PILOT PROGRAM Development of a Pharmacist-Led Emergency Department Antimicrobial Surveillance Program Christopher A. Knefelkamp, PharmD, BCPS; and Christopher M. Degenkolb, PharmD, BCPS An analysis of antimicrobial agents used for urinary tract infections found effective initial empiric coverage, but fluoroquinolones were used in a majority of patients who may have been better served with narrower spectrum agents.

Dr. Knefelkamp is a Clinical n September 18, 2014, President helped to decrease the defined daily doses Specialist, Acute Barack Obama signed an executive of broad-spectrum per 1,000 pa- Care, and Dr. Degenkolb is a Clinical Pharmacy Oorder that made addressing - tient days nearly 36%, from 532 in 2012 to Specialist, Internal resistant bacteria a national security policy.1 343 in 2015, as well as decrease the days of , both at This legislation resulted in the creation of of fluoroquinolones per 1,000 pa- Richard L. Roudebush VA Medical Center in a large multidepartment task force to com- tient days 28.75%, from 80 in 2012 to Indianapolis, Indiana. bat the global and domestic problem of an- 57 in 2015. Additionally, the program Correspondence: timicrobial resistance. The order required showed a significant decrease in the stan- Dr. Knefelkamp hospitals and other inpatient health care dardized antimicrobial administration ratio, (christopher.knefelkamp @va.gov) delivery facilities, including the Department a benchmark measure developed by the of Veterans Affairs (VA), to implement ro- CDC to reflect a facility’s actual antimicrobial bust antimicrobial stewardship programs use to the expected use of a similar facility that adhere to best practices, such as those based on bed size, number of intensive care identified by the Centers for Disease Con- unit beds, location type, and trol and Prevention (CDC). More specifi- affiliation.2 cally, the VA was mandated to take steps While the RRVAMC antimicrobial stew- to encourage other areas of health care, ardship team has been able to intervene on such as ambulatory centers and most of the inpatients admitted to the medi- outpatient clinics, to adopt antimicrobial cal center, the outpatient arena has had few stewardship programs.1 This order also re- antimicrobial stewardship interventions. inforced the importance for VA facilities to Recognizing a need to establish and expand continue to develop, improve, and sustain pharmacy services and for improvement efforts in antimicrobial stewardship. of outpatient antimicrobial stewardship, Prior to the order, in 2012 the Richard L. RRVAMC leadership decided to establish a Roudebush VA Medical Center (RRVAMC) pharmacist-led outpatient antimicrobial sur- in Indianapolis, Indiana, implemented an in- veillance program, starting specifically within patient antimicrobial stewardship program the emergency department (ED). that included thrice-weekly meetings to re- Clinical pharmacists in the ED setting view inpatient records and make stewardship are uniquely positioned to improve pa- recommendations with an infectious dis- tient care and encourage the judicious use eases champion and clinical phar- of antimicrobials for empiric treatment of macists. These efforts led to the improved urinary tract infections (UTIs). The CDC’s use of antimicrobial agents on the inpatient Core Elements of Outpatient Antibiotic Stew- side of the medical center. During the first ardship recommends pharmacist availabil- 4 years of implementation, the program ity in the ED setting, and previous literature

