Prostate Screening: The Decision Nobody Can Make For You

The advice that prompted this

In June 2026 Jeremy Clarkson announced he was in remission from prostate cancer, an aggressive tumour caught early, and used the moment to tell other men to get tested. If a GP turns you away because you have no symptoms, he said, invent some — tell them you are up half the night needing to urinate.

The impulse is decent and I understand it completely. A man who believes a test saved his life wants other men to have the test. But the specific advice is wrong in a way that is worth explaining, because it reveals what the argument is actually about.

Inventing symptoms does not get you screened. It moves you out of the screening pathway and into the symptomatic one, where different clinical reasoning applies, different thresholds are used, and the doctor in front of you is now working from a history that is not true. You have not obtained a test you were being denied. You have degraded the information your clinician is using to decide what to do with you.

What changed in 2026

In May 2026 the UK National Screening Committee published its final recommendation. It does not recommend screening all men, on the grounds that PSA is not reliable enough as a screening test and the evidence does not show that screening everyone would do more good than harm.

What it does recommend is a narrow targeted programme: men aged 45 to 61 who both carry a BRCA2 variant and have a family history of prostate, breast, ovarian or pancreatic cancer, offered a PSA test every two years. In June 2026 the UK government accepted this for England, with rollout planned for 2027. Scotland, Wales and Northern Ireland are still considering it — so if you are reading this in Edinburgh, as I am writing it, nothing has yet changed here.

One detail is worth pausing on. The November 2025 draft also included men with BRCA1 variants. The final recommendation removed them after consultation and further evidence. A committee narrowing its own proposal in public, on evidence, is not a failure of the process. It is the process.

What PSA actually is

Prostate specific antigen is a protein made by prostate tissue — normal prostate tissue as well as cancerous. It is specific to the prostate. It is not specific to cancer, and the difference between those two sentences is the whole problem.

PSA rises with benign prostatic enlargement, with prostatitis, with urinary infection, with recent ejaculation, with vigorous cycling, and simply with age. It can also be unremarkable in a man who has aggressive disease. Among men who get a raised PSA and go through further investigation, roughly 72 to 80 per cent turn out not to have prostate cancer at all.

Richard Ablin, who discovered PSA in 1970, spent years publicly criticising its use as a population screening test, describing that use in 2010 as a public health disaster. He was not disowning the protein — PSA is genuinely useful for monitoring a man already known to have cancer. He was objecting to a test that cannot distinguish the cancers that kill from the cancers that never would being used to decide who gets investigated. That criticism is now sixteen years old and predates MRI-first pathways, which is worth remembering when it is quoted, as it very often is.

There is also a measurement problem that almost nobody raises. Any marker has biological variation — a range it moves within, in the same person, for no clinical reason at all. Before you can say a PSA has genuinely risen between two draws, the change has to exceed that variation plus the assay’s own imprecision. A single number, read against a fixed threshold, tells you less than the confidence with which it is usually delivered.

What the trials show — the honest version

Two large randomised trials carry most of the weight, and they are often quoted selectively by both sides. Here is what each actually found.

ERSPC, 23-year follow-up

The European Randomized Study of Screening for Prostate Cancer followed a core group of 162,236 men aged 55 to 69 across eight countries. At a median of 23 years:

Finding	Result
Prostate cancer mortality	13% lower in the screened group (rate ratio 0.87, 95% CI 0.80–0.95)
Absolute risk reduction	0.22% (95% CI 0.10–0.34)
Cancer incidence	30% higher in the screened group (rate ratio 1.30, 95% CI 1.26–1.33)
Men invited per death prevented	456 (95% CI 306–943)
Men diagnosed per death prevented	12 (95% CI 8–26)

Both sides can quote this and both are being accurate. A 13 per cent relative reduction sounds substantial. An absolute reduction of 0.22 per cent sounds negligible. They are the same finding.

The number I would put in front of anyone weighing this is the last one: twelve men diagnosed with prostate cancer for every one death prevented. Diagnosis is not a neutral event. It brings biopsy, surveillance or treatment, and treatment carries real risks of urinary incontinence and erectile dysfunction. Eleven of those twelve men carried a cancer diagnosis, and whatever followed it, without their death being the thing that was averted.

