Quality assessment of blood glucose testing in general practitioners' offices improves quality.

Measurement of blood glucose is a frequent test in Danish physicians' offices. We describe here a new design of external quality assessment wherein fresh, unstabilized whole-blood samples were analyzed by the physicians' office methods and by a hospital laboratory comparison method (glucose dehydrogenase assay, calibrated against NIST 909a). This approach was used in the offices of 171 general practitioners in our county during a 5-year period. The first survey, in 1992, revealed unsatisfactory performance in terms of our criterion that no difference between the physician's office and the hospital laboratory's single result should exceed +/-20% of the laboratory result. After initiation of a program of consultation and assistance, the proficiency testing rounds of 1994 and 1996 showed considerable improvement. Thus, whereas in 1992 12% of the values were outside the acceptance limits, in 1994 and 1996 the respective values were 4% and 3%. We believe the effect was mainly related to improvements in choice of technology, procedures, and handling of specimens. We conclude that the use of bedside instruments should be restricted to diagnosis of only severe hypo- or hyperglycemia and to monitoring glucose >5 mmo/L in already diagnosed diabetics. Our program differs from previous approaches to quality assessment and could also be useful for large hospitals' in-house proficiency testing.