Impact of Antimicrobial Stewardship Program Intervention Acceptance on Hospital Length of Stay

Abstract

Background: Inappropriate antibiotic use is a major public health concern. Excessive exposure to antibiotics results in the proliferation of multidrug-resistant bacteria, increase in potentially avoidable adverse drug reactions, healthcare utilization, and cost. Currently, systematic reviews and controlled trials assessing the effects of antimicrobial stewardship programs (ASP) on hospital length of stay (LOS), mortality, and cost-savings are conflicting. Some studies reported a significant cost-savings driven by shorter hospital LOS while the others found no effect and, in some cases, prolonged LOS. Shortening the time to appropriate therapy and reducing unnecessary days of therapy have been shown to reduce hospital LOS. Objective: The purpose of this study was to evaluate the effects of prescriber acceptance to ASP interventions on hospital LOS. Methods: Between January 2018 and December 2019, 764 charts were retrospectively reviewed for patients who received antimicrobial treatment and in whom an ASP intervention was performed. Patients were allocated into 2 groups: those whose ASP interventions were accepted and those whose were rejected. Provider responses were then documented within 24 hours of being communicated. The primary outcome was hospital LOS. Secondary outcomes included 30-day readmission rates and inpatient antimicrobial duration of therapy (DOT). Results: There were 384 patients with an accepted ASP intervention and 380 with a denied intervention. Baseline characteristics were similar between both groups, except for a difference in the types of intervention performed (P < 0.001). The median hospital LOS for patients in the accepted intervention group was 6.5 days compared to 7 days in the rejected intervention group (P = 0.009). Antimicrobial DOT was also shorter in the accepted intervention group (5 vs 7 days; P < 0.001). There was no difference in 30-day readmission rates (P = 0.98). Conclusion: Prescriber acceptance to ASP interventions decreases hospital LOS and antimicrobial DOT without affecting 30-day readmission rates.

Introduction

Antimicrobial resistance is a growing concern for health care institutions around the world. While the Centers for Disease Control and Prevention (CDC) has observed fewer deaths from antimicrobial resistance since 2013, there is an increase in infections caused by multidrug-resistant organisms. ¹ Inappropriate use of antimicrobial agents, including unnecessary broad-spectrum therapy or inappropriate durations of therapy, additionally contributes to increased antimicrobial resistance, and preventable adverse events.^{2 -4} Optimization of the use of antimicrobial agents is necessary to effectively treat infections, decrease harms caused by unnecessary therapy, and reduce antimicrobial resistance. Since its mandate by the White House executive order in 2014, antimicrobial stewardship programs (ASP) have been shown to improve clinical outcomes, minimize adverse events, and reduce antimicrobial expidenture.^5,6 Despite the reported benefits of ASPs, there has been conflicting evidence regarding their effects on key measures, including hospital length of stay (LOS), readmission rates, and cost-savings. ⁷

Our institution is a 240-bed community hospital, including a 24 bed intensive care unit. Initiated in 2013, our institution’s ASP, led by an infectious diseases (ID) physician and an ASP pharmacist lead, utilizes a pharmacist-driven prospective audit and feedback system. Intervening pharmacists include staff pharmacist, clinical pharmacist, and PGY-1 pharmacy residents with interventions overseen by ID-certified pharmacists. The antimicrobial stewardship key core elements include shortening time to appropriate therapy, optimizing duration of therapy (DOT), antimicrobial streamlining, ensuring optimal dosing, and facilitating appropriate conversion from intravenous (IV) to oral antimicrobial agents. ⁸ An automatic list of patients on antimicrobials for at least 48 hours, or who have a reported positive culture is generated and reviewed daily by de-centralized pharmacists. The pharmacist then documents an ASP progress note documenting current antimicrobial therapy, type of infection, current days of therapy, and clinical status, including white blood cell count (WBC), presence of fever, pertinent diagnostic tests, and source of infection if known. Interventions are based on available culture susceptibilities, diagnosis, and clinical status. Interventions include escalation (inappropriate coverage), de-escalation (streamlining), or discontinuation (no indication for antimicrobial therapy or completion of therapy), and recommendations are proposed to physicians via telephone or in-person. Guideline-directed duration of therapy is communicated to physicians, and once interventions are accepted/declined, intervention outcomes are documented and stop dates entered when accepted. Pharmacists ensure antimicrobial therapies are appropriate with adequate source control and that the patients are clinically stable. In addition to assessing appropriateness of antimicrobial therapy, our institution conducts daily clinical reviews to ensure optimal dosing, including renal dose adjustments and use of appropriate extended infusion beta-lactams, such as piperacillin/tazobactam and meropenem per institutional policy.

