Scribd
Upload
114 views30 pages

Errors in Epidemiological Studies

This document discusses various types of errors that can occur in epidemiological studies, including random error, systematic error (bias), and confounding. It provides examples of selection bias, information bias, and confounding. Selection bias can result from non-representative sampling or differences between study groups. Information bias occurs when exposures or diseases are inaccurately measured. Confounding is when a third variable is associated with both the exposure and disease being studied.

Uploaded by

FYM
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
114 views30 pages

Errors in Epidemiological Studies

This document discusses various types of errors that can occur in epidemiological studies, including random error, systematic error (bias), and confounding. It provides examples of selection bias, information bias, and confounding. Selection bias can result from non-representative sampling or differences between study groups. Information bias occurs when exposures or diseases are inaccurately measured. Confounding is when a third variable is associated with both the exposure and disease being studied.

Uploaded by

FYM
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 30

Errors in epidemiological studies

What can go wrong ??


• sampled from the target population ?
• sample large enough ?
• measured precisely ?
• measured accurately ?
• disease correctly diagnosed ?
• absence of disease correctly diagnosed ?
• exposure accurately determined ?
• measurement error same in different groups ?
• groups comparable ?
• Errors can be differential (systematic) or
non-differential (random)
• Random error: use of invalid outcome
measure that equally misclassifies cases
and controls
•
• Differential error: use of an invalid
measures that misclassifies cases in
one direction and misclassifies controls in
another
• Term 'bias' should be reserved for
differential or systematic error
Types of Error

• ‘random error’
= imprecision
• ‘systematic error’
= bias unbiased + unprecise unbiased + precise

biased + unprecise biased + precise


Random error is the
divergence due to chance
alone, of an observation on
sample from the true
population value, leading to
lack of precision in
measurement of association
Random error
• Sources of Random Error
– Sampling error
– Biological variation
– Measurement error
Sample Size Calculations
Variable to consider
– Required level of statistical significance of the 
expected result
– Acceptable chance of missing the real effect
– Magnitude of the effect under investigation
– Amount of disease in the population
– Relative sizes of the groups being compared
What is Bias?
• *Any trend in the collection, analysis, 
interpretation, publication or review of data 
that can lead to conclusions that are 
systematically different from the truth (Last, 
2001)
• *A process at any state of inference tending to 
produce results that depart systematically from 
the true values (Fletcher et al, 1988)
• *Systematic error in design or conduct of a study 
(Szklo et al, 2000) 
What is meant by bias in
research?

• Bias is the term used to


describe differences
between the study findings
and truth
Bias is a systematic error in  an 
epidemiologic  study  that 
results  in  an  incorrect 
estimation  of  the  association 
between  exposure  and  health‐
related event
Systematic Error (Bias)
• Occurs when there is a tendency to produce results 
that differ in a systematic manner from the true 
values 
• A study with a small systematic error is considered 
highly accurate
• Accuracy is not affected by sample size
• Principal biases
– Selection Bias
– Measurement (Classification) Bias
• Confounding
Types of Bias
Selection bias
• Unrepresentative nature of sample 
Information (misclassification) bias
• Errors in measurement of exposure of disease
Confounding bias
• Distortion of exposure ‐ disease relation by some 
other factor
• Types of bias not mutually exclusive
(effect modification is not bias)
Selection Bias

• Occurs when there is a systematic difference 
between the characteristics of the people 
selected for a study and the characteristics of 
those who are not
• Distortion of effect resulting from the way 
participants are accepted into studies
• Healthy Worker Effect – risk for certain illnesses 
in industrial working populations is lower than in 
the general population
Selection Bias

