Impact of an Unsolicited, Standardized Form-Based Antimicrobial Stewardship Intervention to Improve Guideline Adherence in the Management of Staphylococcus aureus Bacteremia.

BACKGROUND: Antimicrobial stewardship programs (ASPs) improve Staphylococcus aureus bacteremia (SAB) management. The objective of the current study was to evaluate the effect of unsolicited prospective audit and feedback (PAF) using a standardized SAB bundle form on the management of SAB.
METHODS: Multicenter, pre-post quasi-experimental study of inpatients with SAB. The ASP developed an evidence-based SAB management bundle that included recommendations for infectious diseases consultation, blood culture clearance, appropriate empiric and definitive therapy, echocardiography, adequate treatment duration, and source control where applicable. ASP pharmacists performed PAF using a standardized form outlining bundle components. The primary outcome was bundle component adherence. Secondary outcomes were length of stay, 30-day readmission rate, and in-hospital and 30-day mortality rates.
RESULTS: A total of 199 patients were included (preintervention group, 62; intervention group, 137). Bundle implementation with PAF resulted in significant improvements in infectious diseases consultation (56.5% in preintervention vs 93.4% in intervention group), appropriate definitive antibiotic therapy (83.9% vs 99.3%), ordering echocardiography (72.6% vs 95.6%), and adequate treatment duration (87.0% vs 100%) (all P < .001). Overall bundle adherence increased by 43.8% (P < .001). Readmission and 30-day mortality rates decreased, but this difference did not reach statistical significance.
CONCLUSIONS: Unsolicited PAF using a standardized SAB management bundle significantly improved adherence to evidence-based recommendations. This simple yet effective ASP-driven intervention can ensure consistent management of a highly morbid infection.

Staphylococcus aureus bacteremia (SAB) is a serious infection associated with high morbidity and mortality rates, ranging from 10% to 40% [1]. Management of SAB requires a multipronged approach with evidence-based interventions. These include documenting blood culture clearance [1], prescribing appropriate antibiotic therapy [2], timely source control [2], assessing for infective endocarditis with echocardiography [3], and, in particular, infectious diseases (ID) consultation [4]. Multiple studies have demonstrated a mortality benefit from formal ID consultation [4-16]. Despite strong evidence, prescribers may not recognize the gravity of SAB and the need for ID involvement. In previous studies, the baseline ID consultation rate has ranged from 13.3% to 85.4% [4, 17]. A standardized method of informing providers about evidence-based SAB quality-of-care measures and ensuring ID consultation could greatly improve care.

Antimicrobial stewardship programs (ASPs) promote appropriate antibiotic use in order to optimize patient outcomes. Given that ASPs are now required organizational practices, designating SAB management to ASPs can ensure universal optimization of one of the most morbid infectious diseases. Prospective audit and feedback (PAF) by ASPs can be used to target SAB management.

In previous studies, both ID physicians and ASPs have developed and implemented an SAB care bundle [18-20]. However, securing dedicated personnel (ID physicians or pharmacists) or automated computer systems (eg, electronic medical records [EMRs]) to address bacteremias may be challenging. Combining existing resources with a simple intervention may address these deficiencies. A form outlining the evidence-based SAB management measures was developed to assist with PAF implementation by our ASP (see Supplementary Material). The objective of the current study was to assess the impact of this unsolicited, standardized form–based, antimicrobial stewardship SAB intervention on adherence to an evidence-based bundle and patient outcomes.

METHODS

Study Design

This pre-post quasi-experimental study was conducted at 2 acute care hospitals (combined 620 beds). The ASP servicing these hospitals comprises a medical director (ID physician) and 2 ASP pharmacists (1 at each site) who provide unsolicited recommendations to medical teams based on either antibiotic ordered or infectious diagnosis. An independent ID consultation service was available at each site throughout the study period but required formal consultation by the medical team before engagement in patient care. The study was approved by the University of Alberta Research Ethics Office, Covenant Health Research Center, and the Covenant Health Antimicrobial Stewardship Committee with endorsement from the Divisions of Infectious Diseases and Cardiology.

