Post details: Hospital Transparency and Mortality Rates

Google
 

Hospital Transparency and Mortality Rates

December 31st, 2007

by Nick Jacobs

From the local paper, " Hospital A's mortality rates improved, while Hospital B, Hospital C and Windber Medical Center received poorer marks. The truth, the whole truth and nothing but the truth, . . . or so it would seem.

As many of you know, we are very passionate about transparency in health care and, as a Planetree Hospital, we are committed to demystifying health care. One of the problems that we face, however, is a nuance problem.

We have a palliative care unit, a hospice that is utilized by a five county area. It is a center that provides pain control, respite and end of life care. My opinion of this service is that every hospital should offer it to every family. Bottom line though is that each year, our hospital is penalized statistically because of the number of deaths that occur. Why does this happen? It is because the conventional thinking in this country is still that, even if a patient has a do not resuscitate order, death is not acceptable and must be statistically noted as something bad.

Consequently, even though patients elect to come to our unit to die surrounded with dignity and peace and embraced by their loved ones, their death shows up as a negative State statistic without differentiation.

If the terminal patient was there because of heart failure, the ultimate end of that condition is not life, it is, in fact, death. Unfortunately, the statistics will show an inordinately high number of deaths for heart failure in the category that graphically depicts our medical center's death rate. Then the newspaper will cover this statistic, and we will, once again, attempt to respond to the public by explaining what hospice services are and how their impact on our numbers should be calculated.

As Ronald Reagan once said, "Here we go again."

Transparency in death rates must be carefully monitored so as not to penalize those facilities that help families by providing transitional hospice services. We know of some heart centers that will not operate on patients with high risks because it will skew their statistics.

Numbers can do whatever you want them to do, and we want them to be honest and carefully depicted to demonstrate truth and clarity. We're not blaming the press, but we are, once again, questioning the Health Care Cost Containment Council's mechanism for production of these statistics.

Comments:

Comment from: Terence Coughlin [Visitor]
Excellent points. Here in New York, the majority of mortality-based measures/reports utilize data from the Dept. of Health's statewide administrative discharge dataset (AKA "SPARCS"). Hospitals in NY are instructed to only submit the acute portion of any inpatient stay to DOH. Patients transferred to in-hospital hospice beds are administratively discharged, and anything that occurs with those patients beyond that point is essentially invisible from the state data submittal point of view. Hence, deaths occuring in hospice do not count against the hospital's mortality rates.

This wasn't always the case, and many NY hospitals battled the same issue of mortality perspective as you wrote about - facilities with hospice were being penalized in mortality scores for the services they were providing for end-of-life care.

I'm not familiar with the Council's specific methodology, but are they not collecting any information that would allow a hospital to identify hospice patients, or adjusting their measures to account for this? Having to explain yourself year in and year out after-the-newspaper-article fact must no doubt be maddening.
Permalink 12/31/07 @ 12:27
Comment from: Judy Volkar MD [Visitor]
"There are lies, damn lies and statistics"
Disraeli
Permalink 01/05/08 @ 12:14
Comment from: Bill Walker, MD [Visitor]
There are only 2 ways to improve actual mortality rates. When there is a process problem (not the case for hospice care) that should be addressed. The other way is to 'get them off the books' (transfer so they are invisible as in the comment by Coughlin above).
There are only 2 ways to improve expected vs actual mortality rates. Lower actual as above, or raise expected through more explicit documentation resulting in codes that convey greater attributable mortality. Unspecified codes default to the lowest risk category.
What is most maddening about observed versus expected "quality" scores is that while they almost certainly represent communication/ documentation failure, clinical failure is assumed.
Permalink 01/17/08 @ 14:30

Leave a comment:

Your email address will not be displayed on this site.
Your URL will be displayed.
Allowed XHTML tags: <p, ul, ol, li, dl, dt, dd, address, blockquote, ins, del, a, span, bdo, br, em, strong, dfn, code, samp, kdb, var, cite, abbr, acronym, q, sub, sup, tt, i, b, big, small>
URLs, email, AIM and ICQs will be converted automatically.

authimage

Options:
 
(Line breaks become <br />)
(Set cookies for name, email & url)

Google