Commentary: No simple and absolute ethical rule exists for every conceivable situation

BMJ 1997; 314 doi: (Published 12 April 1997) Cite this as: BMJ 1997;314:1083
  1. Y K Seedat, chairman of postgraduate committeea,
  2. Y K Seedat, (chairman)a,
  3. M Adhlkaria,
  4. M H Cassimjeea,
  5. V Gathirama,
  6. V B Jogessara,
  7. J Moodleya,
  8. D J Pudifina,
  9. J V Robbsa,
  10. S R Thomsona,
  11. J R van Dellena,
  12. E M Barkera,
  13. E M Barkera,
  14. S Downesa,
  15. R Gcabaa,
  16. D J McQuoid-Masona,
  17. M E de Haasa,
  18. U Govinda
  1. a Medical School University of Natal Durban South Africa


    The obvious ethical problems posed by this study concern (a) the fact that all patients admitted to the intensive care unit over a six month period were included in the study without their knowledge or consent, and (b) the fact that blood samples obtained from all patients were tested for HIV infection without the consent of the patients, with the information that blood samples had been tested for the infection being given to patients only after the test had been done.

    At first sight the decision to override the patients' right to full information, and to give or to refuse consent to inclusion and testing, seemed to all members of the research ethics committee to be so fundamentally at variance with the ethical principles governing research involving patients that it seemed impossible to give ethical approval for the study. However, during lengthy discussions with the investigators several considerations emerged.

    Firstly, the information being sought by the investigators was clearly going to be of crucial importance to the community, not only in South Africa or in Africa as a whole but also worldwide. The importance of the study was perceived to be twofold. If it showed that a patient's HIV status significantly worsened his or her chances of a favourable outcome from intensive care–to a degree comparable to the poor prognosis associated with criteria already established for non-acceptance for admission to an intensive care unit–then the clinicians who have to make decisions on allocating the community's scarce intensive care resources would have to include HIV positivity among the criteria for non-acceptance. If, on the other hand, the study showed that HIV positivity per se did not adversely affect a patient's likelihood of a favourable outcome, then the current widespread tendency to include HIV positivity among the criteria for non-acceptance into intensive care facilities would become manifestly unjust. Such information gained from the study would be of life and death importance to the large and increasing numbers of people who are HIV positive.

    Secondly, the study entailed no interventions of any sort different from those that are necessary and are carried out in standard intensive care. The blood samples that would be tested for HIV infection would be aliquots of samples taken for other necessary clinical purposes. Apart from the HIV testing, the study did not depart from normal standard of care and consisted essentially of analysis of data that would be recorded even if the study were not undertaken.

    Thirdly, apart from the HIV testing, the “injury” that would be done to the patients as a result of not being given the opportunity to consent to or to refuse inclusion in the study was considered to be so small as to be virtually not appreciable and entirely analogous to the “injury” to patients whose hospital records are reviewed for retrospective research projects. Given the importance of the study, the failure to ask patients for permission to analyse data necessarily generated during their clinical care did not seem to be material.

    Fourthly, testing the patients' blood for HIV infection without their consent and only informing them afterwards posed an important ethical dilemma. In considering this aspect of the study, the committee took into account several considerations. It agreed that there is no such thing in ethics–and particularly in the increasingly complex field of bioethics–as a simple and absolute ethical rule that must be observed in every conceivable situation. Virtually every ethical dilemma necessarily poses the problem of competing and conflicting ethical obligations. There are no absolutely satisfactory resolutions of ethical dilemmas, and the best that one can hope to achieve is to accord, with justice, preference to those ethical considerations (or “rules”) that seem in the particular circumstances to be of preponderant weight.

    The committee was also at pains to satisfy itself that the effective performance of the proposed study could not be achieved if any of the subjects were not to have their blood tested for HIV infection. Unless it could be shown, scientifically, that it was absolutely essential to include all admitted patients in the study, the committee would not have considered the proposed testing of blood without consent as ethical.

    The committee was also strongly influenced by the fact that the results of the HIV tests would remain strictly confidential to only one investigator and that all potential linkage of the results of the tests to identifiable individuals was to be destroyed at the end of the study. The situation, as the committee saw it, was analogous to the anonymous and unlinked testing of attenders at antenatal and sexually transmitted disease clinics, for epidemiological purposes. This testing, to be of value, has to include all attenders, and for this reason consent to testing of aliquots of attenders' blood samples taken for other purposes is not obtained. This practice has ethical approval throughout the world, on the basis that the community's need for reliable epidemiological data outweighs by far the almost imperceptible injury done to the patient's autonomy. From a practical point of view, the clinic attender is in the same situation as he or she would have been if HIV testing had not been done at all. Similarly, for the patients in this study the end result was the same as it would have been if their blood samples had never been tested, with the sole difference that, if they so wished, they would be informed of the outcome of the test. Weighing the importance of the study in terms of the welfare of the community against the almost imperceptible injury that would be inflicted on patients, the committee was satisfied that the proposed method of obtaining complete data regarding the patients' HIV status was ethically acceptable.

    In the outcome, it seems that the committee's view on the ethics of this study was vindicated by the fact that no patient expressed any objection to the fact that his or her blood had been tested in this fashion. Furthermore, the fact that only two patients elected to be informed of the result of the test is in keeping with the general reluctance of well people to undergo HIV testing and suggests that if inclusion in the study depended on a patient's consent to HIV testing–even if effectively performed anonymously–then it is quite likely that the study would not have produced a reliable outcome.


    The University of Natal's ethics committee is a subcommittee of the postgraduate committee.

    View Abstract