• Selective differences between comparison 
groups that impacts on relationship 
between exposure and outcome
• Usually results from comparative groups 
not coming from the same study base and 
not being representative of the 
populations they come from
Selection Bias Examples
• Case‐control study:
• Controls have less potential for exposure than 
cases
• Outcome = brain tumor; 
• exposure = overhead high voltage power lines
• Cases chosen from province wide cancer registry
• Controls chosen from rural areas
• Systematic differences between cases and 
controls 
Selection Bias Examples 
Self‐selection bias:
• You want to determine the prevalence of 
HIV infection
• You ask for volunteers for testing
• You find no HIV
• Is it correct to conclude that there is no 
HIV in this location?
Selection Bias Examples
• Healthy worker effect:
• Another form of self‐selection bias
• “self‐screening” process – people who are unhealthy 
“screen” themselves out of active worker population
• Example: 
• ‐ Course of recovery from low back injuries in 25‐45 year 
olds
• ‐ Data captured on worker’s compensation records
• ‐ But prior to identifying subjects for study, self‐selection 
has already taken place
Selection Bias Examples
• Also occurs before subjects are identified for study
• Diagnostic or workup bias:
• Diagnoses (case selection) may be influenced by physician’s 
knowledge of exposure
• Example:
• ‐ Case control study – outcome is pulmonary disease, 
exposure is smoking
• ‐ Radiologist aware of patient’s smoking status when reading 
x‐ray – may look more carefully for abnormalities on x‐ray 
and differentially select cases
• Legitimate for clinical decisions, inconvenient for research
Sources of Selection Bias
• Volunteers for studies are almost always 
selective
• Paid participants may be selectively different 
from the general population
• Hospital and clinical data are based on a 
selective population
• Disease or factor under investigation makes 
people unavailable for study
Measurement Bias
• Occurs when individual measurements or 
classifications of disease or exposure are 
inaccurate
• Sources:
– Quality of laboratory analysis
– Recall bias
• Information / Measurement / 
Misclassification Bias
• Method of gathering information is inappropriate 
and yields systematic errors in measurement of 
exposures or outcomes 
• If misclassification of exposure 
(or disease) is unrelated to disease 
(or exposure) then the misclassification
is non‐differential
• If misclassification of exposure (or disease) is related 
to disease (or exposure) then the misclassification is 
differential
• Distorts the true strength of association
Sources of information bias 

• Subject variation
• Observer variation
• Deficiency of tools
• Technical errors in measurement 
•
Information / Measurement / 
Misclassification Bias
• Recall bias: 
• Those exposed have a greater sensitivity for
recalling exposure (reduced specificity)
• - specifically important in case-control studies
when exposure history is obtained retrospectively
• cases may more closely scrutinize their past
history looking for ways to explain their illness
• - controls, not feeling a burden of disease, may
less closely examine their past history
Information / Measurement /
Misclassification Bias
• Reporting bias: 
• Individuals with severe disease tends to have 
complete records therefore more complete 
information about exposures and greater 
association found
• Individuals who are aware of being 
participants of a study behave differently 
(Hawthorne effect)
Confounding
• Occurs when another exposure exists in the 
study population and is associated with both 
the disease and the exposure being studied
• When the effects of two exposures (risk 
factors) have not been separated, and 
incorrect conclusions are drawn that the 
effect is due to one rather than the other 
variable
Confounding

• Effect of a third variable that accounts for all or


some of the association between exposure and
disease: Confounder

Exposure Disease
Definition of Confounding

A confounder:
1) Is associated with exposure
2) Is associated with disease
3) Is NOT a consequence of exposure (i.e. not 
occurring between exposure and disease)

Conf.
Exposure Disease
Confounding

• Confounder is not a result of the exposure
– e.g., association between child’s birth rank 
(exposure) and Down syndrome (outcome); 
mother’s age a confounder?
– e.g., association between mother’s age 
(exposure) and Down syndrome (outcome); birth 
rank a confounder?
Confounding

• To be a confounding factor, two conditions


must be met

• Be associated with exposure


- without being the consequence of
exposure
• Be associated with outcome
- independently of exposure (not an
intermediary)
Controlling for confounders
– RESTRICTION of subjects according to potential 
confounders (i.e. simply don’t include 
confounder in study)
– RANDOM ALLOCATION of subjects to study 
groups to attempt to even out unknown 
confounders
– MATCHING subjects on potential confounder 
thus assuring even distribution among study 
groups

You might also like