Bias, Confounding and Fallacies in Epidemiology

Bạn đang xem bản rút gọn của tài liệu. Xem và tải ngay bản đầy đủ của tài liệu tại đây (1.01 MB, 67 trang )

Trang 1<div class="page_container" data-page="1">

Bias, Confounding and Fallacies in Epidemiology

M. Tevfik DORAK

o/epi

</div>Trang 3<div class="page_container" data-page="3">

Bias is one of the three major threats to internal

</div>Trang 4<div class="page_container" data-page="4">

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 Bias?

</div>Trang 5<div class="page_container" data-page="5">

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

Bias is systematic error

</div>Trang 8<div class="page_container" data-page="8">

Chance vs Bias

Chance is caused by random errorBias is caused by systematic error

Errors from chance will cancel each other out in the long run (large sample size)

Errors from bias will not cancel each other out whatever the sample size

Chance leads to imprecise resultsBias leads to inaccurate results

</div>Trang 9<div class="page_container" data-page="9">

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)

This classification is by Miettinen OS in 1970sSee for example Miettinen & Cook, 1981 (www)

Types of Bias

</div>Trang 10<div class="page_container" data-page="10">

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

</div>Trang 11<div class="page_container" data-page="11">

Selection Bias Examples

</div>Trang 12<div class="page_container" data-page="12">

Selection Bias Examples

</div>Trang 13<div class="page_container" data-page="13">

Selection Bias Examples

</div>Trang 14<div class="page_container" data-page="14">

Selection Bias Examples

</div>Trang 15<div class="page_container" data-page="15">

Selection Bias Examples

Selective survival (Neyman's) bias

</div>Trang 16<div class="page_container" data-page="16">

Selection Bias Examples

Case-control study:

Controls have less potential for exposure than casesOutcome = brain tumour; exposure = overhead high voltage power lines

Cases chosen from province wide cancer registryControls chosen from rural areas

Systematic differences between cases and controls

</div>Trang 17<div class="page_container" data-page="17">

Case-Control Studies: Potential Bias

Schulz & Grimes, 2002 (www) (PDF)

</div>Trang 18<div class="page_container" data-page="18">

Selection Bias Examples

Cohort study:

Differential loss to follow-up

Especially problematic in cohort studies

Subjects in follow-up study of multiple sclerosis may differentially drop out due to disease severity

Differential attrition  selection bias

</div>Trang 19<div class="page_container" data-page="19">

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?

</div>Trang 20<div class="page_container" data-page="20">

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 populationExample:

- 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

</div>Trang 21<div class="page_container" data-page="21">

Selection Bias Examples

Diagnostic or workup bias:

Also occurs before subjects are identified for studyDiagnoses (case selection) may be influenced by physician’s knowledge of exposure

- 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 casesLegitimate for clinical decisions, inconvenient for research

</div>Trang 22<div class="page_container" data-page="22">

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

</div>Trang 23<div class="page_container" data-page="23">

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

Distorts the true strength of association

</div>Trang 25<div class="page_container" data-page="25">

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 retrospectivelycases 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

Those who develop a cold are more likely to identify the exposure than those who do not – differential misclassification

- Case: Yes, I was sneezed on

- Control: No, can’t remember any sneezing

</div>Trang 26<div class="page_container" data-page="26">

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)

</div>Trang 27<div class="page_container" data-page="27">

Controlling for Information Bias

- Blinding

prevents investigators and interviewers from knowing case/control or exposed/non-exposed status of a given participant

- Form of survey

mail may impose less “white coat tension” than a phone or face-to-face interview

- Questionnaire

use multiple questions that ask same information acts as a built in double-check

- Accuracy

multiple checks in medical records

gathering diagnosis data from multiple sources

</div>Trang 28<div class="page_container" data-page="28">

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

</div>Trang 30<div class="page_container" data-page="30">

Cases of Down Syndrome by Birth Order

</div>Trang 31<div class="page_container" data-page="31">

Cases of Down Syndrome by Age Groups

</div>Trang 32<div class="page_container" data-page="32">

Cases of Down Syndrome by Birth Orderand Maternal Age

</div>Trang 33<div class="page_container" data-page="33">

• A third factor which is related to both

exposure and outcome, and which accounts for some/all of the observed relationship

between the two

• Confounder 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

</div>Trang 34<div class="page_container" data-page="34">

Exposure Outcome

Third variable

To be a confounding factor, two conditions must be met:

