Discussion

Our survey results show that most clinicians rely on non-specific UTI signs and symptoms (ie, cloudy or foul-smelling urine), and that these characteristics are more commonly used by APPs and emergency department clinicians. We found that clinicians with lower confidence in clinical skills for management of UTIs in men were more likely to choose a longer treatment duration. Similarly, those who perceived that peers were choosing longer durations were more likely to treat with longer durations themselves.

Most clinicians endorsed dysuria (96%) and increased urinary frequency (88%) as indicative of a UTI, which is consistent with evidence-based symptoms and signs for UTIs in women. ^{Reference Bent, Nallamothu, Simel, Fihn and Saint7,Reference Medina-Bombardó, Jover-Palmer, Medina-Bombardó and Jover-Palmer17} Interestingly, although hematuria is also an evidence-based sign of UTI in women, it is only used by half of clinicians, and it is unclear why this differs from the other evidence-based symptoms and signs. ^{Reference Bent, Nallamothu, Simel, Fihn and Saint7} In general, clinicians used an array of symptoms and signs, ranging from 20% to 96% for the 14 surveyed variables, highlighting a lack of consensus.

A notable proportion of clinicians (~50%) relied on non-specific, non-evidence-based urine characteristics (ie, changes in urine odor or cloudy appearance) which lack specificity for symptomatic UTI compared to symptoms like dysuria. ^{Reference Bent, Nallamothu, Simel, Fihn and Saint7,Reference Nicolle, Gupta and Bradley13,Reference Midthun, Paur and Lindseth18} The reliance on these non-specific characteristics to diagnose UTI in men probably reflects the pervasive but incorrect belief that these symptoms and signs are reliable predictors of UTI. We found that emergency department clinicians were more likely to rely on non-specific urine characteristics, consistent with studies showing overdiagnosis of UTI in general in the emergency department. ^{Reference Tomas, Getman, Donskey and Hecker19} Use of these non-specific characteristics may be related to increased diagnostic pressures and high patient turnover in the emergency department. Increased use of these characteristics by APPs may be due to their disproportionately increased involvement in both emergency room moonlighting and same-day urgent care visits at our institution.

Clinician respondents tend to omit key components of the physical examination in diagnosing UTIs in men. While clinicians commonly evaluated for suprapubic or flank/costovertebral angle tenderness, they rarely evaluated the prostate or genitals, which could lead to missed diagnoses of prostatitis and epididymo-orchitis. The limited use of DRE may be due to a perceived lack of benefit or concerns for causing harm in acute prostatitis. A qualitative evaluation of the primary care clinician view of DRE in evaluating men with UTI deemed it “painful and irrelevant in the management strategy.” ^{Reference Soudais, Gallais and Schuers11} Another study reported that DRE was performed in only 6.6% of outpatient men presenting with a UTI. ^{Reference Soudais, Lacroix-Hugues, Meunier, Gillibert, Darmon and Schuers20} However, a tender prostate is present in 63%–90% of cases of ABP and is one of the few symptoms or signs which may distinguish ABP from other UTI syndromes in men. ^{Reference Millán-Rodríguez, Palou and Bujons-Tur21} It is possible clinicians are fearful of precipitating bacteremia, a common refrain in review articles going back decades. ^{Reference Nickel22} Notably, no citation in these articles supports this claim and these reviews do recommend a gentle prostatic exam, just not a prostatic massage. Another contributing factor to the limited use of DRE may be perceived self-efficacy by clinicians—only 33.6% of physicians felt confident their training to perform DREs was sufficient in a recent study. ^{Reference Wong, Drossman and Bharucha23} Overall, the limited use of DRE represents an opportunity to improve clinical practice as DRE is a low cost, low risk intervention that may help distinguish ABP from other UTI syndromes in men.

Of our respondents, only 13% reported prescribing longer than 7 days of therapy for UTIs in outpatient men, significantly less than the 65% of clinicians who treated men with UTI for over 7 days in a 2009 observational VA cohort. ^{Reference Drekonja, Rector, Cutting and Johnson8} This reported willingness to treat with shorter courses of antibiotics may reflect a nationwide trend of decreasing antibiotic usage within the Veterans Health Administration (VHA) or could be related to social desirability bias in answering survey questions. ^{Reference Kelly, Jones and Echevarria24} The behavioral factor (based on the Theory of Planned Behavior framework) which contributed the most to continued prescription of longer durations was subjective norms, as the perception that peers prescribed longer antibiotic durations was significantly associated with intention to prescribe longer antibiotic durations. Self-efficacy, as evidenced by clinician confidence questions, was another contributory factor in which respondents who felt lower self-efficacy were more likely to treat for a longer duration. The inverse relationship between confidence and treatment duration has been reported in prior studies of intensive care physicians specifically and more generally with antimicrobial prescribers at acute-care hospitals. ^{Reference Janssen, Oerlemans, Van Der Hoeven, Ten Oever, Schouten and Hulscher25–Reference Pandolfo, Horne and Jani27}

Our survey uncovered significant heterogeneity in antibiotic selection, with only two antibiotics deemed appropriate for empiric treatment by more than half of clinicians, and none by more than 70%. This reflects not only a difference in preference but a lack of agreement on appropriate empiric therapy among providers. Less than half of the surveyed clinicians preferred fluoroquinolones (48%) for UTIs in outpatient men, and over half endorsed nitrofurantoin (54%). These preference patterns may be due to overall decreased fluoroquinolone use for UTI treatment in the U.S. and increased recognition that men can have uncomplicated UTIs (and thus nitrofurantoin may be an effective choice). ^{Reference Trautner, Cortés-Penfield and Gupta5,Reference Umarje, Alexander and Cohen28} The lack of treatment consensus may also reflect the lack of applicable guidelines for treatment of uncomplicated UTI in men. ^{Reference Trautner, Cortés-Penfield and Gupta5,Reference Gupta, Hooton and Naber6}

Limitations of our study are that this was a single-center survey, although in a very large, tertiary care medical center with 12 outlying referral clinics serving over 131,000 Veterans with over 2 million outpatient visits per year. Additionally, we surveyed clinicians about their self-reported behavior, meaning that they may manage men with UTI differently than what they reported in the survey. This limitation may be particularly relevant for the reported duration of treatment for men with UTI, as responses may have been shaped by social desirability bias. Social desirability bias inherent to survey studies may have been further exacerbated by the face-to-face interactions incorporated into our survey distribution and collection.

In conclusion, our survey showed significant heterogeneity in clinician approaches to diagnosis and treatment of UTIs in outpatient men. For diagnosis, the impact of evidence gaps is noteworthy, leading to reliance on non-specific UTI symptoms and signs and minimal use of important physical examination tools (ie, DRE). For treatment, more recent evidence supporting the use of shorter durations of therapy has been reflected in local practice. Subjective norms and limited self-efficacy play a role in overtreatment of UTIs in outpatient men; both factors can be addressed in antibiotic stewardship programs tailored for UTI in men.