Pre-analytical factors and measurement uncertainty.

Pre-analytical factors are an important source of variation or errors in clinical laboratory measurements. Based on the new accreditation standards, medical and laboratory professions now seek to develop tools to deal systematically with these diverse factors. Several obvious pre-analytical uncertainty components were estimated in pragmatic experiments and combined with data on analytical variation and literature knowledge on biological variation, to estimate the measurement uncertainty of most common chemical and haematological examinations in clinical laboratories. The main aim was to assess quality specifications for regional laboratory services. The expanded measurement uncertainties (level of confidence 95%) of serum cholesterol, albumin and potassium remained within 13-16%. The major uncertainty component for cholesterol was biological variation, whereas those for albumin and potassium were sample collection and pretreatment. The measurement uncertainties for serum free thyroxin, thyrotropin and C-reactive protein, 20%, 42% and 125% respectively, were largely due to their biological variation. The measurement uncertainties of basic erythrocyte parameters (erythrocyte count and mean corpuscular volume, blood haemoglobin concentration) were less than 10%. Larger measurement uncertainties were obtained for thrombocyte and leukocyte counts, 24 and 31%, respectively, and for the reticulocyte fraction, 41%.