And — against my own instinct — that ratio has improved with time. At 16 years it was 628 invited and 18 diagnosed per death prevented. At 23 years it is 456 and 12. Screening looks better the longer you follow it, because the benefit accrues slowly while the harms arrive at once. Anyone citing the older figures to argue against screening is using out-of-date numbers, and I have seen it done.

CAP, the British trial

The Cluster Randomised Trial of PSA Testing enrolled 415,357 men aged 50 to 69 across 573 primary care practices in England and Wales, and tested something different: a single invitation to a PSA test, which is much closer to what actually happens when a man asks his GP once.

At 15 years, 0.69 per cent of the invited group had died of prostate cancer against 0.78 per cent of the control group — a rate ratio of 0.92 (95% CI 0.85–0.99). A real reduction, and a small one. All-cause mortality showed no difference at all: 23.2 per cent against 23.3 per cent.

But the finding that matters most is what the screening actually found. Compared with no invitation, PSA screening increased detection of low-grade disease (Gleason 6 or below: 2.2% against 1.6%) and localised disease. It did not increase detection of intermediate-grade, high-grade, locally advanced or metastatic disease.

That is the overdiagnosis argument stated as cleanly as any trial has ever stated it. A single PSA invitation found substantially more of the cancer that was never going to kill anyone, and no more of the cancer that does.

The other half of the trade-off: what treatment does

Any discussion of screening that stops at mortality is only half a discussion. The ProtecT trial randomised 1,643 men with localised prostate cancer to active monitoring, surgery, or radiotherapy with hormones, and followed their own reported function for twelve years. It is the best evidence we have on what treatment actually costs.

The figures below are widely quoted, and almost always quoted wrongly — so read the comparison columns, not just the first one.

Outcome	Prostatectomy	Radiotherapy	Active monitoring
Urinary leakage requiring pads (years 7–12)	18–24%	3–8%	9–11%
Erections sufficient for intercourse (year 7)	18%	27%	30%
Faecal leakage (year 12)	6%	12%	6%
Nocturia, twice or more nightly (year 12)	34%	48%	47%

These are serious, common, and largely permanent harms. Roughly one man in five needs pads after surgery. Four in five do not have erections adequate for intercourse at seven years.

But look at the monitoring column. At seven years, 30 per cent of men who had no radical treatment at all reported erections sufficient for intercourse — against 18 per cent after surgery. The difference attributable to the operation is around twelve percentage points, not eighty. And by year twelve, the paper reports, all three groups had converged to low levels of potency.

I want to be careful about what that does and does not mean. It does not soften the harm: a man made impotent at 60 by an operation he did not need has lost years he cannot get back, and telling him the untreated group caught up by 72 is no comfort at all. What it does mean is that “four in five men are left impotent by surgery” overstates the causal contribution of the surgery, because a substantial part of that decline was going to happen anyway.

I am flagging this because I have seen the 82 per cent figure used, correctly quoted and stripped of its comparison group, to argue against screening — which is precisely the move I criticise when the same trick is used to sell a supplement. Same trial, same table, opposite conclusions, depending on which column you print. If you only ever check the numbers that support the position you already hold, you are not checking.

One thing the trial found that deserves more attention than it gets: over years seven to twelve, generic quality-of-life scores were similar across all three groups, as were anxiety and depression measures. The functional harms are real and they show up in the function data. They did not show up as a difference in how men rated their lives overall.

The all-cause mortality problem

Here is the strongest argument against screening, and it deserves stating properly. No trial has convincingly shown that PSA screening reduces deaths from all causes. CAP found 23.2 per cent against 23.3 per cent. The prostate-cancer-specific reduction is real; the overall reduction has never been demonstrated.

Some of that is competing mortality — a man whose prostate cancer would have killed him at 84 may die of something else at 84 instead. Some of it may be treatment harm offsetting benefit.

But the counter is important and it is usually left out. No cancer screening programme has ever demonstrated an all-cause mortality benefit — not breast, not bowel, not cervical. Detecting an absolute effect of around 0.2 per cent against a background where roughly a quarter of participants die of something else would need a trial far larger and longer than any that exists. Setting all-cause mortality as the bar does not single out prostate screening as uniquely unproven. It abolishes screening as a category.

