Table 4.

Predictive model of multidrug-resistant gram-negative bacteria in patients with complicated urinary tract infection: a mixed-effects logistic regression model

Factors	OR	95% CI	p-Value
(Intercept)	0.1	0.06–0.16	< 0.001
Gender (male)	1.66	1.20–2.29	0.002
Acquisition in a medical facility	2.59	1.80–3.71	< 0.001
Indwelling urinary catheter	1.44	0.99–2.10	0.06
UTI within the previous year	1.89	1.28–2.79	0.001
Antibiotics within the previous 30 days	1.68	1.13–2.50	0.011

OR odds ratio, CI confidence interval, UTI urinary tract infection

Potential predictors included in the predictive model were age, sex, source of infection, place of residency, functional capacity score, personal history of myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, ulcer disease, diabetes mellitus, chronic kidney disease, hemiplaegia, solid tumor, liver disease, metastatic tumor, Charlson score, infection acquisition site, presence of indwelling urinary catheter, urinary retention, organ transplant, kidney organ transplant, immunosuppressive therapy, active chemotherapy, corticosteroid therapy, UTI within the previous year, previous 30-day antibiotic treatment (including previous treatment with quinolones, penicillins, cephalosporins, carbapenems, and other antibiotics), infection severity, neurogenic bladder, obstructive uropathy, other urinary tract modification, and chronic renal impairment