Antimicrobial Stewardship at a Large Tertiary Care Academic Medical Center: Cost Analysis Before, During, and After a 7-Year Program
Background. An antimicrobial stewardship program was fully implemented at the University of Maryland Medical Center in July 2001 (beginning of fiscal year [FY] 2002). Essential to the program was an antimicrobial monitoring team (AMT) consisting of an infectious diseases–trained clinical pharmacist and a part-time infectious diseases physician that provided real-time monitoring of antimicrobial orders and active intervention and education when necessary. The program continued for 7 years and was terminated in order to use the resources to increase infectious diseases consults throughout the medical center as an alternative mode of stewardship.
Design. A descriptive cost analysis before, during, and after the program.
Patients/setting. A large tertiary care teaching medical center.
Methods. Monitoring the utilization (dispensing) costs of the antimicrobial agents quarterly for each FY.
Results. The utilization costs decreased from $44,181 per 1,000 patient-days at baseline prior to the full implementation of the program (FY 2001) to $23,933 (a 45.8% decrease) by the end of the program (FY 2008). There was a reduction of approximately $3 million within the first 3 years, much of which was the result of a decrease in the use of antifungal agents in the cancer center. After the program was discontinued at the end of FY 2008, antimicrobial costs increased from $23,933 to $31,653 per 1,000 patient-days, a 32.3% increase within 2 years that is equivalent to a $2 million increase for the medical center, mostly in the antibacterial category.
Conclusions. The antimicrobial stewardship program, using an antimicrobial monitoring team, was extremely cost effective over this 7-year period.
Antimicrobial resistance is increasing throughout the United States. Furthermore, there are fewer new antimicrobials being developed to treat these resistant organisms.1 This has led the Infectious Diseases Society of America, the Society for Health Care Epidemiology, and the Centers for Disease Control and Prevention to recommend that hospitals develop antimicrobial stewardship programs (ASPs) to help protect our existing armamentarium as long as possible.2 Despite this recommendation, many medical centers have not established active programs. Such programs require resources during times when competition for finances is great and return of investment is uncertain.
The University of Maryland Medical Center (UMMC) developed an ASP using an antimicrobial monitoring team (AMT) in calendar year 2001. However, after 7 years the program was discontinued in favor of using the resources to provide additional infectious diseases physicians to enhance infectious diseases consultation throughout the medical center. The rationale for this was that the infectious diseases physicians, via consultations, would provide the necessary stewardship, making the AMT superfluous. This article documents the marked cost savings that resulted after the implementation of the ASP and the AMT in 2001 and the major cost impact that resulted after the program was discontinued.
The ASP was established at UMMC in 2001, in response to the perceived need for more appropriate use and the escalating costs of antimicrobial agents. It was continued through June 2008. At the time of the implementation of the ASP, UMMC had approximately 28,700 patient admissions per year; this has steadily increased, to 38,590 admissions in fiscal year (FY) 2010. The medical center has a very active cancer center, trauma center, and transplantation program. It has approximately 175 intensive care beds in the facility.
At the time of initiating the program, UMMC had a regulated formulary with restricted antimicrobials that required preauthorization for their use and an active pharmacy and therapeutics committee with an antimicrobial subcommittee. Despite this, annualized antimicrobial purchasing costs increased from $4.7 million in calendar year (CY) 1997 to over $8 million in CY 2000. Additionally, hospitals in Maryland were to be reimbursed for the disease entity treated (a drug-related group [DRG] system) instead of for the expenses incurred. This provided an additive monetary incentive for initiating the ASP.
Developing the ASP and the AMT
Leadership of the initiative was established within the infection control program and by working closely with the pharmacy department. The program was developed by the medical director for infection control and the infectious diseases clinical pharmacist and vetted at a number of meetings with key hospital personnel, including the director of pharmacy, the chief of infectious diseases, the vice president of quality for the medical center, and the chief medical officer. Essential and central to the ASP was the establishment of an AMT. This team consisted of (1) an infectious diseases physician with an initial dedicated effort to the program of 25% that subsequently increased to 50% and (2) a clinical pharmacist with infectious diseases training with a dedicated effort to the program of 80%. A data analyst with a dedicated effort of 5% and direction of the program was incorporated in the infection control program and required no additional resources.
