Published in the July 2018 issue of Today’s Hospitalist
HOSPITALS ON THE HOOK for readmission penalties have this chronic frustration: Too many patients discharged to SNFs come back as either repeat ED visits or full-blown readmissions. SNFs, it seems, just don’t have the staff or resources to keep patients from bouncing back once they’re discharged.
But what if SNFs could identify patients at high risk of being readmitted as a first step toward targeting those patients with needed interventions? That’s the idea behind the development of the SNF Prognosis Score, which was detailed in research published in the March issue of the Journal of the American Geriatrics Society. The score—which consists of five separate factors (see “SNF Prognosis Score“)—was developed based on data about Medicare fee-for-service patients discharged from hospitals to SNFs.
See our latest article “Your patient is going to a SNF. Should you follow her there?” (July 2019) for more on handling SNF transfers.
The score “demonstrated very good discrimination,” the authors wrote, at identifying SNF patients at high risk for one of three outcomes: readmission to the hospital (the most common), mortality or a long SNF stay of more than 100 days. (The authors found that 28.8% of patients had one of those outcomes, with 19.4% readmitted.)
“Of the five factors that make up our final model, function is the most important.”
~ Robert Burke, MD, MS
Corporal Michael J. Crescenz VA Medical Center
While risk-scoring models have been developed to predict readmissions of patients being discharged home, “I’m not aware of any other standardized score that might help a SNF figure out how likely a particular patient is to get into trouble,” says hospitalist and lead author Robert Burke, MD, MS. Dr. Burke finished the research while he was at the Denver Veterans Affairs Medical Center, and he is currently practicing at Philadelphia’s Corporal Michael J. Crescenz VA Medical Center. He also is the director of research for the hospital medicine section at the University of Pennsylvania.
While publishing such a risk score may help SNFs and their patients, it may also prove to be very timely. That’s because SNF readmissions are about to become a high-stakes topic.
While SNFs have been publicly reporting their readmission rates, they will as of FY 2019—which kicks off this October—start feeling the pain of financial penalties for excessive readmissions. That’s been a reality for hospitals since 2012.
“If you’re a skilled nursing facility with a readmission problem,” says Dr. Burke, “the pressure is really on to figure out what to do about that before this October.” Dr. Burke spoke to Today’s Hospitalist.
Why is such a score important?
We’re in a payment environment where hospitals and post-acute care facilities are increasingly being held responsible for patient outcomes. And SNFs are really a resource-constrained setting. It’s to their financial advantage to accept as many patients as they can, yet they have limited resources and staff to appropriately care for them.
That’s particularly true in terms of nursing: The predominant nursing staff in SNFs are certified nurse assistants, while patients in hospitals are assessed hourly by an RN. That’s a big step down in treatment intensity. If I’m a SNF medical director, I’m thinking about how to best allocate my or my nurses’ limited time toward higher-risk patients and away from lower-risk ones.
So this is a scoring system that SNFs would use, right? Not hospitals at discharge?
One of the score’s variables, the Barthel index, is not routinely measured at discharge, but it is used at SNF admission. If I was a hospitalist envisioning a program to improve care transitions to SNFs, I would have to import these assessments.
That said, of the five factors that make up our final model, function—as measured by that index—is the most important. If hospitalists had time to focus on only one thing in determining whether a patient being discharged to a SNF might be at high risk of being readmitted, they should consider patients’ level of function.
You write that the point of scoring is to better match monitoring and interventions to the patients who need them. Is it your sense that SNFs now aren’t providing those?
My impression is that hospitals and SNFs are now coming to the table to decide together how to improve outcomes and deliver higher-value care, driven by financial pressures.
There are a lot of ad hoc solutions, which can be valuable, but there aren’t a lot of standardized or generalizable ones. I think this score can contribute to having a common understanding about the types of patients who are transitioning.
What types of interventions could SNFs offer high-risk patients?
I can think of two examples. One is SNF Connect, a heart failure readmission program in Colorado to improve heart failure disease management in SNFs.
The other is INTERACT, which many nursing homes—and it’s for all nursing homes, not just SNFs—have sought to implement. It’s a complicated multi-part intervention that had a negative finding in a recent randomized trial because it was difficult to implement. But SNFs that successfully implemented INTERACT significantly reduced their hospitalization rates. A growing number of SNFs may consider investing in training their staff to implement that program.
Are you using this score yet in your own practice?
Not yet, but the University of Pennsylvania, like many institutions, is developing a SNF network of preferred providers. I anticipate using this score in that work, and it would be easy to embed in our EHR as a SNF screening tool. We could then work together to pair high-risk patients with interventions.
The study mentions the possibility of incorporating other factors into the risk score, like gait speed or vital-sign stability. Do you plan to continue to refine it?
We’re looking for other data sets that we may be able to use. But we’re going to reach a point of diminishing returns, and skilled nursing facilities and hospitals may prefer to turn instead to using the score right now to identify high-risk patients, rather than in getting the score to be even more predictive.
With penalties looming, the tolerance among SNFs for the perfect score is probably a lot lower. They’re looking for something they can operationalize right now.
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
SNF Prognosis Score
PUBLISHED IN the March 2018 issue of the Journal of the American Geriatrics Society, the SNF Prognosis Score is designed to help SNFs differentiate between high- and low-risk patients in terms of readmissions, mortality and long SNF stays. The score is based on the following five variables that are routinely collected by most SNFs within the first week of admitting patients:
1. The Barthel index. This scores how well patients function in terms of mobility and daily-living activities like grooming, bathing and self-feeding.
2. Charlson-Deyo comorbidity score. This scoring system is pegged to about a dozen different diagnoses. Most—including MI—are scored as 1 point, although moderate or severe liver disease counts for 3 points and AIDS counts for 6.
3. Hospital length of stay.
4. Heart failure diagnosis.
5. The presence of an in-dwelling catheter.