a b s t r a c t

Introduction: sepsis identification and treatment is a priority for emergency department (ED) providers and payors alike. However, aggressive metrics aimed at improving sepsis care could have unintended consequences for patients who do not have sepsis.

Methods: All ED patient visits for a one month period before and after a quality initiative to increase early antibi- otic use in septic patients were included. Overall broad spectrum (BS) antibiotic use, admission rates, and mor- tality were compared in the 2 time periods. A more detailed chart review was performed on those who received BS antibiotics in the before and after cohorts. Patient were excluded for pregnancy, age < 18, COVID-19 infection, hospice patients, left ED against medical advice, and if antibiotics were given for prophylaxis. In BS antibiotic treated patients, we sought to determine mortality, rates of subsequent multidrug resistant (MDR) or Clostridium Difficile (CDiff) infections and rates of non-infected patients receiving BS antibiotics.

Results: There were 7967 and 7407 ED visits in the pre- and post-implementation periods, respectively. Of those, BS antibiotics were administered in a total of 3.9% pre-implementation and 6.2% post-implementation (p <= 0.00001). Admission was more common in the post-implementation period, but overall mortality was unchanged (0.9% pre-implementation and 0.8% post-implementation, p = 0.41). After exclusions, 654 patients treated with BS antibiotics were included in the secondary analyses. Baseline characteristics were similar be- tween the pre-implementation and post-implementation cohorts. There was no difference in the rate of CDiff in- fection or the proportion of patients receiving BS antibiotics who were not infected, but there was an increase in the post-implementation period in MDR infections after ED BS antibiotics, 0.72% vs. 0.35% of the entire ED cohorts, p = 0.0009.

Conclusions: We found that a QI sepsis initiative was associated with an increase in the proportion of patients who received BS antibiotics in the ED, and a small absolute increase in associated subsequent MDR infections, with no apparent effect on mortality in all ED patients or the subset treated with BS antibiotics. Further research is needed to assess the impact on all patients affected by aggressive sepsis protocols and initiatives, rather than only those with sepsis.

(C) 2023

1. Introduction

Sepsis is a leading cause of morbidity and mortality among hospital- ized patients [1-3]. Early identification and treatment of sepsis may decrease mortality and is strongly advocated by national and interna- tional guidelines [3]. While many guidelines and metrics are

* Corresponding author at: 1701 N. Senate Avenue, Indianapolis, IN 46202, USA Room AG008, Department of emergency medicine, Indiana University Health.

E-mail address: brhunter@iu.edu (B.R. Hunter).

controversial [4-7], early antibiotic administration and sepsis bundle compliance have been associated with improved outcomes for patients with severe infections [3,8-10].

Metrics emphasizing early antibiotic administration may benefit septic patients, but may also have negative effects [11-13]. Broad spec- trum (BS) antibiotics, such as those used for sepsis, can cause unwanted effects, such as Allergic reactions, development of multi-drug resistant (MDR) bacteria, and Clostridium difficile (CDiff) infections [11-13]. Met- rics emphasizing very early administration of these antibiotics may force providers to treat patients more indiscriminately, or before

https://doi.org/10.1016/j.ajem.2023.06.013

0735-6757/(C) 2023

confirmation of a true bacterial infection, leading to treatment of patients who will not benefit but could suffer the potential harms of BS antibiotic administration [11,12,14]. Many studies have tried to elucidate the benefits to septic patients of quality improvement (QI) initiatives designed to increase compliance with sepsis recommenda- tions, bundles, and metrics [3,6,9,10,15], but far fewer have attempted to determine the extent of harm associated with such initiatives.

The objective of this retrospective observational study was to deter- mine whether implementation of a sepsis QI initiative was associated with changes in rates of administration of BS antibiotics in the ED, BS antibiotic use in non-infected patients, ED mortality, or subsequent development of MDR or CDiff infections.

