1. Successful
Scien+fic
Wri+ng
Eugene
Elbert,
MS
(Johns
Hopkins
University,
U.S.)
Special
thanks
to
:
Paul
Siegel
MD,
MPH
9-­‐10
August
2012

2. Biological
Threat
Reduc+on
Program
of
the
Defense
Threat
Reduc+on
Agency
(DTRA)
2

3. Biological
Threat
Reduc+on
Program
• Consolidate
especially
dangerous
pathogens
(EDPs)
into
one
or
two
safe,
secure
central
reference
laboratories
or
repositories
• Build
and
sustain
long-­‐term
partnerships
through
interna+onal
scien+fic
engagement
and
coopera+on
• Improve
capacity
to
detect,
diagnose
and
report
outbreaks
and
poten+al
pandemics
by
providing
training
to
personnel
of
the
appropriate
facili+es
3

4. Biological
Threat
Reduc+on
Program
(BTRP)
• EDPs
for
human
and
animal
health
include:
o Avian
and
pandemic
influenza
(influenza
viruses)
o Crimean-­‐Congo
Hemorrhagic
Fever
(CCHF
virus)
o Anthrax
(Bacillus
anthracis)
o Brucella
(Brucella
species)
o Tularemia
(Francisella
tularensis)
o Botulism
(Clostridium
botulinum)
o Tick
Borne
Encephali+s
(TBE
virus)
o Plague
(Yersinia
pes6s)
o Foot
and
Mouth
Disease
(FMD)
o Glanders
o Newcastle
Disease
Virus
o Rinderpest
o Pox
viruses
(goat
and
sheep
pox)
o Swine
fevers
(African
and
Classical
Swine
Fever)
• Although
the
BTRP-­‐provided
training
focus
on
these
pathogens,
the
knowledge
and
skills
learned
and
prac+ced
are
applicable
to
a
broad
range
of
other
infec+ous
diseases
and
public
and
animal
health
concerns
4

5. 5

6. BTRP-­‐Provided
Training
Courses
include:
• Disease
recogni+on;
• Laboratory
equipment
use
and
maintenance;
• Biosafety
and
security;
• Laboratory
safety;
• Laboratory
quality
systems;
• Respiratory
protec+on
program;
• Purchasing
and
inventory
control;
• Introduc+on
to
microbiology;
• Introduc+on
to
molecular
biology;
• Introduc+on
to
immunology/serology;
• Diagnos+c
assays
for
specific
EDPs;
• Laboratory
management;
• Sample
collec+on
and
processing;
• Basics
of
epidemiology;
and
others
6

7. Conceptual
design
of
approved
TIPME
Training
Facility
7

8. BTRP
Summary
• Enhancement
of
exis+ng
surveillance
capacity
through
expansion
of
generic
skills
• Development
of
capacity
for
rapid
detec+on
(PCR
and
ELISA),
which
contributes
to
public
health
• Improved
biosafety
and
biosecurity
for
laboratory
personnel
• BTRP-­‐provided
training
complements
the
Ministry
training
requirements
for
specialists
8

9. Successful
Scien+fic
Wri+ng
9

10. Introduc+on
Objec+ves
of
the
workshop:
• To
introduce
basic
concepts
of
scien+fic
approach
• To
detail
the
structure
and
format
of
scien+fic
papers.
• To
compare
examples
of
different
research
designs.
• To
examine
components
of
a
scien+fic
paper.
• To
cri+cally
examine
published
examples
of
scien+fic
wri+ng.
• To
apply
new
wri+ng
skills
to
draging
an
abstract.
• To
learn
about
the
submission
process
for
publica+ons,
funding
proposals,
and
presenta+ons
10

11. Why
do
we
publish?
• Presen+ng
research
• Reaching
global
scien+fic
community
• Advancing
science
• Educa+on
• Funding
and
credibility
11

12. 12

13. Repor+ng
Scien+fic
Research
• Hypothesis
or
research
ques+on
• Planned
research
• Ethics
– Plagiarism
– Misuse
of
data
and
informa+on
– Conflict
of
interest
– Integrity
– Human
subject
research
13

14. Process
of
scien6fic
wri6ng
Submiing
Hypothesis
Wri+ng
Study
plan
ar+cle
Having
Experiment
journal,
audience
in
mind
Results
Data
processing
genera+on
14

15. General
Guidelines
for
Scien+fic
Papers:
Style
and
Content
EASE
guidelines
• Complete,
concise
and
clear
• For
effec+veness
of
interna+onal
coopera+on
all
publica+ons
should
be:
• COMPLETE,
CONCISE
AND
CLEAR!
• IMPORTANT
15

16. General
Guidelines
for
Scien+fic
Papers:
Style
and
Content
• Do
not
include
irrelevant
informa+on
• Informa+on
should
not
be
repeated
• Include
only
necessary
tables
and
figures
• Cap+ons
–
informa+ve
but
concise
• Delete
redundancies
• Define
abbrevia+on
at
first
use
• Do
not
over-­‐generalize
• Numbers
for
all
numerals
16

17. Content
• Study
should
be
planned
in
advance
• The
journal
and
the
audience
should
be
chosen
• Informa+on
should
be
organized
• All
the
components
of
scien+fic
ar+cle
should
be
present
and
sa+sfy
the
guidelines
for
a
chosen
journal
17

18. Repor+ng
Guidelines:
Content
• Dis+nguish
your
original
ideas
• Paraphrase
text
from
other
sources
• Proper
terms
(plant
community
vs.
phytocoenosis)
• Define
every
uncommon
term
• Avoid
ambiguity
• Be
clear
what
you
regard
as
100%
when
repor+ng
%
• SI
units
(interna+onal
system
of
units;
metric)
• Decimal
point
• Remember
that
the
text
will
be
read
by
foreigners
18

19. Repor+ng
Guidelines:
Content
• Make
posi+ve,
objec+ve
asser+ons,
directly
supported
by
the
results,
with
necessary
qualifica+ons
and
caveats
• Don’t
oversell:
“This
result
clearly
proves
that
the
neural
network
approach
is
superior
and
will
revolu+onize
research
methods”.
• Don’t
base
substan+al
claims
on
unpublished
data
or
on
“experience”
without
objec+ve
suppor+ng
evidence.
• If
you
rely
on
a
reference
to
draw
a
conclusion,
be
sure
the
reference
supports
the
idea,
and
say
where
the
support
may
be
found
in
the
reference.
19

20. A
Dic+onary
of
Useful
Research
Phrases
• "It
has
long
been
known..."
• I
didn't
look
up
the
original
• "It
is
believed
that..."
references
• "It
is
generally
believed
• I
think
that..."
• My
friends
think
so,
too
• "A
sta+s+cally
oriented
projec+on..."
• Wild
guess
• “Typical
results
are
shown”
• Best
results
are
shown
• “Obviously,
we
will
need
• I
don’t
understand
anything
addi+onal
studies”
• “Authors
thanks
Joe
in
• Joe
did
the
work
and
conduc+ng
experiment
and
George
explained
it
to
me
George
for
helpful
comments”
20

