When less is more for the struggling clinical reasoner

Introduction

Guiding students who struggle with clinical reasoning is one of the greatest challenges faced by clinician-educators [1]. These students’ difficulties with clinical reasoning typically become apparent when they are unable to tell a cohesive history or articulate and defend a prioritized differential diagnosis in their oral presentations.

A large body of scholarship addresses the acquisition of clinical reasoning skills; this work has led to important insights about the remediation of clinical reasoning [2, 3]. Formal remediation programs are effective [4], but they are labor intensive, require a longitudinal, institutional response, and are not well suited to the real-time clerkship environment where the unexpected struggling student is one of many learners on the teacher’s team [5].

When students and faculty work together for a short period of time (e.g. a few clinic sessions or 2 weeks on a ward team), teachers often struggle with the enormity of the problem and find it difficult to focus on a specific, tangible skill that can be refined from one clinical encounter to the next. In this essay we describe an approach that we are developing in our work with struggling reasoners that devotes less time to comprehensive feedback and more time to deliberate practice with the goal of improving the crux of clinical reasoning: the assessment.

Less is more

The student’s oral presentation is the standard vehicle for assessing clinical reasoning in clerkship settings and is a common starting point for formative feedback. In working with students on our internal medicine clerkship and pre-clerkship elective, we previously employed the common approach of providing feedback on all areas of the oral presentation – starting with the chief complaint and ending with the assessment and plan – in one sitting. We came to see that two aspects of our extended conversations hindered the development of presentation skills in the struggling student. First, as cognitive load theory suggests, this volume of feedback can be overwhelming for the learner [6]. Second, by concluding the discussion without having the student act on the feedback (and waiting until the next presentation to see if improvements were made), we forfeited a valuable opportunity to ensure that the student was able to incorporate the feedback.

Audétat outlined a taxonomy of five categories of reasoning difficulties that present in clinical teaching settings [7]. All deficiencies but one (data collection) are manifest in the assessment section of an oral presentation. Accordingly, we began limiting our feedback to the heart of the assessment: the problem representation and the prioritized differential diagnosis of the patient’s main problem.

Practice, practice, practice

On rounds, we listen closely to the student’s presentation but give minimal feedback in the moment. Later that day we deliver feedback limited to the assessment, and the student is asked to immediately present a new version of the problem representation and the prioritized differential diagnosis. We provide corrective feedback on this new version, and the student tries again (see Table 1). These feedback loops are short, with each iteration incrementally moving the student closer to an acceptable product. Real time deliberate practice is continued as many times as needed until the student presents a competent diagnostic argument, which replaces the memory of failure with the memory of success.

This approach embodies the core principle of deliberate practice: isolating a single high-yield skill (in this case, making a persuasive assessment) and using iterative feedback from a coach until the gap between current and desired performance closes [8]. In many ways, this exercise is similar to a sports drill or music rehearsal, where the teacher observes performance, provides corrections, and then says “let’s take it from the top.” Critically, in both instances, the instructor does not wait for another day to see if improvement materializes.

We were initially reluctant to ignore errors made in data collection, presentation structure, or management plans. However, we have found that once students attain a clear understanding of the components of a cogent diagnostic argument, they more naturally understand the key pieces of data needed to develop this argument. In our experience, data gathering (knowing what to ask) and data presentation (knowing what to highlight in the history, examination, or studies) will improve even before specific attention is paid to these domains.

There is no expectation that the student will have radically transformed their medical knowledge in the short time frame that students and attending physicians typically work together. Instead, the goal is for the student to develop a basic understanding of the essential components of a persuasive diagnostic argument. Success is marked by their ability to give a well-defended prioritized differential diagnosis for a single problem.

The set-up

In order to model, analyze, and teach clinical reasoning, clinician-educators must become facile with its basic lexicon: problem representation, illness scripts, and the prioritized differential diagnosis [9, 10] (Table 2). Using this framework, attendings can share their expectation that the major problem be clearly defined and analyzed with a differential diagnosis where competing diagnoses are compared, contrasted, and prioritized. With these well-defined expectations in place and a brief orientation to closely spaced iterative feedback loops, students can embrace the process that will bring them closer to the articulated standard.

While the initial conversation setting up the coaching plan does not entail substantial faculty time, the repeated one-on-one feedback cycles surrounding each presentation require dedicated blocks of time for deliberate practice. We have found that effective practice sessions require approximately 20–30 min each. We employ this remediation approach for a few students each academic year.

To complement in-person coaching sessions, attendings may consider electronic communications. One professor of English described how he helps his students develop writing skills by using email to solicit multiple early drafts and engage in repeated feedback cycles on key aspects of the paper before its formal submission for a final grade [11]. Clinician-educators can adapt this idea, with students securely emailing just the problem representation and prioritized differential diagnosis. Faculty can then reply using track changes and comment boxes and request re-writes until the student reaches competency. In our experience, students welcome this electronic feedback.

Conclusions

Clinical reasoning is challenging to remediate in real time in busy clinical settings. When a teacher’s time and a struggling student’s cognitive bandwidth are constrained, both parties should focus their efforts on an objective that is specific, relevant to the workplace, and achievable in one feedback session (Table 3). Our approach turns the teacher into a reasoning coach and gives the learner the opportunity to mark clear progress in a foundational skill.

While it may be difficult for the teacher to limit feedback to the assessment and let errors in the remainder of the presentation go uncorrected, we have seen clarity develop organically in other areas of the presentation once students fully grasp the diagnostic assessment. Conversely, when a sea of feedback is provided, struggling students can often drown.

We have not yet collected data or utilized formal assessment tools to evaluate our approach. Rather, we have drawn on observations of our own students, which are subject to confirmation bias. Formal study with a validated oral presentation scoring system is needed to confirm our early experiences. Also, we have focused on what one clinical teacher can do with one struggling student during the short time they have together. However, deliberate practice is most effective when it is sustained; we expect that improvements in clinical reasoning would be maximized when a series of clinician-educators use this approach longitudinally with a student.

We suspect that this singular focus on the assessment may refine the early organization of students’ illness scripts, which underpin all clinical reasoning skills. When students have a better understanding of what data are needed to make a coherent clinical argument, they are better equipped to select the right kinds of information to encode in their growing library of illness scripts. Instead of memorizing every aspect of a disease, students come to prioritize and encode features that they now realize are needed to make a convincing diagnostic case. Sparking this insight, more than anything else, may ultimately catalyze these students’ journey toward competence.