Antimicrobial control programs have become commonplace in healthcare. Specific controls vary, but may include formulary restrictions, prior authorization programs, clinical pathways, consultation services, streamlining and provider education. Hospitals are the most common site of antimicrobial control programs, although some health systems have adopted these measures.[1]

The purposes behind antimicrobial control policies are as varied as the types of interventions. Antimicrobial controls have been used to augment traditional infection control programs with the hope of influencing antimicrobial susceptibilities. Economic principles have also driven the development of many control policies. Finally, programs have been instituted to improve the quality of care of individual patients, including clinical cure and minimization of adverse drug events.

Evaluation of the success of multiple programs with multiple goals can be a difficult task within an institution or health system. Most frequently, the documented success of programs has been largely evaluated with respect to a single outcome — prescribing habits, antimicrobial resistance, cost, or patient outcomes. Investigations have inconsistently reported each of these effects. The effectiveness of changing antibacterial prescribing behavior has been reviewed in detail elsewhere.[2] While brief reflections on economics and bacterial susceptibility changes are required for a complete overview of the issues pertaining to antimicrobial control, the current review will focus on the impact of antimicrobial control programs on patient care.

A comprehensive description of each antimicrobial control program is beyond the scope of this review. Rather, our aim is to offer an overview of the various types of antimicrobial control programs that are used, with an emphasis on the reported effects of these programs on patient-specific outcomes. In addition, particular limitations to the current body of published research examining antimicrobial control will be addressed. Finally, the particular role of the pharmacist in antimicrobial control (as well as avenues for expansion of this role) will be discussed.

Studies published in English were identified through a Medline search from 1966 to June 2003. Search terms included the MESH headings ‘anti-infective agents’ or ‘bacterial infections and mycoses’ with terms including: ‘critical pathways’, ‘disease management’, ‘drug therapy’, ‘drug utilization’, ‘formulary’, ‘guidelines’, ‘medication systems’, ‘mortality’, ‘outcome’, and ‘pharmacy’. A review of all studies that reported clinical outcomes was performed. Further studies were obtained by review of the bibliographies of relevant articles.

1. Description of Antimicrobial Control Programs

The antimicrobial control policies that will be considered in this review have been grouped into seven major categories. While some of these categories overlap, they have been divided below to best allow further evaluation of each component. A more comprehensive analysis of reported effects on outcomes for each type of control program is presented in section 6.

1.1 Formularies

Maintenance of a formulary of available antimicrobials offers broad control over prescribing. Formularies restrict the classes of antimicrobials available to clinicians as well as the agents available within those classes.[3] Because antimicrobial mechanisms of action and bacterial mechanisms of resistance vary among drug classes, it is uncommon for formularies to exclude an entire class of antimicrobials. However, individual agents within a class are often limited for reasons relating to cost, adverse effect profile, and the potential to induce bacterial resistance. Therapeutic interchange — the automatic replacement of a non-formulary prescribed agent with an agent contained within the formulary — is an active form of formulary control. In addition, antimicrobial order forms are often employed to limit prescribing to a predetermined antibacterial, usually linked to a specific indication. Antimicrobial cycling programs (programs that entail systematic use of only certain antimicrobials for treatment of particular pathogens for a predetermined period of each year) are considered as another type of formulary system. Most institutions with formulary controls have policies in place for the exceptional cases where non-formulary antimicrobials are required.

1.2 Restricted Access

Restricted access to certain antimicrobials limits the prescribers who may utilize them. In the case of antimicrobials, this is often a restriction to infectious diseases physicians or to specialists treating patients with special needs (e.g. bone marrow transplant recipients, pediatric patients). Prior approval, often from clinicians holding no restrictions on use, is required before other clinicians may utilize these antimicrobials. Although restricted access is related to the formulary control system, it is considered as a separate system because its potential impact on patient care appears to differ from the mere presence or absence of antimicrobials on a formulary.