21. Example
“In
order
to
provide
analy+c
control
during
forensic-­‐
chemical
inves+ga+on,
it
is
customary
to
use
highly
sensi+ve
and
specific
analysis
methods.
Very
popular
in
the
prac+ce
of
chemic-­‐toxic
studies
is
the
TLC
method
in
view
of
its
accessibility,
ease
of
conduc+ng
and
expressiveness.
Due
to
the
possibility
of
changing
not
only
sorbents
but
also
solvents,
it
is
possible
to
quickly
solve
the
problems
of
separa+on”
21

22. Repor+ng
Guidelines:
Text
Structure
• Simple
sentences,
should
not
be
very
long
• Avoid
passive
voice
• Text
should
be
cohesive,
logically
organized
• Each
paragraph
should
start
with
a
topic
sentence
• Use
text
tables
• Make
figures
and
tables
understandable
by
themselves
• Explain
your
figures
and
charts,
and
jus+fy
their
inclusion.
Do
not
just
show
them
with
no
stated
reason.
22

23. Text
tables
Original
sentence:
• Iron
concentra+on
means
(±standard
devia+on)
were
as
follows:
11.2±0.3
mg/dm3
in
sample
A,
12.3±0.2
mg/
dm3
in
sample
B,
and
11.4±0.9
mg/dm3
in
sample
C.
Modified:
• Iron
concentra+on
means
(±standard
devia+on,
in
mg/dm3)
were
as
follows:
• sample
B
12.3±0.2
• sample
C
11.4±0.9
• sample
A
11.2±0.3
23

24. Replace
phrases
with
a
single
word
• Considering
this
fact
• In
the
rela+on
to
• Exceeding
number
• In
the
previous
case
• In
the
absence
• In
large
number
of
cases
24

25. Passive
Voice
“Have
you
ever
been
told
to
use
passive
voice”
or
“Did
anyone
tell
you
to
use
passive
voice”
Examples:
• “James
Watson
was
awarded
the
Nobel
Prize
for
discovering
the
molecular
structure
of
DNA.“
vs.
• "The
Nobel
CommiSee
awarded
James
Watson
the
Nobel
Prize
for
discovering
the
molecular
structure
of
DNA."
25

26. Passive
voice
Nobody
takes
responsibility
in
passive
voice:
“Mistakes
were
made
during
the
experiment”
vs.
We
made
mistakes
during
the
experiment
“It
is
shown
in
the
table”
vs.
The
table
shows
26

27. Example
Common
dysfunc+on
of
the
immune
system
was
shown
in
the
trials
on
humans
and
animals
__________________________________
Trials
on
humans
and
animals
show
a
common
dysfunc+on
of
the
immune
system
27

28. Correct
Use
of
Passive
Voice
• When
the
ac+on
is
more
important
than
the
agent
of
it
(as
in
Materials
and
Methods)
• In
order
to
emphasize
somebody
other
than
the
ac+ng
agent
• When
the
agent
is
unknown
28

29. Repor+ng
Guidelines:
Language
• Use
commonly
known
words,
but
not
idioma+c
expressions
• Define
abbrevia+ons
(avoid
them
in
abstract)
• Spelling
• Past
tense
in
body,
present
in
general
statements
• Refer
to
the
author
as
“we”
or
“I”
not
“the
author”
29

30. Repor+ng
Guidelines:
Language
Transforma2on
of
verbs
into
nouns
Obtained
es+mates
–
es+mated
Gained
improvement-­‐
improved
Showed
growth
–
grew
Made
a
decision
–
decided
30

31. Common
Fallacies
in
Wri+ng
• Non
Causa
Pro
Causa
Fallacies
—
No
Cause
for
Cause
• Asempts
to
establish
a
causal
rela+onship
– Cum
Hoc,
Ergo
Propter
Hoc
– Post
Hoc,
Ergo
Propter
Hoc
– The
Regression
Fallacy
– Texas
Sharpshooter
Fallacy
31

32. Fallacies
in
Wri+ng
Cum
Hoc,
Ergo
Propter
Hoc
—
With
This,
Therefore
• African
American
popula+on
is
more
likely
to
experience
metabolic
consequences
of
Chronic
Kidney
Disease
(CKD)
before
reaching
the
eGFR
<60
ml/min
threshold
…
that
these
observa+ons
support
a
need
to
adapt
clinical
prac+ce
guidelines
shiging
screening
for
CKD
to
a
higher
eGFR
threshold
specifically
for
African
Americans
(1)
• The
assump6on
that
the
measured
clinical
parameters
in
this
representa6ve
popula6on
are
physiologically
linked
to
CKD
in
African
Americans
is
simplis6c
and
ignores
the
effects
of
a
combina6on
of
gene6c
and
physiologic
adapta6ons
superimposed
on
a
background
of
social
and
environmental
factors
that
account
for
minority
health
dispari6es
(2)
• Lesson:
Adjustment
for
possible
confounders
and
other
sources
of
bias
32

33. Fallacies
in
Wri+ng
Post
Hoc,
Ergo
Propter
Hoc
—
AAer
This,
Therefore
Because
of
This
• “Since
that
event
followed
this
one,
this
event
must
have
caused
that
one.”
It
also
is
referred
to
as
“false
cause”
or
“coincidental
correla+on.”
• 7
women
in
California
developed
ovarian
cysts
taking
the
new
mul+phasic
oral
contracep+ve
pills
which
led
to
case
series
report
and
media
prin+ng
the
story
[1].
• No
associa6on
was
shown
in
follow-­‐up
studies
[2]
• Lesson:
Checking
for
possible
confounders,
conduc+ng
valida+on
studies
before
jumping
to
conclusions,
repor+ng
on
it
in
wri+ng
33

34. Fallacies
in
Wri+ng
Texas
Sharpshooter
Fallacy
Outbreak
foci?
• In
medical
research,
this
fallacy
occurs
when
inves6gators
select
certain
data
to
demonstrate
a
cause-­‐effect
rela6onships.
34

35. Fallacies
in
Wri+ng
The
Art
of
Argumenta
– Argumentum
ad
Ignoratum
(Appeal
to
Ignorance):
Absence
of
evidence
is
not
evidence
of
absence
Width
of
Confidence
Interval(±w)
Sample
Size(n)
0.01
9612
0.02
2403
0.03
1068
0.05
384
0.10
96
0.15
43
Sample
sizes
required
to
es2mate
a
true
prevalence
of
0.50
with
95%
confidence
intervals
of
different
widths
(±w)
Lesson:
Making
sure
that
the
sample
size
is
large
enough.
Recognizing
beneficence
and
non-­‐maleficence
35

36. Fallacies
in
Wri+ng
Argumentum
ad
Verecundiam
(Appeal
to
Authority):
Users
of
this
fallacy
ogen
call
upon
the
published
works
of
others
to
bolster
their
arguments,
without
ques+oning
the
accuracy,
reliability,
or
validity
of
those
sources
• Quote
from
an
editor
as
a
condi+on
for
publica+on
highlights
the
problem:
“you
cite
Leukemia
[once
in
42
references].
Consequently,
we
kindly
ask
you
to
add
references
of
ar6cles
published
in
Leukemia
to
your
present
ar6cle”
(1)
• Editors'
incen+ve
to
inflate
impact
factors
through
self-­‐
cita+on
• Survey
found
that
having
a
tenure
posi6on
also
increased
coercion
• Lesson:
Being
true
to
your
work
36

