Epidemiology

• Study of distribution and determinants of disease frequency in human populations.

• Based on two assumptions that diseases does not occur at random and casual and
preventive factors can be identified through systematic investigations of different
populations in different places at different times.

Aims of Epidemiology
• Describing distribution and magnitude of health and disease problems in human
populations
• To identify the causes/related factors of disease
• Providing data essential for planning, implementation and evaluation of services for
prevention, control and treatment of diseases and to prioritize those services

Types of measurements
• Count/frequency – number of affected individuals
• Proportion – Count related to source population in which the count is taken.
• Percentage – Proportion multiplied by 100
• Rate – Events or cases occurred in a defined population in a defined time, usually expressed
in multiples of 10
• Ratio – Value obtained by dividing one quantity by another. (male:female ratio), numerator
is not a part of denominator

Measures of Disease frequency – Depends on counting cases in a defined population within a

defined time period. For that there should be a case definition, which should be clear,
unambiguous, easy to measure in a standard manner under a variety of circumstances and easy to
use. For the population, we should use population at risk, if it not available we can use total
population as an approximation.
𝑵𝒐.𝒐𝒇 𝒆𝒙𝒊𝒔𝒕𝒊𝒏𝒈 𝒄𝒂𝒔𝒆𝒔 𝑜𝑓 𝑎 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑜𝑟 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛 𝑎𝑡 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑒𝑑 𝑝𝑜𝑖𝑛𝑡 𝑖𝑛 𝑡𝑖𝑚𝑒
• 𝑃𝑟𝑒𝑣𝑎𝑙𝑒𝑛𝑐𝑒 (𝑃) = ×𝑘
𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛−𝑎𝑡−𝑟𝑖𝑠𝑘 𝑖𝑛 𝑡ℎ𝑒 𝑑𝑒𝑓𝑖𝑛𝑒𝑑 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑝𝑜𝑖𝑛𝑡 𝑖𝑛 𝑡𝑖𝑚𝑒

• 𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝐼𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 (𝐶𝐼) =

𝑵𝒐.𝒐𝒇 𝒏𝒆𝒘 𝒄𝒂𝒔𝒆𝒔 𝑜𝑓 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑒𝑑 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑑𝑢𝑟𝑖𝑛𝑔 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑡𝑖𝑚𝑒
×𝑘
𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛−𝑎𝑡−𝑟𝑖𝑠𝑘 𝑑𝑢𝑟𝑖𝑛𝑔 𝑡ℎ𝑎𝑡 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑

𝑃𝑒𝑟𝑖𝑜𝑑 𝑃𝑟𝑒𝑣𝑎𝑙𝑒𝑛𝑐𝑒
𝑵𝒐. 𝒐𝒇 𝒆𝒙𝒊𝒔𝒕𝒊𝒏𝒈 𝒄𝒂𝒔𝒆𝒔 𝑜𝑓 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑒𝑑 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑎𝑡 𝑡ℎ𝑒 𝑏𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 𝑜𝑓 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑝𝑒𝑟𝑖𝑜𝑑
+
𝑵𝒐. 𝒏𝒆𝒘 𝒄𝒂𝒔𝒆𝒔 𝒅𝒊𝒂𝒈𝒏𝒐𝒔𝒆𝒅 𝑑𝑢𝑟𝑖𝑛𝑔 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑
= ×𝑘
𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑 𝑎𝑡 − 𝑟𝑖𝑠𝑘 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 ′𝑚𝑖𝑑 ′ 𝑡𝑖𝑚𝑒 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙
Period prevalence combines both point prevalence and incidence as for some disease (mental
illness), it is difficult to determine those values.

