Saturday, January 19, 2008
(Editor's note: this post was originally published Thursday Jan 18 but was never picked up via the email update feed so I reposted it on Saturday for those who rely on the email for Pallimed updates. Thanks/Sorry.)
Archives of Internal Medicine has an article about the development of a prognostic index for COPD. The data came from 12 randomized drug trials for COPD involving ~8800 patients; this analysis extracted data from those trials (mortality, hospitalizations, etc.) and created several prognostic indices from them (they used the first 2/3 of patients temporally enrolled in each trial as the derivation group, with the final 1/3 as the validation group). The methods are quite complicated but essentially were: they identified significant patient characteristics in the trials which predicted death, hospitalization, or COPD exacerbations (e.g. body mass index, FEV1, age, health related quality of life); they made 3 separate indices to predict those outcomes as well as a composite index of all 3 indices; then validated these indices separately for the individual outcomes, as well as the composite index for the individual outcomes. With some tweaking they more or less found that the composite index was pretty decent at predicting all 3 outcomes and the final index they present gives likelihood - at 3 years - of death, hospitalization, as well as expected number of exacerbations.
How useful this is going to be for palliative medicine professionals is unclear. The final index contains mostly readily available information (age, BMI, FEV1, history of cardiovascular disease, gender, history of ED visits) but also needs quality of life data from one of two quality of life indices (they provide a web-based calculator for the index, albeit one that is not currently functional, and it's unclear if the calculator will assume one has access to the QOL indices or not). (The other recent COPD prognotic index - the BODE index - also contains a less than routinely available element - the 6 minute walk test.) Also, similar to the BODE index, this one is unable to predict short term mortality (and, in fact, the highest possible score on the index predicts 'only' a 31% 3 year mortality). I don't want to sniff at a 1/3 3 year mortality, but I think we are all hoping for a reasonably accurate way of identifying patients likely to die within one year, and this is not it. I've done some philosophical pondering about COPD in the past on this blog, although I can't find the post currently, but I have long wondered if COPD is a sufficiently inherently unpredictable disease and it might be a positivistic folly to think we can systematically do better than this (for individual patients, at times of course, we certainly can).
That aside, there is good reason to believe this index overestimates prognosis for COPD patients, as all subjects involved were healthy enough to participate in a drug trial, and it would be helpful to see further validation (more like re-calibration) studies in more real-life COPD patients, particularly those which HPM professionals tend to see (applying it to a series of patients at the time of hospitalization for a COPD exacerbation, for instance). Despite the relatively good survival of the worst group, they were predicted to have 9 exacerbations in that time: these are patients in need of help, support, and advance care planning.
Briggs, A. (2008). Development and Validation of a Prognostic Index for Health Outcomes in Chronic Obstructive Pulmonary Disease. Archives of Internal Medicine, 168(1), 71-79.(The same issue also has an interesting discussion of whether participants in phase I cancer studies should be considered 'vulnerable' or not. The authors conclusions are more or less that most phase I patients are white, middle-class, English speakers with good performance statuses and therefore can't be considered vulnerable.)
Annals of Internal Medicine has published an 'Update' in palliative medicine, as well as a couple other palliative care reviews/guidelines.
The update, which is to the best of my knowledge the first one ever published by Annals for palliative medicine, is part of an ongoing series Annals publishes on all the internal medicine subspecialties/disciplines (cardiology one month, hospital medicine the next, etc.). The intended audience is non-palliative care internists/subspecialists, and it is great to see HPM included in the mix. Maybe ACP Journal Club will be next.
The first guideline is one on improving pain, dyspnea, and depression at the end of life. The second is a systematic review of improving palliative care at the end of life. The first is a fine one for the teaching file, although there's likely not much new there most HPM professionals. The second is a good, broad overview of improving care at the end of life, and is well worth a read. I found this paragraph particularly interesting about how we even go about defining 'end of life' care:
"The literature used various approaches to identify patients at the end of life. Some used clinician assessment of "active dying" or "patient readiness," but no precise definitions or performance characteristics of these terms have been published. Many studies used specified clinical characteristics, survival prediction rules, or physician judgment. Although prognostic tools usefully characterize subpopulations (for example, heart failure), many patients with fatal conditions have substantial probabilities for 2- or 6-month survival, even in their last week of life. Patients with metastatic cancer, who have an estimated 10% or greater chance of dying within 6 months, are more likely to prefer to avoid resuscitation, even when survival was much less likely than they acknowledged. Thus, clinicians might define the end of life as having a fatal condition, risking death with the next exacerbation, or beginning to acknowledge the seriousness of the situation. Asking clinicians "Would it be a surprise if this patient were to die within 6 months?" is being used widely but also has had no rigorous testing. The studies emphasize that acknowledging death risk is important for decision making. Waiting for near-certainty would fail to identify most dying people, so palliative approaches need to be regularly incorporated for people living with serious illnesses."
The same issue also has a prospective study on the prevalence and severity of pain in adults with sickle cell disease. Disturbing findings.
American Journal of Medicine has a report on sudden death and methadone use. It's based on an autopsy series comparing findings in patients with sudden death who had methadone in them at the time of death (known overdoses were excluded) with those who weren't taking methadone and found that cardiac abnormalities (e.g. coronary artery disease) was more common in those not taking methadone (suggesting increased rates of death from cardiac arrhythmias in the methadone group). The level of evidence to support this conclusion is weak (based on the study design), but it is still another cautionary study about methadone (there has seemingly been a half dozen or so of these in the last couple years).
A recent Current Heart Failure Reports has a casual article about long-term use of inotropes in 'end of life heart failure care.' It is a general discussion of their use, but also discusses a single institution's experience with home inotrope use specifically in a palliative/end of life context. While the paper doesn't formally present data, it looks like however this program wasn't terribly successful in keeping patients at home: 36 patients, mean survival 3.2 months, 46 hospitalizations (7 for line sepsis), and 42% died in the hospital. (To be fair of course it may not have been the patients' or clinicians' goals to keep the patients out of the hospital, and while I don't have any ready figures I wonder if 42% in-hospital death is relatively low for CHF....).
And in case you missed it here is the info on the Pallimed Happy Hour at AAHPM:
The '2nd Annual Pallimed Happy Hour' will be at Backjack's, Friday February 1st at 7:30PM until at least 8:30PM, but could go longer if anyone hangs around. Map here - note it is just a short, tree-lined bridge south of the convention center (601 S Harbour Island Blvd - Backjack's is the bar associated with a restaruant - Jackson's Bistro).