Prostate cancer screening could save hundreds of men. So why is there a debate about it? Why has the UK National Screening Committee, chaired by highly respected Professor Sir Mike Richards, a man himself affected by prostate cancer, failed to recommend it nationwide? To answer, we must understand a little more about tests.
What does a test do?
We’d like to think a test tells us a truth about the world that we didn’t know before. I don’t know whether it’s raining or not, so I’ll stick my hand out the window, and if it comes back wet, it’s raining. Simple, right? Well, no. What if there were a sprinkler on outside?
It’s a silly example, but it nonetheless shows that there isn’t such a thing as a perfect test. Tests do not tell us the truth; they give us evidence that hints at a truth. If I stick my hand out the window and it gets wet, that’s a positive test result, which is evidence for the truth I’m interested in, the truth of rain, but it doesn’t guarantee it.
Whether a positive test result correctly identifies the truth we’re interested in, depends on two things, the quality of the test, and the things tested, the sample. In the case of the hand-rain-test, the sample would be the environment in which my hand, the window, and the possible rain exist, so let’s explore this a little more.
Why the sample matters
If I live in a single storey house, alone, and (crucially) without sprinklers, the hand-rain-test is quite likely to be accurate. There are very few things that could make my hand wet other than rain, so if my hand gets wet, I can safely assume it’s raining. I can trust the test result.
However, if I decide to start a cottage garden, and I place a sprinkler outside my window, then the waters are, well, sprinkled. In this situation, the accuracy of the test depends on a few factors. If the season is particularly rainy and I never need to turn on my sprinkler, then a wet hand will mean rain.
On the other hand, in the dry season, where there’s little rain, I’ll turn the sprinkler on to compensate, and a wet hand may mean sprinkler, not rain. In this instance, the test result would be what’s known as a ‘false positive’. Again, this seems like a silly example, but it is inching towards an important concept in screening: ‘prevalence’.
Prevalence is the amount that the truth we are interested in actually happens. In the rainy season, the prevalence of rain is high; in the dry season, the prevalence of rain is low. Where prevalence is low, a test is more likely to give a false positive, all else being equal.
The quality of the test
Clearly, there are problems with the hand-rain-test, as evidenced by its potential for false positives, particularly in the dry season. If, all else being equal, a test has a propensity for giving false positives, then it is described as lacking ‘specificity’. A positive result is not very specific, as it can mean several different things. But there’s another quality to a test that’s vitally important, its sensitivity.
Sensitivity describes how likely a test is to detect the truth when the truth is sitting in front of it. For the hand-rain-test, my hand might remain dry during the rain if I don’t keep it out the window long enough, thereby missing scanty drops. In this instance, my dry hand, a negative result, would be what is known as a ‘false negative’.
If, all else being equal, a test tends to yield false negatives, it is said to lack sensitivity. I could increase the sensitivity of the hand-rain-test by keeping my hand out the window for longer. Indeed, I could increase the test’s sensitivity to 100% (never missing a single rainy day) by simply declaring rain every day, without fail. However, I hope you can see that if I were to do this, any gains in sensitivity would mean losses in specificity; I would be declaring a lot of rainy days when no rain was falling, a lot of false positives. Oh dear. Things are getting a little complicated.
Thankfully, there are ways of increasing both sensitivity and specificity, ways of simply making a better test. In the case of the hand-rain-test, the solution would be to look out the window, too.
What does a medical test do?
Like any other test, a medical test is simply a way of detecting something. It’s like having an additional sense. Imagine if, in addition to sight, hearing, touch, taste and smell, you could also sense whether someone was anaemic just by looking at them. Seems simplistic, but there are dogs that can literally smell when someone is about to have a seizure, so perhaps not so far from the truth.
Indeed, during a physical examination, your GP uses their senses to detect medical conditions, including anaemia, by looking for pallor in the skin or the mucous membranes around the eyes. But, just like the hand-rain-test, medical tests do not tell us the truth; they simply give us evidence. Whether or not a result happens to coincide with the truth depends, as with all tests, on the quality of the test (its sensitivity and specificity) and the prevalence of the truth we’re interested in.
