Saturday, November 20, 2010
For those of you wondering what to do with all of your free time now that the boards are over, I highly recommend immersing yourself for an hour or two in the maps, graphs, and tables of the Dartmouth Atlas’ first comprehensive report on end of life care for patients with advanced cancer. The premise of the Dartmouth Atlas is that the large local variations in health care services which Dartmouth has documented, are caused by supply (more surgeons=more surgeries; more hospital beds=more hospital days) and by local patterns and customs of care, NOT by evidence-based care delivered to achieve better outcomes. The Dartmouth Institute researchers assert that
“Our findings point to important opportunities to improve not only the quality of care (by ensuring that effective care is reliably delivered) but also to reduce the costs of care (by reducing avoidable hospitalizations and unnecessary specialist visits).”
For years, the Atlas has had a number of measures of end-of-life care and of care of chronic illness in the last two years of life (all for Medicare patients, since the data comes from Medicare files.) The newly released report is the first to examine care specifically for Medicare cancer patients with a poor prognosis, looking at ten measures of care in the last month of life such as percent of patients admitted to ICU in the last month of life, percent of patients admitted to hospice during the last month of life, and percent of patients admitted to hospice within three days of death.
I don’t think the major findings will surprise any Pallimed readers:
1. Across many regions and academic medical centers, over one third of patients with poor prognosis cancer spent their last days in hospitals and intensive care units. A significant proportion of patients received advanced life support interventions such as endotracheal intubation, feeding tubes and cardiopulmonary resuscitation (CPR).
2. The use of chemotherapy in the last two weeks of life overall was about 6% of patients, but in some regions and academic medical centers the rate exceeded 10%.
3. The use of hospice care varied markedly across regions and hospitals. In at least 50 academic medical centers, less than half of patients with poor prognosis cancer received hospice services.
4. In some hospitals, referral to hospice care occurred so close to the day of death that it was unlikely to have provided much assistance and comfort to patients.
What may surprise you is just how large the variation really is from area to area. For instance, the percent of patients admitted to intensive care varied more than sevenfold across hospital referral regions and the number of days in ICU in the highest use regions were more than 25 times higher than the number of days in the lowest regions. Hospice use in the last month varied three-fold. While some of this variation may be due to appropriate matching of care to regional variations in patient preferences, I suspect that most of it is driven by the local culture of medicine and health care. I hypothesize that more effective delivery of patient-centered and patient-preference driven care will actually reduce variation in these measures – but perhaps that can be the premise for a few performance improvement projects in some of these regions with (can I say it) shockingly poor performance on these measures.