High and low standard deviations around the mean. Source: National Library of Medicine.
Back when he was program director of our ID fellowship, Dr. David Hooper would give the applicants a description of our program. One of the key parts was his estimating the workload — in particular, the number of new consults per day.
“We average three to four consults a day,” he said. “But there’s a high standard deviation around the mean.”
That last part he said humorously, with a smile and a shrug. It was a wonky joke, but everyone got it since it’s well known that consult volume is unpredictable — nothing different about our program compared to any other. But this variance is a critically important part of consultative medicine, one I’d argue is one of the key drivers of physician stress, especially for trainees.
What do I mean? Join me in this thought exercise. You’re an ID fellow with a weekend off, and it’s Sunday night. You’ll be picking up a new service on Monday — first day of a new rotation! — and will be responsible for learning the details of the cases on your team and catching up on weekend events.
Not only that, you’ll also be seeing the new consults that get called in that day. You know Monday will be busy, but how busy?
Let’s take two scenarios for the number of new consults — which would you choose if given the option?
- A day when you will get four new consults — no more, no less. Once you’ve done the fourth, you’re done with new consults for the day.
- A day where your consult volume that day is uncertain. You know that the average number of consults/day in your program is between three and four; however, light days (one or two consults) are balanced out with very busy days of five, six or, rarely, even seven new consults.
I suspect most of us would choose the first option, even though the “average” outcome of choice #2 is fewer consults.
This says a lot about psychology, risk perception, and our decision-making strategies. In studies of decision science, participants often choose the “sure thing” over a potentially more valuable but uncertain outcome — a phenomenon often referred to as “loss aversion.”
An important point is that the relationship between consult volume and stress does not increase linearly — it’s more like on a log scale, which means that going from five to six consults is much more difficult than going from three to four, even though both just add a single new case. And what this additionally means is that getting six consults is much more than twice as stressful as getting three.
(How about those math skills. Impressive, eh?)
Finally, there is something inherently stressful about living through the amorphous blob of work potentially coming your way in choice #2. The day could start out relatively peacefully, with just a single consult, making you cautiously optimistic but still vulnerable. But then, an hour or so after lunch, the chief resident in orthopedics could page you and say they’ve just accepted in transfer two patients with infected prosthetic joints — both of whom will need your attention when they arrive (whenever that will be).
That hypothetical day still didn’t yield more consults than in choice #1, but the unpredictability of the way they came in made the day seem so much more tense.
(Note: This discussion must seem foreign to ID doctors in private practice, where consult volume directly links to personal revenue. But try to imagine yourself back in the days of your ID fellowship, however, and you’ll get what I mean!)
I thought of this challenge recently since we recently had quite the week when it comes to consult-volume variance. Afterwards, I sent a note commenting about this to Dr. Daniel Solomon, our fellowship’s current Associate Program director. His response:
I think the hardest thing about being on service (and in particular first-year fellowship when they are on the front lines holding the pager) is not the cumulative volume of work. It’s the unpredictability of each day. It is hard to make reliable plans with friends and family when the variance is so high.
I couldn’t agree more.
Solutions? One thing we proposed was to unload some of the simpler cases to an eConsult system, where inpatient medical and surgical teams received clinician-to-clinician advice from us after a discussion, record review, and our writing a brief note in the chart. The issue? As I wrote previously — no one has figured out how to pay for these things. Proposal rejected.
Some say that instituting a “cap” on new consults solves this variance-in-volume problem, and there’s definitely some truth to that. Such caps limit the burden of a high consult day on the ID fellows, much as a cap on admissions does the same for interns and residents. Plus, that uncertainty factor is greatly reduced.
This solution isn’t straightforward to implement, however. First, who sets the right number? ID training programs have a wide range of expected new cases per ID fellow per day. I’m very much aware that our daily average of three to four per day isn’t the same as other programs, some of which have considerably higher volume.
Also, should the cap be the same regardless of the number of patients you’re already following, or the complexity of the service? Should it be the same for general ID consults as it is for transplant and oncology services, which have an average complexity per case that’s much higher? And how do we account for variable team structures? Some regularly have rotating medical residents and/or students on board to help defray some of the work, while others rarely have these learners.
One other issue with a cap relates to the inherent value of clinical volume for volume’s sake. There’s a cliché in clinical medicine that goes, “The more you see, the more you see.” Since many ID programs (ours, for example) have only 1 year of intense inpatient clinical training, why not make the most of it, provided the volume isn’t too brutal? We all know that there’s no better way to learn about a clinical entity than to care for a patient who has it — the first-year ID fellow who sees CNS nocardiosis, or falciparum malaria during pregnancy, or disseminated histoplasmosis will never to forget those distinctive but relatively rare diseases, to choose just three that recently popped up on our inpatient service.
Although it doesn’t seem so at the time — an understatement — even doing a consult on “routine” cases brings value. Seeing many examples of Staph aureus bacteremia, or infected abdominal collections, or osteomyelitis under sacral pressure ulcers cumulatively helps develop an approach to these common entities, and to appreciate the wide variability in clinical presentation and management.
So far this discussion about the cap looks at it from the fellow perspective. It doesn’t address the fundamental cause, which is that the consulting services need our help caring for their patients, and it’s our mission to help them. That the volume of these requests is unpredictable isn’t their fault. Hence, once a trainee is capped, this work then must get done by someone else, right? Is it the on-service attending, who is still responsible for staffing the fellow-seen cases on this already busy day? Some other faculty member waiting in the wings, eager to do a late afternoon bunch of consults? Who has those faculty?
In summary, there are pros and cons to putting a cap on new consults for ID fellows. I’d be interested to hear — do you have a cap on new consults in your fellowship program? If so, what is it, and how did you decide on a number? Who’s responsible for doing the work over the cap? Share your thoughts in the comments section.
And enjoy this quite remarkable video, which somehow escaped my notice when it first appeared. Glad they remembered to press the record button!