Practicing medicine like an adult

[This post was later discussed with host Bryan Boling in an episode of the Critical Care Practitioner podcast.]

Early training in critical care is dominated by the details: which fluid to select? What ventilator mode? How to resuscitate the septic patient?

These details need to be learned, but they aren’t the end of the road. Once comfortable with such practical daily concerns, the skilled provider needs to learn to step back and evaluate things on another, higher level. More than anything, this is what distinguishes the master from the journeyman. Some people learn this early, some late, and some hardly at all.

It’s the ability to think a little deeper, see a little further, and take responsibility for difficult decisions that distinguishes this level of practice. In short, it involves practicing medicine like an adult rather than a rambunctious teenager. Like much of adulthood, it’s not particularly easy and often not very fun. However, somebody needs to do it.

Here are a few of the tricks that characterize the transition to adulthood.

Make diagnoses

Don’t say: “This patient has hypotension and hypoxia.

Say: “This patient has septic shock from community-acquired pneumonia.

It’s easy to look at a sick patient and list their abnormalities: this number is low, that parameter is high, these metrics are wrong. But you don’t need any training to do that, just the ability to recognize when lab results are the wrong color, and a trained monkey (or the EMR itself) could do that job. It’s the hallmark of the student or early intern to evaluate a patient and then report only the bare facts about them.

Instead, make diagnoses, just like you learned to do in school. This can be difficult, because it requires some actual cognitive commitment. You can’t be wrong if you merely say that someone’s troponin is elevated; it’s simply a fact. You can be wrong by calling it a type II NSTEMI. But one of those problems has answers guided by physiology, guidelines, and evidence, while the other is just commentary. By giving things a name, you inform the plan of care in a way that you never will by merely describing derangements.

It’s okay to be uncertain and to express that. To say that a diagnosis “seems to be this” or is “consistent with that” are reasonable remarks when the picture is not clear. Having no opinion at all, however, makes you a putz.

Which leads us to…

Have opinions

Don’t say: “Plan: continue to monitor.

Say: “Plan: Trial of HFNC for one hour followed by intubation if hypoxemia is refractory; stop vancomycin; start stress-dose steroids.

The next step after making a diagnosis is creating a treatment plan. And here again, it’s easy to merely select the low-hanging fruit: we’ll do the obvious things, most of which are already being done, and then see if other obvious things present themselves.

That’s not medicine and certainly not critical care. Your diagnosis informs your treatment. You’re a clinician, not a historian; your role at the bedside is not to follow the path of least resistance, but to change the direction of care for the better. Make decisions (or suggestions) regarding what should happen, regardless of whether or not it is happening. It should be unusual to see a patient and have nothing to add.

Failure to follow either of these principles is how the phenomenon of the “useless consult” was born, wherein a specialty service fully evaluates a patient, tells you what you already know, recommends what you’re already doing, and then leaves. “But what do you think?” you cry in vain. “What’s wrong with him, and what should we do?” Without taking a stand on those questions, anyone seeing the patient has added very little.

Again, taking a stand means that you will sometimes be wrong. But rendering no diagnoses and developing no plan isn’t right either, unless you’ve become so nihilistic you don’t believe that anything works (and that’s an opinion too).

Consider details within their larger context

Don’t say: “The patient’s plateau pressures have risen and his blood pressure has dipped. We increased his pressors and adjusted the vent.

Say: “This patient has bullous lungs and has been on positive pressure ventilation; we need to urgently evaluate for a pneumothorax.

Or even better, say: “This patient is end stage. Let’s talk to the family; this latest development may be the sign that it’s time to start redirecting care towards comfort.

Developing a plan that’s more intelligent than a game of whack-a-mole requires thinking that goes beyond what’s in front of your face. Abnormalities exist on their most obvious level, true, but they also exist within the larger context of the patient’s pathology, hospital course, and personal history. Learn to address the details without losing perspective on them; otherwise you can fix every leaf without realizing you’ve become lost in the woods.

If it needs to be done, do it immediately and aggressively

Don’t say: “He’s septic with a MAP of 45, so we’re giving a bit of fluid; we’ll see how he responds.

Say: “He’s unstable; we’ll immediately bolus fluids, start pressors, and place lines, all without delay.

Almost every indicated intervention should be done as soon as possible.

For critical patients, this can be life-saving; more patients have died from the right treatment applied too late than from receiving the wrong treatment.

For more stable patients, the sooner things are done, the faster they make progress. It may not be the end of the world if it takes an extra hour to give metoprolol or an extra day to start a diet, but even at their most benign, delays beget more delays downstream, leading to prolonged hospital courses and a slow, flabby healthcare system. And at their worst, treacherous complications can develop while you were dragging your feet.

If it doesn’t need to be done, don’t do it or stop it immediately

Don’t say: “The cultures came back sensitive to nafcillin, but let’s just finish off a course of vancomycin and Zosyn anyway.