1.3 Clinical Pathways

The adoption of clinical pathways is increasing for commonly encountered infectious diseases.[4–6] Clinical pathways for the purposes of antimicrobial control include peer-reviewed guidelines for the treatment of a particular infection (e.g. nationally adopted guidelines for the management of community-acquired pneumonia), use of standardized antimicrobial order forms for the treatment of a given infection or pathogen, and institution-specific care plans for a given infection. The aim of such programs is generally to limit the available options for the management of a particular infection. Similarly, automatic stop orders are sometimes employed to limit the use of prophylactic or empiric antibacterials that may require only transient use unless reordered. The use of electronic prescribing aids (i.e. computerized systems that may be used by prescribers to assist in antimicrobial selection) are also considered an example of a clinical pathway.

1.4 Antimicrobial Management Services

An antimicrobial management service (AMS) generally oversees hospital-wide antimicrobial use, and evaluates such use for appropriateness and/or adherence to institutional restrictions. These services may be traditional physician consult services, or multidisciplinary teams. In addition to their involvement in other measures of antimicrobial control (e.g. approval of restricted agents, streamlining, intravenous to oral switches), antimicrobial management services review the use of non-restricted antimicrobials and recommend therapy alterations based upon review of patient data.

1.5 Streamlining

Antimicrobial streamlining is the process of matching antimicrobial therapy to culture and susceptibility data. As microbiology data become available, individual patients are screened to ensure that antimicrobial coverage that is appropriate for the micro-organism(s) isolated is employed. The individuals who screen the data typically make suggestions (oral and/or written) to the primary medical team regarding any deficiencies or duplications in current therapy. An additional aid to streamlining therapy is the selective release of antibacterial susceptibility data for specific organisms. For example, broad-spectrum antibacterial susceptibility is often withheld from general reporting when a narrow spectrum agent may be most appropriate for a given organism. This would indirectly narrow perceived antibacterial choices for a given infection.

1.6 Intravenous to Oral Conversion

Targeted antimicrobial adjustments made in response to individual patient characteristics constitute another form of antimicrobial control. The most common adjustments include dosage form changes and dose changes, which do not constitute an overt antibacterial ‘control’. Intravenous to oral conversion, however, is a commonly targeted adjustment that does serve as a true control measure. In this system, reports are generated that list those intravenous antimicrobials that may be changed to an oral regimen with similar antibacterial activity. Those involved in management of the control system (who are often pharmacists) evaluate the suitability for oral therapy and contact the prescribing clinicians to suggest the change.

1.7 Provider Education

Various educational programs have been devised to improve antibacterial selection. These include educational mailings, direct prescriber education and peer review of prescribing habits.[2] All of these methods attempt to influence clinicians to adopt and maintain the institutionally approved best practices.

2. Prevalence of Institutional Antimicrobial Control Programs

The prevalence and scope of antimicrobial control policies used in hospital settings have been reported over time.[7] A 1983 report from 108 hospitals in the US with affiliated medical schools identified antimicrobial restriction in 57% of hospitals and a closed formulary in 68%.[8] A 1994 report from 431 hospitals in the UK indicated that 79% of hospitals used a formulary system, with 77% of hospitals having further restrictions on the use of specific antimicrobials.[9] A US study in 1996 reported similar findings, with 81% of hospitals utilizing restrictions.[10] Finally, a 2000 investigation surveyed 47 hospitals involved in data collection regarding antimicrobial resistance in the US.[11] All hospitals used a formulary, with 40% of hospitals having further restrictions. Clinical practice guidelines were used in 70% of hospitals. Infectious diseases consultation services were utilized in 70% of hospitals, while 64% used a formal policy of antimicrobial streamlining. Although the above studies surveyed different institutions in various geographic regions, there appears to be a trend towards more common use of control programs than in the past.

3. Pharmacy Involvement in Antimicrobial Control Programs

Pharmacists, as specialists trained in the evaluation, selection, and monitoring of patient-specific drug therapy, have several potential roles in antibacterial control programs. These roles can be divided into both administrative and clinical duties.

3.1 Policy Development

Pharmacists can play a vital role in the decision-making process of the antibacterial formulary. These decisions are generally made through multidisciplinary hospital Pharmacy and Therapeutics (P & T) committees. Pharmacists have a unique role to aid in determining the selection of antibacterials based on safety, tolerability, efficacy, cost, pharmacokinetic profile, and resistance selection potential. Similarly, clinical pathways and prescriber restrictions are typically reviewed by the hospital committees, again allowing for pharmacy involvement in the antimicrobials utilized.

