Influence of a UTI empiric treatment pathway on physician prescribing in an academic medical center

In an attempt to improve patient outcomes and control costs, a guide to empiric antimicrobial therapy, including a urinary tract infection (UTI) algorithm, was distributed to hospital physicians at an academic medical center. A retrospective study was conducted to assess the impact of the guide on physician prescribing of empiric antimicrobial therapy for UTIs. Prior to the implementation of the guide, 45% (n=55) of patients with UTIs were treated consistently with the algorithm. Although not statistically significant, consistency increased to 51% (n=45) after implementation of the guide. The initial publication of a guide to empiric antimicrobial therapy did not significantly influence physician prescribing patterns regarding UTIs. Educational sessions at the time of implementation along with reinforcement of the guidelines may further impact prescribing habits and influence algorithm compliance.




After the implementation of the guide, 23 patients (51%) were treated consistently with the algorithm, which is not a statistically significant improvement in algorithm compliance (P=.35). Of the 5 patients with asymptomatic bacteriuria, none of the empiric treatments were consistent with the algorithm. Empiric treatments included cephalexin, levofloxacin, and SMX/TMP. The algorithm does not support antimicrobial use for asymptomatic bacteriuria, with the exception of pregnant patients. No patients were pregnant in this asymptomatic bacteriuria group. Of the 9 patients with an uncomplicated UTI, 3 patients (33%) received SMX/TMP in accordance with the UTI algorithm for empiric antimicrobial selection. With the exception of 1 patient administered nitrofurantoin, the remaining patients with an uncomplicated UTI were prescribed levofloxacin. Regarding the treatment of the 24 patients with complicated UTIs, 15 patients (64%) were empirically treated with levofloxacin in accordance with the UTI algorithm. As with the pre-algorithm implementation group, the empiric treatment selection was most varied in the treatment of complicated UTIs. In the acute pyelonephritis group, 3 patients (43%) received levofloxacin empirically, which is consistent with the algorithm. The other 4 patients with acute pyelonephritis were empirically treated with either cefazolin or piperacillin/tazobactam, which are not included in the algorithm for empiric treatment of acute pyelonephritis. The empiric antimicrobial treatment selections for UTIs post-algorithm implementation are depicted in Figure 3.

The most commonly prescribed antimicrobial for the empiric treatment of UTIs in both study populations was levofloxacin. Prior to the implementation of the guide, 62% of patients received levofloxacin as empiric treatment for UTIs. Although levofloxacin use decreased to 56% after implementation of the guide, the difference was not statistically significant (P=.53). The use of SMX/TMP also increased from 5% to 16%; however, this difference was not statistically significant (P=.09). Prior to the implementation of the guide, no patients received cephalexin for uncomplicated UTIs. Although not statistically significant, the use of cephalexin increased by 4% after the implementation of the guide. The use of cefazolin, which is not an option in the algorithm, was 5% prior to implementation and was reduced to 0 post-implementation.

Bacterial cultures were not obtained in all patients with suspected UTIs. In both groups, only 47% of patients had a urine culture with identified bacterial isolates tested for antibiotic susceptibilities. The susceptibility profiles of infecting organisms were consistent with the institutional antibiogram from previous years. Patients without cultures were diagnosed with a UTI based on a positive urinalysis and symptoms.


After implementation of the guide, a slight, statistically insignificant improvement in physician prescribing habits for the overall empiric treatment of UTIs was noted. Contributing to the overall improved adherence, compliance with the algorithm for uncomplicated UTIs similarly increased. However, immunocompetent, non-pregnant patients with asymptomatic bacteriuria were still being treated with antimicrobials empirically even though the algorithm did not indicate treatment. Although not statistically significant, the decrease in the use of levofloxacin and the increase in the use of SMX/TMP and cephalexin for uncomplicated UTIs post-implementation of the guide are encouraging.

In the treatment of complicated UTIs and acute pyelonephritis, physicians' empiric treatment selections were the most diverse. Physician compliance with the UTI algorithm was greatest in the treatment of complicated UTIs. Although only a small number of patients, the increase in adherence to the acute pyelonephritis pathway trended toward significance. The most significant factor contributing to algorithm adherence in both of these study populations was the inclusion of levofloxacin in the respective pathways. Of note, the empiric antimicrobial treatment recommendations included in the institutional treatment algorithm remained appropriate when considering the susceptibility patterns of the bacterial pathogens specifically isolated from the urine of study patients.

