Physicians’ Influence on Patient Use of Hospice

End-of-life care has become controversial, partly due to long-standing cost concerns and partly due to ludicrous assertions that the reform law created “death panels”.   Hospice is an increasingly recommended and used option for end-of-life and/or palliative care.  Research in Health Affairs finds that use of hospice appears strongly tied to physician preferences and other characteristics.   (Health Affairs Article)   The authors first note that patient preferences and characteristics appear to account for only about 10% of variation in hospice use.  They then attempted to examine the influence of physicians, using data on about 200,000 Medicare beneficiaries, around 132,000 of whom ended up in hospice, with a poor prognosis for cancer in the years 2006 to 2011.  Physician and patient characteristics were analyzed for association with hospice use.  The physician characteristics included percent of patients with these prognoses using hospice, specialty and hospital economic status.  The patients who ended up in hospice were more older, more likely to be white and female and to be higher income than the non-users.  Use of inpatient, ER and home health services was similar between the two groups and they had similar disease burdens.

There was substantial variation in hospice use across hospital referral regions but no clear regional patterns.  Hospice use generally increased over the study period.  Physicians who seemed to practice mostly at non-profit hospitals had more patients utilize hospice than did those at for-profits.  Medical oncologists had higher numbers and percents of patients enroll in hospice than did internists, family practitioners, medical subspecialists or surgeons.  Doctors who had higher proportions of patients with poor prognoses were, understandably, more like to have greater numbers use hospice.  But by far the strongest factor was the overall percent of a physician’s poor prognosis patients that went to hospice.  In other words, the more of doctor’s prior patients in this category went to hospice, the more likely the next one would as well.  The results would indicate, as would seem logical, that a doctor’s views on hospice use strongly affect the patient’s decision to choose that option.  No attempt was made to understand why some physicians seem to more strongly favor use of hospice, and that would be an interesting followup.