3.2 Clinical Practice

Although clinical pharmacy skills are certainly required in the administrative processes, additional training in infectious diseases pharmacotherapy is imperative for advanced practice activities. In the US, postgraduate pharmacy residencies and fellowships in infectious diseases pharmacotherapy are available.[12,13] These programs train pharmacists in patient care, antibacterial stewardship, and infectious diseases research. Pharmacists trained in infectious diseases are important members of antimicrobial management services and are often the primary practitioner involved in streamlining and medication route of administration changes. Pharmacists are often involved in approval of restricted antibacterials. Additionally, trained pharmacists often lead educational efforts to improve the quality and cost of antibacterial usage. Infectious diseases pharmacists can aid in antibacterial decisions made for individual patients, as well as contribute to hospital-wide policies.

3.3 Documentation

Pharmacists have a role in documenting the effectiveness of control programs on clinical outcomes, economics, and resistance. This involves the identification of current problems and investigating the changes that occur over time. Documentation and research into these effects are imperative to the development and maintenance of effective controls.

4. Antimicrobial Control Programs and Antimicrobial Resistance

Numerous antimicrobial control programs have been implemented in order to curtail the progression of bacterial resistance. The theory supporting this idea holds that if antimicrobial usage is optimized, and, for example, use of agents with a particularly broad spectrum of antimicrobial activity is limited, selection pressure will be minimized and bacterial susceptibilities will be maintained or even improved. Antimicrobial restriction programs are often implemented with this goal in mind. For example, limitation of vancomycin use will theoretically limit the proliferation of vancomycin-resistant enterococci, while limitation of the use of certain cephalosporins may reduce the incidence of various β–lactamase-producing bacteria that are often resistant to multiple antimicrobials.[14,15] Of note, the available evidence supporting the notion that a reduction in antibacterial use results in a decrease in antimicrobial resistance is surprisingly sparse. Most available data suggest that such a relationship exists, but is not conclusive.[16]

While a reduction in bacterial resistance is obviously an important goal, interpretation of the effects of bacterial resistance on patient outcomes is problematic. First, a direct link between bacterial resistance and adverse clinical outcomes is often difficult to establish. In a meta-analysis, Cosgrove et al.[17] found that bacteremia caused by methicillin-resistant Staphylococcus aureus was associated with significantly higher mortality than bacteremia caused by methicillin-susceptible S. aureus. Various other studies have found that infection with a resistant bacterium leads to increased mortality and added costs; these adverse outcomes are often attributed to the fact that presence of a resistant organism naturally leads to a reduction in the efficacy of empiric antimicrobial therapy.[18] However, several studies examining the impact of resistance have found that infection with a resistant bacterium does not necessarily lead to increased mortality.[18] Methodologic concerns, especially a lack of appropriate patient controls, are said to be responsible for many of these equivocal findings.

Lesch et al.[19] conducted a retrospective analysis of the relationship between antimicrobial usage and the prevalence of bacterial resistance. A significant correlation between usage of certain antimicrobials and resistance prevalence was found. For example, usage of ceftazidime correlated with the prevalence of ceftazidime-resistant Pseudomonas aeruginosa. In addition, de Man et al.[20] examined the effect of using alternating standardized regimens for the treatment of neonatal sepsis, and found that a particular regimen (amoxicillin + cefotaxime) resulted in diminished resistance among Gram-negative organisms. These results have been corroborated in numerous other studies.[21–24] However, in each of these studies, the impact of resistance on subsequent patient outcomes was not evaluated.

When evaluating the impact of control programs designed to reduce resistance, a major limitation is the inability to link control efforts with patient outcomes through the intermediate outcome of a reduction in resistance. Even if an antimicrobial control program is effective in limiting bacterial resistance, a subsequent improvement in clinical outcome might rarely be detectable in a given individual patient. Rather, improved outcomes will generally be observed on a hospital-wide scale.[25] However, linking these three variables (antimicrobial control, bacterial resistance, patient outcomes) then becomes difficult because of the presence of countless confounding factors that preclude valid correlations.