37. Fallacies
in
Wri+ng
Argumentum
ad
An;quitatem
(Appeal
to
Tradi2on
or
History)
“(Talking
about
acupuncture)
I
think
it
is
insul+ng
to
say
that
Chinese
people
would
carry
on
with
some
sort
of
mys+cal
belief
when
it
didn’t
work”
“Well,
you
know
–
acupuncture
is
one
of
those
amazing
things.
I
mean
it
has
been
around
for
several
thousand
years
.
.
.
there
is
a
huge
amount
of
validity
to
what
it
represents,
and
there
has
to
be
–
or
it
wouldn’t
have
survived
such
a
long
+me
“
Lesson:
Not
making
unsupported
claims
37

38. Fallacies
in
wri+ng
• Argumentum
ad
Populum
(Appeal
to
the
People
or
Popularity)
• 4
from
5
den+sts
recommend
sugar-­‐
free
“Trident”“
chewing-­‐gum!
• The
adver+sement
“forgot”
to
men+on
“If
pa+ents
INSIST
to
use
chewing-­‐gum”.
They
also
hid
each
5th
den+st
recommended
to
avoid
the
use
of
chewing-­‐gum.
• «Thus
based
on
the
assessment
of
leading
Russian
clinics
“Sangviri+n”
is
one
of
the
effec+ve
modern
an+microbial
drug
of
local
and
common-­‐
resorp+ve
ac+on
for
preven+on
and
treatment
of
different
infec+ous
diseases
[14–17].»
7/28/2012

39. Fallacies
in
Wri+ng
Myths
of
Beneficence
An
analysis
of
60
adver+sements
that
had
appeared
in
the
Bri+sh
Medical
Journal
between
1999
and
2001
demonstrated
that
drug
adver+sing
uses
strong
imagery
to
fabricate
mythical
associa+ons
between
medical
condi+ons
and
branded
drugs,
and
that
drug
adver+sing
manipulates
readers’
percep+ons
by
subtle
appeal
to
ancient
and
modern
mythological
founda+ons
of
humanism
and
Western
psychology.
39

40. Fallacies
in
Wri+ng
False
Dichotomy
This
is
also
called
a
false
dilemma,
an
either-­‐or
fallacy,
fallacy
of
false
choice,
or
black-­‐and-­‐
white
thinking.
Most
wide-­‐spread
false
dichotomy
in
scien+fic
repor+ng:
Sta+s+cal
significance
P
=
0.049
vs.
P
=
0.051
40

41. Fallacies
in
Wri+ng
Essen2alism
Some
argument
in
print
or
spoken
word,
some
“essen+al
feature”
is
proposed
as
a
defining
characteris+c
of
an
otherwise
complex
issue
or
larger
problem
Each
scien+fic
specialty
looks
at
disease
differently.
For
example,
cancer
from
the
perspec+ve
of
a
general
surgeon,
a
pathologist
or
an
acupuncturist
are
completely
different.
Lesson:
To
be
aware
of
specialized
terminology
and
body
of
knowledge
when
repor+ng
41

42. Fallacies
in
Wri+ng
Редукционизм
Efforts
to
simplify
the
problem
to
the
simple
rela+ons
(O’Connor
et
al.
2011):
“Reduc+onist
methods
of
disease
control
involve
the
removal
of
infec+on
or
the
infec+ous
agent,
implemen+ng
barriers
to
direct
and
indirect
transmission
or
by
increasing
inherent
or
acquired
immunity
to
the
infec+ous
agent.
However,
for
those
diseases
which
evade
such
methods
of
conven+onal
control,
a
more
comprehensive
understanding
of
the
complex
interac+ons
amongst
biological
(agent
and
host(s)),
environmental,
economic
and
social
factors
which
can
affect
the
emergence
and
spread
of
an
infec+ous
disease
is
required.”
42

43. Things
to
avoid:
• Plagiarism
• Fishing
expedi+ons
–
research
must
be
hypothesis
driven
• Do
not
plan
your
study
in
order
to
use
your
results
to
pool
evidence
against
the
same
problem
(e.g.
meta-­‐analyses.
• Do
not
fail
to
take
into
account
heterogeneity,
uncertainty
and
dependence
• Do
not
fail
to
have
a
robust
exploratory
data
analysis
(EDA)
before
proceeding
into
any
confirmatory
tes+ng
(John
Tukey
teachings)
• Do
not
discount
the
importance
of
internal
and
external
validity
when
interpre+ng
results
• Do
not
underes+mate
the
sta+s+cs.
The
absence
of
evidence
is
not
the
evidence
of
absence
–
your
study
may
not
have
enough
power
to
detect
anything
unless
you
have
large
numbers
43

44. Things
that
annoy
reviewers
– Poor
English
– Repe++on
– Lack
of
structure
in
the
text
– Sentences
that
are
too
convoluted
and
long
– Lack
of
asen+on
to
detail
(a
premature
drag
with
typographical
errors,
etc.)
– Not
well
thought
out
statements
(make
each
word
count)
– Obscure
methods
or
not
well
described
– Oversta+ng
the
results
– Too
long
of
a
paper
44

45. Repor+ng
Guidelines:
Structure
• IMRaD
standard
(Introduc+on,
Methods,
Results,
and
Discussion)
• Design
Specific
–
EQUATOR
network
• Journal
-­‐specific
• General:
– Title
Page
– Conflict
of
Interest
No+fica+on
Page
– Abstract
– Introduc+on
– Methods
– Results
– Discussion
– References
45

46. Standardizing
Health
Repor+ng
EQUATOR
(Enhancing
Quality
and
Transparency
of
Health
Research)
network:
“Too
oaen,
good
research
evidence
is
undermined
by
poor
quality
repor6ng”
• Raising
awareness
of
the
crucial
importance
of
good
repor+ng
of
research
• Becoming
the
recognized
global
center
providing
resources,
educa+on
and
training
rela+ng
to
the
repor+ng
of
health
research
and
use
of
repor+ng
guidelines
• Assis+ng
in
the
development,
dissemina+on
and
implementa+on
of
repor+ng
guidelines
• Monitoring
the
status
of
the
quality
of
repor+ng
across
health
research
literature
• Conduc+ng
research
rela+ng
to
the
quality
of
repor+ng
46

47. Guidelines
for
Repor+ng
Common
Study
Types
• CONSORT
–
Consolidate
Standards
of
Repor+ng
Trials
• STROBE
–
Strengthening
the
Repor+ng
of
Observa+onal
studies
• STARD
–
Standards
for
repor+ng
of
Diagnos+c
Accuracy
• QUOROM
–
Quality
of
Repor+ng
of
Meta-­‐
analyses
(under
CONSORT)
47

