3. Bias and Experimental Studies Flashcards
Define Bias.
Systematic deviation of results or inferences from truth.
What is selection bias?
Selection bias is the bias introduced by the selection of individuals, groups or data for analysis in such a way that proper randomization is not achieved, thereby ensuring that the sample obtained is not representative of the population intended to be analyzed. It is sometimes referred to as the selection effect.
Questions to be considered when selecting a population:
- What population were the sample meant to represent, do they represent it?
- How were they selected?
What is information bias?
It arises from measurement error and is defined as: “A flaw in measuring exposure, covariate, or outcome variables that results in different quality (accuracy) of information between comparison groups. The occurrence of information biases may not be independent of the occurrence of selection biases.”
What are the two types of information bias?
Differential and non-differential.
What is differential information bias?
Measures that are more likely to be incorrect for some respondents than others.
What is non-differential bias?
Measurements are randomly incorrect.
What is confounding bias?
The distortion of a measure of the effect of an
exposure on an outcome due to the association of
the exposure with other factors that influence the
occurrence of the outcome.
- Associated with exposure
- Associated with outcome
- Not on the causal pathway
Could you summarise biases?
• Bias is systematic error often caused by:
– Selecting respondents who do not accurately represent
the population of interest
– Collecting information that is incorrect
– Confounding the effect of the exposure of interest with
the effect of another exposure
How can we address selection and information bias?
We ask ourselves…
– What is our population of interest ? (PICO)
– Are there systematic differences between this
population and our study sample?
– Are we measuring exposures and outcomes
appropriately?
How do we address confounding bias?
- Through analysis: using statistical methods, provided all
confounders are measured properly. (“we adjusted for
socio-economic status…”) - By carrying out experimental studies called Randomised
Controlled Trials (RCTs)
Why do we use RCT?
To address known and unknown confounders.
Why do we randomise?
Randomisation helps to create intervention and control
groups that are similar in every way, except that one group
gets the intervention
This happens because each person or cluster of persons has
an equal chance of being in in the intervention or control arm:
that chance is decided randomly, not by doctors or politicians
If your sample is large enough, both known and unknown
confounders will be equally distributed between arms
What are two important factors of RCTs besides randomization and confounding bias eradication.
Blinding to allocation (intervention or control):
– Of assessors – minimise measurement bias
– Of analysts – minimise temptation to fudge results
– Of intervention workers (not always possible)
Reporting loss to follow-up
