Table 1.
Summary of Included Studies
| Study Citation | Study Design | Intervention Summary | Primary and Key Secondary Outcomes |
|---|---|---|---|
| Yadav et al. 2018 [14] | Single-center, quasi-experimental study of incorporation of institutional EP for duration of therapy into preexisting ASP rounds | Institutional EP for duration of antimicrobial therapy developed and approved by hospital committees. EP reinforced on ASP rounds. Preexisting ASP rounds included prospective audit and feedback, restriction program, and de-escalation rounds. |
Primary outcomes: mean antimicrobial DOTs administered inpatient and prescribed outpatient for patients discharged with ICD-10 codes for UTI, SSTI, PNA, VAP in 12 months before and 12 months after implementation of EP • Change in mean DOTs: UTI, –1.4 (–2.3 to –0.6; P = .001); SSTI, –2.2 (–3.3 to –1.0; P < .001); PNA, –2.0 (–3.2 to –0.9; P = .001); VAP, –9.6 (–16.0 to –3.3; P = .003) Secondary outcomes: total antibiotic exposure (sum of total milligrams of antibiotics administered inpatient plus prescribed outpatient) • Change in antibiotic exposure: UTI, –3718 (–5185 to –2252; P < .001); SSTI, –5404 (–8227 to –2582; P < .001); PNA, –9430 (–12 028 to –6833; P < .001); VAP, –34 246 (–57 507 to –10 986; P = .004) |
| Thom et al. 2019 [15] | Multicenter, quasi-experimental, pre- and postintervention study | Provider-driven ATOs were implemented across 11 units located in 6 hospitals. Providers were prompted to complete paper ATO tool on antibiotic days 3–5 without study or stewardship input. |
No difference between hospital DOT per admission or total DOT per admission before or after controlling for study unit and season • Average hospital DOT 12.7 vs 12.2 and total DOT 18.9 vs 18.2 • Multivariable analysis showed no association between intervention and number of times regimen was modified or discontinued on antibiotic days 3–5 (OR, 1.0; 95% CI, 0.85–1.19) • Multivariable analysis showed that the ATO was inversely associated with receipt of inappropriate antibiotics on antibiotic days 3–5 (OR, 0.58; 95% CI, 0.48–0.69), as was having undergone a surgical procedure (OR, 0.70; 95% CI, 0.54–0.90) |
| Foolad et al. 2018 [16] | Multicenter, quasi-experimental study |
1) Update and dissemination of institution-specific CAP guidelines via pocket cards and hospital intranet sites. 2) Multiple educational sessions to prescribers and pharmacists regarding appropriate management of CAP, focusing on DOT, updates to the institution-specific guidelines, and the stewardship initiative. 3) Targeted prospective audit with feedback and intervention by ID pharmacists Monday–Friday. |
Decrease in median antibiotic DOT • Historical 9 (IQR, 7–10) days vs intervention 6 (IQR, 5–7) days; P < .001 Improvement in guideline-concordant therapy • Historical 5.6 % vs intervention 42%; P < .001 Decrease in median excess antibiotic days • Historical 3 (IQR 2–5) days vs intervention 1 (IQR 0–2) days; P < .001 No significant difference in clinical outcomes 30 days postdischarge, No. (%) • CDI: historical 0 (0) vs intervention 0 (0); P = not reported • Re-presented to emergency center or clinic with pneumonia: historical 20 (6.8) vs intervention 13 (4.4); P = .22 • Readmission with pneumonia: historical 21 (7.1) vs intervention 11 (3.8); P = .075 |
| Musgrove et al. 2018 [17] | Multicenter, single pre- and postintervention, quasi-experimental study | Clinical microbiology laboratory changed wording in reports on non-pathogen-containing respiratory cultures to emphasize no Staphylococcus aureus, MRSA, or Pseudomonas aeruginosa. |
• Mortality: historical 7 (2.3) vs intervention 3 (1); P = .233 Primary outcome • De-escalation: 39% vs 73%; P < .001 Secondary outcomes • Discontinuation of anti-MRSA therapy: 37% vs 71%; P < .001 • Discontinuation of antipseudomonal therapy: 32% vs 70%; P < .001 • Acute kidney injury: 31% vs 14%; P = .003 • In-hospital, all-cause mortality: 30% vs 18%; P = .52 |