Responsibilities of the AMT
The duties of the team were to provide an active computer-assisted real-time review of antimicrobial orders for the designated restricted antimicrobials and to provide active intervention when necessary. During their review, the team attempted to (1) identify ineffective or excessive antimicrobial coverage, (2) assure that the orders adhered to policies and guidelines, (3) discontinue unnecessary double coverage, (4) determine patients whose treatment could be converted safely from parenteral to oral therapy (IV-PO), and (5) suggest infectious diseases consults for difficult and complex cases. The team prioritized for review those patients receiving restricted antimicrobial agents and those areas of the medical center not served by specialized infectious diseases physicians such as were present in the trauma center.
Preauthorization Using the “BUGS Beeper”
Prior to the stewardship program, preauthorization for the use of certain antimicrobial agents was required, and this consent was provided by the infectious diseases consult service. To improve the process, a “BUGS beeper” program was developed. In this a beeper, activated by dialing “BUGS,” was carried by the on-duty infectious diseases fellows on a rotational schedule. The fellow was supervised by the infectious diseases–trained faculty member assigned to the antimicrobial team. If preauthorization was not provided, the attending physician on the requesting service could directly call the attending physician on the antimicrobial team. If the conflict was still not resolved, a dose of the antimicrobial agent would be administered and an infectious diseases consult would be obtained in order to clarify the situation. Initially, an on-duty fellow carried the beeper until 10:00 p.m., but this was soon extended to 24 hours, 7 days a week. The program was initiated in late January 2001 and became fully implemented by July 2001 (beginning of FY 2002).
Development of Guidelines and Policies
To provide consistency for the physicians ordering the antimicrobials and for the antimicrobial team monitoring their use, guidelines and policies were developed where applicable. The AMT frequently provided leadership when formulating these guidelines, but attempts were made to have those most involved with their use participate in the development. When the policies and guidelines were written, they were approved through the normal channels, including the antimicrobial subcommittee of the pharmacy and therapeutics (P&T) committee, the parent P&T committee, and the medical executive committee, when appropriate.
The financial goals for the antimicrobial program, which were based on those of other programs using an active interventional model at the time, were purposefully conservative: to save 10%–20% of the costs of antibiotics, based on an antimicrobial budget of $6 million per year over a 3-year period (a savings of $600,000–$1,200,000 over the 3-year period).3
The actual costs for the antimicrobials were determined by the pharmacy service, and they represented the purchasing cost for the drug per unit. These costs were obtained by the pharmacy administration, independent of the antimicrobial team. The data program captured the utilization of antimicrobials, which was defined as those antimicrobials ordered by the provider who was caring for the patient and dispensed by the pharmacy. Any unused medication returned to the pharmacy was subtracted from this amount. Initially, these utilization data were captured using the Mega Source program. The data in this program were transferred using Monarche into Access, so that they could be categorized and manipulated into a usable database. On October 7, 2002 (second quarter of FY 2003), PharmNet was initiated. During this changeover, there were no data collected for the first 7 days of that quarter. Instead, data for this period were estimated by extrapolating costs for the remaining portion of the quarter. Data from the PharmNet program were transferred into Access so that they could be categorized in a manner similar to the data from the Mega Source program.
Beginning in May 2004, PharmWatch (Cereplex; now owned by Premier), a decision-support program designed to assist in antimicrobial utilization, was used to evaluate for a 3-month period one-half of the patient population monitored by the antimicrobial team. Results of this evaluation have been reported previously.4 The use of this program was subsequently expanded to include the entire hospital.
Defined Daily Doses (DDDs)
DDDs were determined according to dosages recommended by the World Health Organization (http://www.whocc.no/atc_ddd_index/). When no daily dosage was suggested by that organization, one was assigned that was thought to represent a typical daily dosage for adults with normal renal and hepatic functions. When a dosage amount was assigned, it was not changed throughout the evaluation period. The DDDs are expressed per 1,000 patient-days.
