Lecture 25: Selection Bias Flashcards
What is bias?
“…any systematic error in an epidemiologic study that results in an incorrect estimate of the association between exposure and risk of disease.”
What are the effects of selection bias and when is this important?
Findings could be:
* Over-estimated
* Under-estimated
* Not affected
Important when:
* Critically appraising scientific literature
* In evidence-based practice
* Considering studies reported in the media
* Undertaking research
What can we do about bias?
No magic wand….
* Selection and information bias can only be controlled during the design and data collection phases of a study
- So… at the planning stages of a study investigators must:
➢Identify the potential sources of bias in the proposed study
➢Identify possible ways to minimise these potential biases
What is selection bias?
“…occurs when there is a systematic difference between the people who are included in a study and those who are not, or when study and comparison groups are selected inappropriately or using different criteria”
How does selection bias arise?
- How do we recruit people into epidemiological studies?
- Does everyone agree to participate?
- Does everyone remain in the study?
How do we minimise loss to follow up?
➢Alternative contact details obtained at the start of the study
➢Maintaining regular contact e.g. study newsletters
➢Making several attempts to contact people
Selection bias and different study designs?
“While all analytic studies are subject to bias, each design has particular types to which it is inherently most vulnerable”
Selection bias for cross sectional studies?
Cross-sectional studies
* Exposures and/or outcomes assessed at one point in time
* Measuring prevalence
Considering selection bias:
➢Who entered the study?
➢Is the sample representative of the source population?
➢What is the response rate?
If the sample was not representative of the source population: Is the prevalence likely to be:
* Over-estimated?
* Under-estimated?
* Not affected?
Selection bias for case control studies?
Case control:
- Participants are selected on the basis of their outcome status.
- Exposure and outcome have already occurred
- Cases and controls are selected separately
Selection bias:
- If this selection is in some way dependant on their
exposure status then bias can occur
So…
Selection of cases and controls must be independent of the exposure status
What are some potential biases in selection of cases in a case control study? How are these controlled?
If the identification or participation of cases is related to exposure status there can be bias
Controlled by:
- Ensure high participation
- Clearly defined population of interest
- Reliable way of ascertaining all cases or a representative
sample of cases
What are some potential biases in selection of controls in a case control study? How are these controlled?
Recap:
what is the purpose of the control group in a case-control
study? … to provide an estimate of the odds of exposure in the population from which the cases came
How can selection bias in controls occur?
- If controls are not representative of the population which gave rise to cases (in place or time)
- If inclusion/exclusion criteria differ between cases and controls
- If non-participation of eligible controls
How can this be controlled?
- Ensure controls are from the same defined population as
the cases over the same time period
- Same inclusion and exclusion criteria for cases and controls
- Ensure high participation
What are three steps to consider for bias in the exam?
- Who does it affect?
- How does it affect them?
- What effect does it have on the measure of association?
- Is the measure of association under- or over-estimated?
- AND also specify the direction in relation to the null value:
i.e. has the measure of association been biased towards
or away from the null value? (or state whether it is biased numerically upwards or downwards)
How can selection bias arise in cohort studies?
Loss to follow up
* Study goes over time: people may not remain in the study
* If loss to follow-up is related to both the exposure and the outcome this can lead to bias
If comparison group selected separately from exposed group can lead to bias
How can selection bias occur in a RCT?
RCT randomly assigns participants to minimise selection bias however it can still arise if there is loss to follow up.
How can bias be controlled for in the analysis phase of a study?
It is not possible to control for bias in the analysis phase of a study
During the design phase of the study, the researchers will try and identify potential biases that may arise and
design the study in such a way as to avoid or minimise the bias.
If a particular bias is unavoidable, or has been identified once the data have been collected, the researchers
will attempt to assess the extent to which the bias may affect the results of the study (i.e. are the results biased towards/away from the null value?).
Researchers may also conduct sensitivity analyses (page 170 of the text book)
