Robert C. Owens Jr.
Introduction
Microbes account for 90% of the 1014 cells in the human body [1]. Thus, it should not be surprising that the well-known benefits of antimicrobials must be balanced with an appreciation of their unintended effects on beneficial microbiota and the evolution of resistance. In essence, of use antimicrobials is a double-edged sword. Their life-saving benefits and disease-modifying effects can be nothing short of miraculous; however, the opposing edge of the sword can be just as sharp. Examples include serious patient harm in the form of adverse drug events (e.g., hepatotoxicity, torsades de pointes, anaphylaxis, renal failure), development of potentially life-threatening Clostridium difficile-associated disease (CDAD), and the emergence of resistance [2,3]. While the first two can occur during therapy and are more tangible to the practicing clinician, the emergence of resistance often occurs late and may be less obvious.
It has been said that those who move forward without taking a moment to look back are doomed to repeat history. In 1956, noted microbiologist Ernest Jawetz once wrote:
On the whole, the position of antimicrobial agents in medical therapy is highly satisfactory. The majority of bacterial infections can be cured simply, effectively, and cheaply. The mortality and morbidity from bacterial diseases has fallen so low that they are no longer among the important unsolved problems of medicine. These accomplishments are widely known and appreciated…[4].
He goes on to state the intentions of his paper: “the author wishes to call attention to the abuse of antibiotics, its causes and results…” [4]. In the current antimicrobial era, some 50-plus years after these initially optimistic observations were made, an increasing number of infections are not easily treated, morbidity and mortality are appreciable, and many infectious diseases have become unsolved problems of modern medicine. The theme of Jawetz's latter plea remains a challenge facing many infectious diseases experts today, calling attention to the abuse of antibiotics and its causes and results.
In fact, published data note that many prescribers today still do not fully value the importance of preserving these therapeutic resources. Twenty-five million pounds of antibiotics are produced yearly for human consumption and are administered to 30–50% of hospitalized patients with nonhospitalized Americans receiving 160 million courses [5]. Yet data suggest that as much as 50% of all antimicrobial use is inappropriate. Stewarding these precious
P.180
resources has become a priority for many organizations, including the Infectious Diseases Society of America (IDSA), the Society of Healthcare Epidemiology of America (SHEA), the Society of Infectious Diseases Pharmacists, the Alliance for the Prudent Use of Antimicrobials, the Centers for Disease Control and Prevention (CDC), and the World Health Organization. In fact, joint guidelines endorsed by multiple societies recently have been published that reiterate the need for proactive, programmatic efforts to optimize the use of antimicrobials in healthcare settings [6].
Since the initial work of Finland and McGowan, a variety of interventional strategies has been shown to reduce unnecessary antimicrobial use, to optimize the dose and duration, and to minimize the collateral adverse effects of their use [7,8,9]. Most studies have evaluated the impact of interventions on inpatient antimicrobial use and, to a lesser degree, outpatient antimicrobial use. This chapter shall reinforce why the judicious use of antimicrobials is essential and the implementation (including potential barriers) of an institutional approach is necessary to optimizing antimicrobial use. It shall also discuss the collaborative nature of these programs with infection control and prevention programs and the microbiology laboratory.
Rationale for Optimizing Antimicrobial Use
Antimicrobial Resistance
Optimizing antimicrobial use through appropriate selection, dosing, and duration can be viewed as a strategy to minimize the development of resistance among clinically important pathogens [10]. Factors promoting resistance are complex, numerous, and extend beyond the use of antimicrobial agents in humans; as such, it is not surprising that they do not allow for a prompt resolution [11]. Although antimicrobial resistance has been present on this earth since the days of the primordial soup, its practical onset began in the 1920s with the observation that Pfeiffer's bacillus (now Haemophilus influenzae) showed a natural resistance to penicillin before its introduction for human use [12]. With the introduction of Gerhard Domagk's sulfa drugs in the 1930s, strains of Neisseria gonorrhea and Streptococcus pneumonae were noted to have developed so-called “insensitivity” [12]. Observations from the laboratory moved to the clinic in the 1940s shortly after the introduction of penicillin for the treatment of human infections. The miracle drug, penicillin, while initially effective for the treatment of Staphylococcus aureus bloodstream infection (BSI), began to fail to treat infections caused by penicillinase-elaborating strains as reported in Timemagazine on May 15, 1944.
Important to this discussion is the fact that studies have established a strong relationship between antimicrobial use and resistance. Levy et al. developed a biologic model that showed a clear relationship between antimicrobial use and the selection of resistance in humans [13]. Additionally, supportive data can be gleaned from in vitro investigations, ecologic studies correlating drug exposure with resistance, controlled trials in which patients with prior use of antimicrobials were more likely to be colonized or infected with resistant bacteria, or prospective studies in which drug use was associated with the development of resistant flora [13,14,15]. With many pharmaceutical companies no longer supporting anti-infective development and fewer new chemical entities being identified, we, now more than ever, should be concerned about the consequences of antimicrobial resistance [16].
Data from surveillance studies demonstrate that for community-acquired respiratory pathogens, resistance among S. pneumoniae or H. influenzae can be an obstacle in selecting and dosing the ideal regimen [17,18]. For healthcare-associated infections (HAIs), resistance is an important impediment to treating patients with the correct antimicrobial and dose in a timely fashion. The effects of selecting the wrong antimicrobial and at the incorrect dose have measurable effects on patient outcomes as highlighted by several recent studies [19,20]. Thus, an important value provided by a programmatic approach dedicated to the oversight of antimicrobial use and employed by the healthcare system is to “quarterback” and operationalize efforts to maximize the benefits of antimicrobials. Antimicrobial stewardship programs (ASPs) are able to drive the multidisciplinary development of locally customized disease-based guidelines, protocols, and order sets and to provide real-time human and, where available, computerized decision support in addition to their day-to-day interventions.
