Capillary blood glucose measurements in hospital inpatients using portable glucose meters

Background: The use of portable glucose meters by nursing staff to perform bedside capillary blood glucose measurements is standard practice in the management of diabetic hospital inpatients. Few studies, however, have examined the practical limitations or the cost of this technology. Aim: To investigate the performance and cost of capillary blood glucose measurements using portable glucose meters in a hospital inpatient population being managed for diabetes mellitus. Methods: The setting was a 500 bed metropolitan University Teaching hospital, with 22 meters (Glucometer-M) in routine use by about 450 accredited nursing staff. The Glucometer-M was also compared with an operator-independent meter (Hemocue) to assess the effect of operator bias on the overall efficacy and cost of the programme. Results: Retrospective analysis of Glucometer-M reagent strips and comparison of measurements (n = 72) with the Hemocue revealed a marked operator bias which diminished accuracy and increased costs. The significant proportion of low haematocrits (< 30%) in the hospital population limited the applicability of the Glucometer-M which only operates reliably over the haematocrit range 35–50%. The excess of blood glucose measurements was highlighted by both a hospital ward audit and the frequency which exceeded that of routine electrolyte assays. Conclusions: Reliable bedside estimation of capillary blood glucose levels in hospitals requires a meter which is accurate, has negligible operator bias, is largely unaffected by haematocrit, and has insignificant risk of cross-contamination. At present only the Hemocue fulfils these specifications. Irrespective of meter choice, it is necessary to develop criteria for glucose measurements and monitor adherence. (Aust NZ J Med 1993; 23: 667–671.)