1. Methods
   1. Setting and subjects

This cohort study took place at a level 1 trauma and tertiary referral academic hospital. The annual ED census is approximately 100,000 pa- tients. In 2019, hospital leadership identified early treatment of sepsis as a quality priority. A sepsis quality improvement (QI) process was ini- tiated with a primary goal of increasing the proportion of patients seen in the ED and ultimately diagnosed with sepsis who received broad spectrum antibiotics within 3 h of ED arrival. To meet this goal, several initiatives were taken. First, an “ED sepsis orderset” was created and embedded within the electronic medical record (EMR). The orderset contains commonly ordered labs and radiology tests for patients with possible or suspected sepsis, as well as resuscitation orders (fluid boluses and vasopressors) and antibiotic regimens tailored to specific sources of infection (Ex: pulmonary, urinary, intra-abdominal, un- known). Orderset antibiotic orders were labeled as a “STAT” priority for pharmacy and nursing. A more aggressive screening protocol was implemented and integrated into the EMR. At triage, nurses were instructed to fill out checkboxes for all patients that included Shock Index, age, and risk factors for sepsis. Any 2 checks prompted an imme- diate alert to a clinician. Additionally, a sepsis screening tool based on SIRS criteria was embedded in the ED EMR tracking board. Additionally, providers were sent monthly emails reminding them of the importance of early aggressive sepsis care. Cases of sepsis were identified at hospital discharge and in cases where antibiotics had not been given within 3 h, the ED providers were sent an email alerting them to the “miss” and en- couraging them to review the case to identify any contributors to the delay. Providers had a target of 80% of sepsis patients would be treated with BS antibiotics within 3 h and that the orderset should be utilized whenever there was any suspicion for sepsis. Orderset usage was also tracked, with an expectation that it be utilized in at least 80% of cases ultimately diagnosed with sepsis.

This study compares one month of patients before, and one month of

patients after, the implementation of the above described sepsis QI measures. Patients were identified by an electronic medical record (EMR) pull of all patients registered in the ED during the 2 study pe- riods. All patients receiving BS antibiotics, as defined by the Centers for Medicare and Medicaid Services SEP-1 bundle [16], were identified through an EMR orders search over a one-month period both before and after the quality improvement initiative, which began in July of 2019. We chose May 2019 as the “before” month. The “after” month was May 2021. This was chosen for several reasons. The rollout of the initiative was in steps and took several months. Additionally, it was felt that it would take several months to achieve a steady state of impact. Since ED volumes and the prevalence of some infections can vary seasonally, we chose to use May as the comparator “post-imple- mentation” month. However, in May 2020, the COVID-19 pandemic had severely altered the makeup of ED volumes and patients across the country, including our institution. We therefore chose May 2021 as the post-implementation comparator month.

• 1. Outcomes and definitions

The primary outcome was the proportion of all ED patients seen that received BS antibiotics, with a hypothesis that more patients would be treated with BS antibiotics in the ED after the rollout of the QI initiative. We hypothesized a 30% relative increase in BS antibiotic use in the 2021 period compared to 2019. Estimating a baseline rate of BS antibiotic use in 2019 of 5%, then a sample size of 10,000 ED patients would provide 90% power with an alpha of 0.05 to detect a 30% difference in BS antibi- otic use. Assuming each month of data would provide between 7500 and 8000 patients, we chose to include one month of patients in each cohort. We recognized this would leave the study underpowered for most secondary outcomes, but were limited by the amount of time re- quired for duplicate manual chart reviews. Mortality and admission rates were compared in the two cohorts as well. These outcomes were pulled directly from the EMR and it’s associated “cube” data.

We then performed a more detailed manual chart review of the patients who received BS antibiotics in the 2 time periods. Patients were excluded for: 1. Age < 18 years; 2. Pregnancy; 3. Received antibi- otics for prophylaxis (Ex. Open fractures, animal bites, Cirrhotic patients with GI bleed); 4. Documented COVID-19 infection; 5. Left ED against medical advice; 6. Hospice patient. Data extracted included age, vital signs in the first 3 h after ED arrival, admission status, duration of anti- biotic therapy in the hospital, development of CDiff infection within 2 months of ED visit [17,18], cultures showing multi-drug resistant (MDR) bacteria within 3 months of ED visit, and in-hospital mortality. These data were extracted manually using individual patient visit identification numbers provided by the EMR orders search and entered into an excel database. All authors participated in the data extraction. For CDiff and MDR outcomes, each patient was searched on the Indiana Health Information Exchange, an EMR system that encompasses results from all of the major healthcare systems in the state. MDR was defined as resistance to 3 or more different antibiotic drug classes, any extended-spectrum beta-lactamase, carbapenem-resistant enterobacteracea, methicillin-resistant Staphylococcus aureus, or vanco- mycin resistant enterococci. Antibiotic duration was classified as either

<=48 h, 48 h to 7 days, or >7 days. For all patients at least 2 authors de-

termined whether the patient was likely infected or not infected, based on strictly defined criteria (Appendix). In cases of disagreement, a third author was the tie breaker. At least one author was blind to the study hypothesis for all cases since some subjectivity was inevitable in the determination of “likely infected” vs “not infected”. The secondary endpoints were 1. Development of MDR cultures within 3 months of the initial ED visit; 2. CDiff infection within 2 months of initial ED visit; 3. Proportion of patients receiving ED BS antibiotics who did not have a bacterial infection.

