Confounding and Bias

Question 1

What is validity in research?

Answer 1

Validity is about the truth. It refers to how close a study finding (observed association) comes to the truth (true association).

Question 2

What does a valid measure of association describe?

Answer 2

A valid measure of association describes the true (real) situation accurately.

Question 3

What is internal validity?

Answer 3

Internal validity refers to the extent to which the study findings accurately reflect the true situation within the current context or study, free from biases and errors.

Question 4

What is external validity?

Answer 4

External validity refers to the extent to which the study findings can be extrapolated to other groups or settings, also known as generalisability.

Question 5

Why is validity important in research?

Answer 5

Validity is important because it determines the accuracy and truthfulness of the study findings, ensuring that the results genuinely reflect the reality being studied.

Question 6

What are the types of validity in research?

Answer 6

The types of validity in research are internal validity (current context/study) and external validity (generalisability to other groups or settings).

Question 7

What is reliability in research?

Answer 7

Reliability is about the consistency and repeatability of measurements, often described as precision, repeatability, consistency, and stability.

Question 8

Why is reliability important in research?

Answer 8

Reliability is important because it ensures that similar information is elicited when the measurement is repeated, demonstrating that the results are stable and consistent over time

Question 9

How does reliability differ from validity?

Answer 9

Reliability refers to the consistency and repeatability of measurements, while validity refers to how accurately the measurements reflect the true situation or association. Reliable results may not always be valid, but valid results must be reliable.

Question 10

What are the sources of invalidity that can affect internal validity?

Answer 10

The sources of invalidity that can affect internal validity include chance (random error), bias (systematic error), and confounding (influence of a third variable).

Question 11

How does chance (random error) affect internal validity?

Answer 11

Chance (random error) affects internal validity by introducing variability in the data that is not due to the true association being studied. This randomness can lead to untrue associations or mask true associations.

Question 12

How does bias (systematic error) affect internal validity?

Answer 12

Bias (systematic error) affects internal validity by introducing consistent, directional errors that distort the true association. This can result from flaws in study design, data collection, or analysis

Question 13

What is confounding and how does it affect internal validity?

Answer 13

Confounding occurs when a third variable influences both the independent and dependent variables, creating a false association or masking a true association. It affects internal validity by distorting the observed relationship

Question 14

What are random errors in research?

Answer 14

Random errors are errors of measurement or population selection that occur due to chance and introduce variability in the data that is not related to the true association being studied.

Question 15

Why is it challenging to conduct studies with the entire population?

Answer 15

It is very difficult and often impossible to conduct studies with the entire population due to practical constraints, so researchers select a sample of participants instead.

Question 16

What issue can arise when selecting a sample for a study?

Answer 16

It is possible that a particular sample does not adequately represent the source population, which can affect the validity of the study finding

Question 17

How can random errors impact study findings?

Answer 17

Random errors can cause findings to occur due to chance, usually diluting the results and reducing the accuracy of the study’s conclusions.

Question 18

How can biostatistics help address random errors?

Answer 18

Biostatistics can be used to quantify the role that chance could play in the study findings, helping to determine the reliability and significance of the results.

Question 19

How to prevent random errors

Answer 19

• big enough sample size
• test questionnaire/ equipment

Question 20

How do researchers quantify the role of chance in study findings?

Answer 20

Researchers quantify the role of chance in study findings using p-values and confidence intervals.

Question 21

What is a p-value?

Answer 21

A p-value is a numeric value ranging from 0 to 1 that indicates how likely it is that an observed association is due to chance.

Question 22

How are p-values interpreted in research?

Answer 22

P-values are typically interpreted with a cutoff of 0.05. A p-value less than 0.05 suggests that the observed association is unlikely to be due to chance (reject the null hypothesis), while a p-value greater than 0.05 suggests that the association could be due to chance (cannot reject the null hypothesis).

Question 23

What is a confidence interval (CI)?

Answer 23

A confidence interval is the range of values within which the true association is likely to fall, usually expressed as a 95% confidence interval.

Question 24

Why are confidence intervals important?

Answer 24

Confidence intervals are important because they provide a range of values that likely contain the true association, giving an estimate of the precision and reliability of the study findings.

Question 25

What does a 95% confidence interval indicate?

Answer 25

A 95% confidence interval indicates that there is a 95% probability that the true association falls within the specified range of values.

Question 26

Narrow CI

Answer 26

precise estimate, not alot of variability

Question 27

Wide CI

Answer 27

imprecise and alot of of variability

Question 28

CI crosses 1

Answer 28

so not statistically significant

Question 29

What is bias in research?

Answer 29

Bias is a systematic error in the design, conduct (collection of data), or analysis of a study that leads to an incorrect estimate of the exposure-disease association, making the results invalid.

Question 30

What are the types of bias in research?

Answer 30

Types of bias include selection bias, measurement bias, confounding bias, and publication bias, each influencing study outcomes in different ways.

Question 31

How does a researcher influence bias in research?

Answer 31

Researchers can introduce bias through the collection, analysis, interpretation, publication, or review of data, consciously or unconsciously skewing results towards a particular outcome.

Question 32

What characterizes bias in research?

Answer 32

Bias is patterned, often resulting in either overestimation or underestimation of the association between exposure and disease, distorting the true relationship being studied.