Admitted patients ≥18 years of age with a positive blood culture for S. aureus during the year before intervention (1 October 2014 through 30 September 2015) or the 2-year intervention (1 October 2015 through 30 September 2017) period were included. Patients were excluded if they were transferred from another facility and lacked a positive S. aureus blood culture at 1 of the study hospitals, were transferred or discharged from a study hospital before completion of all SAB pertinent investigations, died or had care withdrawn within 48 hours of the positive blood culture, or left against medical advice before completing antibiotic therapy for SAB.

Microbiological Studies

Each blood culture set comprised 2 separate venipunctures totaling 30 mL of blood, divided equally (10 mL per bottle) among 2 aerobic BACTEC Plus Aerobic/F bottles and 1 anaerobic BACTEC Anaerobic/F media bottle (Becton Dickinson Instrument Systems). After collection, the microbiology laboratory incubated blood culture samples in BACTEC FX automated blood culture systems (Becton Dickinson Instrument Systems) for 5 days. When the growth index of a bottle was positive, a broth aliquot was collected for standard identification studies, including Gram staining (with immediate notification of results to the attending physician), routine subculture, and matrix-assisted laser desorption–ionization time-of-flight mass spectrometry identification using the VITEK MS PLUS system (bioMérieux). For rapid and accurate presumptive identification of S. aureus, a tube coagulase test was used, as well as both SaSelect chromogenic agar (Bio-Rad) and Denim Blue agar (Chromogenic MRSA Screening Agar; Oxoid) when a blood culture smear contained gram-positive cocci in clumps.

Intervention

An evidence-based SAB management bundle was developed by the ASP after review of current literature and aligned with the quality-of-care measures recommended by the Infectious Diseases Society of America methicillin-resistant S. aureus guidelines [21]. The bundle components included ID consultation, obtaining repeated blood cultures until negative, initiation of appropriate empiric and definitive antibiotics, echocardiography, appropriate treatment duration, and source control if applicable (refer to Supplementary Material for the standardized form). An ID consultation was recommended for all patients with SAB, regardless of complexity of infection, and repeated blood cultures were suggested until blood cultures were negative. Before SAB susceptibilities became available, the suggested empiric therapy was vancomycin. If the isolate was confirmed to be methicillin-susceptible S. aureus (MSSA), therapy was optimized to either cefazolin or cloxacillin. Less commonly, if the primary source of infection was polymicrobial, a broader-spectrum β-lactam (eg, piperacillin-tazobactam) may be suggested.

An echocardiogram was recommended for all patients with SAB. A transthoracic echocardiogram was deemed acceptable if the patient was considered low risk for infective endocarditis (absence of permanent intracardiac device, follow-up blood cultures negative by 2–4 days, no hemodialysis, nosocomial acquisition of SAB, no secondary foci of infection, and no clinical signs of infective endocarditis), or was too frail or high risk for transesophageal echocardiograpy. The suggested duration of therapy was based on the complexity of the infection, with ≥14 days recommended for uncomplicated bacteremia, 4–6 weeks for osteoarticular infections, and 6 weeks for endovascular infections (refer to Supplementary Material for further details). Recommendations for source control were made at the time of initial assessment, and adherence was expected to be as soon as medically feasible.

During the preintervention period, ASP activities included PAF of Clostridium difficile infections, and all carbapenem, daptomycin, linezolid, and tigecycline prescriptions. No recommendations regarding SAB management were made in the preintervention SAB population. Throughout the preintervention and intervention periods, physicians continued to receive notification of all positive blood cultures from the laboratory, per standard protocol. During the intervention period, the laboratory also notified the ASP of isolates presumptively identified as S. aureus by email twice daily. On notification, an ASP pharmacist performed a prospective audit of the patient with SAB, reviewed the case with the ASP medical director, and provided unsolicited written and verbal recommendations to the attending physician. ASP recommendations for SAB management were made on weekdays during regular business hours. Culture results on weekends and statutory holidays were addressed the following business day. Recommendations were recorded on an SAB assessment form (outlining the bundle components and supporting evidence) and placed in the patient’s paper medical record. No further recommendations from ASP were made thereafter unless antibiotic susceptibilities were not available at the first encounter, in which case a repeated encounter occurred the following day to optimize definitive therapy. Compliance with bundle components was assessed until clearance of bacteremia and duration of therapy was established.

