The Hospitalist-Patient Conundrum


Hospitalist-patient continuity is primarily a function of the number of consecutive days worked by a hospitalist, but the way new referrals are distributed can also affect continuity. This month, I will discuss both.

For my first few years as a hospitalist in the late 1980s and early 1990s, my one partner and I generally worked a schedule of 21 days on and seven days off. While I wouldn’t recommend that anyone try that today, it wasn’t as bad as you might think, because our patient volumes weren’t terribly high and, on about a third of the worked days, I was done shortly after lunch.

While working that schedule, I became aware of its benefit to hospitalist-patient continuity. I can remember many patients with hospital stays of more than two weeks whom I saw every day myself. As you might imagine, my partner and I talked periodically about working fewer than 21 days at a stretch and handing a service over to one another more frequently. But we were concerned that this would make us inefficient because more of our worked days would involve getting to know a new list of patients. In effect, we’d work more hours without an increase in income or patient volume.

While still working the 21-day schedule, I came to know another practice and was stunned that these doctors had taken essentially the opposite approach to scheduling. They worked 24-hour shifts on site and never worked more than one shift at a time. (If your shifts are 24-hours long, you probably can’t or shouldn’t work more than one at a time.) This schedule meant that a patient would see a different hospitalist each day. I couldn’t believe that either the patients or the hospitalists would think this was a reasonable thing to do, but the doctors were convinced it worked well. Later I learned that this group had been started by an emergency medicine practice, and it seems they had made the mistake of inserting an emergency department (ED) physician schedule into a hospitalist practice—and 24-hour shifts for ED doctors were more common then than now.

While the number of consecutive days of work is the most important variable in determining continuity, the way a group distributes new referrals can also have an impact.

So, early in my career, the first two schedules I became acquainted with sat on opposite ends of a continuum that has since been filled in by many other options. Both the practice I was part of and the 24-hour-shift practice abandoned their original schedules within a few years and moved on to other alternatives. In fact, I have since worked nearly every schedule you can imagine, including the seven-on/seven-off schedule, which I think is a suboptimal choice for most groups. (See August 2006 “Career Management,” p. 9.) With each variation in my work schedule, I’ve thought a lot about its effect on continuity.

While there isn’t a great deal of research to prove it, improved continuity is probably associated with improvements in things like:

  • Quality care and safety;
  • Patient satisfaction (and probably hospitalist satisfaction also);
  • Hospital resource utilization; and
  • Hospitalist efficiency.

When hospitalists design a schedule, I recommend that the doctors think first about what will allow them a sustainable lifestyle while ensuring the necessary coverage—for some practices, this means keeping a doctor in the hospital around the clock. Ideally, they will come up with several options that satisfy these two metrics. In many cases, the option that results in the best continuity is the one they should choose.

While the number of consecutive days of work is the most important variable in determining continuity, the way a group distributes new referrals can also have an effect. If a hospitalist doesn’t take on any new referrals on the day before starting a series of days off, there will be fewer patients to hand over to the doctor who takes over the next day. If every member of the practice is exempted from taking on new patients on the day before rotating off, then fewer patients will have a new hospitalist, which means continuity is better. Another way to think of this is that new referrals are “frontloaded” into the earlier days a hospitalist works in each series, and no new patients are assigned on the last day—or maybe more—of a hospitalist’s consecutive days of work.

In my experience, most groups try to distribute patients equally to each doctor each day. This means that a doctor who will be off starting tomorrow takes the standard portion of new referrals to the practice today. The fact that all of those patients will be turned over to a new hospitalist tomorrow, however, adversely affects continuity.

Note that exempting a group member from taking on new patients on the last day of a rotation means that the doctors can’t all rotate on and off on the same day. For example, if a group has four doctors working each day, and all rotate off on the same day—a common practice—then of course it is impossible to exempt all the doctors from new referrals on the day before going off. Instead, it will work best if only one doctor rotates on or off at a time. That leaves all the other doctors available to accept new referrals, while exempting the one about to rotate off.

You can see how adjusting the number of consecutive days worked affects continuity by using this formula:

Continuity = [days — (LOS — 1)]/days x 100

Continuity = the percent of patients who have the same hospitalist for every day of their stay

Days = number of consecutive days worked by an individual hospitalist

LOS = length of stay (best to use median LOS, but mean LOS still allows a reasonable estimate)

Note: The continuity that is derived using this formula can be improved if doctors in the practice are exempted from accepting new referrals on the day before rotating off.

