Exam Revision Flashcards
What is the appropriate measure to use in cohort studies?
Risk ratio
How do you calculate risk ratio?
Risk in exposed = a/a+b
Risk in unexposed = c/c+d
RR = Risk in exposed/risk in unexposed
How do you interpret risk ratios?
RR>1 - harmful effect
RR = 1 - no effect
RR<1 - positive effect
What would be required to fully interpret risk ratio?
95% confidence interval is required to determine if the result is statistically significant
What are the advantages of cohort studies?
- Temporality (exposure precedes the outcome)
- No recall bias involved
- Can study multiple outcomes which are associated with a rare exposure
- It is possible to estimate all measures of incidence (rate, risk) and effect (risk difference, risk ratio)
What are the disadvantages of cohort studies?
- Requires a large investment of time, human and financial resources
- Requires a large sample size
- Reproducibility is hard, unless someone has a similar study
- Loss to follow up - bias introduced is difficult to control for (selection bias)
- Inefficient for rare diseases
- Uncontrolled confounding (unmeasured day-to-day exposures)
- Information bias due to misclassification of exposure or outcome
What is an example of a prospective cohort study?
Framingham study (cardiovascular cohort study)
What are the advantages and disadvantages of a prospective cohort study?
Advantages
- They can be followed up to any point in the future
Disadvantages
- Time consuming
- Expensive
- Defined cohort so requires detailed exposure records
- Records must be maintained well for both exposure groups
What is an example of a retrospective cohort study?
Asbestos factory and mesothelioma
What are advantages and disadvantages of retrospective cohort studies?
Advantages
- Can identify a cohort/determine exposure and continue to follow up
- Faster answers
- Dont have to employ people for regular follow up so is therefore cheaper
Disadvantages
- Since exposure is determined after the fact, quality of records must be checked
What is an example of a case-control study?
Probability of lung cancer in smokers
What is the appropriate measure for case-control studies?
Odds ratio
How do you calculate odd ratio?
a x d/b x c
What are the advantages of a case-control study?
- Allows for the study of rare diseases (case control useful for identifying current cases and looking at historical data as opposed to doing a large cohort study which would be expensive)
- Design allows to look at multiple risk factors at once
- Can be helpful when disease outbreaks occur, and potential links and exposures need to be identified
What are the disadvantages of a case-control study?
- Selection bias which is when the selection of individuals, groups or data is in a way that proper randomisation has not occurred
- Information bias which occurs when there are errors in the measurement of characteristics and the consequences of the errors are different for different subjects
- Recall bias
- Observer bias in data collection but can be solved by ‘blinding’
- Misclassification bias
- Confounding
What are advantages of randomised controlled trials?
- Intervention and control groups will be similar in all respects except the intervention minimising selection bias and confounding
- If participants are ‘blind’ to the treatment allocation, reporting bias is minimised
- Also, if the investigators are ‘blind’ to the allocation, observer bias is minimised
- RCTs carry less risk of bias and confounding than other study designs and so can provide powerful evidence of causal relationship between the intervention and the outcome
- RCTs allow for multiple outcomes to be assessed
- RCTs allow for the incidence rate of the outcome to be measured
What are disadvantages of randomised controlled trials?
- Expensive to conduct as they may require a large study team, perhaps at several sites and may require a long follow up period
- In some situations, intervention studies are impossible to conduct for ethical or logistical reasons
- Recruitment is difficult and time consuming (trials can take up to years to do)
- Unequal at baseline
- Lack of blinding
- Loss to follow up
What is the difference between case-control and cohort study design?
Cohort:
1. Select population without disease and classify into exposed/unexposed
2. Measure disease incidence in exposed v unexposed
Case-control:
1. Select cases (diseased) and controls (undiseased)
2. Measure exposure in cases and controls
Define cross-sectional studies and their design?
They estimate frequency or outcome at a particular point in time
Either descriptive or analytical
DESCRIPTIVE = describe frequency of exposure in a defined population
ANALYTICAL = simultaneously collect information on both the outcome and potential risk factors in a defined population. Then compare the prevalence of the outcome in the people exposed to each risk factor with the prevalence in those not exposed
What are disadvantages of cross-sectional studies?
- Selection bias
- Information bias (recall bias)
- Measures prevalence rather than incidence cases which are of limited value for investigating aetiological relationships
- Can be difficult to establish the time-sequence of events in a cross-sectional study, the exposure may have occurred as a result of the outcome (reverse causality)
- Not good for studying rare diseases
- Unable to investigate the temporal relation between outcome and risk factors
What are the advantages of cross-sectional studies?
- Easy, relatively quick and economical
- No ethical difficulties
- Provides important information on the distribution of burden of exposures and outcome relationship
- Can be used as the first step in the study of a possible exposure-outcome relationship
- Data on all variables are only collected at one time point
- Multiple outcomes and exposures can be studied
- Easy for generating hypotheses
Write short notes on ecological fallacy?
