Why Ordering More Tests Is Not Automatically Safer

Does ordering more tests make care safer?

Not automatically, and often the opposite. Every test has a false-positive rate, and when the chance a person actually has the disease is low before testing, even an accurate test produces mostly false alarms. Those false positives are not harmless paperwork. They set off biopsies, repeat scans, referrals, and weeks of fear in people who were never sick. The instinct to "just check" feels cautious, but the math and the outcome data both show that testing without a reason can create the very harm it was meant to prevent.

This is one of the least intuitive ideas in medicine, so it is worth walking through slowly. A test result does not tell you the truth about a body. It shifts a probability. How much it shifts, and in which direction, depends on where that probability started.

Pretest probability sets the meaning of the result

Before any test is run, a clinician can estimate the pretest probability: the chance this person has the condition, based on symptoms, history, and how common the disease is in people like them. As the American Society for Microbiology explains in its primer on pretest and posttest probability, a test result then moves that starting estimate up or down to a posttest probability. The same positive result means very different things depending on where you began.

Two properties describe how a test performs. Sensitivity is how well it catches disease in people who truly have it. Specificity is how well it clears people who truly do not. A highly specific test rarely flags healthy people, so a positive result tends to confirm disease. A highly sensitive test rarely misses, so a negative result tends to rule it out. Neither property is fixed by the test alone in the way most people assume, because the usefulness of a positive result also depends on how many people being tested actually have the disease.

Why low probability plus imperfect specificity breeds false positives

Here is the part that trips up smart people. Imagine a test with 99 percent specificity, meaning it wrongly flags only 1 healthy person in 100. That sounds nearly perfect. Now apply it to 1,000 people among whom only 5 truly have the disease. The test correctly flags most of those 5. But it also wrongly flags about 1 percent of the 995 healthy people, which is roughly 10 false positives. So out of about 15 positive results, two-thirds belong to people with no disease at all.

Nothing was wrong with the test. What changed the outcome was the low pretest probability. When disease is rare in the group being tested, false positives outnumber true ones, and a positive result carries far less weight than the raw specificity number suggests. The ASM primer makes this point directly: in a low-prevalence setting, the chance of a false positive rises, and testing people with very low pretest probability tends to add confusion rather than clarity. The number that captures this is positive predictive value, the share of positive results that are real, and it collapses as pretest probability falls.

The false positive is the start of a cascade, not the end

A false positive rarely resolves with a shrug. It opens what clinicians call a cascade of care: one unexpected or borderline result leads to another test, then a specialist, sometimes a biopsy or a procedure. A national survey of US physicians published in JAMA Network Open in 2019 found that almost all responding internists had experienced these cascades after incidental findings, and that many cascades led nowhere clinically useful while still causing patients psychological harm, physical harm, and financial burden. Notably, about a third of physicians said the test that started their most recent cascade may not have been necessary in the first place.

The harm is concrete when you follow a real screening pathway. In its 2018 recommendation on prostate cancer screening, the US Preventive Services Task Force reported that approximately 1 in 6 men screened at least once had one or more false-positive results, that follow-up biopsies carry risks including pain, infection, and occasional hospitalization, and that 20 to 50 percent of cancers found through screening may be overdiagnosed, meaning they would never have caused symptoms. Treatment of those cancers carries lasting consequences, including urinary and sexual dysfunction. This is why the Task Force does not tell every man to get screened. It grades screening for men aged 55 to 69 as an individual decision to be made after discussing benefits and harms, and recommends against routine screening in men 70 and older, where false positives and downstream harm rise further.

Restraint is a clinical skill, not neglect

None of this argues against testing. It argues for testing with a reason. The value of a test comes from having a real question it can answer, which usually means a meaningful pretest probability that a positive result would raise and a negative result would lower in a way that changes what happens next. A test ordered to rule out an anxiety rather than a diagnosis often does neither.

The Choosing Wisely campaign, launched by the ABIM Foundation, built an entire framework around this idea. Working with more than 80 specialty societies, it published hundreds of recommendations naming common tests and procedures that patients and clinicians should question, guided by whether care is supported by evidence, free from harm, not duplicative, and truly necessary. The campaign reframes the conversation away from "more is safer" toward "is this test going to help this person," which is the question the probability math has been pointing at all along.

For a reader, the practical takeaway is a question worth asking out loud: what will we do differently depending on the result? If the honest answer is nothing, or if a positive result is more likely to be a false alarm than a real finding, then the safer choice may be the test not ordered. This is educational information, not medical advice, and specific decisions belong with a person and their own clinician who knows their history.