Patient Safety
Whether being used appropriately or inappropriately, antimicrobials have the potential for causing serious harm to patients. For example, macrolides, ketolides, or fluoroquinolones are associated with QT interval prolongation; macrolides or ketolides are associated with metabolic liability in the form of cytochrome P450 3A4 inhibition; trimethoprim-sulfamethoxazole is associated with Stevens-Johnson Syndrome; and β-lactams are best known for hypersensitivity reactions [3,21,22] while all antimicrobials are associated with CDAD [23]. Disturbingly, the rate and severity of CDAD is increasing and our traditional treatment options appear to be less effective [24,25]. The potential harm caused by antimicrobials should incentivize even the most temerarious clinicians not to casually prescribe antimicrobials in the setting of a nonbacterial infection or to stop therapy in a timely manner and carefully monitor patients off antimicrobials [2]. In addition, antimicrobials are unique and unlike any other drug class. One can develop an infection due to a resistant pathogen without ever having received the particular antimicrobial it is resistant to; hence, antimicrobials are considered societal drugs because their use has societal consequences [26].
P.181
Programs to Optimize Antimicrobial Use
Hospital-Based ASPs
A variety of studies has evaluated the impact of interventions on antimicrobial use in healthcare systems. These studies have been conducted using a wide range of resources, methodologies, interventions (often multiple), and outcome measures (usually considering cost, antimicrobial consumption, patient safety, and less frequently resistance). The culmination of these studies was considered and led to a guidance document developed jointly by IDSA and SHEA to provide the framework for developing, implementing, and monitoring the impact of ASPs [6].
IDSA/SHEA Guidelines for Developing an Institutional Program to Enhance Antimicrobial Stewardship
An effective ASP is financially self-supporting and is aligned with patient safety goals [8,27,28,29,30,31,32,33]. For these reasons, there should be no excuse for an institution not to have a formal program dedicated to improving the quality of antimicrobial use. Realizing that institutions vary in size and type of specialty services offered, the ASP should be customized accordingly.
Interventional Strategy
Two major interventional styles have evolved over recent years. The first is prospective audit and feedback (“back-end” program). This entails obtaining a daily (or every-other-day for smaller hospitals) list of patients receiving antimicrobials and determining interventions such as pharmacodynamic dosage adjustment, streamlining the deescalation and identification of redundant therapy based on culture and susceptibility results, parenteral-to-oral conversions, drug interaction identification, guideline/protocol compliance, and recommendation of more cost-effective treatments (Figure 13-1). Recommendations are provided to the prescriber in either written form or by direct conversation. Written forms of communication usually take place on nonpermanent forms placed in the patients' medical record that are removed at discharge. This allows flexibility in what can be written and allows the ASP team member to communicate educational messages effectively and to provide citations or references on why the intervention is being recommended. The benefits of this type of program are its customizability to smaller- [29] or larger-size healthcare facilities [9,27]; it avoids taking away the prescriber's autonomy, which also increases productive “educational” dialogue; and it circumvents the potential for delays in initiating timely antimicrobial therapy since the antimicrobial already is prescribed. The downside is that recommendations are optional (although there are ways to correct repeated unaccepted recommendations by communicating with either the department chief or an institutional committee [e.g., medical executive committee, pharmacy and therapeutics committee, patient safety committee]). Our program at Maine Medical Center (MMC) has employed this primary strategy for >5 years [8,27,33], and others have operated for longer periods of time [28].
|
Figure 13-1 Concurrent review program workflow diagram. |
|
Figure 13-2 Multidisciplinary involvement and core team members. |
The second chief strategy is a preauthorization or “front-end” program that restricts most antimicrobials to an approval process. A team member carries a pager or telephone and receives approval requests for restricted antimicrobials. At the time of interaction, the antimicrobial either is justified and approved, or an alternative recommendation is given. The University of Pennsylvania [34,35], The University of Pittsburgh [36], and others [37] have used preauthorization as their primary strategy for several years. The benefits of this strategy include the ability to funnel all initial antimicrobial prescribing through experts versed in antimicrobial therapy and the typical demonstration of these programs for immediate and significant cost savings. The potential downsides to this strategy include the loss of prescriptive autonomy that may lead to “gaming the system” [38] and the fostering of potentially adversarial relationships (if not properly implemented with buy-in from important and opinionated prescribers), the potential for delaying initial therapy, time- and resource-intensive plans (usually 7-days per week with contingency plans for night coverage), and the necessity to make decisions when the least amount of information is known about the actual infection (culture and antimicrobial susceptibility results are not available for 2–3 days and the quality of information relayed to the ASP team member by prescriber can be variable [39]).
In reality, although ASPs may lean toward one of the two primary strategies, overlap often exists. For example, the program at Maine Medical Center, while relying primarily on prospective audit and feedback, does incorporate a limited number of antimicrobials that require approval [8]. One of the most valuable aspects of an ASP is the institutionwide responsibility for overseeing the use of antimicrobials. Although in moderate-size or larger hospitals, an infectious diseases (ID) physician consultation service, ID pharmacist, and infection control department often are present and co-exist or collaborate on specific areas of interest, responsibility at the institutional level for antimicrobial stewardship usually is not assigned. An administratively supported ASP aligns resources from these various specialties and assigns responsibility to them.
