He arrived with his PSA on his phone.

He is 58, has a PSA of 4.6 and three tabs open on his phone: a forum, a clinical trial and a press release that contradict one another. He reaches the consultation knowing more than his father did at his age and deciding worse, because the evidence does not give a single answer, that is, because the benefit is real and the harm is real too, and they fall on different men. In October 2025 the European trial ERSPC published its 23-year follow-up, and the numbers on his screen deserve a slow read.
Inviting men aged 55 to 69 to repeated PSA screening reduced prostate cancer mortality with a rate ratio of 0.87, 13% lower. In absolute terms, the reduction was 0.22 percentage points. To avoid one death, 456 men had to be invited and 12 diagnosed, with wide intervals. At 16 years those figures were 628 and 18, so the benefit improves over time, but the excess of diagnoses does not disappear: cumulative incidence was 30% higher in the invited group. That excess is not, one for one, measured overdiagnosis, and the protocol predates MRI; for that reason the authors themselves propose risk-based screening.
The other half of the story comes from ProtecT, the British trial that followed for 15 years 1,643 men with localised cancer detected by PSA, assigned to active monitoring, surgery or radiotherapy. Forty-five died of prostate cancer, 2.7%: 17 with monitoring, 12 with surgery and 16 with radiotherapy, with no statistical difference. And here precision matters: the absence of a difference does not prove equivalence. Radical treatment did reduce progression and metastases, 9.4% with monitoring versus 4.7% with surgery and 5.0% with radiotherapy. What ProtecT shows is that, in that mostly low- and intermediate-risk population, some men could defer treatment without paying for it with their lives: 24.4% of those monitored were still alive and untreated at the end.
Read together, the two trials tell him something uncomfortable: PSA saves lives, few in absolute terms, and in exchange diagnoses many; and if a localised cancer is found, there will be men for whom watching is reasonable and men for whom it is not, and nobody will know in advance which group he is in. That is not solved with another tab. It is solved with twenty minutes of conversation with someone who can explain what 0.22% means and what 12 to 1 means.
For those who manage, the consequence is one of design. A pathway that turns every raised PSA into a biopsy, every biopsy into a diagnosis and every diagnosis into treatment is not individualising by risk, nor using MRI where it belongs, nor giving him the time to decide. That conversation has a cost, and almost no fee schedule pays for it. Whether financing it reduces unnecessary procedures is a hypothesis an organisation can test; it is not a demonstrated saving.
Hyper-informed, we decide under uncertainty. The question is who pays him for the time to decide.
Sources: Roobol MJ, de Vos I, Månsson M, et al. European Study of Prostate Cancer Screening — 23-year follow-up. N Engl J Med 2025;393(17):1669-80. https://doi.org/10.1056/NEJMoa2503223 (rate ratio 0.87, 95% CI 0.80–0.95; absolute reduction 0.22 pp, 95% CI 0.10–0.34; 456 invited, 95% CI 306–943; 12 diagnosed, 95% CI 8–26; incidence ratio 1.30, 95% CI 1.26–1.33). · Hamdy FC, Donovan JL, Lane JA, et al. Fifteen-year outcomes after monitoring, surgery, or radiotherapy for prostate cancer. N Engl J Med 2023;388(17):1547-58. · MINSAL. Clinical guideline or GES problem for prostate cancer and its position on screening.