The biomedical model views causes (etiologic agents) as leading inevitably to their consequences (disease). Within this model, chance and uncertainty have a very small role. In the 1950s, doctors downplayed notions of probability and statistical inference from epidemiological data. The conventional wisdom was that epidemiology dealt with populations, and as such, was incompatible with clinical medicine, which dealt with individuals. The 1965 edition of DeGowin’s Introduction to Clinical Medicine stated that “…statistical methods can only be applied to a population of thousands … [T]he relative incidence of two diseases is completely irrelevant to … diagnosis. A patient either has or has not a disease” [13].

My clinical tutors similarly rejected the application of statistical inference to clinical practice by using expressions such as: “Nobody is 70 % pregnant,” and “Every patient is unique; I learn from a single case more than from epidemiological studies of a thousand patients.” The deterministic reasoning of the biomedical model also downplayed notions of uncertainty. I was taught that “Nothing is left to chance if the patient is properly worked up.” As late as the 1980s, several authors claimed that students were encouraged to ignore uncertainty rather than accept and deal with it [14–16], and a 1992 review of the sociological literature concluded that, “denial of uncertainty was one of the most consistent observations made by sociologists studying medical training” [17].

One of the first indications that this deterministic approach might be inadequate were observations that such conditions, as diabetes and smoking, increased the risk of ischemic heart disease, although they were not etiologic agents of atheromatosis. This observation suggested that disease was not the result of a single cause, but rather of a convergence of risk factors. Today, we speak of risk indicators rather than of etiologic causes; a patient is more or less likely to have a disease, rather than either have it or not. In 1993, DeGowin’s Diagnostic Examination replaced the rejection of the application of statistics to individual patients with the ambiguous statement “…Using the probability theory to evaluate the sensitivity and specificity of clinical findings … may serve to strengthen the credibility of these parts of the diagnostic examination as future clinical trials demonstrate their validity” [18]. In the 2009 edition, phrases, such as “relative disease probabilities,” “more or less common diseases,” and “incidence and prevalence of diseases in the patient’s age group” appear repeatedly in the chapter on diagnosis [19].

A change occurred in doctors’ attitudes to diagnostic tests . In the 1950s, teaching emphasized the importance of a thorough accrual of a patient’s history, and physical and ancillary data. Doctors believed that the more data obtained, the better the chances of making a correct diagnosis. Early detection of disease was thought to always improve the likelihood of cure, and therefore, students were required to perform a thorough review of systems and a head-to-toe physical examination . Today, doctors are aware that false-positive test results may confound diagnostic reasoning , and consequently, are selective in the choice of diagnostic tests and in screening asymptomatic persons for early detection of disease. Paradoxically, the increase in biomedical knowledge seems to have enhanced the awareness of its limitations and led to a transition from right/wrong determinism to acceptance of chance and uncertainty in clinical practice.

The Webster defines “intuition” as a perception that is independent of reasoning. A judgment is said to be intuitive when it is made rapidly and without apparent effort. In medicine, the term “intuition” has acquired several meanings. The first one equates intuition with an ability to recognize patterns/configurations of disease manifestations already encountered. The second meaning of intuition refers to judgments arising from simplifying heuristics (mental shortcuts) elaborated in Chap. 14. Finally, the term intuition is used to denote mastery of the “art of medicine,” which is a mystical ability to make clinical decisions that are not amenable to analysis or explication.

To prevent confusion I shall use the terms “pattern recognition” to refer to an ability to recognize previously encountered disease manifestations; “heuristics,” to refer to cognitive shortcuts; and “intuitive reasoning” or “art of medicine” to refer to the ability to make rapid decisions that are not amenable to explication. In the 1950s, this latter ability was characterized by confidence in its accuracy, precluding any attempt to improve clinical reasoning . Statements by experienced doctors beginning with “in my judgment” implied the closure of a clinical debate.

Such authoritarian attitudes are hard to comprehend in today’s climate of evidence-based practice. To understand doctors’ attitudes to the art of medicine, one must consider its strengths. First, the belief that in the indeterminate realm of clinical practice there is an absolute truth is extremely appealing. Conformity with authority has been identified as a means by which medical students and residents control anxieties generated by the complexity of clinical practice [20, 21]. Second, intuitive reasoning satisfied clinical needs. In the 1950s, doctors were not aware that there was anything wrong in clinical decision-making, or that it was in need of improvement.

