The mainly rural population served by the laboratory is approximately 230,000 individuals within 72 general practices. The laboratory of the Galway University Hospitals (GUH) analyses all urine samples from the practices in this region. The study required that practices submitted samples for every patient with suspected UTI on clinical grounds. In order to minimise the study impact on laboratory workload and established GP practice, practices with already high numbers of submitted samples were invited to participate. Practices were ranked according to the number of submitted urine samples in 2007 and based on samples size calculations, economic and time constraints, the 25 highest ranking practices were invited to participate in the study. Of these, 22 practices participated, two practices were not included as they did not have computerised records and one practice declined (they never participated in research).
From the 14th of September to the 9th of November 2009 (8 weeks) practices were requested to send a urine sample from all adult patients presenting with symptoms of urinary tract infection. Notices outlining the study were put up in all practice waiting and consultation rooms. A website was set up with information on antimicrobial resistance and detailed information on the study (http://www.antibiotics.nuigalway.ie). All adult patients (≥18 years of age) from whom a urine sample was received in the laboratory were informed of the study by a letter explaining the study. Patients could opt-out of the study by returning the included (freepost) opt-out form, by filling in the opt-out form on the website or by telephone. All practices were visited and the charts of all participating patients were viewed for information on demographic variables and all antimicrobial prescribing in the year previous to the urine samples was recorded. Patients with indwelling catheters and pregnant women were excluded from the study.
Ethical approval was obtained from the Research Ethics Committee of the Irish College of General Practitioners. More detail on the ethical approval, the selection of practices as well as on the inclusion of patients has been described in a previous paper .
The receiving laboratory is accredited to the ISO 15189 quality standard. Laboratory diagnosis was based on microscopy and semi quantitative culture of a urine sample. Pyuria was defined as the presence of increased leukocytes (greater than 20 white cells/μl) in the urine. For the purpose of this study, bacteriological confirmation of UTI was defined as a pure or predominant growth of relevant organisms at the level of 105 colony forming units (cfu) per ml . Antimicrobial susceptibility testing was performed in accordance with the disk diffusion method of the Clinical and Laboratory Standards Institute (CLSI) . For further statistical analysis, samples with intermediate resistance were categorised as resistant.
Antimicrobial prescriptions were recorded according to the major groups: penicillins, β-lactam/β-lactamase inhibitor combinations, cephalosporins, trimethoprim, fluoroquinolones (including nalidixic acid and fluoroquinolones such as ciprofloxacin), tetracycline, nitrofurantoin and macrolides. Information on previous urine samples was also included. Other variables available for patients were age, gender, number of visits in the previous year (more or less than 10). Medical card eligibility depends on income and age and can be interpreted as a proxy measure of socio-economic status. At the time of the study, about 30% of the population was eligible for a medical card including all pensioners over the age of 70 years. Medical card patients have free medical care and medication while other (private) patients pay for both.
For each patient the first visit during the study period at which a urine sample was obtained, was used as the index visit. The antimicrobial therapy prescribed in a period of 7 days around the date of the sample, was identified to be the medication given for this episode, with the closest prescription to the date of urine collection regarded as the initial prescription.
Appropriateness of different treatment approaches was assessed by evaluating the treatment prescribed by the GP with the subsequent laboratory report of the urine sample. Overall, antimicrobial prescription for patients without bacteriological confirmation of infection was considered inappropriate. The prescription of an antimicrobial to which the cultured organism was resistant was also considered inappropriate. Appropriate treatment included not prescribing an antimicrobial to patients with a negative culture and treating patients with an isolate susceptible to the antimicrobial prescribed. The antimicrobial prescriptions for patients with an isolate which was not E.coli were individually assessed. A simulation of different treatment options, i.e. if all patients who received an antimicrobial were treated with nitrofurantoin, trimethoprim or ciprofloxacin, is also presented. Based on the relative numbers of E.coli in each treatment group, an additional analysis was performed to assess at what trimethoprim resistance level the trimethoprim scenario would reach similar levels of appropriate treatment as nitrofurantoin and fluoroquinolones. The price of treatment was calculated from the average price of all the prescriptions of these practices for this antimicrobial group. Prices were according to the manufacturing cost of the medicine as recorded in the HSE-PCRS (year 2010).
Statistical analysis was performed with SPSS (version 18.0) for univariate comparisons and WinPepi for the comparison of proportions .