Thursday, May 7, 2015
Measuring the Value of Hospice
by Kristina Newport
Hospice is better care at a cheaper price, right? Prior studies support cost savings but are critiqued for methodological problems. The most problematic is selection bias: hard to know whether cost savings are truly due to hospice enrollment or, instead, due to the patients’ underlying preference for less aggressive (and cheaper) care.
Hospice is better care at a cheaper price, right? Prior studies support cost savings but are critiqued for methodological problems. The most problematic is selection bias: hard to know whether cost savings are truly due to hospice enrollment or, instead, due to the patients’ underlying preference for less aggressive (and cheaper) care.
Gozalo et al attempted to clear up this issue in their May 2015 NEJM
paper that rigorously evaluates Medicare cost changes in nursing home
patients with and without hospice. They
apply a new statistical analysis, difference-in-differences cross-temporal
matching design (which definitely was not covered in Statistics 101) in attempt
to more accurately compare hospice and non-hospice users, without selection
bias. Since this is not a statistics
blog and I’m sorely unqualified to do so, I will not argue the merits of this
method but will instead assume that it is valid. At the very least, it seems to better
approach the goal of comparing apples to apples than previous studies.
The findings of the paper are striking enough that they were
broadcast in the popular press on the day of publication:
Hospice in the nursing
home does not save money but instead
costs an average of $6,761 more per
patient.
Not to be overlooked are other findings in their study
showing that hospice patients had less hospitalizations, fewer feeding tubes,
less ICU time and fewer burdensome transitions.
Hospice use in the nursing home has significantly increased,
doubling from 23% to 47% from 2000 to 2012.
Most of this increase occurred due to increased enrollment of patients
with non-cancer diagnoses who have longer length of stays(LOS) and
harder-to-predict prognoses. In this
study, LOS increased from 72 days to 93 days, in contrast to other instances of increasing short-stay enrollments.
Initially, these data are surprising, and maybe even
disappointing. Those of us who routinely
enroll nursing home patients in hospice do so with the belief that it is the
best option not only for the patient (which remains well supported by the evidence
in this paper) but also for “the system”. The question now at hand is, what is the
balance of patient benefit and system cost?
Is it really all
about the money?
Is hospice still the right thing even if it does not save
money? At its inception, the Medicare hospice benefit was intended to provide
high quality end of life care, but was also expected to save costs or be budget
neutral. As appreciation for the benefits of quality end of life care grows,
should cost savings remain a primary goal?
When do we stop making the argument that we have to do more for less
and simply argue that we need to do more?
Although the triple aim of
improving care experience, population health and per capita cost remains
important, the balance of the three factors needs to be maintained. As Jon
Keyserling states in the NHPCO
response to this article, “cost savings should always be secondary to the
goal of delivering high quality care”. Good
care is not always the least expensive care- do you ever hear anyone touting
cost savings when promoting the latest targeted cancer treatment or surgical
cardiac procedure? Patients and families never
choose hospice with a goal of saving money for the system. They choose hospice care because it meets
their needs. It’s great that hospice and
palliative care can save money in some settings, but even when it doesn’t,
maybe it’s still the right thing.
The population of people we are really examining is the
group for whom hospice services have been “expanded”, mostly nursing home patients
with dementia, without cancer. The
finding that hospice may be less cost-efficient in this group could be a
stimulus to develop a unique program of care for this population: one that can achieve
the goals of avoiding aggressive care and addressing patients’ unique priorities,
but with a model different than hospice, perhaps a less costly one. May this be
a challenge for the field: Will one of
you complete a study to examine which elements of hospice are beneficial to
this patient population? Are there additional
helpful interventions that are not part of the hospice benefit?
The Bottom Line:
Hospice enrollment of patients in nursing homes does not
result in Medicare cost savings, due in large part to patients with non-cancer
diagnoses who have long lengths of stay. But, hospice enrollment does decrease aggressive care and
burdensome transitions at end of life.
What’s Next?
-
Will experts weigh in on the statistical tool
utilized in this study and confirm the validity of the findings?
-
Can we develop better measures to compare the
benefits of hospice to the costs and make the case that, at times, hospice care
remains the best care even when costs increase?
-
Is there a different program that would be
better suited to care for patients with dementia or other non-cancer diagnoses
that offers similar benefit without the same daily cost of hospice?
For
now, we’ll keep taking care of patients with the programs that best suit their
needs and continue to work to prove that it is the right thing to do.
Graphiccredit: fistfuloftalent.com