38 • FEDERAL PRACTITIONER • JULY 2018 mdedge.com/fedprac has demonstrated pharmacist utility in TABLE Top 5 Urine Culture Results ED postdischarge culture monitoring and surveillance.3-5 Organisms Cultures, No. This article will highlight one such pro- gram review at the RRVAMC and dem- Escherichia coli 36 onstrate the need for pharmacist-led No growth 16 antimicrobial stewardship and monitoring in the ED. The purpose of this study was Entercocus faecalis 14 to test the hypothesis that pharmacist in- tervention would be necessary to prospec- Kiebsiella pneumoniae 13 tively check for “bug-drug mismatch” and assure proper follow-up of urine cultures < 10,000 CFU, GPC: not identified 7 at this institution. The project was deemed to be quality improvement and thereby Abbreviations: CFU, colony-forming units; GPC, Gram-positive cocci. granted exemption by the RRVAMC Institu- tional Review Board. not within the direct practice area during the final 61 days of the project but was in a dif- METHODS ferent area of the hospital and available for This project took place at the RRVAMC, consultation. a 229-bed tertiary academic medical cen- Primary data collected included urine cul- ter that serves > 60,000 patients annually. ture results and susceptibilities, empiric an- The RRVAMC ED has 20 beds and received timicrobial choices, and admission status. about 29,000 visits in 2014. Patients were Other data collected included duration of eligible for initial evaluation if they had a treatment and secondary antibiotics chosen, urine culture collected in the ED within the each of which specifically evaluated those 91-day period from September 1, 2015 to patients who were not admitted to the hos- November 30, 2015. Patients were included pital and were thus treated as outpatients. for data analysis if it was documented that Additional data generated from this study they were treated for actual or clinically were used to identify empiric antibiotics uti- suspected, based on signs and symptoms, lized for the treatment of UTIs and assess for uncomplicated UTI, complicated UTI, or appropriate selection and duration of therapy UTI with pyelonephritis. Patients did not within this institution. need to have a positive urine culture for in- clusion, as infections could still be present RESULTS despite negative culture results.6 Patients During the study period, 722 urine cultures with positive cultures who were not clini- were collected in the ED and were included cally symptomatic of a UTI and were not for initial evaluation. Of these, 127 were treated as such by the ED provider (ie, treated by the ED provider pursuant to one asymptomatic bacteriuria) were excluded of the indications specified and were in- from the study. cluded in the data analysis. Treatment with Data collection took place via daily chart an antimicrobial agent provided adequate review of patient records in both the Com- coverage for the identified pathogen in puterized Patient Record System and De- 112 patients, yielding a match rate of 88%. centralized Hospital Computer Program As all included cultures were collected in medical applications as urine cultures were suspicion of an infection, those cultures performed. Data were gathered and assessed yielding no growth were considered to have by a postgraduate year-2 been adequately covered. Overall, the most pharmacy resident on rotation in the ED frequently grown pathogen identified via who reviewed cultures daily and made in- culture was Escherichia coli (E coli), appear- terventions based on the results as needed. ing in 36 of 127 cultures (28%) (Table). The pharmacy resident was physically pres- Susceptibility analysis of the E coli cultures ent within the ED during the first 30 days revealed a fluoroquinolone susceptibility of the project. The pharmacy resident was rate of 75%.

mdedge.com/fedprac JULY 2018 • FEDERAL PRACTITIONER • 39 Antimicrobial Surveillance

FIGURE 1 Empiric Antibiotic Choices for Outpatient Treatment

Ciprofloxacin Cefuroxine 17% Levofloxacin 1% and 3% Sulfamethoxazole/trimethoprin 48% 7% Amoxicillin/clavulanate Nitrofurantoin Cephalexin

11%

15%

1%

Nearly half (45%) of treatment plans rotation in the ED. All culture results were included a fluoroquinolone. Of those addressed on the day the organism’s identifi- treated on an outpatient basis, fluoroqui- cation and susceptibility data was released by nolones were even more frequently used, the microbiology lab (Figure 2). comprising 50 of 82 (61%) courses. Cip- rofloxacin was the most frequently used DISCUSSION treatment, used in 39 of the 82 outpa- Empiric antibiotic selection for the treat- tient regimens (48%). Cephalexin was the ment of UTIs continues to be the corner- second most common and was used in stone of antibiotic management for the 14 outpatient regimens (17%), followed treatment of such a disease state.7 The by levofloxacin (15%) (Figure 1). The noted drug-bug match rate of 88% in this average duration of treatment for outpa- study demonstrates effective initial em- tient therapy was 8.3 d (range: 3-14 d). Of piric coverage and ensures a vast major- the 50 patients who received a fluoroqui- ity of veterans receive adequate coverage nolone for outpatient treatment, 33 could for identified pathogens. Additionally, this have received a narrower spectrum anti- rate shows that the current system seems microbial for adequate treatment based on to be functioning appropriately and re- final sensitivities. futes the author’s preconceived ideas that Mismatched cultures, or those where the the mismatch rate was higher at RRVAMC. prescribed antibiotic did not provide ade- However, these findings also demonstrate quate coverage of the identified pathogens a predominant use of fluoroquinolones for based on susceptibilities, occurred at a rate empiric treatment in a majority of patients of 12%. Follow-up on these cultures was de- who could be better served with narrower termined largely by the patient’s admission spectrum agents. Only 2 of the outpatient status. The majority of mismatched cultures regimens were for the treatment of pyelone- were addressed by the inpatient team (10/15) phritis, the only indication in which a fluo- upon admission. Of those patients who were roquinolone would be the standard of care not admitted, 3 cultures were addressed by per guideline recommendations.7 the patient’s primary care physician’s clinic, These findings were consistent with a sim- 1 was addressed by a specialty clinic, and ilar study in which 83% of ED collected urine 1 was addressed by the pharmacy resident on cultures ultimately grew bacteria susceptible

