Healthy Skepticism Library item: 20468

Warning: This library includes all items relevant to health product marketing that we are aware of regardless of quality. Often we do not agree with all or part of the contents.

Publication type: Journal Article

Ed.
And now all this
The Lancet 1997 Jan 4; 349:(9044):1
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736%2897%2921001-9/fulltext

Abstract:

A patient consulting a doctor is looking for knowledge and predictions: what is the diagnosis and, with or without treatment, will the disease go away? Does either party see the near-impossibility of this situation? Hippocrates had it right when he said that “it is unwise to profess either death or recovery”. Over 2000 years later, the philosopher Isaiah Berlin, writing about political rather than clinical judgment, looks at the dilemma from the other side: “A true science must be able not merely to rearrange the past but to predict the future”.
We take the view (Dec 7) that the public should be told about uncertainty when data with public-health implications are preliminary or inconclusive. Let us ask here the tough perennial question about how doctors advise the individual when uncertainty is in the air. Our end-of-year supplementary review showed that medical knowledge does advance. How does knowing what proportion of patients improve on a particular treatment help the doctor advising an individual?
Epidemiological research on cigarette smoking and health risks shows that smoking is dangerous but cannot pinpoint the subgroup who will succumb to their habit. Is that why so many individuals are willing to take the risk? Meta-analyses often show small but statistically significant benefits for a whole population of patients, but cannot tell the doctor which patient is suited for that particular intervention. A 10% average treatment benefit is generous for a meta-analysis in oncology, for example (such statistical techniques would not be needed if all trials showed a massive benefit), but that means ten patients will be exposed to the potential side-effects of an intervention so that one might survive longer than would otherwise have been the case. An important psychological hazard here for the individual, of course, is that all ten will be exposed to the hope of being the one.
While clinical and epidemiological research marches on, individuals are failing to benefit from proven remedies. Aspirin greatly reduces the risk of a second heart attack, for example, yet this cheap intervention is prescribed in only about half the eligible patients. Condoms and needle-exchanges, which reduce the risk of HIV transmission, are not “politically favourable” for President Clinton in his impossible strategy to cure AIDS (see News, p 39). Paclitaxel is an expensive drug and caught the public’s eye because its source was an old-forest American tree. Yet it has cured no-one of cancer and can hardly “win the war” against cancer. The latter was President Nixon’s campaign, yet US presidents and other advocates of sloganeering medicine (there seems to be a “Day” for every disease now) have yet to learn that the doctor/patient relationship is more complex than a catchphrase. Too many slogans and overuse of “breakthrough” in (but not exclusively by) the media can only encourage individual patients to expect too much, thus heightening the cautious clinician’s dilemma.
And so what of 1997? Epidemiology and clinical trials are generally uninformative at the level of the individual. Therefore researchers must use new tools that focus the results of research onto the individual. An important first step, already started in some trials, is the compulsory inclusion of quality-of-life outcome measures. As long as meta-analyses are needed, then similar measures should be compulsory there too. To that we would add study of the validity of database analyses to ensure that all useful results really have been found. Bayesian thinking, which is intimately linked to the individual, is long overdue more formal use. Finally, has anyone ever compared the counselling habits of research clinicians compared with their clinically full-time peers? If the former have solved the problem of translating summary statistics from population and intervention studies into individual messages, they would do well to pass on the secret.