Fiddling As HM Burns


It’s been a hectic week, as the Annals of Internal Medicine paper regarding hospitalist outcomes was published.1 I cannot escape the fallout of the paper showing that the hospitalist model is associated with increased costs of care. The Internet, the phone, my email, the radio, the hallways all are abuzz with excitement about the implications of this paper. Everyone, it seems, has an opinion. The viewpoints range from “the article is methodologically flawed” to “yeah, but that data is old and things are different now,” to “I knew the model was bunk” to “it’s time to bring back the traditional model of care.”

Moreover, nobody is afraid to share.

Wherever you stand on this continuum, it isn’t hard to find a supporting opinion. NPR covered it, newspapers reported it, and bloggers blogged it. Thousands of words were typed, printed, tweeted, spoken. However, one word seemed conspicuously absent. That word? Thanks.

This study lifted the veil on what happens after discharge—and the findings have rocked the foundation of the HM field.

The Study

The study authors examined more than 58,000 admissions at 454 hospitals and compared the impact of hospitalist versus PCP care on in-hospital and post-discharge outcomes. Predictably, hospital length of stay (LOS) was shorter (0.64 days) and less costly ($282) with the hospitalist model. This has been shown, albeit generally with more robust outcomes, in nearly all-previous studies. Importantly, however, this study lifted the veil on what happens after discharge—and the findings have rocked the foundation of the HM field.

Hospitalist patients were less likely to follow up with their PCP, more likely to go to a skilled nursing facility, more likely to return to the ED, and had higher rates of 30-day readmission. All of this post-discharge care came with a price tag—$332 more than the PCP model—making the bundled in- and outpatient costs of care about $50 more per patient in the hospitalist model.

And this is where the controversy—and the words—begins. Connecting the earlier discharge, the added SNF utilization, and the higher readmission rate could only mean only one thing to those that favor the traditional model—a cost shift. Clearly hospitalists, motivated by saving money, are shifting the financial costs just beyond the hospital confines, discharging patients so early that they require nursing home and, ultimately, more ED visits and hospital care.

On the other side of the ledger, HM supporters have pointed out that the patients in the two arms were not the same. The HM patients were more likely to be admitted from a nursing home, more comorbid, poorer, and more likely to be admitted on a weekend—all valid points, which are hard to control for in an observational study. They argue that patients in an SNF are, of course, less likely to see their PCP than patients at home and, therefore, more likely to be sent to an ED (and admitted) when issues arise. Perhaps, the argument goes, in this scenario the system is actually working. Without indicators of quality of life and functional status, it’s hard to know that HM patients didn’t do better. Sure, there were more readmissions and it cost more, but perhaps that’s the cost of better, longer-term outcomes.

My take: Let’s move beyond debating the study merits and its implications. HM is here to stay. No matter how much we conjure Osler, we aren’t going back to the traditional model. In the debate we miss the point. Rome is afire; it’s time to stop fiddling.

So, let’s put our preconceived biases, the potential methodological flaws, the conspiratorial overtones, the vitriolic banter, and the fruitless debates behind us. This study was generally well done. It focuses on a (perhaps, the) crucial issue for HM. And its findings are plausible. For that I say “thanks”: for exposing this issue so we can tackle it head on by moving in at least three distinct directions—quality, training, and retraining.

Now let’s put down the fiddle and pick up the extinguisher.

$50: The Price of Quality?

First, I hope this study finally pushes our field beyond the cost discussion. We simply can and should not be a field that is about saving money. Yes, it is great to save money. But more important, we have to enhance the quality and safety of hospital (and post-hospital) care.

The piece missing from the Annals paper is any significant look at quality metrics, beyond perhaps readmission. It is possible, and maybe probable, that hospitalists in this study reduced complications, avoided harm, and improved inpatient mortality. Perhaps this, and not a zeal for too-early discharges, is what fueled the lower hospital cost and shorter LOS.

How much is that worth? It’s hard to say, but I’d venture much more than the $50 more per patient associated with the hospitalist model. Quality, even at higher costs, needs to be our primary focus moving forward. We must improve the quality of care to levels that, if necessary, Medicare would happily pay more for. This must be our singular goal. I’d also argue that we include the post-discharge period in our quality reach.

More QI, Less MEN

We cannot continue to train square pegs and struggle to cram them into round holes. We need more systems thinkers. We do a tremendous job teaching our students and residents about the interplay of pathophysiology and pharmacology, but spend very little time with the interplay between our patients and the system. We simply must triage process improvement, quality, safety, and efficiency training closer to the top of our medical curricula.

As such, it shouldn’t be that surprising that two groups of providers with the exact same background and training should have similar outcomes as seen in this study. The reality is that these providers came from very similar training backgrounds. Yes, they have chosen different practice models, but we all learned how to treat pneumonia and heart failure. It’s not about the model as much as what you do with the model. Cohorting patients to providers who just care for hospitalized patients will lead to efficiencies, but if we want to fundamentally improve patient outcomes, both during and after their hospital stay, we need to train hospitalists to transform that model through systematic process improvement.

And I firmly believe that hospitalists should lead this sea change. Our teaching brethren are perfectly positioned to develop hospitalist-focused training models that better prepare future hospitalists to fundamentally improve not just transitions of care but indeed all systems of care. Training that emphasizes systems thinking, mentored process improvement, and patient safety across the continuum of care.

BOOSTing Outcomes

This paper highlights HM’s Achilles’ heel. It has always been transitions of care—specifically, communication with PCPs. HM is by its very nature a fractured care model. And that discontinuity results in information drop on transitions. A PCP who knows a patient and admits and follows them after discharge is better positioned to reduce readmissions because there is no information drop in that model. The success of the HM model hinges on hospitalists efficiently and effectively approximating that level of knowledge transfer to PCPs. And to be honest, we don’t need an NIH-funded study to tell us that we have not been doing a great job with this.

This is not necessarily from a lack of effort but rather because we lack systems that simplify information transfer on transitions. It is incumbent these systems be built. It is incumbent we lead this. Whether you choose Project BOOST, Project RED, or a homegrown solution, it is no longer acceptable to ignore the transitions of care issue.

And with that I’ve just typed my 1,319th word about an article that has already commanded too many words. Don’t get me wrong, the discussion is important, and for that we must thank the study’s authors. But it’s time to move beyond the discussion, the debate, the words.

It’s time to turn words into deeds.

Dr. Glasheen is physician editor of The Hospitalist.


  1. Kuo YF, Goodwin JS. Association of hospitalist care with medical utilization after discharge: evidence of cost shift from a cohort study. Ann Intern Med. 2011;155:152-159.

Next Article:

   Comments ()