| García-Rodríguez et al. 2019 [18] | Single-center, quasi-experimental, pre- and postintervention study | A multidisciplinary antimicrobial stewardship team was implemented with prospective follow-up of meropenem use. An ID physician reviewed the EMR for each case and provided antibiotic treatment recommendations to the prescribers, with adherence to or rejection of the recommendations from the ID physician assessed at 24–48 hours postrecommendation. |
Improved rates in appropriate justification of meropenem use • Pre-intervention (2014) 47.3% vs postintervention (2017) 76.8%; P = .001 • Reduction in meropenem consumption (DDD/100 OBDs) • During 2015–2017, meropenem consumption decreased compared with 2012–2014 (RR, 0.67; 95% CI, 0.58–0.77; P < .001) |
| Kulwicki et al. 2019 [19] | Retrospective, single-center cohort analysis | Addition of an emergency medicine pharmacist into the ED to provide antimicrobial stewardship. Adherence to empiric treatment recommendations for CAP and community-acquired IAIs was examined pre-EMP and post-EMP. A secondary analysis was undertaken to examine adherence to these same guidelines in the early phases of implementation of an ASP compared with the established program. |
Significant difference in total appropriate empiric antibiotic selection with the EMP vs without the EMP • 78% vs 61%; P = .001 Significant difference in CAP treatment with the EMP vs without the EMP • 95% vs 79%; P = .005 Significant difference in community-acquired IAIs treatment with the EMP vs without the EMP • 62% vs 44%; P = .025 Significant difference in guideline-directed antibiotic prescribing in the established ASP period compared with the pre-ASP period • 82.5% vs 60%; P < .001 |
| Sacco et al. 2019 [20] | Single-center, quasi-experimental pre- and postintervention study | Following development of a validated risk stratification algorithm to guide testing and antibiotic use in patients with penicillin allergy. Health care professionals were educated on its use. The algorithm was intended to guide patient assessment and antibiotic selection. Data were assessed pre– and post–educational initiative. |
Antibiotic use • Cephalosporins +121.2%; P = .03 • Penicillins +256%; P = .04 • Vancomycin –67.2%; P = .04 • Fluoroquinolones –33.3%; P = .31 • Carbapenems –81.9%; P = .08 • Aztreonam –73.8%; P = .18 EMR documentation of type of adverse reaction to penicillin in the admission note • Pre 4.8% vs education 64.9%; P < .001 Use of the test-dose procedure • 8/27 patients Occurrence of adverse drug reactions • None Length of hospital stay • Pre 2.33 days vs education 2.07 days |
| Lee et al. 2018 [21] | Retrospective, single-center quasi-experimental cohort analysis | A fluoroquinolone restriction policy was implemented in 2005. Fluoroquinolone susceptibility was analyzed in a pre-implementation period (1998–2004) and a postimplementation period (2006–2016). Five Gram-negative organisms were included in the analysis: Escherichia coli, Klebsiella pneumoniae, Enterobacter cloacae, P. aeruginosa, and Acinetobacter species. |
Fluoroquinolone use decreased from 173 DOT in the pre-implementation period to <60 DOT in the postimplementation period Fluoroquinolone susceptibility increased for: • Acinetobacter species (RR, 1.038; 95% CI, 1.005–1.072) • E. cloacae (RR, 1.028; 95% CI, 1.013–1.044) • P. aeruginosa (RR, 1.013; 95% CI, 1.006–1.020) Susceptibility did not change significantly for K. pneumoniae (RR, 1.002; 95% CI, 0.996–1.008) E. coli susceptibility continued to decline postimplementation (RR, 0.981; 95% CI, 0.975–0.987) |
| Keller et al. 2018 [22] | Single-center, prospective time series analysis | To reduce the ordering of urinalyses and urine cultures in patients without symptoms of a UTI, a series of interventions including the distribution of educational materials and implementation of CDS alerts in the EMR was implemented. CDS alerts were placed on all orders for urinalyses, urine cultures, and for antibiotics commonly used for treating UTIs (nitrofurantoin, trimethoprim-sulfamethoxazole, ciprofloxacin, cefazolin, cephalexin, and ceftriaxone). |
