Crozet Annals of Medicine: Gestalt—In which I go head to head with Dr. Google


Spring is here! I don’t need to go outside to know it, I recognize all the signs in my interns (first year out of medical school trainees). There is a new spring in their step after a long dark, depressed winter. They now know they are going to survive this formative year and come June will become upper-level residents.

This has been a hard year for them. They came to the hospital in June, rested, optimistic, eager to meet each other and us, the attending physicians. Their first month was filled with orientation lectures, retreats, team-building exercises and introductory shifts in the Emergency Department. For that first month they were welcomed, supported and they forged new friendships with each other and the nurses.
After that, they were thrown to the wolves. Scattered to all corners of the hospital on various solitary off-service rotations, each with its own rules, culture and expectations. Just when they began to get the hang of each rotation they were switched to another new service, over and over again. Rotations like Cardiology, Pediatric Surgery, Trauma ICU, Medical ICU, Anesthesia, General Medicine and others. Lots of night call, lots of new material and decision-making, lots of mistakes, some serious. No one was going to the gym anymore, social outings became rare and were cut short by the prospect of pre-dawn rounds.
And so the interns were looking pretty ragged through December and January. Chronically tired with blunted emotions and borderline depression.

But as the days lengthen in February and March, the interns are blossoming like the daffodils. They have learned enough to survive. They know which residents to seek out for help with hard cases. They know which nurses will give them enough space to catch their breath before requesting more decisions and orders.

They know which attendings to question and which attendings are better left unquestioned. Above all this they have learned a vast amount of medicine and from that primarily flows their newfound equipoise. They are competent, they are confident, they can take care of sick people.

The second- and third-year upper-level residents have made similar progress and my supervisory role has become largely ceremonial in the simpler cases. This was brought home to me on a recent shift when my second-year resident approached me. He proposed that he would manage all of his patients that day without any input from me. If he had questions he would seek me out, but otherwise he wanted to see if he could do it.

I was charmed with this arrangement. One less person to talk to. I could see his orders on the computer and follow the results. And of course I saw all his patients surreptitiously while he was busy doing other things.This went well for a while.

I monitored one case from behind the curtain as the 50-year patient told his story in Spanish. I caught several words, mareado (dizzy), una semana, (one week). Pretty simple case. Benign vertigo. Dolor en pecho. What’s that, chest pain? Not so simple. I would need to take a closer look, which I did after the resident finished up. I satisfied myself that the chest pain was also benign. So I was surprised when the resident approached me for guidance on this case.

“I am not sure what to do with this patient.”

I sat silently.

“He looks pretty benign. The chest pain was an afterthought. I could let him go home.”

I sat expressionless.

“I could admit him to Cardiology.”

I inclined my head noncommittally.

“His HEART Score is 5,” He added. “I Googled it.”

The HEART Score is a clinical decision rule that helps doctors gauge the risk of death of any given patient with chest pain. Each letter stands for a risk factor: History, ECG, Age, Risk factors like smoking, Troponin (a blood test for heart attack).

Each of the five components is awarded zero, one, or two  points depending on the degree of abnormality. Five is a high score, corresponding to a risk of 12-16 percent for a bad outcome.

“That is a high score,” I allowed.

“I could admit him to Cardiology.”

“You could.”

“Or I could send him home. His troponin is normal.”

“True,” I replied giving him little guidance. Hey, he was the one who wanted to make all the decisions.

“His heart score is 5. I guess I answered my own question. I’ll admit him to cards.”

“Good luck.”

I knew Cardiology would shoot down the admission. This man was not having cardiac chest pain, his chest wall muscles were sore from repetitive lifting.

Which brings me to an observation about a generational shift in medical decision-making. My generation read lots of papers, saw lots of patients and made the best decisions we could based on what we now call gestalt. We also probably made a lot of mistakes.

Gestalt is a tricky concept. Merriam-Webster defines it as something that is made of many parts and yet is somehow more than or different from the combination of its parts.

It is sort of the opposite of the clinical decision rule described above where a thing is just the sum of its parts. My gestalt of the patient was that putting those numbers together did not truly account for his diagnosis or prognosis.

The latest generation of physicians, the generation I am currently training, relies not as much on gestalt but much more heavily on decision rules, algorithms and guidelines to diagnose patients. They have generated hundreds if not thousands of decision aids using research and statistics. They take comfort in numbers and data and guidelines (and working in groups). Hundreds of decision aids can be found with a google search, which is where this resident found the HEART score

My generation of physicians, used to making decisions on their own data, deride this new paradigm as “cookbook medicine” and claim medicine is too complicated to be reduced to a recipe to be applied to each patient. So who is right?

Well, it is too soon to say. Of the last 171 decision aids published in the Annals of Emergency Medicine, our major specialty journal, only 21 were compared directly to physician judgment or gestalt. In the remaining 150, no comparison was made or could be identified in the external literature.
In the 21 comparisons, the decision aid was clearly superior to clinician judgment in only two.Of the two comparisons favoring the decision aid, one was too unwieldy for widespread use and the other performed well in widespread clinical use.

Conversely, six decision aids clearly underperformed when compared to clinician judgment, and the remainder were a wash. Examples of popular decision instruments either inferior to or no different than clinician judgment included scores for appendicitis, scores for pediatric head injury, rules for pulmonary embolism (blood clots in the lung), and rules to evaluate patients who have fainted to determine benign or serious causes.

Back to the case at hand. In the end, the resident, intrigued by my equanimity about the patient’s outcome in the face of this alarming HEART score, went back and spent a lot of time with him at the bedside puzzling out his story. Ignoring the HEART score, he eventually reached the same conclusion I had, which was actually also his initial gut reaction to the patient before he applied the HEART score. And thus, is gestalt developed.

We sent the patient home and he did fine. Take that Google!


Please enter your comment!
Please enter your name here