That does not make the point wrong. It is a genuine and unresolved weakness in the case for screening, and anyone considering a PSA test should know it. It just is not the knockout argument it is often presented as.

The case for screening, at its strongest

I have to give this properly, because the organisations making it are serious and the argument has moved.

The strongest version is that the harms measured in these trials are the harms of an old diagnostic pathway. Both ERSPC and CAP largely used the sequence of raised PSA leading straight to systematic biopsy. Modern practice puts MRI first, and only biopsies where the scan suggests it — which substantially reduces both the number of unnecessary biopsies and the detection of insignificant disease, and cuts biopsy-related infection markedly. If overdiagnosis is a function of the pathway rather than the test, then a trial that measured the old pathway is measuring harms that no longer fully apply.

The advocacy organisations also argue that the NHS already pays for a large volume of opportunistic PSA testing — men who ask, and get tested, with no structured pathway around them. Their point is that this is the worst of both worlds: the costs and harms of screening without the safeguards, distributed by who happens to ask rather than by who is at risk.

That is a good argument and I think it is largely right. It is also worth naming that it comes from organisations whose purpose is to advance prostate cancer detection. That is not a criticism — charity-authored evidence is no more automatically wrong than industry-authored evidence is automatically right — but the same standard applies in both directions. Disclose the interest, then assess the argument on its merits. On its merits, the MRI-first point is strong and the UK NSC will have to keep answering it.

The committee has, notably, committed to a “live” model with iterative updates as evidence arrives, rather than fixed multi-year review cycles. Whatever you think of the current decision, that is the right structural change, and it is the same principle I try to apply here: publish the position, publish the corrections, revisit when the evidence moves.

What this means if you are 50-something and wondering

I cannot tell you whether to have a PSA test, and I would be suspicious of anyone who tells you they can. What I can do is set out what the decision actually consists of.

• Your risk is not average. Family history, and in particular a father or brother diagnosed young, changes the arithmetic. So does being Black — prostate cancer is substantially more common in Black men, and the current targeted programme does not address that, which is a live and legitimate criticism of it.
• A raised PSA is the beginning of a pathway, not an answer. Before you test, work out what you would do with each possible result, including the ambiguous middle. Most of the distress in this area comes from people who tested without having thought that through.
• Ask what the pathway looks like where you live. MRI-first is the thing that changes the harm profile. It is a reasonable question to put to your GP before you agree to the test rather than after.
• Active surveillance is a real option, not a fudge. A diagnosis of low-grade disease does not automatically mean treatment. Knowing this in advance changes how frightening a diagnosis is.
• In England, if you have a BRCA2 variant and a relevant family history, the position has changed. Speak to your GP or genetics service. Elsewhere in the UK, it has not changed yet.

Where I sit, and where I do not

I do not screen for prostate cancer and I am not qualified to interpret a PSA result. A PSA can be included in a blood panel; what it cannot be is something I adjudicate. A raised result goes to your GP, promptly, and the whole subsequent pathway — MRI, biopsy, staging, treatment decisions — belongs to urology and oncology. There is no functional medicine reading of PSA, and anyone offering you one is selling something.

What I can legitimately do sits alongside that and never instead of it: the metabolic and cardiovascular picture, which is what most men in this age group will actually die of; strength and conditioning, which matters enormously for anyone who does end up on androgen deprivation therapy; and the unglamorous business of sleep, alcohol and body composition.

I came into this expecting the evidence to support my instinct. It half did. The overdiagnosis problem is real and CAP demonstrates it about as clearly as a trial can. But the mortality benefit is also real, it is larger than I expected, and it has been getting larger as follow-up lengthens. Anyone telling you PSA screening does nothing is overstating it in exactly the direction I would have found comfortable.

What would change my mind further

The TRANSFORM trial is the big one — a large UK randomised trial testing MRI-first and other strategies directly, rather than inferring their effect from trials that used a different pathway. If it shows that an MRI-led approach retains the mortality benefit while substantially reducing overdiagnosis, the argument for a broader programme becomes much harder to resist, and I will say so here.

Equally, evidence on Black men specifically, and on risk-stratified rather than age-based invitation, could change the shape of the recommendation without changing its logic.