Vancomycin dosing and monitoring remained the responsibility of the attending team (with support from the team pharmacist) during both study periods. ASP did provide recommendations to optimize vancomycin dosing as part of the SAB bundle, as necessary. A newsletter outlining optimal SAB management, including descriptions of each of the bundle components, was widely posted in the hospitals and distributed through email to all prescribers in advance of the intervention.

Data and Outcomes

Data were collected by retrospective record review for preintervention patients, and by prospective audit for intervention patients, and entered into the Research Electronic Data Capture (REDCap) application. Data collected included age, sex, weight, allergies, and comorbid conditions. The Charlson comorbidity index was used to document conditions associated with mortality in patients with SAB [22]. The following clinical parameters relevant to SAB were also collected: acquisition (community, healthcare associated, or nosocomial acquired according to Friedman criteria [23]), source of bacteremia, simple or complicated bacteremia, and S. aureus antibiotic susceptibility results. A bacteremia was classified as simple if all of the following were met: exclusion of infective endocarditis, no implanted prosthesis present, follow-up blood cultures negative 2–4 days after initial culture, afebrile status within 72 hours of effective therapy initiation, and no evidence of metastatic infection (defined as the presence of a secondary infectious focus due to hematogenous spread of S. aureus [4]) [21]. All other bacteremias were deemed complicated.

The primary outcome was complete adherence to the bundle as well as adherence to the individual quality-of-care measures within the bundle. Empiric and definitive treatment were deemed appropriate if the antibiotic chosen was effective against the S. aureus species identified and the regimen was suitable based on patient characteristics. If MSSA was identified, any antistaphylococcal β-lactam antibiotic (eg, cloxacillin or cefazolin) was considered the appropriate definitive antibiotic, unless contraindicated, based on studies demonstrating the superiority of β-lactams over vancomycin for MSSA bacteremia [24]. Appropriate treatment duration was defined as a minimum of 14 days of intravenous antibiotics from the first negative blood culture for simple cases and at least 4–6 weeks for complicated cases. Only patients with a removable infectious foci were evaluated for adequate source control, including debridement, abscess drainage, and catheter or hardware removal (if possible).

Secondary outcomes included acute length of stay and 30-day readmission, in-hospital mortality, and 30-day all-cause mortality rates. The acute length of stay was calculated as the difference between the hospital admission date and the date of discharge or designation as awaiting long-term care placement with no further active issues. The number, type, and acceptance rate of ASP pharmacist recommendations were recorded.

Statistical Analyses

Categorical variables were compared using χ2 or Fisher exact tests where appropriate. Continuous variables are listed as mean and standard deviation and were compared using the Student t or Mann-Whitney U tests, as appropriate. The Newcombe-Wilson method without continuity correction was used to calculate the confidence interval (CI) for the difference of 2 proportions [25]. For all performed tests, differences were considered statistically significant at P < .05 (2-sided test). Multivariate analysis was performed with logistic regression analysis using the enter method. All models were adjusted to include covariables showing statistically significant differences between the preintervention and intervention periods. All statistics were calculated using SPSS Statistics software, version 24.0. Because several studies have demonstrated a mortality benefit with ID consultation [4-16], this component of the bundle was prioritized for the sample size estimate. Assuming a 50% rate of ID consultation in the preintervention group and 70% in the intervention group, 49 patients were required in each group to achieve a statistical power of 80% and a 2-sided significance level of .05.