Be associated with exposure

- without being the consequence of exposure

Confounding

</div>Trang 35<div class="page_container" data-page="35">

Birth Order Down Syndrome

Maternal Age

Maternal age is correlated with birth order and a risk factor even if birth order

is low

</div>Trang 36<div class="page_container" data-page="36">

Birth Order

Down SyndromeMaternal Age

Confounding ?

Birth order is correlated with maternal age but not a risk factor in younger mothers

</div>Trang 37<div class="page_container" data-page="37">

Coffee CHD

Smoking is correlated with coffee

drinking and a risk factor even for those who do not drink coffee

</div>Trang 38<div class="page_container" data-page="38">

Coffee

CHDSmoking

Confounding ?

Coffee drinking may be correlated with smoking but is not a risk factor in

</div>Trang 39<div class="page_container" data-page="39">

non-Alcohol Lung Cancer

Smoking is correlated with alcohol consumption and a risk factor even for

those who do not drink alcohol

</div>Trang 40<div class="page_container" data-page="40">

Smoking CHD

Yellow fingers

Not related to the outcome

Confounding ?

</div>Trang 42<div class="page_container" data-page="42">

Imagine you have repeated a positive finding of birth order

association in Down syndrome or association of coffee drinking with CHD in another sample. Would you be able to replicate it? If not why?

Imagine you have included only non-smokers in a study and examined association of alcohol with lung cancer. Would you find an association?

Imagine you have stratified your dataset for smoking status in the alcohol - lung cancer association study. Would the odds ratios differ in the two strata?

</div>Trang 43<div class="page_container" data-page="43">

Imagine you have repeated a positive finding of birth order

association in Down syndrome or association of coffee drinking with CHD in another sample. Would you be able to replicate it? If not why?

You would not necessarily be able to replicate the original finding because it was a spurious association

due to confounding.

In another sample where all mothers are below 30 yr, there would be no association with birth order.

In another sample in which there are few smokers, the coffee association with CHD would not be

replicated.

</div>Trang 44<div class="page_container" data-page="44">

Imagine you have included only non-smokers in a study and examined association of alcohol with lung cancer. Would you find an association?

No because the first study was confounded. The

association with alcohol was actually due to smoking. By restricting the study to non-smokers, we have

found the truth. Restriction is one way of preventing confounding at the time of study design.

</div>Trang 45<div class="page_container" data-page="45">

Imagine you have stratified your dataset for smoking status in the alcohol - lung cancer association study. Would the odds ratios differ in the two strata?

The alcohol association would yield the similar odds ratio in both strata and would be close to unity. In confounding, the stratum-specific odds ratios should be similar and different from the crude odds ratio by at least 15%. Stratification is one way of identifying

confounding at the time of analysis.

If the stratum-specific odds ratios are different, then

</div>Trang 46<div class="page_container" data-page="46">

If the smoking is included in the statistical model, the alcohol association would lose its statistical

significance. Adjustment by multivariable modelling is another method to identify confounders at the time of data analysis.

Imagine you have tried to adjust your alcohol association for smoking status (in a statistical model). Would you see an

association?

</div>Trang 47<div class="page_container" data-page="47">

For confounding to occur, the confounders should be differentially represented in the comparison groups.

Randomisation is an attempt to evenly distribute

potential (unknown) confounders in study groups. It does not guarantee control of confounding.

Matching is another way of achieving the same. It ensures equal representation of subjects with known confounders in study groups. It has to be coupled with matched analysis.

Restriction for potential confounders in design also prevents confounding but causes loss of statistical power (instead stratified analysis may be tried).