In addition, a separate list of patients on restricted antimicrobials, including carbapenems, novel beta-lactam/beta-lactamase inhibitors, IV antivirals, and IV antifungals is generated and reviewed daily for appropriateness. Recommended interventions are then communicated to physicians to assist in antimicrobial therapy decision-making, and responses are documented on the ASP progress note in the patient’s electronic medical record (EMR). If a response is not received from a physician by the end of that pharmacist’s work day, the recommended intervention is documented as rejected.

The purpose of this study is to investigate whether prescriber acceptance to ASP interventions has an effect on hospital LOS.

Methods

This retrospective study received our institution’s internal review board (IRB) approval.

A retrospective chart review of patients admitted from January 2018 to December 2019 was conducted. Patient data were extracted from the patient EMR and all core measure reports were obtained from Crimson Continuum of Care^®, an online source used to assess hospital quality and assessment outcomes. (For more information, refer to https://www.optum.com/business/solutions/provider/data-analytics/physician-engagement.html) Crimson is based on a severity-adjusted methodology using the 3M^™ All Patient Refined Diagnosis Related Groups (APR-DRG) Classification System^®, a national standard for inpatient cases. Many patient attributes (such as patient age, gender, admit and discharge date, and ICD-9 Diagnosis and Procedure Codes) are to assign each inpatient case an APR-DRG, severity level, and risk of mortality level. (For more information, refer to https://apps.3mhis.com/docs/Groupers/All_Patient_Refined_DRG/Methodology_overview_GRP041/grp041_aprdrg_meth_overview.pdf)

Patients were included if they were at least 18 years of age, received at least one dose of antimicrobial therapy and had an ASP intervention performed. Patients were excluded if they received more than 21 days total of antimicrobial therapy as these patients were considered complex and required more than antimicrobial treatment alone. Only the first ASP intervention was evaluated, and patients were then allocated into those who had the intervention accepted or denied. We evaluated the first ASP intervention since it has the most potential impact on time to appropriate therapy, hospital LOS, and inpatient antimicrobial DOT.

The collected baseline demographics included age, gender, types of infection, types of intervention performed, admission severity and mortality risk. Admission severity and mortality risk were categorized via the 3M APR DRG Classification System^®. Types of interventions done included antimicrobial escalation, de-escalation, and discontinuation of therapy. The primary endpoint was hospital LOS. Secondary endpoints included inpatient antimicrobial DOT and 30-day readmission rate for any reason.

An initial mean of hospital LOS for patients who received antimicrobial therapy and in whom an ASP intervention was performed was calculated from a random sample of 50 patients who were admitted to our hospital from January 2018 to December 2019. Based on a mean hospital LOS of 11 days obtained from this sample of patients, 380 patients were needed in each group to provide a power of 80% to detect a difference of 1 day in the primary outcome between the 2 groups. The a priori alpha level of significance was set at P < 0.05. Statistical analyses were performed using SigmaStat 4.0. Categorical variables were compared using Chi-square or Fisher’s exact test as appropriate, whereas continuous data were analyzed using student t-test or Mann-Whitney U test as appropriate.

Results

From January 2018 to December 2019, a total of 2110 ASP interventions were documented with 544 denied and 1566 accepted interventions. Overall intervention acceptance was 74.2% during this period. Of the 2110 patients who had an intervention made, 776 were screened on a 1:1 basis for inclusion and exclusion criteria until the desired number of patients were obtained to meet power. Seven hundred sixty-four patients were included in the final analysis. Of these 764 patients, 384 had an accepted initial ASP intervention, and 380 did not (Figure 1).

Figure 1.

Patient allocation.

Demographic data between the 2 groups were similar, except for more de-escalation interventions in the denied group (49% vs 71%, P = 0.001), and more discontinuations (31% vs 20%, P < 0.001) and escalations (20% vs 9%, P < 0.001) in the accepted group (Table 1). Median age was 67 years, and close to 50% were male. Both groups had similar admission severity, risk of mortality, and source of infection. The most common infections were pneumonia and urinary tract infection, followed by intra-abdominal and skin and soft-tissue infections.

Table 1.

Patient Demographics.

Baseline Characteristics	Accepted intervention group (n = 384)	Rejected intervention group (n = 380)	P value
Median age, years (IQR)	68 (55-82)	65 (53-78)	.22
Male, n (%)	188 (49)	175 (46)	.34
Infection type, n (%)
Pneumonia	115 (30)	102 (27)	.38
Urinary tract	110 (29)	113 (30)	.8
Intra-abdominal	57 (15)	69 (18)	.26
Skin and soft tissue	58 (15)	61 (16)	.79
Osteomyelitis	20 (5)	28 (7)	.28
Line infection	18 (5)	14 (4)	.61
Other	34 (9)	28 (7)	.54
Type of intervention, n (%)
Escalation	78 (20)	33 (9)	<.001
De-escalation	188 (49)	270 (71)	.001
Discontinuation	118 (31)	77 (20)	<.001
Severity type, n (%)			.15
Minor	13 (3)	24 (6)
Moderate	113 (29)	97 (26)
Major	142 (37)	158 (42)
Extreme	114 (30)	98 (26)
Unknown	2 (1)	3 (1)
Mortality risk, n (%)			.34
Minor	56 (15)	59 (16)
Moderate	95 (25)	113 (30)
Major	130 (34)	126 (34)
Extreme	101 (26)	79 (20)
Unknown	2 (1)	3 (1)

