Analyser comparability lookup

How the answer is worked out

Each entry is one published, open-access comparison. Where the study ran the same patient samples on two analysers, the calculator uses its regression line: expected = slope × result + intercept, and the line inverted for the other direction. Where the study measured several systems against a reference, the calculator converts through each system's mean bias: expected = result × (1 + bias of the target) / (1 + bias of the source).

The comparisons

• Two central-lab chemistry analysers (one hospital laboratory). Patient serum and plasma, one US hospital laboratory, including dialysis patients; Deming regression. [Bush et al., 2020]
• Two haematology analysers (hospital network, 183 samples). 183 EDTA patient samples, each run on both analysers within six hours; Passing-Bablok regression. [Ciullini Mannurita et al., 2026]
• Two haematology analysers (one hospital, 296 samples). 296 EDTA patient samples, each run on both analysers within four hours; Passing-Bablok regression. [Bhola et al., 2023]
• Enzymatic creatinine on three platforms (four EQA schemes). Commutable EQA samples from four schemes, targets traceable to a reference measurement procedure, at 61 and 85 µmol/L. [van der Hagen et al., 2021]
• Creatinine peer groups in a provincial EQA programme (after IDMS). Provincial EQA results after IDMS standardisation; mean bias of each peer group against the target. [Lee et al., 2017]
• HDL cholesterol with four homogeneous reagents. Fresh sera from people without disease, against the CDC reference measurement procedure; mean bias per reagent. [Miida et al., 2017]

This site does not name analyser makers or models; the cited papers do, and each link opens the paper.

What it does not do

A published line describes one study: its laboratory, reagent lots, calibration and patients. Your analysers can differ from it, more so far from the study's average sample. The line also gives only the average difference: one sample can sit well above or below it, so a converted value close to a decision limit can fall on either side. Use the lookup to size a change before you plan it, then verify it on your own paired samples. To see whether a difference between two results is more than chance, use Is this change real?; to see what a creatinine difference does to eGFR, use eGFR after a creatinine method change.

To fit and check your own comparison line, see the guide to method comparison regression; for the whole switch, the analyser change checklist.

Sources

1. Bush V, Smola C, Schmitt P (2020). Practical Laboratory Medicine. pmc.ncbi.nlm.nih.gov/articles/PMC6909053/ (opens in a new tab)
2. Ciullini Mannurita S, Romeo C, Bonari E, et al. (2026). Diagnostics. pmc.ncbi.nlm.nih.gov/articles/PMC13511784/ (opens in a new tab)
3. Bhola A, Fudaly G, Rastogi P, et al. (2023). Indian Journal of Hematology and Blood Transfusion. pmc.ncbi.nlm.nih.gov/articles/PMC11065844/ (opens in a new tab)
4. van der Hagen EAE, et al. (2021). Feasibility for aggregation of commutable external quality assessment results to evaluate metrological traceability and agreement among results. Clinical Chemistry and Laboratory Medicine. doi.org/10.1515/cclm-2020-0736 (opens in a new tab)
5. Lee E, Collier CP, White CA (2017). Creatinine assay attainment of analytical performance goals following implementation of IDMS standardization. Canadian Journal of Kidney Health and Disease. pmc.ncbi.nlm.nih.gov/articles/PMC5347424/ (opens in a new tab)
6. Miida T, Nishimura K, Hirayama S, et al. (2017). Homogeneous assays for LDL-C and HDL-C are reliable in both the postprandial and fasting state. Journal of Atherosclerosis and Thrombosis. pmc.ncbi.nlm.nih.gov/articles/PMC5453684/ (opens in a new tab)