Elon Musk, the tech visionary behind Tesla and SpaceX, once observed that the single most important factor in professional growth is to have “a feedback loop, where you’re constantly thinking about what you’ve done and how you could be doing it better.”

I suspect most of us would agree. Yet in medicine, as in other fields, concrete measures of performance can be elusive. For instance, patient surveys help us deliver more respectful and satisfying care, but because the evaluations are subjective, they do little to improve the technical quality of our care.

Musk’s quote holds especially true for trainees, who are still fine-tuning their skills. For medical residents, the accreditation council stresses the need for a steady stream of “formative evaluation feedback” in daily practice. Meanwhile, attending physicians are busier than ever with patient care, paperwork and various other tasks. In one study at Boston Children’s and Boston Medical Center, only 6 percent of residents reported that they had received feedback on every rotation.

In light of this, we in the school of medicine have been looking at new ways to provide more personalized instruction and performance assessment—not just in the clinic, but also for the learners in our graduate biomedical programs and preclerkship medical courses. Historically, medical schools have taken a one-size-fits-all approach to classroom instruction. We hand out a syllabus at the beginning of the semester, and all students follow along at the same pace. We evaluate their grasp of the material retrospectively, using broad assessments that lump students into pass or fail categories.

Now we are exploring more precise, data-driven modes of instruction. With the leadership of Roy Ziegelstein, vice dean for education; associate deans Peter Espenshade, Jessica Bienstock and Nancy Hueppchen; and Harry Goldberg, assistant dean and director of academic computing, we are looking at ways to implement what we call precision education. (See Celebrating 125 Years of Medical Education at Johns Hopkins.)

Analogous to precision medicine, precision education allows us to optimize learning in a similar fashion. Moving forward, informatics will influence not only what we teach our students—competencies related to the use of big data and artificial intelligence—but also how we teach them. Precision education involves harnessing the power of information technology to provide learners with objective analysis of their progress and skills. In this system, we use learner data to shape the curriculum so that students can progress at their own pace, get individual remediation in an area that presents a challenge and even learn in their preferred modality. (After all, some students learn better through videos or interactive lectures than by reading textbooks.)

A recent experiment in our Department of Surgery illustrates the promise of this approach. Venous thromboembolism (VTE), which occurs when a blood clot forms in a deep vein, is a common, life-threatening complication following surgery. Depending on a patient’s risk factors, we take varying degrees of precaution, ordering anything from compression stockings to blood-thinning drugs. At Johns Hopkins, we began providing individualized feedback to surgery residents on how often they prescribed the appropriate VTE prophylaxis, based on the patient’s profile.

In this effort, led by health informatics expert Brandyn Lau and surgeon Elliott Haut, residents received an electronic scorecard detailing their performance and how it compared with that of other residents. Some low performers also got one-on-one coaching. In the wake of the intervention, there were no preventable complications among patients, and the number of patients getting appropriate interventions increased significantly.

Imagine if we could scale this “scorecard” model across all specialties for many common conditions and procedures. 

Our trainees have a genuine appetite for self-improvement and for observations that can help them grow. They know that ongoing feedback is as essential to learning as any seminar, textbook or lab rotation. With that, it feels appropriate to end this column with a request for comment. If you have thoughts on the column or suggestions for future topics, please email [email protected].