FUrTIHF: French urinary tract infections in healthcare facilities – five-year historic cohort (2014–2018)

Factors associated with urinary diversion and fatality of hospitalised acute pyelonephritis patients in France: a national cross-sectional study (FUrTIHF-2)

“…The incidence of in-hospital AP increased slightly over the 5-year period, which could be related to an ageing population. However, incidence varied substantially by sex and age, likely reflecting specific risk factors for UTIs as suggested by our and other studies [2, 5, 7, 10], in that elderly males, sepsis, and comorbid conditions were primarily associated with AP hospitalisation [2, 11]. Some studies have also reported a proportion of patients hospitalised to be lower than 20% among young people, especially females, and higher in the elderly [2, 10, 22].…”

“…Moreover, hospital stays are sometimes coded by various healthcare practitioners, leading to variation in the coded information [18, 20]. However, validation of the algorithm showed excellent PPV (91%) for AP, which was higher than our UTI algorithm [10], and excellent PPV and sensitivity for urinary diversion (100% and 91%, respectively). Indeed, these situations rely on each physician’s assessment, leading to wide heterogeneity according to the healthcare practitioner and their specialty: a surgeon will likely not have the same judgement as an infectious disease specialist or an intensivist [18, 25, 29].…”

“…This study is, to date, one of the only estimations of the incidence of AP-related hospitalisations at this scale. A robust computerised tool using a hospital discharge database algorithm previously validated for UTIs [ 10 ] has been applied for AP and urinary diversion, after confirming its performance. In a real-life setting, the incidence estimate of AP-related hospitalisations was 19/10,000 cases, along with diversion recourse and other risk factors resulted in a case fatality of 5.6%.…”

“…An algorithm to select patients with UTI was previously validated with a predictive positive value (PPV) of 70.4% [ 10 ]. A validation step, focusing on a sample of 150 AP cases, was performed by a clinical reviewer with a second physician check if necessary to define the performance parameters of the specific AP algorithm: PPV 90.6%, with a 95% confidence interval (95% CI) ranging from 84.8% to 94.8%.…”

“…Similar variation in the incidence of AP has been recorded in the United States, Canada, and South Korea [ 5 , 7 , 8 ]. In France, AP generated 88,159 visits to emergency departments in 2019 [ 9 ], and nearly one-quarter of in-hospital UTIs were due to AP [ 10 ]. Despite the related morbidity, few studies have described a care pathway for AP, and the frequency of hospitalisations is poorly documented, especially where a diversion was necessary.…”