2. Observational study
A study in which no intervention is made. Such
studies provide estimates and examine
associations of events in their natural settings
3. Retrospective vs. Prospective
• Retrospective = measurement obtained after
outcome occurred
• Prospective = measurement obtained before
outcome occurred
4. What is an observational study?
There are three main types of observational
studies:
– Cohort (prospective and retrospective),
– Case control
– Cross-sectional
5.
6. Which Design to choose?
• Depending on the study question
Object Common Design
Prevalence Cross Sectional
Incidence Cohort
Cause (in order of
reliability)
Cohort, case-control, cross
sectional
Prognosis Cohort
Treatment effect Controlled Trial
7. Study Design
• Research Question
– P patient
– I intervention
– C comparison group
– O outcome
– T time
• Literature Search
9. When should observational studies be
performed?
– Clinical trials difficult in certain areas, e.g Surgery or
Anesthesiology
– Exploratory (generate hypothesis)
– Exposure cannot be manipulated by the investigator
– Interventions for very rare diseases or rare adverse effect
– Studies about accuracy of diagnostic tests
13. Measures Disease Frequency
Cross-Sectional Study
• Prevalence: The number of cases in a population at a
given point in time.
• Notice any proportion can be estimated in the 2X2
table
Not ExpExp
Disease
No Disease
a
d
E
b
nE
c
N
nD
D
14. Cross Sectional Study
• Cross sectional studies are also used to infer
causation.
• At one point in time the subjects are assessed
to determine whether they were exposed to the
relevant agent and whether they have the
outcome of interest. Some of the subjects will
not have been exposed nor have the outcome
of interest
15. Advantages
• Subjects are neither deliberately exposed,
treated, or not treated and hence there are
seldom ethical difficulties.
• Only one group is used, data are collected
only once and multiple outcomes can be
studied
• There is no loss to follow up
16. Advantages
• Quick and cheap- less resources are required
to run the study
• Cross sectional studies are the best way to
determine prevalence and are useful at
identifying associations that can then be more
rigorously studied using a cohort study or
randomized controlled study
17. Disadvantages
• The most important problem with this type of
study is differentiating cause and effect from
simple association. For example, a study finding an association
between low CD4 counts and HIV infection does not demonstrate whether
HIV infection lowers CD4 levels or low CD4 levels predispose to HIV
infection.
• Only a snap shot (At a specific time)
18. Disadvantages
• Often there are a number of plausible
explanations. For example, if a study shows a negative
relation between height and age it could be concluded that
people lose height as they get older, younger generations are
getting taller, or that tall people have a reduced life
expectancy when compared with short people.
• Cross sectional studies do not provide an
explanation for their findings
19. Disadvantages
• Rare conditions cannot efficiently be studied
using cross sectional studies because even in
large samples there may be no one with the
disease. In this situation it is better to study a
cross sectional sample of patients who already
have the disease (a case series).
20. Population
Risk factor
Disease present
Risk factor
No disease present
No risk factor
Disease Present
No risk factor
No disease
Selected sample
Cross sectional study
Summary
21. Key Points
Cross Sectional Studies
- The best way to determine prevalence.
- Are relatively quick & cheap
- Can study multiple outcomes.
- Do not themselves differentiate between cause and effect or
the sequence of events.
- Not for studying rare conditions
- The principal summary statistic is the odds ratio
Cross sectional study
Summary
22. Cohort studies
A group of individuals who share a common
characteristic and who are then followed up at
specified intervals to determine outcomes.
23. • Cohort studies look forwards in time by following up each
subject
• Subjects are selected before the outcome of interest is
observed
• They establish the sequence of events
• Numerous outcomes can be studied
• They are the best way to establish the incidence of a disease
• They are a good way to determine causes of diseases
• The principal summary statistic of cohort studies is the
relative risk ratio
Cohort Study -Summary
25. Cohort Study
= sampled = observed
Not ExpExp
Disease
No Disease
a
d
E
b
nE
c
N
nD
D
Note that E and nE are fix by design.
Relative risk (RR)= a/(a+b)
C/(C+D)
26. A prospective cohort study
• A group of people is chosen who do not have the
outcome of interest (e.g MI)
• The investigator then measures a variety of variables
that might be relevant to the development of the
condition.
• Over a period of time the people in the sample are
observed to see whether they develop the outcome of
interest (that is, myocardial infarction).
• One group has been exposed to or treated with the
agent of interest and the other has not, thereby acting
as an external control.
27. Cohort Study - Prospective
= sampled = observed
EXPOSED
UNEXPOSED
Present Future
D nD
D nD
28. A retrospective cohort study
• If the data are readily available then a
retrospective design is the quickest method.
• These use data already collected for other
purposes.