• 1. Data analysis

Descriptive statistics were used to characterize baseline characteris- tics and the Kolmogorov-Smirnov test was utilized to test normality. Continuous data were compared using chi-square test and dichotomous data were compared utilizing the Mann-Whitney U test or paired t-test, as appropriate. A p-value of 0.05 was used to determine significance. For likelihood of MDR or CDiff infection, we also report Relative risk with 95% confidence intervals (CIs). For the data extraction of “infected vs. not infected” agreement between the first 2 reviewers was measured and is reported as a Kappa statistic. All analyses were conducted using Minitab statistical software v20.3.

1. Results

A total of 7967 and 7407 patients were seen in the ED during May 2019 and May 2021, respectively. Significantly more patients received BS antibiotics in the post-implementation period compared to pre-

Table 1

Outcomes for all ED patients.

Time period	May 2019	May 2021	P value
total ED visits	7967	7407
BS Antibiotics (%)	307 (3.85%)	458 (6.18%)	<0.00001
Admissions (%)	1959 (24.6%)	2077 (28.0%)	<0.001
Mortality (%)	71 (0.9%)	57 (0.8%)	0.41

implementation, 6.18% vs. 3.85%, p <=0.0001. Mortality was similar in the 2 time periods, 0.9% in 2019 and 0.8% in 2021, p = 0.41, but hospital ad- mission was more common in 2021 at 28.0% vs 24.6% in 2019

(p <=0.0001). See Table 1.

There were a total of 765 patients who received BS antibiotics in the 2 time periods, of which 111 were excluded, most commonly because they received Prophylactic antibiotics (Fig. 1). For the secondary analy- ses, Table 2 shows baseline characteristics of the patients who received BS antibiotics in the 2 time periods. The cohorts were nearly identical in age, vital signs, initial white blood counts, and history of MDR bacteria prior to the ED visit.

Secondary outcomes among the antibiotic treated cohorts are displayed in Table 3. There was no statistically significant difference in the proportion of antibiotic treated patients who were deemed “not infected” after chart review. Agreement between reviewers on this out- come was substantial (K = 0.66). There was also no statistically signifi- cant difference in the proportion of patients with longer or shorter antibiotic treatment duration or in mortality among antibiotic treated patients. There were 53 ED antibiotic-treated patients that developed subsequent MDR infections in 2021 compared to 28 in 2019. This differ- ence was not statistically significant as a percentage of BS antibiotic treated patients (13.6% vs 10.6%, p = 0.25), but was significant as a per- centage of all ED patients (0.72% vs 0.35%, p <= 0.001; Relative risk 2.0; 95% CI = 1.3 to 3.2). CDiff infections were also more frequent in 2021 (6 vs 2) but the difference was not statistically significant as a percentage of ei- ther BS antibiotic treated patients or the full ED populations (Table 3).

1. Discussion

In this cohort study evaluating ED patients before and after the implementation of a QI initiative to improve early BS antibiotic admin- istration in septic patients, we found that more patients in the post im- plementation cohort received BS antibiotics, and that this was associated with an increase in subsequent MDR infections with no change in hospital mortality among ED patients. Most sepsis literature is focused on the effect of interventions on patients who actually meet some definition of sepsis. Similarly, our initiative did improve antibiotic metrics in patients diagnosed with sepsis. time to antibiotics in these patients decreased from 127 to 98 min, and the percentage of sepsis pa- tients given BS antibiotics within 3 h of arrival increased from 63% to 79%. However, in this study, we sought to examine overall effects of ag- gressive antibiotic administration among all ED patients, since not only those with sepsis are affected by such initiatives.

Since the early 2000s, when early goal directed therapy (EGDT) first showed great promise [19], there has been exceptional interest in defin- ing and achieving sepsis core metrics aimed at decreasing morbidity and mortality from this common life-threatening condition. Although strictly followed EGDT has since fallen out of favor [20], many of the underlying tenets still live on in current metrics [3]. “Bundles” of sepsis care have been associated with improved metrics, but data to support patient important benefits is low quality and inconsistent [5,6,8- 10,15]. No sepsis metric is more widely accepted than the need for antibiotics. Although there is little doubt about the importance of antibi- otic administration in patients with sepsis, there remains considerable uncertainty around the importance of “early” administration, usually defined as within either 1 h or 3 h of recognition [4,7,21].