</div>Trang 48<div class="page_container" data-page="48">

Randomisation, matching and restriction can be tried at the time of designing a study to reduce the risk of

confounding.

At the time of analysis:

Stratification and multivariable (adjusted) analysis can achieve the same.

It is preferable to try something at the time of designing the study.

</div>Trang 49<div class="page_container" data-page="49">

Effect of randomisation on outcome of trials in acute pain

</div>Trang 50<div class="page_container" data-page="50">

Obesity Mastitis

In cows, older ones are heavier and older age increases the risk for mastitis. This

</div>Trang 51<div class="page_container" data-page="51">

If each case is matched with a same-age control, there will be no

association (OR for old age = 2.6, P = 0.0001)

</div>Trang 52<div class="page_container" data-page="52">

No Confounding

</div>Trang 53<div class="page_container" data-page="53">

Cases of Down Syndrome by Birth Orderand Maternal Age

If each case is matched with a same-age control, there will be no association. If analysis is repeated after stratification by age, there

will be no association with birth order.

</div>Trang 55<div class="page_container" data-page="55">

Confounding or Effect Modification

Birth WeightLeukaemiaSex

Can sex be responsible for the birth weight association in leukaemia?

- Is it correlated with birth weight?

- Is it correlated with leukaemia independently of birth weight?

- Is it on the causal pathway?

- Can it be associated with leukaemia even if birth weight is low?

- Is sex distribution uneven in comparison groups?

</div>Trang 56<div class="page_container" data-page="56">

Confounding or Effect Modification

Birth WeightLeukaemiaSex

Does birth weight association differ in strength according to sex? Birth WeightLeukaemia

OR = 1.5

</div>Trang 57<div class="page_container" data-page="57">

Effect Modification

In an association study, if the strength of the

association varies over different categories of a third variable, this is called effect modification. The third

variable is changing the effect of the exposure.

The effect modifier may be sex, age, an environmental exposure or a genetic effect.

Effect modification is similar to interaction in statistics. There is no adjustment for effect modification. Once it

is detected, stratified analysis can be used to obtain stratum-specific odds ratios.

</div>Trang 58<div class="page_container" data-page="58">

Increases knowledge of biological mechanismAllows targeting of public health action

</div>Trang 61<div class="page_container" data-page="61">

HOW TO CONTROL FOR CONFOUNDERS?

• IN STUDY DESIGN…

– 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

</div>Trang 62<div class="page_container" data-page="62">

HOW TO CONTROL FOR CONFOUNDERS?

• IN DATA ANALYSIS…

– STRATIFIED ANALYSIS using the Mantel Haenszel method to adjust for confounders

– IMPLEMENT A MATCHED-DESIGN after you have collected data (frequency or group)

– RESTRICTION is still possible at the analysis stage but it means throwing away data

</div>Trang 63<div class="page_container" data-page="63">

Effect of blinding on outcome of trials of acupuncture for chronic back pain

</div>Trang 64<div class="page_container" data-page="64">

WILL ROGERS' PHENOMENON

Assume that you are tabulating survival for patients with a certain type of tumour. You separately track survival of patients whose cancer has

metastasized and survival of patients whose cancer remains localized. As you would expect, average survival is longer for the patients without metastases.

Now a fancier scanner becomes available, making it possible to detect

metastases earlier. What happens to the survival of patients in the two groups? The group of patients without metastases is now smaller. The patients who are

removed from the group are those with small metastases that could not have been detected without the new technology. These patients tend to die sooner

than the patients without detectable metastases. By taking away these patients, the average survival of the patients remaining in the "no metastases"

group will improve.