To monitor the safety of the stewardship program, we monitored selective quality indicators for the medical center including length of stay, readmissions within 30 days, and 30-day mortality. The DRG case mix index was monitored to ensure that changes in outcomes were not related to this index.
The χ2 test was used to compare the annual and cumulative reductions of antibiotic usage overall and then after introduction and after discontinuation of the AMT. To examine the trend in the prevalence of antimicrobial utilization over time, the results were compared using the χ2 test for trend. For all analyses, the threshold for establishing statistical significance was set at . Statistical analyses were completed using the SPSS statistical package (ver 16).
Cost Savings after Implementation of the ASP
The utilization costs by quarter for the medical center from FY 1998 through FY 2010 are presented in Figure 1. (For example, the fiscal year for 2001 extends from July 1, 2000, through June 30, 2001.) The overall upward trend of costs that occurred from FY 1998 through the third quarter of FY 2001 (prior to the time of implementation of the ASP) is readily apparent. Similarly, the downward trend that occurred after the program was fully implemented (beginning in FY 2002) is marked, particularly in the first 3 years. Thereafter, a relatively stable period exists from FY 2004 until the first quarter of FY 2009, at which point the program was terminated. Following this termination, the costs for antimicrobials increase dramatically for the next 2 years.
The yearly costs before the stewardship program was implemented (FY 2001), for the 7 years of the program’s existence (FY 2002–FY 2008), and for 2 years after it was terminated (FY 2009–FY 2010), including dollars per patient-days, are presented in Table 1. These costs are further separated into antimicrobial categories. The total expenditures for antimicrobial agents were reduced from $44,181 to $23,933 per 1,000 patient-days (45.8%; ) over the duration of the program. In terms of costs for the hospital, a reduction of $2,949,705 occurred for the medical center within the first 3 years after implementation of the ASP.
|FY 2001||FY 2002||FY 2003||FY 2004||FY 2005||FY 2006||FY 2007||FY 2008||FY 2009||FY 2010|
|Total per 1,000 patient-days||44,181||35,974||30,951||27,718||27,031||28,146||27,363||23,933||27,833||31,653|
|Savings (loss) from previous year||1,284,357||822,338||843,010||(269,917)||(132,690)||(88,358)||539,185||(1,093,101)||(873,184)|
By antimicrobial category, for FY 2001 (baseline), almost one-half ($3.7 million) of the entire budget of $7,774,588 went to pay for antifungal agents, with an additional $3.5 million (45%) paying for antibacterial agents. Together, these 2 categories were responsible for 93% of the antimicrobial costs. After the program began, the majority of the cost savings occurred in the antifungal category, which were reduced by $2,251,976 (60.7%) over the 3-year period, driven primarily by treatment guidelines for fungal infections in the medical center (). However, costs of antibacterial agents were also reduced by $513,044 (14.6%) over this 3-year period after the program was implemented (). Reductions are noted for the beta-lactam antibiotics as a group (10.2%), which include the carbapenems (primarily imipenem), the cephalosporins, and the penicillins (see Table 2). There was a slight reduction in the use of the antipseudomonal penicillins, primarily piperacillin/tazobactam, whereas use of the aminopenicillins (primarily ampicillin/sulbactam) increased. Reductions were also noted for the miscellaneous antimicrobial agents as a group, including vancomycin, quinipristin/dalfopristin, and metronidazole. The cost of the quinolones was reduced by 47% over the 3-year period. From FY 2004 to FY 2008, costs appeared to stabilize, decreasing by only $48,220 (1%); however, when patient-days are considered, this was a decrease of $3,785 per 1,000 patient-days, or 13.7%
|FY 2001||FY 2002||FY 2003||FY 2004||FY 2005||FY 2006||FY 2007||FY 2008||FY 2009||FY 2010|
Switch from Intravenous to Oral Delivery
An early intervention initiated by the stewardship program was the switch from intravenous to oral routes of delivery when the oral intake of other drugs was apparent and when the bioavailability of the antimicrobial agents permitted it. This was subsequently instituted as a policy. The reduction of costs resulting from this initiative was $179,285 in FY 2002 compared with baseline (FY 2001). The savings resulting from this switch were most apparent for fluconazole ($142,534) and linezolid ($19,597).