RESULTS

A total of 253 patients were eligible for the study. Fifty-four patients were excluded, including 21 who died (preintervention, 5 patients; intervention, 16) or had care withdrawn (2 patients) within 48 hours of diagnosis, 17 who were transferred to or from another facility, 13 who left against medical advice, and 1 who was discharged before completion of investigations. The remaining 199 patients were included (preintervention group, 62; intervention group, 137). Baseline patient demographics and comorbid conditions in the 2 study periods were similar except that the preintervention group had more patients with diabetes mellitus and a Charlson comorbidity index ≥4 (Table 1). A higher rate of complicated bacteremia on SAB diagnosis was observed in the intervention group owing to the presence of a metastatic infection or associated septic arthritis or osteomyelitis. SAB was community acquired in nearly 75% of instances, with the majority of cases due to MSSA. The most common identifiable sources of SAB were skin and soft-tissue infection, septic arthritis, and osteomyelitis.

Table 1.

Baseline Characteristics of Patients with Staphylococcus aureus Bacteremia

Variable	Patients, No. (%)		P Value
	Preintervention Group (n = 62)	Intervention Group (n = 137)
Age, mean (SD), y	68.3 (17.15)	62.9 (20.4)	.07
Male sex	39 (62.9)	79 (57.7)	.49
Intravenous drug use	6 (9.7)	16 (11.7)	.68
Comorbid conditions
Charlson comorbidity index ≥4	46 (74.2)	76 (55.5)	.01
Diabetes mellitus	30 (48.4)	42 (30.7)	.02
Chronic pulmonary disease	18 (29.0)	28 (20.4)	.18
Chronic kidney disease	10 (16.1)	11 (8.0)	.08
Acquisition			.39
Community acquired	45 (72.6)	105 (76.6)
Healthcare associated	4 (6.5)	13 (9.5)
Nosocomial	13 (21.0)	19 (13.9)
Complicated bacteremia	25 (40.3)	94 (68.6)	<.001
Metastatic at onset	14 (22.6)	51 (37.2)	.04
Source of bacteremia
Cardiac	5 (8.1)	5 (3.6)	.29
Joint and osteomyelitis	4 (6.5)	27 (19.7)	.02
Central catheter line infection	4 (6.5)	12 (8.8)	.78
Respiratory	15 (24.2)	15 (10.9)	.02
Skin and soft tissue	9 (14.5)	26 (19.0)	.44
Intravenous drug use	5 (8.1)	14 (10.2)	.63
Unknown	13 (21.0)	31 (22.6)	.79
Other	7 (11.3)	7 (5.1)	.14
MRSA	10 (16.1)b	22 (16.1)	.99
MSSA	53 (85.4)	116 (84.5)	.99

Abbreviations: MRSA, methicillin-resistant Staphylococcus aureus; SD, standard deviation.

Data represent no. (%) of patients unless otherwise specified.

One patient had both MRSA and methicillin-susceptible S. aureus infection.

Vancomycin was prescribed as empiric therapy for 75.8% of patients in the preintervention and 75.9% in the intervention group. For patients with MSSA bacteremia, an appropriate β-lactam antibiotic was prescribed for 96.2% of patients in the preintervention and 98.3% in the intervention group. The ASP pharmacists made 148 recommendations during PAF pertinent to the SAB bundle with a 92.6% acceptance rate. Almost half of the interventions were for ID consultation (44.6%), followed by suggesting echocardiography (19.6%), repeated blood cultures (16.9%), and adjustments to definitive therapy (14.9%). An additional 19 recommendations to optimize the dose and antibiotic spectrum were made with a 100% acceptance rate.

Compared with the preintervention period, overall adherence to the SAB bundle increased from 29.0% to 72.8% (P < .001) during the intervention. The intervention was associated with statistically significant improvements in obtaining an ID consultation (56.5% in preintervention vs 93.4% in intervention group), prescribing guideline concordant definitive therapy (83.9% vs 99.3%), assessing for endocarditis by echocardiography (72.6% vs 95.6%), and, where care was not withdrawn, providing an appropriate antibiotic duration of therapy (87.0% vs 100%) (all P < .001) (Table 2). A multivariate logistic regression analysis for overall bundle adherence, adjusted for age and the statistically significant baseline differences between the preintervention and intervention group (Charlson comorbidity index [including diabetes mellitus] and complicated bacteremia at diagnosis) confirmed that our intervention was associated with improved overall bundle adherence (adjusted odds ratio [OR], 5.63; 95% CI, 2.81–11.26; P < .001) (Table 3).