This approach has benefits that go beyond improving continuity. It means a hospitalist’s workload the day before rotating off will be lighter than on other days. This provides extra time to “tee the patient up” for the next hospitalist—maybe write a more detailed note or dictate a discharge summary in anticipation of the patient leaving the next day—and/or it means a shorter day of work, assuming the doctors can leave the hospital when their work is done and aren’t required to stay around until the end of a pre-defined shift. This shorter day can mean an opportunity to do things like scheduling a flight to vacation late on the last day of work rather than the next morning, which might mean an extra day out of town.

And if a doctor isn’t assigned any new patients on the day prior to rotating off, there will be fewer patients to get to know for the doctor coming in the next day. The first day back on service will be less stressful and, arguably, more efficient.

A principle benefit of having all hospitalists switch on the same day is that the system quickly becomes transparent to the hospital staff. They learn that every Monday—or whatever day is decided upon—all of the hospitalists are getting to know patients for the first time and will probably round much more slowly. Many important processes, such as discharges, may be delayed until later in the day. If only one hospitalist switches at a time, hospital staff will have a hard time keeping up. Any inefficiencies caused by the switch will be spread over many days, however, and this may make it easier to handle. And if patients are teed up better for the new hospitalist and there are fewer patients on the list, the inefficiency that is caused by a new doctor getting to know the patients can be offset by the smaller number of patients. That means, for example, that discharges are much less likely to be delayed because of the smaller patient load on the doctor’s first day on service.

Up to this point I’ve been discussing continuity during a single admission. What about continuity from one admission to the next? Nearly all groups assign patients based on when they are admitted, and the hospitalist who cared for the patient during a prior admission may have little influence on which hospitalist admits them this time. If a patient is readmitted within a week or two, and the previous attending hospitalist is working, some groups will try to pair them once again through a bounce-back system.

It’s worth thinking about whether your group could make an effort to always have the same hospitalist care for a patient unless that physician is off—even if the admissions are months apart. This system would mean that on first admission to the practice a patient would be assigned to the hospitalist who is up next. In this way, each hospitalist in the group would develop his or her own panel of patients. This would be particularly valuable for patients who are admitted frequently; however, it would be difficult for a doctor to control how labor-intensive his patient panel might become. One person might have the bad luck to collect far more medically and socially complex patients than others in the group, and workloads might become unbalanced, making the whole group less efficient. I’m hopeful that a group will come up with a way to overcome these problems and create a workable system of good continuity from one admission to the next, but, as far as I’m aware, no group is doing this now. If you have a workable system, please let me know.

One group I worked with years ago addressed continuity from one admission to the next by using a system that matched each hospitalist with a panel of referring doctors. For example, the same hospitalist would always admit the patients “belonging to” a cadre of primary care physicians (PCPs), and another hospitalist in the group would always admit patients from another set of PCPs. The patient would see the same hospitalist each admission, and the hospitalist could develop a close working relationship with the panel of PCPs. The hospitalist and the PCP became familiar with each other’s practice styles, schedules, and days off, and memorized one another’s phone and fax numbers, the names of office and support staff, and so on, making for a very smooth working relationship that could benefit patient care. If the assigned hospitalist was off when a particular PCP’s patient needed admission, then a partner would provide interim care and turn the patient over when that hospitalist returned.

As you can imagine, this can be a difficult system to implement because there are many days on which a patient might be hospitalized when the assigned hospitalist is not around. Additionally, it is nearly impossible to divide PCPs and their patients equitably so that each hospitalist has a reasonable workload and patient complexity. I can imagine this group meeting periodically to match hospitalists and PCPs in a fashion similar to a fantasy football draft: “I’ll take PCP Smith and Williams from you, if you take PCP Wilson off my hands.”

I’m interested in hearing any additional ideas groups have developed to facilitate good continuity. The number of consecutive days worked by each hospitalist and the way new patients are assigned are significant ways to influence continuity, but there may be others that we should all keep in mind. And remember, maximizing continuity is not only good for patients; it enables the hospitalist practice to function more efficiently because it minimizes the number of new patients each hospitalist will have to get to know. TH

Dr. Nelson has been a practicing hospitalist since 1988 and is a co-founder and past-president of SHM. He is a principal in Nelson/Flores Associates, a national hospitalist practice management consulting firm. This column represents his views and is not intended to reflect an official position of SHM.

Next Article:

   Comments ()