Antimicrobial susceptibility testing (AST) is an essential diagnostic procedure to determine the correct course of treatment for various types of pathogen infections. Patients are treated with broad spectrum antibiotics until AST results become available, which has contributed to the emergence of multidrug resistant bacteria worldwide. Conventional AST methods require 16-24 h to assess sensitivity of the bacteria to a given drug and establish its minimum inhibitory concentration (MIC). A rapid AST assay can assist clinicians in making an informed choice of targeted therapy and avoid unnecessary overprescription. Here, we have developed a highly parallelized droplet microfluidic platform that can screen four antibiotics/pathogens simultaneously and assess antibiotic sensitivity in 15-30 min. The device consists of four integrated microdroplet arrays, each hosting over 8000 docking sites, which can be operated individually or jointly for greater flexibility of operation. Small numbers (1-4) of bacterial cells were entrapped in droplets of 110 pL volume and monitored dynamically over 2 h. This imaging-based AST approach was used to determine the growth rates of four types of clinically relevant bacteria known to cause urinary tract infection (UTI) in millions of patients. We quantified doubling times of both Gram positive ( Staphylococcus aureus, Enterococcus faecalis) and Gram negative bacteria (e.g., Escherichia coli, Klebsiella pneumoniae) with varying levels of antibiotic resistance. Six concentrations of bactericidal and bacteriostatic antibiotics (oxacillin and tetracycline, respectively) were tested to determine the MIC of the strains as well as the heterogeneity in growth profiles of bacteria at single cell resolution. The MIC determined from phenotypic analysis in droplets matched the MIC obtained from broth microdilution method for all strains. The advantages of the proposed droplet-based AST, including rapid drug sensitivity response, morphological analysis, and heterogeneity in antibiotic-resistance profiles, make it an excellent alternative to standard phenotypic AST with potential applications in clinical diagnostics and point of care testing.