In both study groups, levofloxacin was the most prescribed empiric antimicrobial agent for all types of UTIs. After the implementation of the guide, the use of levofloxacin decreased by 6%. This decrease in levofloxacin use corresponded with an increase in the use of SMX/TMP and cephalexin, although no change was statistically significant. Reflecting these changes, potential cost savings, especially in the uncomplicated UTI pathway, can be predicted. For example, the acquisition cost of SMX/TMP 800/160 mg po bid is $0.36, while the cost of levofloxacin 500 mg po qd is $6.44. Conversion of the antimicrobial regimen from levofloxacin 500 mg po qd to SMX/TMP 800/60 mg po bid produces a cost savings of $6.08 per day. When applicable and appropriate, utilizing the more cost-effective SMX/TMP could potentially produce a significant cost savings to the institution as well as the patient.

Since the guide directs the empiric treatment selection of UTIs, urine cultures and sensitivities are critical for ensuring that appropriate treatment continues. Both before and after the implementation of the algorithm, the majority of study patients did not have a positive urine culture. Often no urine culture was ordered or the urine culture yielded mixed organisms necessitating a second urine specimen, which was rarely collected, for possible identification of any infecting organisms. Furthermore, of the patients who did have a positive urine culture available, in general, empiric treatment selections were not changed to a more cost-effective antimicrobial therapy once antimicrobial susceptibilities were reported. The failure to obtain urine cultures was not improved by the implementation of the guide, which states urinalysis, urine gram stain, and culture as part of the necessary laboratory tests for diagnosis.

The Guide to Empiric Antimicrobial Therapy was designed to be a resource for physicians aiding in the selection of the most appropriate antimicrobials based on typical infecting organisms and institutional resistance patterns. In July 2002, the Guide to Empiric Antimicrobial Therapy was distributed to all hospital physicians at the academic medical center. While there was an educational session at the time of implementation, no consistent reinforcement was provided to the physicians throughout the year. The lack of significant change in the physicians' prescribing habits after the implementation of the guide emphasizes the need for continued education regarding treatment pathways. In addition, the cyclic turnover of physicians, specifically medical residents, is routine at academic medical centers. The medical resident turnover may negatively impact adherence with treatment algorithms and further supports the need for ongoing educational efforts. The small patient population, retrospective design, and yearly medical resident turnover limit study results. The exclusion of patients who did not receive antimicrobial therapy limits the analysis of the asymptomatic bacteriuria treatment, as withholding treatment in non-pregnant patients is in accordance with the algorithm.

Over the past decade, there has been an increased sense of awareness of the value of utilizing practice guidelines. Even with this heightened awareness, academic medical centers often struggle in implementing practice guidelines, as dissemination of the guidelines requires the use of healthcare dollars and resources to be successful. Many studies have analyzed compliance with guidelines after implementation to find less than satisfactory use of the guidelines.9 Despite low compliance, physicians agree that utilizing guidelines is of value.9,10 Implementation strategies for practice guidelines should include educational programs at the time of implementation as well as a method for reminding physicians to utilize the guidelines.9 Reporting results of audits or medication use evaluations to the affected physician groups may further influence future consistency with existing algorithms or guidelines.9

Initially driven primarily by the principles of evidence-based medicine, guidelines are now being implemented to increase cost-efficient care and optimize health outcomes.10 The resources required to implement these guidelines can be justified if the guidelines actually produce cost-effective medical practices.11 An additional opportunity for increasing the cost-savings associated with UTI treatment at the University of Kentucky Chandler Medical Center may be provided by encouraging physicians to use available urine cultures to guide changes in empiric therapy to the most cost-effective antimicrobial agent.

Simple distribution of the first edition of the guide without reinforcement was not sufficient for truly impacting physician prescribing habits in the empiric treatment of UTIs. Noting the need for additional education regarding the empiric treatment pathways, the availability of the second edition was publicized in a feature article of the institutional Drug Information Center newsletter, which was distributed to all hospital physicians along with the revised guide. Additionally, the pathways are now available electronically through the pharmacy website. With the implementation of computerized physician order entry (CPOE), linking the selection of an antimicrobial with a request for a corresponding diagnosis becomes a possible mechanism for reinforcing established treatment pathways. Periodic evaluations of compliance with the algorithm may facilitate identification of particular services or individual physicians and prompt opportunities for further education or discussion.