Team Members
The IDSA/SHEA guidelines for developing an institutional program to enhance antimicrobial stewardship are very clear about the following: the ASP is directed or codirected by the two core team members, an ID physician and an ID-trained pharmacist, both receiving remuneration for their time. The pharmacist should have formal training in ID or
P.182
be knowledgeable in the appropriate use of antimicrobials with training being made available to maintain competency. Other team members optimally include a dedicated computer information support specialist, microbiologist, and an infection control professional/hospital epidemiologist. Figure 13-2 illustrates an optimal schematic for collaboration and partnership that Maine Medical Center uses and was adapted from previous models [40]. Administrative and committee support (e.g., pharmacy and therapeutics committee) is critical. The particular interventional philosophy, responsibilities, remuneration, and reporting measures should be discussed in advance of implementation to address expectations and resources. Effective communication between the ASP, administration, and an appropriate committee should be maintained to facilitate dialogue over time as the healthcare environment continues to change.
Prospective Audit and Feedback and Preauthorization Studies
Prospective Audit and Feedback Strategy
Fraser et al. [27] designed a prospective randomized controlled study of interventions for targeted antimicrobials in hospitalized patients. The team included a part-time ID physician and a PharmD with antimicrobial expertise. The intervention group (N = 141) received suggestions (written or verbally) whereas the control group did not (N = 111). Controlling for severity of illness between groups, outcomes were similar with respect to clinical and microbiologic response to therapy, adverse events, inpatient mortality, or readmission rates. Interventions included change to oral therapy (31%), regimen or dosing changes (42%), stopping therapy (10%), or ordering additional laboratory tests (18%); 85% of the suggestions were instituted. Multiple logistic regression models identified randomization to the intervention group as the sole predictor of lower antimicrobial expenditures. A conservative annualized reduction in antimicrobial expenditures of $97,500 was realized. The intervention group also showed a trend toward reduced mean length of stay compared to the controls (20 days versus 24 days, respectively). Fifty percent of patients receiving targeted regimens had their treatment refined on the third day of therapy resulting in narrower spectrum therapy and lower antimicrobial costs; most important, reducing antimicrobial use did not adversely impact patient outcomes. This study later was used as a platform to implement an ASP that is more
P.183
robust in types of activities and numbers of patients served by the program. The team currently includes a part-time ID physician (2 hours per day, 5 days per week) and a full-time ID PharmD. Close collaboration exists with the Department of Epidemiology and Infection Prevention, the ID division, the pharmacy, administration, the patient safety officer, and the pharmacy and therapeutics committee.
Srinivasan et al. studied the impact of an antimicrobial management program on antimicrobial expenditures at Johns Hopkins Hospital. Before the introduction of a comprehensive ASP, the hospital used a closed formulary system and employed prior approval requirements for several antimicrobials. The ASP consisted of a hospital-funded ID physician, ID PharmD, and data analyst. The team concurrently reviewed antimicrobial therapy in all areas of the hospital except pediatrics and oncology. Their interventions included a survey and the use of institution-specific guidelines, concurrent antimicrobial review, and educational sessions. A “knowledge, attitude, and beliefs” survey was used to determine awareness of antimicrobial use and resistance and sense deficiencies in knowledge that could lead to targeted education among house staff [41]. Interestingly, only 18% viewed the program as an obstacle to patient care and 70% wanted additional feedback on antimicrobial choices. Hospital guidelines were published and updated annually. Antimicrobial therapy interventions occurred before culture and susceptibility results were available only when actively solicited or when called for prior authorization of an antimicrobial agent. For all others, interventions were suggested at the time the microbiologic data became available. Compliance with suggested recommendations by the ASP was 79%. Costs for antimicrobial agents for the covered areas decreased by 6.4% the first year and 2.2% the second year. Assuming a steady inflation rate of 4.5%, savings translated to $224,753 and $413,998 for fiscal years 2002 and 2003, respectively.
Bantar et al. demonstrated the impact of their ASP's interventional component on antimicrobial use, cost savings, and antimicrobial resistance [42]. The ASP consisted of an ID physician, two pharmacists, a microbiologist and laboratory technologist, an internal medicine physician, and a computer systems analyst. In six-month periods, four consecutive intervention strategies were unveiled. During the first six months, an optional antibiotic order form (ID diagnosis, pertinent epidemiologic data) was introduced and baseline data were collected (i.e., bacterial resistance, antibiotic use, prescribing practice, HAI, and crude mortality rates). In the second period, an “initial intervention” period consisted of transforming the optional order form to a compulsory form and providing feedback to clinicians based on a review of the data collected in the first period. In the third period, called the “education” period, clinicians were verbally engaged with each new antimicrobial order by members of the multidisciplinary team. The fourth or “active control” period was similar to the third period, but the ASP team member modified prescriptions if necessary. During the four periods, no antimicrobial agent was restricted. To estimate
P.184
the rates of use of a particular drug in relation to other drugs, an index was calculated (e.g., rate of cefepime use to that for third-generation cephalosporins—ceftriaxone and ceftazidime—equaled cefepime/consumption of ceftriaxone and ceftazidime × 100). Consumption data were measured in defined daily doses (DDD). The program periods were associated with declining cost savings as time advanced (periods 2, 3, and 4 were associated with a reduction of $261,955, $57,245, and $12,881, respectively). Comparing antibiotic order forms from period 1 (voluntary form and pre-intervention, N = 450) with period 4 (mandatory form with active intervention, N = 349), the ASP identified an increase in microbiologically based treatment intent (27% versus 62.8%, respectively, P < 0.0001). The team intervened on twenty-seven percent of the period 4 antibiotic order forms. Of the interventions, either the dose or duration (not specified) was reduced in 11.5%, 47% involved streamlining therapy to a narrower choice, and 86.1% was associated with cost reduction. HAI impact (e.g., length of hospitalization and mortality; only length of stay was impacted significantly [P = 0.04]). The increased rate of cefepime use relative to third-generation cephalosporins was associated with declining third-generation cephalosporin resistance rates among Proteus mirabilis or Enterobacter cloacae but not to E. coli or K. pneumoniae. The increased rate of aminopenicillin/sulbactam use relative to the third-generation cephalosporins in conjunction with a sustained reduction in vancomycin use was associated with a reduction in MRSA rates. In addition, P. aeruginosa resistance rates to carbapenems declined to 0% and were strongly associated with the reduction in carbapenem consumption.
