In the pursuit of Evidence-Based Practice (EBP), healthcare professionals must critically appraise a wide range of clinical research. While randomized controlled trials (RCTs) are often considered the gold standard, observational studies, such as cohort studies, provide invaluable information, especially for questions of prognosis, harm, or etiology. However, the validity of any study, particularly observational research, can be threatened by systematic errors. Understanding and identifying bias (systematic error in design or conduct) and confounding (distortion of an association by a third variable) are fundamental skills for any PharmD student. This ability to critically appraise research is a “transcending concept” essential for translating evidence into sound clinical practice. This MCQ quiz will test your knowledge on cohort studies and the critical concepts of confounding and bias.
1. A cohort study is a type of observational study where researchers:
- A. Compare a group with a disease (cases) to a group without the disease (controls) and look back at past exposures.
- B. Analyze data from a whole population at a single point in time.
- C. Identify a group of individuals (a cohort), classify them by exposure status, and follow them over time to determine outcome incidence.
- D. Randomly assign participants to an intervention or control group.
Answer: C. Identify a group of individuals (a cohort), classify them by exposure status, and follow them over time to determine outcome incidence.
2. What is the key difference between a prospective cohort study and a retrospective cohort study?
- A. Prospective studies are always smaller than retrospective studies.
- B. Retrospective studies are a type of randomized trial.
- C. In prospective studies, the investigator identifies the cohort and follows them forward in time; in retrospective studies, the investigator uses existing historical data to identify a cohort and determine outcomes that have already occurred.
- D. Prospective studies are immune to bias, while retrospective studies are not.
Answer: C. In prospective studies, the investigator identifies the cohort and follows them forward in time; in retrospective studies, the investigator uses existing historical data to identify a cohort and determine outcomes that have already occurred.
3. Bias in a clinical study is defined as a(n):
- A. Random error that occurs by chance.
- B. The presence of a third variable that is associated with both the exposure and outcome.
- C. Systematic error in the design, conduct, or analysis of a study that results in a mistaken estimate of the exposure’s effect on the outcome.
- D. The failure to achieve statistical significance.
Answer: C. Systematic error in the design, conduct, or analysis of a study that results in a mistaken estimate of the exposure’s effect on the outcome.
4. Confounding occurs when a third variable (the confounder) is:
- A. An intermediate step in the causal pathway between the exposure and the outcome.
- B. Associated with the exposure and independently associated with the outcome, but is not on the causal pathway.
- C. Only associated with the exposure but not the outcome.
- D. A result of systematic error in data collection.
Answer: B. Associated with the exposure and independently associated with the outcome, but is not on the causal pathway.
5. In a study examining the association between coffee drinking (exposure) and pancreatic cancer (outcome), smoking is a classic confounder because:
- A. Smoking is caused by coffee drinking.
- B. Coffee drinking causes smoking.
- C. Smoking is associated with both coffee drinking and pancreatic cancer independently.
- D. Pancreatic cancer causes people to start smoking and drinking coffee.
Answer: C. Smoking is associated with both coffee drinking and pancreatic cancer independently.
6. The primary measure of association calculated from a cohort study is the:
- A. Odds Ratio (OR)
- B. Relative Risk or Risk Ratio (RR)
- C. Prevalence Rate
- D. P-value
Answer: B. Relative Risk or Risk Ratio (RR)
7. “Recall bias” is a major concern in which type of study design?
- A. Randomized controlled trials
- B. Prospective cohort studies
- C. Retrospective case-control studies
- D. Meta-analyses
Answer: C. Retrospective case-control studies (where cases with the disease may recall past exposures differently than healthy controls).
8. “Selection bias” occurs when:
- A. The way subjects are selected or retained in a study leads to a systematic difference between the groups being compared.
- B. Interviewers ask questions differently to cases and controls.
- C. The outcome is measured inaccurately.
- D. A known confounder is not controlled for in the analysis.
Answer: A. The way subjects are selected or retained in a study leads to a systematic difference between the groups being compared.
9. A major strength of a prospective cohort study is its ability to:
- A. Be completed quickly and inexpensively.
- B. Be efficient for studying rare diseases.
- C. Establish temporality (i.e., that the exposure preceded the outcome).
- D. Eliminate all forms of bias.
Answer: C. Establish temporality (i.e., that the exposure preceded the outcome).
10. “Interviewer bias” can be minimized in a study by:
- A. Using interviewers who are aware of the study hypothesis and each participant’s status.
- B. Using standardized questionnaires and blinding the interviewers to the participants’ exposure or disease status.
- C. Allowing interviewers to ask questions in any way they see fit.
- D. Only interviewing participants who are likely to confirm the hypothesis.