40 • FEDERAL PRACTITIONER • JULY 2018 mdedge.com/fedprac Antimicrobial Surveillance

FIGURE 2 Culture Disposition

1 culture addressed by pharmacist 112 cultures with adequate antibiotic coverage 127 cultures treated 1 culture addressed by by emergency specialty care clinic department provider 15 cultures without 722 urine cultures adequate antibiotic collected 595 cultures not coverage 10 cultures addressed treated by emergency by inpatient medicine department team provider

3 cultures addressed by primary care clinic

to empiric treatment.8 This number was nolone use, to be similar to those collected in similar to the current study despite the lat- the ED alone (78.5% hospital-wide suscep- ter study consisting of predominantly female tibility vs. 75% ED susceptibility). This was patients (93%) and excluding patients with expected, as similar studies comparing E coli a history of benign prostatic hypertrophy, resistance patterns from ED-collected urine catheter use, or history of genitourinary can- cultures to those institution-wide also have cer, which are frequently found within the found similar rates of resistance.8,9 These VA population. Thus, despite having a differ- findings are of particular importance as ing patient population at the current study’s E coli resistance is noted to be increasing, var- facility with characteristics that would clas- ies with geographic area, and local resistance sify most to be treated as a complicated UTI, patterns are rarely known.7 Thus, these find- empiric coverage rates remained similar. The ings may aid ED providers in their empiric lower than anticipated intervention rate by antimicrobial selections. the pharmacist on rotation in the ED can be Ciprofloxacin was the most frequently directly attributed to this high empiric match used medication for the treatment of UTIs. rate, which could in turn be attributed to the While overall empiric selections were found extensive use of broad-spectrum antibiotics to have favorable resistance patterns, it is dif- for treatment. ficult to interpret the appropriateness of cip- Empiric antimicrobial selection is based rofloxacin’s use in the present study. First, largely on local resistance patterns.7 Of par- there is a distinct lack of US-based clinical ticular importance is the resistance patterns practice guidelines for the treatment of com- of E coli, as it is the primary isolate respon- plicated UTIs. As the majority of this study sible for UTIs worldwide. Thus, it is not un- population was male, it is difficult to directly expected that the most frequently isolated extrapolate from the current Infectious Dis- pathogen in the current study also was E coli. eases Society of America treatment guide- While clinical practice guidelines state that lines for uncomplicated cystitis and apply hospital-wide antibiograms often are skewed to the study population. Although recom- by cultures collected from inpatients or those mended for the treatment of pyelonephritis, with complicated infection, the current study it is unclear whether ciprofloxacin should be found hospital-wide E coli resistance pat- utilized as a first-line empiric option for the terns, specifically those related to fluoroqui- treatment of UTIs in males.

mdedge.com/fedprac JULY 2018 • FEDERAL PRACTITIONER • 41 Antimicrobial Surveillance

