Thanks to some of the posts on this site I was very kindly invited to visit the Mater Dei hospital in Malta. This was a great opportunity to meet like minded people and to savour some of the winter sunshine in the Mediterranean!

If you would like to learn more about teaching offerings on this, or any topic of uncertainty, please feel free to contact me through this site

The three days covered a wide range of topics and saw 17 delegates discuss MU from every angle of pathology. In the next few weeks I will be posting updates from that course on the topics discussed. It certainty gave plenty of inspiration from diverse topics including cell pathology, histopathology, bacteriology and mycology. Of course haematology, biochemistry and immunology were also covered.

The themes from the three days seemed to be the continued lack of guidance for dealing with manual techniques and other disciplines that don’t immediately lend themselves to “simple” MU assessments. To follow on from this I intend to try to approach some of those topics on this site.

Figure 1: Agenda for the morning session of day 1

Day one started with some introductory material. As with any course everyone comes to the topic with different exposure to MU so it is helpful to cover where we are coming from and where we want to end up. The benefit of this session was the discussion around some of the statistics that we use. Although many are familiar, it is easy to apply the statistics without fully appreciating what we are doing and why. I felt it very important to set the scene for the rest of the course at this point as we were going to be using these statistics throughout the course.

A course like this is most useful when active discussion is encouraged. The group really got into the spirit during the early session which made it lively and engaging. It is clear to see that although medical laboratories have widely different specialist interests, there are still large areas of commonality that exist. Finding this common ground led to us being able to openly discuss issues in departments many had not worked in before. The added benefit of these sessions was engaging the right people together so that moving forward a support network is built locally.

Figure 2: The programme for the afternoon session of day one

The afternoon was where the magic started to happen. This is where we got our teeth stuck into what we need to be considering for the very first steps of expressing our measurement uncertainty. We first looked at the definition of the measurand.

What do we need to consider?

What is the minimum we need to record?

How far should we go?

How do we record it?

How often does it need reviewed (if at all)?

We covered all of these things and hopefully got to the point where we were able to recognise exactly what we needed to consider and record. Following on from that, once we know how our measurand is defined we needed to think about what aspects of our assays should we be considering as potential contributors to our overall uncertainty. Again, with the wealth of knowledge in the scientists in the room we were able to look at things in a slightly less blinkered way – as we often do when thinking about assays we work with day in and day out. As before there are a few “must have” contributors that are applicable irrespective of what discipline we work in. But, thinking about possible others, we were able to include, and exclude as well, their significance in propagation to our final results.

We concluded a very busy day with discussion about bias uncertainty and its application in uncertainty analysis. This included talking through:

The concept of bias

Its uncertainty

How do we detect it?

How much matters?

How do we know if it matters?

What do we do if we do decide it matters?

What do we need to do after that?

The was a pretty heavy day. However, it was extremely successful in developing a lot of discussion about the subject of uncertainty. the next day was more focused on the analytical stages of uncertainty quantification. How do we do the calculations and how can we tie them all together to somehow represent our overall uncertainty. We will discuss the content next time….