Tuesday, May 11, 2010
A Randomized Controlled Trial of "Death Panels"
...or rather, a randomized controlled trial of an intervention very similar to Respecting Choices, which is the program developed by the Gundersen Lutheran Medical Foundation in Wisconsin. (See Christian's post here for more on relationship between the Gundersen Lutheran program and the death panel "controversies.")
Researchers in Melbourne, Australia conducted a single institution prospective randomized controlled trial to investigate the impact of this advance care planning (ACP) intervention, (Respecting Patient Choices) on end of life care in older patients. BMJ publishes the results.
The intervention was delivered in a hospital setting primarily by a nurse or allied health worker who conducted a conversation to evoke the patient's goals, values, and beliefs, and to discuss and document their future choices about health care. Family members, especially potential surrogates, were encouraged to participate. The creation of advance directive documents was encouraged. In this study, the discussions lasted a median of 60 minutes spanning 1-3 visits. The intervention appears similar to POLST programs in the United States.
The participants included 309 medical inpatients aged 80 or more who were followed for six months or until the time of death. Of the 871 patients screened, reasons for ineligibility included a lack of decision making capacity (n=368), expected death during the hospitalization, previous advance care planning, no relative, non-English speaking patients, and the six patients who refused to participate.
At the time of enrollment, intervention and control groups were similar with respect to age (median 84-85), reason for hospital admission, code status, presence of appointed surrogate, and overall preference for life prolonging treatment. There seemed to be a trend towards having more males in the intervention group (54% vs 41%), but p values aren't reported.
Of the 154 patients assigned to the intervention group, the ACP intervention was delivered to 125 (81%). Of those receiving the intervention, 70 completed a written document and 38 expressed wishes verbally (which were then recorded by the facilitator for chart documentation).
The primary outcome examined was the proportion of patients who died at six months and whose end of life wishes were known and respected by their doctors. Compliance was determined by review of patients' charts after death to identify any documentation of patients' wishes, what type of care patients received, and how the patient's wishes impacted medical decisions. It was determined that of 25/29 patients who died in the intervention group, wishes were documented and followed versus 8/27 in the control group (p<0.001).
Yes, six month mortality was...ahem... no different between the two groups.
Even with attempts at blinding to intervention, it was probably evident upon chart review which patients had received the intervention. Whether this is problematic becomes irrelevant in my mind after reviewing some of the secondary outcomes:
- Reduced risk of PTSD in survivors as evidenced by a reduced Impact of Events Scale scores in intervention group family members (including 0% vs 15% who had scores >30, which places them at higher risk of developing post-traumatic stress disorder, p 0.03)
- Reduced Hospital Anxiety and Depression Scale scores (including 0% in intervention group having a depression or anxiety score >8, which indicates clinically significant depression or anxiety, vs. 30% and 20%, respectively, in the control group).
- No patients died in ICU in intervention group vs. four patients in control group. Otherwise, the intervention didn't significantly impact place of death (hospital vs. home vs. palliative care/hospice setting).
- Family members reported greater satisfaction in a survey regarding the patient's "quality of death" (83% very satisfied in the intervention group vs. 48% in control group).
The secondary outcomes are quite reminiscent of the French ICU study on physician-family communication in the ICU that showed a reduction in risk of PTSD, anxiety, and depression in family members in the intervention group. (See Drew's commentary on that study here.) When combined with the many recent Coping with Cancer study results and the fact that results are now coming in from a variety of countries (US, France, now Australia), we should be confident in these secondary outcomes.
The fact that this is not a physician-led intervention should be highlighted. Physicians were drawn into the discussions to help clarify diagnosis, treatment options, and prognosis, but they were not central to the process. This is an important consideration as we confront workforce issues in palliative care and is an example of a "disruptive innovation" in medicine. By no means does this diminish physician involvement in expert palliative care teams nor does it relieve other physicians of their duty to be mindful of advance care planning.
Also notable is that this is not a primary care, office based intervention. It makes sense in some ways that the hospital would be the locus for this type of intervention rather than an office. Patients are captive in terms of having time to spend thinking about this and they are also more likely to see the importance of it (ie in an office, "why are you bothering me with this now? I'm just here for a cold.") Family members are more likely to be around in the hospital. Patients who are most likely to benefit from the intervention are "batched" in the hospital making it easier for those applying the intervention.
Only about 1/6 of patients had died by the six month mark. I'd be quite interested to know if the results hold as a greater proportion of patients die. If so, it only strengthens the above conclusions in my mind. If not, it would be interesting to know why (patient preferences changed? conversations need to be reiterated periodically?)
In the article's conclusion, the authors discuss the fact that advance directives were once much ballyhooed and more recently have been derided as ineffective at changing care at the end of life. They propose reasons why their intervention may be more effective than other previous attempts:
"We specifically focused on a model with five key elements identified by others as crucial to successful advance care planning: trained facilitators, patient centred discussions, involvement of family in discussions, correctly filed documentation, and systematic education of doctors.....Four of these five key elements were absent from the SUPPORT study, a project that relied on a research nurse to relay information verbally to the treating doctor about a patient’s preference for cardiopulmonary resuscitation, but which did not facilitate patient centred discussions, involve the patient’s surrogate, and ensure that the treating doctor was aware of the patient’s advance care plan when end of life decisions were made."Should hospitals implement programs like this? If so, what role should palliative care teams play in developing and implementing such programs? The palliative care team is a natural home for the direction of interventions like this--these conversations are the bread and butter of our daily work, although the patient population we see is generally sicker than the population included in this study. An intervention like this may represent an opportunity for earlier formal palliative care involvement (through triggers) versus an opportunity to insert palliative principles into the patient's care through separate but related programs. Physician and non-physician leaders alike can play an integral role in the development of programs and the training of other health care professionals to conduct these conversations (without needing to insert ourselves into every advance care planning event, which is, of course, impossible). We'll still be needed for the more complex scenarios, no doubt.
