatment
Improved
Not
improved
Current treatment
Improved
Not
improved
Components of a Randomized Trial
9


Randomization
Blinding (Masking)
Randomization
10



Every participant has equal probability of being
selected into the study arm
Randomize AFTER participants are screened
Makes control and treatment groups similar for
factors known and unknown
 Use
statistical tools to compare means or proportions of
these variables by group (ANOVA, Chi-Square test)

Removes investigator bias in the allocation of
participants
Blinding (Masking)
11

Single blinded trials
 Subjects
do not know which treatment group they are
assigned to

Double blinded trials
 Subject
and data observers (and analysts) are
unaware of treatment assignment
Validity
Results of a study are valid if they reflect the
truth
 Internal
validity refers to the design, methods,
analysis and results of an individual study. Did
the study do what it intended to do?
 External validity refers to whether internally
valid findings are also applicable
(generalizable) to people other than those
included in an individual study.
Source Population
External Validity
External Validity:
Generalization of study
to larger source
population
Study Population
RANDOMIZATION
Internal Validity
Exposed to
factor
Event
No event
Internal Validity:
Ability to reach correct
conclusion from study
- Reliability of data
- Compliance
- Loss to follow-up
Not exposed to
factor
Event
No event
Important Note

A study that is internally invalid cannot ever be externally
valid because it is fundamentally flawed and will result in
biased (i.e., wrong) results
Measures of Association: Relative
15

Ratio of two measures of disease incidence
 Risk
Ratio (Relative Risk)
CIexposed
CIunexposed
 Rate
=
a / (a + b)
c / (c + d)
Ratio
IRexposed
IRunexposed
=
a / Person Timeexp
c / Person Timeunexp
> 1 Exposed/treated has more risk than unexposed/control
= 1 Exposed/treated has same risk as unexposed/control
< 1 Exposed/treated has less risk than unexposed/control
Outcome
Exp
+
-
+
a
b
-
c
d
Total
Total
TOTAL
Measures of Association: Difference
16

Difference of two measures of disease incidence
 Risk
Difference
CIexposed – CIunexposed = [a/(a+b)] – [c/(c+d)]
 Efficacy
CIunexposed – CIexposed
Outcome
Exp
+
CIunexposed
+
a
b
-
c
d
Total
Total
TOTAL
The RCT as “gold standard”


Advantages
Strongest evidence for
causality
If blinded, less observer
bias




Disadvantages
Not “real life”
High cost
Inappropriate or
unethical for many
questions
Subject to poor
compliance, loss to
follow-up
Today. . .
18

In-class activity re: findings from DASH (Dietary
Approaches to Stop Hypertension) Study on dietary
patterns & blood pressure
Group 1
Group 2
Group 3
Group 4
Group 5
19
Cousineau, Cody
ccousz
Ma, Heyun
heyunma
Zhang, Shukai
shukaiz
Burr, Laura
lburr
De Jong, Maxwell
mgdejong
Empey, Margaux
empeym
Liang, Chen
ccliang
Dave, Ashka
Group 6
Sheth, Alexandra
assheth
Kilian, Tess
tkilian
Parkinson, Patrick
pparkins
Serpuja, Gita
gserpuja
Pont, Cassidy
casspont
Yaros, Thomas
tfyaros
addave
Zhu, Ziwei
zwzhu
Espana, Katherine
kvespana
Craig, Nathan
nacraig
Fan, Lingxiao
lingxfan
Compton, Emily
eacompto
Spicer, Rosanna
rrspicer
Metz, Liza
lizamm
Gao, Chao
gchao
Amin, Shivali
shivali
Davis, Ellen
emdavi
Ma, Ying
yingma
Hamilton, Shannon
shanlham
Mulenga, Lukonde
lukondem
Sidhar, Shubhum
ssidhar
Yue, Xubo
maxyxb
Brewer, Aubrey
brewerau
Cloeter, Elyse
cloetere
Flynn, Nicole
flnicole
Amaya, Kerent
kerentam
Robinson, Myshelle
marobi
Eck, Hannah
eckh
Babcock, Alexandra Maccarthy
ambab
Enache, Emma
enacheem
Urdahl, Nicole
nurdahl
Fucinari, Juliana Eva
jfucinar
Tierney, John
jomckayt
Shackelford, Mira
mirasha
Novoa, Alfredo
novoa
Fang, Fang
ffa
Sharp, Whitney
sharpwe
Higgins, Madeline
mlucille
Group 7
Group 8
Group 9
Group 10