The particular study by Bantar et al. is different than others in that it used a staggered approach to implementation. Although the cost reduction appeared to dwindle significantly in each newly introduced period, one cannot ignore the cumulative impact of the overall impact on cost. In addition, the final period offers a comprehensive mechanism for long-standing success and serves as a template to introduce other initiatives as deemed necessary. Part of the success related to reduction in resistance rates noted by this program is related to the high rate of carbapenem or ceftriaxone use and the “seldom” ordered cefepime or aminopenicillin/sulbactam in conjunction with the types of problem pathogens noted at their hospital (e.g., ampC phenotypes and carbapenem-resistant P. aeruginosa). Penicillin-based inhibitor combinations or cefepime have been noted to more often favorably impact the environment, in contrast to high usage rates of carbapenem or third-generation cephalo-sporin [9,43,44].
Another study demonstrated the impact of a multidisciplinary ASP using a blend of interventions including minimal formulary restrictions, comprehensive education (e.g., direct communication, antibiograms, peer feedback every six months), rounding with medical teams, and introduction of guidelines (appropriate initial empirical therapy, transitional therapy, duration of therapy) [31]. All adult patients admitted to the medicine service were consecutively evaluated before the introduction of the program (N = 500 patients) and postimplementation. Using defined daily dose (DDD) data and hospital expenditure data, is study showed a 36% reduction in overall antimicrobial use (P < 0.001), intravenous antimicrobial use (46%, P < 0.01), or overall expenditures (53%, P = 0.001) without compromising the quality of patient care (determined by inpatient survival, clinical improvement/cure, duration of hospitalization, and readmission rates ≤ 30 days). These benefits were sustained for the four-year period evaluated.
Carling et al. [28] evaluated their ASP over a 7-year period. It consists of a physician (one-quarter time support) and PharmD (full-time support), both with specialty training in ID. Antimicrobial consumption was measured by using DDD/1000 patient-days for targeted antimicrobial agents. This program operated 8 hours per day and 5 days per week, and during this time, new orders are typically evaluated within 4 hours of their entry. Orders falling outside the 8-hour day are reviewed as a priority the next time the PharmD is on duty. Informal written notes are generated when the team identifies a problematic regimen and then is placed in the patient's chart. “Academic detailing” also occurs between the PharmD and the prescribing clinicians to supplement the written recommendations. They evaluated their impact on vancomycin-resistant enterococcus (VRE), methicillin-resistantStaphylococcus aureus (MRSA), and CDAC by means of internal benchmarking and externally benchmarking themselves with similar hospitals within the CDC's National Nosocomial Infections Surveillance (NNIS) system. Over the 7 years, a 22% reduction in parenteral broadspectrum antibiotics occurred (P < 0.0001) during which time was a 15% increase in the acuity of their patient population observed. Reductions in HAIs caused by C. difficile (P = 0.002) or resistant enterobacteriaceae (P = 0.02) were reported. MRSA rates remained unaffected.
A smaller hospital (120-bed community hospital) also has successfully implemented an ASP using a prospective audit and feedback strategy [29]. The ASP involved an ID physician, a clinical pharmacist, and representatives from infection control and the microbiology laboratory. The ID physician was involved approximately 8–12 hours per week. Antimicrobial therapy was reviewed three days a week in patients receiving targeted drugs or prolonged durations of therapy. Recommendations were conveyed using a form that was temporarily placed in the patient's chart and by telephone, if necessary. During the first year, 488 recommendations were made with a 69% acceptance rate; antimicrobial expenditures were reduced by 19%, saving an estimated $177,000. Common interventions were the discontinuation of redundant antimicrobial therapy, the discontinuation of treatment due to inappropriate use or excessive duration,
P.185
the transition from intravenous to oral therapy, or the substitution or addition of an antibiotic to the regimen.
Preauthorization Strategy
White et al. [37] implemented an ASP restricting the use of antimicrobials based on cost and/or spectra of activity. A 24-hour per day, 7-day per week on-call system was established via a dedicated pager that clinicians would call to receive approval for restricted agents. In their quasi-experimental study, patients in the preimplementation period were similar to the post implementation period in severity of illness. Outcome measures that were not statistically significant between groups were survival (P = 0.49), infection-related length of stay for BSI (P > 0.05) or, more important, time to administration of the antimicrobial (P > 0.05). Benefits received in the post-ASP implementation group were improved susceptibilities to a number of bug-drug combinations, primarily involving nonfermenting Gram-negative rods or enterobacteriaceae, significant reduction in the use of a number of broadspectrum agents, and a significant reduction in annualized antimicrobial costs ($803,910) and costs per patient-day ($18 to $14.4).
Various well-conducted studies at the University of Pennsylvania over the last two decades have contributed to our current knowledge of ASPs in general [40]. Gross et al. [34] initially employed a dedicated beeper schedule for weekdays during normal business hours that was covered by an antimicrobial management team (AMT) member (an ID PharmD or ID physician). Second-year ID fellows covered evenings and weekends. At night, restricted drugs were released pending next morning follow-up. Taking advantage of their existing program. The AMT evaluated interventions performed by the ID fellows versus those made by the ID PharmD and/or ID physician. They concluded that interventions performed by the veteran ASP team members (ID PharmD and/or ID physician) were more cost effective and resulted in narrower spectrum therapy compared with interventions made by ID fellows. Based on the results of their study, the ID fellows have been more fully incorporated into their ASP and work with the PharmD and ID physician more directly. The AMT also have published their intranet/Internet resource list of restricted antimicrobials and guidelines on a Web site that can be accessed, at least in part, by others (www.uphs.upenn.edu/bugdrug). In addition to the preauthorization method for active interventions. The AMT also work closely with the hospital epidemiologist, are involved in establishing guidelines for antimicrobial use and dosing, are proactively involved in the antimicrobial formulary, work closely with the pharmacy and therapeutics committee, provide education, and continuously evaluate antimicrobial consumption trends [40].
