Uncomplicated urinary tract infections are probably one of the diagnoses that Emergency Physicians handle the worst - if they come to the ER, they're likely to get some sort drawn-out testing, whereas, if they went to their regular physician or called the nurse hotline, there would be antibiotics waiting for them at the pharmacy before they finished talking.
This is a prospective study in which patients with possible UTI were referred to a triage kiosk to complete a standardized computer questionnaire. 624 patients with possible UTI interacted with the kiosk - and unfortunately, only 103 qualified for the study by having enough features of typical, low-risk illness. Patients were then randomized to protocolized antibiotic prescription as reviewed by a triage physician or usual care.
The good news - the kiosk saved a lot of time (89 minutes vs. 146 minutes). The bad news - there were only 41 patients followed-up in the intervention group and 26 followed-up in the control group, so we end up with only a tiny number of patients in each arm. The kiosk group received more antibiotics for negative urine cultures than the control group (93% vs. 67%), so there is some additional element of harm secondary to antibiotic exposure - and, with a limited protocol, there are potential misses - and this study isn't large enough to identify them.
But, really, uncomplicated, typical UTI symptoms in women shouldn't be rocket science - and you shouldn't necessarily be doing any testing. I would say the computer is a better physician - except, it would be absolutely simple for a physician to simply narrow their approach to match the efficiency of the kiosk with, in theory, some added skill.
"A Randomized Trial of Computer Kiosk–expedited Management of Cystitis in the Emergency Department"