Editor’s note: This is the first of a three-part series.
My experience is that some, maybe even most, hospitalists tend to assume there is a standard or “right” way to organize things like work schedules, compensation, or even the assignment of patients among the group’s providers. Some will say things like “SHM says the best hospitalist schedule is …” or “The best way to compensate hospitalists is …”
But there really isn’t a “best” way to manage any particular attribute of a practice. Don’t make the mistake of assuming your method is best, or that it’s the way “everybody else does it.” Although scheduling and compensation are marquee issues for hospitalists, approaches to distributing new patients is much less visible. Many groups tend to assume their method is the only reasonable approach. The best approach, however, varies from one practice to the next. You should be open to hearing approaches to scheduling that are different from your own.
Assign Patients by “Load Leveling”
I’ve come across a lot—and I mean a lot—of different approaches to distributing new patients in HM groups around the country, but it seems pretty clear that the most common method is to undertake “load leveling” on a daily or ongoing basis.
For example, groups that have a separate night shift (the night doctor performs no daytime work the day before or the day after a night shift) typically distribute the night’s new patients with the intent of having each daytime doctor start with the same number of patients. The group might more heavily weight some patients, such as those in the ICU (e.g., each ICU patient counts as 1.5 or two non-ICU patients), but most groups don’t. Over the course of the day shift, new referrals will be distributed evenly among the doctors one at a time, sort of like dealing a deck of cards.
This approach aims to avoid significant imbalances in patient loads and has the potential cultural benefit of everyone sharing equally in busy and slow days. Groups that use it tend to see it as the best option because it is the fairest way to divide up the workload.
Practices that use load-leveling almost always use a schedule built on shifts of a predetermined and fixed duration. For example, say the day shift always works from 7 a.m. to 7 p.m. This schedule usually has the majority of compensation paid via a fixed annual salary or fixed shift rate. One potential problem with this approach is that the doctor who is efficient and discharges a lot of patients today is “rewarded” with more new patients tomorrow. Hospitalists who are allergic to work might have an incentive to have a patient wait until tomorrow to discharge to avoid having to assume the care of yet another patient tomorrow morning. Hospital executives who are focused on length-of-stay management might be concerned if they knew this was a potential issue. Of course, the reverse is true as well. In a practice that doesn’t aggressively undertake load-leveling, a less-than-admirable hospitalist could push patients to discharge earlier than optimal just to have one less patient the next day.
Another cost of this approach is that the distribution of patients can be time-consuming each morning. It also offers the opportunity for some in the group to decide they’re treated unfairly. For instance, you might hear the occasional “just last Tuesday, I started with 16 patients, compared with 15 for everyone else. Now you want me to do it again? You’re being unfair to me; it’s someone else’s turn to take the extra patient.”
Assignment by Location
Groups that use “unit-based” hospitalists distribute patients according to the unit the patient is admitted to—and the hospitalist covering that unit. The pros and cons of unit-based hospitalists are many (see “A Unit-Based Approach,” September 2007), but there is an obvious tension between keeping patient loads even among hospitalists and ensuring that all of a hospitalist’s patients are on “their” unit. Practicality usually requires a compromise between pure unit-based assignment and load-leveling.
Some groups assign patients according to a predetermined algorithm and employ load-leveling only when patient loads become extremely unbalanced. For example, Dr. Jones gets all the new referrals today, and Dr. James gets them tomorrow. The idea is that patient loads end up close to even over time, even if they’re unbalanced on any given day.
A system like this allows everyone, including the hospitalists themselves, ED staff, etc., to know who will take the next patient. It decreases the need to communicate the “who’s next” information time after time during the course of the day. In small- to medium-sized practices, it could mean no one needs to function as the triage doctor (i.e., the person who inefficiently answers the service calls, scribbles down clinical information, then calls the hospitalist who is due to take the next patient and relays all the pertinent patient info). This system allows the
hospitalists to know which days will be harder (e.g., taking on the care of new patients) and which days will be easier (e.g., rounding but not assuming care of new patients). Allowing uneven loads also eliminates the need to spend energy working to even the loads and risking that some in the group feel as if they aren’t being treated fairly.
Uncommon yet Intriguing Approaches
Pair referring primary-care physicians (PCPs) with specific hospitalists. I’ve encountered two groups that had hospitalists always admit patients from the same PCPs. In other words, hospitalist Dr. Hancock always serves as attending for patients referred by the same nine PCPs, and hospitalist Dr. Franklin always attends to patients from a different set of PCPs. It seems to me that there could be tremendous benefit in working closely with the same PCPs, most notably getting to know the PCPs’ office staff. But this system raises the risk of creating out-of-balance patient loads, among other problems. It is really attractive to me, but most groups will decide its costs outweigh its benefits.
Hospitalist and patient stay connected during admission. I’m not aware of any group that uses this method (let me know if you do!), but there could be benefits to having each patient see the same hospitalist during each hospital stay. Of course, that is assuming the hospitalist is on duty. Hospitalist and patient could be paired upon the patient’s first admission. The hospitalist could form an excellent relationship with the patient and family; the time spent by the hospitalist getting to know patients on admission would be reduced, and I suspect there might be some benefit in the quality of care.
This method, however, likely results in the most uneven patient loads, and load-leveling would be difficult, if not impossible. Even if hospitalist and patient did form a tight bond, there is a high probability that the hospitalist would be off for the duration of the patient’s next admission. So despite what I suspect are tremendous benefits, this approach may not be feasible for any group.
In next month’s column, I will discuss issues related to the way patients are distributed. TH
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 (www.nelsonflores.com). He is co-director and faculty for SHM’s “Best Practices in Managing a Hospital Medicine Program” course. This column represents his views and is not intended to reflect an official position of SHM.