What about the other group? The group of patients with metastases is now larger. The additional patients, however, are those with small metastases. These patients tend to live longer than patients with larger metastases. Thus

the average survival of all patients in the "with-metastases" group will

</div>Trang 65<div class="page_container" data-page="65">

Cause-and-Effect Relationship

</div>Trang 67<div class="page_container" data-page="67">

M. Tevfik DORAK

Paediatric & Lifecourse Epidemiology Research GroupSchool of Clinical Medical Sciences (Child Health)

Newcastle UniversityEngland, U.K.

o

</div>

Bias, Confounding and Fallacies in Epidemiology

<b>Bias, Confounding and Fallacies in Epidemiology</b>

<b>M. Tevfik DORAK</b>

<b>Bias is one of the three major threats to internal </b>

<b>Any trend in the collection, analysis, interpretation, publication or review of data that can lead to conclusions that are systematically different from </b>

<b>the truth (Last, 2001)</b>

<b>A process at any state of inference tending to produce results that depart systematically from </b>

<b>the true values (Fletcher et al, 1988)</b>

<b>Systematic error in design or conduct of a study (Szklo et al, 2000)</b>

<b>What is Bias? </b>

<b>Errors can be differential (systematic) or non-differential (random)</b>

<b>Random error: use of invalid outcome </b>

<b>measure that equally misclassifies cases and controls</b>

<b>Differential error: use of an invalid measures that misclassifies cases in one direction and misclassifies controls in another</b>

<b>Term '</b><i><b>bias</b></i><b>' should be reserved for differential or systematic error </b>

<b>Bias is systematic error</b>

<b>Chance vs Bias</b>

<i><b>Selection bias</b></i>

<b>Unrepresentative nature of sample</b>

<i><b>Information (misclassification) bias</b></i>

<b>Errors in measurement of exposure of disease</b>

<i><b>Confounding bias</b></i>

<b>Distortion of exposure - disease relation by some other factor</b>

<b>Types of bias not mutually exclusive</b>

<i><b>(effect modification is not bias)</b></i>

<b>Types of Bias </b>

<b>Selection Bias</b>

<b>Selective differences between comparison groups that impacts on relationship between exposure </b>

<b>and outcome</b>

<i><b>Usually results from comparative groups not coming from the same study base and not being </b></i>

<b>representative of the populations they come from</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Case-Control Studies: Potential Bias</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<b>Selection Bias Examples</b>

<i><b>Selection bias</b></i>

<b>Unrepresentative nature of sample</b>

<i><b>** Information (misclassification) bias **</b></i>

<b>Errors in measurement of exposure of disease</b>

<i><b>Confounding bias</b></i>

<b>Distortion of exposure - disease relation by some other factor</b>

<b>Types of Bias </b>

<b>Information / Measurement / Misclassification Bias </b>

<b>Information / Measurement / Misclassification Bias </b>

<b>Information / Measurement / Misclassification Bias </b>

<b>Controlling for Information Bias </b>

<i><b>Selection bias</b></i>

<b>Unrepresentative nature of sample</b>

<i><b>Information (misclassification) bias</b></i>

<b>Errors in measurement of exposure of disease</b>

<i><b>** Confounding bias **</b></i>

<b>Distortion of exposure - disease relation by some other factor</b>

<b>Types of Bias </b>

<b>Cases of Down Syndrome by Birth Order</b>

<b>Cases of Down Syndrome by Age Groups</b>

<b>Cases of Down Syndrome by Birth Orderand Maternal Age</b>

<b>• A third factor which is related to both </b>

<b>exposure and outcome, and which accounts for some/all of the observed relationship </b>

<b>between the two</b>

<b>• Confounder not a result of the exposure</b>

<b>– e.g., association between child’s birth rank </b>

<b>(exposure) and Down syndrome (outcome); mother’s age a confounder?</b>

<b>– e.g., association between mother’s age (exposure) </b>

<b>and Down syndrome (outcome); birth rank a confounder?</b>

<b>Confounding</b>

<b>Exposure Outcome</b>

<b>Third variable</b>

<b>To be a confounding factor, two conditions must be met:</b>

<b>Be associated with exposure</b>

<b> - without being the consequence of exposure</b>

<b>Confounding</b>

<b>Birth Order Down Syndrome</b>

<b>Maternal Age</b>

<i><b> Information (misclassification) bias </b></i>

<i><b> Confounding bias </b></i>