Antimicrobial Stewardship Programs as a Means to Optuimize Anitmicrobial Use

Robert C. Owens, Jr., Pharm.D.; Gilles L. Fraser, Pharm.D., FCCM; Patricia Stogsdill, MD


Pharmacotherapy. 2004;24(7) 

In This Article

Practical Considerations

For those considering the initiation of a formal antimicrobial stewardship program, it is fundamentally important to appreciate the framework built by initial communications. Preliminary discussion should involve developing goals and a formal plan, selecting the main strategy (prior approval or concurrent review, or a hybrid of both), and determining the members involved, their time commitment, and a funding stream for the program. In addition, the importance of assembling a multidisciplinary committee to provide a venue for continuing dialogue, and marketing the program to clinicians, should be stressed.

Program Strategies

Two main programmatic, multidisciplinary strategies exist: prior authorization, and concurrent review with feedback. Prior authorization relies on scheduled coverage (usually 24 hrs/day, 7 days/wk) or normal business hours with contingencies for off-hour antibiotic orders. Off-hour orders typically are handled through a dedicated pager and call system with a list of antimicrobial agents that require approval from the service before dispensation.

Two programs serve as models for the prior-authorization approach. One group constructed a list of restricted antimicrobials based on cost and/or spectra of activity.[22] The list included, for example, intravenous formulations of amikacin, ceftazidime, ciprofloxacin, ofloxacin, fluconazole, ticarcillin-clavulanate, piperacillin-tazobactam, and aztreonam. Members of the infectious diseases division were on call to receive approvals for these agents 24 hours/day, 7 days/week by dedicated pager.

Another group initially employed a dedicated beeper schedule for normal weekday business hours that was covered by an antimicrobial management team member (a physician or pharmacist trained in infectious diseases).[23] Second-year infectious disease fellows covered evenings and weekends. At night, restricted drugs were released pending next-morning follow-up. Based on the results of their study ( Table 1 ), the infectious disease fellows were incorporated into the antimicrobial management team and work more directly with the pharmacist and infectious disease attending physician. They also have published their list of restricted antimicrobials and guidelines on a Web site that can be accessed, at least in part, by outside institutions (available from The list of restricted drugs can be large or small and can be based on cost, resistance-evoking potential, or spectra of activity, as well as safety and tolerability.

The latter antimicrobial stewardship program strategy relies on less restrictive antimicrobial policies. Rather, it involves reviews or making rounds for patients identified from a computer-generated list of those receiving specific (or all) antimicrobial agents on scheduled days (e.g., Monday-Friday). The list can be modified to contain enough information to make general assessments related to appropriateness of combination therapy, dosing, route, and selection (if culture and susceptibility information is available). With some programs, recommendations are made only when culture and susceptibility results are available, typically by the third day of treatment.

From the initial assessment, a chart review and/or discussion with other health care team members usually ensues. When intervention is deemed necessary, nonpermanent notes usually are left in the medical record with educational vignettes, literature citations, and/or local guidelines; verbal communication is used for more urgent matters. One of the more widely quoted published antimicrobial stewardship programs of this type originated as a randomized study of clinical and economic outcomes involving an infectious disease physician and a pharmacist.[18] Therapy was reviewed and interrupted (when necessary) in hospitalized adults receiving one of 10 targeted parenteral antibiotics (based on cost and spectrum of activity) for more than 3 days.

