Screening is a mainstay of modern preventive medicine. Physicians routinely screen patients for breast or colorectal cancer, as well as hypertension, hypercholesterolemia, and abdominal aortic aneurysm. In each of these cases, randomized trials demonstrated that screening saves lives or prevents major morbidity. As coronary artery disease remains the number one cause of death in Western society, there is interest in screening for the disease itself, as opposed to just its risk factors. Some investigators have argued for routine screening of adults with coronary artery calcium testing; proponents have used direct-to-consumer advertising, legislative lobbying, and appeals in the mass media. Although it is true that coronary calcium scores correlate well with clinical events, the evidence base has not yet reached the level of established tests. We may fall into a too-often ignored trap, where strategies based on logic or risk markers turn out to be ineffective or even dangerous. When a seemingly healthy person undergoes a screening test and is found to have disease, that person becomes a patient subject to clinical thinking. If this translates into improved outcome, clinical thinking is appropriate. We must remember, though, that screening is supposed to prevent clinical events. When it merely detects subclinical disease, leading to unnecessary tests or procedures, it instead exposes people to the risks of thinking clinically. We must have the ambition, willingness, and patience to develop the evidence base needed to introduce new screening technologies to the clinic at the right time.
Genetic analysis of coronary artery disease using tree-based automated machine learning informed by biology-based feature selection