48. Example
–
STROBE
checklist
Item No Recommendation
Title and abstract 1 (a) Indicate the study’s design with a commonly used
term in the title or the abstract
(b) Provide in the abstract an informative and balanced
summary of what was done and what was found
Introduction
Background/rationale 2 Explain the scientific background and rationale for the
investigation being reported
Objectives 3 State specific objectives, including any prespecified
hypotheses
Methods
Study design 4 Present key elements of study design early in the paper
Setting 5 Describe the setting, locations, and relevant dates,
including periods of recruitment, exposure, follow-up, and
data collection
Participants 6 (a) Cohort study—Give the eligibility criteria, and the
sources and methods of selection of participants. Describe
methods of follow-up
Case-control study—Give the eligibility criteria, and the
sources and methods of case ascertainment and control
selection. Give the rationale for the choice of cases and
controls
Cross-sectional study—Give the eligibility criteria, and
the sources and methods of selection of participants
(b) Cohort study—For matched studies, give matching
criteria and number of exposed and unexposed
Case-control study—For matched studies, give matching
criteria and the number of controls per case
48

49. Study
Designs
in
Public
Health
Experimental
(Interven2onal)
Studies
Observa2onal
Studies
Randomized
Trials
Case
reports
Community
Trials
Case
Series
Descrip+ve
Therapeu+c/Preven+ve
Trials
Cross-­‐sec+onal
Studies
Surveillance
Cohort
Studies
Analy+c
Case-­‐Control
49

50. Observa+onal
Descrip+ve
Studies
• Case
Reports
–
detailed
presenta+ons
of
a
single
case
or
a
handful
of
cases.
“Normal
Plasma
Cholesterol
in
an
88-­‐Year-­‐Old
Man
Who
Eats
25
Eggs
a
Day
—
Mechanisms
of
Adapta+on”
[Kern
J,
NEJM
1991;
324:896–899]
• Case
Series
–survey
of
a
group
of
individuals
with
a
par+cular
disease
performed
at
a
single
point
of
+me.
“Pneumocy+s
pneumonia:
Los
Angeles”
[MMWR
Morbidity
and
Mortality
Weekly
Report
1981;30:250-­‐252]
50

51. Cross-­‐Sec+onal
Studies
• Describes
health
of
popula+ons
(both
exposed
and
non-­‐exposed)
• Iden+fies
prevalent
cases
• Finds
associa+on,
not
causa+on
• Best-­‐suited
for
lisle
disability,
pre-­‐symptoma+c
studies
• Surveys
• Good
for
planning
health
care
– Na+onal
Health
Surveys
are
a
good
example
51

52. Surveillance
• An
ongoing,
systema6c
collec6on,
analysis
and
interpreta6on
of
health-­‐related
data
essen6al
to
the
planning,
implementa6on,
and
evalua6on
of
public
health
prac6ce
• Detec+on
and
no+fica+on
of
health
events
• Collec+on
and
consolida+on
of
data
• Inves+ga+on
of
cases
and
outbreaks
• Rou+ne
Repor+ng
• Feedback
U.S.
CDC:
Ears,
EWIDS,
NTSIP,
ESP,
NEDSS,
FluNet,
BRFSS,
FoodNet,
etc.
Australia:
NNDSS
U.S.:
ProMED,
HealthMap
Canada:
FluWatch,
GPHIN
France:
GPs
Sen+nelles
Network
Asia:
APEC
EINet
WHO:
GOARN
Europe:
MedlSys
52

53. Case-­‐Control
Studies
• Comparison
of
cases
versus
non-­‐cases
(controls)
• Retrospec+ve
for
exposure
• Matching
all
popula+on
characteris+cs
of
cases
to
those
of
controls
(including
biases)
• Mostly
for
prevalent
cases
(but
could
be
for
incident
cases,
too)
53

54. Cohort
Studies
• To
support
the
rela+on
between
the
cause
and
disease
• Presence
or
absence
of
risk
factor
is
determined
before
outcome
occurs
• Longitudinal/prospec+ve/incidence
studies
• Cohorts
are
free
of
disease
at
baseline
• Cohorts
should
be
comparable
• Diagnos+cs
and
eligibility
should
be
defined
54

55. Cohort
vs.
Case-­‐Control
COHORT
STUDY
DATA
COLLECTION
Sick
Exposed
Sample
of
Not
Sick
disease-­‐free
individuals
Sick
Not
Exposed
Not
Sick
Exposed
Develop
Illness
Not
Exposed
Popula+on
Exposed
Don’t
Develop
Not
Exposed
Illness
Case-­‐Control
Data
Collec+on
55

56. Experimental:
Control
Study
Controlled:
– Inves+gator
decides
on
interven+on
Randomized:
– Gold
Standard
in
Epidemiological
research
– Controls
for
confounding
– Prevents
selec+on
Bias
Therapeu+c
vs.
Preven+ve:
Pa+ents
with
Disease
vs.
Popula+on
at
Risk
56

57. Experimental:
Controlled
Studies
DATA
COLLECTION
Exposure
COHORT
(Observa+onal)
occurs
naturally
Sick
Exposed
Sample
of
Not
Sick
disease-­‐free
individuals
Sick
Not
Exposed
Not
Sick
Inves+gator
CONTROLLED
(Interven+onal)
Determines
Exposure
57

58. Randomized
Clinical
Trial
•
Sample
size
should
be
sufficient
•
Possibility
to
follow
up
during
the
trial
•
Par+cipants
should
be
informed
of
risks/
benefits/
blinding/
placebo
•
Inclusion
Criteria
Reference
Popula+on
Reference
Popula+on
Experimental
Experimental
Popula+on
Popula+on
Study
Popula+on
Internal
Validity
External
Validity
58

59. Randomized
Clinical
Trial
• Design
– Simple
– Cross-­‐over,
factorial
• Sampling
• Eligibility
criteria
• Blinding:
single
vs.
double
• Alloca+on:
Randomiza+on
• Follow-­‐up
• Analysis
• Therapeu+c
vs.
Non-­‐therapeu+c
59

60. Randomized
Trial:
CONSORT
Flow
Eligible
Non-­‐eligible
Declined
Alloca+on
using
randomiza+on
scheme
Follow-­‐up
Included
in
analysis
60

61. Protocol of clinical study
(typical errors)
• During
development
of
CS
protocol:
– Fail
to
jus+fy
the
study
of
given
drug
by
the
given
indica+ons;
– Absence
of
pre-­‐clinical
and
clinical
(if
applicable)
trials;
– The
objec+ves
of
study
are
not
listed
(primary
and
secondary
objec+ves),
hypothesis
of
study;
– Mixed
concep+on
of
primary
objec+ve
of
study
and
criteria
of
efficacy;
– Sta+s+cs!
Instead
of
sample
size
jus+fica+on
and
sta+s+cal
power:
“the
assessment
will
be
performed
with
PC,
Excel,
Student’s
methods,
etc.”;
– Vague
procedures
and
methods,
allowing
ambiguous
interpreta+on;
– No
dates,
no
versions

62. Protocol of clinical study
(typical errors)
• While
repor+ng
of
CS:
– Vague
descrip+on
of
study
popula+on,
that
unable
the
formula+on
of
conclusion
about
homoscendacity;
– No
sta+s+cal
assessment
inclusion/exclusion
criteria
of
lost
follow-­‐up
pa+ents;
– No
side
therapy
details
and
its
effect
in
sta+s+cal
analysis;
– No
severity
and
resolving
of
side
effects
(e.g.
2
pa+ents
presented
the
head
ache
–
no
terms,
methods
od
treatment,
outcome,
etc.);
– No
pa+ents’
compliance
data;
– Separate
reports
from
each
center
instead
of
all-­‐centers
consolidated
report
…

