Yesterday I posted on a JAMA study that showed that critical access rural hospitals (CAHs) have higher adjusted mortality for 3 common conditions. 30 day mortality for patients admitted for AMI, CHF and pneumonia remained higher in CAHs even after controlling for fewer resources (less likely to have an ICU, cardiac cath lab, Electronic Health Records), reduced clinical personnel, membership in a hospital or health care system, and degree of rurality.
The authors note that they were unable to control for patient health behaviors that are known to be higher/worse in rural areas (obesity, smoking, exercise). Further, in the Medicare program, patients have absolute choice of what hospital they use for health care services, and there are always worries about patient selection effects (patient choices being related to mortality in ways that are not obvious). Of course in the case of emergencies (AMI is one of the conditions studied) choice may be irrelevant. Do these unmeasured individual variables explain the observed mortality differences?
Typically when you have an observational study and concerns that unmeasured indivdiual variables could explain the differences in observed outcomes, the holy grail is a randomized control trial (RCT) to determine whether outcomes and quality are truly worse in CAHs. However, in this case, I don’t think a RCT could really provide a meaningful answer (which is good news, because you could never do the trial). The reason is that people don’t live in clinical trials. Some of them live in rural areas–about 60 Million Americans. They choose to live there and we have to have a health care system that works for such citizens.
There are two types of questions that need to be addressed to inform policy:
- Can we determine whether unmeasured patient characteristics explain the outcome differences? I am unsure of how to best investigate this question, especially on a national level. We likely need more targeted micro studies in states or parts of individual states.
- If patient factors do not explain all of the differences, how big of a outcome/quality difference are we willing to tolerate to allow CAHs to remain in rural America? I have some thoughts about this that I will address tomorrow.