We’re all guinea pigs, and it’s not so bad

The outrage provoked by Professor Anthony Mathur’s suggestion that patients should sometimes be obliged to enrol in clinical trials (discussed already on this blog by Steve Clarke: 11 December) continues to rage on.  Armies of shrill autonomists tell us that all our worst nightmares are real: that Mengele walks again.


But is it really so bad? Surely every patient is always and inevitably a subject of medical research. And the moon hasn’t turned to blood.Every patient is a part of a clinical trial that every doctor conducts throughout his professional life. Each doctor calls that trial ‘My experience’. Every doctor is constantly comparing patients, treatments and outcomes.  The trial breaches many of the conventional  rules of ethical clinical trials. Every time we go to see our doctor, he makes use of information held by him about other patients. Those patients have given no explicit consent to that information. But the doctor will be remembering how comparable to you those previous patients were, and how they responded to the therapy he is considering for you. Suppose that he gives us treatment X. Suppose that the outcome is bad. He will use that result in determining his approach to other patients. And he will do it whether we like it or not. It doesn’t matter whether we are formally part of a project designed to evaluate X: our data will affect that clinician’s future conduct. And quite right too.

The more experience a doctor has, the more patient information he will be using. And the more we will want him to be our doctor.

Of course nobody objects to the use of information in this way. Everyone is happy about being a member of a clinical research project in this sense. Now let’s see where this leads us.

Surely it follows that there can be no sensible objection to the use of our anonymised information in the form of clinical audit data. A doctor’s own recollection of and reflection on his practice is a type of auditing no different in ethical nature to the formal collation of audit statistics.

What’s the difference between anonymised audit data and any other kind of anonymised medical information? Consider three cases:

(a)    A doctor says to himself ‘I saw that condition last week in patient A, and it responded  to X: so I’m going to use it again for patient B.

(b)   A  doctor writes an anonymised case report in the BMJ on patient A, saying that X worked well, and commending its use in patients like B.

(c)    A doctor gathers together 10 years’ experience of cases like, and including, A and B, does some statistics on them, and publishes the results.

What is the difference between these cases? In (a), have the autonomy rights of patient A been violated? Well, his information has been used. There was no express consent to that use, but the patient could only prevent the doctor from using that experience by refusing to see the doctor in the first place, or by insisting that the doctor retired from practice at the end of the consultation.  But was there implied consent to its use? Perhaps: but it seems rather artificial – a typical lawyer’s pigeonholing – to say that there was. The real situation was that the patient was forced by his illness to sign a contract with the doctor: it was a term of that contract that the doctor would be able to use the patient’s information in the treatment of other patients.

Case (b) is arguably rather different. But only arguably. What are the differences of substance?   I suggest that there is none. All that has happened is that the private information that was already being used for the benefit of other patients is being disseminated rather more widely, and accordingly that more patients will benefit. The patient has already agreed to limited use of his information (as demonstrated in case (a)). Case (a) shows that after all clinical consultations the cat will creep out of the bag. In case (b) it simply runs a little further from the bag. Is that slightly longer excursion (which seems to do no real ham to the patient) a reason to deprive other patients of the potentially colossal benefit of the cat’s presence?

The difference between case (b) and case (c) is simply one of scale. No new ethical principle comes into play.

So I can see no good argument against the use of anonymised patient information in published case studies or in studies involving the manipulation of anonymised patient data from cohorts.

Can we conclude anything from this about the propriety of the sort of study proposed by Mathur?  Indeed we can.

In such studies patients typically get either the conventional, baseline treatment (BT) or the Revolutionary New Treatment (RNT). In enrolling in the trial they might, depending on the trial design, be randomly allocated into either the BT or the RNT group, or choose the RNT. The trial is only allowed by the Research Ethics Committee when there is good reason to suppose that the results of the RNT will be at least as good as the BT. Accordingly patients in randomised controlled trials tend to hope that they will be in the group getting the RNT. They’re in it primarily for their own good. The fact that the RNT is being trialled, though, means that the clinicians couldn’t be labelled negligent if they treated the patient with the BT. BT is Bolam-proof. And so the patient can’t say with a straight face that they have any right to the RNT.

