How refreshing to read these two cases! No conjoined twins, fantastical chimeras, or other incredible scenarios at the fringes of medical reality. Each case highlights the practical and theoretical difficulties that doctors face in their everyday practice.

Case 1: In case 1, the patient, who had declined an HIV test, changed his mind and requested an HIV test on the request form without informing the doctor. The pathologist performed the tests and returned the results to the unsuspecting doctor, who duly investigated the matter. The doctor’s anxiety at the situation, as well as the worries of the entire practice and pathology laboratory, must have been deeply unpleasant. Thus the patient, no doubt unintentionally, caused much distress by requesting the HIV test. He was, at the very least, guilty of a lack of foresight and consideration. As a minimum, he should have informed the doctor of his last minute alteration. This would have enabled the doctor to discuss the implications of the test, to organise appropriate counselling, and, most importantly, to receive the results in full knowledge of the situation.

As Dr Fraser points out, the experience of case 1 reinforces the importance of taking good medical notes. Even if the patient’s refusal had been documented in the notes, however, this would not have prevented him from requesting the HIV test, and indeed other tests, on the blood form. Pharmacists may still occasionally be faced with situations when they are unsure whether the doctor prescribed 10 capsules of a drug, or 110 capsules, as the patient may have added a “1” in front of the doctor’s original prescription. Similarly, how can doctors and pathologists guarantee that patients will not modify the form as soon as they leave the surgery?

Whatever their immediate consequences, mishaps in medicine often provide opportunities to improve practice. Once the causal factors are identified, changes can be made and implemented. So what can be done to prevent reoccurrences of incidents such as those in case 1?

First, patients can be taught to refrain from ordering tests without prior approval by their doctor. The doctor/patient relationship, like all relationships, is not one sided. Patients have duties too. In general, they should be truthful, cooperative, courteous, considerate, and trustworthy. All of these duties should have compelled the patient in case 1 to inform the doctor of his alteration. Whether—and, if so, how—to instil these duties in the doubtless small number of patients who are unaware of them is beyond the scope of this brief commentary. None the less, whatever the efforts to raise awareness of patients’ duties, it would be unrealistic to expect all patients to abide by them. Our solution should operate successfully in a non-ideal world where some patients are, as one general practitioner put it, “a pain”.

The key concern arising from the case relates to the content of pathology request forms. If in possession of a written form, it may be easy for patients to order additional tests without the doctor’s knowledge. Doctors trust patients to deliver the request forms and specimens from their office to the next step of the delivery process without tampering with them. To reduce the likelihood of tampering and to increase detection if it does occur, computer generated forms may be preferable to handwriting. Delivery of forms and specimens without involvement of the patient would also reduce the possibility of unauthorised changes. Whether such changes are practically desirable, given the time constraints and the rarity of abuse, is an issue for consideration. If pathologists suspect a discrepancy in authorship—for example, because of different handwriting—they should contact the general practitioner for confirmation. Alternatively, separate forms could be used for HIV or other “sensitive” tests, as is already done in some areas of the United Kingdom.

Although case 1 exposes a systemic weakness in the completion and transit of pathology request forms, its main interest lies in the important but neglected questions it raises about the role and duties of patients.

Case 2: Case 2 illuminates a common dilemma for doctors: when should a possible but very uncertain diagnosis be communicated to a patient?

The doctor had two opportunities to convey the possibility of HIV. The first was when he investigated the granuloma. Because there is a link between pyodermic granuloma and Kaposi’s sarcoma (KS) and, in turn, between KS and HIV, the doctor could have informed the patient of this unlikely possibility. Now, it is probable that Dr Fraser did not even suspect KS at this point, let alone HIV. The probabilities of each, in light of the clinical evidence and the doctor’s knowledge of the patient, would have seemed minuscule.

Although one often reads of “fully informed consent” in the medical literature, such a state of total knowledge and understanding is a myth. It is impossible for doctors to disclose all the risks and possible side effects of a drug, treatment, or procedure. Not only would attempting to do this impossible task be practically disastrous, prolonging waiting times to unacceptable levels, but it would be medically and psychologically undesirable. A headache is very rarely a brain tumour, and there seems little point in alarming all those with headaches of this grim possibility. Only when the likelihood reaches a certain threshold of significance should the suspicion be communicated. By anyone’s account, the threshold for disclosing the possibility of HIV was surely not reached at the time of investigating the pyodermic granuloma. Even if Dr Fraser had entertained the possibility, it would arguably have been inappropriate to reveal this thought for fear of unduly distressing the patient.

The second opportunity arose following the suspected KS. A few points should be considered. Firstly, the pathology report only queried KS. The diagnosis was thus far from positive. Secondly, there is no necessary link between KS and HIV. Kaposi’s sarcoma may occur independently of HIV. Thirdly, the gravity of KS varies considerably from relatively harmless lesions to fairly rapid death—as in African KS—depending on the type of KS, the extent of organ involvement, and various patient characteristics.¹

Dr Fraser’s decision to withhold the suspicion was based on three reasons:

1. Discussing the issue of HIV would very probably cause distress.

2. The probability of HIV was low, hence discussion of this possibility would be unnecessary.

3. The clinical course would not be altered by the delay to obtain confirmation of KS.

Of these, only the second reason is valid. The fact that a disclosure is likely to distress a patient is not, in itself, a sufficient reason to withhold information. Unless the anguish caused by the revelation would be extremely harmful, the crucial importance of information for autonomous decision making surely outweighs the benefits of non-disclosure. Furthermore, it is difficult to know how others will respond to bad news and, even when it is highly distressing, the deleterious effect of adverse disclosures is often temporary. In cases of sexually transmitted diseases, the communicable nature of the disease provides additional weight to the prima facie duty to disclose.

Although a delay may not alter the clinical progression of the disease in the individual patient, HIV can be transmitted to others. Informing patients of this possibility can help reduce the likelihood of transmission by allowing them to control their behaviour. For this reason, suspicions of communicable diseases should be shared with the patient at a lower threshold of certainty than non-communicable conditions.

Without detailed knowledge of the patient, his history, and his lifestyle, and the probability of the initial diagnosis of KS, I can only come to a tentative decision about what I would have done myself in Dr Fraser’s situation, but, based on the information provided in the case description, I am inclined to support his conclusion. In light of the queried test results, the probabilistic link between KS and HIV, and the patient’s “risk free” lifestyle, the probability of HIV probably fell below the threshold of likelihood requiring disclosure, even for a condition as serious and transmittable as HIV. One caveat, however, about the dangers of assuming the absence of risk factors: even a prolonged friendship with a patient may not reveal an unavowed penchant for drugs or hazardous sexual behaviour. A good doctor should never rule out these possibilities, however unlikely they may appear.

William Osler called medicine “an art which consists largely in balancing probabilities”.² This case is a wonderful illustration of this.