Surveillance of antimicrobial use is essential for predicting the emergence and preventing the spread of drug-resistant bacteria. This retrospective study aimed to estimate denominators in surveillance and correlation of carbapenem use, expressed as either days of therapy (DOTs) per 1,000 patient-days or DOTs per 1,000 admissions, and compare them with the resistance rates of Pseudomonas aeruginosa in 14 hospitals in Tottori prefecture. The DOTs per 1,000 patient-days data format was significantly correlated with the imipenem/cilastatin (IPM/CS; R2 = 0.393, β = 0.008, P = 0.016) and meropenem (MEPM; R2 = 0.407, β = 0.007, P = 0.014) resistance rates of Pseudomonas aeruginosa. Moreover, there was a strong correlation between the DOTs per 1,000 admissions and resistance rates of IPM/CS (R2 = 0.621, β = 0.011, P = 0.001) and MEPM (R2 = 0.706, β = 0.011, P < 0.001). When exploring the association between carbapenem use and resistance rates of Pseudomonas aeruginosa, the best fitted linear regression model (R2) was obtained for use in admissions. Our findings suggest that carbapenem use per admissions may be a more useful indicator than that of per patient-days in predicting carbapenem resistance in Pseudomonas aeruginosa.
