1

Analytical studies

By :Dr. Meluda Rajab AlHa madi

Fa mily & Community medicine department

MR-2015 ‫صفحة‬1
 2

Epidemiology:
• It is the basic science of community medicine .
• It is defined as : the study of the distribution and determinants of health-related
states in specified populations, and the application of this study to control of
health problems."
• It Provides insight regarding the nature, causes, and extent of health and disease.
• It Provides information needed to plan and target resources appropriately.

Kinds of epidemiology:
1- Descriptive : Study of the occurrence and distribution of disease, Answers the four
major questions: what, who, where, and when?
2- Analytical : Study of the determinants or underlying cause of disease., Answers
two other major questions: how? and why?
3- Experimental : To confirm an etiological hypothesis

Analytical epidemiology:

Analytic studies are the second major type of epidemiological studies. They differ from
descriptive ones that: -

1- Individuals are the subjects of interest not the whole population. The inference
is to be applied on the population from which the subjects are selected.

2- The aim is to test the hypothesis and not to formulate it.

Analytic studies are two types :

1-Case control studies (retrospective).
2-Cohort studies (prospective).

From these studies we can determine: -

1) Whether or not a statistical association exists between a disease and a
suspected factor
2) If the association exists, what is the strength of it.

MR-2015 ‫صفحة‬2
 3

1-Case control study:

Case control study has 3 distinct features:
1- Both exposure & outcome have occurred before the start of the study.
2- Study proceeds backwards from effect to cause.
3- Use of a control group to support or refute an inference.

Basic steps of case-control study: -

1) Selection of cases
2) Selection of controls
3) Matching
4) Measurement of exposure
5) Analysis and interpretation

MR-2015 ‫صفحة‬3
 4

The basic design of case-control study is 2X2 contingency table:

Suspected risk Cases Control
factor Disease present Disease absent
Present a b

c d
Absent
Total a+c b+d

1- Selection of cases

-Definition of a case: is a crucial step that involves two specifications;

• The diagnostic criteria: criteria of the disease and its stage.
• The eligibility criteria : including & excluding criteria e.g., only new
diagnosed cases (incident) within specific period of time. Old and advanced
cases should be excluded.
• A Case definition usually includes four components:
1- Clinical criteria ( usually include confirmatory laboratory tests, if
available, or combination of symptoms (subjective complaints),
signs(objective physical finding), and other findings
2- Characteristics of person
3- Place
4- Time ( incident, prevalent)
-Sources of cases:

1. Hospitals: it is very convenient to select cases fro m hospitals, one or group

of hospitals. Selection is done on sampling methods.

2. General population: All cases occurring in defined geographic area during

specific period of time. Data is collected by a survey, disease registry or
hospital network.
MR-2015 ‫صفحة‬4
 5

2-Selection of controls:
Controls (comparison group) must be similar to cases in all aspects except the
disease in question under study, they must be free from it.

-Sources of controls:
1- Hospitals controls:
patients can be selected as controls having different diseases but not the disease
under study. Many hospital patients may have diseases that are also influenced by
the factor under study, resulting in what called selection bias. For exa mple if we
want to study the relationship between smoking and myocardial infarction. We
should not choose bladder cancer as controls.

2- Relatives: as spouses and siblings. Siblings are not suitable for genetic
diseases.
3- Neighborhood controls: these are taken from persons living in the same
locality, workers in the same factory, or from children in the same school.

4- General population controls: taking a random sample from people free from
the disease, living in a defined geographic area.

The number of controls is one-to-one if we have large number of cases. If the

cases are small (say under 50). it is better to chose 2, 3 or even 4 controls for
every case under study.

3-Matching:
Matching is the process by which the selected controls are similar to cases with
regards to certain variables such as age, sex, and social class, which are known to
influence the outcome of the disease and distort the results.