Despite the lack of disease-specific rec- providers in an area of larger uncertainty ommendations for ciprofloxacin, recommen- with regards to treating both complicated and dations exist regarding its use when local uncomplicated cystitis. Given the enhanced resistance patterns are known.7 It is currently warnings on fluoroquinolone use, it is un- recommended that these agents not be used known whether prescribers would gravitate when resistance rates of E coli exceed 20% to utilizing similar options as their peers as for trimethoprim-sulfamethoxazole or 10% alternatives to fluoroquinolones. Similarly, for fluoroquinolones. As this study demon- duration of therapy with nonfluoroquinolone strated a nearly 25% resistance rate for E coli agents is unclear as well; as the present study to fluoroquinolones in both the ED and in- demonstrated a large range in treatment du- stitution-wide sample populations, it could ration of outpatients (3-14 days). While the potentially be ascertained that ciprofloxacin average observed duration of 8.3 days is in- is an inappropriate choice for the empiric tuitively fitting, as the majority of cases were treatment of UTIs in this patient population. in males, no published guideline exists that However, as noted, it is unknown whether affirms the appropriateness of this finding. this recommendation would still be ap- Such uncertainty and potential inconsistency plicable when applied to the treatment of between providers affords a large opportunity complicated cystitis and greater male pop- for developing a standardized treatment path- ulation, as overall rates of susceptible cul- way for the treatment of UTIs to ensure both tures to all organisms was similar to other effective and guideline concordant treat- published studies.8,9 ment for patients, specifically with regards While there is scant specific guidance re- to antimicrobial selection and duration of lated to the treatment of complicated UTIs, treatment. there is emerging guidance on the use of flu- It is noteworthy to mention that all fol- oroquinolones, both in general and specifi- low-ups on positive cultures inadequately cally related to the treatment of UTIs. In July covered by empiric therapy took place on 2016, the FDA issued a drug safety commu- the day organism identification and suscep- nication regarding the use of and warnings tibility data were released. This finding was for fluoroquinolones, which explicitly stated somewhat surprising, as it was originally the- that “health care professionals should not orized that most ED-collected urine cultures prescribe systemic fluoroquinolones to pa- were not monitored to completion by a phar- tients who have other treatment options for macist and that would be necessary in order acute bacterial sinusitis, acute bacterial exac- to ensure proper follow-up of culture results. erbation of chronic bronchitis, and uncom- What is not clear is whether there is a robust plicated UTIs because the risks outweigh the process for the follow-up of urine cultures in benefits in these patients.”10 the ED. Most of the bug-drug mismatches co- This guidance has the potential to impact incidentally were admitted to the inpatient fluoroquinolone prescribing significantly at teams where there were appropriate person- RRVAMC. Given the large number of fluo- nel to follow up and adjust the antibiotic se- roquinolones prescribed for UTIs, the down- lection. If there was a bug-drug mismatch, stream effects that this shift in prescribing and the patient was not admitted, it is un- would have is unknown. As most nonflu- clear whether there is a consistent process for oroquinolones used for UTI typically are follow-up. narrower in antimicrobial spectrum (eg, tri- Given the limited number of mis- methoprim/sulfamethoxazole, nitrofuran- matched cultures that required change in toin, cephalexin, etc) the possibility exists therapy, it is unknown if this role would ex- that the match rate for empiric therapy may pand if more narrow-spectrum agents were decrease. Thus, a larger need for closer fol- utilized, theoretically leading to a higher low-up to assure adequate coverage may mismatch rate and necessitating closer arise, posing a more expanded role for an follow-up. Furthermore, given the common ED-based pharmacist than was demon- practice of mailed prescriptions at the VA, strated in the current study. it is all the more imperative that the cul- This new guidance also may place tures be acted upon on the day they were

