TECHNICAL ACCURACY OF TEN SELF-MONITORING BLOOD GLUCOSE DEVICES COMMONLY USED IN DHAKA CITY OF BANGLADESH
	
Background
Introduction of Self-Monitoring Blood Glucose (SMBG) Devices is a landmark in the history of diabetes management and prevention; however, the potential of this revolution may not be realized if the technical accuracy of the devices is not ensured. Due to unregulated nature of the marketing in many countries, consumers are victims of unethical promotion of SMBG systems. Evidence are thus required on the technical accuracy of these Devices which can help the Policy Makers and Implementing Executives to protect and guide the public to choose devices with acceptable quality within a reasonable price range.

Aim
to evaluate the technical accuracy of ten most commonly used SMBG Devices marketed in Dhaka City of Bangladesh.

Method
Under an experimental study design 10 SMBG Devices commonly sold in Dhaka City, selected though a Retailer-based survey, were subject to the accuracy test. One hundred type 2 diabetes mellitus (T2DM) subjects (aged 30-60 yrs) were recruited from the BIHS Hospital (a tertiary care hospital).  Female subjects with pregnancy or lactation were excluded and also any subject with an acute or chronic complication of DM or any other clinically evident major disorders were excluded. After standard preparation the experiments were performed in the morning. Blood glucose measurement by SMBG Devices was done following the enclosed instruction of the specific Devices. The Laboratory measurement of blood glucose was done by enzymatic method using the Dimension RXL automated chemistry analyzer and average results from duplicate measurements were taken and expressed in a blinded manner. Hematocrit was measured using Sysmex XT 2000 hematology autoanalyzer. Data were analyzed by paired t-test and Pearson’s correlation analysis.

Results
The results of all devices generated from the two equipment differed from each other significantly (p<0.05 to p <0.001) at both prandial states. In all cases the blood glucose values from the SMBG Devices were lower than those of the laboratory values. All meters showed significant variation from lab results in both fasting and ABF samples. When all the meters were analyzed together the Device and laboratory values showed highly significant correlation. However, all of the Devices failed to show on acceptable accuracy at the lower levels (5-10%) of deviation from the corresponding lab values. At the 15% accuracy level, 6 (60%) out of 10 devices (1,2,4,5,6,7) showed deviated results at both the prandial states. At 20% accuracy level, 3 (30%) of the devices (2,6,7) failed to show accurate results in fasting samples and 2 (20%) of the devices (2, 4) failed to show accurate results at the postprandial state. On pooling together data from two prandial states, at 15% and 20% accuracy level, 70% and 20% of the devices failed to show accurate results from the corresponding lab values respectively. A wide variation was found in the market price of both the devices (from US$ 7.5 to US$ 30) and the strips (from US$ 0.15 to US$ 1.02 per strip). On analysis of results and costs, higher price of a device or strip was not found to be necessarily associated with the quality of the result.

Discussion
Blood glucose results from a substantial proportion of SMBG Devices, commonly sold in Dhaka City, do not have minimum level of technical accuracy and many others are not optimum in their accuracy levels. %.  When the bare minimum of 80% accuracy (as per ISO 2007 Criteria) is considered, a good proportion (20%) of the Devices still fall outside the acceptance limit. Since pre- and post- analytic errors are almost avoided and analytic errors are substantially avoided in the present study, the inaccuracy of results seems to originate in the devices themselves.  It should be noted that the quality of the results do not have any straightforward relation with the cost of the devices or the strips. Patients’ awareness and education as well as appropriate intervention by regulatory authorities should urgently be promoted to address this alarming situation.

Conflict of Interest
I have no potential conflict of interest to disclose

Technical accuracy of ten self-monitoring blood glucose devices commonly used in Dhaka City of Bangladesh