The State of Prognostic Art

The state of the prognostic art leaves much to be desired. ‘Of the trilogy of disease, diagnosis, prognosis, and treatment, prognosis is the most difficult to evaluate. The accurate prediction of things to come is often most baffling, perplexing, and problematic. Caution is essential. The less said the better. Remember that we are endowed with two eyes and two ears but with only one tongue. The implication must be apparent.’¹²² This is but natural, for prognosis as a subject has been least touched upon in medical literature. In A Medical Bibliography complied by Morton ¹²³ and published in 1970 in its 3rd Edition, there are 7,534 entries dating back to the time 2250 B.C., of the great Hammurabi; of these, only one entry is on prognosis, viz. Depraesagienda vita et morte aegrotantium by Prospero Alpino, published in 1601. In Familiar Medical Quotations, edited by Strauss¹²⁴ and published in 1968, there are more than 7,000 quotations on over 400 subjects - ‘from Cathay’s Huang Ti, five thousand years ago to present day opinions on transplantation and birth control.’ On ‘prognosis,’ there are eighteen quotes of which only two are from specific works on prognosis; one from Hippocrates ‘On the prognostics, and the other from a small editorial by Robbins,¹²² published in 1961 in the Archives of Internal Medicine, and quoted from, as above. All told, medical prognostication is more of an art than a science.

Prognosis in Cancer

One who makes a prognosis in cancer, blissfully unaware of his limitations and rather too sure of the ideas of early and late cancer, indulges in two extremes: (1) offering hope when hope may not be rational, or (2) presenting hopelessness when hopelessness may not be warranted. And he can get away with either. The first measure allows him to be condoned on the grounds of his benevolence; the latter measure provides a subtle defense for him, for rarely does a patient surviving longer-than-expected have the heart to find faults with his prognosticator. There are doctors who, to show their worth and to be sure of an excuse, make bad seem worse and of the worse make a disaster.

’¹²⁵ The reasons why a cancer prognosticator feels so sure of himself are many. His diagnostic and therapeutic skills register advances everyday. His assessment of a cancer case is seemingly complete - clinical examination, endoscopy, an arsenal of investigations, and the findings at the operation.

Yet despite this impressive array of aids to prognosis, the prognosticator encounters unreliability, at every stage. What he has always thought to be an early cancer has rarely been so. He takes the small size of a tumour as his guide, but that cannot help him: The general assumption that all dwarf-sized cancers must be biologically young is no more valid than the assumption that all human dwarfs must be young because they are small.’¹²⁶ The mode of cancer growth renders usefuless any attempts at detecting the silently growing tumours measuring less than half a centimeter in diameter.

Cancer cytology is highly arbitrary and therefore unreliable. Histology does not fare any better. ‘Contrary to the experience of some workers, we have not found that the histology of biopsy specimens offers any useful guide to prognosis or management.’¹⁰⁴ This1970 generalization by Ho¹⁰⁴ - in whose series most cancers were undifferentiated - is similar to the 1960 generalization by Sutherland⁹⁵ that, at present, prognostically different cancers are often morphologically indistinguishable. The grading and staging of cancers represent, the valiant efforts by prognosticators at playing the prophet, depending on apparently objective criteria; but’ a given carcinoma may be graded II one day and III next, or vice versa, depending on the functional tone of the gastrointestinal tract of the pathologist or the barometric pressure.’¹²⁷ And what if the grading were to be precise? Writing on the grading of the adenocarcinomas of the colon and rectum, Boyd¹²⁸ comments that while statistically it is possible to establish some agreement between the grade, lymph node involvement, prognosis, and so on, ‘this does not mean that it is of prognostic value in the individual patient.’

Prognosis is cancer, like in other branches, is a judgement based on circumstantial evidence, but no judgement can be respected when the evidence is largely suspect. To the prognosticator , cancer is what and where he sees it. However, his detecting cancer at one or more sites is no guarantee that the cancer is not additionally present elsewhere. Moreover, what he sees as cancer is an independent, biologically predetermined behavioural entity, that does not permit him to tell: What really is the cancer? Where else is it lurking? What will it do? And when? And when will the patient die of something totally different?

Despite such ignorance, an all too common pitfall is the urge to make favourable prognoses on the basis of ‘early ‘ treatment. It was as early as in 1936 that Nathanson and Welch¹²⁹ reported that in their series of breast carcinoma, ‘patients with the shortest delay of the treatment have the worst prognosis.’ Not infrequently, a prognosis of doom proves wrong, and the patient survives, showing that the prognosticator had seen a disaster greater than was in store for the patient.

Role of Statistics

‘In individual prognosis,’ Hyman¹³⁰ remarked, ‘statistics function as a weather-vane. From them the practitioner recognizes the wind direction; he knows nothing of wind velocity, or of weather conditions such as temperature, humidity or visibility.’ The prognosticating physician is, by and large, unaware of the weather-vane-nature of statistics which come to him as definite, reliable, proved-correct-generation-after-generation figures in authoritative writings on diabetes mellitus,¹³¹ coronary heart disease,¹³² hypertension,¹³³ or various cancers^.20,134. The prognosticator has nothing to guide him in an individual case - for whom he can, at best, retrognosticate or be wise after the event.

