Well, say goodbye to hospice, you cruel, malicious heel. Zimmerman and colleagues, reporting in the April 9 JAMA (subscription required to get to the full article) found that there were 22 randomized control clinical trials evaluating ‘palliative care,’ which was defined as a service that provides or coordinates comprehensive care for patients with terminal illness. Only 4 of 13 studies found that there was an improvement in quality of life, 1 of 14 found that there was symptom improvement and 1 of 7 found that there were cost savings. Yet, palliative care is a growth industry with approximately 4,500 standing programs, of which 46% are for profit, involving 1.3 million patients. It’s been covered by Medicare since 1982, along with the vast majority of other commercial and government health insurance programs.
The Disease Management Care Blog has already fussed over the “evidence-based” double standard being applied by many of the health care illuminati to ‘population-based care approaches’ to care versus yearly check-ups, the medical home and integrated delivery systems. While that cognitive dissonance can be a source of great fun, it’s time to be serious.
The DMCB thinks it’s difficult for the vast majority of hospice programs to accommodate randomized clinical trials (RCTs) let alone concoct one. It’s simply “out of reach” in terms of their business model or available resources. What’s more, the vast majority of hospice programs leaders, their workers, their patients, the families and the physicians wouldn’t want to bother with RCTs because the questions they answer wouldn’t add to their knowledge. Sound familiar?
How can we reconcile what we know and accept about hospice versus the lack of evidence? The DMCB commends an excellent essay authored by Don Berwick in a preceding issue of JAMA (subscription again). In it, he points out RCTs are better suited for the circumscribed evaluation of tests, drugs and procedures. They fall short, however, in evaluating complex multi-component interventions that depend on ‘context’ and the ‘mechanism.’ It’s not a matter of getting to the right answers, it’s a matter of getting to the right questions.
He suggests new evaluation methodologies are necessary when the question is not if something works, but why and under what circumstances. He also recommends that a) the high threshold used to decide whether change is warranted needs to shift if the status quo is unacceptable, b) sources of bias are better accommodated than eliminated and c) academician-front-line caregiver gap needs to be bridged.
Can you think of a big health care policy debate that would benefit from this approach? Hint here.