Epidemiology
Epidemiology 12
The study of the distribution and determinants of a disease frequency in human populations
epidemic/outbreak
The occurrence of a disease in members of a defined population
More than the normal amount of cases in a population
Pandemic
Throughout the population of a country, people, or world
Epidemiologic methods purpose
Measure how common health problems are.
Find out what causes diseases.
Help decide how to best use public health resources.
Check if prevention strategies are working through ongoing monitoring.
History of epidemiology
5th century B.C Hippocrates
suggested that the development of human disease may be related to the external and personal environment of individual
1662 John Graunt
analyzed weekly reports of births and deaths
First to quantify patterns of disease
# of men that were born/died
High infant mortality rate
Seasonal variations (more die in winter)
William farr 1839
System of routine complication of numbers and causes of death
Compared mortality rates with several different characteristics
John snow (2 decades later) used williams data
Formulated and tested hypothesis’ concerning origin of cholera epidemic
Cholera: bacterial infection of the small intestine
Symptoms - severe diarrhoea and vomiting, muscle cramps, dehydration, and depletion of electrolytes
Suggested it came from contaminated water
Charted frequency and distribution of water
Found out cause of cholera
Components of epidemiology principles and methods
Distribution: where, when who
Comparisons between different populations at a given time, or between subgroups, or different periods of observation
Describes disease patterns
Formulates hypotheses to cause and preventative factors
Disease frequency
Quantification of occurrence of disease in human population
Determinants of disease
Distribution and disease frequency
Necessary to test epidemiologic hypotheses
Hypothesis
A statement derived from a theory that predicts the relationship among variables representing concepts, constructs, or events
What researcher expects to find
Key assumption of epidemiology
Majority of human disease doesn't occur at random
Casual and preventative factors of human disease can be identified through different populations or subgroups
Primary units of concern
Interested in groups of people - not individuals
Groups must be studied to find the cause and prevention of a disease
Key concepts of epidemiology
Quantitative science
Count
# of people studied who have particular disease
Ex. 40 of 90 KP290 students have asthma
Ratio
Relationship between 2 numbers
Ex. 100 males : 150 females
Proportion
Special type of ratio
The numerator is part of the denominator and the resulting proportion is expressed as a percentage
Ex. 100 (males) / 250 (females + males) x100%
Rate
Certain kind of proportion
Frequency of how much an event occurs over a certain time
Prevalence rate (point prevalence rate)
Can only exist at one point in time
# of preexisting cases of a condition within a specific population and time per 100 of population at risk
Denominator includes everyone
= # of cases (specified time)/ population (everyone) x100
Incidence
# of new cases
Incidence rate
# of new cases of a starting during a specific time/ population - pre existing cases x 100 = population at risk
Epidemiologic study designs
Descriptive study designs
Concerned with disease distribution
Used for initial hypothesis about exposure disease relationships
Who, where, when (for hypotheses)
Who: what populations do not develop disease
Where: what location is more or least common
When: how does the frequency of the disease occurrence vary over time
Cross-sectional and ecological
Cross-sectional design
Measures cause and effect at a certain point and look at relationships
Advantages
Looks at individuals rather than groups
Can control potentially confounding variables
Limitations
Outcome and exposure are measured at same point in time
Cannot make statement about cause and effect
Cannot examine continuous relationships
Ecological design
Use existing data sources to understand the relationship between outcome and exposure at a population level
Analytic study designs
Focus on disease causes by testing hypotheses, formulated from descriptive studies
The goal is to determine if exposure to a certain factor prevents disease
Cohort studies and case-control studies
Cohort studies (prospective studies)
Begin with a small group of people and follow them over a period of time
Amount of exposure to each will vary
Try to figure out cause and effect
Try to figure out the rate of X occurs in relation to the amount of exposure to Y
All cohort members have no health problems at the start of study
Groups divided into 2 or more
Sample of cohort selected from population grouped into cases and non cases
Advantages
You know exactly the time between the exposure and outcome
Good for rare exposure
Good for understanding multiple effects of a single exposure
Limitations
Difficult to do (large sample)
Costly
Forget to follow up (can last years)
Some diseases are very rare, may be difficult to get a sufficient number of cases for analysis
How to select a cohort
Accessibility (easy selection)
History of previous exposure (random sample)
Medical records
Case-control studies (retrospective studies)
Participants from a group with a disorder and compare cases without disorder
1st group who already have X (aka cases)
2nd group who doesn't have X (aka controls)
Need to determine how 2 groups differed in their exposure to Y
How to select cases
Physicians
Employers
Healthcare providers
Medical records
Dental records
Subjects themselves
Advantages:
Results can be seen in a short period of time
Efficient for studying rare diseases ( don't need large population or long follow up periods)
Enable hypothesis testing for multiple exposures for a single disease outcome
Can be used to get more detail about exposure
Limitations
Exposure information is obtained after disease has been diagnosed
Could be recall bias
Can be challenging to recruit control group (can affect odds ratios)
Can be selection bias
Errors in data collection
Clinical observations (disease presence/ absence)
Physical examinations rounding of B.P
Medical history interview
Disease reporting
Clinical diagnosis
Different criteria for making same diagnosis
Death certificates + mortality statistics
Medical chart review
Missing info
Illegible
Laboratory data
Ingestion of certain drugs affects blood constituents
Responses to questionnaire
Non response
Inconsistent response
Overreporting symptoms
Understanding factors
Establishing causation
Temporal sequence
The thing that causes disease came before disease
Consistency
Strength of association
Specificity of effect
Biological gradient
Existing data and theory
Problems of error
Bias
Selection bias
Information bias
Confounding bias
Random variation
Chance differences between groups
Random misclassification
Subject could be in wrong group
Exposed person with non exposed (vice versa)
How to control errors?
Matching variables
Homogeneous grouping
Stratified sample (1:1 ratio)
Post-stratification
Sample first, then strata