Published in the July 2014 issue of Today’s Hospitalist
FEELING PRESSURE from administrators to keep piling on more patient encounters? Now, for the first time, hospitalists have some ammunition to push back: study findings that hospitals pay a steep price when physician workloads are too high.
Hospitalist researchers in a large private group with the Christiana Care Health System in Wilmington, Del., took a comprehensive look at what happens in the two hospitals they cover to length of stay and costs per case as daily census rises.
Their retrospective study encompassed more than 20,200 admissions between 2008 and 2011. Their findings were published in the May issue of JAMA Internal Medicine.
First, the good news: Researchers did not find any association between higher hospitalist workloads and greater mortality, more seven-day and 30-day readmissions, lower patient satisfaction scores, or how often rapid response teams had to be activated.
But there was bad news: Length of stay and hospital costs both shot up as hospitalist census increased beyond a daily mean of 15.5 patient encounters. At the higher end of census range that researchers considered “from 11 to 22 patients per day “length of stay increased as much as two days, while costs per patient rose between $5,000 and $7,500.
The study also mapped the interplay between hospitalist workload and hospital occupancy levels, based in part on how many non-ICU medicine beds were filled. Occupancy levels were categorized as either low (less than 75%), medium (75% to 85%) or high (greater than 85%).
At low-occupancy levels, LOS increased linearly from 5.5 to 7.5 days across low to high physician workloads. At medium occupancy, however, LOS stayed stable in the lower workload range but then rose sharply above a census of 15. And with high hospital occupancy, LOS increased dramatically as census rose above approximately 19 patients.
Why? Researchers concluded that the data suggest that “as hospitals reach capacity, LOS is affected more by hospital factors, such as demands for nursing support or other ancillary services, than by physician factors at lower ranges of hospitalist workload. However, hospitalist workload becomes a driving factor above these thresholds.”
Researchers also concluded that their data imply that focusing solely on productivity may miss the much bigger picture. “Hospitalists should be prepared to increase services to handle high clinical loads … perhaps through additional care coordination or discharge support for providers,” the authors concluded. “At a minimum, incentive programs should balance productivity, efficiency, and quality measures.”
Today’s Hospitalist spoke to lead author Daniel Elliott, MD, MSCE, a hospitalist who is Christiana Care’s associate chair of research in the department of medicine.
Why tackle the issue of hospitalist workload?
It was a very practical question to try to answer. We have a lot of internal discussions about how many patients physicians feel comfortable seeing, as many hospitalist groups do.
But we’re also struggling as a health care system to really understand what value looks like and how we can regularly deliver the highest-value care. Workload seemed to be a question with more than personal interest. It’s a piece of the equation that may have implications for whether we can deliver high-value care day in and day out.
The study is quite complicated, assessing not only the impact of workload on cost and length of stay, but the role of hospital occupancy levels as well.
We understand that the hospital environment is dynamic and that a lot of things affect efficiency, length of stay and costs, not all of which are under the doctor’s control. We know from our personal experience and from talking to other hospitalists that when the hospital gets busy, things happen differently than when the hospital is less busy. We felt that we had to find ways to incorporate these other variables, and we knew that capacity was going to be a big factor.
At lower occupancy levels, the rise seen in both length of stay and workload is almost linear. But that’s not the case at medium and high occupancy. Does that mean that hospitalists ” at least those in your group “may get used to working higher workloads and aren’t affected until they pass a certain census threshold? Something happens as hospital occupancy reaches those middle-range or higher levels where, particularly when doctors’ census is lower, we don’t see workload having an impact on length of stay. But when you hit that critical threshold “which, again is a range “we see a dramatic increase in length of stay as we get to higher workloads in a higher occupancy situation.
What do you think that means?
It suggests that the strength of association between workload and length of stay at lower hospital occupancy levels is more direct.
If you have less busy hospitalists but a very busy hospital, the hospitalists may have less impact on length of stay. At higher occupancy levels, there may be something about the hospital system and throughput that attenuates that relationship, even when hospitalists have lower workloads.
But when you have both a busy hospital and a busy hospitalist, that’s when costs and length of stay, at least in our data, go up at a much higher rate.
Potentially, there are hospitalist-dependent factors and hospital-dependent factors. As you get to medium and high occupancy levels, hospital-dependent factors may affect efficiency of care more than hospitalist-dependent factors. What those factors are “possibly nursing services or ancillary care, for instance “would be conjecture. We did not explore that in our study.
Obviously, your finding that higher workloads did not adversely affect quality was good news.
I think it’s incredibly reassuring news. It appears that hospitalists are able to provide safe and quality care across the spectrum of workload.
But it will be important to look at the relationship between workload and outcomes in different environments. One caution of a single observational study that answers a question that hasn’t been answered before is that people think it answers that question for everybody. The impact of occupancy may vary by hospital or hospitalist group, and by work structure, back-up and support within each group.
I think we answered the question in a rigorous way for our group in our environment. But really understanding what this relationship looks like in other environments with other practice structures is a key next step.
Phyllis Maguire is Executive Editor of Today’s Hospitalist.