Lecture 4: Epidemiological methods Flashcards
Define epidemiology
the study of the distribution, determinants, and deterrents of morbidity and mortality in human populations
Define descriptive epidemiology
describes the distribution of morbidity (disease) or mortality (death) by person, place, or time variables.
This is looking at 1 variable.
Ex. Obesity in Canada
Define analytical epidemiology
studies the association or causes of disease.
This is looking at 2 variables, an exposure, and the outcome.
Ex. comparing physical activity and weight loss
What two types of studies are used in a descriptive epidemiology study design
- case series
2. cross-sectional
What are the two types of analytical epidemiology study design
- experimental
2. observational
What is an experimental design
one variable in the stud is manipulated by the investigator
What is an observational design
investigator observes without intervention
What are the three types of studies used in an observational design
- cohort
- case-control
- case-sectional
What is a case series study
Looking at a group of individuals that share a similar characteristic
What is a cohort study
looking at an exposed population and following them through time to observe the outcome
Ex. Have a population of smokers and observe if they get lung cancer in later years
Wha is a case-control study
looking at the outcome of a population and looking for the exposure in the past.
Ex. Have a population of people with lung cancer and looking back 5 years to see if they smoked
What is a cross-sectional study
Simultaneously looking at the exposure and the outcome (a snapshot in time)
Ex. Is the individual a smoker and do they have lung cancer
What is the hierarchy of evidence
A pyramid list that rates a study design based on their quality of evidence
List the hierarchy of evidence from best quality to worst quality
- systematic review (looks at multiple studies)
- randomised control trials
- cohort
- case-control (can be difficult for participant to remember the past)
- cross-control (no time element)
- ecological (looking at averages)
- case series
- expert opinions
What are the two main problems with randomized control trials
- Ethical: We cannot ask someone to smoke because we want to see if they get lung cancer
- Demographics: we cannot randomize someone to a specific ethnic group (ethnicity and health outcome)
What are descriptive epidemiology measures
Measures of disease frequency:
- Incidence
- prevalence
What are analytical epidemiology measures
Measures of association:
- risk difference
- relative risk odds ratio
What is the prevalence rate
The proportion of the population that has a given disease or condition at a specified time
point prevalence rate = X/Y
point PR = individuals with disease/population
What is the incidence rate
The proportion of the population at risk that develops a given disease or condition during a specific time period
incidence rate = X/Y
IR = new cases of disease/population at risk
population at risk = total population - individuals already with the disease in the population
What is the relationship between incidence and prevalence
A decrease in prevalence is either due to:
- a decrease in the incident rate
- an increase in the resolution (recovery from disease or death)
What is relative risk
The ratio of the risk of the outcome in an exposed group to the risk of the outcome in the unexposed group (referent group).
Provides an estimate of the increased or decreased risk of an outcome due to a particular exposure
relative risk = X/Y
RR = incidence rate of exposed group/incidence rate of unexposed group
How do you turn a relative risk ratio or an odds ration into a percentage
subtract the ratio number by 1 and multiply by 100%
If I had a relative risk of 1.5 for smoking and heart disease, how would I present that in a sentence
The smokers have 1.5 times the increased risk of heart disease compared to non-smokers
What is the odds ratio
used in place of the relative risk in case-control studies and cross-sectional studies
Case-control:
Odds ratio = X/Y
OR = prevalence of exposure in cases/prevalence of exposure in controls
Cross-sectional
Odds ratio = X/Y
OR = prevalence of outcome in exposed group/prevalence of outcome in unexposed group
If I had an odds ratio of 2.2 for economically unstable children and protein deficiency, how would I present that in a sentence
Economically unstable children have 2.2 times increased odds (or likelihood) of protein deficiency compared to economically stable children
What does a value of 1 mean for relative risk and odds ratio
no difference between exposed and unexposed
What does a value less than 1 mean for relative risk and odds ratio
a decreased risk compared to the referent group
What does a value greater than 1 mean for relative risk and odds ratio
an increased risk compared to the referent group
Give a range for a weak, moderate, and strong relative risk
weak: 1.01 - 1.50
moderate: 1.51 - 3.00
strong: > 3.01
What is Germ Theory for cause and effect
Thought disease has one single causative agent
What is Black Box
- understand the risk factors and the disease but not necessarily the mechanics
- a change in thinking since developed countries shifted to one dominated by chronic disease.
Describe cause and effect
an association between two factors does not necessarily imply cause and effect, a change in the exposure must result in a corresponding change int he outcome.
What are Hill’s 9 criteria to help determine if associations are causal
- Temporality (exposure occurs before the disease
- Strength of association
- consistency
- specificity (germ theory)
- dose-response relationship (linear relationship between exposure and outcome)
- biological plausibility (a model exist that can explain the association)
- coherence (does not conflict with any known facts about the history of the disease)
- experiment (has been tested)
- analogy (similarities between associations)