Answer: B. Using standardized questionnaires and blinding the interviewers to the participants’ exposure or disease status.
11. Which of the following is the BEST method for controlling for both known and unknown confounders in a clinical study?
- A. Stratification
- B. Matching
- C. Randomization
- D. Multivariable analysis
Answer: C. Randomization (This is a key advantage of RCTs over observational studies).
12. In a cohort study, “loss to follow-up” can lead to which type of bias if the losses are different between the exposed and unexposed groups and are related to the outcome?
- A. Recall bias
- B. Interviewer bias
- C. Selection bias (specifically, attrition bias)
- D. Confounding
Answer: C. Selection bias (specifically, attrition bias)
13. A Relative Risk (RR) of 2.0 for developing lung cancer in smokers vs. non-smokers means that:
- A. Smokers are 20% more likely to develop lung cancer.
- B. Smoking causes lung cancer in 2% of people.
- C. Smokers have twice the risk of developing lung cancer compared to non-smokers.
- D. Non-smokers have twice the risk of developing lung cancer.
Answer: C. Smokers have twice the risk of developing lung cancer compared to non-smokers.
14. A Relative Risk (RR) of 1.0 indicates that:
- A. The exposure is strongly protective against the outcome.
- B. The exposure is a major risk factor for the outcome.
- C. There is no association between the exposure and the outcome.
- D. The study results are invalid.
Answer: C. There is no association between the exposure and the outcome.
15. In the analysis phase of a study, which technique can be used to control for confounding?
- A. Blinding
- B. Using a placebo
- C. Stratification or multivariable regression analysis
- D. Increasing the sample size
Answer: C. Stratification or multivariable regression analysis
16. A significant weakness of prospective cohort studies is that they are:
- A. Inefficient for studying rare exposures.
- B. Prone to recall bias.
- C. Inefficient for studying rare diseases and can be very time-consuming and expensive.
- D. Unable to establish a temporal relationship.
Answer: C. Inefficient for studying rare diseases and can be very time-consuming and expensive.
17. “Misclassification bias” refers to:
- A. Errors in selecting participants for a study.
- B. Errors in classifying a subject’s exposure status or outcome status.
- C. The effect of a third variable on the exposure-outcome relationship.
- D. The bias that occurs when participants are lost to follow-up.
Answer: B. Errors in classifying a subject’s exposure status or outcome status.
18. To be a confounder, a variable must be associated with the exposure AND:
- A. Be an intermediate step in the causal pathway.
- B. Be a result of the outcome.
- C. Be an independent risk factor for the outcome.
- D. Be measured inaccurately.
Answer: C. Be an independent risk factor for the outcome.
19. How does “matching” in a case-control study help to control for confounding?
- A. It ensures that the cases and controls are systematically different with respect to the matching factor.
- B. It ensures that the cases and controls have an equal distribution of the potential confounding factor (e.g., age, sex).
- C. It eliminates all forms of bias.
- D. It increases the sample size of the study.
Answer: B. It ensures that the cases and controls have an equal distribution of the potential confounding factor (e.g., age, sex).
20. A study finds that people who carry lighters in their pocket have a higher risk of lung cancer. In this scenario, what is the most likely confounder?
- A. The brand of lighter
- B. The presence of lung cancer
- C. Smoking status
- D. The type of pocket
Answer: C. Smoking status (Smokers are more likely to carry lighters and also more likely to get lung cancer).
21. A major difference between bias and confounding is that:
- A. Bias can be controlled in the analysis phase, while confounding cannot.
- B. Confounding can be controlled in both the design and analysis phases, while bias is best prevented in the design and conduct of the study.
- C. Bias and confounding are interchangeable terms for the same problem.
- D. Confounding only occurs in RCTs, while bias only occurs in cohort studies.
Answer: B. Confounding can be controlled in both the design and analysis phases, while bias is best prevented in the design and conduct of the study.
22. The “healthy worker effect” is a type of selection bias where:
- A. The general population is healthier than a working population.
- B. Working populations are often healthier than the general population, which can lead to an underestimation of risk if the working population is compared to the general population.
- C. All workers in a study are guaranteed to be healthy.
- D. Workers accurately recall their occupational exposures.
Answer: B. Working populations are often healthier than the general population, which can lead to an underestimation of risk if the working population is compared to the general population.
23. Which of the following is NOT a method to control for confounding in the design stage of a study?
- A. Randomization
- B. Restriction
- C. Matching
- D. Stratified analysis
Answer: D. Stratified analysis (This is a method used in the analysis stage).