Table 2.

Adherence to the Evidence-Based Bundle

Bundle Component	Patients, No. (%)		P Value	Proportion Difference (95% CI), %
	Preintervention Group (n = 62)	Intervention Group (n = 137)
Overall bundle adherence	18 (29.0)	99/136 (72.8)	<.001	43.8 (29.11–55.68)a
Infectious diseases consultation	35 (56.5)	128 (93.4)	<.001	37.0 (24.16–49.71)
Repeat blood culture	57 (91.9)	135 (98.5)	.03	6.6 (0.72–16.13)
Empiric treatment concordance	52 (83.9)	122/136 b (89.7)	.24	5.8 (0–17.63)
Empiric antibiotic	57 (91.9)	131/136 b (96.3)	.29	4.4 (2.13–14.80)
Empiric dosing regimen	53 (85.5)	122/136 b (89.7)	.39	4.2 (0–15.79)
Definitive treatment concordance	52 (83.9)	136 (99.3)	<.001	15.4 (7.55–26.49)
Definitive antibiotic	60 (96.8)	137 (100)	.10	3.2 (0–11.02)
Definitive dosing regimen	52 (83.9)	136 (99.3)	.02	15.4 (7.55–26.49)
Echocardiography	45 (72.6)	131 (95.6)	<.001	23.0 (12.34–35.44)
Appropriate duration of therapy	40/46 (87.0)	112/112 (100)	<.001	13.0 (5.36–25.67)
Source control when applicable	13/19 (68.4)	41/55 (74.5)	.60	6.1 (0–30.52)

Abbreviation: CI, confidence interval

aThe number needed to treat (ie, the number of forms distributed to increase bundle adherence by 1 patient) was 3 (95% CI, 2–4).

bOne value missing.

Table 3.

Multivariate Regression Examining the Antimicrobial Stewardship Program Intervention as a Correlate of Overall Staphylococcus aureus Bacteremia Bundle Adherence

Outcome	Bundle Adherence (n = 117)	No Bundle Adherence (n = 81)	Adjusted OR (95% CI)	P Value
Age, mean (SD), y	63.3 (19.8)	66.9 (18.9)	1.01 (.99–1.04)	.30
Charlson comorbidity index, mean (SD)	4.21 (2.9)	5.59 (3.6)	0.86 (.74–1.00)	.04
Complicated bacteremia, No. (%)	78 (66.7)	40 (49.4)	1.30 (.67–2.51)	.44
ASP intervention, No. (%)	99 (84.6)	37 (45.7)	5.63 (2.81–11.26)	<.001

Abbreviations: ASP, antimicrobial stewardship program; CI, confidence interval; OR, odds ratio; SD, standard deviation

With respect to clinical outcomes, the intervention was associated with reductions in 30-day readmission and 30-day mortality rates, but these did not achieve statistical significance (Table 4). A multivariate analysis adjusting for age, Charlson comorbidity index, and complicated bacteremia at diagnosis confirmed lack of mortality benefit (adjusted OR, 0.91; 95% CI, .42–1.93; P = .80). Owing to the 36.9% increase in obtaining an ID consultation, a post hoc analysis was conducted, which demonstrated a decrease in 30-day all-cause mortality rate with ID consultation (14 of 36 [38.9%] vs 34 of 163 [20.9%]; P = .02); however, this was not confirmed by multivariate analysis (adjusted OR, 0.47; 95% CI, .19–1.18; P = .11).

Table 4.

Outcomes in Patients with Staphylococcus aureus Bacteremia

Outcome	Patients, No. (%)a		P Value
	Preintervention Group (n = 62)	Intervention Group (n = 137)
Acute length of stay, median (IQR), d	18.68 (12.80–26.68)	19.22 (11.80–39.27)	.42
Readmission within 30 d	11 (17.7)	12 (8.8)	.07
In-hospital death	17 (27.4)	29 (21.2)	.33
Death within 30 d (all cause)	18 (29.0)	30 (21.9)	.28

Abbreviation: IQR, interquartile range

Data represent no. (%) of patients unless otherwise specified.