Question 33

How does bias affect the validity of a study?

Answer 33

Bias undermines the validity of a study by producing results that do not accurately reflect reality, potentially leading to incorrect conclusions and ineffective or harmful interventions.

Question 34

What is selection bias?

Answer 34

Selection bias refers to systematic errors in the selection of study participants, which can lead to a sample that does not adequately represent the target population, affecting the generalizability of study findings.

Question 35

What is information bias?

Answer 35

Information bias, also known as measurement or misclassification bias, refers to systematic errors in the collection of information from study participants. This can distort the measurement of exposure or outcome variables, leading to inaccurate study results.

Question 36

Ways to prevent selection bias

Answer 36

■Understand your study population and how to access them and keep them in your study;
■Define your population

Question 37

types of selection bias

Answer 37

sampling bias
self selection bias
response bias
diagnostic bias
admission (Berkson’s) bias

Question 38

Sources of information bias

Answer 38

subject variation
observer variation
deficiency of measurement tools

Question 39

subject variation

Answer 39

■Example: recall bias in case control study
■where you are interviewing mothers of children with congenital birth defects

Question 40

observer variation

Answer 40

■Example: unblinded outcome assessment in RCT where the doctor knows which patients are on the drug

Question 41

Deficiency of measurement tools

Answer 41

Example: Questionnaire poorly designed with poor validity

Question 42

Methods to prevent selection bias

Answer 42

–Ensure sample is representative
–Clear protocols
–Clear inclusion and exclusion criteria
–Minimize non-response and loss to follow-up

Question 43

Methods to prevent information bias

Answer 43

–Clear ways of determining the exposure and the outcome
–Ensure tools are valid, reliable, standardized & calibrated (NB questionnaire lecture)
–Blind subjects and observers
–Train all staff well
–Validate data, test tools and pilot study

Question 44

What is confounding in research?

Answer 44

Confounding is a situation where a third variable, not accounted for in the study design or analysis, distorts the association between the exposure (E) and outcome (D), leading to incorrect conclusions about their relationship

Question 45

What role does confounding play in research?

Answer 45

Confounding can create the appearance of an association between exposure and outcome when none exists (false positive) or mask a true association (false negative), thereby compromising the validity of study findings.

Question 46

How does confounding impact the validity of research?

Answer 46

Confounding leads to overestimation or underestimation of the association between exposure and outcome, reducing the validity of study results by providing inaccurate estimates of the true relationship.

Question 47

How can researchers address confounding in their studies?

Answer 47

Researchers can address confounding by using study designs that control for potential confounders (e.g., randomized controlled trials), matching or stratifying participants based on confounding variables, or statistically adjusting for confounders in data analysis.

Question 48

Why are confounders considered context-specific?

Answer 48

Confounders vary depending on the specific study context, including what the exposure and outcome variables are. This variability reflects how different factors can influence the relationship between exposure and outcome differently in various settings.

Question 49

What role do confounders play as risk factors?

Answer 49

Confounders often act as risk factors for various health outcomes, influencing the likelihood of developing certain conditions or experiencing particular health effects. Understanding and controlling for these factors are critical in epidemiological and clinical research.

Question 50

Example of cofounders

Answer 50

Age (categories)
■Sex /Gender (Male or Female)
■SES (Rich or Poor)
■Smoking status
■Alcohol
■Diet, etc.

Question 51

A variable is a cofounder if

Answer 51

1. it is a risk factor for the outcome (associated with disease)
2. it is associated/ correlated with the exposure
3. not in causal pathway between exposure and disease

Question 52

controlling cofounding during design phase of a study

Answer 52

• randomization
• matching
• restriction

Question 53

controlling cofounding during data analysis “statistical adjustment’

Answer 53

• stratification
• multivariable regression

Question 54

How does randomization help control confounding during the design phase?

Answer 54

Randomization ensures that participants are assigned to exposure groups (e.g., treatment and control) randomly. This helps achieve equal distribution of both known and unknown confounders between the exposed and unexposed groups, reducing the potential for bias in estimating the exposure-outcome relationship.

Question 55

How does matching control confounding in case-control studies?

Answer 55

Matching involves selecting controls that are similar to cases with respect to potential confounders. Group matching ensures the proportion of controls with specific characteristics matches that of cases, while individual matching pairs each case with a control who is similar on confounding variables. This method can lead to analytical complexities and should be approached with expert input

Question 56

How does restriction control confounding?

Answer 56

Restriction limits participant entry into the study based on confounding variables. For example, if age is a confounder, the study may restrict enrollment to only younger or older participants. This approach prevents variation in the confounding variable within the study population, simplifying the analysis but potentially limiting generalizability.

Question 57

Why is expert input important for using these methods?

Answer 57

Analytical complexities can arise when implementing matching techniques, requiring expertise to navigate potential pitfalls and ensure robust study design. Expert input helps maximize the effectiveness of confounding control methods and minimize bias in research outcomes.

Question 58

How does statistical adjustment help control confounding during analysis?

Answer 58

Statistical adjustment involves including confounding variables as covariates in regression models or other statistical analyses. By doing so, researchers can estimate the exposure-disease association while accounting for the potential influence of confounding factors.