Statistical analysis of metrological characteristics: a method of diosmin quantitation

The aim of the work was a statistical analysis of the metrological characteristics of the method of diosmin quantitation by UV spectrophotometry in medicinal formulations.

Material and methods. Medicinal formulations with diosmin were used as reference materials. The quantitation of diosmin was done by UV spectrophotometry at wavelengths of 268 nm and 370 nm. Statistical processing of quantitation data was carried out by variational statistics, correlation analysis.

Results. The values of specific absorption indices for diosmin in sodium hydroxide solution 0.02 mol/L at wavelengths of 268 nm and 370 nm were determined, they amounted to 463.0±24.6 and 259.0±9.9. The methods of spectrophotometric quantitation of diosmin in medicinal formulations at wavelengths of 268 nm and 370 nm do not statistically significantly differ in reproducibility. The pharmacopoeial algorithm using the t-test for analytical data by UV spectrophotometry of diosmin shows the failure of a systematic error. Relative errors of the average diosmin concentration and systematic errors at analysis on wavelengths of 268 nm and 370 nm differ statistically significantly. The analysis of the correlation dependencies "relative error of the average value – analytical wavelength" and "systematic error – analytical wavelength" shows a relationship between these parameters. With an increase of the systematic error of diosmin quantitation a nonlinear and insignificant fluctuation in relative error of the average diosmin concentration was observed.

Conclusion. Comparison of the average values of the diosmin concentration to medicinal formulation at two analytical wavelengths is not a sufficient statistical test to verify the correctness of analytical method, since subsequent statistical analysis shows a significant difference between systematic errors and relative error of the average diosmin concentration.