• The methodology is the same as prospective
cohort
29. Cohort Study – Retrospective (Historical)
= sampled = observed
EXPOSED
UNEXPOSED
Past of Past Past
D nD
D nD
30. Advantages
• A randomised controlled trial may be unethical; for e.g
cigarette smoke or asbestos. Thus research on risk factors
relies heavily on cohort studies
• The study can demonstrate that these “causes” preceded
the outcome, thereby avoiding the debate as to which is
cause and which is effect.
• Allow to calculate the incidence of disease
• A single study can examine various outcome variables e.g
smokers can simultaneously look at deaths from lung,
cardiovascular, and cerebrovascular disease
• Calculation of the effect of each variable on the probability
of developing the outcome of interest (relative risk).
31. Disadvantges
• Expensive and time consuming
• Inappropriate for rare diseases – disease with
a long latency
• Loss of follow up: incidence = cases/per period
of time), it can be seen that the loss of a few
cases will seriously affect the numerator
• Presence of confounders
32. Confounding Variable
• A confounding variable is independently associated
with both the variable of interest and the outcome of
interest. For example, lung cancer (outcome) is less
common in people with asthma (variable). However, it
is unlikely that asthma in itself confers any protection
against lung cancer.
• The only way to eliminate all possibility of a
confounding variable is via a prospective randomized
controlled study. In this type of study each type of
exposure is assigned by chance and so confounding
variables should be present in equal numbers in both
groups.
33. Retrospective Vs. Prospective Cohort
studies
• Retrospective studies are much cheaper as the
data have already been collected.
• Unlikely that all the relevant information will
have been rigorously collected.
34. Population Selected Sample
Variable Present
Conditions
Developed
Conditions do not
develop (internal
controls)
Variable Absent
External Controls
Conditions
Developed
Condition does
not develop
Cohort Study -Summary
35. Key Points
Cohort Studies
- Cohort studies describe incidence or natural history.
- They analyze predictors (risk factors) thereby enabling
calculation of relative risk.
- Cohort studies measure events in temporal sequence
thereby distinguishing causes from effects.
- Retrospective cohorts where available are cheaper and
quicker.
- Confounding variable are the major problem in analyzing
cohort studies.
- Subject selection and loss to follow up is major potential
cause of bias.
Cohort Study -Summary
37. Case-control studies
• Retrospective
• A population with a particular outcome is compared with a
matched population without the outcome. Researchers can
then look at variables “retrospectively” in each group to
see if there are any factors that are associated with having
or not having the outcome
• Where the outcome is rare, case-control studies may be the
only feasible approach. As some of the subjects have been
deliberately chosen because they have the disease
• Only one outcome is studied
• Case-control studies can however be used to calculate odds
ratios
38. Case-control Study
= sampled = observed
Not ExpExp
Disease
No Disease
a
d
E
b
nE
c
N
nD
D
One may estimate:
OR
Odds (Exp|D)
Odds (nE|D)
=
ad
bc
=
Odds (Exp|D) = P (Exp|D)
1−P (nE|D)
Odds (D|Exp)
Odds (D|nExp)
=
39. Advantages
• When conditions are rare
• When there is a long latent period between
an exposure and the disease,
• The only feasible option e.g of new variant
CJD and possible aetiologies
• Requires few study subjects
• Less funds – Quick results
40. Disadvantages
• Can not generate incidence data
• Difficult if records are inadequate
• Selection bias
• Recall bias
41. Disadvantages
Sampling bias:
• The patients with the disease may be a biased
sample
• Observation and recall bias : As the study
assesses predictor variables retrospectively
there is great potential for a biased
assessment of their presence and significance
by the patient or the investigator, or both.
43. Key Points
Case Control Studies
- Case control studies are simple to organize
- Retrospectively compare two groups
- Aim to identify predictors of an outcome.
- Permit assessment of the influence of predictors on
outcome via calculation of an odds ratio
- Useful for hypothesis generation
- Can only look at one outcome
- Bias is a major problem
45. • Why do a randomized clinical trial?
• generally reserved for mature questions
• considered as "gold standard" for demonstrating
efficacy
• incorporates design features to maximize outcome
ascertainment and minimize bias
• minimize risk of confounding
• Why not do a randomized clinical trial?
• major ethical or feasibility issues (e.g. vaccine, toxic
agents)
• cost and logistics concerns
• unrealistic time horizon
47. Randomized Controlled Clinical Trials
• Phases of a clinical trial
• Equipoise
• Randomization
• Blinding
• Intention to treat analysis
48. Definition
A prospective planned experiment designed to
assess the effect of a treatment by comparing
the observed outcomes in a group exposed to
an intervention with those in a comparable
group receiving a control or placebo treatment.
e.g. drug vs. placebo, drug vs. standard of care,
medicine vs. surgery; low dose vs. high dose,
etc. The observed differences are attributed to
the intervention. The outcome must be clearly
defined prior to commencing the study.
49. The Principle of Equipoise
Equipoise indicates genuine, community-based
uncertainty. A consensus regarding the uncertainty
of the risk/benefit of an investigational agent.
