Friday, September 30, 2011

Quality Outcomes in Palliative Care, The Swedes, and The Dawn of ACO Patient Outcomes

by Lyle Fettig

The Journal of Clinical Oncology published a Swedish study which examines correlation between patients being informed of imminent death status and other palliative care quality outcome measures. In the study, 91% of patients were informed of their imminent death status.  How was this measured?  Sweden has a national palliative care registry which is described in the study text (click the little Union Jack at the top right of the page to view the front page in English) :

This online register was completed by the physician and/or nurse responsible for the patient's care during the last week of life, although HCPs were not aware of the research question of this study. The design of the SRPC Web questionnaire does not allow any missing data at submission. The Web questionnaire is based on the 11 principles constituting good death as defined by the British Geriatrics Society and was designed to be used in any care setting, independent of diagnosis, with the intent to look retrospectively at important aspects of care delivered during the last week of life. The 25 items on the SRPC include 14 questions that cover background data of patient and reporting unit and three questions that cover items not considered relevant to the aim of this study (ie, use of visual analog or numeric rating scales during last week of life, frequency of pressure ulcers, and staff reflection). The remaining eight questions covering, for example, degree of autonomy, information about imminent death, symptom control, parenteral as needed (ie, pro re nata or PRN) prescriptions, and preferred location at time of death, were included in the analysis in this study.
The ACP Hospitalist Blog commented on this study in more detail right after the study was released.  I'm going to keep my comments on the results brief.  The bottom line was no surprise.  Being informed of imminent death status correlated positively with more prn symptomatic medication orders, family being informed of imminent death status as well as being offered bereavement services, and patients recorded as dying in their preferred location.  Being informed did not correlate with increased anxiety or pain.  Minor methodological issues aside, this is yet more data (need we more?) to suggest that we should get over the worry about the effect of prognostic information on patients, even at the very end. When we don't sensitively explore patients' wishes for prognostic information and then give them info when they want it, it's usually because of our own discomfort rather than a protective effect for the patient.

What really strikes me about this study though is the existence of the database which made the study possible.  Considering the fact that I would struggle to access many of these data at my own institution, it's amazing that Sweden has a national database and that over 13,000 subjects could be included in the study.  Sweden is not unique in their attempt to measure various outcomes on a population level near the end of life.  See this Canadian study looking at the trajectory of performance status and symptom scores in patients with cancer in the last six months of life. 

Would something like this be possible in the United States? Could we develop a common repository for palliative care outcome measurement for individual patients? What are the true patient-centered outcomes which matter most to our patients and families when confronted with serious illness? 

These questions don't exist in a vacuum and should be viewed within the context of the Accountable Care Act and evolving healthcare system.  The Department of Health and Human Services recently released final rules for Accountable Care Organizations.  The rules document describes 33 quality outcomes which ACO's will need to report in order to qualify for "shared savings" above and beyond Medicare Part A and B reimbursements (see page 324). In the first year of the program, merely reporting all 33 outcomes will be sufficient to qualify for the "shared savings" payments.  In subsequent years, the ACO will need to meet a certain standard to qualify.

Take note of the first ten quality outcomes (all but the last three will be measured by patient survey data):
  • Getting Timely Care, Appointments, and Information
  • How Well Your Doctors Communicate
  • Patients' Rating of Doctor
  • Access to Specialists
  • Health Promotion and Education
  • Shared Decision Making
  • Health Status/Functional Status
  • Risk-Standardized, All Condition Readmission
  • Ambulatory Sensitive Conditions Admissions: Chronic Obstructive Pulmonary Disease
  • Ambulatory Sensitive Conditions Admissions: Congestive Heart Failure
The majority of these outcomes are near and dear to palliative care- don't we aim for these day-in, day-out? Palliative care teams can directly influence all of these outcomes for individual patients but also indirectly through local palliative care educational and quality improvement projects.  What other measurable palliative care outcomes correlate with each of these outcomes.  Could some of those other outcomes become standard in future ACO rules?  If you want to help the field (and the American healthcare system) take one small step in that direction, make sure your program applies for TJC Advanced Certification.  In the process, your program will be required to collect/analyze at least four palliative care performance measures on a continuous basis.  High penetrance of TJC Palliative Care Advanced Certification may lead TJC to build palliative care standards into the usual accreditation process for hospitals, etc.   

While I'm rambling on about everything under the sun, the Swedish Palliative Care Registry has a second questionairre that is filled out yearly at a program level (rather than specific to individual patients).  We do already have a registry similar to this: The CAPC National Palliative Care Registry.  Your program should sign up for that, too, if it hasn't already.

Pallimed | Blogger Template adapted from Mash2 by Bloggermint