Backed by an implicit faith in the truth of large numbers, the prognosticator finds it convenient to extrapolate the herd data to an individual, ignoring the fact that the extrapolation is fraught with Heisenbergian uncertainty. ‘In biological problems, variable factors of considerable complexity often are present, the necessary consideration of which distinguishes biometry from statistology.^’135 Such warnings escape the eyes of the prognosticator and so he continues to prognosticate despite Heisenberg or McDonald. Even when he employs statisticalprognosis, the prognosticator probably neglects cautionary statements often appearing at the very beginning of the text. ‘It is true in diabetes mellitus as in other chronic diseases that the prognosis for the patient is extraordinarily individual.¹³¹ The cautionary note on an individual is followed by one on a group: ‘Generalization with regard to prognosis may be based on averages in special groups and for special complications; nevertheless, wide variations are found in the duration of life and the presence or absence of diabetic sequelae within each group.’¹³¹

How realistic would it be for the patient were the prognosticator to admit that all that he is offering prophetically is statistical! How unburdening would it be for the prognosticator to realize that, at the level of an individual he need not prognosticate at all!

The Need for Prognosis

Notwithstanding his crass ignorance on the whether when- how-and-why of the art, the physician must prognosticate. Brooke,⁸⁸ writing on cancer, described prognosticating as ‘perhaps the most important act in medicine.’ And perhaps, this is true for altogether different reasons, namely to share with the patient and his near- ones the usually unacknowledged medical ignorance on cancer and to let the patient know that cancer can be as kind as it can be cruel. Set below are a few suggestions:

1. The time of prognosticating is the time to talk things over with the patient. It is the time to act as a patient’s friend by providing him with the drive to dare the disease, and to live with it.

2. Prognosis involves exploring and exposing areas from where assurance can be had and destroying areas from where unwarranted fear stems. Cancer patients often live in depression and it is for the prognosticating physician to pull them out into living a year- saying life that meets with the Kiplingian urge to fill every irretrievable minute with sixty seconds’ worth of distance run.

3. Prognosticating includes protecting the patient against the tyranny of lay and medical articles rich in well-intentioned scare- mongering.

4. Prognosticating does not include,. despite Hippocrates, guilt- pointing and fault-finding. Carcinoma of the lung in a smoker, or carcinoma of the stomach in a gourmet, or carcinoma of the cervix in a woman who has loved live, is no Dostoevskian story of Crime and Punishment. The occurrence of a lethal laryngeal cancer, in that sage from Dakshineshwar, Ramakrishna Paramahansa - whom Wilson⁸⁷ calls a great mystic, a God-intoxicated saint - was certainly no retribution from a wrathful God.

5. Active Patient-Participation. What I do not know, is unfathomable. What I do know, is shareable. Our ‘the-more-we-know-about- cancer-the-less-we-seem-to-understand-it’ predicament has attained sufficient magnitude to enforce the prognosticator to practise the above code of conduct vis-a-vis his prognosee.

6. Herd-realism, normal distribution, Gompertz function, curves of disease-specific mortality, etc., are subtle and inexorable indicators of the fact that an individual, despite all his unprecedented, unparalleled and unrepeatable uniqueness, is herd-dependent with respect to many features. In his chapter on the ‘Statistical study of tumours,’ Willis²⁰ emphasizes that the age distribution of a sufficiently large series of cancer deaths, in a population, provides ‘ a smooth ideal curve’ of normal distribution. This normality of distribution is a herd-function, and, at an individual level, depends on the point of the curve one falls on so as to die of cancer at the age of eighteen, or ninety-eight years. To the present set to die of cancer at eighteen, ( as well as to his near ones), it is ‘chaotic’ that he should be so ‘victimized’ by Nature. But if he and the others realize that his ‘chaos’ is a part of the orderly ‘ideal’ curve, the sense of persecution is likely to be minimized.

7. Prognostication in cancer should include retrognostication consisting in explaining to the patient that his cancer, dis-easing him now, has been with him for five to fifteen years. Further, that the earliness or lateness of a cancer lies in the mind of the clinician, and not in the cancer.

8. The patient’s cancerrealism that he is harbouring a cellular phenomenon of which even the prognosticator is only as wise as he himself can make him an equally important participant in the fight against the disease. In the absence of such realism, the patient suffers from a sense of singular victimization out of the frustration that medicine and medical men are not offering him his due.

9. Prognosticating includes admitting investigational limitation and therapeutic impotence.

10. The cancer-can-be-cured syndrome is no different from the well-recognized ICCU (or ICU) syndrome.^6,136,202 (ICCU stands for Intensive Coronary Care Unit. The syndrome is suffered by heart patients admitted to the ICCUs, as also by the medical and paramedical staff attending to them. The ICCU syndrome symbolizes people’s and medical men’s faith in the marvels of modern medicine bought at an enormous psychic and monetary expense, without good done to anyone.)

    The cancer prognosticator must see to it that his patient’s body, mind and soul are not additionally burdened with the above syndrome, and that the syndrome does not kill the patient’s family while medicine is fruitlessly trying to save the patient.

11. ‘Prognosis is a continual process that may extend beyond the patient’s death, for the bereaved ones.’⁶ This can go a long way towards assuaging the unhappiness, anger and bitterness of the patient’s near-ones.²⁷⁹

12. An excellent way of winning the confidence of a patient, while prognosticating cancer, is to allude to the not-very-uncommon event of the cancer-patient outliving his cancerologist: Evarts Graham, the famed St. Louis surgeon who introduced ‘curative’ operation for lung cancer - called ‘Graham’s operation’ - operated on a doctor who survived to see Graham dying of lung cancer ‘not diagnosed until it was too late to apply the operation that he had developed.’²⁸⁰ Such an exercise in medical humility, whereby the healthy-looking physician admits that the diseased prognosee could well outlive him, may find a starting point in an old skip-rope song: