Epidemiology Chapter 1 Flashcards
Define risk
The risk is the probability of getting a disease.
Risk must lie between 0 and 1 inclusive
Denote with the greek letter pi
Risk = number of people in the population with the disease/ number of people in the population
Examples
pi =0 : the risk is 0, the disease does not exist in the population
0
Define relative risk
Relative risk compares the risk for two groups.
Range of possible values [0, positive infinity)
Example
risk for males/risk for females = 2, then males are twice as likely to have the disease than females
Define odds
Odds = number of people in the population with the disease/ Number of people in the population without the disease
Examples
7/3 - the odds are 7 to 3 on that they have the disease
1/49 - the odds are 49 to 1 against that they have the disease
Interpretation
odds = 0, the odds are zero, it is impossible to have the disease in the population
odds = (0,1), the odds are against an individual having the disease
odds =1, the odds are even, an individual is as likely as not to suffer from the disease
odds = (1,positive infinity), the odds are in favour of the individual having the disease
odds = positive infinity, infinite odds, an individual chosen at random is certain to have the disease
Define odds ratio
Compares the odds for two groups
Range from [0.positive infinity)
Example
odds for males/odds for females = 2, the odds of the males having the disease are approximately twice those for the females
Define log odds ratio
Taking logs ensures coverage of the whole real line
risk (pi) Odds. Log odds
0 0 - infinity
0 - 0.5. (0,1) (-infinity,0)
0.5. 1 0
0.5 -1 (1,+infinity). (0,+infinity)
1 + infinity. + infinity
What is the relationship between odds and risk
Odds= risk/1-risk = pi/1-pi Risk = odds/1+odds = pi
If the risk of a disease is small (disease is rare) then the risk and odds will be approximately the same.
If the risk is small for both subgroups then the relative risk is also approximately the same as the odds ratio
Describe the design of a case-control study
A case-control study is an observational rather than an interventional study. They are retrospective.
Two groups of individuals are collected:
- Case group: individuals with the disease
- Control group: individuals without the disease
There needs to be a least as many in the control group as in the case group. Often matching is used to match the individuals in each groups in terms of age, gender, social class …..
We look back to see how many in each group were exposed to some risk factor.
The study design must be precise as we are asking people to recall the past (recall bias).
We cannot estimate risk, relative risk or odds but we can estimate the odds ratio which will be approximately relative risk for rare diseases.
What are the advantages and disadvantages of a case-control study
Advantages
- results are obtained relatively quick and cheaply
- good for study of a rare disease
- can look at multiple exposures in the one study
Disadvantages
- Cannot establish cause and effect
- selection and recall biases are potential problems
- cannot estimate risk and odds
Describe the design of a cohort study
Cohort studies are observational rather than interventional. Cohort studies are generally prospective (looking forward).
Begin by dividing our subjects into two groups: those that have been exposed and those who haven’t.
We then observe our groups going forward and record how many got a particular illness and how many did not.
A cohort study is an example of a longitudinal study.
Here we can estimate risk.
What are the advantages and disadvantages of a cohort study
Advantages
- can estimate the risk, relative risk, odds and odds ration
- less susceptible to bias than a case-control study
Disadvantages
- More time consuming and expensive than a case-control study, subjects have to be followed up over several years
- More susceptible to bias than a randomised clinical trials
- Not as good as case-control studies for rare diseases
Describe how cross-sectional study differs from a cohort study
In a cohort study patients are observed over time whereas in a cross-sectional study patients are only observed once. In cross-sectional studies subjects are selected without regard to whether they are affected or unaffected
What is meant by cofounding
When the data makes it impossible to split a factor like gender from the treatment.
For example if treatment A is given to all males and B to all females the gender and treatment effects are cofounded.
Another example if we have a case-control study and the cases group has a much higher age.
What is incidence and prevalence?
Describe the difference between them
Incidence and prevalence are two ways of measuring the frequency of disease in a population.
Incidence - measures the number of new cases over a period of time
Prevalence - measures the number of current cases of a disease in a population
Point prevalence
Point prevalence rate = Number of cases in the population at a particular time/ Number of individuals in the population
- probability or risk
- time specific
- often underestimates risk
Period prevalence
Period prevalence = (The number of cases at the start of the period+The number of new cases over the period)/The size of the population mid-period