The increases in costs by antimicrobial category after the program was discontinued also is noted in Table 1, with the effects on specific antimicrobials presented in Table 2. There was an immediate increase in cost of $1 million during the first year after the program was discontinued and an additional increase of $873,184 during the second year, which represents an increase in cost of 41.2% for the 2-year period over the last year of the program (FY 2008; ). After the discontinuation of the program, the increased costs primarily occurred in the antibacterial category, particularly the agents acting against gram-positive organisms, including vancomycin, linezolid, and daptomycin (see Table 2; ). Also during this 2-year period, the costs of the antifungals as a class increased by approximately $100,000 (9.9%; ).
Utilization Costs by Hospital Location
Cost savings after the program was implemented were observed in all 3 major areas of the medical center: the cancer center, the shock trauma center, and the main hospital (Table 3). However, the savings that occurred during the first 3 years of implementation were most apparent in the cancer center (∼$2,000,000). Cost savings also occurred in the medical intensive care unit (MICU), the surgical ICU (SICU), and the transplantation service.
|FY 2001||FY 2002||FY 2003||FY 2004||FY 2005||FY 2006||FY 2007||FY 2008||FY 2009||FY 2010|
Discontinuation of the Program
After the program was discontinued, cost increases were most marked in the main portion of the medical center ($1,585,957), with very little added cost occurring in the cancer center. Costs in the shock trauma center also increased, but a new 12-bed ICU opened in this area at that time. Increases in costs after the ASP was terminated were also noted in the MICU and the SICU.
DDDs per 1,000 Patient-Days
Since 2004, we were able to obtain DDDs per 1,000 patient-days for all of the antimicrobial agents. These are indicated in Table 4 for selected antimicrobials and for the antimicrobial categories. Overall, there was a significant decrease in DDDs from FY 2004 through FY 2008, when the program was terminated. Total antimicrobial DDDs per 1,000 patient-days decreased by 439 (29%; ), and for antibacterial agents they decreased by 323 (27.5%; ). DDDs per 1,000 patient-days for antifungals and antivirals were also reduced, by 24% and 57%, respectively (). After the program was terminated, the overall DDDs per 1,000 patient-days for all antimicrobials increased minimally during the 2-year period (5.2%; ). However, there was an increase in the uses of cefepime (14.3%), piperacillin/tazobactam (10.5%), and antibacterials against gram-positive organisms, including linezolid (21.2%), daptomycin (113%), and vancomycin (32.3%). Echinocandin usage also increased, by 121%.
|During program||After program|
|FY 2004||FY 2005||FY 2006||FY 2007||FY 2008||FY 2009||FY 2010|
Quality indicators, including length of stay, readmissions, and mortality, are noted in Table 5. There were no significant changes in these parameters before, during, or after the program. There were no significant changes in the DRG case mix index after the program ended, indicating that changes in antimicrobial use were not caused by a change in the severity of our cases.
|FY 2001||FY 2002||FY 2003||FY 2004||FY 2005||FY 2006||FY 2007||FY 2008||FY 2009||FY 2010|
|% of admissions||5.30||4.50||4.10||4.90||4.96||4.99||5.80||6.15||6.02||5.75|
|% of admissions||3.0||3.1||3.1||2.9||2.9||2.6||2.8||2.7||2.7||2.5|
ASPs have been recommended to prevent antimicrobial resistance, decrease disease from Clostridium difficile infections, and curb adverse reactions to antimicrobials.5-8 Nevertheless, establishing stewardship programs requires resources. This report focuses on the cost analysis of such a program and demonstrates that the establishment of an ASP incorporating an AMT can be very cost effective in a large tertiary care teaching medical center.