24. A cohort study follows a group of nurses for 20 years to examine the association between oral contraceptive use and breast cancer risk. This is an example of a:
- A. Retrospective cohort study
- B. Case-control study
- C. Cross-sectional study
- D. Prospective cohort study
Answer: D. Prospective cohort study
25. A researcher uses a company’s employment records from 1980-2000 to identify workers exposed to a chemical and then checks death records up to 2020 to see if they developed a specific cancer. This is an example of a:
- A. Prospective cohort study
- B. Retrospective cohort study
- C. Randomized controlled trial
- D. Case series
Answer: B. Retrospective cohort study
26. If the results of a cohort study are biased, increasing the sample size of the study will:
- A. Eliminate the bias.
- B. Decrease the bias.
- C. Not correct the bias; it may only increase the precision of the incorrect result.
- D. Make the bias less important.
Answer: C. Not correct the bias; it may only increase the precision of the incorrect result.
27. A key advantage of cohort studies over case-control studies is that cohort studies can be used to:
- A. Study rare diseases efficiently.
- B. Calculate the true incidence rates and relative risk of a disease.
- C. Be completed very quickly and with minimal cost.
- D. Avoid any potential for confounding.
Answer: B. Calculate the true incidence rates and relative risk of a disease.
28. If an exposure is protective against an outcome, the Relative Risk (RR) would be expected to be:
- A. Greater than 1.0
- B. Equal to 1.0
- C. Less than 1.0
- D. Equal to zero
Answer: C. Less than 1.0
29. The main purpose of blinding (or masking) participants and investigators in a study is to reduce:
- A. Confounding by age
- B. Loss to follow-up
- C. Performance bias and detection (information) bias
- D. The cost of the study
Answer: C. Performance bias and detection (information) bias
30. Which of the following must be true for a variable to be a confounder?
- A. It must be a direct cause of the outcome.
- B. It must be on the causal pathway between exposure and outcome.
- C. It must be associated with both the exposure and the outcome.
- D. It must be measured without any error.
Answer: C. It must be associated with both the exposure and the outcome.
31. A cohort study is initiated to investigate if Drug X causes a rare adverse effect. What is a major limitation of this study design for this research question?
- A. It is difficult to assess the temporal relationship.
- B. The study would need to be extremely large and long to observe enough outcomes, making it inefficient.
- C. Recall bias will be a major problem.
- D. The study cannot assess multiple outcomes.
Answer: B. The study would need to be extremely large and long to observe enough outcomes, making it inefficient. (Case-control is better for rare diseases).
32. A study reports an association between gray hair and risk of heart attack. Age is a confounder in this relationship because:
- A. Gray hair causes heart attacks.
- B. Heart attacks cause gray hair.
- C. Age is associated with both gray hair (older people have more) and risk of heart attack (older people are at higher risk).
- D. The association is purely coincidental.
Answer: C. Age is associated with both gray hair (older people have more) and risk of heart attack (older people are at higher risk).
33. The key characteristic that defines a cohort study is:
- A. The study population is selected based on their disease status.
- B. The study population is selected based on their exposure status and followed for development of disease.
- C. An intervention is randomly assigned by the investigator.
- D. Data is collected at a single point in time.
Answer: B. The study population is selected based on their exposure status and followed for development of disease.
34. “Differential misclassification” of exposure occurs when the error in classifying exposure status:
- A. Is the same in both the diseased and non-diseased groups.
- B. Is different in the diseased and non-diseased groups.
- C. Is random and not systematic.
- D. Can only be corrected by increasing the sample size.
Answer: B. Is different in the diseased and non-diseased groups. (This can bias the results towards or away from the null).
35. If a study finds a crude relative risk of 1.5, but after adjusting for age, the relative risk becomes 1.0, this suggests that:
- A. The study had significant selection bias.
- B. Age was a positive confounder in the relationship.
- C. The exposure is truly a risk factor.
- D. The sample size was too small.
Answer: B. Age was a positive confounder in the relationship.
36. A major advantage of using a retrospective cohort study design compared to a prospective cohort study is that it is:
- A. Less prone to confounding.
- B. More efficient for studying rare exposures.
- C. Generally faster to complete and less expensive, as it uses existing data.
- D. Not subject to loss to follow-up.
Answer: C. Generally faster to complete and less expensive, as it uses existing data.
37. Which of the following is NOT a source of bias?
- A. Loss to follow-up
- B. Recall bias
- C. Interviewer bias
- D. Random error (chance)
Answer: D. Random error (chance) (Random error affects precision, while bias affects accuracy).
38. “Restriction” is a method used in the design phase to control for confounding. It involves:
- A. Limiting study participants to those with a specific characteristic (e.g., only enrolling non-smokers to eliminate confounding by smoking).
- B. Analyzing the data in different strata or layers.
- C. Randomly assigning participants to groups.
