SOCRATES: Hello Geoffrey, did you get some rest?
ROSE: I’m fine, but I sure would like to pick up the pace and quit getting bogged down into ridiculous minutiae.
SOCRATES: I have a feeling this will happen. You are now going to share with me your perspective on what you call the “high-risk” strategy of prevention. If I understand correctly, this prevention essentially aims at “lopping off the tail” of the distribution curve, sort of like this:
ROSE: That’s essentially it. It consists of identifying and focusing the preventive efforts on the patients at high risk of developing disease or those with early stage or sub-clinical disease. This is the common strategy that doctors and patients are most accustomed to and comfortable with.
It requires putting in place some screening test or tests, agreeing on a cut-off point to establish what ‘high-risk’ means, and have a plan to intervene on the patients identified in this way. This was described in detail in a wonderful 1968 WHO report by Wilson and Jungner.
SOCRATES: What are the important features of that report?
ROSE: Wilson and Jungner provided a very clear but exhaustive examination of all the relevant parameters that impact on such a screening strategy, particularly as it applies to the kinds of chronic diseases that affect developed countries. That report is truly a pleasure to read and remains relevant today.
Incidentally, we can see in the report that Pickering’s insight about unimodal diseases had already influenced our thinking. When Wilson and Jungner discussed the problem of “borderline cases,” they make reference to both possibilities as can be seen in their Figure 2:
SOCRATES: They make the point that in the case of uni-modal diseases (which you claim most diseases are, as we discussed), the “borderline” cases can be larger in number than the “diseased.”
ROSE: That’s exactly right!
SOCRATES: And for you, that’s a problem.
ROSE: Naturally, especially since by definition, mild or borderline cases tend to have a good prognosis, although a small number of individuals will suffer clinical consequences. This point emerged forcibly in the Medical Research Council trial of treatment for mild hypertension. The relative risk of stroke increased steeply with increasing blood pressures, and treatment effectively reduced it, but for any one individual the absolute risk was low, and overall it needed 850-person-years of treatment in order to prevent one stroke (1).
So depending on how you pick your dividing line for disease, you can treat needlessly a lot of people. But if you move the line higher up, you will fail to prevent bad outcomes in reasonably large numbers.
SOCRATES: That’s frustrating.
ROSE: Especially if you recognize that under this screening strategy, prevention becomes “medicalized.”
SOCRATES: What do you mean?
ROSE: Consider, for example, that a man who went to see his doctor because he had a pain in his neck walked away for that encounter bearing a label of ‘hypertensive patient’, which he must now wear for the rest of his life. Having hitherto perceived himself as healthy, he now has to see himself as someone needing to take pills and to see the doctor regularly. He was, he thought, normal; now he is a patient. This may be unavoidable and justified by the benefits, but it is a major cost (2).
SOCRATES: And your strategy avoids this problem?
ROSE: I believe it does.
SOCRATES: Before we go there, let’s finish our exploration of the “high-risk” strategy. I will place here the summary recommendations by Wilson and Jungner:
Do you agree with these, as far as screening strategies go?
ROSE: Yes I do. In my book, I elaborate on points (2), (3), and (9). First of all with point (3), I think that if one is going to adopt a screening strategy to identify high risk patients, it is imperative that resources be in place to provide advice and long term care. Experience shows repeatedly that effective interventions, particularly for the chronic diseases of developed countries, require education, involvement of the practitioner, and sustained follow-up.
It follows that a policy of mass screening for risk identification presupposes a medical care system which is able to provide continuity of long-term personal care for everyone. This is a major obstacle to effective preventive care in countries, such as the USA, which lack a general practitioner system covering the whole population. In Britain, and in other countries fortunate enough to possess such a system, at least the potential is there for long-term preventive care, but its realization requires substantial additional investment in staff, training, and organization (3).
SOCRATES: You’ll be interested to hear that the USA is now actively trying to emulate the British system and establish so-called “medical homes” for each person, where the primary care physician would be in charge of orchestrating preventive services.
SOCRATES: But don’t be too optimistic…neither the NHS nor the American system are in a position to provide the resources that you think are necessary…
What about points (2) and (9)?
ROSE: For point (2), I simply stress that we go after reversible risk factors, not risk factors about which we cannot do anything. Furthermore, the effectiveness of an intervention must be judged in terms of its absolute, not relative benefit. For example, treating an elevated cholesterol level in a patient with normal blood pressure and no other risk factors offers a trivial benefit to the individual or the population. Policy decisions must therefore be founded on absolute, not relative, risk estimates, and they should take account of those other factors which modify the risk of a particular exposure (3).
