Epidemiological methods
Objectives
At the end of the week students should be able to:
 Differentiate between different types of data.
 Describe the structure of an epidemiological
dataset
 Define and calculate measures of disease
occurrence and measures of association
 Describe the basic features of the main types of
epidemiological studies
 Explain the main features of bias, confounding,
chance
 Be able to discuss causality of the association
Epidemiology
 The study of the distribution and
determinants of the frequency of healthrelated
outcomes in specified populations
 Quantitative discipline
 Measurement of disease / condition / risk
factor frequency is central to epidemiology
 Comparisons require measurements
Much of epidemiological research is
taken up trying
 to establish associations between
exposures and disease rates
 to measure the extent to which risk
changes as the level of exposure changes
 to establish whether the associations
observed may be truly causal (rather than
being just consequence of bias or chance)
 Epidemiology has a major role in developing
appropriate strategies to improve public
health through prevention
◦ public health has wider meaning in this sense; it is
about the health of the whole population.
◦ it does not cover only classic areas, such as
immunization or monitoring of diseases, it also
covers factors such as poverty, smoking, nutrition
 In this sense, epidemiology has a crucial role
in trying to put into perspective the effects on
population health of different risk factors.
Variables (outcomes/risk factors)
 Binary
◦ Deaths (y/n)
◦ Disease (y/n)
 Categorical (ordinal or nominal)
◦ Frequency of drinking (never, 1-3 times a month,
1-3 times a week, 4 times a week or more
often)
◦ Severity of pain (none, some, a lot)
 Continous
◦ BMI, blood pressure etc
What type of variable is…
 Self-rated health
◦ Very poor, poor, average, good, very good
 Total cholesterol concentration
 Economic activity
◦ Employed, unemployed, housewife, pensioner
 Risk of CVD death in the next 10 years
(SCORE)
 Ethnicity
 Quartile of income
 Sex
 Marital status (married, divorced, ever single,
widowed)
Binary outcomes:“cases” vs.
“non-cases”
 Persons with disease = “cases”
 Definition of case is crucial
 E.g.
◦ Obesity: BMI≥30
◦ Hypertension: SBP≥140 mm Hg or DBP≥90 mm
Hg or treatment
◦ High cholesterol: ≥6.2 mmol/L
 Must always be clearly specified
Measures of disease frequency
 Used for binary outcomes
 Require a numerator and denominator
number of persons with disease
--------------------------------------------------
number of persons examined
 expressed as X per 1000 persons (or per 100,000 etc)
Numerators and denominators
 The number of cancer cases in the UK is 247,667
whereas in Belgium it is 47,948.
 The UK has a bigger problem in numerical terms.
 But do Belgians have lower risk of getting cancer?
◦ Numerators alone are meaningless
◦ We need both numerators AND
denominators
Numerators and denominators
 The number of cancer cases in the UK is 247,667
whereas in Belgium it is 47,948.
 The UK has a bigger problem in numerical terms.
 But do Belgians have lower risk of getting cancer?
◦ Numerators alone are meaningless
◦ We need both numerators AND
denominators
 UK: 247 667 / 60 000 000 = 0.00413 = 413 per 100
000
 Belgium: 47 948 / 10 000 000 = 0.00479 = 479 per
100 000
Prevalence
 number of existing cases / population of
interest at a defined time
• number of new cases in a given time
period / total population at risk
Incidence
Prevalence
 number of existing cases / population of interest at a
defined time
◦ Unable to work now for health reasons
◦ Injury ever in the past
◦ Ever wheezing or whistling in the chest
NOTE a denominator is needed for prevalence
Adult prevalence by BMI status
Health Survey for England (2008-2010 average)
Adult (aged 16+) BMI thresholds
Underweight: <18.5kg/m2
Healthy weight: 18.5 to <25kg/m2
Overweight: 25 to <30kg/m2
Obese: ≥30kg/m2
© NOO 2012
Healthy weight
40.8%
Underweight
2.1%
Overweight
32.2%
Obese
24.9%
Women
Healthy weight
31.8%
Underweight
1.7%Overweight
42.4%
Obese
24.1%
Men
Incidence rates
 In 2014, 55,222 new cases of breast cancer were
diagnosed in the UK.
 Approximately 65M people in the UK
 Most cases in women (only 389 cases in men)
 Population at risk?
 Cumulative incidence of breast cancer in the UK
in 2014 in females was ?
???
------------
???
Incidence rates
 In 2014, 55,222 new cases of breast cancer were
diagnosed in the UK.
 Approximately 65.5M people in the UK
 Most cases in women (only 389 cases in men)
 Population at risk?
 Incidence of breast cancer in the UK in 2014 in
females was ?
55222-389 54833
-------------------- = ---------------- = 0.001674=167.4/100,000
65.5M/2 32.75
Incidence rate example:
3-year study with a sample size of 100, outcome of
interest was fatal heart disease.
year 1 year 2 Study ends
developed outcome 6 5 4
dropped out 4 10 sample
at risk 90 75 71
 10 participants were followed for 1 year
 15 participants were followed for 2 years
 75 participants were followed for 3 years
Total person-years:
Rate =
Incidence rate example:
3-year study with a sample size of 100, outcome of
interest was fatal heart disease.
year 1 year 2 Study ends
developed outcome 6 5 4
dropped out 4 10 sample
at risk 90 75 71
 10 participants were followed for 1 year
 15 participants were followed for 2 years
 75 participants were followed for 3 years
Total person-years of follow up = (10x1) + (15x2) + (75x3) = 265 person-years
at risk
Incidence rate = 15 / 265 = 0.057 = 57 cases per 1000 person-years
Relationship between prevalence and
incidence
 The prevalence of a health-related outcome
depends both on the incidence rate and the
time between onset and recovery or death.
 Prevalence = Incidence x Average
disease duration
 E.g. volume of water in water tank depends on
◦ Inflow
◦ Outflow
Mortality
 number of deaths / total population
 Rate (or risk)
 the number of deaths in a specified population, divided
by the number of that population, per unit time.
 If the mortality rate is to be calculated in a given year,
the mid-year population is usually used as the
denominator.
 Mortality rate is always expressed as deaths per X (e.g.
1000 persons per year). E.g.
◦ A city has a population of 900,000, 30,000 deaths occur in a 3-year
period.
◦ Mortality rate for the period = 30 000 / 900 000 = 0.0033 or 33
deaths per 1000 per 3 years
◦ = 11 deaths per 1000 per year.
Mortality rates can be:
 All-cause mortality rates: refers to the total
number of deaths per 1000 people per year.This
is also usually referred to just as all-cause
mortality.
 Cause-specific mortality rate refers to total
number of deaths due to a specific cause.
Mortality rates can be:
 Crude mortality rates – no care has been
taken for age structure of the population
 Standardised mortality rate refers to a
mortality rate which is age-standardised in order
to permit comparisons between different
countries, regions etc.
Case fatality
 Case fatality rate is the rate of death among
people who already have a condition, usually in a
defined period of time. usually measured as a
decimal or as a percent.
 Survival rate is the proportion of people who
remain alive for a given period of time after
diagnosis of disease. E.g. breast cancer has 5-year
survival rate around 70%.