Assessment of analytical performance of a glucometer in Indian adults at a tertiary care hospital

To determine whether glucometer-based glucose measurements are analytically and clinically comparable to the standard laboratory Glucose Oxidase-Peroxidase (GOD-POD) method in adult inpatients at a tertiary care hospital. Self-monitoring of blood glucose (SMBG) is the primary focus of Diabetes management as it helps the patient to monitor their plasma glucose levels on their own and maintain strict glycaemic control. Their affordability, portability, and convenience of use give them an edge over traditional laboratory-based reference methods.

We conducted our study (cross-sectional, method-comparison study) in a Tertiary Care Hospital affiliated to Rajiv Gandhi Medical College, Thane. Venipuncture was done for the collection of whole blood venous sample from the Median Cubital Vein randomly and the samples were analysed for glucose concentration by fully automated analyser in the central biochemistry laboratory. Simultaneously, capillary whole blood samples were analysed by the glucometer. This helped us to rule out bias of inherent difference. Sample size for this study was determined according to the International Organization for Standardization (ISO) guideline 15197:2013, Accordingly, a total of 104 paired capillary and venous blood samples were collected. Agreement was evaluated through the Bland-Altman plot. Parke's error grid and Surveillance error grid were plotted for the evaluation of clinical efficacy of the glucometer. The two methods were contrasted using the Passing-Bablok regression analysis and Deming Regression analysis. A scatter plot with a regression line was plotted to present the outcomes of the analysis.

Statistical difference between the results of Glucometer (109.4 ± 56.54) and Glucose Oxidase-Peroxidase (120.4 ± 55.19) method was significant with p value < 0.001. The dissemination of results in Surveillance Error Grid were as follows with 47.1% of results showing no risk while 26% & 23.1% results showed slightly lower and slightly higher risk respectively. Around 3.8% of results show moderately lower risk. Only 41.3% of results met ISO 15197:2013 accuracy criteria; however, 96.1% fell within clinically acceptable Parke's zones A and B. Although most discrepancies had minimal impact on treatment, caution is warranted in critical care settings.

Our study demonstrates that a substantial gap exists in the analytical performance of the glucometer compared to the standard laboratory analyser when evaluated against ISO 15197:2013 criteria. However, error grid analysis suggests that the majority of discrepancies observed had minimal impact on clinical decision-making in our study setting.