Needed to justify carrying out an RCT (i.e.
randomizing treatment)
50. Choice of Participants
INCLUSION CRITERIA
appropriate to the study
question
reflect the potential
benefit the intervention
generalizable
ease of recruitment
EXCLUSION CRITERIA
safety concerns
intervention likely
ineffective
conditions that could
compete with the
outcomes (e.g. early
death)
contradictions to
intervention of interest
adherence issues
51. Typically, and inert agent (i.e. placebo) is ideal but
might not always possible (e.g. "sham" surgery,
side effects).
May want to consider and active therapy as a
control (e.g. standard of care).
Note: Placebo is not ethical in virtually all studies
that involve diseases with a proven/established
treatment.
Remains ethical in trials where no proven
treatment is established
Choice of Control
52. • Easy to diagnose, collect or observe
• Clinically relevant
• Chosen before the start of data collection
• Outcomes:
– Primary
– Secondary
– Surrogate
– Composite
Choice of Outcomes
54. Randomization
The process of assigning participants to
treatment and control
– Assuming that each participant has an equal chance of being
assigned to any group.
– Assignment of participants to treatment by a random process such that
neither the investigator nor the participant knows the treatment to be
assigned at the start of the study.
– Study participants are assigned to treatment
groups by chance (not by choice).
• Random number generator (table or computer)
• Centralized and secure system
55. Why is it important?
In a clinical trial, you need comparability among study groups and
the best way to assure comparability is by randomization.
Features:
- Balances patient baseline characteristics to all arms of the study.
- Equal distribution of all known and unknown confounding
variables i.e. minimizes the possibility that the observed association
between the exposure and the outcome is in reality caused by
another factor.
- Allows for an unbiased comparison between groups and forms the
basis for the derivation of statistical tests
56. Refers to masking treatment assignment
from patients, investigators and/ or
end point assessors.
• Single blind - participants are not aware
of treatment assignment
• Double blind - both participants and
investigators unaware
• Triple blind - various meanings
– - persons who perform tests
– - outcome adjudicators
– - safety monitoring group
I C
Blinding (Masking)
57. If an outcome can conceivably be affected by patient or
investigator expectations, then maintaining blinding is important.
Maintains balanced groups during follow-up and minimizes
biased outcome ascertainment (participant) and measurement
(investigator).
Can be assessed:
- ask participants to guess treatment
- ask study staff to guess treatment
Why Blinding?
58. Key Features: In the attempt to reduce bias and to improve the quality of evidence:
- Prospective
- Comparison (control)
- Pre specified outcomes and analysis
- Randomization
- Blinding
Notes:
- treatment is assigned by randomization into an experimental and control (comparison) group.
- course is tracked prospectively and differences in outcomes are compared.
- differences are attributed to the intervention.
OutcomeInterventionPopulation
timesample
treatment group
control group
randomization
Basic ("parallel") Design
blinding
60. Factorial Design
sample
treatment A +
placebo B
randomization
OutcomeIntervention
Population
treatment A +
treatment B
placebo A +
treatment B
placebo A +
placebo B
A B
C D
2 X 2 Design
(simplest)
treatment 1
treatment2
controlactive
controlactive
Advantages
- can examine main effects
and interaction effects
Disadvantage
- may be difficult to
disentangle the
interaction effect and
interpret results
61. Analysis of RCTs
• Intention to treat analysis: Once randomized
always analyzed
• Everyone on treatment (by assignment or
choice) versus everyone not on treatment
• Per protocol analysis: analyze only subjects
that complied to the assigned treatment
61
62. # Subs. Length Purpose Success
Phase I 20 – 100 Several months Mainly Safety 70%
Phase II Up to
several 100
Several
months- 2 yrs.
Short term
safety; mainly
effectiveness
33%
Phase III 100s –
several 1000
1-4 yrs. Safety, dosage &
effectiveness
25-30%
Summary of Phases
63. RCTs vs. Observational studies
RCTs Observational Studies
Selection bias If randomization is done properly,
negligible
Likely to occur
Measurement bias If blinding is done properly in double
blinded studies, less likely
Possible especially with subjective
outcomes
Statistical Analysis Simpler – usually unadjusted simple
analysis is sufficient (t-tests or ANOVAs)
More complicated – adjusted
analysis is necessary with logistic or
linear regression or cox proportional
hazard models (survival)
Unmeasured confounders Usually balanced if sample size is
appropriate
Usually imbalanced
External validity Usually limited as restrictions of patients
who can participate in the trial
The real-world scenario makes this
design more externally valid
Ethical issues Placebo use can be an ethical issue Minimal ethical issues usually
Costs High Usually lower
Power Might be under powered due high costs Usually high power (large cohorts)
Follow-up Limited Usually long
from Merit Cudkowicz, MD
64. Object Common Design
Prevalence Cross Sectional
Incidence Cohort
Cause (in order of reliability) Cohort, case-control, cross
sectional
Prognosis Cohort
Treatment effect Controlled Trial