Potential Barriers
The literature reports some pitfalls that some programs have experienced. Delays in the approval for a necessary antimicrobial agent can be detrimental to critically ill patients who need initial broadspectrum antimicrobial therapy. White et al. showed no delay in the administration of antimicrobial agents before or after the introduction of its program; however, approval times and time to antibiotic administration must be monitored as a process measure [37]. The perception of threatened autonomy can be a significant impediment to the efficacy of the program. The experience of LaRocco et al. [29] and the author [8] using the prospective audit and feedback strategy with few restricted antimicrobials promotes education at the point of intervention, neutralizing negative emotions. Thus, regardless of approach, constant communication with frontline prescribers and education are vital. The concept of gaming the system cannot be ignored and is a function of human nature. One program reported an HAI outbreak following introduction of its ASP [38]. A 30% relative increase in HAI documentation in the medical record occurred (HAI incidence increased from 11 to 14.3 per 1000 patient-care days, P < 0.05) [38]. After further investigation into this counterintuitive finding, the HAI outbreak was termed a “pseudo-outbreak.” Clinicians were required to document infection in the medical record to justify antimicrobial use; thus, more clinicians were documenting infections in an attempt to use particular restricted antimicrobials.
The perception that ASPs are solely financially driven also can be an impediment. However, the guidelines of the IDSA/SHEA and other authorities endorse these programs not based on the potential for cost savings but as a means to improve patient safety and to reduce the selective pressure exerted by unnecessary antimicrobial use that facilitates the evolution of antimicrobial resistance. Typically, as a side effect of interventions to optimize antimicrobial use to improve efficacy and reduce resistance, a cost saving is observed that financially justifies the program. Administrators need to be cognizant of this when helping to develop ASPs. Program funding can be a barrier for some institutions, but as mentioned in SHEA/IDSA guidelines and as numerous studies in the area of ASPs point out, ASPs typically pay for themselves as a side effect of their existence.
Supplemental Programs to the Primary ASP Strategy
Formulary Interventions
A survey of teaching hospitals suggests that 80% of them limit prescriber access to antibiotics using a variety of mechanisms [45]. Formulary restriction is the most direct way to influence antimicrobial use and is central to
P.186
the primary preauthorization ASP strategy. Most hospitals, regardless of having an ASP, employ this strategy by limiting access to the numbers of antimicrobial drugs within a class. This limitation is a passive intervention strategy, whereas enforcement through ASPs using either preauthorization or prospective audit and feedback strategies shifts the ASP strategy to an active intervention.
Careful selection of drugs within a class involves an in-depth analysis of not only the basics (efficacy, safety, and cost) but also should include an evaluation of resistance-evoking potential when possible as well as a pharmacodynamic evaluation of the drug. A benefit of having an ASP allows this process to be centralized, and working closely with the pharmacy and therapeutics committee helps to establish an optimal formulary stocked with the best drugs for the given institution. The cost evaluation should extend beyond purchase prices, although leveraging contracts is a useful tool [46]. An important factor is that the overall cost of care should be considered when an antimicrobial is evaluated, but this is not always done because of the compartmentalization of costs within an institution (also referred to as the “silo” mentality). For example, although 8–10 fold more expensive than intravenous vancomycin in purchase cost, the use of oral linezolid has been shown to decrease the length of stay and improve discharge dynamics for patients with MRSA infections [47,48,49,50]. This is particularly financially appealing for institutions from the perceptions of those operating at maximal census (because bed costs far outweigh drug costs) and with high MRSA rates and of, the infection control of reduced transmission dynamics when length of stay is shortened, and, most important, of the patient who can be at home rather than hospitalized. In the Maine Medical Center's institution, ASP, working with the care coordination and infection control departments, has taken advantage of efficiently transitioning the patient after becoming clinically stable to an oral MRSA therapy, such as linezolid and, in some cases, minocycline. The purposes of this transition is to reduce the barrier to discharge posed by home intravenous therapy arrangement or skilled nursing facility placement if solely for administering intravenous vancomycin [33].
Practical, evidence-based examples of preferentially replacing drugs with increased resistance-evoking potential (e.g., ceftazidime) with a member of the same class exist, but these examples have demonstrated a reduced ability to select for resistance (e.g., cefepime) [8,9,42,51]. A number of studies of hospitals still characterized by high third-generation cephalosporin use have demonstrated that the replacement of third generation cephalosporins with either cefepime or piperacillin/tazobactam is an effective strategy (particularly in concert with infection control interventions) to minimize the selective pressure that facilitates the appearance of problematic β-lactamases (e.g., AmpC enzymes, extended spectrum β-lactamases) or VRE [9,52,53,54]. The contrast between vancomycin and daptomycin is another example that highlights the point that not all drugs are created equal in their potential to select for antimicrobial resistance. Less high-level vancomycin-resistant S. aureus strains have been reported than daptomycin nonsusceptible S. aureus strains. The difference is that vancomycin has been used for severely ill patients with BSIs, endocarditis, or meningitis, for > 3 decades. As many [6] daptomycin-resistant strains were selected during therapy in a single trial [55]. Also, it has been observed that linezolid resistance can occur more readily in enterococci than in staphylococci [56]. However, staphylococcal resistance to linezolid remains low after six years of use. Nuances related to the propensity for the antimicrobial to become resistant to the pathogens of interest within a particular institution's patient population should be considered from a formulary perspective, and monitoring the drug's susceptibility performance in a perpetual manner is an important component of an ASP working directly with the hospital epidemiologist and the microbiology laboratory [57].