Say: “We treat things for a reason, not from momentum. We’ll stop these antibiotics and use nafcillin for the remainder of the course.

You will meet some providers who prefer to overtreat and many others who prefer to undertreat. Neither group will self-identify this way, but that’s how their personalities unpack. In general, they’d rather be wrong by doing either too much or too little.

The truth is that both of these sins are equally evil. If something needs doing, it should be done without delay, and if it doesn’t need doing, it should be avoided or discarded even quicker. Getting it wrong from either bias leads to morbidity and mortality.

The problem, of course, is that you often don’t know which error you’re committing until after the fact. A fine example: the patient with borderline respiratory failure. Perhaps you can put them on NIPPV and avoid intubation. Or perhaps you’re simply causing them harm by delaying the inevitable. Deciding which is the case without the benefit of hindsight requires experience, acumen, and a good bit of luck.

Predict problems before they arise

Don’t say: “She seems to be doing well, so let’s downgrade her to the floor.

Say: “She’s improving, but her mental status is still borderline and she’s handling her secretions poorly. Let’s keep her another day or two for aggressive pulmonary toileting so she doesn’t bounce back with pneumonia.

It’s easy to see what’s in front of you. Guessing what will happen next is another matter.

Only with experience (mostly born of mistakes) can you start to recognize pitfalls before they occur. The reality of medicine is that there are countless right ways to handle most situations, so it doesn’t require a genius to come up with a plan. Instead, genius lies in foreseeing a hidden danger in one of those seemingly correct plans, and steering toward a safer option instead.

When there’s no right decision, make one anyway

Don’t say: “He still has a leukocytosis, but he’s clinically stable and hasn’t had a positive culture in a week. He’s on antibiotics but I don’t know how to narrow them or when to stop. In fact, I’m not sure if there’s an infection at all. I don’t know what to do.

Say: “I don’t know either, but we can’t do nothing forever. Let’s stop antibiotics today; if he gets worse, we’ll have an answer. If not, we’ll get some clean, sterile cultures off antibiotics and go from there. If they’re positive we’ll consider a tagged white blood cell scan.

ICU patients are often complex, with many challenging, competing, and confounding problems. A clear diagnosis is sometimes impossible, and treating one problem often worsens another. Many times, there are no right answers.

Nevertheless, a decision needs to be made, and the ability to make a hard decision — even a wrong one — instead of waffling forever is what distinguishes the clinician from the technician.

Don’t ask what — ask when, where, and how

Don’t say: “He needs to be intubated.

Say: “He needs to be intubated, but he’s high risk. Let’s further resuscitate him first, and then bring him to the OR where we can have surgery standing by in case his airway is difficult.

Deciding what to do is one thing, but the devil is in the details. After you have a broad scheme in mind, a second level of decisions calls to you, many of them rife with danger: How, when, and where should it be done?

This is the flip side of the “do what needs to be done immediately” rule. If the only reason to delay is because you’re being hesitant, stop being afraid and strike aggressively. However, rushing to act without necessary preparation is a mistake as well. Thoughtfully assembling the logistics of your maneuver can often make as much of a difference as knowing the general direction.

Take responsibility

Don’t say: “He has heart failure, so we consulted Cardiology. He has an infection, so we consulted Infectious Disease. He has toe pain, so we consulted Rheumatology. And he’s been upset about his care, so we asked the patient representative to see him.

Say: “We’re the primary team, and we’ll seek help when we need it, but we are responsible for the patient’s problems. I’ll go talk to him and address his concerns, and then we’ll decide how to treat him.

A dozen services and far more providers will likely see your patient. Most of them don’t consider it their job to make decisions — and they’re right. Of the rest, many feeble-minded milquetoasts will still try to avoid the burden.

Responsibility doesn’t roll downhill, or at least it shouldn’t, because in most cases nobody’s at the bottom to catch it (if there is, it’s often the person least equipped to do so). Responsibility rises to the top. Delegation is appropriate, and seeking assistance from others is vital, but it all begins with someone willing to step up and take the lead.

When you show up at work, as your default approach, when you perceive a decision to be made — make it.

When you discover a puzzle, solve it.

When you encounter a problem, address it head on.

That’s how you steer the trajectory of a challenging patient upward and forward, rather than turning in endless circles.


So our list of the key traits of a critical care clinician who is able to lead, inspire, and cut through the thick mess of modern medicine?

  1. Make diagnoses
  2. Have opinions
  3. Consider details within their larger context
  4. If it needs to be done, do it immediately and aggressively
  5. If it doesn’t need to be done, don’t do it or stop it immediately
  6. Predict problems before they arise
  7. When there’s no right decision, make one anyway
  8. Don’t ask what — ask when, where, and how
  9. Take responsibility