7 January 2019 (Tuesday) - A Reflection on a Career in EQA

Here's something thought-provking...

4 August 2011 (Thursday) - D.I.

Last week I was wondering what a D.I. was. I know now. It’s a deviation index. Basically it’s a measure of how many standard deviations a result is from an expected target.

Which is far simpler than I thought it was going to be….

1 August 2011 (Monday) - QC Again...

Here’s an interesting dilemma. One of my IQA blood count parameters (the top one pictured above) has a bias. The results are clearly within the range of acceptability, but they are not bouncing about the expected mean – they are ever so slightly low. About two per cent low, which in practical terms is utterly insignificant, but in real terms worries me – I want to be having results on the mean, not below, no matter how slightly below.

However the weekly EQA (the lower one pictured above) is absolutely spot-on.

Should I tweak up the calibration to fix the IQA, thus forcing the EQA to be high? Should I leave it alone (bearing in mind the deviation isn’t actually significant)?

What I actually did was to increase the calibration very slightly. The IQA is still slightly biased (but not as much), and I shall see what the EQA has done when it’s EQA day next week.

26 July 2011 (Tuesday) - E.Q.A.

I’ve been asked to take over overseeing the haematology EQA work for the Trust (as well as doing all my other stuff). It should be interesting, and I spent this morning seeing what I will actually be taking on. It looks straight forward in principle: take everyone else’s EQA returns, look at the D.I.s and mark them “Good”, “Adequate” or “Poor “ accordingly.

If I can be given some dedicated time to do the job I think I’ll enjoy doing it.

There are several areas I can see I’d like to make immediate changes. Actually getting the results in the first place. I’m told this can be problematic. I don’t see why this should be so – all I need is a few passwords and I can access the data directly from NEQAS themselves.

I suspect chiefs might be a tad reluctant to let me have passwords. We shall see…

I’m also not happy on marking a blood count as “Good”, “Adequate” or “Poor “. Given that two parameters have D.I. over two, then we have a “Poor “, as oppose to only one D.I. being over two and less than three which is classed as “Adequate”. However a red cell count or a haemoglobin or an MCV being slightly squafty could make the MCH and MCHC squafty (since they are calculated parameters) and so we’d get a ““Poor “ where really we should get the better result of “Adequate”.

Similarly coagulation investigations and haematinic assays are lumped together. I’d much rather scrutinise each measured analyte. It might be more work for me to do, but I can’t help but feel that technically it’s a much better thing to do.

I’ll squabble this one with the Head BMS and the boss consultant…

Coming on from that, despite having a degree in mathematics, I’m still rather vague about what a D.I. actually is. After six years of mathematical study I’ve only ever encountered D.I. on an NEQAS return.

I shall email NEQAS to ask them…

And it had to be said that I will need to seriously reflect on how I mark the blood film morphology. It’s easy to mark numeric assays – you can’t argue with a number. And the blood transfusion EQA already comes with (effectively) “Good”, “Adequate” or “Poor “ already. But the morphology is far more open to interpretation.

For example take a case of a megaloblastic anaemia. We are only asked to report the five most salient features. In fact we can only report five – there is no more space on the return form for more. Given the presence of right shifted neutrophils, Howell Jolly bodies, reduced platelet count, megaloblasts, anisocytosis and poikilocytosis, which one would you choose not to report? Because they would all be there, and whichever one you left off the report, you’d be criticised for missing something.

I shall discuss this with those who scrutinise blood films in the first instance…