Discussion

The IDSA's recommendations [1] were published in 1999 to provide physicians with a standard protocol for the treatment of uncomplicated UTI. Our study demonstrates that physicians in the United States have not altered their prescribing patterns in response to these guidelines. We observed that TMP-SMX is not the preferred agent in the treatment of uncomplicated UTI, although its use is suggested by the IDSA guidelines, and that ciprofloxacin is not only preferred, but its rate of prescription is increasing.

In general, the rates of antibacterial use that we report are consistent with those previously reported. In our study, 30% of uncomplicated UTIs were treated with TMP-SMX from 1996 through 2001. Other studies have reported a rate of 37% from 1996 through 1999 [7], 24% from 1997 through 1998 [8], and 30% in 2002 [9]. On the basis of these studies, as well as on the basis of our study, the 3 most commonly prescribed antibacterial agents for the treatment of uncomplicated UTI are TMP-SMX, fluoroquinolones, and nitrofurantoin. Furthermore, all studies demonstrated that antibacterial preference was significantly influenced by race, payment method, and physician specialty.

McEwen et al. [7] analyzed treatment preferences for both acute and chronic UTIs using the insurance claims from 1996 to 1999 of employees of a company that had preferred provider organization health care insurance coverage. Trends were not analyzed. Duration of therapy, which could not be ascertained with our data, was reported in the study by McEwan et al. [7]. Average duration of therapy for acute UTI was 8.6 days, which exceeded the recommended 3-day course. It should be noted that acute UTI was defined as “not recurrent” and that the study was unable to identify visits as “uncomplicated” because data about comorbid conditions and pregnancy were unavailable.

Huang and Stafford [8] examined the NAMCS data from 1989 to 1998. They reported a statistically significant increase in the use of fluoroquinolones, as well as a decrease in the use of TMP-SMX. Because NHAMCS data were not included in this study, prescribing patterns for hospital clinics were unavailable, and conclusions were drawn only from data obtained from private offices.

Kallen et al. [9] examined NAMCS and NHAMCS data from 2000 to 2002. They reported a “nonsignificant trend” toward increasing fluoroquinolone use. This was likely caused by the smaller number of study years included in the analysis. Our study, which spanned a longer period of time, and the study by Huang and Stafford [8] were able to demonstrate statistically significant trends in antibacterial use.

McEwen et al. [7] and Huang and Stafford [8] incorporated data obtained prior to 1999, before the publication of the IDSA guidelines [1]. These studies were, therefore, unable to evaluate physicians' compliance with the guidelines. An analysis performed by Kallen et al. [9] using more-recent data from the NAMCS and the NHAMCS corroborated that quinolones were used more frequently than sulfa agents [9], but it made no comparison with data from years prior to the publication of the IDSA guidelines [1]. Our study is unique, in that we evaluated physicians' compliance with the IDSA guidelines by analyzing and modeling trends of antibacterial use for uncomplicated UTI with data compiled from the NAMCS and NHAMCS both before and after publication of the guidelines.

The deviation from the IDSA guidelines may be explained, in part, by many physicians' lack of awareness of them. Of greater concern, however, are physicians who are aware of the guidelines but who continue to prescribe independently of the published recommendations. It is possible that these physicians are influenced by external factors, including substantial marketing—with attendant perquisites—from pharmaceutical companies that promote the use of the company's drugs. Differences in cost are not trivial. During the study period, a 3-day course of TMP-SMX cost approximately US$1.83, whereas a 3-day course of ciprofloxacin cost approximately US$53.56 [6]. Other possible explanations for physicians' nonadherence to the guidelines may include an unwillingness to change because of greater familiarity with ciprofloxacin and personal bias against TMP-SMX because of its adverse effect profile.

Physicians in hospital clinics exercised much greater compliance with the IDSA guidelines [1] than physicians in private offices. Providers from hospital clinics may be more conscientious in practicing evidence-based medicine. A greater number of uninsured patients (who cannot afford more-expensive medications) or fewer external influences (such as pharmaceutical marketing) may explain better compliance with the guidelines by hospital-based clinic practitioners.

Our observation that nonwhite and uninsured patients were more likely to receive TMP-SMX may also reflect the impact of economic status on treatment preferences. These patient populations may be more likely to be affected by differences in antibacterial costs. Interestingly, the treatment prescribed to patients who had Medicare or Medicaid coverage was similar to that prescribed to patients who had private medical insurance coverage. These populations would unlikely be affected by differences in drug price, because their treatment regimens would be covered by their respective health care plans.

The preference of general surgeons for TMP-SMX and of obstetricians and gynecologists for nitrofurantoin may reflect the influence of physician “culture” in determining treatment preferences. The prescribing pattern observed for obstetricians and gynecologists to use nitrofurantoin more commonly for UTI during pregnancy is likely because of its profile as a class B drug. The habit of prescribing nitrofurantoin for UTI may, therefore, have been extended to include their nonpregnant patients.

The influence of local habit and culture, whether in the treatment of patients in a private office versus a hospital clinic or by physician specialty, is apparent. Measures to improve practices should include efforts to increase awareness of IDSA guidelines for the treatment of UTI. Education should target private offices. In addition, the development of local practice guidelines can help to assure physicians that these recommendations pertain to their community. Community antibiograms, although often not readily available, may help to alter location-specific practices in an effort to develop compliance with national guidelines.

Several study limitations warrant discussion. We identified cases of uncomplicated bacterial UTI by isolating observations in which the health care provider entered “cystitis” or “UTI” as a diagnosis on the survey. It is possible that, in some cases, the provider used these terms but, in actuality, treated a more complicated infection, such as recurrent UTI or pyelonephritis. In these circumstances, the diagnoses of pyelonephritis or recurrent UTI would have been excluded from our data analysis and TMP-SMX would not have been recommended as an empiric, first-line therapy. Other factors that were not elucidated but that may have affected the “complicatedness” of the infection included recent hospitalizations, recent antibiotic use, and abnormalities in urinary tract anatomy.

Finally, culture results may have been available to guide therapy decisions. In these cases, treatment would not have been empirical, but directed. However, these data were unavailable to us. We attempted to minimize these possibilities by excluding patients who received diagnoses or had an underlying condition that would preclude a diagnosis of uncomplicated UTI, such as patients with long-term indwelling urinary catheters, as well as by repeating logistic regression analyses after excluding potential risk factors for complicated UTI, such as diabetes mellitus.

According to these data, ∼7 million patient visits are made to health care facilities for cases of uncomplicated UTI in the United States every year. Given the volume of patients seen for this problem, physicians' choices regarding treatment are important and will have a major impact on yearly drug costs, as well as on the emergence of drug resistance. Our findings demonstrate a general lack of adherence to the IDSA guidelines for treatment of uncomplicated UTI [1], particularly among physicians in private offices, and raise concerns as to what factors motivate a physician to change his or her prescribing practices.