Finally, determining when an antimicrobial may fit into order sets and guidelines and how its use will be monitored completes a comprehensive evaluation of how the antimicrobial will be most effectively used within the institution. For follow-up of an antimicrobial's use in the institution, the support of the pharmacy and therapeutics committee is vital because it provides a mechanism to report back inappropriate use of the drug and has the power (in many facilities) to be an effective countermeasure to correcting inappropriate use by intervening.
The ASP's involvement in the antimicrobial formulary is not limited to drug evaluations but should work periodically with the pharmacy to evaluate pricing contracts, which can be complicated and may fluctuate. Having someone with an expertise in antimicrobials working with the pharmacy buyer can greatly improve the institutional purchase costs and ensure that the hospital/healthcare system is receiving competitive pricing. Also, frequent communication with the pharmacy buyer has been a requisite over the last several years due to antimicrobial shortages. The ASP can facilitate preparation for shortages when advanced notice is given can be helpful to maintain par levels of necessary drugs, can provide insight into alternative drugs that will likely be used in their place, and can communicate these shortages with alternatives to the prescribers. In fact, an ongoing nationwide shortage of piperacillin/tazobactam has periodically challenged the Maine Medical Center, which developed a product that can be easily substituted. The combination product (cefepime combined with metronidazole in the same mini bag) can be given as a single administered product due to its stability and compatibility in combination, is administered on average two fewer times per day (cefepime and metronidazole can be given every 12 hours for most infections), and is approximately 30% less expensive than piperacillin/tazobactam [58]. Because MMC has an institutionally supported ASP, it was
P.187
able to creatively dedicate resources to an idea that both is beneficial to patients and cost effective.
Pharmacodynamic Dose Optimization
Dose optimization interventions are likely to be one of the most common interventions by an ASP. Although formerly viewed as a means to efficiently trim excess drug exposure secondary to renal dysfunction, the modern application of pharmacodynamic principles is important to maximize drug exposure for organisms with elevated minimum inhibitory concentrations (MICs), patients with excess body mass indices, and for closed-space or otherwise difficult to penetrate sites of infection (e.g., meningitis, endocarditis, pneumonia, and bone and joint infections). A recent paper provides a more in-depth review of the subject and serves as a primer for all ASPs that are incorporating optimal dosing strategies [59]. Although MMC approached its program with the thought that many patients would receive downward dose adjustments for renal impairment, what we found was a significant proportion of patients requiring increased drug exposure [8,33]. Other examples of pharmacodynamic dose-optimization include regimens intended to more effectively treat higher MIC pathogens such as continuous or “prolonged” infusion of short half-life β-lactams (e.g., piperacillin/tazobactam, cefepime, meropenem), and extended-interval aminoglycoside dosing. As previously mentioned, MMC has exploited the fact that metronidazole with its prolonged half-life (~10 hours) and active metabolite can be given every 12 hours instead of every 6–8 hours for non-C. difficile, noncentral nervous system infections. As mentioned, MMC has created a clinical program that integrates cefepime and metronidazole into a single administration product that can be infused twice daily, mimicking the spectrum of activity provided by piperacillin/tazobactam [58].
Educational Efforts
The first step in any process leading to change are the development of pertinent information and its dissemination. Early attempts at influencing prescribing behaviors relied heavily on educational efforts: It was simplistically believed that the reason physicians frequently inappropriately prescribed antibiotics was that they were “therapeutically undereducated” [60]. The assumption was that the misuse of antibiotics was more often the result of insufficient information rather than inappropriate behavior.
Over the years that MMC has taught antibiotic principles and specifics of therapy, it has been impressed by the intense interest of both physicians in training and established practitioners in learning more about antibiotics. Equally as impressive is the laissez-faire and even fatalistic attitude toward retaining and applying lessons learned in these educational sessions. Without direct application to current patients, prescribers often refer to antibiotics as “alphabet soup” and “impossible to understand.” These impressions are supported in the literature. Although a supplemental cornerstone to any ASP, educational efforts when applied alone are the least effective and certainly the shortest lasting way to affect prescriber behaviors. Active intervention that is supplemented by education is a synergistic method for changing behavior.
Computer-Assisted Decision Support Programs
Direct computer-based physician order entry (CPOE) is rapidly becoming the standard of care and has been adopted as one of the Leapfrog initiatives to avoid medication errors and improve the quality of care [33]. Computer-assisted decision support programs have been designed to provide real-time integrated patient and institutional data including culture and antimicrobial susceptibility results, laboratory measures of organ function, allergy history, drug interactions, cumulative or customized location-specific antimicrobial susceptibility data, and cost information. These programs provide therapeutic choices for clinicians and allow for the incorporation of clinical judgment by overriding suggestions. Autonomy is preserved while ensuring that important variables in the choice of antimicrobial therapy are considered.
Almost all published data on the effect of computer-assisted decision support programs on antibiotic use are from researchers at the Later Day Saints Hospital in Salt Lake City, Utah. The approach of these studies has been associated with reductions in antibiotic doses, inappropriate orders, costs, treatment duration, and associated adverse drug events [61,62,63]. This degree of computer sophistication is not universally available but has been made available through a variety of commercial systems [64,65]. MMC has used its own CPOE system to design a logic-based algorithm to optimize the treatment of pneumonia (community, healthcare, and hospital acquired). A recent randomized control trial of clinical decision support on the appropriateness of antimicrobial prescribing demonstrated improved appropriateness and reduced overall use of antimicrobials for respiratory tract infections [66]. In this rural outpatient setting, handheld personal digital assistants (PDAs) and paper forms of decision support supplemented the prescribing decision and choice of therapy. Patient-specific data had to be entered into the PDA, which provided a logic-based recommendation that was measurably useful in the prescribing process.