Given these problems, in this review we do not consider a reduction in resistance to be a patient-specific outcome resulting from antimicrobial control.

5. Antimicrobial Control Programs and Economics

Because cost containment has been another major reason to implement antimicrobial control programs, considerable data are available on the economic impacts of these programs. Reduction in drug expenditures has been the primary subject of multiple investigations.[26–28] Results of studies of general formulary control and restricted access programs have suggested economic advantages.[29,30] Similarly, it is relatively simple to evaluate the before and after antimicrobial expenditures of programs such as streamlining and consultation services. However, the economic outcomes of antimicrobial control programs certainly extend beyond antimicrobial cost-minimization and often extend into cost-benefit analyses incorporating differences in patient outcomes.[31] Because the economic impact of antimicrobial control programs is not the focus of this review, we do not review investigations reporting only economic outcomes. However, we do examine any reported clinical outcomes that have been related to these cost-savings interventions.

6. Antimicrobial Control Programs and Patient Outcomes

Patient outcomes in the treatment of infections are best described in terms of clinical cure or clinical failure. The resolution of the infection and its accompanying signs and symptoms at a given timepoint constitute clinical cure. Clinical failure can be defined by continuation of infectious signs and symptoms, with death attributable to infection being the most extreme negative outcome. Also inherent in patient outcomes are the presence or absence of iatrogenic events, most notably adverse drug reactions. Unfortunately, the effects of antibacterial control programs on individual patient clinical outcomes and/or adverse drug events have not been adequately collected in most investigations. ‘Appropriate’ antibacterial use, based upon available guidelines or microbiological results, is often used as a surrogate marker for clinical outcomes. Although there are data to support a link between appropriate drug use and favorable outcomes, this correlation is not absolute. Patients receiving the ‘correct’ drug are not guaranteed clinical cure.

In this article we included only studies that reported clinical response, adverse events and mortality as outcome measures. Although mortality related to infectious diseases is not always easily extracted from comorbid conditions, it is an absolute endpoint commonly reported in most investigations. Assignment of mortality attributable to infection has been attempted, but is not universally accepted and infrequently reported. Matching of severity of illnesses has been attempted to determine infection-related mortality.[32–35] In these studies, crude mortality overestimated mortality attributable to infection by 2–18 times.[32–35]

Notably, we excluded studies that used only length of stay as an outcome measure for antibacterial control programs. Length of stay is a questionable surrogate marker of clinical outcomes.[36–38] There are little data to suggest that hospital length of stay is highly correlated with resolution of infection. Without rigorous matching for disease severity and comorbid illnesses, length of stay becomes imprecise as an outcome measure.

6.1 Formularies

Few trials exist that have systematically examined the impact of a formulary system on such endpoints as clinical outcomes or antimicrobial resistance. Rather, many studies examining implementation of a formulary system have simply measured subsequent compliance with that system, as measured by prescribing practices, or have measured reductions in cost and/or the number of doses of a particular antimicrobial as primary outcome measures.[39] For the purposes of this review, we did not consider these outcomes.

Several trials of formulary programs have included bacterial resistance (and/or recovery of a resistant bacterium from patients studied) as the only outcome measure.[21–24] For example, Moss et al.[24] found that institution of an antibacterial cycling program (consisting of three 3-month periods using distinct broad-spectrum antibacterial regimens for all empiric therapy) resulted in nonsignificant reductions in the prevalence of patients colonized with antibacterial-resistant bacteria, and in the prevalence of bacteremias caused by resistant micro-organisms. Similarly, Gerding et al.[23] showed that use of amikacin as the aminoglycoside of choice resulted in a reduction in gentamicin and tobramycin resistance, and that this resistance returned upon an increase in gentamicin usage. Of note, several so-called ‘cycling’ programs, such as that described by Gerding,[40] did not fulfill the strict definition of cycling, which mandates the use of planned antibacterial rotation periods of a fixed duration.

