CMO Brief: Inpatient Continuity of Care

Last time, we completed our discussion of the impact of outpatient continuity of care, so now we’ll begin to look at the available literature on inpatient continuity of care.  This is not as well studied as the outpatient world, so we have less to work with, but there are still some take home points.

When we think about hospitalist continuity, the first challenge is how to measure this.  The easiest way to think about this/measure this is to examine hospitalist discontinuity.  There are two common ways this is measured in the literature, as in O’Leary, et al from 2015.  The first is NPI or Number of Physicians Index.  This is simply the “total number of unique hospitalists caring for the patient”.  Obviously, as this number goes up, discontinuity goes up.  The second is UPC or Usual Provider of Care.  This is the “proportion of encounter with the most frequently encountered hospitalist”.  The lower this number is, the greater the discontinuity.

There is an assumption, that while hospitalists provide excellent care for their patients, the hospitalist model has led to an increase in the discontinuity that patients experience.  Surprisingly, Fletcher, et al found that while discontinuity had been increasing between 1996 and 2006, this did not appear to be related to the hospitalist model of care.  In fact, patients who received all of their care from hospitalists saw fewer generalist physicians than those who didn’t.  They attribute this decrease in continuity, not to hospitalists, but to other system level factors (such as PCPs in large multispecialty groups that may rotate call in the hospital).

So, we have tools to measure continuity, discontinuity has been increasing, and this increase in discontinuity does not appear to be strictly related to the use of hospitalists.  Even though it may not be caused by hospitalists, there are definitely effects of increasing discontinuity, and next time we’ll begin to look at what those might be (and thus what could potentially be improved by decreasing discontinuity).