Research into Clinical Reasoning
There are three types of research into clinical reasoning: clinical judgment, decision theory, and process tracing. Clinical judgment research attempts to identify the criteria used by clinicians in making decisions. Decision theory explores the flaws and biases that deflect accurate clinical judgment. Process tracing elucidates the progressive steps of naturalistic reasoning. The first two types are statistical and prescriptive, the third is normative.
Clinical Judgment & Decision Theory
According to the “lens” model of clinical judgment, each patient exhibits a set of symptoms, signs, or criteria that the clinician weighs and combines to reach a decision (eg, whether the patient is at risk for suicide or whether he or she should be hospitalized). Researchers attempt to “capture the policy” of the expert decision maker in order to construct mathematical models that replicate clinical judgment.
Given the fuzzy nature of clinical data, medical decisions have to be probabilistic. Accordingly, decision theorists base their research on Bayes’ theorem. This theorem states that P (D/F) (the probability that a diagnosis is present given a clinical finding) is a function of P (F/D) (the probability that a finding will be associated with that diagnosis), P (D) (the probability of the diagnosis in that population), and P (F) (the probability of the finding in that population). Thus
The intuitive clinician gauges P (F/D) from theoretical knowledge and experience. However, P (F/D) must be combined with the local base-rates for both the disease and the finding, base-rates that are often either unknown or ignored by the clinician. Unfamiliarity with Bayes’ theorem and other biases can introduce several errors into clinical reasoning (Table 9-1).
Decision theory has been applied most often to convergent problems, for example, whether or not to hospitalize a patient. The better choice should be the one that has the highest expected utility. Expected utility is the product of the probability of an outcome and its subjective utility (eg, how highly the patient or physician values that outcome). Consider the following case:
A 14-year-old girl is evaluated in a pediatric hospital ward after she has taken an overdose of 50 acetaminophen tablets. She has made one previous suicide attempt. The recent suicide attempt occurred after she was rejected by a boyfriend. The patient is emotionally labile and clinically depressed. She is hostile to her mother and refuses to agree to a “no-suicide contract.” The consequences of hospitalizing the patient versus not hospitalizing her but referring her for outpatient treatment can be represented in Figure 9-1. The clinician is expected to opt for the choice that leads to the greater expected utility.
Clinical decisions are often more complex than can be represented by a simple decision tree of the type shown in Figure 9-1. Multiple branching yes-no decision trees have been constructed to aid diagnostic decision making (eg, Appendix A in Diagnostic and Statistical Manual of Mental Disorders, 4th edition [DSM-IV]) and to encode expert treatment decisions. Utility and probability are also important considerations when cost-benefit analyses are undertaken, for example, concerning the desirability of mass screening procedures.
Research has shown that clinicians do not always follow the expected utility model. For example, decision making may be biased by the readiness with which a particular outcome can be remembered, particularly if it has had an emotional impact on the clinician (eg, a patient’s recent suicide). Decision making is also affected by the way problems are presented. For example, a treatment that saves 800 lives out of 1000 may be preferred to one that sacrifices 200 out of 1000 (although the two situations are equivalent in risk). Furthermore, the probability of an outcome can sometimes affect the subjective estimation of its utility, whereas theoretically the two should be independent.
Comparing experts and novices, researchers have traced the steps of naturalistic reasoning in areas such as chess, physics, mathematics, neurology, family practice, internal medicine, radiology, and psychiatry. A chess expert, for example, has built up from experience the memory of perhaps 50,000 chessboard patterns. Each pattern is associated with possible moves. Rapid pattern matching dynamically linked to good choice of next move explains the capacity of the chess expert to play and defeat many novices simultaneously. Thus pattern recognition is linked to strategic option choice. The expertise of the diagnostician is similar: The recognition of an incomplete clinical pattern that matches, in part, the memory of a diagnostic syndrome is linked to tactical choices for eliciting, evaluating, and integrating further evidence to solve the diagnostic puzzle.
Clinical reasoning is a species of “bounded rationality” in which the clinician converts an open problem (ie, a problem with no clear endpoint) into a series of closed problems (each with a hypothesized endpoint). In other words, the open problem is reframed as an array of closed problems that organize the search for evidence. Furthermore, diagnosis is not a static endpoint but rather a dynamic way station on the road to treatment. The decision pathway toward diagnosis is shown in Table 9-2 and is described in more detail in the sections that follow.
