Quality specifications based on biological variation – where are we?

Consensus recommendations on the quality targets in laboratory medicine supports hierarchy quality model. In this concept, a model higher in the hierarchy should be preferred over lower one. The analytical performance in the majority of laboratories in our country is evaluated according to manufacturer recommendations, which is the lowest model in the hierarchy. Thus, we aimed to compare analytical coefficients of variation (CVa) against criteria for imprecision based on biological variation from five different laboratories. Imprecision of 17 biochemistry parameters (amylase, AST, ALT, bilirubin – total and conjugated, Ca, Cl, CK, CK-MB, creatinine, CRP, glucose, K, LD, Na, urea and total protein) was monitored by analyzing two levels of commercial quality control material, three times per day, over one year. CVa was monitored on four different platforms (AU2700, AU480, Architect ci 4100 and Cobas c501). Obtained CVa for each platform was evaluated against criteria for imprecision based on biological variation. Comparison against biological variation was as follows: all monitored parameters met the optimum quality specifications only on AU480, while CVa monitored on the rest of the platforms (AU2700, Architect ci 4100 and Cobas c501) did not meet quality specifications based on biological variation for four parameters (sodium, chloride, calcium and total protein). According to our results sodium, chloride, calcium and total protein did not meet quality requirements on three different platforms. One of the possible reasons is low within-subject biological variation which is difficult to achieve with available analytical methods.