𝐼𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝐷𝑒𝑛𝑠𝑖𝑡𝑦 𝑜𝑟 𝐼𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑅𝑎𝑡𝑒(𝐼𝐷)

𝑁𝑜 𝑜𝑓 𝑛𝑒𝑤 𝑐𝑎𝑠𝑒𝑠 𝑜𝑓 𝑎 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑑𝑢𝑟𝑖𝑛𝑔 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑝𝑒𝑟𝑖𝑜𝑑
= ×𝐾
𝑇𝑜𝑡𝑎𝑙 𝑝𝑒𝑟𝑠𝑜𝑛 − 𝑡𝑖𝑚𝑒 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠
Incidence density is used because the whole group will not be available through the study period, so
we calculate person-time by multiplying number of person by the time they were in the study.

Use with caution

Examples of incidence rate
• Morbidity rate
• Mortality rate
𝑁𝑜.𝑑𝑒𝑎𝑡ℎ𝑠 𝑓𝑟𝑜𝑚 𝑎 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑖𝑛 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑡𝑖𝑚𝑒
• 𝐶𝑎𝑠𝑒 𝐹𝑎𝑡𝑎𝑙𝑖𝑡𝑦 𝑅𝑎𝑡𝑒 (𝐶𝐹𝑅) = × 𝐾,
𝑁𝑜.𝑑𝑖𝑎𝑔𝑛𝑜𝑠𝑒𝑑 𝑐𝑎𝑠𝑒𝑠 𝑜𝑓 𝑡ℎ𝑎𝑡 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑖𝑛 𝑡ℎ𝑎𝑡 𝑝𝑒𝑟𝑖𝑜𝑑
It indicates the severity of the disease.
𝑁𝑜. 𝑑𝑒𝑎𝑡ℎ𝑠 𝑓𝑟𝑜𝑚 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑐𝑎𝑢𝑠𝑒 𝑖𝑛 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑡𝑖𝑚𝑒
• 𝐶𝑎𝑠𝑒 𝑆𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑀𝑜𝑟𝑡𝑎𝑙𝑖𝑡𝑦 𝑅𝑎𝑡𝑒 = ×𝐾
𝑇𝑜𝑡𝑎𝑙 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
• 𝐴𝑡𝑡𝑎𝑐𝑘 𝑅𝑎𝑡𝑒 (𝐴𝑅) =
𝑁𝑜.𝑛𝑒𝑤 𝑐𝑎𝑠𝑒𝑠 𝑜𝑓 𝑎 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑡𝑜 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑒𝑥𝑝𝑜𝑠𝑢𝑟𝑒 𝑑𝑢𝑟𝑖𝑛𝑔 𝑎 𝑠ℎ𝑜𝑟𝑡 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑡𝑖𝑚𝑒
×𝐾
𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛−𝑎𝑡−𝑟𝑖𝑠𝑘 𝑖𝑛 𝑡ℎ𝑎𝑡 𝑙𝑖𝑚𝑖𝑡𝑒𝑑 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛
It is useful during an outbreak of a disease (e.g- food poisoning) to determine the specific
cause.

Epidemiological
Study Designs

Observational Experimental
Studies Studies

Therapeutic
Descriptive Analytical
trials

Case report and Case control

Preventive trials
case series studies

Cross-sectional Cross-sectional
studies studies

Correlational
Cohort studies
Studies

• Observational studies – investigator observes and measures but does not intervene
• Experimental studies – investigator intervenes in one or more variable in a group and does not
intervene in another group, but not always possible.
• Descriptive studies – Focuses on describing the occurrence of a disease or a health related
event in a population in terms of person (age, sex, religion, marital status, socio economic
status, etc.), place (districts, countries) and time (monthly or annual). Three kinds of change
with time are secular trend (progressive increase or decrease over a long period of time, e.g-
crude birth rate and death rates for Sri Lanka from 1945 to 2003), cyclical trend (periodic
fluctuations on an annual or seasonal basis) and epidemic trend (short term fluctuations). Does
not begin with hypothesis but its results are used to formulate the hypothesis. Uses are trend
analysis, health care planning and hypothesis generation.
• Analytical Studies – Focuses on the determinants of a disease or health related event in order to
establish whether a particular exposure causes or prevents it. Investigator begins with a
hypothesis (usually formulated based on the findings of a cross-sectional study) and designs the