After the implementation of an empiric antimicrobial therapy guide, a slight improvement in physicians' compliance with the empiric antimicrobial treatment algorithm for UTIs was noted. Increasing adherence to the UTI algorithm should maintain or improve patient outcomes while decreasing overall healthcare expenditures. While education of medical staff at the time of guideline implementation is fundamental, reinforcement of adherence to guidelines is equally important in ensuring the successful use of practice guidelines. Emphasizing the benefit of using clinical practice guidelines should continue beyond initial dissemination of any standards. Education regarding the guidelines must be perpetual to promote adherence and stronger evidence-based medical practices.

Dr LeClaire is a clinical specialist, critical care, at Shands at the University of Florida and an assistant clinical professor of pharmacy, University of Florida, Gainesville, Fla. She can be reached at
. Dr Lopez is a clinical specialist, pediatric hematology/oncology, at Children's National Medical Center, Washington, DC. At time of study, Drs LeClaire and Lopez were pharmacy residents at the University of Kentucky Chandler Medical Center, Lexington, Ky. Dr Smith is a clinical specialist, medication use policy, at the University of Kentucky Chandler Medical Center and an associate professor of pharmacy and science at the University of Kentucky. Dr Lewis is a clinical Pharmacy Specialist, Therapeutic Drug Monitoring, at the University of Kentucky Chandler Medical Center and adjunct assistant professor of pharmacy, University of Kentucky. Dr Rapp is manager of clinical services at the University of Kentucky Chandler Medical Center and a professor of pharmacy and surgery at the University of Kentucky. Dr Martin is a clinical specialist, infectious diseases, at the University of Kentucky Chandler Medical Center and an adjunct assistant professor of pharmacy, University of Kentucky.

Disclosure Information: Drs LeClaire, Lopez, and Lewis report no relevant financial disclosures as related to this article. Dr Smith has unrestricted educational grants from Aventis, Roche, Wyeth, and Pharmacia and a research grant from Roche. Dr Martin is on the speakers’ bureau for Ortho-McNeil and Wyeth, has grants from Ortho-McNeil, and consults for Ortho-McNeil and Wyeth. Dr Rapp is on the speakers’ bureau and consults for Pfizer, Ortho-McNeil, and Wyeth.


1. Saint S, Scholes D, Fihn SD, et al. The effectiveness of a clinical practice guideline for the management of presumed uncomplicated urinary tract infection in women. Am J Med. 1999;106: 636–641.

2. Hooton TM, Scholes D, Hughes JP, et al. A prospective study of risk factors for symptomatic urinary tract infection in young women. N Engl J Med. 1996;335:468–474.

3. Krieger JN, Ross SO, Simonsen JM. Urinary tract infection in healthy university men. J Urol. 1993;149:1046–1048.

4. Stamm WE, Norrby SR. Urinary tract infections: disease panorama and challenges. J Infect Dis 2001;183(Suppl 1):S1–S4.

5. Garnick DW, Hendrick AM, Brennan TA. Can practice guidelines reduce the number and costs of malpractice claims? JAMA. 1991;266: 2856–2860.

6. Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluation. Lancet. 1993;342: 1317–1322.

7. Fang E, Mittman BS, Weingarten S. Use of clinical practice guidelines in managed care physician groups. Arch Fam Med. 1996;5:528–531.

8. Warren JW, Abrutyn E, Hebel R, Johnson JR, Schaeffer AJ, Stamm WE. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Clin Infect Dis. 1999;29:745–758.

9. Davis DA, Taylor-Vaisey A. Translating guidelines into practice: a systematic review of theoretic concepts, practical experience and research evidence in the adoption of clinical practice guidelines. Can Med Assoc J. 1997;157:408–416.

10. Costantini O, Papp KK, Como J, Aucott J, Carlson MD, Aron DC. Attitudes of faculty, housestaff, and medical students toward clinical practice guidelines. Acad Med. 1999;74(10)1138–1143.

11. Stone TT, Kivlahan CH, Cox KR. Evaluation of physician preferences for guideline implementation. Am J Med Qual. 1999;14(4):170–177.