Algorithms in Action
In many cases, physicians rely on medical algorithms as part of an ongoing process of patient assessment and care.
Algorithms pervade medicine. They come in forms as simple as the ubiquitous equation for calculating body mass index and as complex as a system that uses patient data to predict the probability of seizures in people with epilepsy. There are thousands to choose from; advocates of integrating carefully vetted algorithms into health care say these tools could help save time and money.
The hypothetical example offered here illustrates how algorithms could help improve care for a patient with ovarian cancer.
A 36-year-old woman complains of swelling, pain and redness in her limbs—classic signs of deep vein thrombosis, a blood clot in the veins. Her physician, having ruled out a number of risk factors, such as smoking and oral contraceptives, considers the possibility that she has Trousseau’s syndrome, which is DVT caused by cancer (tumors can release substances that trigger clotting). But because evaluating a patient for cancer is expensive, it pays to carefully assess the risk of a tumor.
A checklist algorithm can aggregate vague, inconclusive symptoms and raise a red flag if several occur together. Without it, a physician could under- or overestimate the possibility of cancer.