Last month, I discussed ED, or “front end,” throughput. This month I will focus on what can be thought of as the two components of “back end” throughput: effective management of length of stay (LOS) and patient discharge in the morning rather than late in the day.
How many times have you heard a well-intentioned administrator ask: “We can’t get patients from the ED to ward more quickly unless we can discharge patients earlier in the day to make beds available. So please round on your potential discharges and get the D/C order written early”?
Easier said than done. But not impossible, if you’re willing to make some adjustments in how you organize your work.
I provided a list of strategies in my April 2009 column (see “Early-Day Discharge Strategies,” p. 48), and in May 2010 I described in greater detail two ideas that can help with early-morning discharge. The easiest of these is to increase the portion of patients who discharge early in the day. To do this, hospitalists should write in the order section of the chart (not just the progress notes where it can be missed by nursing staff) an order like “possible disch tomorrow” or “Probably discharge Thurs or Fri.” That will help hospital staff anticipate and prepare for discharge, and there is little cost if the patient isn’t ready on the day forecast.
More difficult, but more effective, is fully preparing a patient’s discharge a day ahead of time. I do this on about half or more of my patients and, despite having no rigorous data to prove it, I’m convinced that it makes for better discharges and transitions, and it’s a real stress-reliever for me. My mornings are much less hectic, as I rarely have to devote 30-60 minutes to a discharge while other patients are waiting to be rounded on. And it helps me uncover loose ends like the need to get additional chest imaging to evaluate a possible lung lesion early enough that I can order the additional test without delaying discharge (it can be done the night before).
Note that when billing the discharge visit only, the time spent on the day of discharge counts as billable time. So preparing everything the day before will mean that nearly all discharges will be billed at the lower level: 99238 rather than 99239. You will have to decide for yourself whether losing the ability to bill some discharges at the higher level is worth it. Most hospitals probably will be willing to make up the lost professional fee revenue if it led to a meaningful improvement in the portion of patients with discharge orders early in the day.
One CMS administrator told my consulting partner, Leslie Flores, that discharge summaries can’t be dictated on a day prior to discharge because they have to include all information related to the hospital stay, including any relevant information from the night before discharge. The administrator said that CMS would view this as a surgeon dictating an operative report before performing the operation. (Are there surgeons who are actually doing that?) I think care is better when I do a discharge summary unhurried and with few interruptions in the evening before discharge rather than the busy morning of the discharge day. And I can always add an addendum (and often do) at the time of actual discharge. So you might want to ask your hospital compliance expert about dictating a D/C summary the day before discharge.
A Tricky Proposition
It is pretty easy for administrators to think that the time of day a discharge order is written is nearly always up to the doctor and how they organize their day. To them, it might seem reasonable to expect more than 50% of discharge orders to be written before 10 a.m. or earlier. But wait … it is more complicated than that.
The same administrators care a lot about LOS, which is most commonly reported in days but can also be measured in hours. So the time of day a discharge order is written, arguably the most important determinant of when a patient will vacate a room, is a determinant of LOS. If we’re really managing LOS optimally, we should discharge a patient at the first moment it is clinically appropriate, which means discharging at any hour of the day or night. (Of course, no one is proposing that we discharge and expect patients to vacate rooms in the middle of the night. Yet.)
But that would mean only a small portion of patients would end up with early-morning discharge orders. So excellent LOS management and a high portion of discharge orders written early in the day are incompatible. Setting the target percentage of early-a.m. discharge orders too high probably will increase LOS and defeat the original objective.
Benefits and Costs of Improved Throughput
Let’s say you’re sold on the value of improving throughput. It should lead to improved efficiency and financial performance for your hospital. I suspect it will improve quality and reduce iatrogenesis for some patients, but risk more readmissions and quality lapses or errors for others. And as every department tries to improve their own throughput, there will be a tendency to push problems off on others. For example, it is easy to improve ED throughput if the ED doctor just does a lot less evaluation and sends patients upstairs without much of a workup. Example: “The patient has fever and low sats, so I know he’s going to be admitted. Why should I keep him in the ED to do a workup?”
In fact, I think we should move away from using the ED as a triage unit and send some patients directly from ED triage to the inpatient unit. But we’ll need to put in place systems that make that safe and ensure good care. I don’t think any hospital has such systems in place now.
Let’s say that by 2013 most hospitals have dramatically improved their throughput, have short ED admit decision time to ED departure, and many patients are discharged and vacate their rooms early in the day. And let’s say we’ve been able to do that while maintaining or improving quality of care (we must!). We’ll then have happy patients and hospital administrators, and can bring up this improved performance when negotiating for hospital financial support for our practice. Everyone is happy, right?
Well, won’t this increase marginal or unnecessary admissions and readmissions, and lead to overall increased hospital utilization? After all, if it becomes really quick and easy to admit a patient to the hospital, won’t we do it more often? I think we will, but would never propose that we maintain poor throughput to keep a lid on costs and inappropriate utilization. But we’d better come up with other ways, or improved throughput will just be one more factor contributing to escalating healthcare costs.
Dr. Nelson has been a practicing hospitalist since 1988 and is co-founder and past president of SHM. He is a principal in Nelson Flores Hospital Medicine Consultants, a national hospitalist practice management consulting firm.