Cohort studies Practice Questions

Q: A researcher wants to determine whether there is an association between CRP values and the risk of MI or cancer. Four relative risk (RR) values were plotted $(0.5,1.5,1.7,1.8)$ with respect to CRP levels. What conclusion can be drawn?

CRP increases disease/cancer risk. ***CRP increases disease/cancer risk*** - A **relative risk (RR)** greater than 1 indicates an increased risk of the outcome (MI or cancer) in the exposed group (higher CRP levels) compared to the unexposed group. - The plots show RRs of 1.5, 1.7, and 1.8, all of which are greater than 1, consistently indicating that higher CRP levels are associated with an elevated risk for MI or cancer. - The overall trend across the four intervals demonstrates a positive association between CRP and disease risk. *CRP has no relationship* - This conclusion is incorrect because three of the four plotted RR values (1.5, 1.7, 1.8) are above 1, indicating a positive association or increased risk. - An RR of 1 signifies no relationship, but the majority of values clearly deviate from 1, showing a definite association. *CRP decreases & disease decreases* - While one RR value (0.5) suggests a decreased risk, the majority of the given RRs (1.5, 1.7, 1.8) are greater than 1, indicating an increased risk. - This option would only be true if all or most RR values were less than 1, implying a protective effect, which is not the overall trend here. *No association in first interval* - The first interval shows an RR of 0.5. An RR of 1 indicates no association, while an RR of 0.5 actually indicates a **decreased risk or protective effect**, rather than no association. - Therefore, stating "no association" for the first interval is inaccurate given the definition of relative risk. *CRP shows protective effect in first interval* - While the first interval RR of 0.5 does suggest a protective effect in isolation, this option fails to capture the **overall conclusion** from all four data points. - When interpreting multiple RR values together, the predominant pattern (three values >1) indicates an overall increased risk, making this a misleading conclusion for the study as a whole.

Q: A researcher wants to study the carcinogenic effects of a food additive. From the literature, he finds that 7 different types of cancers have been linked to the consumption of this food additive. He wants to study all 7 possible outcomes. He conducts interviews with people who consume food containing these additives and people who do not. He then follows both groups for several years to see if they develop any of these 7 cancers or any other health outcomes. Which of the following study models best represents this study?

Cohort study . ***Cohort study*** - This study design involves selecting a group based on their **exposure status** (consumers vs. non-consumers of the food additive) and **following them forward in time** to observe the incidence of outcomes (cancers). - It is ideal for studying **multiple potential outcomes** from a single exposure and for establishing the **temporal relationship** between exposure and disease. *Case-control study* - This design starts by identifying individuals with a particular **outcome (cases)** and comparing them to individuals without the outcome (controls) to look back for **past exposures**. - It is efficient for studying **rare diseases** or when multiple exposures are suspected for a single outcome, which is inverse to the scenario described. *Cross-sectional study* - This study measures both **exposure and outcome simultaneously** at a single point in time, providing a snapshot of prevalence. - It cannot establish a **temporal relationship** between exposure and outcome and is less suitable for studying incident diseases that develop over time. *Randomized clinical trial* - This design involves **randomly assigning participants** to an intervention group or a control group and following them for outcomes. - It is primarily used to evaluate the **efficacy of interventions** or treatments, not to study the carcinogenic effects of naturally occurring exposures. *Crossover study* - In a crossover design, participants **receive all interventions** in a specific sequence, making each subject serve as their own control. - This design is generally used for evaluating **short-term effects of treatments** in chronic, stable conditions and is unsuitable for observing the development of diseases like cancer over extended periods.

Q: A 52-year-old man comes to the physician for a follow-up examination 1 year after an uncomplicated liver transplantation. He feels well but wants to know how long he can expect his donor graft to function. The physician informs him that the odds of graft survival are 90% at 1 year, 78% at 5 years, and 64% at 10 years. At this time, given that the graft has already survived 1 year, the probability of the patient's graft surviving to 10 years after transplantation is closest to which of the following?

