1
Q
Definition of bias
A
Systematic error in the design, conduct or analysis of a study that results in a deviation from the truth
2
Q
Random deviation (error) definition
A
- RE’s are equally and randomly distributed across the comparator groups and are independent of exposure or disease status
- Tend to lead to a smaller observed effect size
- Deviates in different directions
- Not precise (deviating from actual value)
- -> The result of random error would be an imprecise estimate of person’s ____
3
Q
Definition of selection bias
A
Commonly defined in two ways:
- Error due to systematic differences in characteristics between those who are selected into a study and those who are not
- > Selection bias INTO A STUDY => the study population cannot be compared to the general population - Error due to systematic differences in characteristics between those who are selected into one of the groups within an analytical study (those who are selected into each comparison group within an alaytic study)
- > Selection bias into groups WITHIN an analytic study => the groups cannot be VALIDLY compared with each other
“Occurs when the way in which groups of subjects are selected for a study means that the groups cannot be validly compared with each other/ the general popn - diffs between baseline characteristics of he groups that are compared”
4
Q
Definition of sampling bias
A
- T ype of selection bias
- A systematic error that occurs when each potential member of the population being studied does not have an EQUAL chance of selection
- Error due to systematic differences in chara. between those who are selected into a study and those who are not
- Sampling bias and selection bias can sometimes be used interchangeably
5
Q
Definition of allocation bias
A
- Type of selection bias
- Term used in RCTs to refer to bias as a result of systematic differences in the chara. between those who are in one trial arm (study group) and those in another trial arm.
- Occurs when all participants do not have the same probability of entering any branch in the trial
- Bias in how parts. are assigned to treatment and comparison groups in clinical trial
6
Q
Definition of spectrum bias
A
- Type of selection bias
- Systematic error in diagnostic test accuracy studies that arises when specific groups of patients are inappropriately excluded - concerned with the INCLUSION of pts into a DTA study
- e.g. ‘difficult to diagnose’ patients
- Results in index test appearing more accurate than it actually is
7
Q
Definition of response/ responder/ participation bias
A
- Systematic error due to difference in chara. between those who choose to volunteer to participate in a study or who choose one study group compared to another, and those who do not
8
Q
Definition of loss to follow up/ follow up/ attrition bias
A
- Type of selection bias
- Systematic error due to the diff. in the chara. of those participants who were lost to follow up compared to those who completed the study
- Unequal loss of participants
9
Q
Definition of healthy worker effect
A
- Typically seen in observational (occupational epidemiology) studies of occupational exposures with improper choice of comparison group
- Less healthy workers are more likely to reduce their workplace exposures
- makes exposed workers appear healthier than control popn - using general popn wrong because it includes workers and non-workers
- describes the deficit of both morbidity and mortality when workers
and the general population are compared. - Workers usually exhibit lower overall death rates
than the general population because the severely ill and chronically disabled, who form part
of the generally population, are usually not in employment.
10
Q
Definition of information/ observation/ measurement bias
A
A systematic error in the way that data is gathered from the comparison groups that results in differential quality (Accuracy) of information between comparator groups
- The source of measurement bias is either the measurer or those being measured
“Occurs when the way in which the data are obtained from groups being compared differs systematically”
11
Q
Definition of interviewer bias
A
- Type of information bias
- Aspects of the interviewers and the way in which they ask questions and respond to answers
- error due to interviewers’ subconscious/conscious gathering of data that
differs systematically between cases and controls. If the interviewer is aware of whether the
person being questioned about exposure is a case rather than a control, they might probe
the cases more for evidence of exposure to the factor of interest.
12
Q
Definition of detection bias
A
- Type of information bias
- Systematic diff. between groups in how outcomes are determined / assessed (this could relate to ascertainment of outcome, diagnosis or verification of outcomes, for example)
- can occur in trials when groups differ in the way outcome info is collected or the way outcomes are verified
- can cause over or underestimate of size of effect
- need to blin - blinding of outcome assessors can be especially important for assessment of subjective outcomes, e.g. degree of post op pain
13
Q
Definition of performance bias
A
- Systematic diff between groups in the care that is provided, or in exposure to factors other than the intervention of interest
- This type of bias particularly affects RCTs
- one group gets more attention from investigators than another group
14
Q
Definition of reporting bias
A
- Type of information bias
- Systematic diff. between groups of reported and unreported findings
- Within a published report, those analyses with stats sig diffs between intervention groups are more likely to be reported than non-sig diffs
15
Q
Definition of verification bias
A
- Systematic error in diagnostic test accuracy studies that arises when not all the participants receive both the index test and the reference standard
- This is most likely to occur with patients with negative index test results
16
Q
Definition of review bias
A
- Systematic error in diagnostic test accuracy studies that arises when interpretation of the index test is not independent of the reference standard (and vice versa)
- This has the potential to make the index test appear more accurate that it actually is
17
Q
Definition of reflexive’ stance of researcher
A
- In qualitative studies, the explicit reflection of the researchers on how their background, professional role or personality, influences their relationship with the research participants, or interpretation of the data, and thus affect their research
18
Q
Definition of attrition bias
A
- Systematic diff. between groups in withdrawals from a study
- type of selection bias due to systematic differences between study groups in the number and the way participants are lost from a study
19
Q
Examples of selection bias
A
- Sampling bias
- Response/ responder/ participation bias
- Healthy worker effects
- Healthy reproducer effect
- Allocation bias
- Attrition/ loss to follow up bias
- Spectrum bias
20
Q
Examples of information bias
A
- Interviewer bias
- Recall bias
- Recording bias
- Social acceptability bias
- Detection bias
- Performance bias
- Reporting bias
- Verification bias
- Review bias
- Reflexive’ stance of researcher
21
Q
What is social acceptability bias?
A
- Form of response bias where people prefer to answer questions on how their answers will be interpreted by others rather than truthfully
- selective suppression or revealing of information, that relates to
the exposure of interest, so that the most socially acceptable answer is given.
22
Q
What is recall bias?
A
- Parts do not remember previous events or experiences accurately or omit details
- AKA response bias/ responder bias/ reporting bias
- systematic error due to differences in accuracy or completeness of recall of
prior exposure of interest between controls and cases.
23
Q
How can detection bias be prevented?
A
Prevent by blinding of outcome assessment
24
Q
How can performance bias be reduced?
A
Blinding the participants and people treating the participant
25
What are the main types of biases in RCTs?
1. Selection bias
2. Performance bias
3. Detection bias
4. Attrition bias
5. Reporting bias
26
What should we look out for to identify selection bias?
Is there:
1. Random sequence generation
2. Allocation concealment
27
What should we look out for to identify performance bias?
Is there
| 1. Blinding of participants and personnel
28
What should we look out for to identify detection bias?
Blinding of outcome assessment
29
What should we look out for to identify attrition bias?
Incomplete outcome data
30
What should we look out for to identify reporting bias?
Selection outcome reporting
| - Some outcomes have been measured but results not given
31
Definition of recording bias and when would you be most likely to see this?
- error that arises due to the information of interest being systematically
recorded in more detail by the cases than the controls.
- Usually in medical records
- Or in studies using self-completed questionnaires, as cases may be more meticulous in the providing info than healthy controls are
32
What are the two main types of bias
1. Information
2. Selection
