Multiple Testing and Overdiagnosis

More tests, more chance of an "abnormal" flag

Reference ranges are usually set so that 5% of healthy people fall outside them. [1] Add more tests and the chances add up. A clinical review calculates that, for independent tests, 64.2% of "normal" people will have at least one abnormal result in a panel of 20 tests. [1] The calculation assumes the tests are independent; real panels include correlated tests, so actual figures differ, and the number describes statistical chance, not disease.

The practical point is that comprehensive direct-to-consumer panels, longevity clinic panels and frequent self-testing are likely to flag something in a healthy person. See False Positives, Multiple Testing and Incidental Findings and Reference Ranges vs "Optimal" Ranges.

What happens after a flag

The same review stresses that all results should be interpreted in clinical context so that no one is given a diagnosis based only on an abnormal value. [1] The National Cancer Institute describes the typical sequence for a false positive: anxiety, then more tests and procedures that carry their own risks. [2] That cascade is a cost of testing, not just an inconvenience.

What overdiagnosis means

Overdiagnosis occurs when a true abnormality is found, but detecting and treating it does not benefit the patient. [3] Put differently, it is the identification of conditions that would otherwise never have caused significant symptoms or harm. [4] The NCI uses the term for screening-detected cancers that would have gone away or never caused symptoms, and notes it can make survival look longer when it is not. [2]

Overdiagnosis is different from a false positive. In a false positive nothing is wrong; in overdiagnosis something real is found, but finding it does not help.

Drivers of overdiagnosis

• Looking harder in healthy people. Screening is meant for apparently healthy people who may have increased risk, and the NHS notes screening tests are not perfect and can lead to difficult decisions. [5]
• Lowering thresholds. A review describes how redefining a disease or risk threshold can create overdiagnosis, giving the cholesterol threshold change as an example and estimating 14% to 33% of people overdiagnosed with hypertension under certain guidelines. [4]
• Genetic and chronic-condition testing. Overdiagnosis is encountered in cancer screening, genetic disease and chronic conditions when definitions broaden. [3]

Consumer testing and self-tracking

Many of the tests in this library, such as home blood kits or biological-age scores, are not designed or validated as screening programs. Screening programs are chosen for groups likely to benefit and judged by outcome evidence; the NHS says an expert committee advises on which programs to offer. [5] Compare with Overdiagnosis and Screening vs Diagnosis, Home Blood-Sampling Kits and Biomarkers vs Clinical Outcomes.