An audit of ciprofloxacin use at Southmead Hospital, Bristol was carried out for forty patients treated in early 1992 employing a modified Delphi technique with six assessors. Most patients assessed (20/40, 50%) had urinary tract infections (UTIs), 5/40 (12.5%) had chest infections, 4/40 (10%) had bacterial gastroenteritis and 3/40 (7.5%) had either bacteraemia or infection following an orthopaedic procedure. A likely bacterial pathogen was isolated from 32/40 (80%) of patients; 14/32 (44%) had Pseudomonas aeruginosa infections and from the remainder Enterobacteriaceae including Salmonella spp. (non-typhoid) were cultured. Oral therapy with ciprofloxacin was used in 37 (93%) of the 40 patients, and the three others received iv treatment. In 21/35 (60%) of patients where an assessment was made by majority scoring, a quinolone was felt to be clinically justified. A quinolone was least likely to be thought justified if the patient had a chest infection. The assessors had few concerns about the effectiveness or toxicity of ciprofloxacin but for 41% (14/34) of patients, where there was a majority opinion, a cheaper alternative was felt to be available; most of these patients had hospital-acquired UTIs caused by Enterobacteriaceae. The duration of therapy was felt to be too long in 35% (10/29) of patients, mainly because of prolonged treatment of UTIs. In some cases of P. aeruginosa infection the assessors would have used higher doses than those prescribed. Ciprofloxacin was the quinolone of choice in 24/32 (75%) of assessable cases. Norfloxacin was chosen to treat UTI due to multi-resistant Enterobacteriaceae in 6.2% (2/32) cases.(ABSTRACT TRUNCATED AT 250 WORDS)