1/34

Statistics

Types of Studies

Shaheena Bashir

FALL, 2019
2/34
Outline

Study Designs

Observational Studies
Surveys
Cross-Sectional
Cohort
Case-Control Studies

Experimental Studies
Randomized Controlled Trials(RCT)

o
3/34
Study Designs

o
 4/34
Study Designs

Background
I Designing a research study is not a simple task
I The success of a study is determined by appropriate planning
of the study before initiating the study
I Knowing what kind of information the study should collect is
a first step in determining how the study will be carried out
(also known as the study design).
I Successful studies need to be reliable, i.e., should provide the
same results if the same study parameters are applied
I Successful studies need to be valid, i.e., should correctly
answer the question it asks
I Basic Question: ’Are treatment (exposure) & outcome
(disease) linked’ ?
o
5/34
Study Designs

o
6/34
Study Designs

o
7/34
Observational Studies

I Well-designed observational studies have been shown to

provide results similar to randomized controlled trials,
challenging the belief that observational studies are
second-rate.
I The investigator does not intervene and rather simply
’observes’ and assesses the strength of the relationship
between an exposure and disease variable called association.
I Observational studies can be misleading about
cause-and-effect relationship due to confounding (another
variable is indirectly influencing the relationship between
cause-and-effect, e.g., age in the study of risk of CHD due to
smoking).

o
8/34
Observational Studies
Surveys

o
9/34
Observational Studies
Surveys

I The most common observational study is the survey.

I Surveys are questionnaires that are presented to individuals
who have been selected from a population of interest
I The survey results provide a ’snapshot’ of a population.
I Surveys are a useful tool for gauging the health of a
population or to monitor effectiveness of a preventative
intervention or provision of emergency relief.
I A survey is considered a cross-sectional study.

o
 10/34
Observational Studies
Surveys

Examples

I Population-based survey studies are conducted by NCHS and

other agencies
I Collect information on health and other characteristics (such
as nutrition) of U.S. household population
I For example:
I National Health Interview Survey
I National Health and Nutrition Examination Survey
I China’s National Prevalence Survey
I Pakistan Demographic and Health Survey (2017-18 PDHS)
was implemented by the National. Institute of Population
Studies (NIPS)

o
11/34
Observational Studies
Cross-Sectional

o
 12/34
Observational Studies
Cross-Sectional

Cross-Sectional Studies

I An observational study in which exposure and disease are

determined at the same point in time in a given population
I Provides a snapshot of variables in the study, at one particular
point in time.
I Reveals how those variables are represented in a cross-section
of the population.
I However, the temporal relationship between exposure and
disease cannot be determined

o
 13/34
Observational Studies
Cross-Sectional

Design

I Begin with defined population.

I Gather data on exposure and disease

Disease No Disease Totals Incidence Rates

a
Exposed a b a+b a+b
c
Non-Exposed c d c +d c+d

Prevalence of disease in Exposed compared to Not Exposed:

a c
(a+b) vs (c+d)

o
 14/34
Observational Studies
Cross-Sectional

Example: Smoking & CVD

CVD No CVD Totals Incidence Rates

84
Smokers 84 2916 3000 3000 = 0.028
87
Non-Smokers 87 4913 5000 5000 = 0.0174

2.8% of smokers have CVD as compared to 1.74% of Non-smokers

who are with CVD

o
 15/34
Observational Studies
Cross-Sectional

Examples

I The comparison of school systems which require the

bachelor’s degree of teachers versus those which do not;
I The comparison of students in classes given speed-reading
training versus those not given it;
I The comparison of those who heard a certain TV program
with those who did not, etc

o
16/34
Observational Studies
Cohort

o
 17/34
Observational Studies
Cohort

Cohort Studies

I The term ‘cohort’ is derived from the Latin word cohors

I Set of people followed over a period of time
I Cohort studies can be retrospective or prospective

o
 18/34
Observational Studies
Cohort

Types of Cohort Studies

I Retrospective Cohort Studies also known as historical

cohort studies, are carried out at the present time and look to
the past to examine medical events or outcomes.
I The primary disadvantage of this study design is the limited
control the investigator has over data collection. The existing
data may be incomplete, inaccurate, or inconsistently
measured between subjects.
I However, because of the immediate availability of the data,
this study design is comparatively less costly and shorter than
prospective cohort studies.

o
 19/34
Observational Studies
Cohort

Types of Cohort Studies

I Prospective Studies are set up to collect the data from a

predetermined sample from the present time into the future.
I Because prospective studies are designed with specific data
collection methods, it has the advantage of being tailored to
collect specific exposure data and may be more complete.
I The disadvantage of a prospective cohort study may be the
long follow-up period while waiting for events or diseases to
occur.
I This study design is inefficient for investigating diseases with
long latency periods and is vulnerable to a high loss to
follow-up rate.