A number of possibilities exist to incentivize hospitals/providers to institute interventions like this. For instance, I wonder if there's an opportunity to include this somehow in the "national pilot program on payment bundling" (included in the health care reform bill, PPACA, Sec. 3023). See Dr. Brad Stuart's recent commentary at Geripal on issues related to health care reform and palliative care.
Many locales throughout the United States that have implemented POLST interventions, and I don't practice in one of them. Feel free to leave any ideas you have below, especially if I'm missing some major demonstration or pilot project.
See also Alex Smith's thoughtful commentary on this and a related article at Geripal.
5 Responses to “A Randomized Controlled Trial of "Death Panels"”
May 13, 2010
It's an interesting trial, and along with the recent NEJM paper about ADs, it's got me wondering if my deep pessimism about them is as justified. I do think the devil's in the details with things like this - clearly having an AD is insufficient (to improve medical care at life's end, etc) - it's how it's implemented, etc.
Another thought I had about this trial is that it involved patients who were very old and hospitalized: ie this is a very sick cohort - pre-selected (by virtue of those simple demographics) to have a very high short-term mortality (compared to e.g. hospitalized 65 year olds, or healthy community dwelling 80 year olds). I wonder if doing this intervention at a time close to death is what made the impact - the AD, the discussion with clinicians and family, etc. were fresh, and timely (a patient's impending death was more apparent to those involved, at least for many patients, as opposed to being an abstraction which 'will happen to all of us one day.') Could this be part of what caused the AD process to have such a real effect on the care these patients received as they died?
May 13, 2010
Oh yeah and very happy, but not surprised, that mortality was the same. Take that death-panel-mongerers!
May 14, 2010
This study extends Bud Hammes' "Respecting Choices" experience in LaCrosse WI, a community that for over 25 years has been steeped in the culture of advance care planning. Anyone familiar with the LaCrosse results (the majority of patients showing up in EDs have ADs, and hospital physicians follow them over 99% of the time) has to wonder whether the poor results of ADs in prior studies are associated mostly with the fact that "it takes a village" to really make them work. This was demonstrated in Oregon where POLST originated, and if I remember correctly, Joan Teno's editorial in JAMA, where some of the first POLST results were reported, emphasized the fact that it's all about the community, not just the form. California is now a POLST state, and we're using it in our system-wide Advanced Illness Management program. The HCR law, and Ctr for Medicare Innovation pilots soon to come out of HHS, should allow us to parlay that experience into accountable care and bundled payment initiatives.
For years I've shared Drew's pessimism about ADs. However, I'm now betting that advance care planning in general and ADs in particular will be found to achieve better and better actualization of patient wishes in studies over the next few years. We're getting beyond the flat part of the acceptance curve, which should now slope upward at an accelerating rate, at least in aware and motivated communities.
May 14, 2010
As both of you indicate, merely filling out the paperwork is insufficient. The conversations between patients/potential surrogates and healthcare providers are key. The obvious reason cited is that if the conversation occurs, patients, family members, providers will have greater clarity regarding what the patient values and prefers. I also wonder if excellent ACP processes help to reduce the mutual pretense between patients/family members, patients/physicians, and eventually surrogates/physicians that "everything is going to be ok" and "we're going to avoid death through our wonderful modern technology." There is probably an indirect normalization of death that (aside from knowing patients' preferences) helps everyone forgo or let go of non-beneficial interventions in a more timely manner.
I wonder if there's a "tipping point" in life/disease trajectory where it begins to feel more appropriate for most people to have conversations that would lead to POLST completion. (the circumstances of death begin to appear more concrete in context of chronic/life limiting illness, the benefits of having the conversations begin to outweigh the burden of having a potentially emotionally charged conversation, etc) If so, where is the tipping point for the majority of people? not sure if it's a helpful question to answer...maybe/maybe not.
Drew- this is a sick bunch in general and they looked at outcomes in the sickest of the sick (those that died in 6 months). In my dream extension of the study, they'd randomize one arm of survivors to no further intervention and the other arm to recurrent discussions/revisiting of ACP. Then, as the whole study population dies, you could see if the results hold in both groups..
May 15, 2010
Nice work from the Austin Hospital in Melbourne. Their success with Respecting Patient Choices (RPC) led to Commonwealth (Department of Health and Aging - DOHA) funding of RPC. A nationwide 'Train the Trainer' project has been going on for a couple of years now and Australia now has many trained facilitators, although "results may vary".
For the cynics out there, I must maintain that reimbursement drives the outcomes in healthcare, 100% of the time. The Aussies have a culture of not putting seriously ill 80 year olds in the ICU to begin with. If this study were conducted here, with seasoned well trained facilitators, we may get better results with physician compliance for the intervention group, but I doubt we would EVER have the result of only 4 patients in the control group NOT ending up dying in the ICU.
I am sorry to say, but we have along road ahead of us with the integration of Palliative Care in to the culture of US healthcare delivery systems. I mean, look at the hospice benefit. We must be separate AND unequal.
Brad is right...the community (of healthcare) has a lot to do with it. A shout out to the Aussies - Coooey!
Post a Comment