However, since the 1950s, clinical decision-making has become more complicated. The number of diagnostic and therapeutic options expanded and the choice between them is no longer easy, even for experienced clinicians. Doctors have to consider the trade-off between the benefit of medical interventions and their risk: the risk of false-positive or false-negative diagnostic test results, and the risk of undesirable side effects of treatment. Health-care expenditures have increased and this necessitates considerations of trade-offs between cost and effectiveness. In the present reality of unlimited demand and finite resources, economic appraisal is increasingly being used to inform health-care decision-making. Finally, doctors have to balance conflicting values. The medical code of ethics has expanded to include not only the principles of non-maleficence “do no harm” and beneficence “do good,” but also “respect a patient’s autonomy” and justice “be fair in distributing health-care resources to those who need them.” Consequently, doctors today are much more likely to confront ethical dilemmas, i.e., to find themselves in situations in which they cannot honor one ethical principle without violating another.

Obviously, these trade-offs can no longer be resolved by intuitive reasoning: first, claims that the “art of medicine” eludes explication are no longer acceptable when patients, students, colleagues, and courts of law ask us to justify our decisions. Second, judgments based on simplifying heuristics may be confounded by cognitive biases [22, 23]. Therefore, since the 1970s, there have been sustained efforts to gain insight into the reasoning of expert clinicians, and to base clinical decision-making on a critical appraisal of risks and benefits.

Another assumption of the biomedical model was that patient care should be deduced from disease etiology and pathophysiology. In the 1950s, the justification for treating patients with left ventricular failure with digitalis was its positive inotropic effect on the heart muscle. The justification for treating patients with peptic ulcer with aluminum salts was their antacid activity. However, in recent decades, doctors have become aware that pathophysiologic rationale does not always produce the expected outcomes. Consequently, there has been a growing tendency to base clinical practice on empirical evidence, and in 1992, Guyatt et al. coined the phrase “evidence-based medicine” (EBM) [24].

Unlike the traditional paradigm, EBM posits that intuition, unsystematic experience, and pathophysiologic rationale are insufficient for clinical decision-making. EBM places a lower value on authority and stresses the value of evidence from clinical research. From deductive reasoning we have moved to evidence-based reasoning: Digitalis treatment of heart failure and antacid treatment of peptic ulcer are justified not by their expected physiological effect, but rather by evidence provided by clinical trials.

The main flaw of the biomedical model was the assumption that all diseases are structural or biochemical dysfunctions of the body, and amenable to treatment only by surgical or pharmacological means. This assumption excluded the patient’s attributes as a person. It is true that, as early as the 1930s, there were attempts to correlate personality types with diseases, such as peptic ulcer or bronchial asthma. However, these attempts were mostly based on anecdotal observations. In the 1950s, the prevailing attitude towards psychosomatic medicine was one of mistrust.

A change in this attitude was brought about by evidence of the association between life events and morbidity [25], and between measures of socioeconomic status (income, education, and housing) and mortality [26–28]. More recent studies have detected an association between income inequality (rather than income per se) and population health [29]. These findings led to the recognition that some risk indicators of disease can be described as psychosocial, and to the adoption of Engel’s biopsychosocial model of clinical reasoning and practice [30]. The latter model encourages physicians to look into the biomedical and psychosocial components of a patient’s predicament and provide support and treatment for both. The model’s premise is that a patient’s complaints cannot be considered in isolation from the psychosocial context, and that disease cannot be divorced from a patient’s surroundings.

The nature of the human response to psychosocial stressors remains a subject of controversy. On the one hand, the theory of psychosomatic specificity assumes a one-to-one relationship between a specific psychosocial configuration and a disease, and views psychosomatic research as a quest for links between stressors and specific diseases. According to this theory, a specific set of circumstances (bereavement and loss of a job) in association with specific personality traits (submissive and manipulative) would result in a specific disease (bronchial asthma and peptic ulcer) [31]. For example, hostility has been linked with hypertension, and fear of separation with bronchial asthma. On the other hand, the association between life events and morbidity , and between socioeconomic state and mortality supports the theory of psychosomatic nonspecificity. According to this theory, any life event may increase the risk of any disease, irrespective of personality traits. The association between disease and psychosocial determinants is still in need of study.