SOCRATES: I think this point is now well accepted in academic medical circles. What about point (9)
ROSE: As far as cost-effectiveness is concerned, I remark that selective screening is always more cost-effective than mass screening, and simple and readily available information may often indicate that risk is more likely to be found in one group than another. This can make it profitable to plan a two-staged process in which one looks for high-risk individuals only within a high-risk sector of the population (4).
SOCRATES: That makes perfect sense to me. In general I don’t find you to be all that set against the high-risk, screening approach.
ROSE: Not at all. It has some important benefits but also some shortcomings.
SOCRATES: We have exposed some potential short-comings. What are the benefits?
ROSE: The main ones are that the intervention is appropriate to the individual (5) and the approach avoids interference with those who are not at special risk (6).
SOCRATES: How’s that?
ROSE: What do you mean how is that? The patients who are not at special risk are not intervened upon.
SOCRATES: Except for the borderline cases which, depending on how you identify them, could be a large number of people at no special risk. Furthermore, you mention yourself that the ability to predict individual course is very poor, particularly for the chronic diseases such as coronary disease or hypertension. Wilson and Jungner identify in (7) that the “natural history of the condition, including development from latent to declared disease, should be adequately understood.” It seems to me that by this criterion, much of cardiovascular disease would be disqualified from becoming the target of a screening strategy.
ROSE: That’s crazy! We have had great success with our screening procedures for hypertension, serum cholesterol, smoking, etc. Coronary disease mortality has plummeted in the last 30 years.
SOCRATES: I don’t dispute that but there is a cost being paid which Wilson and Jungner were aware of in 1968 and which nowadays seems completely overlooked. They say that “In enthusiastically attacking disease at an early stage, the Hippocratic principle …primum no nocere, should not be neglected.” .
ROSE: Socrates, Hippocrates was your contemporary and I suspect that you know full well he did not enunciate this principle.
SOCRATES: Touché! But my point remains that when a screening policy is established, some committee of experts somewhere decides what risk and benefit “justify” the given approach, and I suspect your man walking into the doctor’s office with a stiff neck is not involved in that decision.
ROSE: Well what do you propose, Socrates?
SOCRATES: You could at least ask your man if he wishes to have his blood pressure taken, his cholesterol measured, or his seat belt behavior scrutinized. You could spend the time educating him about his a priori individual risk, and explain to him how that risk would be modified by the screening test, what harm might ensue from a borderline finding, and what the imprecision in predicting outcome really is.
In other words, you could put in place resources similar to those you propose for post-testing counseling before actually applying the test. That would really be an approach adapted to the individual since it would take into account the individual’s personal risk tolerance.
ROSE: But that’s completely impractical! And besides, most patients would defer the decision to the doctor. You know the chime, “Whatever you say, doc!”
SOCRATES: In that case, from an ethical standpoint, you’re off the hook, at least if you accept the principles of non-malfeasance and autonomy. Plus you assume that patients will defer to the doctor, but that may not be necessarily the case. The public’s attitude may change over time as people learn to make decisions for themselves. They may also seek the counsel of others in the agora, and channels that distil collective experience could spontaneously form. Have you heard of something called the internet? It is a remarkable vehicle for dissemination of information.
ROSE: I don’t know what the internet is, but if the scenario you outline materializes, it is likely that very few people will accept the screening, since individual risk is always very low. This would completely reverse all the gains we have made over the last 30 years!
SOCRATES: Who knows, Geoffrey? People may actually decide in large enough numbers that it is worthwhile to get screened. Or even better, the ones at risk will self-select on a hunch or on some occult knowledge that something’s wrong. This would greatly improve the yield of testing!
ROSE: That’s crazy, Socrates, and you know it! But let’s agree that a preventive strategy based on screening for high risk individuals is less than desirable. What don’t we now explore my “population” strategy?
SOCRATES: Excellent idea, Geoffrey, but let’s leave this for another time.
- Rose’s Strategy of Prevention, 2nd. ed. p. 83
- Ibid., p. 81
- Ibid., p. 71
- Ibid., p. 77
- Ibid., p. 78
UPDATE: This series is now complete and consists of a total of 9 posts. Beginning with Post #3 I am aided in this exploration by a legendary cross-examiner who interviews Geoffrey Rose’s directly and in depth in an informal Q&A format. Quick links to the part are here: 1, 2, 3, 4, 5, 6, 7, 8, 9