An appendicitis virus proves the value of statistics

Appendicitis might be caused by a virus. The operation may not always be necessary. This was learned by a hospital examining its own records.
Written by Dana Blankenhorn, Inactive

A new study in the Archives of Surgery offers a fascinating thought.

Appendicitis might be caused by a virus.

(This is a picture from an old public service announcement about appendicitis, from the blog of Julie Luongo, author of the 2008 novel The Hard Way.)

As compelling as that conclusion may be, how it was reached matters more.

Doctors at UT Southwestern in Dallas did a statistical study. They looked at their own records.

Statistical data revealed peaks, which may be outbreaks of appendicitis, in the years 1977, 1981, 1984, 1987, 1994 and 1998. In addition, researchers uncovered some seasonal trends for appendicitis, documenting a slight increase in appendicitis cases during the summer.

Appendicitis is the most common operation performed, still, but it remains a mystery. The analysis looked at records from 1970 to 2006 and also found that the operation may not always be necessary. Outcomes among sailors and children seemed to indicate many ruptures resolve themselves without surgery.

At the heart of health reform, or that portion of health reform that has passed into law -- the stimulus for electronic medical records -- is the idea that you gather the data and it tells you what to do.

We're going to be gathering a lot more data from here, which is highly valuable. With enough cases you can rule out extraneous causes and concentrate on just what you want to know. You can also make firm conclusions based on smaller fluctuations inside the data.

Most medical studies today are done with hundreds or thousands of patients. It is expensive to find the right sample, to screen out anomalies, and to collect relevant records.

But that is going to become routine soon, not just in hundreds or thousands of cases but over millions and tens of millions of cases. When data is collected and collated using standards, like HL7, the cost of statistical studies goes down, while their value goes up.

Expect many more amazing results like this one over the next few years.

This post was originally published on Smartplanet.com

Editorial standards