Accuracy of Glucose Meter among Adults in a Semi-urban Area in Kathmandu, Nepal.

Introduction: Glucose meters are gaining popularity in monitoring of blood glucose at household levels and in health care set-ups due to their portability, affordability and convenience of use over the laboratory based reference methods. Still they are not free of limitations. Operator's technique, extreme temperatures, humidity, patients' medication, hematocrit values can affect the reliability of glucose meter results. Hence, the accuracy of glucose meter has been the topic of concern since years. Therefore, present study aims to evaluate the analytical and clinical accuracy of glucose meter using International Organization for Standardization 15197 guideline.

Methods: A community based descriptive cross-sectional study was conducted in Kapan, Kathmandu, Nepal in April 2018. Glucose levels were measured using glucose meter and reference laboratory method simultaneously among 203 adults ≥20 years, after an overnight fasting and two hours of ingestion of 75 grams glucose. Modified Bland-Altman plots were created by incorporating ISO 15197 guidelines to check the analytical accuracy and Park error grid was used to evaluate the clinical accuracy of the device.

Results: Modified Bland-Altman plots showed>95% of the test results were beyond the acceptable analytical criteria of ISO 15197:2003 and 2013. Park Error Grid-Analysis showed 99% of the data within zones A and B of the consensus error grid.

Conclusions: Glucose meter readings were within clinically acceptable parameters despite discrepancies on analytical merit. Possible sources of interferences must be avoided during the measurement to minimize the disparities and the values should be interpreted with caution.