Use with caution

 study to specifically test that hypothesis. There is always a comparison group. Although it tests
associations between exposure and outcome, but this alone will not imply causality of a
disease. To judge causality we have to consider strength of association, biological credibility of
the hypothesis, consistency of the findings, temporal sequence and the presence of a dose-
response relationship.
• Case report – Carefully detailed report of a single patient.
• Case series – New or unusual collection of individual case reports.
Case report and series help to identify new clinical issues and may lead to development of
hypotheses. But are considered low level evidence as the observations may be subject to bias. If
they are retrospective, depends on availability and accuracy of data records. In case series, they are
subject to selection bias because clinician selects the cases. There is no control group for
comparison.
• Correlational studies – Comparisons are made using the same population at different periods of
time or different groups of populations during the same period of time.
It is the first step in investigating a possible relationship between exposure and disease and
formulating hypothesis. It can be done quickly and inexpensively, often using the already available
data. But we cannot link an exposure to the occurrence of a disease in the same person as the data
refers to the population. So they are not used to test hypotheses. Other limitation is there may be
other factors that are associated with the exposure.
• Cross Sectional surveys – Provide information about frequency and characteristics of a disease
by providing a snapshot of the experience of a population at a specified period or point in time.
Steps taken in conducting the study are
o have a clear objective
o define the study population
o give due consideration to using a sampling method
o ensure adequate response rates
o identify methods of data collection
o carry out appropriate analysis
Great value to public health administrators in assessing the health status and health care needs of a
population. Only prevalence are measured in this study. As both exposure and outcome are
measured at the same time, cost is low and no loss to follow-up. But temporal sequence of exposure
and outcome are impossible to work out.

Hypothesis – Statement of belief or intention, which one expects to prove or disprove. Relates to
certain factor/s which causes or relates to the occurrence of a disease.

• Cohort studies / follow up/incidence studies

Cohort is a group of persons who share a common characteristic or an experience. The direction of
inquiry is from exposure to disease. So can establish the temporal sequence. It can be prospective
cohort studies (investigator is looking forwards from a exposure to outcome) and retrospective
cohort studies (investigator is looking backwards from exposure to outcome). These two types are
based only on the timing of data collection. Steps
1. Identify a suitable cohort – If the exposure is common, general population can be used. If it
is a rare exposure (chemotherapy, x ray treatment), identify a group who have undergone
that exposure, so sufficient number of exposed persons can be collected within a short
period, more complete and accurate information on exposure and good compliance is
possible and allows for identification of aetiological agents in special circumstances.

Use with caution

 2. Select an appropriate comparison group – needs to be similar as possible to exposed group
with respect to all other factors except the exposure. Two types of comparison groups,
internal comparison group (comparison selected from the same cohort) and external
comparison group (when all the individuals in the cohort are exposed, from another group
comparison is selected). By using multiple comparison groups, we can strengthen the study
when using external comparison groups.
3. Collect data on exposure and subsequent outcomes – Data can be collected by surveys with
follow up procedures/interviews, medical and employment records, periodic medical
examination and interviews. In study with fatal endpoints, outcome data can be obtained
from death certificates. But this is not reliable if the outcome is mortality due to specific
cause. For non-fatal end points, outcome can be obtained from physician’s records, BHT,
hospital registers and pathology reports.
4. Approaches to follow up – Main reason for this type of study for it high cost in time and
money.
5. Analysis and interpretation – data can be presented in a two by two table.
Exposure Status Outcome status Total
Present Absent
Yes a b a+b
No c d c+d
Total a+c b+d a+b+c+d

Relative Risk/ Risk Ratio (RR) – estimates the strength of an association between exposure and
outcome
𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑅𝑖𝑠𝑘 (𝑅𝑅)
𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑜𝑟 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 𝑜𝑓 𝑎 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑖𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝 (𝐼1 )
=
𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑜𝑟 𝑖𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑑𝑒𝑛𝑖𝑠𝑡𝑦 𝑜𝑓 𝑎 𝑑𝑖𝑠𝑒𝑎𝑠𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑛𝑜𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 𝑔𝑟𝑜𝑢𝑝 (𝐼0 )

RR of 1.0, indicates there is no association between exposure and the outcome.

