Although Randomized clinical trials (RCT) represent the gold standard to compare two or more treatments, the impact of observational studies cannot be ignored. Obviously, these latter are performed on unbalanced sample, and differences among the compared groups could be detected. These differences could have an impact on the estimated association between our allocation and our outcome. To avoid it, some methods should be applied in the analysis of observational cohort. Propensity score (PS) can be considered as a value which sums up and balances the known variables. It aims to adjust or balance the probability of receiving a specific allocation group, and could be used to match, stratify, weight, and perform a covariate adjustment. PS is calculated with a logistic regression, using allocation groups as the outcome. Thanks to PS, we compute the probability of being allocated to one group and we can match patients obtaining two balanced groups. It avoids computing analysis in unbalanced groups. We compared low protein diet (LPD) and the Mediterranean diet in CKD patients and analysed them using the PS methods. Nutritional therapy is fundamental for the prevention, progression and treatment of Chronic Kidney Disease (CKD) and its complications. An individualized, stepwise approach is essential to guarantee high adherence to nutritional patterns and to reach therapeutic goals. The best dietary regimen is still a matter of discussion. In our example, unbalanced analysis showed a significant renal function preservation in LPD, but this correlation was denied after the PS analysis. In conclusion, although unmatched analysis showed differences between the two diets, after propensity analysis no differences were detected. If RCT cannot be performed, balancing the PS score allows to balance the sample and avoids biased results.