42 • FEDERAL PRACTITIONER • JULY 2018 mdedge.com/fedprac Antimicrobial Surveillance

identified, as the mailing and processing time with regards to antimicrobial stewardship at of prescriptions may limit the clinical util- this facility on the inpatient side, no formal ity in switching from a more broad-spectrum antimicrobial stewardship program exists for agent, to one more targeted for an identified review in the outpatient setting, where liter- organism. While a patient traveling back to ature suggests the majority of antibiotics are the medical center for expedited prescrip- prescribed.3,11 While more robust protocols tion pickup at the pharmacy would allevi- are in place for follow-up of culture data in ate this problem, many patients at the facility the primary care realm at this facility, the pre- travel great distances or may not have read- scribing patterns are relatively unknown. ily available travel means to return to the A recent study completed at a sim- medical center. ilar VA facility found that 60% of antibi- otics prescribed for cystitis, pharyngitis, Future Directions or sinusitis on an outpatient basis were While minimal follow-up was required after guideline-discordant, and CDC guidance a patient had left the ED, this study dem- has further recommended specific focus onstrated a fundamental need for further should be undertaken with regards to out- refinements in antimicrobial stewardship patient stewardship practices in the treat- activities within the ED. Duration of ther- ment of genitourinary infections.3,12 These apy, empiric selection, and proper dosing findings highlight the need for outpatient are key areas where the ED-based phar- antimicrobial stewardship and presents a macy resident was able to intervene dur- compelling reason to further investigate ing the time physically stationed in the ED. outpatient prescribing within primary care The data collected from this study demon- at RRVAMC. strated this and was ultimately combined with other ED-based interventions and uti- Strengths and Limitations lized as supporting evidence in the phar- Strengths of the current study include the macy service business plan, outlining the ability to monitor urine cultures in real necessity of a full-time pharmacy presence time and to provide timely interventions in in the ED. The business plan submission, the event of a rare bug-drug mismatch. The along with other ongoing RRVAMC initia- evaluation of cultures in this study shows tives, ultimately led to the approval for clin- that the majority of cases had a drug se- ical pharmacy specialists to expand practice lected with adequate coverage. The study into the ED. These positions will continue did assure ED providers that, even though to advance pharmacy practice within the guidelines may suggest otherwise, urine ED, while affording opportunities for phar- cultures drawn in the ED at RRVAMC fol- macists to practice at the top of their li- lowed similar resistance patterns seen for censure, provide individualized provider the facility as a whole. Moreover, it is valu- education, and deliver real-time antimicro- able as it captures data that are directly ap- bial stewardship interventions. Further- plicable to the VA patient population, in more, as the majority of the study period which there is little published data with re- was monitored outside of the ED, the proj- gards to UTI treatment and no formal VA ect may provide a model for other VA insti- guidance. tutions without full-time ED pharmacists to A primary limitation of this study is the implement as a means to improve antimi- lack of differentiation between cultures col- crobial stewardship and further build an ev- lected from patients with or without in- idence base for expanding their pharmacy dwelling catheters. However, only including services to the ED. patients who presented with signs and/or Given the large number of fluoroqui- symptoms of a UTI limits the number of cul- nolones utilized in the ED, this study has tures that could potentially be deemed as col- raised the question of what prescribing pat- onization, thus minimizing the potential for terns look like with regards to outpatient UTI nonpathogenic organisms to confound the treatment within the realm of primary care results. This study also did not differentiate at RRVAMC. Despite the great strides made the setting from which the patient presented

mdedge.com/fedprac JULY 2018 • FEDERAL PRACTITIONER • 43 Antimicrobial Surveillance