RR > 1, indicates there is a positive association/ increased risk of disease among exposed
compared to non exposed.
RR < 1, indicates there is a negative association/decreased risk of disease among exposed
compared to non exposed.

Attributable Risk/ excess risk/ absolute risk – difference in disease occurrence between exposed and
non-exposed. It measures the absolute effect of exposure or the excess risk of a disease in exposed
group compared to non-exposed group. But it is only valid when there is a cause-effect relationship
exists between exposure and outcome.
𝐴𝑡𝑡𝑟𝑖𝑏𝑢𝑡𝑎𝑏𝑙𝑒 𝑟𝑖𝑠𝑘 (𝐴𝑅) = 𝐼𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 (𝐼1 ) − 𝐼𝑛𝑐𝑖𝑑𝑒𝑛𝑐𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑛𝑜𝑛 𝑒𝑥𝑝𝑜𝑠𝑒𝑑 (𝐼0 )

AR of 0, indicates there is no association between exposure and outcome

AR > 0, indicates there is an excess risk of disease among the exposed that can be attributable to
that exposure. (Incidence rate of disease that can be eliminated if the exposure was removed)

RR and AR are both measures of associations, but RR provides information about strength of
association between exposure and outcome and AR information about the extent of public health
impact, if the exposure is removed assuming that there is causal relationship between exposure and
outcome.

Use with caution

Advantages
Incidence rates, natural history of disease and
their staging
More than one disease related to the same risk
factor
Cause-effect relationship can be determined as
risk factor precedes outcome
Both relative and attributable risks can be
calculated
Disadvantages
Long term, expensive , time consuming and not
feasible always
Not suitable for rare diseases as large numbers
need to be followed up
Loss to follow up affects validity of conclusions
and samples become less representative of
population
There may be changes in exposure status
Study itself can influence behaviour of people
under investigation
Serious ethical issues when apparent disease
excess before data completion

• Case Control studies

We divide the subjects into two groups (having the disease outcome or not) and take the history of
exposure from them. Compare the exposure pattern of the cases and the controls. Steps in the
design are
1. Select cases and control – Both should be comparable by the baseline risk of developing the
disease other than from the exposure under study. Cases should represent a disease entity
as homogenous as possible. (not hepatitis but hepatitis A). Case definition should be clear
and unambiguous containing inclusion and exclusion criteria. Cases can be selected from a
clinic or hospital or community survey within a specified period of time. Newly diagnosed
cases are preferred than prevalent cases because prevalent cases can have other factors
that would have enable them to survive and will give false information. Controls should be
clearly defined and can be obtained from hospital, general population, from same
population and from family, friends and neighbourhood. Hospital controls are advantageous
because, they are easily identified, readily available, more willing, easily recall exposure
events and when selected from same hospital they would have same background factors.
But the disadvantage is disease of the controls may be associated with exposure under
study.
2. Match cases and controls
3. Measure the exposure status – Exposure data from both cases and controls are collected via
interviews, questionnaires and examination of records.
4. Analysis and interpretation – Relationship between exposure and outcome can be presented
in a two by two table.
Exposure Status Outcome status Total
Present Absent
Yes a b a+b
No c d c+d
Total a+c b+d a+b+c+d

𝑎𝑑
𝑂𝑑𝑑𝑠 𝑟𝑎𝑡𝑖𝑜 = Interpreted as same as the relative risk.
𝑏𝑐

Advantages Disadvanatages
Ideal for rare diseases of diseases with long Possible bias in selecting cases and controls
incubation period