What is screening?
Screening is the process of mass testing of a chosen, asymptomatic population for a particular condition. When we screen, there are two things we control – the test we use, and the population we test. Certain screening programmes, those that are carried out shortly after birth, screen literally everyone. Conditions screened in this way include Cystic Fibrosis.
Most other screening programmes choose a specific population to test. For example, cervical cancer screening is offered on the NHS in England to women aged 25-64 years, whereas screening for bowel cancer goes to people aged 50 to 74, and screening for abdominal aortic aneurysm is for men aged 65 and over.
Case Study – HIV
After finishing my medical degree at Cambridge, I trained to be a GP at St Thomas’. This hospital on the banks of the River Thames is linked to King’s College Hospital in Camberwell, and I worked in the Emergency Departments of both. One of the things you might notice in either of those departments is the signs on the walls telling you about HIV testing.
HIV testing is an ‘opt out’ procedure at St Thomas’ and King’s, meaning that, if you are already having blood tests, you have to tell the staff not to do it, otherwise it’s done by default. As such, all the people who attend the Emergency Department in these hospitals and require blood tests are screened for HIV.
This is not standard procedure throughout the NHS, only in certain places. But why is this? HIV seems like the ideal thing to screen for nationwide. The HIV blood test is very good, with high sensitivity and specificity. HIV is largely asymptomatic until it starts causing real problems, but effective treatments now exist to prevent it from getting to that stage, which means finding it early is very important. Nationwide screening doesn’t exist because HIV is, mercifully, relatively uncommon in the general population.
Unfortunately, this is not the case in South London, where there is much more of it about. Using a word we learned earlier, in the UK in general, the prevalence of HIV is low, but in South London, its prevalence is high.
So, what does this mean for screening? If we think back to our hand-rain-test, we saw that the test gave more false positives when rain was infrequent, when the prevalence of rain was low. The same is true of the HIV blood test. Even though this test is good (way better than the hand-rain-test), the relatively low prevalence of HIV in the UK general population means that if you were to screen someone in that population, and they got a positive result, the test would be more likely to be a false positive than a true positive, even though the test is really good.
You don’t have to take my word for it; this has been done. Pregnant women in Plymouth, a low HIV prevalence area, were screened for HIV: 11,000 were tested, 25 were positive, but only 7 of those had HIV. This meant that, if you had a positive result, there was only a 28% chance you had HIV, with a false positive rate of 72%.
However, if you choose your population more carefully, by picking one with a higher prevalence, the false positive rate drops. In the South London screening programme, the false positive rate was less than 20%, probably significantly less (exact data are hard to find).
What does this mean for prostate screening?
The test currently under scrutiny for prostate screening is called the PSA blood test. It detects a molecule, prostate-specific antigen (aka PSA), released by the prostate gland under certain circumstances (including, but certainly not limited to, cancer). Unfortunately, the PSA test isn’t great. Because it can be positive in conditions other than cancer, when used in a population where prostate cancer prevalence isn’t high, it can give a lot of false positives. This is less of an issue if a population with a higher prevalence is tested, for the reasons we’ve explored earlier.
But false positives aren’t the only problem. If screening in the general population were to take place, it’s thought that many of the additional cancers correctly diagnosed and therefore treated would never have needed treatment if not detected. They would be the kind of slow-growing cancer a man might die with, rather than of, never having realised it was there. This phenomenon is known as ‘overdiagnosis’.
Therefore, the UK National Screening Committee (UKNSC) has decided to go ahead and screen for prostate cancer using the PSA blood test, but, in order to reduce the rate of false positive results, it will only do so in a very specific population of men, where prevalence is especially high. They have chosen “men aged 45 to 61 who have a pathogenic (able to cause disease) BRCA2 variant with a family history of breast, ovarian, pancreatic, or prostate cancer”. It is also thought that overdiagnosis will be less of a problem for men in this group, as they are more likely to get more aggressive forms of prostate cancer.
