Home Coding Not too long, not too short, just right

Not too long, not too short, just right

September 2014

Published in the September 2014 issue of Today’s Hospitalist

I’LL HIT THE 10-YEAR MARK AS A HOSPITALIST this coming April. Although I’m not exactly an old-timer, I’ve lived through some of the specialty’s major developments. (I’ve also lost some hair and started to turn grey. There’s probably a connection.)

Back in the day, hospitalists were trying to land on a financial model to justify their existence. Hospitals were forking over big bucks to employ us, and it wasn’t entirely clear why we were superior to independent community providers who rounded on their own patients for “free.”

Hospital length of stay (LOS) turned out to be the lever. We could discharge patients whenever, whereas community providers tended to empty beds in the morning when they rounded before clinic. Hospitalists, it was argued, produced lower lengths of stay, which would improve hospitals’ bottom lines.

At first, the evidence was a little scant, but now the LOS differential is pretty clear. A meta-analysis in the Jan. 1, 2012, American Journal of Managed Care, for example, found that hospitalists reduced LOS by at least 0.44 days per case. That’s only about 11 hours, but it quickly adds up on an annual basis.

Medicare DRGs
Medicare’s current inpatient prospective payment system emerged in the early 1980s as a catalog of diagnosis-related groups (DRGs). Each DRG represents a class of patients, all of whom need similar care and services. Treatment of pneumonia thus became a commodity that Medicare could purchase from any hospital in the country for a specified list price.

For example, “simple pneumonia and pleurisy without complications and comorbidities or major complications and comorbidities” is DRG 195. Medicare anticipates an average LOS of 2.9 days and pays roughly $4,100 per case.

While DRGs are different than the codes we use to bill for our professional services, the two are very much related. Take the patient with a primary diagnosis of pneumococcal pneumonia (ICD-9 code 481). We might submit a charge for a mid-level H&P (E/M code 99222) to be reimbursed for our work. The hospital, in turn, will take that same ICD-9 code and bill for DRG 195 to cover all other aspects of the patient’s care.

Why shorter is better
On its face, shorter LOS would seem to be a bad thing. In the hotel industry, for instance, longer LOS juices revenue. Holiday Inn charges by the night, and extending your stay produces a larger bill at check-out.

But Medicare pays a flat rate for each DRG, regardless of LOS. If you can discharge a patient with pneumonia in 1.9 days, your hospital probably makes money because it takes fewer resources to deliver the necessary services in a shorter time. Break-even is around 2.9 days, and beyond that, your hospital risks losing money.

This is why inpatient care managers obsess over LOS. They either want the patient discharged as you approach Medicare’s average LOS, or they pester you to add data to the medical record to support a DRG that pays more. For example, “simple pneumonia and pleurisy with complications and comorbidities” (DRG 194) “average LOS of 3.8 days “is worth about $5,700.

“Doctor, doesn’t that patient with pneumonia also have COPD?”

Why longer is better
But shorter LOS, while good for hospitals and hospitalists, isn’t necessarily great for patients and Medicare. Take decompensated heart failure. While hospitalists might rack up shorter lengths of stay, these patients are readmitted at a dreadfully high rate: 50% or more within six months.

Not surprisingly, Medicare got sick of paying for a fresh DRG for each bounceback and imposed readmission penalties instead. Hospitals have since implemented all sorts of countermeasures: post-discharge telephone calls, better handoffs to ambulatory providers and timely follow-up appointments in outpatient clinics, just to name a few.

I’m a little dubious, mainly because a systematic review in the Oct. 18, 2011, Annals of Internal Medicine found poor underlying evidence for the effectiveness of many of these interventions. The review ultimately concluded that, “No single intervention implemented alone was regularly associated with reduced risk for 30-day rehospitalization.”

Maybe it’s time for a very counterintuitive solution: increasing length of stay. This seemingly half-baked idea actually has some merit.

In the February 2014 issue of Medical Care Research and Review, Kathleen Carey, PhD, and MengYun Lin, both with Boston University School of Public Health, analyzed a large administrative data set from California for patients with acute myocardial infarction (AMI) and heart failure. Starting with an index hospitalization, the researchers looked at the subsequent rate of readmissions and did some statistical hocus-pocus dealing with LOS. They studied roughly 70,000 cases.

For heart failure and AMI, 30-day readmission rates were 12% and 16% respectively. Increasing LOS by one day reduced the probability of readmission by 6.5% for AMI and 1% for heart failure.

The effect was even more pronounced when the researchers looked at short hospitalizations, defined as a length of stay of seven days or less. Increasing LOS by one day for those hospitalizations reduced the readmission probability for AMI patients during the first week after discharge by a whopping 18%. The corresponding figure for heart failure was 8%.

Just right
These data suggest that hospitalists may need to change their posture toward LOS. For some conditions ” particularly those with high readmission rates “it might be smarter to keep patients longer. The key is finding the sweet spot between making a buck on shortening LOS and avoiding readmission penalties by increasing it.

What might that spot look like? Let’s take garden-variety heart failure (DRG 293, average LOS of 2.6 days), which pays about $3,900 per admission. Suppose your hospitalist group found a way to reduce LOS by 0.25 days, which is about 10% or $390 per case.

But your group also has the opportunity to hold onto patients for one day longer “a 40% increase in LOS that potentially adds $1,500 more to that cost per case. Which is better: shorter or longer?

At first blush, the roughly $1,900 difference would seem to favor shorter LOS. But the math is much more complicated. What if that shorter LOS resulted in a higher readmission rate, while longer LOS reduced bouncebacks? It might be worth the immediate financial hit associated with longer LOS to avoid eventual readmission penalties.

In the end, the answer is probably “both/and.” The challenge is finding low-risk cases to discharge sooner and high-risk cases to sit on longer.

I’ll let you struggle with how to do this. As for my group, we just went live with a new electronic health record, which promises some sophisticated analytics. The EHR may even lead us to the holy grail: figuring out for this patient with these characteristics on that DRG, what the optimum LOS should be.

David A. Frenz, MD, is a hospitalist for HealthEast Care System in St. Paul, Minn., and is board certified in both family and addiction medicine. You can learn more about him and his work at www.davidfrenz.com or via LinkedIn.