63. General
Guidelines
For
Selec+on
of
Study
Type
Study
objec2ve
Study
type
Study
of
rare
diseases
Case
control
studies
Study
of
rare
exposure,
such
as
exposure
to
Cohort
studies
in
a
popula+on
group
in
industrial
chemicals
which
there
has
been
exposure
(e.g.
industrial
workers)
Study
of
mul+ple
exposures,
such
as
the
Case
control
studies
combined
effect
of
oral
contracep+ves
and
smoking
on
myocardial
infarc+on
Study
of
mul+ple
end
points,
such
as
Cohort
studies
mortality
from
different
causes
Es+mate
of
the
incidence
rate
in
exposed
Exclusively
cohort
studies
popula+ons
Study
of
covariables
which
change
over
Preferably
cohort
studies
+me
Study
of
the
effect
of
interven+ons
Interven+on
studies
63

64. Costs
of
different
types
of
bias
for
different
study
designs
Ecological
Cross-­‐ Case-­‐ Cohort
study
sec2onal
control
study(and
study
study
RCT)
Selec+on
N/A
2
3
1
bias
Recall
bias
N/A
3
3
1
Loss
to
N/A
N/A
1
3
follow-­‐up
Confounding
3
2
2
1
Time
1
2
2
3
Required
Costs
1
2
2
3
1-­‐slight;
2-­‐moderate;
3-­‐high;
N/A=
not
applicable
64

65. Introduc+on
sec+on
Purpose:
to
convince
the
reader
that
your
study
will
yield
knowledge
or
know-­‐how
that
is
new
and
useful
• Iden+fy
a
gap
in
knowledge
or
know-­‐how
(study
problem)
o Provide
key
background
(scope/nature/magnitude
of
the
gap)
o Be
clear
that
filling
this
gap
will
be
useful.
o Describe
the
relevant
limita+ons
of
previous
studies
• Present
your
approach
to
filling
the
gap
(study
purpose)
o Be
clear
that
your
approach
is
new
o Emphasize
that
your
approach
addresses
the
limita+ons
of
previous
studies
in
a
logical
and
compelling
way
Oaen
requires
just
three
paragraphs
65

66. Introduc+on
Checklist
Background Statement:
Scope nature magnitude of the gap
Be clear that filling the gap is useful
Problem Statement
Describe relevant limitations
Study Statement
Be clear that your approach is new
Emphasize that your approach addresses limitations
Summary Statement
Summarizes the study
66

67. Introduc+on
sec+on
• No
major
difference
in
introduc+on
sec+on
between
study
types
• Some+mes
summary
statement
is
omised,
or
becomes
part
of
the
study
statement
• STROBE:
Introduc+on
Background/ra+onale 2 Explain
the
scien+fic
background
and
ra+onale
for
the
inves+ga+on
being
reported
Objec+ves 3 State
specific
objec+ves,
including
any
pre-­‐specified
hypotheses
67

68. Introduc+on
sec+on
The
next
four
slides
detail
the
introduc+on
checklist
process
for
four
separate
studies:
• Background
statement
• Problem
statement
• Study
statement
– General
descrip+on
of
the
surveillance
system
• Summary
statement
68

69. Background
The
treatment
of
human
immunodeficiency
virus
(HIV)
infec+on
has
undergone
Statement:
considerable
change.
Protease
inhibitors
and
non–nucleoside-­‐analogue
reverse-­‐transcriptase
inhibitors,
when
used
as
part
of
combina+on
drug
regimens,
can
profoundly
suppress
viral
replica+on,
with
consequent
reple+on
of
CD4+
cell
counts.
Mul+ple
clinical
trials
have
shown
the
virologic
and
immunologic
efficacy
of
the
newer,
highly
ac+ve
an+retroviral-­‐drug
combina+ons
by
measuring
the
plasma
load
of
HIV
RNA
and
CD4+
cell
counts.
In
addi+on,
prophylac+c
medica+ons
are
now
being
used
rou+nely
to
prevent
disseminated
Mycobacterium
avium
complex
infec+on
Problem
Several
reports
have
described
reduc+ons
in
mortality
and
in
the
rate
of
Statement
hospitaliza+on
of
HIV
infected
pa+ents;
however,
such
reduc+ons
have
not
been
clearly
related
to
specific
therapeu+c
regimens.
Study
Statement
We
analyzed
data
collected
over
42
months
in
the
HIV
Outpa+ent
Study.
During
this
period,
rates
of
chemoprophylaxis
against
opportunis+c
infec+on
remained
rela+vely
constant
even
while
paserns
of
an+retroviral
therapy
were
changing
Summary
This
report
outlines
the
changes
in
death
rates
and
the
incidence
of
Statement
opportunis+c
infec+ons
in
a
large
group
of
HIV-­‐infected
outpa+ents,
many
of
whom
had
previously
received
extensive
treatment.
69

70. Background
Among
the
few
diseases
claimed
to
occur
more
ogen
in
non-­‐smokers
than
Statement:
smokers
1
2
that
of
greatest
poten+al
importance
is
Alzheimer's
disease,
which
accounts
for
most
of
the
demen+as
of
later
life
in
Britain
Problem
The
published
epidemiological
evidence,
although
sugges+ve
of
an
inverse
Statement
rela+on
with
smoking,
is
not
conclusive
either
about
Alzheimer's
disease
or
demen+a
in
general.
Much
of
the
evidence
derives
from
small
retrospec+ve
studies
of
uncertain
reliability,
many
of
which
excluded
vascular
demen+a.
Prospec+ve
studies,
in
which
smoking
habits
are
recorded
before
the
onset
of
demen+a,
should
be
more
informa+ve
about
the
overall
effects
of
smoking,
par+cularly
if
they
concern
large
numbers
and
prolonged
follow
up.
Only
a
few
such
studies
have,
however,
been
properly
reported
(none
of
which
had
prolonged
follow
up)
Study
We
sought
evidence
from
the
cohort
of
Bri+sh
doctors
who
have
been
Statement
followed
since
1951,
with
their
smoking
habits
reviewed
every
six
to
12
years.3
4
Many
have
died
from
or
with
some
type
of
demen+a
over
the
past
two
decades.
Summary
Statement
70

71. Background
Alcohol
was
first
implicated
as
a
possible
risk
factor
for
stroke
in
1725(1)
Statement:
Several
epidemiological
studies
now
suggest
a
U-­‐shaped
associa+on
between
alcohol
intake
and
stroke(2).
Problem
Previous
studies
have
been
cri+cized
for
not
differen+a+ng
between
Statement
nondrinkers
who
were
lifelong
abstainers
and
those
who
had
given
up
drinking(3-­‐7)
By
asking
specifically
about
previous
regular
drinking
habits
we
have
been
able
to
dis+nguish
between
the
two
groups.
The
level
of
alcohol
consump+on
at
which
this
possible
protec+ve
effect
is
lost
and
alcohol
becomes
a
risk
factor
for
stroke
are
unknown.
Study
We
report
the
findings
of
a
case-­‐control
study
that
examines
the
contribu+on
Statement
of
alcohol
to
the
risk
of
stroke
in
moderate
and
heavy
drinkers
(both
currently
and
previously),
lifelong
abstainers
(those
who
have
never
drunk
alcohol),
and
current
abstainers
(those
who
had
formerly
been
regular
drinkers
but
who
currently
do
not
drink
alcohol),
using
validated
measures
of
alcohol
consump+on.
Summary
Statement
71