This program resulted in a marked decrease in costs immediately after the program was initiated and a decrease in costs of 37% within the first 3 years. In terms of dollars, this was equivalent to a $3 million decrease in yearly costs over this period of time. Most of this decrease was related to the antifungal category of antimicrobials and was centered in the cancer center and facilitated by guidelines developed addressing this area of concern. Nevertheless, there were significant cost savings with other antimicrobial categories as well. This included antibacterial agents, for which costs decreased over $500,000. The IV-to-PO-switch therapy program resulted in a decrease of $180,000 within the first year of the program. Nor was the decrease in costs entirely centered within the cancer center. The shock trauma center, which has a team of infectious diseases physicians who see the majority of the trauma patients, also experienced reduced antimicrobial costs, particularly of antibacterial agents. The main hospital also experienced reduced costs, including in the MICU, the SICU, and the transplantation service.
Although the cost of antimicrobials appeared to remain stable from 2004 to 2008, an additional reduction of 13.7% occurred when patient-days are considered, in spite of drug price inflation.9 Also noted during this period was a decrease in DDDs per 1,000 patient-days, from 1,512 to 1,073 (29%) for all antimicrobials.
The benefit of the antimicrobial program was apparent following its introduction, but the strength of this observation is enhanced by the rapid increase in antimicrobial costs that occurred after the program was terminated. Within 2 years these costs increased by 41.2%, or almost $2 million; however, this increase was not related to an increase in the use of antifungals (which were noted to decrease on initiation of the program) but instead primarily involved antibacterial agents. Costs increased with piperacillin/tazobactam, carbapenems, and many of the agents with activity against the gram-positive bacteria, including vancomcyin, daptomycin, and linezolid. External factors may have also partially contributed to this increase in cost. The guidelines for monitoring vancomycin, which called for an increase in its dosing, were published in January 2009, 6 months after our program was discontinued.10 Also, there was an increase in the number of clinical culture isolates of vancomycin-resistant Enterococcus (but not bloodstream infections) in FY 2009 and FY 2010, after the program ended, which also could have contributed to an increase in the use of daptomycin and linezolid.
In 1999, Carling et al evaluated the cost of parenteral antimicrobials in 14 acute care hospitals and found that those 5 facilities that included a system for active prospective intervention that involved a clinical pharmacist and a staff-level infectious diseases–trained physician cost (per 1,000 patient-days) 3%–30% below the mean for all of the hospitals analyzed, and all 5 had costs that were below the those of the 9 hospitals that used only passive measures.3 Passive measures included measures such as automatic stop orders, antimicrobial order forms, limited formularies, measures to control contact between pharmaceutical representatives and prescribers, educational intervention (eg, institutional guidelines for antibiotic use), and restricted antibiotic susceptibility reporting by the microbiology laboratory. It was partially on the basis of this report that we established an active program with an AMT for active prospective antimicrobial monitoring and intervention.
Prior to initiating the program with the team, there was restriction of antimicrobials and a preauthorization process was in place. Although this process was made simpler using a “BUGS” beeper to call for preauthorization, the basic process that was in use prior to the initiation of the program remained in effect. The addition of the AMT, through real-time monitoring, assured that this preauthorization process was followed and that the information given at the time of the request was accurate. Furthermore, after the program was terminated, use of the “BUGS” beeper for authorization continued but the antimicrobial team was no longer present to assure that the information given to request the antimicrobials was accurate and that the release of the antimicrobials adhered to the policies of the medical center and were appropriate in the treatment situation. Thus, the major implementation in the program was the addition of the AMT.
