Monday, February 25, 2008

Reasons for PAS; Prognosis in TBI; More

I'm traveling the rest of the week & this will be my only post until March, which I'm hoping will host fewer record-breaking snowfalls in the upper Midwest than January and February have.

ResearchBlogging.org1)
The Journal of General Internal Medicine has an article about family perceptions of why their loved ones pursued physician assisted suicide. It involved interviews of ~80 patients who had pursused PAS in Oregon (about 50 actually received prescriptions and 30 went through with ingestion of the lethal prescription). Family members were interviewed between 3 months and 3 years after the patient died. It's somewhat unclear to me how these people were identified: it looks like an advocacy group has a database of people who had inquired about PAS and the family members were identified that way (it doesn't seem like it was an exhaustive list of everyone who pursued to whatever extent a lethal prescription in Oregon in the time frame of interest for the study). 38% of family members approached participated in the survey.

The general demographic profile of the patients underscores previous findings that people who pursue PAS are white, predominantly male, and overwhelmingly receiving hospice services (87%). Family members perceived their loved ones pursued PAS for reasons of autonomy/control and fear of future suffering: wanting to control the circumstances of death, not wanting to be a burden in the future, fearing loss of independence and dignity in the future, and fear of symptoms in the future. The importance of current symptoms (current at the time the patient initiated pursuing PAS) were, however, minimal.

There are a lot of limitations to the study: retrospective, relies on family reporting long after the fact, and a low response rate. Nevertheless the findings are consistent with other studies examining why people pursue assisted suicide - it is very much for existential reasons - people who cannot conceive that the decline and indignities of dying could be meaningful or valuable in any way - and not because people's current physical suffering is driving them towards 'desperate' measures. I think this has importance to the debate about PAS, especially since its proponents often tout it as a necessary safe-guard against dying in agony. Besides not being true, that's also not why people seek it, and the only debate about it which makes sense to me is one that centers on rights & autonomy (the rights of an individual to decide what constitutes a good death for them by controlling its circumstance) but not a debate about preventing (physical) suffering. The vast majority of physician proponents of PAS of course are not promoting it as a safe-guard against physical suffering, but some of the patient/'consumer' advocates of it are, and aren't afraid to bring out the scare-mongering language.

I'll also note that this article calls PAS 'physician assisted death' (as does the AAHPM position statement - although I'd encourage you to click that link then actually look at what the URL of that page says!) throughout. While the article, and the AAHPM statement, are explicitly about the practice otherwise known as PAS, others use 'PAD' (or at times, 'assisted death') as a blanket term to include both assisted suicide and euthanasia, making the use of PAD strictly for PAS confusing. I agree with the AAHPM statement that the word suicide is emotionally charged but that doesn't necessarily make it inaccurate and insofar as there are others using 'PAD' to include euthanasia one wonders if changing the language in this way is actually helpful. Curious as to readers' thoughts about this. I'm sticking with PAS for the time-being.

Ganzini, L., Goy, E.R., Dobscha, S.K. (2008). Why Oregon Patients Request Assisted Death: Family Members Views. Journal of General Internal Medicine, 23(2), 154-157. DOI: 10.1007/s11606-007-0476-x

2)
BMJ has a study about developing a prognostic model for traumatic brain injury. The data were prospectively gathered and come from an international study of TBI (10,000 patients with a 20% 14 day mortality and a 37% 6 month 'poor outcome' across the cohort). Poor outcome here is defined as death or severe disability & severe disability is defined as 'able to follow commands/unable to live independently' or worse. All the usual prognostic model derivation and validation hoops were jumped through (well, most of them) and they created a couple prognostic models (a simple one and a more complex one using specific CT scan findings). Given the large size of the studythey achieved a tight fit between their derivation and validation cohorts. The simple model uses easily available data: Glasgow coma score, age, presence of major extracranial injury, and pupil reactivity upon presentation. They created a handy web-based calculator for 14 day mortality and 6 month 'poor outcome' prediction.

While death is an easily agreed upon 'poor outcome,' the 6 month outcome is more problematic. Assuming this model is incredibly accurate, and you could use it to predict a 95% chance of a poor outcome at 6 months, this still represents a wide spectrum of cognitive and functional disability (from death to persistent vegetative state to, presumably, a cognitively intact or nearly intact patient whose medical issues were sufficiently complex to prevent a return home). I imagine this is going to be insufficiently precise to be sufficiently helpful to many family members and clinicians for improving decision-making (it would be more helpful if it gave an outcome for PVS and death alone, for instance). What we need, of course, is a prognostic model which will tell your family if your outcome will be sufficiently positive that you'd want ongoing life-prolonging care after a severe TBI; this is also known as a crystal ball and one has to forgive these researchers for not coming up with it.

One of the accompanying editorials talks about the limits of the model, and cautions against using it in clinical care (instead promoting it for risk stratification of patients for further research). I think the cautions are valid, however prognosis is always limited, human, and incomplete and suggesting that we need perfect prognostic powers before making decisions is strange to say the least. A quote:

Far greater caution is needed if such a model is to be used for making decisions about treatment in individual patients. Estimates of outcome probability from the 10 000 patients in the CRASH trial are based on collective clinical experience beyond that achievable by any individual clinician, and these estimates may help educate clinicians and support clinical decision making. They cannot be used in isolation, however. Models can estimate the probability of a given outcome for a group of clinically similar patients with a high degree of accuracy, but they cannot reliably predict outcome for individuals. At least in the context of deciding whether or not to treat individual patients, it is important to continue to acknowledge, as did physicist Niels Bohr, that "prediction is very difficult, especially about the future."
Thanks to Dr. Bob Arnold for alerting me to this.

3)
And very briefly:
  • JAMA has a study about survival of in-hospital cardiac arrest. The paper uses prospectively gathered data on ~80,000 in-hospital arrests (from a national registry) and compares outcomes for pulseless cardiac arrests that occur during weekdays vs. those that occur on nights or weekends. Overall survival to discharge was 18%; 14% survived and had good neurologic outcome (how this was actually determined is unclear to me). Pursuant to a recent conversation on the blog, the best outcomes were intraoperative arrests during the day (night ones weren't too good presumably because these were emergency surgeries on unstable patients): 36%. (For the record, outcomes are modestly better during weekdays than nights/weekends.) These figures are in the ballpark of previous studies. I'm curious as to whether in the next decade or so rates of in-hospital arrest will start improving due to better patient selection (fewer resuscitations of the otherwise dying); that of course assumes such trends are actually occurring....
  • Scanning through the references from the JAMA piece I noticed this one from a few years back that I had never seen. It caught my attention because they evaluated outcomes of arrests for people with 'limited resuscitation' status (e.g. 'partial - as in not 'full' - code status: people whose code status limits intubation, shocks, or chest compressions but not all three) - I had never seen a study looking at this. None of the patients who had a 'limited code' status survived an arrest to discharge, which is at least a wee bit of evidence to support the long held adage that partial codes make no sense. A long time ago I had a (very sick, but smart and seemingly well-informed) patient who could not understand why her code status couldn't be "one shock, nothing else."
  • The Journal of Clinical Gastroenterology has a study about prognosis in advanced pancreatic cancer based on retrospective analysis of data from a Korean cancer center.
  • To the Best of Our Knowledge, a public radio show, this week has an hour long segment on the good death. It's worth listening to (you can stream it online; there's also a podcast available at iTunes).

Pallimed | Blogger Template adapted from Mash2 by Bloggermint