There was a significant reduction in the median hospital LOS in the accepted intervention group (6.5 vs 7 days; P = 0.009), and in the median antimicrobial DOT (5 vs 7 days; P < 0.001). There was no difference in all-cause 30-day readmission rate (17% vs 17%, P = 0.34).

Discussion

Our study found a significant reduction in both hospital LOS and antimicrobial DOT in patients with accepted ASP interventions (Table 2).

Table 2.

Primary and Secondary Outcomes.

Outcomes	Accepted intervention group (n = 384)	Rejected intervention group (n = 380)	P value
Median hospital LOS, days (IQR)	6.5 (3.5-9.5)	7 (4.5-9.5)	.009
Median antimicrobial DOT, days (IQR)	5 (3-7)	7 (4.5-9.5)	<.001
All-cause 30-day readmission, n (%)	65 (17)	65 (17)	.34

Although there has been mixed evidence regarding the effect of ASP on LOS, a majority of studies, 85% of 146 studies included in a systematic review reported a significant decrease in hospital LOS after implementation of ASPs. ⁷ The results of our study further strengthen the support for ASPs in improving key clinical outcomes. A previous study conducted using an audit-and-feedback system, similar to our institution, evaluated a pre- and post-implementation ASP. Authors found a significant decrease in median LOS, as well as antimicrobial DOT in non-ICU units, with no difference in other clinical outcomes. ⁸ Much like our results, a study that evaluated difference in accepted and rejected intervention groups found similar results for reduction in hospital LOS, although much larger (10.2 vs 16.6 days). ⁹

Our finding of reduced antimicrobial DOT is consistent with previous literature that found similar significant reductions in antimicrobial DOT with ASP intervention acceptance. In a retrospective study conducted at a Veterans Affairs acute care center, an ASP program which aimed to optimize antimicrobial usage was implemented, and antimicrobial DOT before and after implementation was analyzed. The authors found a significant decrease in in median duration of antimicrobial DOT (11 vs 9 days), as well as a decrease in percentage of patients on antimicrobial therapy (36.8% vs 24.8%). Overall, ASP intervention acceptance rate was 81.6% in this study, which was similar to our study. ¹⁰

Previously reported studies that measured ASP effects on 28- or 30-day all-cause readmission rates have varied findings ranging from reductions or no difference, to increased all-cause readmission rates,^{9 -11} while most studies that measured infection relapse-related readmission found that ASP interventions were associated with a decrease.^12,13 Our findings support previous literature that evaluated the effect of ASP interventions on readmission rates. A study done at a community teaching hospital, who prospectively audited restricted antimicrobials and then communicated interventions to physicians similarly to our institution, assessed cost, mortality, and 30-day readmission after implementation of an ASP program. There was no difference in patient outcomes, including 30-day readmission. ¹⁴ In addition, a study conducted at a tertiary care referral center also found no difference in survival, mean LOS, or 30-day readmission. ⁶

While our study did not include a cost analysis, the reduction in hospital LOS observed could potentially lead to reductions in hospital stay-related costs. Additionally, the significant reduction in unnecessary duration of antimicrobial therapy should also result in decreased antimicrobial cost, minimize potential adverse effects, and decrease the risk of antimicrobial resistance associated with prolonged antimicrobial use as reported by previous studies.^6,9,11,12

Our study is limited by the nature of the retrospective design, the difference in the level of training the intervening pharmacist team member, as well as the difference in communication methods (in-person vs telephone communication). Despite the variability in team members and/or communication methods, our institution was still able to impact relevant ASP outcomes.

This study did not assess other non-infectious related confounding factors which could prolong hospital LOS, including concurrent comorbidities, disposition issues, in-hospital complications, or total antimicrobial DOT if treatment in the outpatient setting was necessary. Additionally, culture data and antimicrobial agents used were not evaluated, however therapies that were recommended were deemed appropriate by the ASP Pharmacist Lead. Lastly, due to the retrospective nature of the study and clinical report limitations, we could not assess the reason for readmission within 30 days or if the patients were re-admitted elsewhere.

Conclusion

This study was able to observe a significant reduction in hospital length of stay and inpatient antimicrobial days of therapy in patients with accepted ASP interventions. There was no difference in the rate of all-cause 30-day readmission. Further investigation is needed to correlate the study findings, and determine associations with cost-savings.