Seeing a patient who might be having an adverse effect of a medication may lead to a clash of perspectives. For example, from the viewpoint of a cardiologist caring for someone with an acute coronary syndrome, it is all too easy to blame the internist or rheumatologist who prescribed a cyclo-oxygenase-2 (COX-2) inhibitor for a patient’s arthritis. This is yet another example of “practice variation,” or slow knowledge translation. However, such a judgement ignores the complexity of the evidence that putatively links COX-2 inhibitors to heart disease. To be sure, there is evidence that COX-2 inhibitors can increase risk of cardiovascular (CV) disease, but what may be overlooked is that the extent of such risk varies substantively among patients, depending on the drug, dose, dosing frequency, duration, and outcome definition. Furthermore, the balance of benefits and risks for patients depends on their unique clinical states and circumstances,1 the success or failure of alternatives, and the patient’s wishes.2 Applying the …