The impact of ID consultation on bundle adherence in the preintervention compared with the intervention period is summarized in Table 5. In the preintervention period, ID consultation was associated with improvements in definitive treatment concordance (70.4% vs 94.3%; P = .02) and obtaining an echocardiogram (48.1% vs 91.4%; P < .001). During the intervention period, there were similar rates of adherence to all bundle parameters, regardless of ID consultation.

Table 5.

Bundle Adherence With Infectious Diseases Consultation: Preintervention vs Intervention Groups

Bundle Component by Group	Patients, No. (%)		P Value
	ID Consultation	No ID Consultation
Repeated blood culture
Preintervention	33/35 (94.3)	24/27 (88.9)	.64
Intervention	126/128 (98.4)	9/9 (100)	>.99
Empiric treatment concordance
Preintervention	29/35 (82.9)	23/27 (85.2)	>.99
Intervention	115/127 (90.5)	7/9 (77.8)	.23
Definitive treatment concordance
Preintervention	33/35 (94.3)	19/27 (70.4)	.02
Intervention	127/128 (99.2)	9/9 (100)	>.99
Echocardiography
Preintervention	32/35 (91.4)	13/27 (48.1)	<.001
Intervention	124/128 (96.8)	7/9 (77.8)	.050
Appropriate duration of therapy
Preintervention	34/35 (97.1)	22/27 (81.5)	.08
Intervention	128/128 (100)	9/9 (100)	>.99
Source control when applicable
Preintervention	29/35 (82.9)	27/27 (100)	.03
Intervention	115/128 (89.8)	8/9 (88.9)	>.99

Abbreviation: ID, infectious diseases.

Discussion

This study demonstrates the beneficial impact of developing an evidence-based SAB bundle combined with unsolicited, ASP-driven prospective audit in optimizing SAB management. Significant improvements in rates of ID consultation, prescribing guideline concordant definitive therapy, assessing for endocarditis by echocardiography, and ensuring optimal treatment duration were shown with the intervention. Adherence to all 6 evidence-based recommendations increased from 29.0% to 72.8%, and the number needed to treat (ie, number of standardized forms distributed) to increase overall bundle adherence by 1 additional patient was 3, indicating the efficacy of such a simple intervention.

Several key factors contributed to the effectiveness of the intervention. First, all inpatients with SAB were prospectively reviewed, thereby ensuring universal evaluation and an opportunity to provide optimal treatment recommendations, regardless of physician request. Despite being unsolicited, our recommendations were accepted at high rates, and the program was viewed favorably by the medical teams. Second, the simple, guideline-based feedback form served as a standardized education tool and minimized variability in the feedback provided by the auditing pharmacists resulting in consistent implementation. Finally, PAF is one of the most effective antimicrobial stewardship tools for positively influencing the prescribing of antibiotics, according to numerous studies [26-39] and the Infectious Diseases Society of America guidelines [21]. The use of PAF in our study provided the opportunity to educate prescribers at the point of care on SAB management by directly communicating recommendations to them. Anecdotally, as the intervention period progressed, fewer recommendations were required as prescribers gained familiarity with optimal SAB management.

The objective of this study was to demonstrate a difference in adherence to SAB quality-of-care measures rather than clinical outcomes. Thus, the study was underpowered to detect mortality benefit yet still showed a trend toward decreases in both 30-day readmission and 30-day mortality rates. Notably, the intervention group did consist of a greater proportion of patients with complicated bacteremia at diagnosis (68.6% vs 40.3% in the preintervention group), portending a poorer prognosis regardless of intervention.

Arguably, an essential component of SAB management is ID consultation, with multiple studies demonstrating a mortality benefit [4-16]. Although we demonstrated a trend toward improved mortality rates with ID consultation, this difference was not significant by multivariate analysis, probably because our study was underpowered for this subanalysis. Interestingly, baseline ID consultation rates were lower than expected, presumably because physicians underestimated the pathogenicity of SAB, misjudged the complexity of an SAB presentation, or were unaware of the established mortality benefit associated with ID consultation.

