Theodore Dalrymple summarises the issues on screening for diseases. ISO accreditation is a screening test with poorer performance characteristics. It is also a treatment that offers no cure, just endless repetition.
Unproven prevention; no hope of a cure – since accreditation is promised to deal only with lack of confidence.
Think out how the comments on cancer screening apply by analogy to accreditation.
Start asking doctors why they don’t reject accreditation.
Everyone knows, or thinks he knows, that prevention is better than cure. It comes as something of a shock to discover that, in practice, this might actually not be the case; and that, as a result, we might have to abandon a principle that we took for granted, something that none of us finds easy or likes to do.
The logic of screening for breast cancer is clear. Breast cancer is a common fatal disease whose prognosis is better the sooner it is detected. If women at risk of contracting it are screened by mammography, many of them who would otherwise go on to develop incurable and fatal cancer will be saved by early treatment. If enough women are screened regularly, the number of deaths from breast cancer in the population will fall.
And indeed, the number of deaths from breast cancer in the population has fallen, more or less at the same time as a programme of breast cancer screening had been organised. But what happens after something else does not necessarily happen because of that something else; and a study from Oxford University published this week and reported in this newspaper yesterday casts doubt on the role of mammography in the decline in deaths from breast cancer in Britain.
For one thing, death rates have fallen fastest among those who are not routinely screened for breast cancer, namely the under-40s. For another, treatment has improved greatly over the past two or three decades, so that this might account for the fall in the death rates.
It is important to remember, too, that most studies of mammography compare death rates from breast cancer among the screened and the unscreened, but this is the wrong measure. What should be compared is the all-cause death rate of the screened and the unscreened, for two reasons. First, there is little point in saving people from one fatal disease if it is only to allow them to contract another such disease (unless death from the one is much to be preferred to death from the other); and second because the screening process itself might cause deaths from other causes.This in turn is because of the problem of false positives. Some women screened will be discovered to have lesions thought to require treatment that in fact are harmless and would never have killed them. But the treatment for such harmless lesions is not harmless, and in rare cases may even prove fatal. If enough people undergo unnecessary treatment, indeed, the harm that such treatment does might outweigh all the good that detecting genuine cases at an early stage does.
And this is precisely what a well-known professor of surgery recently claimed in the British Medical Journal. According to his calculations, unnecessary treatment as a consequence of mammography was responsible for up to as many as three times the number of deaths as the screening saved.
Of course not everyone agrees with his calculations, to put it mildly. But he was a reputable professor who had spent much of his professional life studying the question, and it is unlikely that his opinion was simply absurd, though his calculations may prove to be mistaken.
In fact, the whole question is fiendishly difficult to resolve. This is partly because researchers are trying always to hit a moving target. The incidence of the disease (the number of new cases per year) may change with time; likewise the nature of the disease, that is to say the degree of its aggressiveness and fatality. The efficacy of the treatment changes over time, but so does the accuracy of the screening procedure. And if we do not continue with the programme, how could it ever be refined to the point at which it definitely does save many lives?
With so many variables, however, it is difficult to assess the precise effect of a screening programme, whether for good or ill. Experts, all of them honest, reasonable and genuinely anxious to do good by finding the truth, may disagree on the correct answer.
This kind of uncertainty in medicine is by no means unusual, especially nowadays when procedures or treatments are recommended on the basis of population studies rather than their effect on individuals. What is true of screening for breast cancer is also true of screening for prostate cancer, another fiendishly difficult question to resolve – for more men die with cancer of the prostate than of it. And it is very unlikely that most patients who take treatment for high blood pressure are aware that the chances of it doing them any good are slender, though the good it might do them is very great (the avoidance of a heart attack or stroke). This is a very different situation from that in which an individual patient has, say, bacterial pneumonia and will almost certainly benefit from antibiotics.
Before patients are subjected to procedures, they are supposed to give informed consent: but how informed is informed? And what is a reasonable decision in the light of the information he or she has been given? I may be willing to take tablets every day and put up with their side effects in order to avoid a 1 per cent risk of a catastrophic event within five years, but you might not be willing. Neither of us can be shown to be right or wrong; we simply have different priorities.
The problems for ordinary doctors in conditions of uncertainty are legion. If experts on a thorny question, all of whom are competent and well-intentioned, disagree, what is the average doctor to recommend to his patient, who requires a yes or no answer, to take or not to take the medicine, or to agree to or not to agree to the procedure? On the one hand, on the other hand will not satisfy him or her.
Moreover, patients differ in the degree to which they want to be involved in any decision. Some want to know every last detail of the rationale and effect of the treatment or procedure proposed, others want the doctor to decide for them and to be told what to do. I remember going once into some detail with a patient about the pros and cons of treatment of his condition, and asked what he wanted.
But what would I have done? I couldn’t answer the question because I didn’t know, never having been in the patient’s shoes. Besides, even if I had been in his shoes from the purely medical point of view, I still wouldn’t have been him.
In the end I advised him to take the treatment, though to this day I am uncertain whether I was treating his condition or my anxiety. The prospect of a remote and statistically unlikely catastrophic event in his case as a result of not treating him was more vividly present in my mind than the far more likely prospect of nothing serious happening to him in the absence of such treatment. (In fact, he was most unlikely to take it precisely as prescribed. Most patients don’t, though on the other hand some are hyper-scrupulous or obsessive about doing so.)
Once, when I was very ill, I found the reassurance of the treating professor that he would find what was wrong with me extremely comforting. He never did find out what was wrong with me despite a plethora of tests, and I got better anyway. But the last thing I had wanted was to be involved in deciding what he should do.