Summary
Activated partial thromboplastin times accumulated from two proficiency testing surveys
were analyzed to determine simultaneously the effects of the method and reagent used.
Prothrombin time results were reevaluated concomitantly for comparison.
A robust two-way analysis of variance was applied to determine the effect of method
and reagent on APTT results. The effect of the reagent and method on the ratio of
abnormal to normal plasma clotting times was determined. We found a substantial difference
in ratios for the PT using different reagents on the same instrument. There was an
even larger effect of reagents on APTT ratios.
Our finding of substantial reagent effects for the PT and APTT clearly support the
need for standardization. We found standardization to be feasible only for the PT,
and only if applied in a form consistent with the inherent error structure of the
data. For the APTT the present methodology and plasma samples did not achieve consistent
standardization.
Keywords
Prothrombin time - Activated partial thromboplastin time - Proficiency testing - Statistic,
robust estimation - Analysis of variance