Adaptation of Locally Customized Published Guidelines
National guidelines by the IDSA and SHEA are available and are useful to construct clinical pathways locally for a variety of infections. In some instances as when sufficient time
P.188
has passed between the publication of national guidelines and the change of the disease process, there should be a mechanism to develop updated evidence-based guidelines. A good example of this is the management of CDAD. Being one of the first identified institutions in North America with a hypervirulent strain (BI/NAP1) of C. difficile, MMC saw the clinical and prescribing impact almost immediately. Shortly thereafter, the center intervened by developing consensus among feuding specialties on the proper approach to managing CDAD and created guidelines, a clinical pathway, and a follow-up order set, all of which could be accessed on its intranet site and/or in its CPOE system [23]. From the identification of the BI/NAP1 strain of C. difficile to its proper management, the ASP in conjunction with Maine Medical Center's department of epidemiology and infection prevention, environmental services, and administration, it spearheaded an institutional approach to managing this high morbidity-associated infection. MMC also evaluated and reported the impact of the CDAD guidelines that supplemented active interventions made by its ASP on the use of nonevidence-based treatment strategies and demonstrated a significant improvement in the treatment variability of this infection [67].
TGuidelines published by the American Thoracic Society and the IDSA for the management of healthcare-associated, ventilator-associated, or hospital-acquired pneumonia suggest a very broad spectrum approach for the empiric treatment of these infections because of the high probability of mortality associated with inadequate therapy. In addition, these guideline recommend shortened durations of therapy. From these recommendations, we developed a consensus and published a locally customized (per our susceptibility patterns) guideline and continue to meet monthly to discuss the tracking of process measures. One difficulty with simply creating and making guidelines available to clinicians is that compliance is voluntary, and we have found that without active follow-up, clinicians fall back on old habits. Thus, one benefit of having an ASP is having the resources to provide active intervention in the intensive care unit (ICU) from whether it is recommending streamlining/deescalation when culture and susceptibility results are known to stopping antimicrobial therapy at day 7–8 instead of the traditional 14 or more days.
Process and Outcome Measurements
The IDSA/SHEA guidelines for developing an institutional program to enhance antimicrobial stewardship recommend that outcomes be measured [6]. This guideline is one reason for having a data system and an information specialist assist the ASP members in quantifying their impact. Without this support, the ASP team members could spend more time justifying their positions and measuring outcomes than on the actual day-to-day functioning of the program and evaluating antimicrobial therapy, which is the team's primary purpose. Antimicrobial consumption can be measured for targeted (or all) antimicrobials. Using antimicrobial expenditure data has significant limitations but is helpful in evaluating where money is being spent. A more meaningful measure of antimicrobial consumption is the use of DDD data; standardized definitions are available at www.whocc.no/atcddd/. Converting grams of antibiotic used to DDD per 1000/patient-days allows for a useful internal and external benchmark of antimicrobial consumption. Other measures include antimicrobial days of therapy. Regardless of mechanism chosen, establishing a baseline of antimicrobial use before implementing a program allows the team to track the progress of interventions on use over time. These measures also can be used to quantify the impact of parenteral to oral conversions. In addition, periodically reporting to the pharmacy and therapeutics committee or other committee structure allows other clinicians and administrators to be aware of both the successes and the challenges the ASP has faced.
Optimizing Outpatient Antimicrobial Use
Although a departure from the institutional approach to programmatically addressing antimicrobial stewardship, discussing the importance of antimicrobial stewardship in the outpatient setting is worthwhile. More casual prescribing of antimicrobials, patient demand, and managed care constraints have plagued the optimal use of antimicrobials in this venue. Many times the prescription of an antimicrobial is issued under less than scientific circumstances and with less information available. Unfortunately, what Jawetz [4] reported in the 1950s it is still true today.
He (the physician) is under great pressure to prescribe the “newest,” “best,” “broadest,” antibiotic preparation, prescribe it for any complaint whatever, quickly, and preferably without worrying too much about specific etiologic diagnosis or proper indication of the drug. The pressures come from several main sources: (a) In lay magazines and newspapers patients read exaggerated, uncritical, and often misleading claims made for newly discovered drugs. “Scientists announce new potent weapon against colds.” “Antibiotic cures and prevents many infections.” “New drug saves lives.” Most of these accounts are quite meaningless, yet patients proceed to demand the new, marvelous drug from their doctors. The physician may be embarrassed to admit that he knows nothing of this supposed discovery (many doctors find it necessary to read medical news in Time, Reader's Digest and similar media to cope with their patients' pseudo-knowledge), or, he may prefer not into lengthy explanations as to why he thinks little of the new drug. It may be simpler and quicker to yield to the patient's insistent demand, and prescribe.
So have we made any progress?
P.189
To Treat or Not to Treat
National guidelines for the management of a variety of upper respiratory tract infections (URTIs) provide objective criteria for when to prescribe antimicrobials [68]. Optimizing treatment begins with this question, “Does the patient need antibiotics?” In both pediatric and adult practices, the patient often extracts the unnecessary antibiotic prescription from the time-strapped clinician. Additionally, national guidelines for the diagnosis and treatment of URTIs provide basic tenants for the treatment of these infections, but clinicians unfortunately appear to frequently ignore them. In a study of two private practices, 71% of pediatricians indicated that a parent had requested an unwarranted antibiotic at least four times within the previous month [69]. In 35% of these instances, the pediatrician admitted prescribing an antibiotic. In 61% of instances, the parent requested a different antibiotic than that selected by the prescriber.
