Heuristics should not be identified as a mere irrational source of errors and, thus, be avoided in decision making. Instead, it would be beneficial to recognize cognitive biases and windows of opportunity not constrained by normative logic. The most promising way to train future doctors in clinical decision making is to improve the use of heuristics by increasing insight, that is, adjusting the capacity to learn from errors associated with practice, which are sometimes unavoidable. This can bolster competencies in differential diagnoses.

Madrid, May 30, 2019. Many studies suggest that most decisions are based on heuristics, or in other words, on shortcuts or mental rules that would violate the foundations of the theory of logic and probability, which should be avoided as a source of error. In the field of medical education, few studies point out to heuristics as one of the essential pillars of decision making, but some advocate its use to design and improve pedagogies aimed at clinical decision making.

This is the case of Markus Feufel, professor at the Department of Psychology and Ergonomics of the Technical University of Berlin, and John Flach, professor emeritus of Psychology at Wright State University in Ohio (United States), who recently published a study in Medical Education betting on the learning of heuristics for "fewer errors that can be potentially fatal for the patient."

These experts argue that the classical theory tries to avoid heuristics, so the students focus on finding a correct diagnosis and eliminating errors. But this perspective, suitable for areas that deal with well-defined problems, does not always work in the clinical setting. Moreover, in clinical practice, problems are vaguer and may not have an optimal solution, while errors are inevitable.

The idea behind the study is to address the problem of medical errors with training that takes advantage of them, helping students to modify their clinical heuristics to be able to differentiate between plausible explanations for a given set of symptoms (differential diagnoses) and reduce opportunities for potentially fatal or costly errors.

Zero tolerance with error leads to a culture of guilt in which errors are seen as a symptom of human weakness, leading to their cover-up. The authors warn that "attempts to cover up mistakes deprive (professionals and health care systems) from the information necessary to continue learning, adaptation, and improvement (of clinical practices, learning from their own mistakes)," noting that the key would be to "tip decision making toward intelligent errors instead of awkward errors." In short, to improve heuristics, errors should be an integral part of both continuing medical education and professional practice.

After an extensive review of classical works blaming heuristics for errors by deviating from logical rationality, Feufel and Flach support a more productive approach to medical education, with research in clinical contexts, in which experts apply heuristics as a strategy of lucid adaptation to the complexities of a specific context.

Therefore, they recommend successful teaching and learning of the use of heuristics instead of replacing them with logic. "Training should increase the perspicacity of the student with regard to heuristic decision making; that is, tune the (recognition) processes that underlie the adaptive selection of heuristics and the management of errors."

According to Lola Lopes, professor of Experimental Psychology at the University of Iowa, poorly defined problems are associated with a differential diagnosis; that is, to observe the important (the actual condition of the patient) of a noisy background (the myriad of possible conditions that the patient may have). In short, when the decision-making process ends, the resulting decision is adjusted to the appropriate standards in each situation and can be implemented with the available resources.

Favoring the physician's insight

For Feufel and Flach the main element would be to become aware that zero error is an unattainable objective, but two mechanisms may mitigate the probability of specific errors and their consequences: the ability to discriminate the patient's condition and to establish the decision criteria that reflect the most desirable balance between alarms and false alarms of errors.
         
According to them, "A physician’s skill in making decisions is more a function of perspicacity (i.e., discrimination processes that are well-tuned to the demands of the problem to be solved) than rationality (i.e., conformity with logical norms)." However, the perfect decision is unreal due to three reasons: the symptoms of different diseases tend to overlap, the inevitability of error, and the complexity to reach the best balance between alarms and false alarms.

With these defined criteria, medical training should improve the discrimination between a set of alternative diagnoses for a symptom. According to these researchers, the more physiopathology physicians know, the better they will discriminate between diseases and select more appropriate treatments. However, medical education is commonly organized around main effects associated with a specific discipline (anatomy, physiology ...), organ (heart, liver ...), or disease (causes, symptoms, and treatments).

Therefore, training should be more focused on interactive relationships between symptoms and diseases in order to reach a differential diagnosis with the support of multiprofessional teams, and integrating technical and nontechnical skills, such as communication and technologies, "may be as important to the quality of clinical decision making as biomedical knowledge and clinical skills."

Three clinical teaching formats

Feufel and Flach analyze several pedagogical approaches (which do not decrease the probability of errors, but reduce their impact) for training health care professionals in decision making to "provide continuous opportunities for differential diagnoses and teamwork (…), at the same time maximizing learning from errors and minimizing potential consequences."

The first is the teaching of students under supervision working with real patients, taking a clinical history, performing a physical examination, diagnosing and prescribing a treatment. Individual events will eventually add to the holistic decision-making capacity, facilitating students' early exposure to the biomedical and sociomaterial complexity of clinical problems. Moreover, according to Feufel and Flach, this method could help increase insight in "noisy" clinical contexts.

A second approach is to present to students narrative cases (vignettes) to solve problems. These should be constructed by increasing the students' awareness in two aspects: the frequency with which they are present in the clinical setting, and the strengths and limitations of the strategies available to treat those cases. In this way, students can appreciate the need to make "intelligent" and not "clumsy" mistakes.

The third learning is based on simulation. The growing loyalty and availability of patient simulators also promote a holistic education focused on problems that do not expose people to risk. In addition to offering hands-on teaching of perceptual motor skills, simulators are used to create realistic scenarios.

Feufel and Flach conclude that simulation should be designed to accommodate errors that illustrate the tension between standard operating procedures and their exceptions so that students get the most benefit from such errors. Also, more importantly, the researchers say that the simulation scenarios are particularly suitable to train students in nontechnical skills related to the workgroup and to prioritize clinical tasks in a framework of limited resources.

Reference
Feufel MA, Flach JM. Medical Education should teach heuristics rather than train them away. Medical Education 2019; 53 (4): 334-344. doi.org/10.1111/medu.13789.