Matching is done as:

 Pairs matching: individual matching for each case a control is chosen and
matched quite closely.
 Group matching: frequency matching in which cases are assigned to sub-
groups based on age, occupation, and social class and then select control
group with a certain characteristic identical to the cases in the same
proportion as it appeared in cases

MR-2015 ‫صفحة‬5
 6

Confounding factor:

It is a factor associated both with exposure and disease, and is distributed

unequally in study and control group.
( associated with exposure & it independent risk factor )
examples:
Breast cancer& OCP + age
alcohol &oesophageal cancer + smoking

4- Measurement of exposure

Data about exposure should be obtained in precisely the same way for cases and
controls. Data can be collected by interview, by questionnaires or by studying past
records.

MR-2015 ‫صفحة‬6
 7

5- Analysis:

Two thing to be find out: exposure rates and estimation of disease risk.
1- Exposure rates:
Estimation of exposure rates to suspected factor in disease and non-disease groups
can be calculated directly from case-control study contingency table.

A case-control study of stroke & drug abuse

Suspected risk factor Cas es Control

stroke Disease absent

Drug abuse 73 18

No drug abuse 141 196

Total 214 214

Exposure rates:
a) Exposure rate among Cases : 73 / 214 = 0.34 (34%)
b) Exposure rate among control Controls: 18 / 214 = 0.08 (8%)

It is clear from the calculation that the frequency rate of stroke is higher among
drug abuser than among non drug abuser.

2- Estimation of risk:
In case-control study we estimate the risk by calculating the Odds Ratio (cross-
product ratio). The Odds ratio measures the strength of association between risk
factor and outcome.

The derivation of odds ratio is based on three assumptions: -

 Disease must be relatively rare.
 Cases must be representative of those with the disease.
 Controls must be representative of the healthy.

MR-2015 ‫صفحة‬7
 8

Odds ratio = a x d / b x c
From the above example odds ratio = (73x196) / (18x141) = 5.64
This indicates that those who have abused drugs were 5.64 times (say 6 times)
more likely to develop stroke

Interpretation of Odds Ratio Value:

1) Odds ratio equal zero = complete protection

2) Odds ratio more than zero and less than one = Here the exposure is protective
with decreasing strength towards the value one.

3) Odds ratio equal to one = there is no risk at all fro m exposure

4) Odds ratio more than one and less than infinity. = Here the exposure is risky
causing disease

5) Odds ratio equals infinity (theoretical) = 100% risk

Bias in case control studies:

1- Bias due to confounding: it is removed by proper matching.

2- Memory or recall bias: when cases and control are asked questions about
their past, it may be more likely for the cases to recall the existence of certain
event than the control who are healthy . so it relies on memory or past record ,
the accuracy of which may be uncertain and the validation of information
obtained is difficult or impossible.

3- Selection bias: the cases and controls may not be representative of cases and
control in the general population ,it avoided by proper selection of study group
and randomization.

4- Berkesonian bias: the bias arises because of different admission rates to the
hospitals.

5- Interviewer bias: it occur when the investigator knows the hypothesis and
knows who cases are. This leads to question of the cases more than control about
the past history and risk factors . this bias eliminated by double blind trial.

MR-2015 ‫صفحة‬8
 9

Advantages of case-control studies

1. Easy to carry out.
2. Rapid and inexpensive.
3. Requires few subjects.
4. Suitable for rare disease.
5. No risk to subjects.
6. Can study several risk factors
7. Risk Factors can be identified & prevention can be established.
8. No attrition.
9. Ethical issue is minimal.