- D. Blinding the investigators.
Answer: A. Limiting study participants to those with a specific characteristic (e.g., only enrolling non-smokers to eliminate confounding by smoking).
39. When interpreting the results of any observational study, what should a pharmacist first consider?
- A. Only the p-value.
- B. Whether the results can be explained by bias, confounding, or chance, before concluding a causal relationship exists.
- C. The marketing claims of the drug company.
- D. How the results can be immediately applied to all patients.
Answer: B. Whether the results can be explained by bias, confounding, or chance, before concluding a causal relationship exists.
40. A cohort study can be used to examine multiple outcomes from a single exposure. A case-control study can be used to examine:
- A. Multiple outcomes from a single exposure.
- B. Multiple exposures for a single outcome.
- C. Only one exposure and one outcome.
- D. The incidence of a disease.
Answer: B. Multiple exposures for a single outcome.
41. In a cohort study on a new medication, patients who adhere poorly to the medication may be systematically different from those who adhere well. If adherence is also related to the health outcome, this can introduce:
- A. Recall bias
- B. Confounding (confounding by indication or healthy adherer effect)
- C. Interviewer bias
- D. No systematic error
Answer: B. Confounding (confounding by indication or healthy adherer effect)
42. The main difference between a confounder and an effect modifier is that:
- A. A confounder is a nuisance to be controlled, while an effect modifier is a real biological phenomenon to be described.
- B. A confounder is on the causal pathway, while an effect modifier is not.
- C. An effect modifier is associated with the exposure but not the outcome.
- D. They are the same concept.
Answer: A. A confounder is a nuisance to be controlled, while an effect modifier is a real biological phenomenon to be described. (Effect modification means the association between exposure and outcome differs across strata of the third variable).
43. To minimize selection bias when establishing a cohort, it is important that:
- A. The exposed and unexposed groups are selected from different populations.
- B. The exposed and unexposed groups are as comparable as possible on all factors except the exposure of interest.
- C. Only healthy individuals are selected for the study.
- D. Participants are allowed to select their own group.
Answer: B. The exposed and unexposed groups are as comparable as possible on all factors except the exposure of interest.
44. A researcher wants to study the long-term effects of a rare occupational exposure on mortality. Which study design would be most efficient?
- A. Case-control study
- B. Cross-sectional study
- C. Randomized controlled trial
- D. Cohort study (either prospective or retrospective)
Answer: D. Cohort study (either prospective or retrospective) (Cohort studies are good for rare exposures).
45. Which of the following is an example of confounding by indication?
- A. Patients are prescribed a drug for a specific reason (the indication), and that reason itself is a risk factor for the outcome being studied.
- B. A drug’s indication is changed by the FDA.
- C. A patient does not have the correct indication for a drug.
- D. The indication for a drug is a confounder.
Answer: A. Patients are prescribed a drug for a specific reason (the indication), and that reason itself is a risk factor for the outcome being studied.
46. If a study investigating the effect of a new blood pressure medication reports a biased result, this means the result is:
- A. Imprecise
- B. Inaccurate (systematically different from the true value)
- C. Not statistically significant
- D. Due to chance alone
Answer: B. Inaccurate (systematically different from the true value)
47. A major challenge in retrospective cohort studies compared to prospective cohort studies is:
- A. The high cost and long duration.
- B. The potential for poor quality or incomplete data on exposures, confounders, and outcomes from existing records.
- C. The inability to study rare diseases.
- D. The difficulty in assessing temporality.
Answer: B. The potential for poor quality or incomplete data on exposures, confounders, and outcomes from existing records.
48. Why can’t a cohort study definitively prove causation?
- A. Because it establishes temporality.
- B. Because as an observational design, it is always susceptible to the influence of unmeasured or unknown confounding variables.
- C. Because it calculates relative risk.
- D. Because it follows people over time.
Answer: B. Because as an observational design, it is always susceptible to the influence of unmeasured or unknown confounding variables.
49. Blinding is a technique primarily used to control for:
- A. Confounding
- B. Information bias (e.g., interviewer bias, recall bias) and performance bias
- C. Selection bias
- D. Random error
Answer: B. Information bias (e.g., interviewer bias, recall bias) and performance bias
50. The best way to deal with identified confounding in a study is to:
- A. Ignore it if the p-value is significant.
- B. Mention it as a limitation in the discussion section without any adjustment.
- C. Control for it in the design stage (e.g., restriction, matching) or adjust for it in the analysis stage (e.g., stratification, multivariable analysis).
- D. Increase the study’s sample size to wash it out.
Answer: C. Control for it in the design stage (e.g., restriction, matching) or adjust for it in the analysis stage (e.g., stratification, multivariable analysis).