72. Background
Between
May
2009
and
May
2010,
Greece
experienced
two
waves
Statement:
of
influenza
A(H1N1)2009
transmission
Problem
Given
the
poten+al
for
worsening
in
the
clinical
severity
of
influenza
Statement
during
the
post-­‐pandemic
influenza
season,
as
was
the
case
for
previous
influenza
pandemics
[7-­‐9],
it
was
cri+cal
to
con+nue
surveillance
with
a
focus
on
severe
cases
and
their
clinical
characteris+c
Descrip2on
of
In
Greece,
influenza
is
annually
monitored
through
the
rou+ne
the
sen+nel
surveillance
system,
which
became
opera+onal
in
1999.
The
Surveillance
sen+nel
surveillance
system,
which
covers
approximately
three
System
percent
of
the
total
Greek
popula+on
in
the
2010/11
influenza
season,
provides
data
representa+ve
of
the
na+onal
popula+on
Summary
This
report
summarises
data
from
influenza
surveillance
in
Greece
Statement
during
the
post-­‐pandemic
2010/11
influenza
season.
72

73. Materials
and
Methods
Purpose:
to
describe
how
you
collected,
organized
and
analyzed
data
(relevant
to
the
study
purpose)
• Clearly
present/define
all
analysis
variables
• Organize
into
logical
subsec+ons
that
illustrate
the
steps
you
took
to
collect,
organize,
and
analyze
the
data:
o Study
popula+on
o Defini+on
of
variables
o Laboratory
methods/
epidemiological
inves+ga+on
o Interven+on
• Describe
what
you
did,
not
what
you
found
(Results)
• Respect
chronology
• Describe
the
original
methods
in
detail;
otherwise
give
references
Length
varies
depending
on
originality
of
methods
73

74. Materials
and
Methods
–
part1
Methods
Study
design Present
key
elements
of
study
design
early
in
the
paper
Seing Describe
the
seing,
loca+ons,
and
relevant
dates,
including
periods
of
recruitment,
exposure,
follow-­‐up,
and
data
collec+on
Par+cipants
and
(a)
Cohort
study—Give
the
eligibility
criteria,
and
the
sources
and
Seing methods
of
selec+on
of
par+cipants.
Describe
methods
of
follow-­‐up
Case-­‐control
study—Give
the
eligibility
criteria,
and
the
sources
and
methods
of
case
ascertainment
and
control
selec+on.
Give
the
ra+onale
for
the
choice
of
cases
and
controls
Cross-­‐sec6onal
study—Give
the
eligibility
criteria,
and
the
sources
and
methods
of
selec+on
of
par+cipants
(b)
Cohort
study—For
matched
studies,
give
matching
criteria
and
number
of
exposed
and
unexposed
Case-­‐control
study—For
matched
studies,
give
matching
criteria
and
the
number
of
controls
per
case
74

75. Materials
and
Methods
–
part2
Clearly
define
all
outcomes,
exposures,
predictors,
poten+al
Variables
confounders,
and
effect
modifiers.
Give
diagnos+c
criteria,
if
applicable
Data
sources/
For
each
variable
of
interest,
give
sources
of
data
and
details
of
methods
of
assessment
(measurement).
Describe
comparability
measurement of
assessment
methods
if
there
is
more
than
one
group
Describe
any
efforts
to
address
poten+al
sources
of
bias
Bias
Explain
how
the
study
size
was
arrived
at
Study
size
(a)
Describe
all
sta+s+cal
methods,
including
those
used
to
Sta+s+cal
control
for
confounding
methods (b)
Describe
any
methods
used
to
examine
subgroups
and
interac+ons
(c)
Explain
how
missing
data
were
addressed
(d)
Cohort
study—If
applicable,
explain
how
loss
to
follow-­‐up
was
addressed
Case-­‐control
study—If
applicable,
explain
how
matching
of
cases
and
controls
was
addressed
Cross-­‐sec6onal
study—If
applicable,
describe
analy+cal
methods
taking
account
of
sampling
strategy
(e)
Describe
any
sensi+vity
analyses
75

76. Study
Design
• Observa+onal
or
Experimental
• Retrospec+ve
or
Prospec+ve
76

77. Seing
and
Par+cipants
• Describe
the
study
popula+on
and
seing:
• Descrip+on
should
involve
relevant
demographic,
environmental,
diagnos+c,
comorbid
factors
• Defini+on
of
cohort/case
• Exclusion/inclusion
criteria
• How
was
consent
obtained?
• Matching
(in
case-­‐control
study)
77

78. Examples
of
seing
and
par+cipants
-­‐-­‐
cohort
Smoking
and
demen6a
in
male
Bri6sh
doctors:
prospec6ve
study
The
cohort
originally
comprised
34,439
male
doctors
on
the
Bri+sh
medical
register,
resident
in
the
United
Kingdom,
who
had
responded
to
a
ques+onnaire
about
their
smoking
habits
in
1951.
Changes
in
such
habits
were
sought
in
1957,
1966,
1972,
1978,
1990,
and
1998,
and
other
personal
informa+on
was
sought
in
1978,
1990,
and
1998.
In
1971,
follow
up
was
discon+nued
for
2459
subjects
(10.1%
of
the
survivors)
who
were
living
abroad
and
218
(0.9%)
for
other
reasons.
Almost
all
of
the
remaining
survivors
have
con+nued
to
provide
informa+on
about
their
smoking
habits*.
78

79. Examples
of
seing
and
par+cipants
–
case
control
Alcohol
and
stroke.
A
case-­‐control
study
of
drinking
habits
past
and
present
Cases
Three
hundred
sixty-­‐four
consecu+ve
pa+ents
hospitalized
for
acute
stroke
in
Newcastle
upon
Tyne
between
August
1989
and
July
1990
formed
the
study
popula+on.
No
pa+ent
refused
to
take
part
in
the
study.
Pa+ents
were
iden+fied
by
daily
contact
with
the
resident
medical
officer
and
completeness
of
case
ascertainment
was
checked
with
data
from
the
medical
records
department
at
each
of
the
three
par+cipa+ng
hospitals
(Freeman
Hospital,
Royal
Victoria
Infirmary,
and
Newcastle
General
Hospital)
Pa6ents
with
primary
subarachnoid
hemorrhage
were
excluded.
79

80. Examples
of
seing
and
par+cipants
–
case
control
(con+nued)
Controls
Three
hundred
sixty-­‐four
community
control
subjects
were
matched
for
age,
sex,
and
family
doctor.
Control
subjects
were
the
next
unrelated
matching
individual
to
the
case
in
the
family
doctor
register.
Control
subjects
with
a
previous
history
of
stroke
were
excluded.
80

81. Examples
of
seing
and
par+cipants
–
cross
sec+onal
Breast
feeding
and
obesity:
cross
sec6onal
study
The
1997
obligatory
health
examina+on
before
school
entry
evaluated
134,577
children
in
Bavaria,
southern
Germany.
At
the
examina+on,
the
parents
of
13,345
children
seen
in
two
rural
regions
were
asked
to
complete
a
ques+onnaire
about
risk
factors
for
atopic
diseases.
Data
collected
by
this
ques+onnaire
were
linked
with
the
data
from
the
school
health
examina+on.
Our
analysis
was
confined
to
children
aged
5
and
6
who
had
German
na+onality.
81

