« Will health services research determine whether health reform succeeds? | Main | Health care from the heart -- my response »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Kevin Outterson

Based on your chart, not many babies would be saved if everyone reduced their mortality below 20% because very few babies are on the left part of the chart. Take out those 3 very small NICUs (~17 admits total; where are they? remote?) and then re-draw the trendline.

In the prior post, you expressed less certainty about the solutions.

Bill Gardner

That is an excellent observation. For reasons that wouldn't be accessible to you as a reader, I think there are many more small NICUs than is suggested by the graph.

These data come from a database compiled by about 40 freestanding pediatric hospitals. However, in my city, there are many small NICUs that are not represented in the database, and my sense is that the situation is similar elsewhere.

You are concerned, like I am, that with a relatively small number of data points, the estimate of the slope is not really stable. Another important point, however, is that similar relationships between volume and mortality are found in other medical procedures, such cardiac surgery.

Sorry for not making this important point clearer.

The comments to this entry are closed.

My Photo

Become a Fan