Calibration of blood test results from different hospitals? How can you match them.

[B-A-Dreamer]

Hi everyone.

This is the first time I have written. I signed up to this group three years ago and although I follow some of the questions and answers, I always feel that my experiences are so different from what other people describe that it could be confusing to join in the discussion. However, this month I have a question about how transferable blood test results are. After years of what felt like appalling treatment (and non-treatment) from three hospital groups in London, I have finally ended up with the Imperial Group and am generally impressed. However, if my GP orders blood tests, they are done by UCLH and don’t go to the Imperial Group and if the Imperial Group do them, they don’t send them to my GP.

Just before Christmas my GP sent me for a blood test and, because he was concerned, contacted my haematologists at the Hammersmith hospital. They told me to go in the same day. UCLH gave me a platelet count of 28, Hammersmith three hours later, 44. Some of my other blood parameters were within normal range at one and outside normal range at another. Since then I have had them repeated. 26 at UCLH and 44 at Hammersmith and 27 at UCLH and 43 at Hammersmith.

I know that within the same hospital my platelet count can drop from 92 to 8 in 3 days and then shoot up again but now I am starting to doubt any of the results.

I remember being told by a doctor at St. Thomas Hospital around 2008 that they weren’t interested in the results from other hospitals but had not expected the disparity to be so great, nearly 50 per cent. Now I am wondering what other people’s experience is.

[itplassne]

I have no specialist knowledge, but could it not just be different counts coming from different machines?I've noticed that there is quite often a discrepancy (sometimes as much as 15) between the platelet count from the analyser in the haematology department, and the platelet count from the lab: this is on the exact same blood sample...

[Calliecookins]

In the US the patient can get their own blood tests done at a reliable lab. The patient must pay out of pocket. This test is inexpensive, so I would have my own test done and get an accurate score. ITP medication depends on a correct result.

[Wintersun007 in reply to Calliecookins]

how are you quantifying a reliable laboratory? How do you know result is accurate? Who checks quality of venipuncture?

[Wintersun007]

Very good questions.  I see three main points.I apologise my reply will be long so I will split it.I am a retired Chemical Pathologist.  I’ll start my story by saying before I went toUniversity to study medicine I worked in hospital laboratory in the UK which was the second laboratory in the UK to have an analyser which would do automatedPlatelet counts.  Before then platelet counts were done by using microscopes.In 2021 after second dose of Pfeizer Covid vaccine, I developed ITP.  Whilst still working itWas very easy for me to just walk into phlebotomy to get a blood test which was fast trackedThrough haematology.  However when my Covid booster was due it was time to retire.My haematology consultant advised weekly platelet counts after both flu vaccine and Covid (booster was also Pfeizer!).  The flu vaccine was before I retired, the Covid booster wasone day before retiring.  Moving between secondary care blood samples to primary care was not easy!  Getting results to my consultant needs manually sending by email, either by GP practice staff or myself (do bear in mind my haematologist was a work colleague).  Getting results across theseBoundaries so we have only one result record should have been sorted out years ago!Pathology networks should have helped (and lowered variation in results) but in my Experience actually made it worse, and one of the reasons I retired!……..more  

[kiwi99 in reply to Wintersun007]

Hi Wintersun007, I'm just curious whether your ITP has resolved and, if not, are you on any medication for it? Also what effect did the third shot (and any subsequent boosters) have on your platelets?

[Wintersun007 in reply to kiwi99]

no treatment. Platelets run about 80, on vaccination (flu or Pfizer covid) platelets fall initially to about 60 (week 2) then rise to 120 week 4 to 6, by 2 months back to 80.

[Wintersun007]

Variation of measurementUncertainty of measurement has two components reproducibility (imprecision or precision, the ability to produce same result when reanalysed) and accuracy, getting true result.  All accredited laboratories will have figures for these for every test accredited.Accuracy will give an indication how results differ between laboratories although if two specific laboratories involved may be best by measuring specific samples in both laboratories, taken at the same time and also measured at same time.  Should also consider age of sample at time it is analysed, and any affect different anticoagulants may have.  The latter may require consideration of cross contamination. May also be worth considering are samples from correct patient?  Although this is less common these days, in my experience it has become more difficult for the laboratory to detect as such errors have increasingly become the test was ordered on wrong patient.

[Wintersun007]

Biological variation For platelets would be better from a haematologist.But an observation of mine is that when poor phlebotomy technique my platelet count is about 10 10^9 /L lower than expected.  Sadly it would be difficult to undertake a formal ethical study as would require deliberate poor technique.Can intake of specific foods have an affect (eg humus) and over what time frame?  What of alcohol or coffee?I remember a patient from many years ago who had two cholesterol measurements two days apart of 8 mmol/L.  For inexplicable reasons GP did a third a week later and got 5.5 mmol/L, and queried it. I suggested that the GP sent another sample but also ask patient about changes.  The fourth sample (two days after third) was 5.6 mmol/L.  GP reported patient stopped drinking alcohol! Hydration status can have a marked affect on some tests, results being higher when dehydrated and lower after rehydration ie drinking water.