82. Examples
of
seing
and
par+cipants
–
cross
sec+onal
Supplementary
feeding
with
either
ready-­‐to-­‐use
for6fied
spread
or
corn-­‐soy
blend
in
wasted
adults
star6ng
an6retroviral
therapy
in
Malawi:
randomised,
inves6gator
blinded,
controlled
trial
The
study
took
place
at
the
an+retroviral
therapy
clinic
of
Queen
Elizabeth
Central
Hospital
in
Blantyre,
Malawi,
from
January
2006
to
April
2007.
Blantyre
is
the
major
commercial
city
of
Malawi,
with
a
popula+on
of
1,000,000
and
an
es+mated
HIV
prevalence
of
27%
in
adults
in
2004.Eligible
par+cipants
were
all
adults
aged
18
or
over
with
HIV
who
met
the
eligibility
criteria
for
an+retroviral
therapy
according
to
the
Malawian
na+onal
HIV
treatment
guidelines
(WHO
clinical
stage
III
or
IV
or
any
WHO
stage
with
a
CD4
count
<250/mm3)
and
who
were
star+ng
treatment
with
a
BMI
<18.5.
Exclusion
criteria
were
pregnancy
and
lacta6on
or
par6cipa6on
in
another
supplementary
feeding
program
82

83. Seing
and
par+cipants-­‐Surveillance
ONGOING
OUTBREAK
OF
WEST
NILE
VIRUS
INFECTION
IN
HUMANS,
GREECE,
JULY
TO
AUGUST
2011
Case-­‐Defini2on
•
A
confirmed
case
is
defined
as
a
person
mee+ng
any
of
the
following
clinical
criteria:
encephali+s,
meningi+s,
fever
without
specific
diagnosis
and
at
least
one
of
the
four
laboratory
criteria:
(i)
isola+on
of
WNV
from
blood
or
cerebrospinal
fluid
(CSF),
(ii)
detec+on
of
WNV
nucleic
acid
in
blood
or
CSF,
(iii)
WNV-­‐specific
an+body
response
(IgM)
in
CSF,
and
(iv)
WNV
IgM
high
+tre,
and
detec+on
of
WNV
IgG,
and
confirma+on
by
neutralisa+on.
83

84. Study
Variables
• Specify
unit
of
measurement
(if
applicable)
• Quan+fy
exposure
• Variable
transforma+ons
• Criteria
for
defini+ons
• Units
of
+me
and
special
categories
84

85. Study
Variables
(examples)
The
children's
height
and
weight
were
measured
as
part
of
the
rou+ne
examina+on.
Body
mass
index
was
calculated
as
weight
(kg)/(height
(m)2).
The
age
specific
and
sex
specific
distribu+on
of
the
body
mass
index
among
all
children
with
German
na+onality
in
Bavaria,
which
had
been
inves+gated
during
the
1997
school
health
examina+on,
was
used
as
the
reference
for
being
overweight
(defined
as
body
mass
index
above
the
90th
cen6le)
or
obese
(defined
as
body
mass
index
above
the
97th
cen6le)
because
these
cen+les
were
higher
than
other
European
reference
values.
85

86. Study
Variables
(examples)
Hypertension
was
iden6fied
by
medical
history
or
posi6ve
screening
results
(systolic
pressure
≥140
mm
Hg).
Pre-­‐hypertension
(asystolic
pressure
of
120–139
mm
Hg)
and
pre-­‐diabetes
(a
fas6ng
blood
glucose
concentra6on
of
6.1–6.9
mmol/L)
were
defined
on
the
basis
of
screened
laboratory
results.
Individuals
were
regarded
as
regular
alcohol
drinkers
if
they
consumed
two
or
more
alcoholic
drinks
a
day
on
three
or
more
days
a
week,
and
occasional
drinkers
if
they
consumed
less
than
regular
drinkers.
86

87. Study
Variables
(con+nued)
Data
from
clinic
visits
were
used
to
calculate
the
number
of
days
of
observa6on
per
quarter
for
each
pa+ent
in
each
of
four
categories
of
prescribed
an+retroviral
therapy.
These
categories,
in
increasing
order
of
intensity,
were
no
an+retroviral
therapy,
monotherapy,
combina+on
therapy
without
a
protease
inhibitor,
and
combina+on
therapy
that
included
a
protease
inhibitor.
The
data
collected
for
each
case,
using
a
standardised
form,
were:
demographic
characteris+cs
(age,
sex),
dates
of
admission
to
the
hospital
and
the
ICU,
the
+me
course
of
illness
including
the
date
of
symptom
onset,
underlying
condi+ons,
complica+ons,
use
of
mechanical
ven+la+on
support
(dates
of
intuba+on
and
extuba+on),
and
an+viral
treatment
87

88. Data
Sources/Management
• How
the
data
were
collected
• If
it
was
part
of
the
registry,
describe:
– Original
purpose
of
the
database
– How
large
the
database
is,
+meliness
– Valida+on,
quality
checks
– Error
rate
• Database
sogware/hardware
• For
surveillance
paper
–
a
diagram
of
the
surveillance
system
is
preferred
88

89. Data
Sources/Management
Pa+ents
(with
a
close
rela+ve
or
significant
other
when
possible)
were
interviewed
and
examined
by
H.R.
(79%)
or
P.D.A.
within
48
hours
of
hospitaliza+on.
Control
subjects
were
interviewed
in
their
homes
by
H.R.
(also
with
a
rela+ve
or
significant
other
when
possible).
Inter-­‐observer
valida+on
studies
between
the
two
interviewers
were
carried
out.
The
propor+on
of
agreement
between
two
observers,
K,
was
0.68.
89

90. Data
Sources/Management
Drinking
frequency
was
recorded
as
a
categorical
variable,
whereas
past
and
present
amounts
of
alcohol
consump+on,
dura+on
of
abs+nence,
and
heavy
drinking
were
recorded
as
con+nuous
variables.
Data
were
transferred
to
Northumbrian
University's
Mul6ple
Access
Computer
(NUMAC).
Following
verifica6on
procedures
to
ensure
accurate
transcrip6on,
data
were
analyzed
using
spss-­‐x
(SPSS-­‐X
Batch
System,
SPSS
Inc.,
Chicago,
Illinois).
90

91. Data
Sources/Management
• Informa6on
in
five
general
categories
has
been
abstracted
from
the
chart
for
each
outpa6ent
visit
and
entered
electronically
by
trained
data
abstracters;
the
data
are
compiled
centrally,
reviewed,
and
corrected
before
being
included
in
the
data
base.
Because
the
study
physicians
are
the
source
of
primary
care
for
these
pa+ents,
all
symptoms,
diagnoses,
and
treatments
since
the
previous
visit,
are
noted
at
each
clinic
visit.
The
categories
of
informa+on
are
as
follows:
demographic
characteris+cs;
symptoms;
diagnosed
diseases;
medica+ons
prescribed;
and
laboratory
values.
91