(eg, community, extended care facility, etc) investigational use of certain drugs. Please review the com- plete prescribing information for specific drugs or drug com- that could have potentially provided guid- binations—including indications, contraindications, warnings, ance on resistance patterns for community- and adverse effects—before administering pharmacologic acquired UTIs and whether this may have therapy to patients. differed from hospital-acquired or facility- References acquired UTIs. Another limitation was the 1. Obama B. Executive order–combating antibiotic- resistant bacteria. https://www.whitehouse.gov/the relatively short time frame for data collection. -press-office/2014/09/18/executive-order-combating A data collection period greater than 91 days -antibiotic-resistant-bacteria. Published September 18, would allow for a larger sample size, thus 2014. Accessed June 7, 2018. 2. Livorsi DJ, O’Leary E, Pierce T, et al. A novel metric to making the data more robust and potentially monitor the influence of antimicrobial stewardship activi- allowing for the identification of other trends ties. Infect Control Hosp Epidemiol. 2017;38(6):721-723. 3. Sanchez GV, Fleming-Dutra KE, Roberts RM, Hick LA. not seen in the current study. A longer data Core elements of outpatient antibiotic stewardship. collection period also would have afforded MMWR Recomm Rep. 2016;65(6):1-12. the opportunity to track more robust clini- 4. Wymore ES, Casanova TJ, Broekenmeier RL, Martin JK Jr. Clinical pharmacist’s daily role in the emergency depart- cal outcomes throughout the study, identify- ment of a community hospital. Am J Health-Syst Pharm. ing whether treatment failure may have been 2008;65(5):395-396, 398-399. 5. Frandzel S. ED pharmacists’ value on display at ASHP linked to the use of certain classes or spec- Midyear. http://www.pharmacypracticenews.com trums of activity of antibiotics. /ViewArticle.aspx?ses=ogst&d_id=53&a_id=22524. Pub- lished February 14, 2013. Accessed June 15, 2018. 6. Heytens S, DeSutter A, Coorevits L, et al. Women with CONCLUSION symptoms of a urinary tract infection but a negative urine Despite the E coli resistance rate to ciproflox- culture: PCR-based quantification of Escherichia coli suggests infection in most cases. Clin Microbiol Infect acin (> 20%), the empiric treatments cho- 2017;23(9)647-652. sen were > 85% effective, needing minimal 7. Gupta K, Hooton TM, Naber KG, et al. International clinical follow-up once a patient left the ED. None- practice guidelines for the treatment of acute uncompli- cated cystitis and pyelonephritis in women: A 2010 update theless, a change in prescribing patterns based by the Infectious Diseases Society of America and the Eu- on recent national recommendations may ropean Society for Microbiology and Infectious Diseases. Clin Infect Dis. 2011;52(5):e103-e120. provide expanded opportunities in antimicro- 8. Lingenfelter E, Drapkin Z, Fritz K, Youngquist S, Madsen bial stewardship for ED-based pharmacists. T, Fix M. ED pharmacist monitoring of provider antibiotic Further research is needed in antimicrobial selection aids appropriate treatment for outpatient urinary tract infection. Am J Emerg Med. 2016;34(8):1600-1603. stewardship within this facility’s outpatient 9. Zatorski C, Jordan JA, Cosgrove SE, Zocchi M, May L. primary care realm, potentially uncovering Comparison of antibiotic susceptibility of Escherichia coli in urinary isolates from an emergency department with other opportunities for pharmacist inter- other institutional susceptibility data. Am J Health-Syst vention to assure guideline concordant care Pharm. 2015;72(24):2176-2180. for the treatment of UTIs as well as other 10. US Food and Drug Administration. FDA drug safety com- munication: FDA updates warnings for oral and injectable infections treated in primary care patients. fluoroquinolone antibiotics due to disabling side effects. https://www.fda.gov/Drugs/DrugSafety/ucm511530.htm. Author disclosures Updated March 8, 2018. Accessed June 13, 2018. The authors report no actual or potential conflicts of interest 11. Llor C, Bjerrum L. : risk associated with regard to this article. with antibiotic overuse and initiatives to reduce the prob- lem. Ther Adv Drug Saf. 2014;5(6):229-241. Disclaimer 12. Meyer HE, Lund BC, Heintz BH, Alexander B, Egge JA, The opinions expressed herein are those of the authors and Livorsi DJ. Identifying opportunities to improve guideline- do not necessarily reflect those of Federal Practitioner, Front- concordant antibiotic prescribing in veterans with acute line Medical Communications Inc., the US Government, or respiratory infections or cystitis. Infect Control Hosp Epide- any of its agencies. This article may discuss unlabeled or miol. 2017;38(6):724-728.

44 • FEDERAL PRACTITIONER • JULY 2018 mdedge.com/fedprac