Causal Inference, Bias, Confounding, Interaction Flashcards
Describe two errors found in studies
- Random error
- Systematic error
Define a random error and an example
Inherent errors which fluctuate due to the unpredictablity or uncertainty in measuring process.
Measuring height affected by minor posture changes
Define systematic errors and give an example
Error that is predicatble and constant or proportional to the true value. Can be eliminated if identified
taring a balance
How can you minimse random errors?
- Repeats
- Lots of Data
- Averages to recude errors
- Statistically analyse
How can you reduce systematic errors
- Careful design
- Comparison of results to independent groups
Difficult to detect and correct
Cannot statistically analyse (as all values are ‘off’ in the same direction i.e. all too high or too low than true value)
Define Bias
Systematic errors in the design or conduct of study that results in incorrect estimate of association between exposure and ourcome
Describe the two main subtypes of bias
Selection Bias
Information Bias
Define selection bias
Results from selection of participants that is not representative of target population
- If the right people don’t participate then the associations observed are not representative to the real world*
- Systematic difference between the characteristics of those selected for the study and those who are not*
Give some examples of selection bias
Self-selection bias (volunteers)
Non-responder
Survival bias
Ending study early when outcomes have been achieved
Prevalence is influenced by _____ and _____ with disease; whilst incidence is influenced by _____ only.
Prevalence is influenced by exposure and survival with disease; whilst incidence is influenced by exposure only.
Define Survival bias
Bias that results from participants surviving the study period - can only be selected and analysed if they survive to study end point
Often seen in cross-sectional studies
Differentiate between point and period prevalence
Point: Prevalence of a disease at one point in time
Period: Prevalence of a disease over a period of time
What are some biases found in cohort studies
Self-selection: more likely to participate if have disease
Loss to follow up (20% max attrition)
Define Hazard Ratio
Hazard in the intervention group ÷ Hazard in the control group
Measure of an effect of an intervention on an outcome of interest over time.
Hazard ratio is reported most commonly in time-to-event analysis or survival analysis (i.e. when we are interested in knowing how long it takes for a particular event/outcome to occur).
How is a hazard ratio different to a Rsik or Odds ratio
Hazard ratio takes into account the total events and the time they occured
RR or OR only looks at the occurence of an event by the end of the study (i.e. ratio of whether or not an event occured:number of events)
