week three Flashcards
Systematic Bias
Anything which inaccurately influences the conclusions in a study
Can be controlled methodologically
Decreases Validity
Random Bias
A type of error that happens by fluke occurrence
Can be controlled for by our sample size (power), statistical analyses
Affects Reliability
Confounding
Another factor that may be influencing the events. (aka taking into account external factors)
Causation
Values of one variable are casually associated with the changing of another variable
- Indicates one event is the result of the occurrence of another event
Association
values of one variable appear to be related to values of another variable
Randomisation
What is it and what does it minimise
Where participants are allocated/assigned to different groups based on chance alone (randomly)
- evenly distributes characteristics in groups
- allocation bias
- selection bias
Blinding
What is it and why is it important
Refers to a practise where study participants are prevented from knowing information that may somehow influence them and hence the results
Clinicians, outcome assessors, data collectors, data analytics
- Detection Bias
- Performance Bias
Intention to treat analysis
A way we can account for missing data
It allows us to overcome attrition bias
What are the types of bias
Before the experiment:
- Selection
- Sampling
During:
- performance
- Attrition
- Information/measurement
- Detection
- Recall
- Responce
What is selection bias, when does it occur and what are its impacts
When researchers manipulate the enrolment of participants into a study.
- Occurs if the researchers manipulate enrolment of participants in the trial
e. g. self-selected, or non-random methods
Impacts:
Unrepresentative of the population
Affects the external validity of the trial
what does selection bias include
sampling bias
How do we overcome selection/sampling bias
- Randomisation, if possible
- Explicit description in how participants were selected
- Ask if groups are similar at the beginning of the study in terms of baseline characteristics?
what is sampling bias
When one sample of the population is more likely to be included than another
Allocation Bias
What is it and what are its impacts
When the allocation of participants to groups is compromised. We want to make sure we conceal what groups people are assigned too
Impacts of studies with inadequate allocation:
- Leads to differences at baseline
- Have larger estimates of the effectiveness of the intervention
how to overcome allocation bias
– Randomisation
– Concealment of randomisation
Performance bias
What is it and why do we want to prevent it
When participants can identify which group they’ve been allocated to in a trial
Why do we want to prevent:
- Protects against patients in the control group seeking other forms of care
- Identify patients in the experimental group who experience a placebo effect
how do we control performance bias
blinding people to what they’ve got (usually using placebos)
Attrition bias
what is it and why is it important.
Relates to withdrawals, or drop-outs, from a study
Reasons for drop outs need to be established to identify:
– Missing data
– Adverse events
– Motivation
how do we overcome attrition bias
intention to treat analysis
information/measurement bias
An error in the measuring outcome
- For example patient reported data may be influenced if the participant knows what group they’re in
- For example an observer may not know how to use equipment properly
Detection bias
What is it and how to we control it
- An investigator may distort or misclassify the outcome measured if the particular group is known
(pushing arthritis example) - To control we blind the outcome assessor
Recall bias what is it
· Differences in accuracy or recollections of events/exposures from participants
-Bias is unintentional and often based on expectation
Responce bias
What is it, when does it occur and what does it affect
- To please the doctor or investigator (tells the investigator what you think they want to hear)
- Often occurs in patient or self reported data
- Affects internal validity of the study