Certain studies (most notably, studies of antimicrobial cycling programs) have included an examination of individual patient outcomes (see table I). For example, Kollef et al.[41] described a cycling program designed to limit the use of ceftazidime and ciprofloxacin for empiric therapy of infections caused by Gram-negative bacteria. Outcome measures examined included the incidence of ventilator-associated pneumonia (VAP) and nosocomial bacteremia. Statistically significant reductions in the incidence of VAP (11.6% to 6.7%, p = 0.028), microbiologically confirmed VAP (9.6% to 5.5%, p = 0.043), and VAP caused by a resistant bacterium (4.0% to 0.9%, p = 0.013), were noted during the cycling period, while nonsignificant reductions in bacteremia, bacteremia caused by a resistant micro-organism, sepsis, severe sepsis, and septic shock were noted. Also of note, a nonsignificant increase in overall hospital mortality was found during the cycling period, while a nonsignificant decrease in attributable mortality was noted.[41] Gruson et al.[42] also found that implementation of a cycling program resulted in a statistically significant reduction in the incidence of VAP (22.1% to 15.7%, p < 0.01), but noted a nonsignificant reduction in mortality.

Table I
figure Tab1

Antimicrobial control programs: impact of formularies on clinical outcomes

Full size image

Raymond et al.[44] published the results of a prospective trial examining the effect of antibacterial cycling on the incidence of infectious complications arising from resistant micro-organisms. The cycling protocol consisted of a quarterly rotating schedule of empiric antibacterial regimens in a university medical center intensive care unit. Data were collected during the 1-year period when the cycling protocol was in use, as well as during the preceding year, when antibacterial therapy was not regulated. The study found that implementation of the cycling program resulted in statistically significant decreases in the number of infections per 100 patient admissions (42.3% to 32.2%, p < 0.0001) and in infections caused by a variety of resistant bacteria. In addition, statistically significant reductions were observed in crude mortality following infection (38.1% to 15.5%, p < 0.0001), and attributable mortality in the intensive care unit (56.4% to 34.8%, p = 0.035).[44]

Importantly, Woodward et al.[40] claimed that institution of a formulary restriction program (that restricted the use of aminoglycosides, cephalosporins, penicillins, and vancomycin) resulted in no increase in hospital mortality or length of stay. However, the methodology used for analysis of these outcomes was not described in this study, nor were data regarding these outcomes included. These investigators did find that the restriction policy led to a reduction in antibacterial costs, although without a controlled examination of subsequent patient outcomes this outcome carries questionable significance.

As a final note, it should be pointed out that, of the antimicrobial control programs discussed in this review, formulary systems in particular may pose a risk of fostering unintended negative outcomes. First, a limited formulary may prevent access to antimicrobials that might be the most appropriate for a given patient. Second, construction of a formulary with the goal of minimizing resistance to excluded agents may in fact result in an increase in resistance to those antimicrobials included on the formulary, or may cause alterations in the bacterial species predominating in a certain institution (the latter consequence is particularly a risk of cycling programs). For these reasons, some have proposed that the typical ‘closed’ formulary may result in more harm than good; these authorities have instead suggested that a completely open formulary be used as the standard form of this system.[46] Regarding cycling programs in particular, further data are required before an overall assessment of their virtue may be formulated.

In general, studies of formulary systems (including cycling programs) have demonstrated that such programs lead to cost savings and reductions in the incidence of bacterial resistance, as well as in such outcomes as the incidence of pneumonia or bloodstream infection. However, these studies did not focus on mortality as an outcome measure, and, for the most part, nonsignificant changes in mortality after implementation of a formulary program were noted. Inclusion of mortality as a primary outcome measure in future studies of formulary systems is warranted.

6.2 Restricted Access

A restricted access program is generally designed to reduce bacterial resistance, adverse drug events, and/or cost through limitation (or outright restriction) of the use of certain targeted antimicrobials. Restriction policies pose many of the theoretical institution-wide risks that are associated with formulary systems. For example, restriction of a certain antimicrobial may simply result in an increased use of a comparable agent from the same or a different class. In addition, the effects of restriction policies on subsequent bacterial resistance appear to be varied, and unintended increases in resistance among certain micro-organisms may result from implementation of a restricted access program.[46] Finally, restriction policies may result in unnecessary delays in therapy if the policy is not efficiently managed. Pharmacists can thus play a leading role in successful implementation of a restricted access program.