Measuring Outcomes

Several studies have evaluated the impact of antimicrobial stewardship programs on a variety of outcomes ( Table 1 ). Although each study used a slightly different approach, and outcome variables were not always defined uniformly (e.g., appropriateness of therapy, antibiotic consumption measurements), the resulting data have been valuable. For instance, studies indicate that oversight of antimicrobial use (whether restrictive or more subtle through concurrent review and feedback) has had a measurable impact on appropriateness of antimicrobial use, antimicrobial consumption, and/or expenditures.[18,19,20,21,22,23,24,25,26,27,28,29,30] Some studies evaluated the impact of antimicrobial stewardship programs on resistance rates, infection rates, and clinical outcomes.[18,22,23,24,26,27,28]

Some studies showed favorable results to widespread drug-organism combinations,[22] and some antimicrobial stewardship programs were more targeted. For example, two studies showed that reducing third-generation cephalosporin use while increasing fourth-generation cephalosporin use through education, concurrent review and feedback, and enhancements in direct computer-based physician order entry resulted in significantly improved susceptibilities among gram-negative bacilli with AmpC β-lactamase phenotypes.[24,26] Some comparative studies used before-and-after methodologies,[20,22,24,25,26,27,28,29] and some used a control group.[18,19,23] From these studies, antimicrobial stewardship programs have been created, some using a blend of strategies customized for local application.

Those interested in initiating an antimicrobial stewardship program should develop goals ensuring that performance outcomes are easily measured and relevant. Economic issues never prevail over improving clinical outcomes but remain important to evaluate.[32]

Measured outcomes and performance indicators in the literature include recommendation acceptance rates, adherence rates with antibiotic use guidelines, microbiologic and clinical response rates, frequency of antibiotic readministration within 7 days, adverse drug events, time to approve antimicrobials and time to their administration to the patient, hospital readmission rates related to infectious diagnoses, length of hospital stay, mortality rates, antimicrobial resistance rates, infection rates, antibiotic expenditures and use rates measured in terms of defined daily dose, associations between antimicrobial use and resistance or infection rates, overall hospital costs, and costs directly attributable to the infectious process.

Outcome measurement should be customized to the institution and discussed (and agreed on) before the program is implemented. Surveys may be used before the introduction of an antimicrobial stewardship program or may be used as a continuing tool to evaluate and assess clinician knowledge and attitude.[20]

Further literature on measurement of population-based antimicrobial use and resistance correlation has been reviewed in detail.[33] In addition, the recent introduction of interrupted time series with segmented regression analysis has resulted in a more sophisticated means of measuring the true impact of interventions on antimicrobial use.[34,35] Most studies to date have not undergone such rigorous statistical analyses. One must remember that none of the above variables is perfect; benefits and limitations exist with each.


The literature points out some of the pitfalls that have been experienced and concerns that need to be addressed. Delays in approval for a necessary antimicrobial agent can be detrimental to critically ill patients in need of initial broad-spectrum antimicrobial therapy. One group experienced no delay in the administration of antimicrobial agents before and after the introduction of their program; however, they still stressed that approval times and time to antibiotic administration should be monitored as an outcome variable.[22]

The perception of threatened autonomy can be a significant impediment to the effort. Previous studies and our own experience have shown that the introduction of education and rationale at the point of communication diverts such an emotional response.[18,25,29] Thus, some mechanism for education and consistent communication should be considered.

The concept of "gaming the system" inadvertently may be encouraged by more restrictive programs. For example, one group reported an outbreak of nosocomial infection after the introduction of their program.[36] A 30% relative increase in documentation of infection in the medical record occurred (incidence of infection increased from 11.0 to 14.3/1000 patient-care days, p<0.05). After investigation, the outbreak was termed a pseudo-outbreak; in other words, the increased rate of infection represented clinicians attempting to justify the use of a particular restricted antimicrobial by documenting in the medical record that infection existed.

Another impediment is the perception that antimicrobial stewardship programs are solely financially driven. Both national and international recommendations point out that mechanisms must be in place in hospitals to monitor treatment with antimicrobial agents and provide feedback regarding their use in order to preserve their long-term utility. Program funding can be a barrier for some institutions. However, as mentioned in the next section, the literature points out that programs (in large or small hospitals) can usually pay for themselves as a side effect of promoting good antimicrobial stewardship.

