A new blood biomarker strongly predicts cardiovascular events in a large cohort study (p < 0.001, hazard ratio 2.4). A hospital committee concludes it should be added to routine screening. Which response best identifies the flaw in this reasoning?
AThe hazard ratio is too small to be clinically meaningful
BA p-value below 0.001 does not indicate statistical significance
CStatistical association with outcomes does not prove the biomarker improves risk classification beyond existing models
DCohort studies cannot be used to validate biomarkers because they lack randomization
The key distinction is between statistical association and clinical utility. A biomarker can strongly predict outcomes yet add nothing beyond what the Framingham Risk Score or Pooled Cohort Equations already classify. Clinical usefulness requires demonstrating discrimination (correctly separating high-risk from low-risk patients) and reclassification improvement — moving people into different risk categories in ways that change management. Many biomarkers that looked promising on hazard ratios failed this more demanding test.
Question 2 Multiple Choice
A 35-year-old and a 65-year-old both have a relative risk of 3.0 for coronary heart disease due to hypertension. For which patient does treating the hypertension prevent more absolute events per 100 people treated?
AThe 35-year-old, because they have more life-years ahead
BThe 65-year-old, because their higher baseline risk means the same relative elevation represents far more absolute events
CBoth equally, because the relative risk is identical
DThe 35-year-old, because hypertension does more cumulative damage in young arteries
This is the crucial distinction between relative and absolute risk. If the 35-year-old has a 10-year baseline CHD risk of 2%, a relative risk of 3.0 raises that to 6% — an absolute increase of 4 percentage points. If the 65-year-old has a baseline risk of 20%, the same relative risk of 3.0 raises that to 60% — an absolute increase of 40 percentage points. Treating 100 people in the first group prevents ~4 events; treating 100 in the second prevents ~40. Absolute risk, not relative risk, determines the cost-benefit calculus of clinical intervention.
Question 3 True / False
A researcher studying 'cardiovascular disease' mortality should ideally analyze coronary heart disease, stroke, and heart failure as a single combined outcome to maximize statistical power.
TTrue
FFalse
Answer: False
Lumping CVD subtypes obscures important subtype-specific differences. Atrial fibrillation is a powerful stroke risk factor but weakly linked to CHD; LDL cholesterol strongly predicts CHD but is a weaker predictor of hemorrhagic stroke; heart failure often occurs downstream of prior CHD or hypertension. Combining these into one outcome can mask heterogeneous associations and lead to incorrect conclusions about which risk factors matter for which diseases.
Question 4 True / False
Atrial fibrillation is a stronger independent risk factor for ischemic stroke than for coronary heart disease.
TTrue
FFalse
Answer: True
Atrial fibrillation causes cardioembolic stroke — irregular electrical activity leads to blood pooling and clot formation in the atria, which can travel to cerebral arteries. This mechanism is specific to the stroke pathway. While AF can contribute to heart failure, it is not a major independent risk factor for atherosclerotic coronary disease in the same way. This is why stroke risk assessment tools (CHA₂DS₂-VASc) weight AF heavily, while Framingham CHD scores do not include it as a primary input.
Question 5 Short Answer
Why must cardiovascular epidemiologists analyze CVD subtypes separately rather than treating cardiovascular disease as a single unified disease category?
Think about your answer, then reveal below.
Model answer: CVD subtypes have distinct etiologies, risk factor profiles, and pathophysiological mechanisms. Coronary heart disease is driven by atherosclerotic plaque in coronary arteries; ischemic stroke by thromboembolic events in cerebral vasculature; hemorrhagic stroke by vessel rupture; heart failure often by downstream effects of prior CHD or hypertension. Risk factors like LDL cholesterol, atrial fibrillation, and hypertension have different magnitudes of effect across subtypes. Pooling them hides these differences and can produce misleading estimates of which exposures matter — and for whom.
Subtype-specific analysis is essential because an intervention optimized against 'CVD' might be excellent for one subtype and harmful or irrelevant for another. Statins markedly reduce CHD and ischemic stroke risk but have little effect on hemorrhagic stroke. Blood thinners for AF dramatically cut stroke risk but add bleeding risk. Epidemiological rigor requires matching the outcome to the mechanism being studied.