Overdiagnosis and Screening vs Diagnosis

Screening and diagnosis answer different questions

Screening looks for a condition in people who have no symptoms. The National Cancer Institute (NCI) describes cancer screening as looking for cancer before a person has any symptoms, and notes that screening tests usually do not diagnose cancer. [1] The NHS describes screening as identifying apparently healthy people who may have an increased risk of a condition; a higher-risk result is followed by diagnostic tests to confirm whether the condition is present. [2]

Diagnostic testing looks for the cause of symptoms or checks a result that screening flagged.

This page is information only. Whether to be screened is a decision to make with a clinician, using current guidelines for your age and history.

What overdiagnosis means

Overdiagnosis is the diagnosis of a condition that would otherwise not have caused symptoms or death. A 2010 review in the Journal of the National Cancer Institute says it needs two things: a silent reserve of disease, and activity, particularly screening, that detects it. [3] NCI explains that screening can find cancers that would have gone away on their own or never caused symptoms, and calls this overdiagnosis. [1]

Overdiagnosis is different from a false positive. A false positive is a wrong alarm. An overdiagnosed cancer is real but would never have mattered. It can lead to overtreatment, with treatment side effects and no benefit. [4]

How big is the problem?

The 2010 review estimated from randomized trials that overdiagnosis accounted for about 25% of mammography-detected breast cancers, 50% of cancers detected by chest x-ray or sputum tests, and 60% of PSA-detected prostate cancers. It also reviewed observational data suggesting overdiagnosis with CT-detected lung cancer and thyroid, melanoma and kidney cancers, and said better estimates are needed. [3] These are older estimates from particular trials and methods, so treat them as mixed evidence on size, not a settled figure.

Why screening statistics can mislead

NCI describes three biases that can make screening look better than it is. [1]

• Lead-time bias. Diagnosis happens earlier, so survival measured from diagnosis looks longer even if the date of death does not change.
• Overdiagnosis. Cancers that would never have caused harm add to the survivors.
• Healthy screenee bias. People who choose screening are often more health-conscious.

That is why "five-year survival" after screening-detected cancer is not proof screening saves lives. The better question is whether deaths from the disease fall in randomized comparisons. NCI notes that a useful screening test decreases the chance of dying from the cancer and has few false results. [1]

A specific example: PSA

NCI says PSA is not recommended for routine prostate cancer screening in the general population. Most organizations now recommend that people considering it first discuss risks and benefits with their doctor. NCI says some of the cancers it finds grow so slowly that they would never cause symptoms, and treating them can cause urinary, bowel and sexual side effects. [4]

Questions to ask before a screening test

• What disease is it looking for, and what happens if the result is positive?
• What are the chances of a false alarm, and of finding something that would never have caused harm?
• Is there evidence from trials that screening reduces deaths or serious illness?
• Is this test meant for someone with my age, risk and symptoms?

For more detail on specific screening tests, see cancer screening tests and multi-cancer detection and early-detection tests and multi-cancer screening research. For the numbers behind false alarms see false positives, multiple testing and incidental findings.