Even in a randomised controlled trial (where the trial managers themselves won’t know which patients are getting the RNT), anecdotal clinical information about the relative efficacy of BT and RNT will become known by the treating clinicians, and used by them in their own future clinical practice in much the same way as if they had read the results of a formal trial.

Suppose that a doctor is interested in RNT. He sees a patient who is a candidate for RNT. He knows that there will be no funds for RNT unless the patient enrols in a clinical trial. The patient wants the RNT. There is nothing unethical about offering the patient the chance of enrolling in the trial. Indeed  if the clinician thought that RNT might be a good thing, it would be unethical not to give the patient that chance.

Now suppose that the facts are identical, with one exception. RNT would be available (although much more expensive than BT) even if the patient did not enter the trial. The patient wants the RNT. The doctor tries to recruit the patient into the clinical trial, but the patient refuses. The doctor then says: ‘Right: if you’re not going to sign the trial consent form, I’m not going to sign the prescription for RNT. Here’s one for BT.’  Is this unacceptable?

This is the situation, detailed by Mathur, that has caused the storm.

The patient has no legal right to the RNT. He has no moral right either. Why might he be refusing to enter the trial? Broadly because he is a selfish, snivelling brat – a position that can be described philosophically in many different ways. However it is put, it is disagreeable, and ought not to be encouraged. But it is also utterly irrational. If the doctor relents, and gives the patient the RNT without enrolling him in the trial, the information about RNT’s performance in that patient will still be used by the doctor in deciding on his own future use of RNT, on what he will say about RNT in lectures, textbooks and so on. Yes, there would be more use of the information if the patient were part of a trial, but if there is any identifiable detriment in the use of the information, the additional detriment is minute. On any sensible view the benefits to others of the patient’s participation in the trial hugely outweigh the additional detriment.

The doctor should be stern: ‘If you want the RNT, Mr. X, into the trial you go.’

  • Facebook
  • Twitter
  • Reddit

One Response to We’re all guinea pigs, and it’s not so bad

  • There may be another kind of right to the RNT you’re overlooking – an institutional/professional right. That is to say, physicians are under a general professional duty to patients to prescribe the treatments he or she thinks will be efficacious. To deny a prescription based on failure to participate is an abrogation of that professional duty. Furthermore, it might constitute physician abandonment if the physician does not refer the patient to a doctor who *will* prescribe the RNT. Moreover, I think this professional duty carries some moral weight – it is a kind of tacit promise when a patient first comes to a physician which must be respected.

    Additionally, it is important to remember that there is more to study participation than simply receiving a certain treatment and then being passively observed. Clinical trials generally involve numerous tests and procedures which impose burdens and risks significantly above what would occur in usual care. This is one of the main reasons for robust regulation of clinical trials in the first place – in a clinical trial, patients can waive a physician’s general professional duty to only recommend medically indicated tests and treatments. This is to be contrasted with the use of anonymized data, where once the information has been gathered no extra burdens or risks will accrue to the subject (indeed, in the US, the use of deidentified tissue or data is not even classified as research on human subjects, and is thus exempt from direct regulation of such). There is, then, a disanalogy between the acceptability of using anonymized data and the acceptability of making usual care contingent on study participation.

    That having been said, the central motivation for Mathur’s claim is still compelling. Those who refuse to participate in clinical trials are indeed acting somewhat selfishly. In particular, they are willing to accept the benefits of previous studies (novel treatments, effective tests, and so on) but unwilling to themselves participate in that necessary step in developing such medical benefits, namely study participation. Mathur’s proposed solution may be untenable due to, e.g., its violation of professional codes, but I think it’s great that we’re looking for solutions to insufficient study participation not from the perspective that participation is morally optional, but that patients often do have a moral duty to participate in research.

Authors

Subscribe Via Email

Affiliations