From here, since my self-created recording is so bad.
Last week, Georgia Gale Grant argued that, rather than praying for the recovery of Fabrice Muamba, it would be better, for humanists, sceptics and aethists, if aged between ‘14 and 35’ to ‘ get regular ECGs to make sure you’re not at any known risk of sudden death’ – and she recommended that you get your friends or children or other relations to go too.
On the first point; I agree. One study in the American Heart Journal from 2006 suggested that prayer didn’t help heart bypass patients recover better.
But what’s the evidence for her other suggestion – screening a chunk of the population with an ECG – that is, an electrocardiogram – colloquially known as a heart tracing, and obtained by attaching electrodes to the limbs and chest wall.
What she’s suggesting is a screening test. Unfortunately, screening is messy, may be counter-intuitive, and is sometimes hazardous. Screening tests are not for people with symptoms. Someone who has chest pain, for example, is not given an ECG as a screening test, but a diagnostic test. They are very different, but often confused, not just by citizens, but politicians and even healthcare professionals. But the difference is crucial. Screening is looking for disease before the person knows they have it.
There’s a great book by Angela Raffle and Muir Grey, called ‘Screening – Evidence and Practice.’ She is a public health doctor, he was the Director of NHS Screening. It starts “All screening does harm. Some do good as well and, of course, some do more good than harm at a reasonable cost’.
So what are the harms of screening? It seems so sensible, so logical – pick up disease early, fix problems before you get complications – doesn’t it? The truth is that screening is rarely so straightforward. Chapter and verse on what makes a great screening test was published by the World Health Organisation by Wilson and Junger in 1968.
I’ll put a link on my website to it – for this document is full of clarity. For example, they decreed that screening should look for important conditions for illnesses where the natural progression of the disease was understood.
Early treatment should have extra benefits compared to later treatment, and the test should be acceptable with the risks, physical and psychological, being less than the benefits.
Some screening tests come with high benefits and small risks. For example, all newborn babies in the UK are screened for the rare genetic condition PKU, phenyl ketone uria. If the child is diagnosed soon after birth, he or she can avoid foods containing phenylalanine and avoid the brain damage that would otherwise occur. It’s a rare condition – 1 in 10 000 births – but because there is a good preventative treatment – avoiding certain foods – that can be recommended – it’s generally thought of as a pretty good screening test. The false positive rate is about 5%, meaning that of every 20 babies with a positive result, one is wrong. False negatives seem very uncommon.
And this is why all children born in the UK have a heelprick test at about 10 days old, which includes a PKU screen. The harms are in false diagnosis – but this is minimised by re-testing, and, of course, the heels of all new born babies have to be pricked with a needle to pick up the small amount with problems.
But this is a validated, well researched and audited screening test. Not every screening test on offer comes with the same balance of risks. You may have noticed the panoply of high street clinics offering health checks, or adverts on TV offering CT scans as ‘peace of mind’, or you may have had – like me – letters through the door encouraging me to attend a church hall near me for a heart tracing, cholesterol testing, ultrasound of the neck or abdominal aorta for a few hundred quid. (I think I have an article in this week’s BMJ about this – I’ll link it
online.) Again, it sounds sensible, responsible, logical.
Catch it early! Why would you want to be ignorant of your ‘risks’? And indeed, in the USA, where they seem to love health screenings, regular medicals drive a huge amount of testing both on Medicaid and with doctors in the paid-for arena, working to commission.
But how much of this does any good? We have to remember those Wilson and Junger criteria. There is no point picking up a problem that you can’t do something useful about. Let’s take ultrasound of the carotid arteries, in the neck, for people who are well, as offered by many private screening clinics. The idea is that you can find vascular disease and treat it before it causes a stroke.
The problem is that in order to prevent a stroke, the surgery offered is carotid endarterectomy – which corrects the narrowing. However, because the death rate from the operation is, relatively speaking high – about 3% – the potential small drop in stroke rates achievable because of the operation doesn’t provide a good risk benefit ratio.
So that’s why both the UK National Screening Committee and the US Preventative Services Task Force say it does more harm than good – and it’s not a recommended screening test.
The other problem that screening produces is to identify conditions that would have never caused a problem during that person’s lifetime. One clear example is breast screening. This is an xray of the breasts taken every 3 years, in the UK, for women from age 47. The Cochrane review – that is, an independent review of all the information on the subject – concluded that 2000 women are needed to be screened for 10 years to prevent one death from breast cancer. But to get this reduction, 10 other women have to be treated with surgery radiotherapy or chemotherapy for a breast cancer that would never have affected them in their lifetime. This is ‘overdiagnosis’ – cancers diagnosed and treated but which would never have maimed or killed.
In medicine, a clear distinction between normal and abnormal doesn’t always exist. One of my deep concerns is that the dilemmas and uncertainties of screening tests aren’t adequately communicated to the people taking them up. There’s an excellent NHS resource for men considering PSA screening, for example, which is used to detect prostate cancer – but which explains that there is no evidence that PSA screening reduces overall death rates, and that treatment can result in impotence and incontinence. But we don’t get the same aid for the other screening tests the NHS offers, and I am afraid this results in a rosy depiction of screening which is, sadly, illusory.
So what about cardiac ECG screening?; well. There has actually been a lot of work on it. It’s clear that ECG screening doesn’t benefit adults with no symptoms – the test doesn’t usefully or reliably pick up abnormalities.
For young people, contrary to the actions Georgia suggested, the National Screening Committee in the UK doesn’t recommend screening.
Why? There are several causes of cardiac death in this group, including abnormalties in the conduction system of the heart, for example, long QT syndrome, and heart muscle problems, such as the cadiomyopathies. A good screening test would pick up all these abnormalities, with no false positives; and it would also mean that effective treatment was available. The result would also come with meaning – there is no point picking up a ‘problem’ if it only rarely would lead to future problems. You have to know that an ECG abnormality really is abnormal. And a good test would not let abnormal hearts slip through the screening. And what do we know?
Less than 1 in 100 000 young adults die of sudden cardiac death per year. Per 100 000 screened with an ECG, you would expect to find 300 with a cardiac problem, of whom
30 would be ‘at risk’, and of whom 0.4 would die per year.
The screening would also generate at least 2,500 abnormal ECGs – possibly double that – and 150 false negatives. For children with certain cardiac abnormalities, medication can help to reduce risk – but there is no such thing as a guaranteed peace of mind. Not all complications could be prevented. The harms would include telling a large proportion of people that they had an abnormality – but the vast majority would never come to harm from it. And about half of ‘at risk’ children would not be picked up – in other words, false reassurance. Many other children would be medicated – or told they had a potentially fatal condition – when the chances were actually very favorable for them never to have an cardiac arrhythmia. Harms can be overt but also subtle – this is the kind of thing that GPs see. And of course, there’s a cost-effectiveness argument too – is this the best way that money can be spent to improve children’s health?
(Addition: this holds true for people without a family history of sudden cardiac death. If there has been, doctors will usually look pretty hard for evidence of a genetic problem in close relatives, and will do not just ECG but often echo scans and other tests. There’s a link in the comments below.)
Like I said, screening is complicated. It comes with numerous caveats and is never a benign act. It’s worth remembering too, the ‘popularity paradox’ – the worse a screening test it, the more false positives you generate, and the more people believe that their lives were saved – when actually, by being diagnosed wrongly, they were potentially harmed, not helped. Screening needs science and attention to the evidence – even when it tells you things you’d rather not know.