Bias and Misclassification

Question 1

Can anything be done to fix bias in a study?

Answer 1

No. If bias has already occurred and the study has ended, then it’s effects on the data cannot be corrected.

Question 2

Define Bias

Answer 2

systematic (non random) error in study or design or conduct leading to erroneous results

Question 3

what are the 2 main categories of effect that bias can have on the source of the study’s findings

Answer 3

Selection-related Bias

Measurement-related Bias

Question 4

Define Selection related Bias

Answer 4

any aspect in the way the researcher selects/acquires study subjects that can lead to a “systematic difference” between groups

Commonly occurs when the study subjects do not come from the same group

Question 5

Define Measurement-related Bias

Answer 5

any aspect in the way the researcher collects information that creates a “systematic difference” between groups

This can cause resultant error (misclassification)

Question 6

How can Bias affect the Magnitude of a study’s findings?

Answer 6

Bias can account “entirely” for a weak association, but is not likely to account entirely for a strong association.

Question 7

How can Bias affect the Direction of a study’s findings?

Answer 7

Bias can over or underestimate the true measure of association

Question 8

Define Recall/Reporting bias

Answer 8

a difference in the level of accuracy/detail that is provided from each group
Ex. Diseased subjects may have greater sensitivity for recalling history (exaggerate)

Question 9

Define Hawthorne (Observation) Effect

Answer 9

people change their behavior during the study because they are aware that they are under observation (aren’t themselves)

Question 10

Define Contamination Bias

Answer 10

when members of the control group are unintentionally exposed to the intervention that is being studied, or some other type of treatment

Ex. A control group subject being prescribed pain killers by a doctor and not mentioning it to the researcher (if the study is on pain relief)

Question 11

Define Compliance/Adherence Bias

Answer 11

when groups being studied have different levels of compliance with the intervention methods being studied

Ex. If the painkiller group forgets to take their painkillers or is not good about taking the medication on a regular basis

Question 12

Define Diagnosis/Surveillance/Expectation bias

Answer 12

different evaluation/classification/diagnoses between study groups due to observer’s expectations
Basically the Hawthorne effect, but from the researcher’s point of view

Question 13

Define Interviewer bias

Answer 13

a “systematic difference” in soliciting, recording, or interpreting on the part of the researcher

Interviewers knowledge may influence the structure/tone/presentation of the questions asked

Interventions/treatments being applied unequally to the groups as a result of the researchers decision making or compliance of staff is also an example of this

Question 14

Define Lost to Follow Up bias

Answer 14

Groups being studied have different withdrawal or follow-up rates.

Question 15

Define Lead-Time Bias

Answer 15

an apparent benefit due to the early detection of a disease, despite and unchanged clinical outcome (AIDS discovered earlier yields longer lifespan than AIDS discovered later in it’s pathological course)

Question 16

define Selection Bias

Answer 16

occurs when the way that subjects are selected for the study creates differences in groups (very common issue)

Healthy worker bias is an example of this bias

Question 17

Define Self-selection/responder/participant bias

Answer 17

Those that wish to participate in the study may be different in some way to the rest of the population that is unlikely to volunteer for the study

Question 18

Define the Healthy worker effect

Answer 18

when only the workers healthy enough to work are included in the study, instead of all injured/sick/death workers being included

Question 19

How can researchers employ control for bias and misclassification?

Answer 19

Blinding/Masking their interpreters/observers

Randomly allocate observers/interviewers for data collection (train them to avoid bias sources)

Build in as many methods as possible to prevent “lost to follow up bias” before the study is conducted

Question 20

What are the 2 mains types of misclassification?

Answer 20

Differential and Nondifferential

Question 21

Define Differential missclassifcation

Answer 21

the bias is present in different levels between the groups
Moves the RR/OR away from 1

brings the value AWAY from it’s true value if there were no miss-classification occurring

Question 22

Define Non-differential misclassification

Answer 22

the bias is equally present in all groups (preferred)
Moves the RR/OR towards 1

brings the value closer to 1

Question 23

What are the 3 main elements of bias impact?

Answer 23

Source/Type: selection related or Measurement related

Magnitude/strength:

Direction: it can over or underestimate the true measure of association