Start with the status quo, which nobody should defend. Routine screening exists for only a handful of cancers. Together, recommended screening programs catch only about one in seven cancers diagnosed in the United States. Nearly 70% of cancer deaths come from cancers for which there is no recommended screening test at all.

Pancreatic, ovarian, liver, stomach, esophageal and blood cancers often reveal themselves only after symptoms appear, when the disease has progressed and treatment options have narrowed.

New multi-cancer early detection, or MCED, tests are designed to help close that gap. Rather than looking for a single cancer in a single organ, they analyze signals in the bloodstream associated with many different cancers. They examine the whole person.

The Food and Drug Administration will soon make an important decision around how it evaluates this technology as it considers Galleri, an MCED blood test developed by GRAIL. The agency should give a careful, clear-eyed, pragmatic look at the evidence. 

These tests are new. They will miss some cancers. Some positive results will lead to scans, biopsies and other procedures that ultimately prove unnecessary. And MCED tests should supplement, not replace, established screening.

But regulators should also judge them against the real-world alternative.

Today, a patient can do everything right, can follow every recommended screening guideline and still have no routine way to detect many deadly cancers. The relevant policy choice is not between a perfect test and an imperfect one. It is between the information a new test can provide and the information patients have without it.

In a recent large study, Galleri produced a false-positive rate of roughly 0.4% — dramatically lower than any established single-cancer screen. Its positive predictive value was about 60%, meaning most positive results reflected actual cancer. The test can also help identify where in the body a cancer signal originated, allowing physicians to pursue a more focused diagnostic workup rather than a full-body fishing expedition.

But some critics want the FDA to demand something much more before such tests become broadly available: proof that screening ultimately reduces cancer mortality.

That may sound like an appropriately high scientific bar. But it applies the wrong standard to a diagnostic tool.

A treatment should be judged by whether it improves patients’ health. A diagnostic test has a different function: producing accurate information that doctors and patients can use to make better decisions. Whether that information ultimately extends a patient’s life also depends on the type of cancer, when it is detected, available treatments and the decisions made after diagnosis.

We do not judge a thermometer by whether it cures a fever. We judge it by whether it accurately measures temperature and helps guide what happens next. Cancer screening should be evaluated with the same basic logic.

Economists have long recognized that better information has value precisely because it changes behavior. Learning that a dangerous cancer is present before symptoms appear can change when treatment begins and what options remain available. Requiring the test itself to demonstrate a reduction in mortality effectively assigns little value to that information until the entire chain of subsequent medical decisions and outcomes has been proven. Notably, existing single-cancer screens did not uniformly face such a standard before adoption.

Holding new technology to a bar that existing technology never cleared would be bad science, leading to inertia that is extremely harmful to cancer patients. 

That means thinking differently about how MCED tests are measured. 

Traditional single-cancer screening emphasizes sensitivity — the probability that a test identifies a particular cancer when it is present. That makes sense when a test is designed to find one disease.

An MCED test has a different purpose. It searches across many cancers simultaneously. One important measure is therefore its overall yield: how many cancers in a screened population are found through screening rather than after symptoms emerge.

A test that materially raises the share of cancers caught by screening rather than by symptoms – while keeping harms low in an otherwise healthy population – is doing its job, whatever its performance on any single tumor type.

The questions worth asking are answerable with the evidence in hand. Does the test find cancer early? Does it catch disease before symptoms appear, and does it increase detection in stages I-III, when curative intent is still possible? Does it detect aggressive, fast-moving cancers with no screening options today? Is it safe to deploy?

There are economic consequences to consider as well. American medicine spends enormous sums treating advanced disease after it has already inflicted substantial damage. Finding cancer earlier may allow treatment when disease is more manageable, avoiding some costly hospitalizations and complications associated with later-stage illness.

Congress has already cleared the way for Medicare to cover FDA-approved MCED tests beginning in 2029. What remains is the FDA decision allowing doctors and patients to do what’s best with the better information at hand.

American medicine has spent decades perfecting a break-it-and-fix-it model: wait for disease to declare itself, then spend heavily to fight it on worse terms. Prevention, prediction and early detection offer a better path.

Earlier knowledge will not cure cancer. But it can give doctors and patients a chance to act before cancer gets the first move.

Mr. Philipson is an economist at the University of Chicago and a senior fellow at Unleash Prosperity. He served as a member and acting chairman of the White House’s Council of Economic Advisers, 2017-20.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.