Should I have a PSA test for prostate cancer?

This is the question where the honest answer has been unsatisfying for two decades, because the benefit is real and the harm is real and the two are close enough in size that the balance depends on the man. What has changed is that the trials are now long enough to be quoted in absolute numbers rather than percentages.

What the longest trial found#

The European Randomized Study of Screening for Prostate Cancer began in 1993 and reported its final long-term analysis in 2025. Its core group was 162,236 men aged 55 to 69 across eight countries, randomly assigned either to be offered repeated PSA testing or not invited at all.

Both halves of that table are the finding. Screening prevents prostate cancer deaths, and the same screening diagnoses roughly a third more prostate cancer than would otherwise have been found — cancer that in many men would never have surfaced in their lifetime.

What a single test achieves#

The CAP trial asked a narrower and more practical question: what does one invitation do? It randomised 573 primary care practices in England and Wales covering 415,357 men aged 50 to 69, offering the intervention group a single PSA test.

At a median 15 years, 0.69 percent of the invited men had died of prostate cancer against 0.78 percent of those not invited — a rate ratio of 0.92, statistically significant at P equals 0.03. The trial reports the reduction and describes the absolute difference as small.

A single test is not the same intervention as a screening programme, and the gap between CAP's 8 percent relative reduction and ERSPC's 13 percent is roughly what a difference in testing intensity would predict.

Why Korea's question may be a different one#

Almost all of the screening debate concerns overdiagnosis: finding too much harmless cancer. That is the debate Korea is famous for, because of thyroid cancer. Its prostate data does not fit the pattern.

• In 27,075 biopsy-confirmed Korean cases from 51 general hospitals, 53.9 percent were high-risk at the moment of diagnosis, and the low-risk share fell from 11.4 percent in 2010 to 7.6 percent in 2020.
• High-risk disease accounted for 55.4 percent of cases among rural residents against 47.7 percent among urban residents.
• Among 166,848 Korean men newly registered with prostate cancer, only 26.7 percent had been PSA-tested at least three times over at least two years beforehand, while for 42.2 percent the first test came within three months of registration.

A man whose first PSA test happens in the same quarter as his cancer registration was not screened. He was investigated because something had already gone wrong. That is the profile of a country finding prostate cancer late, which is the opposite failure to the one the screening trials were designed to prevent.

How to use the numbers#

1. The benefit is real and small in absolute terms. Over 23 years, screening changed the chance of dying of prostate cancer by about two in a thousand.
2. The cost is a substantially higher chance of being told you have cancer. Screening raised diagnoses by about 30 percent, and a diagnosis carries consequences whether or not it is treated.
3. The trials tested organised programmes with repeated testing in men aged 50 to 69. They say nothing directly about a single opportunistic test in a man of 75.
4. ERSPC's authors conclude that future strategies should be risk-based, to keep the benefit while reducing overdiagnosis. That is a recommendation about programme design, not about whether one man should be tested.
5. What a PSA result leads to matters as much as the test. A man who would accept active surveillance for a low-risk cancer faces a different balance from one who would want immediate surgery.

The defensible position is that this is a decision to be made rather than a test to be defaulted into, in either direction. The numbers above are the ones a man is entitled to have before making it.