Final: Lectures 17-18

Question 0

Definition of Bias

Answer 0

• Systemic (non-random) error in study design or conduct leading to erroneous results.
• Distorts the relationship (ass.) between exposure and outcome

Question 1

What are the 3 aspects of a study do researchers evaluate before declaring a statistical association?

Answer 1

• Check for bias
• Check for confounding or effect modification
• Check for statistical significane

Question 2

T/F Bias can be fixed once it has already occurred (after study end).

Answer 2

• False

* Prospective (pre-study) consideration and adjustment can minimize bias and its impact

Question 3

When assessing for bias and its impact, investigators evaluate 3 components:

Answer 3

1. Source/Type
2. Magnitude/Strength: can account entirely for a weak ass. but is not likely to account for very strong ass.
3. Direction: *Bias can over- or under-estimate the true measure of ass. (towards or away from the Null Hypothesis)

Question 4

Information/Observation/Measurement-related biases

Answer 4

• Any aspect in the way the researcher collects info, or measures/observes subjects, which creates a systematic difference between groups in the quality/accuracy of their info
• Errors in measurement or classification

Question 5

Selection-related biases

Answer 5

• Any aspect in the way the researchers selects subjects which creates a systematic difference in the composition between groups
• commonly seen when comparative groups not coming from same population/group or not being representative of full pop

Question 6

Types of Selection Bias

Answer 6

• Healthy-Worker Bias: can easily be seen in prospective Cohort studies
• Self-Selection/Participant (Responder) bias: those that wish to participate may be different in some way to those that don’t volunteer or self-select to participate (may be base on education or income)
• Control Selection bias: easily seen in Case-Control studies

Question 7

Recall (reporting) bias

Answer 7

• Subject-related variation (info bias): a differential level of accuracy/detail in provided info between study groups
• Exposed or diseased subjects may have greater sensitivity for recalling their history or amplify their responses (Negative event tied to bad experience helps you remember, or profoundly positive)
• Individuals can report their “effects” of exposure, disease symptoms or treatment differently b/c they are part of a study, “Hawthorne Effect”

Question 8

Contamination bias

Answer 8

• Subject-related bias: members of the control group accidentally, or outside of the study protocol, receive the treatment or are exposed to the intervention being studied
• Not taking pills, taking other pills that react to the drug being tested ect.
• Likely connected to interventional studies

Question 9

Compliance Bias

Answer 9

• Subject-related bias: groups being interventionally studied have different compliances
• Not taking drugs as recommended due to side-effects, benefit, ect.

Question 10

Lost to Follow-up bias

Answer 10

• Subject-related bias: groups being studied have different withdrawal rates OR there are other differences between those that stay in the study and those that withdraw or are lost to follow-up
• Differential vs Non-Differential

Question 11

Can Cohort studies have lose to follow-up?

Answer 11

•Yes, with prospective, waiting for the disease to occur lost to follow-up

Question 12

Interviewer (Proficiency) bias

Answer 12

• Observer-related variation: a systematic difference in soliciting, recording, or interpreting on the part of their researcher
• Interviewers knowledge may influence the structure, or tone, of questions or follow-up questions which may influence response from the study subject OR..
• The interventions or treatments are not applied equally between groups due to skill or training differences of study personnel or differences in study procedure compliance by staff at different sites

Question 13

Diagnosis/surveillance (Expectation) bias

Answer 13

• Observer-related variation: Different evaluation/classification/diagnosis/ observation between study groups
• Observes may have preconceived expectations of what they should find in examination/evaluation/follow-up
• “Hawthorne-like Effect” from the researchers perspective
• Kappa statistic, agreement between evaluators, making sure they ask the same question in the same way to each patient

Question 14

Controlling for Bias

Answer 14

• Select most precise, accurate, and medically-appropriate easures of assessment and evaluation/observation
• Blinding/Masking
• Use multiple sources to gather all info
• Randomly allocate observers/interviewers for data collection (and train them! use technology!)
• Build in as many methods necessary to minimize loos to follow-up

Question 15

What is the source of info/observation/measurement bias?

Answer 15

•Misclassification (error in classifying either disease or exposure status, both)

Question 16

Non-differential (error in both groups equally) misclassification

Answer 16

• Misclassification of exposure or disease which is unrelated to the other, depending on study design
• Effect = for 2 category variables, bias can move the measure of ass. toward the null hypothesis, it attenuates your effect estimates of ass.
• Ex. RR of .3 moves to .7 (attenuation towards the null)
• Ex. OR of 1.9 moves to 1.2 (attenuation towards the null)

Question 17

Differential (error in one group differently than other)

Answer 17

• Misclassification of exposure or disease is related to the other, depending on study design
• Effect = Bias can move the meausre of ass. in either direction in relation to the null, it can inflate or attenuate your effect estimates of ass.
• Ex. RR of .8 moves to .2 or 1.4 to 2.1 (inflation from null)
• Ex. OR of 2.3 moves to 1.1 or .6 moves to .9 (attenuation towards null)

Question 18

OR =

Answer 18

•ad/bc

Question 19

T/F Non-Differential misclassification has equal changes in numbers in the control and cases.

Answer 19

•True, Look for balances and equal differentiation