Use with caution

Require smaller sample than cohort
Relatively cheap
Results obtained relatively quickly
Attrition (Loss to follow up) not a problem
Can investigate a wide range of possible risk
factors
Recall and interviewer bias and difficulty in
obtaining information
As population at risk is not there, calculation of
incidence rate is not possible
No way to find out if the exposure was dame
for those who died and those who survived
Not possible to find out about the pathology of
other diseases related to risk factor under study
Problems in sorting out the temporal sequence

• Cross sectional study – more appropriate for measuring relationships between permanent
characteristics of individuals and chronic diseases or stable conditions. Impractical for the study
of rare diseases, conditions of short duration and diseases with high case fatality. Data is
presented in a two by two table.

Advantages Disadvantages
No waiting for the outcome to occur No temporal sequence can be established as
both exposure and outcome are measured at
the same time
More feasible, less costly Chicken egg dilemma
Only study to five the prevalence of a disease
or a risk factor

Chi Square Test

It can be used to test null hypothesis.

Exposure Status Outcome status Total
Present Absent
Yes a B a+b
No c D c+d
Total a+c b+d a+b+c+d

Exposure present Exposure absent

Outcome present Outcome absent Outcome present Outcome absent
Observed a B C D
value (O)
Expected 𝑎+𝑏 𝑎+𝑏 𝑐+𝑑 𝑐+𝑑
× (𝑎 + 𝑐) × (𝑏 + 𝑑) × (𝑎 + 𝑐) × (𝑏 + 𝑑)
value (E) 𝑡𝑜𝑡𝑎𝑙 𝑡𝑜𝑡𝑎𝑙 𝑡𝑜𝑡𝑎𝑙 𝑡𝑜𝑡𝑎𝑙
𝑂−𝐸
(𝑂 − 𝐸)2
(𝑂 − 𝐸)2
𝐸
(𝑂 − 𝐸)2
𝜒2 =
𝐸
2 (𝑂−𝐸)2 2
𝜒𝑑.𝑓. =Σ , which indicates the 𝜒 value at a particular degrees of freedom.
𝐸

Calculate the degress of freedom, 𝑑𝑓 = (𝑛𝑜. 𝑜𝑓 𝑟𝑜𝑤𝑠 − 1)(𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑜𝑙𝑢𝑚𝑛𝑠 − 1)

Use with caution

Refer to the chi square distribution table and read the appropriate value for the number of freedom
against p=0.05 If the calculated value is greater than or equal to the value from the table (critical
value), reject the null hypothesis.

Experimental Studies/ Interventional Studies

There are two types – Therapeutic trials/randomized clinical trials/ randomized controlled trials and
preventive trials/field trials (vaccine trials)
Preventive trial can be carried among the whole community, then it is called the community trials.
(e.g- fluoridation of water supply, iodination of salt)

Steps in the process are

1. Select the study population – Identify the refernce population (to which results of trial will
be applicable) and experimental population (actual group in which trial is conducted).
Experimental population should be sufficiently large enough. Informed consent is taken from
study subjects. Subjects are screened under inclusion and exclusion criteria.
2. Allocate study regimens – Randomization is essential. If the sample size is large,
randomization is more successful. and selection bias is minimized. Randomization can be
done by computer generated randomization list or table of random numbers. Sometimes,
study group is compared with a historical group.
3. Maintain and monitor compliance
4. Measure outcomes – in both groups should be assessed in a uniform manner.

Ethical issues in experimental studies

• Informed consent
• Voluntary participation
• Autonomy
• Fair subject selection
• Risks weighed against benefits
• Maintaining anonymity and confidentiality of participant information
• Conflicts of interest
• Ethical clearance from local ERC
• Determining stopping rules (decisions for early termination of a trial)

Bias
Error made in epidemiological studies due to known sources of variation resulting in an incorrect
estimation of association between exposure and outcome.
1. Selection bias
2. Recall bias
3. Interviewer bias
4. Observer bias
5. Loss to follow up

Use with caution

Use with caution