o
 20/34
Observational Studies
Cohort

Relative Risk
A measure of strength of relationship based on prospective cohort
studies. First select exposed & Not exposed groups & then follow
to see whether they develop disease

Then follow to see whether

CVD develops No CVD Totals Incidence Rates
a
Smokers a b a+b a+b
c
Non-Smokers c d c +d c+d

Incidence in Smokers a/(a + b)

Relative Risk (RR) = =
Incidence in Non-Smokers c/(c + d)

o
 21/34
Observational Studies
Cohort

Interpretation of Relative Risk

I If RR = 1, then no association
I If RR > 1, risk in exposed is greater than risk in non-exposed
(positive association, possibly causal)
I If RR < 1, risk in exposed is less than risk in non-exposed
(negative association, possibly protective)

o
 22/34
Observational Studies
Cohort

Cohort Study: Relative Risk Example

Then follow to see whether

CVD develops No CVD Totals Incidence Rates
84
Smokers 84 2916 3000 3000 = 0.028
87
Non-Smokers 87 4913 5000 5000 = 0.0174

Incidence in Smokers 84/3000

Relative Risk (RR) = = = 1.61
Incidence in Non-Smokers 87/5000
Interpretation: Smokers are 61% more likely to develop CHD than
non-smokers (+ve association, possibly causal)

o
23/34
Observational Studies
Case-Control Studies

o
 24/34
Observational Studies
Case-Control Studies

Case-Control Studies

I A case-control study is usually conducted before a cohort or

an experimental study to identify the possible etiology of the
disease. It costs relatively less and can be conducted in a
shorter time
I A case-control study is preferred when the disease is rare
because investigators can intentionally search for the cases
I First select cases (with disease) & controls (without disease).
I Then measure the past exposure

o
 25/34
Observational Studies
Case-Control Studies

Case-Control Studies: Association

I Can’t derive incidence from case-control study. As it begins

with cases (with disease) & controls (without disease). Then
measure the past exposure
I Therefore can’t calculate incidence directly
I But can use another method called odds ratio.

o
 26/34
Observational Studies
Case-Control Studies

Case-Control Study: Odds Ratio

Past exposure cases (with disease) controls (without disease)

were exposed a b
were not exposed c d
Totals a+c b+d

Odds that an exposed person develops disease

Odds Ratio (OR) =
Odds that a non-exposed person develops disease
a/b
=
c/d
ad
=
bc
o
 27/34
Observational Studies
Case-Control Studies

Case-Control Study: Odds Ratio Example

Past exposure CHD cases controls (without disease)

Smokers 112 176
Non-smokers 88 224
Totals 200 400

ad
(OR) =
bc
112 × 224
=
88 × 176
= 1.62

Interpretation: Smokers are 62% more likely to develop CHD than

non-smokers (+ve association, possibly causal) o
 28/34
Observational Studies
Case-Control Studies

When is Odds Ratio a good estimate of Relative Risk

I When the disease being studies occurs at low frequency

I When cases are representative of the diseased population
I When controls are representative of the population without
disease

o
 29/34
Observational Studies
Case-Control Studies

Longitudinal Studies

I A longitudinal study, like a cross-sectional one, is

observational.
I Researchers do not interfere with their subjects
I Researchers conduct several observations of the same subjects
at regular intervals over a period of time, sometimes lasting
many years.
I Reveals changes at the individual level, e.g., teacher turnover,
nurse health indictors.

o
 30/34
Observational Studies
Case-Control Studies

Message

o
31/34
Experimental Studies

I A study that allows the investigator to control the

intervention.
I The investigator controls the exposure and determines who
gets the exposure and how much exposure one gets
I The investigator decides the dosage and allocation of
treatment to participants through randomization process.
I Such studies involve planned research designs pertinent to
causal hypotheses being tested

o
32/34
Experimental Studies
Randomized Controlled Trials(RCT)

I Gold standard of research designs

I Provides most convincing evidence of causal relationship
between intervention and the effect
I Involve randomized study designs
I The importance of randomization is that in the long run
treatment groups will be balanced in known and unknown
prognostic factors.

o
33/34
Experimental Studies
Randomized Controlled Trials(RCT)

o
 34/34
Experimental Studies
Randomized Controlled Trials(RCT)

Examples

I The comparison of AZT versus no treatment on the length of

survival in patients with AIDS
I Evaluating the effectiveness of a new anti-fungal medication
on Athletes foot
I Evaluating hormonal therapy on the reduction of breast cancer
(Womens’ Health Initiative)