Eliciting & Perceiving Salient Cues
Even before the patient is seen, the clinician may have gathered cues, for example, from the referring agent. As the patient enters the office, before the interview begins, the clinician scans the patient’s eyes, face, skin, clothes, gait, coordination, posture, and voice in order to perceive salient cues (eg,"pale, elderly, frail, shabbily dressed, worried-looking woman using a walking stick, favoring her left leg"). The clinician must be alert to pertinent cues, distinguishing them from the immense amount of noise in the perceptual field. Initially, the net is cast widely so as to maximize the chance of correctly recognizing salient cues, perhaps at the expense of perceiving data that turn out to be irrelevant. As the diagnostic process proceeds, however, the gathering of evidence becomes more focused.
The patient sits down and the interview begins. The clinician’s demeanor, receptiveness, and empathic communication encourage the patient to tell her story. More cues are elicited from the patient’s spontaneous account of herself.
Evaluating Cues & Making Inferences
Out of the enormous amount of noise, the experienced clinician knows what to look for. Freckles, for example, are less likely to be pertinent than are blue lips (although, in certain circumstances, freckles could be relevant). Blue lips, however, must be evaluated before they are regarded as significant (ie, abnormal). Has the patient been eating berries, or is the blueness circulatory in origin? If the blueness is circulatory in origin—that is, cyanotic (a clinical inference)—is it central or peripheral in origin?
If a patient says that people are talking about him, the clinician must decide whether this complaint is based on reality, whether it is an exaggeration of reality, or whether it is based on a false conviction (ie, a delusion). The experienced clinician makes tentative inferences, at first, which he or she is prepared to revise if subsequent information does not bear them out.
Assembling Cues & Inferences As a Clinical Pattern
Soon after the clinical encounter has begun, the clinician has begun to form cues and inferences into tentative patterns that form the gist of clinical reasoning, for example: (1) potentially lethal suicide attempt; (2) angry, depressed, disheveled adolescent girl lying in a hospital bed; (3) uncooperative and dismissive toward the examiner; (4) said to have made one previous suicide attempt (the time and lethality of which are uncertain at this point).
The efficiency and accuracy of pattern discernment distinguishes the expert from the novice, as does the efficiency with which the expert matches clinical patterns, incomplete though they may be, against his or her memory of diagnostic syndromes.
Generating Categorical & Dynamic Hypotheses
The pattern prompts hypotheses, and hypotheses organize the subsequent clinical inquiry. The capacity of working memory limits the array of hypotheses to between four and six. The array of hypotheses may be linear or hierarchical (see Figure 9-2 and 9-3 for examples). Hypothetical reasoning prevents premature closure on one diagnosis and spares short-term memory by dividing the information derived from cues and inferences into strategic units, from each of which a systematic search for evidence can be planned. Hypotheses are open to revision in the light of new information derived from the inquiry process.
The diagnostic hypotheses generated are usually a mixture of categorical and dynamic types. Categorical hypotheses are expressed in the familiar terms of DSM-IV or a similar taxonomy. Dynamic hypotheses (eg,"vulnerability to rejection related to abandonment by father") operate in parallel with categorical hypotheses and are not exclusive of them.
Designing an Inquiry Plan & Searching for Evidence
The inquiry plan (ie, history; mental status examination; physical examination; laboratory testing; special investigations; and information from collateral sources, past records, and consultations) has two aspects: standard and discretionary. Clinicians standardize their data collection (eg, past medical history, mental status examination), casting the net widely to gather important cues and evidence in people from particular age, ethnic, or social groups. For example, questions about substance abuse, physical and sexual abuse, suicidal ideation, and antisocial behavior are virtually obligatory for adolescent patients. Similarly, certain urine and blood chemistry and hematologic tests may be part of a standard screen for hospitalized patients. For the most part, however, the inquiry plan is discretionary. It is designed to elicit information relevant to the array of diagnostic hypotheses.
Revising, Deleting, or Accepting Hypotheses
Modifications of the standard history, mental status examination, and physical examination are determined by the diagnostic hypotheses. For example, if lead poisoning is hypothesized (eg, as a cause of childhood hyperactivity), the clinician will inquire about the child’s physical environment (eg, exposure to old paint, batteries, or tetraethyl lead), examine the child’s teeth and gums, and test the child’s blood and urine. The inquiry plan yields data that complete the clinical pattern from which the preliminary hypotheses were derived and allows the clinician to refine hypotheses or to disconfirm them.
Reaching a Diagnostic Conclusion
When enough evidence has been gathered, the clinician weighs and summarizes the evidence supporting or refuting hypotheses that have not already been rejected. Sometimes, a single diagnosis is insufficient, and two or more diagnoses are required to account for a heterogeneous pattern of clinical features. Next, the clinician expands the diagnosis, combining dynamic and categorical diagnoses in a diagnostic formulation.