Research Methods - Representing risk Flashcards
What is incidence?
Incidence measures the frequency of “new onset cases” of a disease (/condition/outcome) (E.g. COVID-19) in a given population over a designated period of time (e.g. 12 months)
Incidence is expressed as a … of a … which is at risk
Incidence is expressed as a proportion of a population which is at risk
Incidence is the number of people who develop a disease over …
number of people who develop disease over given time
What it person-time?
Person years (or person months) are measurements based on the number of people in the study and the amount of time each person spends in the study
Person-time assumes disease probability during study period is …, so that 10 persons/1 year = 1person/10 years
Person-time assumes disease probability during study period is constant, so that 10 persons/1 year = 1person/10 years
Person-time is often invalid - why?
As risk of many chronic diseases increases with age
What is the relationship between incidence and prevalence? (Assuming all other factors e.g. disease prognosis stay constant)
As incidence increases, prevalence increases
Prevalence = …. x (duration of disease)
Prevalence = Incidence x (duration of disease)
The epidemiologist’s bathtub
The relationship between prevalence and incidence
How we might we find data to measure incidence?
Cohort study - groups of people who have something in common
Followed up of over time to see what happens to them
What is a cohort study?
Group of individuals free from a disease or condition is selected, ideally at random
Typically an exposure of interest - participants selected into an exposed and a non-exposed group
In a cohort study, no allocation/manipulation of exposure by the researcher - but groups may be matched on potential … factors
Confounding factors i.e. influential factors not of interest
How is a cohort study conducted?
The groups are followed up over the given period of time - occurrences of disease onset are recorded
At the end of the study, comparisons are made between the non/exposed groups and their relative numbers of new disease occurrences
What is a cohort study - Prospective vs Retrospective
Prospective = exposed and non-exposed groups, outcomes followed up in the future Retrospective = Historical exposed and non-exposed groups, outcomes followed over defined historical period
… = exposed and non-exposed groups, outcomes followed up in the future …. = Historical exposed and non-exposed groups, outcomes followed over defined historical period
What type of epidemiological study is a cohort study? Descriptive or analytic?
Analytic (tests hypothesis, identify and quantify risk or exposure)
… cohort = secondary data only
Retrospective cohort = secondary data only
… cohort = any/all data sources as relevant
Prospective cohort = any/all data sources as relevant
Cohort study data: Secondary (mortality register, hospital/medical records, census data)
Pros vs Cons
+ Cheap
+ If anonymous, minimal ethical/governance approval needed
- Limited by what data already gathered
- Poor accuracy and missing data
Cohort study data: Primary (survey data)
Pro vs cons
+ Gather additional data
- Difficult to achieve representative sample
- More expensive
Strengths of the cohort study (4)
More than one disease (outcome) related to single exposure
Can offer some evidence of cause - effect relationship through establishing temporal precedence
Good when exposure is rare
Can calculate incidence rate and risk
Weaknesses of the cohort study:
Potential for losses to follow-up (attrition may differ in exposed versus unexposed/disease versus non, the longer the follow-up the greater the risk of attrition)
Often requires large sample, long duration, expensive
Less suitable for rare diseases
Less suitable for diseases with long latency
Does not eliminate confounding
Calculating … = number of people who develop disease over a given time / number of people at risk of developing disease over given time
Calculating incidence = number of people who develop disease over a given time / number of people at risk of developing disease over given time
Measuring incidence - relative risk:
A clear advantage of the cohort study is that we can compare incidence between the two groups i.e. to test hypotheses about associations between exposures and outcomes
Relative risk or risk ratio (RR) is the risk of developing a disease in … group compared to developing a disease in …. group
A clear advantage of the cohort study is that we can compare incidence between the two groups i.e. to test hypotheses about associations between exposures and outcomes
Relative risk or risk ratio (RR) is the risk of developing a disease in exposed group compared to developing a disease in unexposed group
… = incidence of disease among exposed / incidence of disease among non-exposed
Relative risk
Cohort study data analysis: Relative risk
RR of 1.5 means the risk of outcome is … higher in the … group than the … group
Risk of outcome 50% higher in exposed than unexposed group
RR of 3.0 means the risk of outcome is … higher in the … group than the … group
RR of 3.0 means the risk of outcome is 3 times higher in the exposed group than unexposed group
RR of 0.8 means the risk of outcome is … higher in the … group than the … group
Risk of outcome 20% lower in exposed than unexposed group
Calculating incidence - disease incidence
Incidence of disease = (…+…)/N
Incidence of disease in exposed = …/(…+….)
Incidence of disease in unexposed = …/(…+…)
Incidence of disease = (A+C)/N
Incidence of disease in exposed = A/(A+B)
Incidence of disease in unexposed = C/(C+D)
Relative risk of 1.94 for risk of tuberculosis for homeless versus non-homeless people in London - what do we conclude?
Homeless people in London have a greater risk of developing tuberculosis
The same relative risk can represent completely different absolute differences in how many people are affected by a disease/die -
depending on absolute prevalence in the population
Rare disease/outcome = … RR not important for public health (important to the individual)
Common disease/outcome = … RR clinically important in public health
Researchers should report both RR and absolute risk
The same relative risk can represent completely different absolute differences in how many people are affected by a disease/die -
depending on absolute prevalence in the population
Rare disease/outcome = large RR not important for public health (important to the individual)
Common disease/outcome = small RR clinically important in public health
Researchers should report both RR and absolute risk
…. …. - precision of our estimates e.g. for people in London, homelessness is associated with relative risk of contracting TB of 1.94 - how confident are we that this 1.94 is precise? -
The confidence interval describes the range of values with a given probability (e.g. 95%) that the true value of a variable is contained within that range.
The … … describes the range of values with a given probability (e.g. 95%) that the true value of a variable is contained within that range.
The confidence interval describes the range of values with a given probability (e.g. 95%) that the true value of a variable is contained within that range.
Relative Risk or Risk Ratio (RR) is the risk of developing a disease in … group compared to developing a disease in … group
Relative Risk or Risk Ratio (RR) is the risk of developing a disease in exposed group compared to developing a disease in unexposed group
Confidence intervals = a way of expressing the precision of …
Compared to the likelihood ‘…’ value in the population
Confidence intervals = a way of expressing the precision of estimates
Compared to the likelihood ‘true’ value in the population
…. study = select exposed and unexposed individuals at outset and measure incidence of outcome over given period of time
Cohort study = select exposed and unexposed individuals at outset and measure incidence of outcome over given period of time
Pros of cohort study
Advantages – can establish temporal precedence of exposure to outcome, suitable for rare exposures
Cons of cohort study
Disadvantages – does not eliminate confounding, less suitable for rare diseases/outcomes