Primary outcome: Urinalysis orders did not significantly decrease • –10.2%; P = .24 Secondary outcome: Orders for urine cultures did significantly decrease • –6.3%; P < .001 Other results • Decrease in simultaneously ordering urinalyses and urine cultures (–5.8%; P < .001) • Decrease in urinalysis orders followed by antibiotic orders within 1–24 hours (–0.56%; P = .021) • Decrease in urine culture results followed by an antibiotic order within 24 hours (–0.24%; P = .036) |
| Lee et al. 2018 [23] | Prospective, multicenter pre/post chart audit | 15-minute education session to clinical staff focusing on the appropriate management of UTI and ASB, complimented by awareness posters and pocket cards summarizing UTI diagnostic criteria. |
Reduction in antibiotic prescriptions for ASB • Pre-intervention 45 of 50 (90%) vs postintervention 22 of 35 (63%); P = .003 Increase in proportion of residents presenting with localizing UTI symptoms • Pre-intervention 21 of 62 (34%) vs postintervention 22 of 50 (44%); P = .273 Reduction in health care costs • 64% reduction for pharmacy • 30% reduction for laboratory |
| Porter et al. 2018 [24] | Retrospective, single-center, before-and-after study | Conventional microbiology communication vs mRDT plus pharmacist-driven reporting protocol for positive blood cultures. |
Significant decrease in time to change in optimal therapy (50 vs 160 minutes; P = .0081) • Significant increase in percent changed to optimal therapy (41.4% vs 15.6%; P = .013) • Nonsignificant change in percent changed to effective therapy (17.2% vs 24.4%; P = .462) • Multivariate regression analysis showed that the intervention group was significantly less likely to have greater time-to-change value and more likely to be changed to optimal therapy (P < .01 for both) |
| Menichetti et al. 2018 [25] | Retrospective cohort comparing those who received ID consult plus intervention vs intervention alone | Restricted use of voriconazole, posaconazole, caspofungin, anidulafungin, micafungin, liposomal amphotericin B, and lipid complex amphotericin B to ID, intensive care, and hematology, plus ID consultation. |
Primary outcomes • In-hospital, 30-day mortality 20% with ID consult vs 37% without; P = .011 Secondary outcomes • Antibiotic consumption (DDD/100 bed-days): increases in fluconazole (3.1 to 4.3), echinocandins (0.22 to 0.35); decreases in voriconazole (0.25 to 0.18), and amphotericin (0.06 to 0.04) • Antibiotic cost: increased by €207 000 during study period |
| Claeys et al. 2018 [26] | Retrospective, single-center, observational study | Validation of a theoretical Verigene GNB treatment algorithm based on institutional antibiogram data, evidence-based management, and ASP practice. |
Significant theoretical decrease in cases receiving appropriate antibiotic therapy vs standard care (88.4% vs 78.1%; P = .014) • Strong level of agreement between reviewers regarding algorithm recommendations (ĸ = .855) • 14.4% appropriate de-escalation and 5.3% appropriate escalation • 4.8% inappropriate de-escalation and 16% unnecessary escalation |
Abbreviations: ASB, asymptomatic bacteriuria; ASP, antimicrobial stewardship program; ATO, antibiotic time-out; CAP, community-acquired pneumonia; CDI, Clostridioides difficile infection; CDS, clinical decision support; CI, confidence interval; DDD, defined daily dose; DOT, days of therapy; ED, emergency department; EMP, emergency medicine pharmacist; EMR, electronic medical record; EP, expected practice; GNB, Gram-negative bacteremia; IAI, intra-abdominal infection; ICD-10, International Classification of Diseases, Tenth Revision; ID, infectious diseases; IQR, interquartile range; mRDT, molecular rapid diagnostic technology; MRSA, methicillin-resistant Staphylococcus aureus; OBD, occupied bed-days; OR, odds ratio; PNA, pneumonia; RR, rate ratio; SSTI, skin and soft tissue infection; UTI, urinary tract infection; VAP, ventilator-associated pneumonia.