Flashcards in Day 2 Deck (33)
What is the prevalence of a disease?
Prevalence is the proportion of people who have a disease at a given point in time.
-It counts the number of people with existing disease (both old and new)
-Takes a 'snapshot' at a given point in time
-Describes the 'burden of disease'
-Useful to determine resource/service allocation
What is the incidence of a disease?
This is the number of new cases of a disease within a given timeframe.
-It focuses on NEW cases only
-Useful when monitoring epidemics eg the number of new cases of the zika virus in the Americas.
When taking a sample, what three things does it need to be?
What are the two types of error they may occur in a study?
Random error (chance)
-This is due to sampling variation
-This will reduce as sample size increases because precision will be increased (reduced uncertainty)
Systematic error (bias)
-Bias is quantified by the difference between the true value and the expected value
-It DOES NOT reduce as sample size increases
What are some sources of bias?
You get selection biases and information biases.
Selection biases incude:
-Study sample (external validity).
For example, this could be where you conduct the study. It is when the study sample is not representative of the entire population of interest.
-Group selection within a study (internal validity).
This means that groups within a study may not be comparable. You have to ensure that other factors are the same in both populations.
-Healthy worker effect
Workers usually exhibit lower overall morbidity then the general population (eg severely ill/disabled are usually excluded form employment.
Information biases include:
Differences in recollection from study participants regarding events or experiences from the past
-Observer or interviewer error
Study observer/interviewer may have preconceived expectations or knowledge that may influence the result
Differences in the measurement of participants
Participants are put in the wrong group (eg diseased when they are not). This usually arises form observational or measurement error
How is prevalence often reported?
Prevalence is often reported as a proportion (not a rate!).
Prevalence = No. of people with the disease / Total population.
How is incidence often reported?
Incidence is often reported as a rate (eg 50 per 100,000 person years)
Incidence = Number of new cases / Person time at risk = events per person per year
What is the equation that links incidence and prevalence?
Prevalence = Incidence x Duration
What is the incidence rate ratio?
The incidence rate ratio compares the incidence rate in one group to another.
IRR = indicence rate in group A / incidence rate in group B
The IRR is a RELATIVE measure between two groups. We have use relative measure such as IRR to compare 3 groups for other factors such as mortality rate ratios.
What is the difference between relative rate and relative risk?
Relative risk is the underlying quantity we wish to approximate whereas the relative rate approximates the relative risk if the underlying disease is rare, or if the time period is short.
What is the equation to work out the odds of an event happening?
If the probability of an event is p then the odds of that event is given by p / 1-p
This is the odds of an event in a single group.
We can use odds in a medical setting but it is more common to use the odds ratio to compare two exposure groups.
It is a RELATIVE comparison of the odds of disease.
What are confounding variables? Why are these an issue?
When comparing groups, the association to effect between an exposure sand outcome is distorted by the presence of another variable.
Eg when looking at the association between gender and cancer of the mouth and pharynx, you must account for alcohol consumption because men are more likely to drink more alcohol and alcohol is a risk factor in cancer of the mouth and pharynx.
What do you do about confounding variables?
You can adjust for differences in known confounding factors. A common type of adjustment is standardisation. This form of analysis uses weighted averages to allow us to compare like for like. However, it is difficult to control confounding factors as many of them are unknown!
How do you conduct an ecological study?
-Identify groups to study
-Define the characteristics to be studied: the exposure and the outcome
-Decide whether analysis is to involve categorical or continuous data
-Gather data on the group-level characteristics
What issues are there with ecological studies?
-Definiton of characteristics
-Confounding variables -ecological fallacy (falsely inferring individual-level association from group-level association
What is an ecological study?
An ecological study is when the population is studied as groups rather than as individuals. They are often used to look at the prevalence or incidence of a disease -especially when a disease is rare.
What is a cross sectional survey?
This is a study that analyses data from a population at a representative subset at a specific point in time.
When conducting a cross sectional survey, what do you think about when generating your sample?
Who do you want to generalise to? (The theoretical population)
What population can you get access to? (The study population)
How can you get access to them? (The sampling frame)
Who is in your study? (The sample)
What specific issues are there when conducing a cross sectional survey?
What is a case control study?
This is retrospective study. You find cases and controls and you then compare and contract the potential causal factors if the disease.
How do you conduct a case control study?
Identify a set of cases and a suitable set of non cases (controls)
Ascertain previous exposure of everyone
Compare levels of exposure in cases and controls
What are the specific issues for case control studies?
Selection bias - controls should reflect the study population and be comparable to cases. It is often difficult to find controls with both these things, especially in hospitals.
Informationbias, especially for exposure -differential misclassification (this is the bigger problem of the two and is when both parties report different things occurring) and non differential misclassification (when both parties recap similar things but it is wrong).
What are cohort studies?
There ate two types of cohort studies: concurrent or prospective studies and historical or retrospective ones.
In a prospective study, you find people who have been exposed (or not) to a disease and then toy compare and contrast the two groups over time for potential outcomes.
`Whereas, in a retrospective study, you have found the exposed and unexposed people a while ago and you are now comparing and contrasting particular outcomes.
How do you conduct a cohort study?
Identify outcome free individuals
Group individuals according to level of exposure
Ascertain outcomes for everyone
Compare incidence rates for each exposure group
What issues are there when conducting cohort studies?
Loss to follow-ups (similar to misclassification) -inc differential loss and survivor bias (not everyone will finish the study, some people will leave)
Information bias, especially for outcome -inc differential and non differential misclassification
How do cohort studies account for people who do not finish the study?
They calculate the incidence rate. This is counts per person-time at risk. They disclose people who dod not finish the study or get the disease
What things do you think about when describing or finding a study?
What is a study design?
P What is the Population to be studied?
I What is the Intervention / exposure of interest?
C What us the Comparison / Control of interest?
O What is the Outcome of interest?
How are the results analysed?
What epidemiological study designs are descriptive?
Descriptive epidemiological study designs include:
-Ecological study (unit of analysis is groups)
-Cross-sectional survey (unit of analysis is individuals)
What epidemiological study designs are analytical?
-Case-control studies (analysis os only odds ratio so, it shows the relative risk)
-Cohort study (analysis can be rate or odds ratio)