Our intervention resulted in a nearly 40% increase in ID consultation in SAB management compared with baseline (increased from 56.5% to 93.4%; P < .001). This increase may account for the difference noted in the rates of complicated bacteremia at diagnosis, because an ID consultation would probably result in a more thorough evaluation of the patient. In addition to garnering nearly universal ID consultation for SAB, our intervention had the benefit of providing timelier implementation of the evidence-based recommendations. Furthermore, the use of standardized recommendations by the ASP pharmacists during PAF minimized any heterogeneity existing in the practice patterns of individual ID physicians. In those instances in which an ID consultation was not obtained during the intervention period, adherence to the individual bundle components was comparable to that in the group with ID consultation, suggesting that the benefits from the intervention extend beyond securing an ID consultation.

The use of ID physicians to implement an SAB management bundle has been studied previously. In the study by López-Cortés et al [19], ID physicians provided recommendations based on 6 quality-of-care indicators for SAB management. The intervention resulted in improved adherence to the quality-of-care indicators and a reduction in both 14-day and 30-day mortality rates. In a study by Borde et al [20], all patients with SAB were seen by an ID physician. An SAB management bundle was emphasized, and, once again, a significant increase in bundle adherence and reduction in in-hospital mortality rate was demonstrated.

In both studies however, the intervention was performed as an unsolicited ID consultation, which may be less favorable for a number of reasons. First, securing a dedicated ID physician to intervene in all SABs in a timely fashion may not be realistic owing to resource limitations or a lack of in-person ID support. Our intervention optimizes SAB management within the confines of existing resources (ie, preexisting pharmacists and an ID consultation service). Second, there are advantages to placing SAB management under the umbrella of ASP. Nguyen et al [18] studied the impact of an ASP-led SAB care bundle. In addition to improving overall bundle adherence (from 56.1% to 84.1%), they demonstrated higher bundle component compliance with ASP oversight than with strictly ID physician–mediated SAB recommendations (87.2% vs 58.6%, respectively). Third, unsolicited ID physician advice may be viewed as unfavorable for sites with high levels of physician autonomy. Conversely, our intervention used the preexisting relationships between pharmacists and physicians to suggest an ID consultation while providing convincing evidence to ensure that the consultation was requested.

In the study by Nguyen et al [18], pharmacists received real-time, around-the-clock notification of SABs. In our study, ASP pharmacists were available to complete PAF for positive S. aureus blood cultures only during regular business hours, resulting in a delay in communicating recommendations for cultures that became positive outside these hours. Despite this limitation, our study still demonstrated benefit in SAB management and its findings reflect the real-world capabilities of many ASPs.

A retrospective study by Wenzler et al [40] evaluated the effect of an automated, pharmacist-driven intervention in SAB management. Pharmacists received notification of SAB cases through an EMR-mediated scoring system. The EMR also permitted review of investigations and antibiotic treatment, after which pharmacist recommendations were made directly within the EMR. This automated intervention demonstrated improved adherence to SAB quality-of-care measures and a mortality benefit. Although an efficient and effective method of SAB optimization, our process may be more applicable to facilities lacking such a comprehensive EMR and represents an easy-to-implement initiative with immediate benefits.

Owing to our study’s pre-post experimental design, its limitations include potential differences in patient populations and standards of practice over time. We also acknowledge the limitations inherent in an unblinded retrospective record review for the preintervention period but believe that the data points collected were well defined and objective, thereby promoting consistency and minimizing bias.

In conclusion, this study highlights an effective ASP-driven initiative that can significantly improve SAB management to minimize poor patient outcomes. Through unsolicited, standardized PAF by our ASP, we were able to enhance bundle adherence, in particular, increasing the ID consultation rate. Our intervention is unique in that it provides a simple and practical method for optimizing SAB management, and can be readily implemented in resource limited sites by drawing on preexisting resources and by the application of a standardized, evidence-based bundle form.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Click here for additional data file.