Why is asymptomatic bacteriuria overtreated?: A tertiary care institutional survey of resident physicians

We set out to determine why antibiotics are overprescribed for ABU by evaluating both
patient and physician factors associated with inappropriate treatment of ABU. About
one-third of ABUs were overtreated against guidelines, in agreement with earlier studies
4]–6]. Female gender, old age, and pyuria, hematuria, and positive nitrite on urinalysis
were the patient clinical factors contributing to antibiotic overuse, while lack of
knowledge, misperception of the clinical implications of bacteriuria, and discordance
between knowledge and practice were the physician-related factors. One strength of
our study is that we have attempted to identify such patient- and physician-related
factors, as most previous studies focused on clinical factors or physicians’ knowledge
and attitudes alone 4]–6], 12], 13]. Furthermore, our data are valuable because of the high response rate to the survey
(74.2 %) compared to previous studies and the participation of physicians from a variety
of departments, all of which can be helpful for developing effective antimicrobial
stewardship programs and interventions 12], 13].

Pyuria was the most important clinical factor related to inappropriate antibiotic
treatment of ABU. However, pyuria itself does not distinguish UTI from ABU nor does
it indicate antibiotic use, being commonly found in patients with ABU, ranging from
30 % in young women to 90 % in the elderly or hemodialysis patients 3], 10], 18], 19]. This finding agrees with the survey result that 29.5 % of respondents prescribed
antibiotics for ABU in actual practice for abnormal urinalysis results, regardless
of the presence or absence of urinary symptoms. We assume that physicians find it
difficult to neglect pyuria despite the absence of urinary symptoms, or that they
find it difficult to distinguish UTI from ABU with pyuria. Likewise, physicians seemed
to have serious concerns about the possibility of UTI developing in the elderly, and
about women with ABU, even though ABU is prevalent in these groups. These results
are consistent with earlier studies which reported pyuria, old age, and type of organism
as factors promoting antibiotic overuse for ABU 4], 5]. Therefore, education about the differential diagnosis of pyuria and the interpretation
of urinalysis is required to improve the understanding and management of bacteriuria.

Our survey demonstrated an overall paucity of knowledge regarding bacteriuria and
considerable knowledge vs. gaps in practice among physicians. This agrees with previous
surveys reporting resident physicians’ (48 %) and healthcare providers’ (56-71 %)
poor knowledge regarding urine testing and guidelines about ABU, although the mean
level of knowledge was lower in our survey (37.3 %) 13], 20]. For the clinical vignettes describing ABU where antibiotics were not indicated,
about one third of the respondents correctly chose both the diagnosis and management.
Most of the physicians who were able to diagnose ABU said antibiotics should not be
used for such cases, whereas those who could not identify ABU opted to initiate antibiotics.
However, for the vignettes describing ABU that required treatment (pregnancy, prior
to transurethral resection of prostate), only half of the respondents who made a correct
diagnosis were aware of the need for antibiotics. Therefore, we suppose that the 32.0 %
rate of inappropriately managed ABU in the retrospective chart review is mainly due
to physicians’ misdiagnosis of ABU as UTI, and may be influenced by various clinical
factors that confuse physicians about the diagnosis. To improve the current situation,
clinical scenario-based education, not just dependence on guidelines or textbooks,
and specific trainee-centered education may be helpful considering the inter-specialty
difference in knowledge demonstrated by the survey.

From the survey, we could see that urine cultures were often ordered regardless of
the probability of UTI for various reasons mostly associated with misunderstanding
of bacteriuria. This effect was also evident in our review of bacteriuria cases, as
the indications for urine culture were often not clarified on the medical records.
Concern about ABU being a risk factor for postoperative infection or symptomatic UTI
lead physicians to check for the presence of bacteriuria, and thereafter to unnecessary
antimicrobial use to eradicate ABU. Although concern about postoperative infection
originating from ABU is prevalent among surgeons, especially in prosthesis implant
surgery, evidence is mounting that neither routine urine culture in asymptomatic patients
nor preoperative eradication of ABU is useful 21], 22]. Although this remains controversial, we generally do not recommend treating ABU
before hip arthroplasty in the study hospital. In addition, a recent study has shown
that suppressing reports of urine culture results ordered for non-catheterized patients
was effective in reducing reflexive antibiotic use for ABU 23]. Therefore, we suggest that reducing unindicated urine cultures may be the first
step to decreasing antibiotic misuse and to improving physicians’ recognition of the
difference between UTI and ABU.

We found considerable discrepancies between knowledge, perception and clinical practice,
which made the physician factor significant in relation to the inappropriate treatment
of ABU. Nearly half of the respondents acknowledged that their own actual practice
was at variance with their knowledge. Although the majority of the respondents agreed
that patients’ symptoms or signs were important for the diagnosis of UTI and none
opted to initiate antibiotics reflexively for abnormal urinalysis results, the lack
of documented explanations of the rationale for antibiotic use for bacteriuria indicates
a prevalent discrepancy between practice and perception. Such attitudes of physicians
were alluded to in a previous survey of the management of bacteriuria by intensive
care unit clinicians 12]. Practice-based audit, and feedback to help physician detect the errors of their
practices, as well as fundamental educational efforts are of critical importance in
overcoming this gap. Further studies are needed to explore the drivers of physician
practice that contradicts their knowledge.

Our partly retrospective, partly prospective study has a few limitations. First, there
may have been some misclassification of episodes of bacteriuria because of the retrospective
part of the study. However, to minimize possible biases incurred from a retrospective
review, three ID specialists independently participated in the classification and
evaluation of the appropriateness of antibiotic use. Second, our study was conducted
in an acute care hospital, so our result may not be generalizable to other situations
such as long-term care facilities where ABU is much more prevalent. Third, as we surveyed
physicians in the year after the care was provided (2012), some of the residents were
not the same as those who actually cared for the episodes of bacteriuria in 2011.
However, we believe that the inappropriate antibiotic use data can still be reasonably
linked and correlated with the survey considering that the majority of the residents
surveyed did practice in 2011.