Dissertation on medical measuring instruments

By Tomas Homburg, thomburg@health.sdu.dk, 1/23/2025

When developing a new instrument to measure health data, such as blood pressure or lung capacity, it must be tested before it is put into use. Oke Gerke's dissertation is about which methods can be used to test new instruments in the health field before they should be used.

The test will show that the new instrument is at least as accurate as the old one. The reason why new measuring instruments are developed in the medical world at all is either to get more precise measurements, develop cheaper measuring instruments or, not least, to develop measuring instruments that cause the least possible inconvenience to the patient.

While there may be slight differences between measurements from the two instruments that are supposed to measure the same thing, it is important to ensure that these differences do not affect the patient's treatment. One way to examine this can be shown in a graph, called a Bland-Altman plot. The graph shows how the measurements from the two instruments are related and how big differences can be expected.

The first part of the dissertation takes a closer look at the Bland-Altman plots and examines:

• What researchers need to describe in their reports to make their studies clear, easy to understand, and reproducible.
• How to calculate how big the differences are if the results are not evenly distributed.
• How Bland-Altman plots have been used in research since 1986, especially to measure calcifications in the arteries of the heart (called Agatston scores).

The second part of the dissertation is about the Agatston score, which shows the risk of cardiovascular disease. Based on Danish data from the DANCAVAS study, the researchers have created reference curves that show how men and women aged 50-75 typically lie. Doctors can use the reference curves to assess a patient's risk of getting sick by comparing their scores to others of the same age.

The last part of the thesis looks at how best to find a boundary that distinguishes between sick and healthy people when using new tests, for example for blood tests.

This entire area of research belongs in diagnostic research. The thesis is relevant for both the use of new medical measuring instruments and artificial intelligence algorithms, which are making their entry into clinical practice to an increasing extent.