Validation in Somatic Medicine
Let us choose as our point of departure someone who, for instance, complains of headaches to his family doctor. The number of physical aberrations that can be “responsible” for this complaint is huge. The physician will try to orient himself, apply a simple diagnoses, and attempt to influence the complaint with simple treatment. He will almost never consider validating an extremely rare disease of which headache is a symptom, for instance echinococcosis of the central nervous system, during the patient’s first visit. That would require a thorough, deep, expensive and not risk-free examination. He will consider looking for and validating rare disorders only when multiple attempts to find simple explanations have failed. He must constantly contemplate which diagnosis is relevant to the complaint. When during an influenza epidemic someone displays the symptoms of influenza, the physician as a rule will assume that this patient also has influenza. Judgment is based on the assessment of different likelihoods. Conversation with the patient is of utmost importance because based on it the doctor can decide whether further validation is necessary. Are the complaints typical of something specific or vague? Do they seem relatively mild or do they indicate something serious?
Even when the physician conducts a number of examinations which he uses to validate his hypothesis about the patient’s complaints, even when he remains on the side of caution by performing or ordering more examinations than he considers strictly necessary (most physicians would rather examine too much than miss a diagnosis ) he will only very rarely wish or be able to do all of the examinations possible regarding a certain complaint. If he tried that the examinations would soon be worse than the affliction. In addition to the risks and complications of the diagnostic measures themselves, the patient would be led to believe that there is really something seriously wrong with him, and perhaps become fixated on his complaints. In addition, it would make health care exorbitantly expensive. This means that the physician is constantly in a process of assessment. How far must he go in examining the complaint? Which risks must he rule out? Which risks can he take?
If he decides to continuing examining by physicochemical means in order to confirm his suspicions or rule out risks, such means can provide him with either correct or incorrect information. All possible forms of examination render a certain percentage of correct results and a certain percentage of false positives and false negatives. The percentage is different for every method of examination. Besides, the normal values of for instance many laboratory findings range broadly, and clear lines between normal and abnormal can seldom be drawn. Also, in any examination, mistakes can be made. Further, it is inherent to every physicochemical examination that the results must be interpreted. This can be extraordinarily difficult and require a high degree of training and acuity, for instance interpreting x-rays and EEGs, and thus be a source of inaccuracy, error, and misunderstanding.
Cochrane et al reported that experienced radiologists, when evaluating a series of photos of the thorax, overlooked 30% of the aberrations shown by the photos, while observing aberrations in 2% of the photos that had none. Davies presented 100 ECGs to nine experienced physicians and one less experienced physician for assessment. There was unanimity on 30% of the ECGs, some difference of opinion on 50%, and gravely differing opinions on 20%. Several weeks later, all the physicians assessed one in eight ECGs differently than they had done the first time. The experienced evaluators succeeded much better than the inexperienced evaluator. Garland lists a large number of examinations, including the following: some laboratory examinations revealed serious errors in 10% to 28% of the results. Erythrocyte counts varied by 16% to 28%. He recommends having x-ray photographs assessed either by two experts or by the same one at two different occasions.
So for a somewhat reliable assessment it is necessary to conduct a large number of examinations, preferably at different times. When a large number of factors deviate in the same direction a certain aberration, or diagnosis, becomes more likely. The more finely the criteria for a certain diagnosis are defined, the less often that diagnosis will occur. Engle et al. state, “Thus, inherent in every diagnosis is a factor of uncertainty, greater in some and less in others. The uncertainties are partially related to our imperfections of knowledge concerning health and disease with all of their manifestations, and also to the most useful way of thinking about classifying and naming them.” A diagnosis is an evaluation of probability. The probability is sometimes high, sometimes lower, and rarely approaches 100%. “Demonstrating” the accuracy of a diagnosis, delivering the “proof” that a certain disorder exists, is in fact: making assumable that the disorder is likely.
From the above we can infer that the used methods are of a physicochemical nature, yet the assessment of the result of such examination must be done by human perception, not only in the sense of taking readings from instruments, but also in interpreting quite complicated patterns. This means that the objectivity of these methods is limited by the possibilities and boundaries of human perception.
Only a physicochemical disorder of the body can be demonstrated, and validated, by a physicochemical method. The patient’s complaint, for instance, his subjective experience of pain, can strictly speaking not be validated this way, nor can the diagnosis.
Whether such physicochemical examination is the most effective and least harmful way of validating is not so easy to determine. Reiser posits that the value ascribed to clinical dialogue in the nineteenth and twentieth centuries fluctuated a great deal. Today, impersonal, “objective” methods are valued most. He explains this as partly due to the fact that research as to the value of anamneses has lagged behind research as to the value of “objective” examination methods, and that also during training the anamnesis is neglected. Engel agrees.
Feinstein states the same regarding clinical observance: there is a classification of diseases and physical aberrations but not of clinical pictures. The preoccupation with “objective” findings led to the dehumanization of medicine which in turn led to “bad therapy and bad science.” He advocates a renewed interest in clinical symptoms and clinical phenomena “to restore the patient, rather than the disease, to his proper place as the center of the universe of clinical science.”
So it turns out that a value judgment is partly the basis of the preference for “objective” validation. The idea that physicochemical results provide better information than careful anamneses and careful clinical examination is a premise that is probably derived from this value judgment. It is tempting to relate this premise to the biomedical concept of disease. When the existence of disease is believed to be linked to the presence of physicochemical aberrations, finding such aberrations is of utmost importance.
Finally we come to the postmortem examination by autopsy. This examination provides general information about organic aberrations that are found after death. It is thus a test of diagnoses made during life. Steffelaar found that in a series of 163 postmortem examinations 69 cases (42%) had unexpected aberrations. In 40 cases (24.5%) either the main diagnosis was wrong or an unexpected complication which contributed to the death was found. He found that 25% of malignant tumors were not diagnosed during life. This margin of error was found in a modern hospital which had access to all the usual validation methods. Prutting mentions more of such research in the United States and France, with similar results. So too does Sanders.
Table of Contents