Thursday, October 30, 2008
Changes in treatment preferences over time
Archives of Internal Medicine has a study looking at the stability of treatment preferences for end-of-life care over time. The data comes from the Johns Hopkins Precursors Study - a decades long longitudinal look of thousands of physicians (originally designed to to look into cardiovascular disease). This analysis looks at how answers to questions about life-prolonging treatments change over time (this is all based on answers to a single scenario asking about respondents' wishes if they had irreversible brain damage and couldn't communicate or recognize people).
The study compares answers 3 years apart (involving ~700 physicians each time; mean age at baseline 69 years; and reflecting that this was a study initiated in the 1950s in medical schools the subjects were 92% white men). They divided responses into 'most aggressive' preferences, 'intermediate' preferences (essentially would take IV fluids and antibiotics but nothing else), and 'least aggressive' (essentially would want no life-prolonging medical therapies). What they found is that treatment preferences were relatively stable over time, however: those originally who were in the most aggressive category were most likely to change (only 41% remained there after 3 years); and the least aggresive subjects were most likely to stay the same (80% did; 6% changed to most aggressive).
Age and comorbidity didn't seem to influence things much however physicians without advance directives at baseline were much more likely to move into the most aggressive category than into the least aggressive category.
I am not sure how generalizable these preferences are given that these were older white physicians and is based on a single scenario (a relatively uncommon one at that but one that tends to provoke strong responses) however it's illustrative of how a substantial minority of patients (whatever the exact numbers) change their preferences over time. I guess it's 'reassuring' that most people change their preferences to less aggressive measures, however as I read this I became very interested in that 20% of the 'least aggressives' who over the course of 3 years changed into the intermediate or most aggressive categories.
What happened there, what was driving that, etc. is something I'd love to know but is certainly unanswerable with this type of study....
Also from Archives: the previous issue has a study looking at the effect of a dementia diagnosis on cancer survival. It's an analysis of data from the SEER database, looks at tens of thousands of older patients, and has a few notable findings. First 5-10% of patients newly diagnosed with prostate, breast, or colon cancer have preexisting dementia. Patients with dementia were diagnosed at a later stage in general and had significantly worse mortality (both overall, non-cancer and cancer-related mortality). These differences persisted after controlling for things like age and other comorbidities. Gross 6 month mortality rates were as follows (with and without dementia): 25% vs 5% for breast, 45% vs 20% for colon, and 29% vs 4% for prostate. It looks like some, but not all, of the 'excess' mortality was from non-cancer diseases. One assumes that the population of patients with comorbid dementia and cancer will grow rapidly in the coming decades and, as treatment toxicities diminish (with 'targeted therapy' which, although at times less toxic than cytotoxic chemotherapy, had not been a magic bullet for cancer treatment), there will be ongoing challenges for us in the profession in helping these patients/families make decisions about cancer treatment and balancing time gained with toxicity and treatment burdens.