In addition, a laissez-faire attitude regarding national treatment guidelines for diagnosing and treating infection exists. In a study conducted by the CDC, pediatricians and family practitioners were evaluated for self-reported versus actual practice regarding antimicrobial use for URTIs [70]. While 97% agreed that the overuse of antimicrobials is a chief factor driving antimicrobial resistance and 83% believed that they should consider selective pressure for antimicrobial resistance when deciding to prescribe antibiotics for URTIs, a large contingent ignored basic tenants of judicious antimicrobial use. For example, 69% considered purulent rhinitis diagnostic for sinusitis, 86% prescribed antimicrobials for bronchitis regardless of the duration of cough, and 42% prescribed antimicrobials for the common cold [70]. In addition, family practitioners were more likely than pediatricians to omit the requirement for prolonged symptoms to diagnose and treat sinusitis (4 versus 10 days, respectively) and to omit laboratory testing for pharyngitis (27% versus 14%, respectively) [70].
Outpatient Interventions to Improve Antimicrobial Use
Several studies have demonstrated successful outcomes in outpatients including a decrease in overall antimicrobial use and improvement in the appropriateness of therapy. Several methods have been used to effect change including education, consensus guidelines, data feedback, medical information system reminders, financial disincentives, and the use of opinion leaders [71,72,73]. Most of the literature evaluating the impact of interventions on antimicrobial use has occurred in the acute care inpatient setting and may not be generalizable to the outpatient venue. With that said, there is an increasing body of literature is evaluating the impact of multiple educational strategies aimed at both the prescriber and the patient.
Razon et al. [74] conducted a one-day seminar on the diagnosis and prudent use of antimicrobials for the treatment of URTIs in children. Using a quasi-experimental study design to determine the impact of the educational intervention, the researchers determined that the appropriateness of treatment improved for both otitis media (OR 1.8, p < 0.01) and pharyngitis (OR 1.35, p < 0.01) [74]. In addition, overall antimicrobial use for otitis media and URTI decreased from (p < 0.05) [74]. No change in appropriateness or antimicrobial use was noted for sinusitis, however.
A statewide educational intervention performed by the Wisconsin Antibiotic Resistance Network (WARN) used a two-pronged approach. On one hand, clinicians were targeted recipients of education at professional conferences, meetings, grand rounds, satellite conferences and multiple mailings and CD-ROM presentations were distributed. On the other hand, the public was educated by means of multilingual brochures and posters, tear-off sheets, coloring sheets, stickers, magnets, and handouts. These educational means were distributed statewide to clinics, pharmacies, childcare facilities, managed care organizations, and community groups. In addition, mass media events included radio and television advertisements. Minnesota was used as the control state. Postintervention (in 2002), Wisconsin clinicians perceived a significant decline in the number of requests by patients for antimicrobials (50% in 1999 to 30%; p < 0.001) and in requests by parents for their children (25% in 1999 to 20%; p = 0.004) [75]. Decisions of clinicians in Wisconsin to treat with antimicrobials were less influenced by nonpredictive clinical findings (purulent nasal discharge, p = 0.044; productive cough, p = 0.010) [75]. For both states in the post intervention period, treatment scenarios involving viral respiratory illnesses in adults were less likely to include antimicrobials; however, the same scenario in pediatric patients was only lower in Wisconsin.
In a somewhat similar intervention, Rubin et al. [76] evaluated their efforts to improve antimicrobial prescribing for URTIs in a rural community. They used patient education materials, a media campaign, a small group session conducted with physicians, and a treatment algorithm for URTIs. Although Medicaid claims data and community pharmacy data demonstrated a reduction in the rate of antimicrobial prescriptions, a third data-source (using medical record review) did not support a reduction in diagnosis-specific rate of antimicrobial use. All three data sources did, however, demonstrate a reduction in macrolide use. In a similar rural community, clinical decision support (on paper and handheld computers) to aid clinicians in the diagnosis and management of acute respiratory tract infections plus a communitywide educational intervention was compared with a communitywide educational intervention alone [77]. The communitywide educational intervention with clinical decision support demonstrated a reduction in overall antimicrobial use and increased the appropriateness of therapy.
P.190
Changing prescribing behaviors is one thing; maintaining those changes is another. As aids to sustaining any process improvement effect, mechanisms available to inpatient facilities (e.g., point-of-use information reminders and computerized decision support as well as formal ASPs) are needed to sustain outcome measures. As Samuel Johnson said >200 years ago, “Men more frequently need to be reminded than informed.” The long-term effectiveness of this strategy has been demonstrated in hospitals [8,9], but the optimal method to sustain initial efforts in the community remains unclear. Of the tested outpatient strategies, combining clinical decision support tools in addition to education appears to provide the most promising opportunity for sustainable antimicrobial stewardship. Perhaps as a result of many of the state and local educational interventions, a study conducted by the CDC documented a significant decline in antimicrobial use in the pediatric population between 1989–1990 and 1999–2000 [78].
Conclusions
The problem of increasing antimicrobial resistance—due in part to suboptimal antimicrobial use coupled with the fact that a growing number of pharmaceutical companies have abandoned anti-infective research and development—has resulted in a growing public health crisis. Because of their intensity of antimicrobial use, both institutional and community settings are target-rich environments for proactive interventions to improve antimicrobial stewardship. Various studies have concluded that programmatic means to steward the use of antimicrobials optimizes patient safety, addresses antimicrobial resistance, reduces unnecessary antimicrobial use and, as a side effect, minimizes direct and indirect costs to the healthcare system. The IDSA/SHEA guidelines for developing an institutional program to enhance antimicrobial stewardship serve as a starting place for institutions considering adopting an ASP. Finally, Calvin Kunin, MD, once 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 [antimicrobial stewardship] can save lives as well as money” [79].
References
P.191
P.192