As in the case of formulary systems, many studies[47–50] of restriction programs have simply reported physician compliance with the program (by evaluating ‘appropriateness’ of use of restricted antimicrobials). Alternatively, acceptance of the program by physicians has been reported.[47–50] Some studies have examined the cost impact associated with implementation of a restricted access program, but on a pharmacy cost basis, and not in terms of patient care. Other studies have examined the incidence of infection caused by certain micro-organisms after implementation of a restricted access program.[51]

Certain trials, however, have examined the impact of antimicrobial restriction on patient outcomes (see table II). For example, White et al.[52] reported the effect of a prior authorization policy on clinical outcomes, including survival and time to discharge after the first positive blood culture. Interestingly, when survival results were stratified according to the micro-organism responsible for infection, survival rates for infection caused by certain bacteria were reduced after implementation of the restriction policy. However, none of these differences reached statistical significance. A nonsignificant reduction in 30-day survival (79% to 75%) resulted from implementation of the restriction policy.

Table II
figure Tab2

Antimicrobial control programs: impact of restricted antimicrobial access on clinical outcomes

Full size image

Bamberger and Dahl[53] found that a prior authorization program (focusing on third generation cephalosporins) resulted in a nonsignificant reduction (32% to 25%) in mortality. Interestingly, Gentry et al.[54] found that implementation of a restricted access program resulted in statistically significant reductions in mortality (8.3% to 6.6%, p = 0.007) and 30-day hospital admission rates (13.2% to 10.8%, p < 0.0001). However, the comparison groups used in this study appear inappropriate, since all patients (rather than only those whose treatment was affected by the restriction program) were included in the analysis.

No other published studies of the effect of restriction programs have found statistically significant differences in the outcomes we have chosen to focus on. As is the case with studies of formulary systems, further trials examining the impact of restricted antimicrobial access on patient-specific outcomes would be valuable.

6.3 Clinical Pathways

Clinical pathways are designed to improve patient care. The use of infectious diseases clinical pathways incorporates antimicrobial selection suggestions based upon local microbiology data and expert clinician input. Unlike some of the previously discussed antimicrobial control programs, intuitively there appears to be little risk of a detrimental effect on patient outcomes arising from the use of a clinical pathway.

As seen in other control programs, many studies evaluating the effectiveness of a clinical pathway have focused on outcomes such as implementation cost and prescriber acceptance.[55,56] However, several prospective trials have examined the impact of clinical pathways (most notably, national guidelines) on patient outcomes (see table III). In general, these trials have found no statistically significant differences resulting from use or implementation of a clinical pathway. However, Al-Eidan et al.[7] found that use of a prescribing protocol for the management of community-acquired lower respiratory tract infections was associated with a statistically significant reduction (31.3% to 7.8%, p < 0.001) in the incidence of treatment failures. A nonsignificant reduction in the number of adverse drug reactions was also noted. Furthermore, Pestotnik et al.[57] found that use of a computerized management program (consisting of a system containing patients’ clinical records with a linked information system and online clinical decision support) resulted in a statistically significant reduction in mortality (3.7% to 2.7%, p < 0.001). In addition, a nonsignificant reduction (26.9% to 18.8%) in the incidence of adverse events was observed. These reductions were seen between the years 1988 and 1994, during which the study was conducted (the computerized management system was instituted in 1988). A limitation of this study lies in the fact that patients from the pre-intervention period were not included; thus, the effect of the antibacterial management system over time was in fact the ‘intervention’ studied. Importantly, however, only patients treated with antibacterials were included in the analysis.[57]

Table III
figure Tab3

Antimicrobial control programs: impact of clinical pathways on clinical outcomes

Full size image

In summary, the clinical pathway studies described in this section consistently included patient-specific outcomes such as mortality in their analyses. Although most studies failed to detect statistically significant differences, the papers published by Al-Eidan et al.[7] and Pestotnik et al.[57] did show that clinical pathway programs may result in statistically significant reductions in mortality.

6.4 Antimicrobial Management Services

An AMS has the potential to improve patient outcomes through selection of appropriate antimicrobials on a case-by-case basis. Since these programs are generally operated by clinicians with an expertise in infectious diseases, an AMS should improve the quality of care. Theoretically, the potential for a negative impact on patient care is low.

