Sunday, July 12, 2009

Survival after in-hospital CPR in the Elderly

NEJM published a paper last week looking at CPR outcomes in the elderly. The main question of the paper addresses whether the rate of survival after in-hospital cardiopulmonary resuscitation (CPR) has improved from 1992 through 2005, and whether there are any patient characteristics that predict survival. The basic results showed that 18% of Medicare beneficiaries who underwent CPR in U.S. hospitals survived to discharge. There was also no increase in the rate of survival from 1992 through 2005, and predictors of poor prognosis included male gender, age, race, multiple coexisting illnesses, or patients admitted from skilled-nursing facilities.

In addition to its primary aim in addressing the temporal trends in survival, this paper adds to the literature out there on the short term outcomes of CPR. This outcome data is vital to have when discussing the role of CPR with patients. This paper though does not attempt to answer all the questions that arise when having these discussions with patients. Complete reliance on mortality/survivability data is not how many people decide on their preference and should not be the sole component in patient education. Preferences are heavily influenced by what the expected quality of life and functional status would be, as well as the social and financial impact that their preferences would have on themselves and their loved ones.

Survival to discharge seems like a very clean outcome for researchers, but medical decision making can not be reduced to one variable (if it could family meetings would not be as hard as they often are). How about instead of focusing on mortality data we focus on other outcome data like cognitive impairment, functional status (ADLs/iADLs), or even 3/6/12 month mortality outcomes???

1 comment:

ken covinsky said...

This is a very interesting paper. I was surprised CPR survival does not seem to have improved since 1992. I dont think the effectiveness of in hospital CPR has improved, but I would have thought that we were better at selecting patients for CPR---specifically avoiding CPR in patients who in whom the chances of survival are nil.

From a research methods point of view, I think the most important caution is that CPR was identified from procedure codes-- ie, whether a hospital noted on the billing claims forms that CPR had occurred. As the authors note, the accuracy of these codes is unknown. It is possible that not all CPR events are coded on the billing forms. It is also possible that CPR may be coded when it does not actually occur. It is fairly common for a "code blue" to be called on a patient who has not had a cardiac arrest, and it is easy to imagine a hospital billing coder recording this as CPR. The authors argue that errors in coding error rates probably have not changed over the years. This may or may not be the case.

With this caveat, I think this study is quite important, and provided important national data on the use and survival of CPR in older persons. But I agree with Eric that the more important questions relate to the long term functioning of patients after discharge. Eric's interest in functional and cognitive outcomes is supported by the landmark work of Terri Fried which shows that these seem to be the outcomes that drive decision making in older persons. Medicare database studies are not able to examine these outcomes. Medicare data should be able to tell us something about long term survival past hospital discharge, and hopefully the authors will extend their work to include this information at some point.