Costs and Funding

Salaries for a part-time physician and a full-time pharmacist, both trained in infectious diseases, are considered the minimal start-up costs for an antimicrobial stewardship program at most moderate-sized to large hospitals. A part-time data analyst also has been suggested as a requisite. However, smaller hospitals, with fewer than 100-150 beds, may be able to manage with part-time support for all positions. Smaller hospitals that may not have the services of infectious disease specialists (physician or pharmacist) may benefit from consultation agreements with them to serve as advisors, educators, and directors of the antimicrobial stewardship program. The financial benefits reported by numerous studies more than justify the costs of the program.[19,20,21,22] Three possible mechanisms for support have been proposed: defined compensation, payment based on time for consultation services, and negotiation for a percentage of the cost savings.[21]

Multidisciplinary Steering Committee

The composition of the multidisciplinary steering committee will be unique to each institution, but the following should be considered for inclusion: physicians and pharmacists trained in infectious diseases, section chiefs (e.g., internal medicine, surgery), nursing leadership, microbiologist, infection control representative, data analyst, and hospital administrator.

Marketing the Program

Finally, each department affected by the antimicrobial stewardship program must be included and their feedback sought. This may be one of the single most important aspects that determines the degree of success of any new program. Department meetings and specialty grand rounds provide venues for presentation of the proposal. Since curmudgeons exist, it is wise to be proactive. In other words, conducting meetings with individuals to address specific concerns before a program is launched is better than placing the antimicrobial stewardship program in a reactive scenario after its implementation. For instance, one university group disseminated its program's antibiotic guidelines to more than 20 faculty members for review to gain consensus before they were published.[20] Newsletters, independent of active communication, are insufficient to notify clinicians of antimicrobial stewardship program implementation.

Future Considerations

Direct computer-based physician order entry, which is rapidly becoming the standard of care, has been adopted as one of the initiatives of the Leapfrog Group for Patient Safety to avoid medication errors and improve quality of care.[37] The availability of computer technology offers an advantage over academic detailing because it can be accessed 24 hours/day when antibiotics are ordered.

Extensions of this technology, in the form of computer-assisted decision support programs, provide real-time integrated patient and institutional data. Examples of these data are culture and susceptibility results (from the previous 5 yrs); laboratory measures of organ function; allergy history; contraindications; drug interactions; and hospital microbial susceptibility, antibiotic information, and cost figures. These computer-assisted support programs provide therapeutic choices for clinicians and allow for incorporation of clinical judgment. Autonomy is preserved, and consideration of important variables in the choice of antimicrobial therapy is ensured.

Almost all published data regarding the effect of computer-assisted decision support programs on antibiotic use are from one group of researchers.[38,39] Their approach has been associated with reductions in antibiotic dosages, inappropriate orders, costs, treatment duration, and associated adverse drug events. This degree of computer sophistication is not widely available but that should not dissuade institutions from adapting their own computer systems to apply lessons learned from these researchers.

Good antimicrobial stewardship has been described as "akin to motherhood and apple pie."[4] To this end, we would agree. The problem of increasing antimicrobial resistance is due, in part, to suboptimal antimicrobial use. In addition, a growing number of pharmaceutical companies have abandoned antiinfective research and development. The result is a growing public health crisis.[40]

Because hospitals characteristically are centers of high antibiotic use, they are target-rich venues for proactive interventions to improve antimicrobial stewardship. Studies of antimicrobial stewardship programs indicate that these programs can be effective in optimizing antimicrobial use. In addition, when implemented by clinicians trained in infectious disease, overall antibiotic use can be reduced without compromising clinical outcomes.

Other benefits are decreased rates of resistance for certain organism-drug combinations. Thus, antimicrobial stewardship programs provide an ideal partner to infection control efforts. A side benefit of optimizing antimicrobial use is a proved reduction in related expenditures, making these programs not only attractive options to promote good antimicrobial stewardship, but financially possible. It has been stated, "... there are simply too many physicians prescribing antibiotics casually...The issues need to be presented forcefully to the medical community and the public. Third-party payers must get the message that these programs can save lives as well as money."[41]