Disadvantages of case control studies:

1) Problem of bias.
2) Selection is difficult.
3) It cannot measure incidence
4) Relative risk is only estimated.
5) Representativeness of cases and controls is a major concern.
6) No distinguish between causes & associated factors

MR-2015 ‫صفحة‬9
 10

II-COHORT STUDY:
Concept of:
In epidemiology, the term cohort: is defined as group of people sharing common
experience or characteristics within a defined period of time . for exa mple : age
(birth cohort) , occupation (exposure cohort),etc)

INDICATION FOR COHORT STUDIES:-

1-When there is a good evidence of association between exposure to risk factors
and disease; derived fro m descriptive studies and clinical observations.
2-When exposure is rare but with high incidence among the exposed
3-when the attrition (drop out) is low which can be minimized by follow up,
stability of the cohort.
4-when funds are available

Design of cohort study:

(2) Then follow to see whether Totals Incidence rates of

disease
Disease develops Disease does not
develop

(1) First select Exposed A b a+b a/(a+b)

Not exposed C d c+d c/ (c+d)

The distinguishing features of cohort study are:

 In cohort study we proceed from cause to effect (risk factor → disease).
 It begins with a group or cohort (a+b) Knows as study cohort exposed to
particular factor thought to be related to the disease occurrence, and group
(c+d) Known as control cohort not exposed to that particular factor.
 The study groups ( exposed & non exposed) are observed for period of time
to determine the frequency of disease among them.

MR-2015 ‫صفحة‬10
 11

In cohort study, the investigator should take some important points into account:

1. The cohort must be free from the disease under study

2. Both study and control cohorts should be equally susceptible to the disease
under study.
3. Both groups should be comparable in respect of all the possible variables, which
may influence the frequency of the disease
4. The diagnostic and eligibility criteria of the disease must be defined beforehand;
this will depend on the availability of reliable methods for recognizing the
disease when it develops.

Types of cohort study:

1. Prospective: current cohort

A cohort study is in which the outco me (disease) has not yet occurred at the time
the investigation begins. It begin the present and continue into futures.

2. Retrospective:
It is called historical cohort study, is in which the outco mes have all occurred
before the start of the investigation.
The investigator goes back in time to select his study groups from existing records
of past employment, medical records and traces them forward through time. Fro m
a past date fixed on the records to the present. Retrospective cohort study is
generally more econo mical and produce results more quickly than prospective .

3- Combination of retrospective and prospective:

The cohort is identified fro m past records and is assessed of date for the outco me,
and then followed up prospectively into future for further assessment of outco me.

Elements of a cohort study:

1- Selection of study subjects
2-Obtaining data on exposure
3- Selection of comparison group
4-Follow-up
5- Analysis

MR-2015 ‫صفحة‬11
 12

1-SELECTION OF STUDY SUBJECTS:

The subjects of cohort study usually selected from:
 General population residing in well-defined geographical, political, and
administrative areas. Or by applying sampling method if population is very
large.
 special groups:
such as select groups on professional ground as doctors, nurses, teachers ect..
Or exposure groups such as industrial workers and radiologist.

2- OBTAINING DATA ON EXPOSURE

Information about exposure may be obtained from:
A) Cohort members, through personal interviews or mailed questionnaires
B) Review of records such as medical records.
C) Medical examination or special tests
D) Environmental surveys.

Data about exposure should be collected in a way that will allow classification of
cohort members; according to whether they exposed to the suspected factor or not
and according to the level or degree of exposure

3- SELECTION OF COMPARISON GROUPS:

a-Internal comparisons:
They are built-in comparisons; subjects are classified, into groups according to
the degrees or levels of exposure to risk before the development of disease in
question. The groups are then compared in terms of their subsequent morbidity
and mortality rates.

b-External comparisons:
When information on degree of exposure is not available. The study and control
cohorts should be similar in demographic and possibly important variables other
than those under study.

c- Comparison with general population rates:

The mortality experience of the exposed group is compared with the mortality
experience of the general population.
This type has its limitations such as the non- availability of rates for the outcome of
required, and the difficulties of selecting the study and comparison groups which
are representative of the exposed and non-exposed segments of the general
population.

MR-2015 ‫صفحة‬12
 13

4- FOLLOW-UP
The procedures, of follow-up, required :-
A] Periodic medical exa mination of each member of the cohort.
B] Reviewing physician and hospital records.
C] Routine surveillance of death records.
D] Mailed questionnaires, phone calls, periodic ho me visits.

