Evidence-based appraisal of two guidelines for the diagnosis of suspected, uncomplicated urinary tract infections in primary care: a diagnostic accuracy validation study

Abstract

Objectives Given the lack of accurate rapid diagnostics for urinary tract infection (UTI) in women, many countries have developed guidelines aiming to support appropriate antibiotic prescribing, but some guidelines have not been validated. We performed a diagnostic accuracy validation study of two guidelines: Public Health England (GW-1263) and Scottish Intercollegiate Guidelines Network (SIGN160). Methods We used data from women with symptoms suggestive of uncomplicated UTI from a randomized controlled trial comparing urine collection devices. Symptom information was recorded via baseline questionnaire and primary care assessment. Women provided urine samples for dipstick testing and culture. We calculated the number within each risk category of diagnostic flowcharts who had positive/mixed growth/no significant growth urine culture. Results were presented as positive/negative predictive values, with 95% CIs. Results Of women aged under 65 years, 311/509 (61.1%, 95% CI 56.7%–65.3%) classified to the highest risk category (recommended to consider immediate antibiotic prescribing) and 80/199 (40.2%, 95% CI 33.4%–47.4%) classified to the lowest risk category (recommended to reassure that UTI is less likely) by the GW-1263 guideline (n = 810) had positive culture. For the SIGN160 guideline (n = 814), the proportion with positive culture ranged from 60/82 (73.2%, 95% CI 62.1%–82.1%) in those for whom immediate treatment was indicated to 33/76 (43.4%, 95% CI 32.3%–55.3%) in those recommended a self-care/waiting strategy. Conclusions Clinicians should be aware of the potential for diagnostic error when using diagnostic guidelines for managing uncomplicated UTI and making antimicrobial prescribing decisions. Infection cannot be excluded on the basis of symptoms and dipstick testing alone.