MEASUREMENT UNCERTAINTY
doesn’t mean deleting all QC that has failed. Rules and the power of detecting them are linked to probability of failure and detect random variability, even extreme variation. However, flyers due to operator errors, missampling or electrical fluctuations can be identified and removed. The statistical methods for outlier detection are numerous and to be decided locally.
To combine or not to combine? A common question with MU is what to do with each QC levels standard uncertainty when we use multiple QC levels? This is compounded when we factor in many analysers, laboratories and even hospitals in the modern network model.
The options are limited and quite simple. Either keep all results separate and report multiple MU across the range or combine them all into a single number for an assay MU. The quality of the result, from a clinical perspective, is most important at a clinical decision limit. It is exactly at this point that imprecision, and therefore MU, has the most significant impact on diagnostic accuracy. Ideally imprecision data is generated at that point. Often that is not the case. To provide MU at a decision point-existing MU is either extrapolated if close enough to the MU or further testing with additional material of different concentrations may be required. When multiple QC levels are used, the closest to the decision point is the closest representation of the decision limit MU. Whether this is an appropriate representation must be made on an assay by assay basis.
MU across lot numbers, measurement systems and laboratories Aggregating all IQC data into a single metric risks including variability that is a function of the IQC data, not reflected in patient samples. It cannot, without objective evidence, be assumed that the variability across lot numbers is constant. CLSI C24 recognises that the variability of the system across lot number changes should be similar, and this is one method for propagating target SD across lot numbers. This may not be the case for when there are matrix influences such as lyophilisation. Lot-to-lot variation (changing of mean target values) introduces short-term biases to the overall imprecision. This variability is not, generally, reflected in individual patient specimens.
n How does this manifest itself? Aggregation of data across lot
WWW.PATHOLOGYINPRACTICE.COM MAY 2024
numbers, including shifts of the mean may overestimate random variability. The consequence would be inflation of the MU in patient samples. It is recommended that different QC levels, and different lot numbers, are all separated and calculated separately. If results appear compatible (the definition of which is a subject of ongoing debate) they may be combined.
n The challenge of pooling standard uncertainties If standard uncertainties must be aggregated the number of data points within each category (lot, analyser, laboratory etc) impact the combined uncertainty. Using the pooled standard deviation method, the weighting is by the number of samples tested (n) and divided by the degrees of freedom. ISO/ TS 20914:2019 does discuss an approach
where a network MU can be applied across multiple measuring systems so that the MU can be applied to any sample being tested on any of those measuring systems, irrespective of which one it was actually measured on. For consideration is that again differences in means will manifest as small biases. If these are deemed clinically significant they are to be corrected for and the uncertainty of that correction applied. Managing the differences in means, bias and overall uncertainty will be in a future article – along with lots of calculations!
Reporting absolute or relative uncertainty
A common question is whether to report uncertainties as standard deviations or as a coefficient of variation. The answer depends on the use of the result and the understanding of the behaviour of
Gain time
S-Monovette® Safe. Fast. Sustainable.
Gain with the S-Monovette®
• Reduce time and costs for personnel and material with a minimised haemolysis rate
• Utilise the advantages of a dual system for all vein conditions
• Reduce your CO2 emissions,
waste volume and disposal costs considerably
Blood
collection with the
S-Monovette® & Tempus600®
Gain with Tempus600®
• Shorten the total turnaround time (ToTAT) for clinical specimens
• Save valuable time for your staff with the easy, immediate transport of specimens and relieve existing transport systems
• Take the next step towards laboratory automation
Learn more about the synergy of S-Monovette® & Tempus600®
s-monovette-tempus600.
sarstedt.com
SARSTEDT Ltd. · Optimus Way · Optimus Point · Leicester LE3 8JR · Tel: +44 116 235 9023 · Fax: +44 116 236 6099 ·
info.gb@
sarstedt.com ·
www.sarstedt.com
anz_gain_time_gm_symposium_marketing_120x168_0124.indd 1 10.01.2024 08:44:58 21
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60
