Laboratory sample stability. Is it possible to define a consensus stability function? An example of five blood magnitudes