Quality Requirements
2016 State of the Art Hematology Performance Specifications
In May 2016, the International Journal of Laboratory Hematology published a set of State of the Art performance specifications for hematology parameters. How close is state of the art to the other goals specified by Ricos, CLIA, Rilibak and others? See here.
2016 State of the Art Analytical Performance Specifications for Hematology
The table below contains an excerpt of the information on quality requirements (analytical performance specifications) offered by JY Vis and A Huisman:
Verification and quality control of routine hematology analyzers.
While Vis and Huisman cover a broad range of quality topics for hematology, their perspective on what is achievable for imprecision and inaccuracy (they avoid the term "bias") is very useful.
| Hematology Analytical Performance Specifications | ||||||||
|---|---|---|---|---|---|---|---|---|
| Test, Analyte, Parameter, Measurand | SOTA Allowable between-batch CV% |
Allowable "Ricos" CV% |
SOTA Allowable Inaccuracy |
Allowable "Ricos" bias |
SOTA |
Ricos TEa% | CLIA allowable total error, TEa% |
Spanish Minimum TEa% |
| Hemoglobin, Hb | 1.0% | 1.43% | 1.3% | 1.84% | 2.95% | 6.3% | 7% | 5% |
| Hematocrit, Ht | 1.4% | 1.35% | 1.8% | 1.74% | 4.11% | 3.97% | 6% | 8% |
| Red Blood Cell Count, RBC | 1.1% | 1.6% | 3.2% | 1.7% | 5.0% | 6.7% | 6% | 4% |
| Mean Cell Volume, MCV | 0.8% | 0.70% | 2.0% | 1.26% | 3.32% | 2.3% | -- | -- |
| Mean Cell Hemoglobin, MCH | 1.5% | 0.70% | -- | -- | 7% | |||
| Red Cell Distribution Width, RDW | 2.0% | 1.80% | -- | -- | -- | |||
| Reticulocytes | 10% | n/a | -- | -- | -- | |||
| White Blood Cell Count, WBC | 4.4% | 9.19% | 15.49% | 15% | 9% | |||
| High level (>10 x 109/L | 1.5% | -- | 6.88% | |||||
| Normal level (1-10 x 109/L) | 2.5% | 5.73% | 8.52% | |||||
| Low level (<1.0 x 109/L) | 6.0% | -- | 14.3% | |||||
| Neutrophils (abs) | 3.2% | 9.25% | 23.35% | -- | -- | |||
| Normal level (0.5-0.8 x 109/L) | 2.5% | 8.55% | 7.32% | |||||
| Low level (< 0.5 x 109/L) | 10% | -- | 19.7% | |||||
| Eosinophils (abs) | 10% | 10.5% | 13% | 29.5% | 37.1% | -- | -- | |
| Basophils (abs) | 20% | 14.0% | 32% | 15.4% | 65% | 38.5% | -- | -- |
| Lymphocytes (abs) | 3.5% | 5.1% | 5.0% | 10.78% | 17.6% | -- | -- | |
| Monocytes (abs) | 8.5% | 8.9% | 15% | 29% | 27.9% | -- | -- | |
| Platelets | 6.4% | 5.9% | 13.4% | 25% | 16% | |||
| Normal range | 3.0% | -- | 11.35% | |||||
| Low (approx. 50 x 109/L) | 4.5% | -- | 13.83% | |||||
| Very low (10-20 x 109/L) | 5.0% | 2.15% | 14.65% | |||||
This table highlights whenever the State of the Art specifications exceed the "Ricos" desirable specification for either allowable imprecision, allowable bias, or allowable total error. Note that the Vis and Huisman didn't calculate an allowable total error based on their component specifications, we simply added that here using the standard equation TEa = bias + (1.65 *CV).
There are clearly a few parameters (Basophils, Hematocrit, MCV, Platelets, and RBC), where the current methods on the marketplace do not meet the biologically-derived specification. But also quite interesting is that many parameters have State of the Art specifications that are well below what we would demand based on biologic variation. That possibly indicates that the marketplace is more than adequate to provide the quality required by those tests.
Another thing to note is how few parameters are covered by many of the resources available to us. CLIA doesn't provide specifications for the differential counts, for example. These State of the Art recommendations are extremely helpful in providing us some additional insight into areas where previous resources have not provided us any help.