Friday, October 23, 2009
NEJM has an important paper about the natural history of advanced dementia. The data come from a prospective study (2003-7) of patients with advanced dementia residing in 22 nursing homes in the Boston area. 'Advanced' dementia here means patients with a Cognitive Performance Scale (part of the MDS) of 5 or greater. A CPS of 5 corresponds to a mean Mini-Mental State Examination of 5 (+/- 5 SD): ie, the vast majority of these patients then had a MMSE of less than 10. Inclusion criteria also included 'profound cognitive deficits (inability to recognize family members), minimal verbal communication, total functional dependence, incontinence of urine and stool, and inability to ambulate independently.'
The data were collected in a variety of ways in the study: after enrollment incident complications were determined by chart review; pain and dyspnea, agitation, and aspiration were determined by regular interviews with the patient's NH care providers and nurses. One of the categories they measured by chart review was 'eating problems' which they defined as: 'documentation of weight loss, swallowing or chewing problems, refusal to eat or drink, suspected dehydration, and persistently reduced oral intake.'
~320 (56% of patients who met the inclusion criteria) were enrolled: mean age 85 years, 89% white, 85% female, most with Alzheimer's-type dementia. The median time subjects had lived in a NH was 3 years.
They present some general survival data, as well as data on 'complications' during the study. Median survival was 478 days, with 54% of patients dying in the 18 month study window, and 25% dying within 6 months. So - a sick group of people - although this doesn't tell you much more since patients were enrolled regardless of how long they had dementia or had lived in a NH.
But things get more interesting. Many patients had dementia associated complications, which were associated with worse prognoses. 41% had pneumonia (leading to a 47% 6 month mortality after an episode of pneumonia), 53% a febrile episode (45% 6 month mortality), and 86% had an eating problem (38% 6 month mortality). These mortality rates were much higher when compared to patients who did not have those 'complications.' In fact the 500 day survival for the (small number of) patients who had no eating problems was over 90%, suggesting to me that eating problems are the hallmark of the 'terminal phase' (exceedingly likely to die within 2 years) of dementia.
Pain, agitation, and dyspnea were all quite common, 'burdensome interventions' (parenteral therapies, hospitalizations, ED visits, tube feeding, etc.) occured in 40% of patients in the 3 months prior to death. Only 30% of patients who died were enrolled in hospice. Which, yes, seems appalling but also reflects the reality that 'hospice-like-care' frequently occurs in NH for patients with dementia without formal hospice enrollment.
Truly appalling was the only 18% of proxies who reported receiving prognostic information from a physician, and only 32% reported receiving counseling about the expected complications of advanced dementia. Breaks your heart. Despite this, about 80% of proxies reported that they thought they understood the expected complications. Subjects whose proxies reported knowing the expected complications of dementia received fewer burdensome interventions in the last 3 months of life, as did those who thought their loved one had less than 6 months to live. They note that reporting a prognostic conversation with a physician did not make any difference in rates of burdensome interventions that they could tell: it was the expectation of a short prognosis and the knowledge itself of complications (regardless of where that knowledge came from) which were associated with the different outcomes.
So, further corroborating data that knowledge of prognosis (in its broadest sense - not just time but what is expected to occur) changes care at the end of life. What excites me about this and similar findings (and there has been a fair amount of research indicating that prognostic knowledge is associated** with better care at life's end) is that it is something we can change. Or, it seems like something we can change, as at least these data indicate we continue to do a piss poor job of telling people what is likely to happen to them or their loved ones. Frankly, prognostic knowledge is one of the only things I know of which has routinely been shown to effect end of life care for the better (I expect to be shouted at in the comments for this statement) which makes me wonder that all the fancy palliative/end of life care quality measures that have recently been promulgated should be reduced to: tell patients what the future likely holds for them, offer options.
**Getting off my bandwagon for a moment, one should reflect on the 'associated' there. As with the Coping With Cancer Study (see 7th paragraph down), there is always the possibility here that this association can be accounted for not by causation (prognostic knowledge --> different decisions --> better end of life care) but by a confounder which makes patients/families much more open to talking about prognosis/reporting prognostic knowledge/conversations/impressions. That is: confounder --> more likely to engage a clinician about prognosis and/or report in a research environment prognostic knowledge + also more likely to choose 'gentler' end of life care --> better end of life care. What that confounder is I'm not sure exactly but I wonder if it's some generalized 'more comfortable with death and dying and talking about it.' The question I'd like to see answered is one that can get at causation. That's unlikely to happen in a controlled trial (at least with real-life terminally ill patients) of course.
All this aside, the more data we have showing that prognostic knowledge changes outcomes (improves them) the better, and the studies like this not only sharpen this knowledge for clinicians but for patients as well (see here).
See also this associated and very pro-palliative care editorial.
Our pals at GeriPal (funny) blogged about the same article as well and talk a little about the Medicare Hospice Benefit and these findings.
The same NEJM issue also had a fascinating study on dialysis in older patients and next week I plan on blogging on that and another paper on dialysis in advanced age. Mag Citrate will return; I know how much everyone loves it so.