92. Data
Sources/Management
92

93. Study
Size
• Specify
the
null
hypothesis
and
whether
it
is
one
or
two-­‐sided
• Specify
the
minimum
difference
in
response
variable
that
is
considered
to
be
clinically
important
• Specify
power
and
alpha
level
for
calcula+ng
sample
size
93

94. Examples
To
detect
a
reduc+on
in
PHS
(postopera+ve
hospital
stay)
of
3
days
(SD
5
days),
which
is
in
agreement
with
the
study
of
Lobo
et
al.
with
a
two-­‐sided
5%
significance
level
and
a
power
of
80%,
a
sample
size
of
50
pa+ents
per
group
was
necessary,
given
an
an+cipated
dropout
rate
of
10%.
To
recruit
this
number
of
pa+ents,
a
12-­‐
month
inclusion
period
was
an+cipated
94

95. Examples
Based
on
an
expected
incidence
of
the
primary
composite
endpoint
of
11%
at
2.25
years
in
the
placebo
group,
we
calculated
that
we
would
need
950
primary
endpoint
events
and
a
sample
size
of
9650
pa+ents
to
give
90%
power
to
detect
a
significant
difference
between
ivabradine
and
placebo,
corresponding
to
a
19%
reduc;on
of
rela;ve
risk
(with
a
two-­‐sided
type
1
error
of
5%)
95

96. Randomiza+on
–
Randomized
controlled
trials
(RCT)
Par+cipants
should
be
assigned
to
comparison
groups
in
the
trial
on
the
basis
of
a
chance
(random)
process
characterized
by
unpredictability
96

97.
Randomized
controlled
trials
(RCT)
-­‐-­‐
examples
• Independent
pharmacists
dispensed
either
ac+ve
or
placebo
inhalers
according
to
a
computer
generated
randomiza+on
list
• For
alloca+on
of
the
par+cipants,
a
computer-­‐generated
list
of
random
numbers
was
used
97

98. Randomiza+on
(con+nued)
• Randomiza+on
sequence
was
created
using
Stata
9.0
(StataCorp,
College
Sta+on,
TX)
sta+s+cal
sogware
and
was
stra+fied
by
center
with
a
1:1
alloca+on
using
random
block
sizes
of
2,
4,
and
6
• Par+cipants
were
randomly
assigned
following
simple
randomiza+on
procedures
(computerized
random
numbers)
to
1
of
2
treatment
groups
98

99. Randomiza+on
-­‐-­‐
Concealment
A
generated
alloca+on
schedule
should
be
implemented
by
using
alloca+on
concealment,
a
c r i + c a l
m e c h a n i s m
t h a t
p r e v e n t s
foreknowledge
of
treatment
assignment
and
thus
shields
those
who
enroll
par+cipants
from
being
influenced
by
this
knowledge.
The
decision
to
accept
or
reject
a
par+cipant
should
be
made,
and
informed
consent
should
be
obtained
from
the
par+cipant,
in
ignorance
of
the
next
assignment
in
the
sequence
99

100. Randomiza+on
(concealment)
The
doxycycline
and
placebo
were
in
capsule
form
and
iden+cal
in
appearance.
They
were
prepackaged
in
bosles
and
consecu+vely
numbered
for
each
woman
according
to
the
randomiza+on
schedule.
Each
woman
was
assigned
an
order
number
and
received
the
capsules
in
the
corresponding
pre-­‐packed
bosle
100

101. Blinding
(RCTs)
The
term
“blinding”
or
“masking”
refers
to
withholding
informa+on
about
the
assigned
interven+ons
from
people
involved
in
the
trial
who
may
poten+ally
be
influenced
by
this
knowledge.
Blinding
is
an
important
safeguard
against
bias,
par+cularly
when
assessing
subjec+ve
outcomes.
EXAMPLE:
Whereas
pa+ents
and
physicians
allocated
to
the
interven+on
group
were
aware
of
the
allocated
arm,
outcome
assessors
and
data
analysts
were
kept
blinded
to
the
alloca+on.
101

102. Laboratory
Methods(Surveillance)
Serum
and
CSF
specimens
were
tested
for
the
presence
of
WNV-­‐specific
IgM
and
IgG
an+bodies
using
commercial
ELISA
kits
(WNV
IgM
capture
DxSelect
and
WNV
IgG
DxSelect,
Focus
Diagnos+cs
Inc,
Cypress,
CA,
USA).
WNV
posi+ve
specimens
were
also
tested
for
the
presence
of
other
flaviviruses:
+ck-­‐borne
encephali+s
virus
(TBEV)
and
dengue
virus
(DENV).
102

103. Sta+s+cal
Methods
• Describe
all
sta+s+cal
methods,
including
those
used
to
control
for
confounding
• Describe
the
comparisons
to
be
made
and
the
sta+s+cal
procedures
to
be
used
for
making
them
• State
whether
the
sta+s+cal
analysis
will
be
on
the
basis
of
inten+on-­‐to-­‐treat
• Control
for
mul+ple
tes+ng
problem
• Report
hypothesis
power
and
level
(if
it
is
not
reported
in
sampling
sec+on)
• Report
all
required
p-­‐values
and
confidence
intervals
103

104. Assessment
of
risk
ra+on
Sick
Not
sick
Cases
Controls
No
history
of
disease
History
of
disease
Exposed
Not
exposed
A
В
A
В
С
D
С
D
In
case
control
study
the
risk
ra+on
has
no
outcome,
odds
ra+on
used
instead

105. Repor+ng
sta+s+cal
methods
in
Cross-­‐Sec+onal
studies
• Standard
descrip+ve
sta+s+cs:
-­‐Simple
prevalence
calcula+on
• Prevalence
of
disease
or
prevalence
of
exposure
• Regression
to
control
confounders
105

106. Cross-­‐sec+onal
study
example:
Sta+s+cal
Methods
Pa+ent
characteris+cs,
adjusted
for
stone
history
and
age,
were
compared
using
linear
regression
for
con+nuous
covariates
and
logis6c
regression
for
categorical
covariates.
Mul6ple
linear
regression
was
used
to
compare
mean
es+mated
GFR
between
stone
formers
and
non-­‐stone
formers.
Covariates
iden+fied
as
poten+al
confounders
in
the
rela+onship
between
es+mated
GFR
and
stone
history
were
adjusted
for.
Mul6plica6ve
interac6ons
between
stone
history
and
age,
gender,
race,
diabetes,
and
BMI
were
formally
tested.
106

107. Cross-­‐sec+onal
study
example:
Sta+s+cal
Methods
Mul6nomial
logis6c
regression
was
used
to
compare
the
rela+ve
risk
of
having
an
es+mated
GFR
in
a
lower
category
rela+ve
to
the
highest
category
between
persons
with
and
without
nephrolithiasis.
Model
based
es+mates
are
reported
as
rela6ve
risk
ra6os
comparing
stone
formers
with
non-­‐stone
formers.
Adjustment
covariates
included
in
the
mul+nomial
logis+c
regression
included
age,
gender,
race,
BMI,
systolic
blood
pressure,
HbA1c,
diabetes,
history
of
cardiovascular
disease,
smoking
status,
health
insurance
status,
and
use
of
prescrip+on
diure+cs.
107