Another component of the ASP, which enhanced the effectiveness of the AMT, was the use of a computer decision support system.4 Indeed, this was developed to do what the AMT was already doing, but to make the team more efficient in their duties. This computer program organized and alerted the AMT when restricted drugs were ordered and indicated where the patient resided, other medications the patient was receiving, and microbiologic laboratory results. In addition to notifications of patients receiving “restricted” antimicrobials, some other alerts included notification if a patient was receiving double antimicrobial coverage or no antimicrobial coverage for an identified pathogen and identification of potential candidates for the switch from IV to oral therapy and patients who had received 5 days of antimicrobial therapy without the isolation of potential pathogens. In a carefully controlled blinded study, it was projected that this system saved the UMMC over $600,000, compared with the use of the antimicrobial team without the decision-support system if it were used in the major portion of the medical facility for an entire year. When the stewardship program was terminated, the decision-support system was also not continued.
The stewardship program was discontinued because of some dissatisfaction over the preauthorization requirements as well as so that the funding for the program could be used to provide personnel for additional infectious diseases consultation throughout the medical center. The rationale for this change was that infectious diseases experts were the best trained individuals to make the necessary decisions for appropriate therapy in this difficult patient population and thereby could provide antimicrobial stewardship, therefore, rendering the AMT redundant. We cannot say for certain that this would not be an effective strategy, given more time. However, costs have continued to increase and the quality markers have remained stable, suggesting that more and more costly antimicrobials are being used, with no obvious increase in benefit.
In summary, the ASP with a preauthorization protocol using a “BUGS” beeper, an AMT to assure appropriate use, and computer decision-support assistance was an extremely cost-effective model for antimicrobial stewardship over a period of 7 years, and its discontinuation has proven to be very costly. On the basis of this information an ASP using an AMT has been restarted, but an automatic infectious diseases consult has replaced the preauthorization requirement for “restricted” antimicrobials.
We would like to thank the following for their help in developing and maintaining the Antimicrobial Stewardship Program: for the development of the program: David Rorison, MD, MBA (senior vice president of medical affairs), Sherry Perkins, PhD, RN (vice president for clinical effectiveness), and Michael Donnenberg, MD (chief of the division of infectious diseases); from the antimicrobial team: Molly Kent, Pharm D; regarding policy and guidelines: Michael Kleinberg, MD (division of infectious diseases); and for data analysis: Marshall Bryant (pharmacy service).
Potential conflicts of interest. All authors report no potential conflicts of interest relevant to this article. All authors submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and the conflicts that the editors consider relevant to this article are disclosed here.
- 1. Boucher HW, Talbot GH, Bradley JS, et al. Bad bugs, no drugs: no ESKAPE! an update from the Infectious Diseases Society of America. Clin Infect Dis 2009;48:1–12.
- 2. Dellit TH, Owens RC, McGowan JE, et al. Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis 2007;44:159–177.
- 3. Carling PC, Fung T, Coldiron JS. Parenteral antibiotic use in acute-care hospitals: a standardized analysis of fourteen institutions. Clin Infect Dis 1999;29:1189–1196.
- 4. McGregor JC, Weekes E, Forrest GN, et al. Impact of a computerized decision support system on reducing inappropriate antimicrobial use: a randomized controlled trial. J Am Med Inform Assoc 2006;13:378–384.
- 5. Carling P, Fung T, Killion A, et al. Favorable impact of a multidisciplinary antibiotic management program conducted during 7 years. Infect Control Hosp Epidemiol 2003;24:699–706.
- 6. Stevens V, Dumyati G, Fine S, et al. Cumulative antibiotic exposures over time and the risk of Clostridium difficile infection. Clin Infect Dis 2011;53:42–48.
- 7. Fishman N. Antimicrobial stewardship. Am J Med 2006;119:S53–S61.
- 8. Singh N, Rogers P, Atwood CW, et al. Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit. Am J Respir Crit Care Med 2000;162:505–511.
- 9. Hoffman JM, Shan ND, Vermeulen LC, Schumock GT, Grim P, Hunkler RJ, Hontz KM. Projecting future drug expenditures: 2007. Am J Health Syst Pharm 2007;64:298–314.
- 10. Rybak M, Lomaestro B, Rotschafer JC, et al. Therapeutic monitoring of vancomycin in adult patients: a consensus review of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists. Am J Health Syst Pharm 2009;66:82–98.