- Ecological studies enable us to make ecological inferences about effects at a group level but do not enable us to make inferences about individual risk
- An ecological fallacy is a logical error which occurs when characteristics of a group are attributed to an individual
- Ecological fallacy can lead to a false or inaccurate conclusion about social phenomena and individuals within this social phenomena
- There are ways to avoid ecological fallacy including careful study design, statistical methods and critical thinking
- An example of when ecological fallacy occurs is in relation to break cancer and high fat consumption. It may be true that countries with higher levels of fat consumption have higher rates of breast cancer but that isnt to say that individuals who consume high contents of fat are more likely to have breast cancer
Write short notes on incidence risk and incidence rate
- There are two ways of measuring incidence: risk and rate
- Incidence risk is the number of new cases in interval divided by the population initially at risk
-Incidence rate is the number of new cases divided by the total person-time at risk - Which measurement to use depends on the study population. Incidence risk is good for static populations where as incidence rate is good for dynamic populations
- Incidence is different from prevalence as incidence describes the number of new cases in a defined period where as prevalence describes the number of existing cases
- Annual incidence describes the number of new cases in a calendar year
- Cumulative incidence refers to the frequency of new cases over a specific time period
Write short notes on prevalence?
- P = I X D (d being duration of disease)
- Prevalence is the number of existing cases in a population at a designated time
- Point prevalence is the proportion of people in a defined population that has the outcome under study at a specific point in time
- Period prevalence is the proportion of people in a defined population that has the outcome under study over a period of time
- Useful for assessing burden of disease within a population
- Valuable for planning
- Not useful for determining what caused the disease
Write short notes on selection bias in surveys?
- Systematic differences in characteristics between those who take part in a study and those who dont, such that the people who are selected to participate in a study are not fully representative of the reference population, or in analytical studies, the comparison groups are not comparable
- In case-control studies, cases selected for the study are not representative of all eligible cases and controls are not representative of the population which produced the cases
- Bias in selection of cases comes from the cases who have been exposed to the exposure of interest are more (or less) likely to be selected for the study compared with unexposed cases
- Bias in selection of controls comes from controls chosen not being representative of the population which produced the cases. The control group should be such that if a control had had the outcome of interest, they would have the potentially been included as a case
- Selection bias in cohort studies can come from the comparison groups not being truly comparable, this can come from poor choice of unexposed group, differences in follow up between comparison groups and missing data
- Selection bias in cross sectional studies can come from individuals who take part having different exposures e.g. alcohol and drug studies - non-responders are greatest exposure
Write short notes on information bias in surveys?
- Information bias is any error in the measurement of exposure or outcome that results in systematic differences in the accuracy of information collected between comparison groups
- There are 2 main types: reporting bias and observer bias
- Reporting bias is when subjects decision to report depends on the direction and magnitude of the findings. In other words, they give the answer they think will please the investigator, or they conceal potentially embarrassing information
- Reporting bias can be avoided or minimised by use of exposure data before outcome, objective data sources where possible, and keep subjects unaware of association under study
- In case control and cross sectional studies, observer bias occurs when the accuracy of exposure data recorded by the investigator differs systematically between subjects in different outcome groups
- In cohort or interventional studies, when the accuracy of outcome data recorded by the investigator differs systematically between subjects in different exposure groups
- Can be avoided in case control studies by ‘blinding’ - this is where the participants and/or observers are not aware of the exposure or outcome to reduce bias
- Can be avoided in cohort or intervention studies, by the people who are responsible for classifying the outcome do not know the subjects ‘exposure’ category
What is cluster randomisation?
This is when pre-existing groups (or clusters) are randomly allocated to the study.
What is stepped wedge design?
This design involves random and sequential crossover of clusters from control to intervention until all clusters are exposed
What is standardisation?
- Often used to control confounding effects of age so that rates of disease or mortality can be compared in populations with different age structures
How do you carry out direct standardisation?
Obtain an age standardised rate which adjusts for the effect of age
How do you carry out indirect standardisation?
Obtain the standard mortality ratio (SMR)
How do you calculate standard mortality ratio (SMR)?
Observed deaths/expected deaths X 100
Advantages and disadvantages of direct standardisation?
Advantages
- In small, subpopulations INDIRECT IS PREFERRED
- Less bias
- Better when comparing two + groups with different age distributions
Disadvantages
- Needs age specific rates for study populations and these are not always available or reliable
Advantages and disadvantages of indirect standardisation?
Advantages
- Only need total observed cases/deaths in study population and its age structure
Disadvantages
- SMRs can only be compared between populations with similar age structure
- SMR depends on age distribution of study
Write short notes on intention to treat analysis?
Compares outcomes for all randomised individuals - even if they stop taking treatment or drop out of the study
Assesses the overall effect of assigning a subject to receive a particular intervention
Analysis is the most important analysis as intervention and control groups compared as originally randomised
More likely to underestimate treatment effect
Often use a modified ITT analysis - all randomised participants with some follow up data