71%. ***71%*** - This question tests understanding of **conditional probability** in survival analysis. - The patient is currently at 1 year post-transplant with a functioning graft. We need to calculate the probability of surviving to 10 years **given survival to 1 year**. - Using the conditional probability formula: P(survive to 10 years | survived to 1 year) = P(S10) / P(S1) = 64% / 90% = 0.711 ≈ **71%** - This represents the probability that a graft that has already "made it" through the first year will continue functioning until year 10. - In **Kaplan-Meier survival analysis**, conditional probabilities are crucial for counseling patients at different timepoints post-procedure. *64%* - This represents the **absolute probability** of 10-year graft survival measured from the time of transplantation (time zero). - However, the question asks "at this time" (1 year post-transplant) for the conditional probability, not the absolute probability from transplantation. - This would be correct if asking a patient at time zero what their 10-year survival odds are. *82%* - This does not represent any valid calculation from the given survival data. - It may result from incorrect manipulation of the probabilities (e.g., incorrectly adding or averaging values). *58%* - This is not derived from proper statistical calculation of the given survival probabilities. - It does not represent either absolute or conditional probability for any relevant timepoint. *45%* - This is incorrect and does not correspond to any valid calculation. - It might arise from incorrectly multiplying probabilities (e.g., 0.90 × 0.50) but has no basis in survival analysis.

Question 1

A clinician wants to select the single cut-off point on a diagnostic biomarker's ROC curve that minimizes the Euclidean distance to the upper-left corner (0,1), thereby simultaneously minimizing the false-negative rate and the false-positive rate. Which of the following best describes this optimal cut-off point selection criterion on the ROC curve?

Accepted Answer

The point on the curve closest to the upper-left corner (0,1)

Answer

The point at which the curve crosses the diagonal reference line

Answer

The point corresponding to the highest specificity regardless of sensitivity

Answer

The point at which sensitivity equals the positive predictive value

Answer

The point at which the area under the curve to the left equals the area to the right

Question 2

A researcher wants to determine whether there is an association between CRP values and the risk of MI or cancer. Four relative risk (RR) values were plotted $(0.5,1.5,1.7,1.8)$ with respect to CRP levels. What conclusion can be drawn?

Accepted Answer

CRP increases disease/cancer risk

Answer

CRP has no relationship

Answer

CRP decreases & disease decreases

Answer

No association in first interval

Answer

CRP shows protective effect in first interval

Question 3

A researcher wants to study the carcinogenic effects of a food additive. From the literature, he finds that 7 different types of cancers have been linked to the consumption of this food additive. He wants to study all 7 possible outcomes. He conducts interviews with people who consume food containing these additives and people who do not. He then follows both groups for several years to see if they develop any of these 7 cancers or any other health outcomes. Which of the following study models best represents this study?

Accepted Answer

Cohort study

Answer

Case-control study

Answer

Cross-sectional study

Answer

Randomized clinical trial

Answer

Crossover study

Question 4

A 52-year-old man comes to the physician for a follow-up examination 1 year after an uncomplicated liver transplantation. He feels well but wants to know how long he can expect his donor graft to function. The physician informs him that the odds of graft survival are 90% at 1 year, 78% at 5 years, and 64% at 10 years. At this time, given that the graft has already survived 1 year, the probability of the patient's graft surviving to 10 years after transplantation is closest to which of the following?

Accepted Answer

71%

Answer

82%

Answer

58%

Answer

64%

Answer

45%

Question 5

The study is performed to examine the association between type 2 diabetes mellitus (DM2) and Alzheimer's disease (AD). Group of 250 subjects diagnosed with DM2 and a matched group of 250 subjects without DM2 are enrolled. Each subject is monitored regularly over their lifetime for the development of symptoms of dementia or mild cognitive impairment. If symptoms are present, an autopsy is performed after the patient's death to confirm the diagnosis of AD. Which of the following is most correct regarding this study?

Accepted Answer

It is a prospective observational study.

Answer

It is a retrospective observational study.

Answer

Relative risk cannot be determined from this study.

Answer

It can provide proof of causation between DM2 and AD.

Answer

It is a case-control study.

Cohort studies — MCQs

Cohort studies — MCQs

On this page

Practice by Chapter

Want unlimited practice?