A method to improve the reliability of a non-invasive diagnostic test of the presence and / or severity of a liver disease, said method comprises: a. compile a diagnostic index, called the Initial Index, which involves the mathematical combination of at least two data, where the Initial Index is selected from blood tests comprising ELF, FibroSpect, APRI, FIB-4, Hepascore, Fibrotest, tests the FibroMeter family, and tests derived from the FibroMeter family where urea has been removed from the markers； b. analyze the reliability of each of the combined data to collect the Initial Index by identifying if at least one data from the Initial Index collected in stage a) is an anomalous data, an inconsistent data, or an inhomogeneous data, or is responsible for a decrease in the Dispersion Index greater than that observed with other data: i. with the comparison of each data with the expected data in a reference population, or ii. in view of the intrinsic or extrinsic reliability predictor (s), or iii. by calculating the Initial Index Scattering Index compiled in stage a), a high Dispersion Index means homogeneous Initial Index data, and a series of calculations of the Initial Index Scattering Index comprising n data where n delete 1 to (n-2) data from the Initial Index, which results in the classification of the data that most reduce the Scattering Index； C. if a data is anomalous, inconsistent or an inhomogeneous data, or a data that reduces a Scatter Index, generate an Event Alert； d. if an Event Alert is generated, calculating new indices, where the anomalous, inconsistent and / or non-homogeneous data, or a data responsible for the lowest Dispersion Index, is deleted to obtain an Alternative Index, or where the data anomalous, inconsistent and / or non-homogeneous is replaced by its average value in order to obtain an estimated index or, if at least two data are anomalous, inconsistent or non-homogeneous, the most discordant is deleted and the other (s) ) (n) is replace
