Skip to content

KevinMD, how could you publish this?!

August 6, 2010
tags: ,

The commentors on this blog post about CME funding and medication errors from KevinMD have pretty much already nailed all the relevant criticisms (i.e.–correlation doesn’t equal causation, Thomas Sullivan seems to have serious conflicts of interest, etc, etc).

I would just like to make one additional point.  The post contains some data:

What made the Nevada hospitals a specific target was that hospital billing records for 2.9 million inpatient visits that have been submitted to the state over the past decade showed 1,363 occurrences, in which patients are harmed or threatened with harm, but Nevada hospitals reported only 402 events for those years.

Just for a moment, set aside other criticisms and arguments to examine these specific numbers.  1,363 patients harmed or threatened with harm* in 2.9 million inpatient visits is equal to 0.047% being harmed/threatened with harm.  Or, put another way, less than 5 in 10,000 patients were harmed or threatened with harm.  Context means everything.  Take those figures in relationship to one another when thinking about them.

As for the under-reporting, much of this may depend on how patients harmed or threatened with harm was defined.  Examining 2.9 million inpatient visits requires the use of algorithms to examine the data for identification of adverse events.  Results all depend on the construction of these algorithms.  Typically, at least two types of algorithms are constructed–conservative and more liberal (these have nothing to do with politics).  Conservative models make sure that everyone who is counted in the “harmed” column was truly harmed.  The more liberal models use assumptions to try to capture the under-counted “harmed patients” that were missed by the conservative algorithm.  It’s important to remember that neither one is probably correct, especially in this situation where “harmed” is difficult to define using such algorithms.

More generally, I don’t understand why KevinMD posted this!  This seems like a poorly constructed, thinly veiled attempt to drum up support for commercial CME.  I enjoy KevinMD because he consistently carries content from diverse sources with a wide range of opinions.  The common thread connecting his content is the quality of the posts.  Guest posts are consistently high quality–well-written with logical arguments and, when appropriate, backed by solid research.  This post, regardless of the opinion presented, represents none of those qualities.

*I don’t understand the term “threatened with harm.”  When conducting a study of medical errors, it seems your outcome variable is binary–“harmed” or “not harmed.”  I am less concerned with patients who were “threatened with harm” but the threat was somehow caught and the patient was ultimately unharmed.  In reality, the researchers probably slid in this catch-all term to “account” for the inherent variability and ambiguity in their search methods to identify “harmed” patients.

No comments yet

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: