25 Flashcards
What is information bias?
“ Observation or information bias results from systematic differences in the way data on exposure or outcome are obtained from the various study groups”
Two ways data is collected in a study
- By participants
- Collected or measured by someone else
How can measurement error occur?
Participants provide inaccurate responses
• E.g. they may forget past exposures
• E.g. they may under or overestimate their exposure
Data is collected incorrectly/inaccurately
• E.g. problem with measuring device
• E.g. person collecting the data doesn’t follow the same procedure for all participants
Measurement error
can be random or systematic
What effect might measurement error have in a descriptive study?
• Could over/underestimate prevalence
What effect might measurement error have in a analytic study?
Can lead to misclassification
- People without the exposure may be classified as having the exposure (and vice versa)
- People without the outcome may be classified as having the
outcome (and vice versa)
Two types of misclassification
- Non-differential misclassification
- Differential misclassification
Non-differential misclassification
There is measurement error but it is‘ Not different’ between the study groups e.g. exposed/comparison group, or cases/controls
Differential misclassification
There is measurement error but it is ‘Different’ between the study groups e.g. exposed/comparison group,
or cases/controls
Non-differential misclassification can only move RR…
Closer to the null
Examples of differential misclassification
-In a cross-sectional study, people with the outcome might report the exposure differently to those without the outcome
-In a case-control study, cases might more accurately recall past exposures compared to controls
- In a case-control study, an interviewer who is aware they are interviewing a case might ask more probing questions about the exposure of interest
-In a cohort study, an interviewer aware of the exposure status may ask more probing questions about the outcome among those exposed compared with those in the comparison group
What kind of bias (information) is case control susceptible to?
Recall
what is recall bias
“ Systematic error due to differences in accuracy or completeness of recall to memory of past events or experiences”
What can recall bias to to the OR
- bring it towards or further from the null
How to minimise recall bias
- Objective measures
- Validate self-reported measures with other information
- Memory aids
Cohort studies what type of bias
Potential for misclassification of exposure and outcome exposure/outcomes
What kind of classification do we need to concider for cohort studies?
Have the participants been correctly classified
Has the outcome status been correctly classified
Cohort studies: Differential misclassification
• If classification of exposure depends on outcome
(BUT outcome not yet happened in a prospective cohort study; can be an issue in historical cohort studies)
• If classification of outcome depends on the exposure
E.g. if interviewer/observer knew the exposure status and examined the outcome differently for those in the exposed group compared with those in the comparison group
Interviewer/observer bias
How to Minimising interviewer/observer bias
- Clearly defined study protocol and measures
- Structured questionnaire and standard prompts
- Training of interviewers
- Blinding
RCT how called information bias occur and how can u minimise it?
• Bias could occur if knowledge of the treatment/exposure
category influences the assessment of the outcome
- Blinding
• Bias could occur if measurements are undertaken differently for different treatment groups
- Ensure measurements undertaken in the same way
How to minimise information bias when Collecting information from participants
Validated survey instruments
Validate using objective measure
How to minimise information bias with Measurement instruments
Use standardised equipment Use calibrated equipment
How to minimise information bias via Collecting information via interviewers/observers
Blinding
Use objective measures
Use structured interviews and standardised ‘prompts’
Training of interviewers
*Clearly defined study protocol
*Well-defined exposures, outcomes and other factors collected in the study
Publication bias
“The result of the tendency of authors to submit, organizations to encourage, reviewers to approve, and editors to publish articles containing ‘positive’ findings…”
Three properties of a potential confounder
- Independently associated with the outcome
- Independently associated with the exposure
- Not on the causal pathway
What can confounding do?
- Over-estimation of a true association
- Under-estimation of a true association
- change direction of a true association (risk factor becomes
protective) - give appearance of an association when there is not one (go
from null to something else)
Identifying potential confounders - if you don’t measure it, difficult to do anything about it later
Use literature to identify known and suspected risk factors for outcome (property 1)
Collect information on factors strongly associated with exposure, regardless if known risk factor (property 2)
Controlling confounding in the study design
- Randomisation
- Restriction
- Matching
Controlling confounding in the study design
- Randomisation
Design study to minimise confounding by selection and allocation of participants
Controlling confounding in the study design - Restriction
All attempt to make groups being compared alike with regard to potential confounder(s)
What kind of confounders does randomisation eliminate and what study? - design phase
- applies to known and unknown confounders
- RCT
Problems with using RCT to eliminate confounding
- Works best with large sample size
- Need equipoise
- Need intention-to-treat analysis
What does restiction mean? What kind of study designs?
Restrict sample to one stratum of potential confounde
Easy and can be applied to all study designs
Problems with restriction
- Can reduce generalisability
- Reduces number of potential participants
- Potential for residual confounding with imprecisely
measured (or broadly defined) confounders - Usually only one potential confounder
What is matching and what study can it be used in?
Choose people to make the control/comparison group have the same composition as the case/exposed group regarding the potential confounder.
Usually used in case-control studies
Individual vs Frequency matching
Individual: Each case matched with one or more controls having the same confounding variable characteristic(s)
Frequency: matching at aggregated level
Positives of matching
- Useful for difficult to measure/complex potential confounders
- Can improve efficiency of case-control studies with small numbers
Cons of matching
- Individual matching can be difficult and limit number of potential participants
- Need special matched analysis for individual matching (Otherwise will under-estimate the measure of association)
- Can’t assess association between potential confounder and
outcome - Can’t assess whether truly a confounder