5- ANALYSIS:

The data are analyzed in terms of:

A] incidence rates of outcome among exposed and non-exposed
B] estimation of risk; relative and attributable risks.

Relative risk = Risk ratio: RR

RR is the ratio of the incidence of disease (or death) among exposed and the incidence
among non exposed.
it measures the strength of association

Incidence of diseaseamong exposed (Ie)

Relative risk 
Incidence of disease amongnon exposed (Io)

Interpretation of Relative Risk (RR):

1- RR=1: No association between exposure and disease
incidence rates are identical between groups

2-RR> 1: Positive association (increased risk)

exposed group has higher incidence than non-exposed group

3-RR< 1: Negative association (protective effect)

non-exposed group has higher incidence

Attributable risk: risk difference = AR

AR is the difference in incidence rates of disease between an exposed and non

exposed group.
A R%=(incidence of disease among exposed - incidence of disease among non
exposed) / incidence of disease among exposed x 100
AR indicates to what extent the disease under study can be attributed to the exposure.

MR-2015 ‫صفحة‬13
 14

Population attributable risk:

It is the incidence of disease in total population minus the incidence of disease among
non exposed to suspected causal factor.
It provides an estimate of the amount by which the disease could be reduced in the
population if the suspected factor was eliminated

The examples below showing how the incidence be calculated and how the risk
could be estimated

Cigarette Developed Did not develop Total

Smoking Lung cancer lung cancer
Yes 70 6930 7000
No 3 2997 3000

Incidence rates:
1- among smoker = 70/7000 = 10 per1000
2- among non smoker = 3/3000 = 1 per 1000
Estimation of risk -:RR= 10/1= 10 times (strength of association (
AR=1-10/10-= 90x 100 = 90% (90% is attributed to smoking)

Death per 100,000

Heavy smokers 224

Non- smoker 10

Total population 74

RR 224/10 = 22.4

Populat ion AR 74 -10 / 74 = 86%

Advantages of cohort study:

1- Incidence can be calculated
2- Several outcomes can be studied
3- Calculate RR
4- Dose response ratio can be calculated

MR-2015 ‫صفحة‬14
 15

Disadvantages of cohort study:

1- Unsuitable for rare diseases
2- Long time
3- Ad ministrative proble m
4- Attrition
5- Cost
6- Alter people behaviour
7- Ethical problem

Exercise :
1-A case control study was conducted to assess the causal association between
exposure to rubella during the first trimester of pregnancy and congenital birth
defects in the new born. The results of the study are as given below:

History of rubella in mother Child with defect No defect

Present 25 5
Absent 35 295
1- How many controls were selected for each case?
2- Estimate the exposure rates to maternal rubella in case and controls?
3- Calculate the increased risk of CRS in children exposed to maternal rubella by
estimating odds ratio?
2-Look at the table below which is showing death rates from lung cancer per 1000 for
persons above the age 34, then answer the following questions:

Exposure Death rates

Smokers 0.98
Non smokers 0.07
1- The R.R will be equal to:
a) 0.014 b) 0.14 c) 1.4 d) 14 e) 140

MR-2015 ‫صفحة‬15
 16

2- The A.R (as%) is:

a) < 20%
b) 50%
c) 60%
d) > 90%
e) Can not be calculated
3- In the above problem the R.R means:
a) All persons below 34 never develop lung cancer
b) All lung cancer cases are caused by smoking
c) The absolute risk of lung cancer
d) That smokers are more likely to die from lung cancer by x times than non
smokers
3-Complete the following table :

Differences between case control and cohort studies:

Case Control studies Cohort studies

1. Proceeds backward from effect to cause 1. Proceeds from cause to effect

2. Starts with people with the disease 2. Starts with people free from disease.

3. Less expensive 3. More expensive

4. ……………………….. 4……….

6. 5.

6. 6.

7. 7.

8. 8.

9. 9.

10. 10

MR-2015 ‫صفحة‬16