Four published studies reported the effects of an AMS on patient outcomes. These reports of AMS reveal the involvement of multiple clinicians including physicians, pharmacists and microbiologists (see table IV). All services reviewed data for selected hospitalized patients, generally those with orders for a predefined group of antimicrobials. All but one of the AMS provided an initial consultation and continued follow-up of these selected patients.[54,64,65] While all of the reports included mortality as an outcome, only one specifically documented clinical or microbiologic outcomes of infection, with no differences observed.[64] The need for antibacterials within 7 days of cessation of initial therapy was significantly decreased in the intervention group. Statistical improvements in mortality were observed by Gentry et al.[54] (see section 6.2). No statistically significant differences in any patient outcomes were reported in the other studies. All studies had a primary focus on the economic advantages of AMS.

Table IV
figure Tab4

Antimicrobial control programs: impact of antimicrobial management services (AMS) on clinical outcomes

Full size image

Antibacterial management services have not been well evaluated for their effects on individual patient clinical response to antibacterial therapy. Mortality differences were not observed in any of these trials. While antibacterial use and economic parameters were largely improved in these studies, it is difficult to evaluate any direct impact on clinical cure.

6.5 Streamlining

Antimicrobial streamlining controls drug usage by recommending changes based upon culture and susceptibility reports. Recommending antimicrobial changes based on microbiology data is aimed at assuring that current therapy is active against the infecting pathogen. This would be anticipated to improve patient outcomes. Little concern exists for a negative impact of such a program on patient outcomes.

Streamlining is often a part of AMS functions, although it can occur as an isolated control measure. Little data regarding the effect of streamlining on patient outcomes are available. Barenfanger et al.[67] analyzed crude mortality (in addition to the primary analysis of cost) related to streamlining of antibacterials using a computer aid. Three analyses were performed between the intervention and control groups that included (i) only patients with an active intervention (n = 76); (ii) diagnosis-related—group (DRG) matched patients regardless of active interventions (n = 378); and (iii) DRG matched and severity matched groups regardless of active interventions (n = 378).

Nonsignificant changes in crude mortality were observed for analyses (i) 7.7% vs 12.5%; (ii) 11.2% vs 10.0%; and (iii) 11.2% vs 12.6%. No specific clinical outcomes were reported. The variability in mortality underscored the differences observed among the groups analyzed. Failure to adjust for severity of illnesses and the use of patients in an intervention group who have their therapy reviewed but not adjusted, can change the outcome measures.

Computer support was also evaluated in two studies from the same hospital (LDS Hospital in Salt Lake City, Utah, USA) that included streamlining of therapy through a computerized notification system to physicians.[57,60] During a 7-year study beginning with the implementation of the computer aid 63 759 patients were discharged after receiving antibacterials. During the intervention period, antibacterial-related adverse events declined from 26.9% to 18.8%, while mortality decreased from 3.7% to 2.7% (p < 0.001).[57] In a small study in the intensive care unit, the computerized clinical support system was associated with a significant reduction in adverse drug reactions (4 per 398 patients vs 28 per 766 patients; p = 0.018).[60] Mortality rates were reduced with this system (18% vs 27%), but only when the suggested regimens were consistently accepted.

Streamlining of therapy did not result in significant clinical differences in patient outcomes. However, it is important to note that no clinical cure rates were collected in any of these large studies.

6.6 Intravenous to Oral Conversion

Switching from intravenous to oral antimicrobials is generally associated with a reduction in the cost of therapy. The effect of such a switch on clinical outcomes has been evaluated in several prospective randomized clinical trials.[68–71] These trials have shown no declines in clinical outcomes with a planned switch to oral therapy. There is a concern that use of intravenous to oral switch programs may compromise care outside of a carefully controlled clinical trial where inclusion and exclusion parameters are strictly upheld. Conversely, there is a potential clinical benefit to a reduction in intravenous medications, as catheter-related infections and complications may be limited.

