Administration of antibiotics in the private hospital "X" adopted a concept called RonaldIrwanto Antimicrobial Steweardship Program (RASPRO). The suggestion of an empirical antibioticcombination in type III stratifi cation community pneumonia patients was meropenem-levofl oxacinor meropenem-amikasin. The aim of the study was to determine the eff ect of empirical antibioticcombination meropenem-levofl oxacin with meropenem-amikacin to RASPRO type III stratifi cationcommunity pneumonia patients towards the LOS and decreased leukocytes. The test sample calculatedusing diff erence between two population proportions formula and analysed using Chi-square method.Diabetes mellitus, immobilisation and geriatrics as confounding variables were controlled by logisticregretion multivariate analysis. The results showed that meropenem-levofl oxacin had a tendency 1.81times to experience LOS < 5 days and 0.92 times to experience decreased leukocytes ≥ 10% comparedto meropenem-amikacin, but both were not signifi cant (p 0.161 and p 0.835). The result control ofconfounding variables were found that geriatrics as a meaningful confounding variable eff ect of LOSand no confounding variables were considered to aff ect the decreased leukocytes. In conclusion, thereis no eff ect of empirical antibiotic combination meropenem-levofl oxacin with meropenem-amikacintowards the LOS and decreased leukocytes to RASPRO type III stratifi cation community pneumoniapatients using statistics after controlling the confounding variables.