Models vs. Experts #10: Lung Function tests
Comparison of reports on lung function tests made by chest physicians with those made by a simple computer program
- Geddes, D. M., Green, M., & Emerson, P. A.
- Thorax, 33, 257-260
- An online version of the paper can be found here
- Want a summary of academic papers with alpha? Check out our free Academic Alpha Database!
The numerical results of 60 sets of pulmonary function tests were submitted to five consultant chest physicians for independent reporting. The chest physicians’ reports were compared with the reports generated by the on-line computer reporting system in routine use in the pulmonary function laboratory at Westminster Hospital. There was good agreement between the reports of the chest physicians among themselves and with the computer. The individual reports were compared with the consensus opinions of the physicians and the computer. The computer’s decisions differed from the consensus opinion in 4.6% of instances. The comparable figures for the five physicians were respectively 2.9%, 2.9%, 2.9%, 4.0%, and 4.6%. Decisions differing from the consensus were due to mistakes or actual sustained disagreements of opinion. The physicians made more mistakes than the computer even though they were performing under test conditions. They made up for this, however, by producing fewer genuine disagreements from the consensus opinions. It is concluded that in routine day-to-day practice the computer report will be as consistently useful as the chest physicians’ reports and more immediately available.
The authors compare the interpretive reports of a simple computer algorithm and experienced chest physicians on the same set of raw test results from lung function tests.
Here is how the test works (my layman interpretation)
- Patients breath in and out of a tube under different test conditions and the computer tabulates all the raw results.
- 60 patients are analyzed.
- The results are given to a group of 5 chest doctors and a computer.
Here are the results from the test. Note, “mistake” for a computer is due to incorrect entry of data or copying of the report from the the display to the form.
- Mistakes by the computer are lower.
- Disagreements from consensus are higher for the computer.
On net, the computer and the doctors are pretty similar. Of course, the computer is much cheaper, more efficient, and won’t call in sick when their buddy invites them to a golf game (I’ve pulled that one before 😉 ).
Thoughts on the paper?
Note: This site provides NO information on our value investing ETFs or our momentum investing ETFs. Please refer to this site.
Join thousands of other readers and subscribe to our blog.
Please remember that past performance is not an indicator of future results. Please read our full disclosures. The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. This material has been provided to you solely for information and educational purposes and does not constitute an offer or solicitation of an offer or any advice or recommendation to purchase any securities or other financial instruments and may not be construed as such. The factual information set forth herein has been obtained or derived from sources believed by the author and Alpha Architect to be reliable but it is not necessarily all-inclusive and is not guaranteed as to its accuracy and is not to be regarded as a representation or warranty, express or implied, as to the information’s accuracy or completeness, nor should the attached information serve as the basis of any investment decision. No part of this material may be reproduced in any form, or referred to in any other publication, without express written permission from Alpha Architect.
Definitions of common statistics used in our analysis are available here (towards the bottom)