It’s time to view hospitalization as a procedure

I previously suggested that transitioning from the traditional inpatient care model to the hospitalist model inadvertently motivated providers to hospitalize more patients, specifically borderline sick patients. Our example was a 74-year-old woman with pneumonia whose path to admission met less resistance with a hospitalist at the helm. The question I posed at the end was this: model aside, should we admit that patient or not? What’s best for her?

Incredibly, the answer is that nobody knows.

How can that be? We admit thousands of patients to hospitals every day.

To be clear, most of the time we know who needs admission. Patients are either sick enough to warrant hospitalization or they have no viable alternative. But if there’s a subset of admissions we might consider further it’s the borderline sick.

Let’s go back to the reason for our patient’s hospitalization. It’s not her pneumonia. It’s because we felt that admitting her would be the safest thing to do. That’s understandable. She’s elderly, frail, and lives alone. We should be wary of sending her home. If you’re a hospitalist like me, you’ve admitted patients under “the safest thing” mantra more times than you can count.

And yet, we all know that hospitalization isn’t always the safest thing, particularly for the frail and elderly. Deconditioning, delirium, and hospital-acquired infections are just a few of the potential downsides. Medication errors at times of transition – like hospital admission – are shockingly common. Some patients never make it out of the hospital. Some that do end up in nursing homes. Others suffer from something called “post-hospital” syndrome, a period of enhanced vulnerability that puts them at risk for further problems. This is not new information. We’ve known these things for a long time.

So knowing all of this, how do we factor the risks of hospitalization into the admission decision?

The answer is: inconsistently. Most of us try, but there isn’t a comprehensive method of quantifying the risks of hospitalization for individual patients. Instead, we’re often overly swayed by what might go wrong if we don’t admit the patient.

Let’s look at a risk assessment tool called the CURB-65. It’s used to help us decide if we should hospitalize patients with pneumonia, perfect for our case. We plug in our patient’s age, kidney function, and other data; then the calculator tells us our patient’s risk of dying from pneumonia over the next 30-days. Patients with a high risk of death warrant hospitalization and those with a low risk don’t. Sounds good, right?

Right … except that it’s only one side of the equation. What about her risk of harm from hospitalization? Will she become bedbound? Will she get C. Diff? Will she die of a medication error? The CURB-65 doesn’t tell us that. Neither does any other gizmo.

In the world of perioperative medicine, we have a more comprehensive tool to help us sort out the risk of procedures. It’s called the National Surgical Quality Improvement Program (NSQIP) Surgical Risk Calculator. Using this tool we enter the procedure type, along with a host of patient factors, including certain co-existing medical problems and an estimation of the patient’s ability to care for themselves. The tool then tells us the patient’s risk of complications associated with performing the procedure, such as blood clots, urinary tract infection, and discharge to a nursing home. It’s easy to use, and in conjunction with real providers who understand the real patients they’re treating, it can be quite helpful, especially when we’re on the fence about performing the procedure.

Sounds better, doesn’t it?

It’s time for us to think of hospitalization as a procedure.

Is another risk calculator the answer to biased decisions and overcrowded hospitals? We’d be foolish to put that much faith into a simple tool, but for borderline sick patients who could go either way, a hospitalization risk calculator could add a much-needed dose of objectivity.