Isn’t this just a cost thing?
Well, no. Cost plays a role, but it’s not the only thing. False positive PSA tests necessitate other tests like MRIs and biopsies to get to the true negative result, while overdiagnosis also does all this, but also leads to unnecessary treatment. None of this is cheap, but more importantly, it can be damaging.
With cases of overdiagnosis, a man might be exposed to the side effects of chemotherapy, radiotherapy, or even surgery to treat a cancer that would never have needed treatment. Even with false positives, harm can be done if a man requires a biopsy (in which a series of needles are placed into the prostate to remove samples), not to mention the psychological burden of having a positive result, even if that turns out to be false later.
The UKNSC summarises it thus, ‘The committee concluded that screening is more likely to cause more harm than good in the whole population and in men with a family member who has had breast, ovarian or prostate cancer but who do not have a BRCA2 variant.
“For black men, there is ongoing uncertainty as to whether screening would cause more good than harm. The main harms of prostate cancer screening include incontinence and erectile dysfunction in men who do not need treatment.”
But where’s the data?
The quotes above are taken from the official UKNSC summary statements on the gov.uk website. But on which data are these statements based? I’ve looked over and read parts of the full report. It is long and dense, but summarised, somewhat helpfully, in an infographic produced by Cancer Research UK, the link to which can also be found on the page above.
This simplification portrays the outcomes of one of the proposed screening scenarios, in which men aged 58 and 60 are offered a PSA test each year nationwide. It estimates that, for every 1,000 men tested, 100 would have a positive PSA test, all of whom then get an MRI. The 34 men who have a positive MRI get a biopsy, and of these, 28 are diagnosed with prostate cancer, with 2 lives being saved as a result. What happens to the other 26 prostate cancer cases is difficult to work out, even from the simple infographic, because the numbers don’t add up. However, it does say that 12 would be overtreated, the worst outcome.
An open door and the future
There is a benefit to screening that I cannot find discussed anywhere in the UKNSC’s statements. Certainly not in the economic analysis. It is the benefit one gets from receiving a negative screening result. The relief you might feel, especially if prostate cancer is at the forefront of your mind. Much of what I do every day as a GP is reassuring people, and it is a valuable thing. The NHS recognises this and has left the door open to PSA screening in concerned, asymptomatic men aged 50 and over, who ask for it.
Here at Wessex Private GP, our door is also always open for discussion, whatever your age. But we, unlike the NHS, also offer the Stockholm 3 test, which can be used in men with a PSA greater than 1.5 to provide a more accurate picture than PSA alone and to inform decisions about further investigation.
The UK National Institute for Health and Care Excellence (NICE) recognises that ‘The evidence suggests that Stockholm3 is more effective at predicting risk of prostate cancer than PSA testing alone for people aged 45 to 74′. This is reflected in the fact that Stockholm 3 is mentioned as an option in both European and American guidance on prostate cancer diagnosis.
A final thought…
All this talk on screening neglects the group of men for whom PSA testing is always a good option, those with symptoms. So, if you’re a man or know a man who has noticed lower back or bone pain, problems urinating, lethargy, erectile dysfunction, blood in the urine, or weight loss, please seek a discussion with a doctor.
That was a long one. Congratulations on getting to the end! I hope you found it useful.
About the author
Dr Jonathan Tham MA MB BChir MRCGP
General Practitioner and Sleep Specialist
GMC 7488747
Dr Jonathan Tham is an experienced GP and sleep specialist. He runs the Sleep Clinic at Wessex Private GP, diagnosing sleep problems and treating chronic insomnia with Cognitive Behavioural Therapy for Insomnia (CBTi). He provides one-to-one clinical consultations and practical self-help guidance.
Read Dr Tham’s full biography →