In the studies reporting clinical response to infection, no statistically significant differences were observed in response rates between the intervention and control groups[72–75] (see table V). The overall clinical response rates in the intervention groups ranged from 83.3–99%.[72–76] Low mortality rates were reported without significant differences.[72–74,77] One study reported a significant increase in 30-day hospital readmissions in the intervention group (29% vs 9.8%, p = 0.02), but not in infection-related readmissions (8% vs 6%).[71] Combined analysis of adverse drug events in two studies from the same authors reported a nonsignificant increase in all events when interventions were followed compared with controls (14% vs 11%).[73,74] No studies reported significant differences in adverse drug reactions.

Table V
figure Tab5

Antimicrobial control programs: impact of intravenous to oral conversion programs on clinical outcomes

Full size image

No significant improvements in clinical outcomes or decreases in adverse events were associated with the intravenous to oral switching of antibacterials. Similar to the controlled clinical trials, no detrimental effects were realized with these programs.

6.7 Provider Education

Multiple studies have looked at the effectiveness of educational programs in changing antibacterial prescribing behavior.[2] A comprehensive review of these studies reported that pharmacist education of individual practitioners, local guidelines, small-group sessions, feedback to clinicians, and computer-assisted care had positive impacts on antibacterial use.[2] The use of a combination of these programs was shown to improve adherence to predetermined antibacterial choices. However, no studies were identified that specifically investigated the impact of these programs on the clinical outcomes of individual patients.

7. Practitioner Opinions on Antimicrobial Control Programs

Several surveys have been published that report on the overall positive attitudes of hospital practitioners towards antimicrobial control programs.[79–81] For example, physicians were surveyed at three hospitals with three different antimicrobial control systems. Respondents preferred an educational AMS consultation to either no controls or a required prior approval system.[80] Overall, the respondents did not believe that control programs compromised patient care. In a second study, microbiologists and pharmacists reported similar beliefs concerning patient and hospital outcome measures with intravenous to oral conversion programs.[81] Finally, a positive overall attitude to antimicrobial controls was observed in both pharmacists and physicians.[79] Both groups believed that antimicrobial controls improved the quality of antimicrobial prescribing. Differences were observed between physicians in different practice sites, with those working in surgical and pediatric departments reporting a less favorable opinion.

8. Pharmacy Perspective

Pharmacists are often charged with being the ‘antimicrobial police’ in institutional settings. Although cost containment continues to be a major impetus for this distinction, maintenance of quality care is paramount. Quality care includes the clinical outcomes of individual patients and the prevention or reduction of antimicrobial resistance.

Pharmacists have been instrumental in the maintenance of antibacterial control programs. The key role of pharmacists was reported in all but one of the AMS studies reporting clinical outcomes.[54,64,66] Similarly, pharmacy involvement was reported in the majority of antimicrobial streamlining and intravenous to oral conversion programs. The effectiveness of these individual programs is an endorsement of the pharmacist’s role in such programs. Furthermore, there are even data to support an improvement of patient care with a pharmacist making clinical decisions instead of a medical fellow.[82]

Position papers from the medical community have questioned the expanding clinical role of pharmacists.[83,84] It is important that the profession continues to work with the medical community to support clinical care. Pharmacists need to utilize their unique perspective on drug therapy and work collaboratively with prescribers to improve antibacterial utilization. Pharmacists should continue to document the impact of these programs and the pharmacy’s role in them.

9. Conclusion

Antimicrobial control programs are common in institutional settings. While economic savings and alterations in antimicrobial resistance are commonly reported outcomes, clinical outcomes have not been well elucidated. While antimicrobial control systems are based on sound medical reasoning and clinical trial data, further studies are required to confirm that positive clinical outcomes result from systematic antimicrobial controls. In particular, it is imperative that well-controlled trials that document the impact of antimicrobial control on clinical outcomes be published.

Clinical pharmacists must continue to have a vital presence in the development, operation, and evaluation of antimicrobial control programs. By virtue of their specific training in pharmacotherapeutic principles, pharmacists are in a key position to serve as vital members of antibacterial control systems. The pharmacy profession as a whole would be well served if pharmacists continued to obtain the training and education necessary to further their participation in antimicrobial control systems. At the same time, pharmacists should strive to work collaboratively with physicians and other healthcare professionals in order to create a successful control program. Finally, ample opportunity exists for pharmacists to perform exactly the type of research we have highlighted in this review as an essential piece of the puzzle proving the worth of antimicrobial control.