Saturday, February 18, 2006
Archives of Internal Medicine has just published an article on differences in health care utilization at the end of life between lung cancer patients and those with COPD. The study--which took place at multiple VA's--compares health care utilization in the last 6 months of life between COPD and lung cancer patients. They found that COPD patients overall use subtantially more resources, mostly due to a dramatic increase in ICU use compared to lung cancer patients. COPD patients were 5 times more likely to have prolonged ICU stays.
They also looked at opioid and benzodiazepine use and found that the COPD patients were much less likely to receive either of these medications, raising the possibility that they were getting less symptom palliation. That, of course, is an impossible conclusion to make based on pharmacy records.
Studies like this are tough because their findings are useful to us in the profession--the idea that patients with COPD are getting too much of one type of care and too little of another type of care at the end of life appeals to us. More palliative care for everyone, right? While I believe the world does need more palliative care (!), and that certainly too many people with COPD don't get the right kind of care at the end of life, articles like this generate more heat than insight. Part of the problem is that underlying articles like this, or at least their interpretation, is an assumption that this extra money spent is 'inappropriate.' Too much aggressive care on these dying people. The problem is that at the point many--not all, but many--of these patients went to the ICU it was unclear that they were going to die in the next 6 months. Many people, even with advanced COPD, go to the ICU, get better, and live for years, and would want to go back to the ICU again. We have indices to predict prognosis in COPD but they are only helpful for distinguishing people likely to die in the next couple years from those likely to live many years. Contrast this to the case with lung cancer when it is much easier to get a general sense of someone's longevity. So when you take a cohort of dead people and examine the 6 months prior to them dying--for most of the COPD patients you wouldn't have been able to tell who would have died at the beginning of those 6 months--whereas for most of the lung cancer patients you would have. They are different patient populations, and comparing them is useful for rhetorical reasons only.
The one good use for these sorts of articles is convincing policy makers and money people that good end of life care is in their interest, too.
The current BMJ has an interesting little study suggesting that sham devices may have more of a placebo effect than placebo pills. Actually, the evidence presented for this is weak, but an intriguing study for those of us interested in these things.