While it is reasonable to assume that early aggressive treatment, in- cluding BS antibiotic administration, of severely ill sepsis patients may benefit those patients, it is more difficult to quantify the associated harms to non-septic patients treated as sepsis because of the pressure to intervene before a diagnosis can be confirmed. Most published sepsis literature focuses on metrics and outcomes only in patients ultimately

Fig. 1. Patient identification.

Table 2

Baseline characteristics of patients receiving broad spectrum antibiotics in the emergency department before and after the sepsis quality improvement initiative.

Baseline Characteristics	May 2019	May 2021	p-value
	(n = 265)	(n = 389)
Age, years (median)	57 (41, 69)	56 (40, 69)	0.72
Male (%)	116 (43.8)	179 (46.0)	0.57
Hospital length of stay, days (IQR)	4 (1, 8)	4 (2, 8)	0.38
ICD 10 sepsis diagnosis code associated with encounter, n (%)	42 (18.1)	65 (16.7)	0.77
Admission Rate, n (%)
Any location:	196 (74.0)	312 (80.2)	0.06
ICU Only:	54 (20.4)	77 (19.8)	0.11
History of multi-drug resistant infection in the previous year, n (%)	39 (14.7)	54 (13.9)	0.56
Highest in the first three hours White blood cell count, x 109 cells/L	10.8 (7.4, 15.8)	10.5 (6.8, 14.7)	0.30
Temperature, ?C	37 (36.7, 37.6)	36.9 (36.6, 37.3)	0.002
Heart rate, beats per minute	102 (85, 117)	102 (88, 119)	0.53
Respiration rate, breaths per minute	19 (18, 23)	19 (17, 24)	0.90
Lowest systolic blood pressure in the first three hours after ED arrival; mmHg	124 (106, 141)	123 (105, 143)	0.85
Vasopressor Ordered in the ED	15 (5.7%)	23 (5.9%)	0.00

% refers to percent of treated patients in that cohort, not to percent of all ED patients.

confirmed to have sepsis. We attempted to quantify effects on an entire ED population after the implementation of our system’s QI initiative. Our findings that a higher proportion of ED patients received BS antibiotics after the initiative are unsurprising and may represent bene- fit in more septic patients receiving early treatment or harm in more uninfected or non-septic patients receiving unnecessary BS antibiotics. Contrary to our hypothesis, we did not find a statistically significant in- crease in the proportion of treated patients to be not infected in the post implementation period, although the proportion was numerically higher, 27.5% vs 24.1%. These overall results are similar to previous find- ings [12] and suggest there was not a large increase in inappropriately treated patients after the initiative.

Subsequent MDR infections are more common among patients treated with BS antibiotics [22] and our findings of increased MDR cultures among ED patients in the post implementation period suggests that there were quantifiable harms associated with the increased antibiotic use. Of note, we did not find an increase in MDR or CDiff infection among BS treated pa- tients in 2021 compared to 2019. This is unsurprising, as we did not hypothesize that BS antibiotic administration would have more complica- tions per patient treated in the post-implementation period. Rather, we hypothesized that there were more MDR infections associated with ED BS antibiotic use out of all ED patients because a higher proportion of all ED patients received BS antibiotics in 2021 compared to 2019. Previous lit- erature has been contradictory on the impact of sepsis protocols on the in- cidence of CDiff infections [11,14]. We found a very low Incidence of CDIff overall, with no statistically significant increase in CDiff after implementa- tion of the sepsis initiative.

There are several limitations to our study. Observational before-and- after study designs are associated with a high risk of bias [23]. However, contrary to most such studies, we were more interested in identifying

increased harms in the interventional period, which may limit the risk of bias to some degree. There were a relatively small number of patients included in the BS antibiotic-treated groups. The study was powered to detect a difference in the primary outcome, understanding this would leave the study underpowered for most secondary outcomes. This was by design as we did not have a way to reliably extract many data points, such as the presence of MDR infections or whether or not the patient was infected, without time consuming manual chart review that included multiple EMR searches. Importantly, we were not able to de- termine the overall rate of MDR or CDIff in all ED patients in the 2 time periods because this information would have required manual chart extraction of approximately 15,000 patients. However, this means the increase in MDR was only among patients whose complica- tions could potentially be attributable to ED initiation of BS antibiotics. The 2-year difference in between cohorts introduces the possibility of temporal confounding in our MDR and CDiff results. We were able to obtain our institution’s overall number of extended-spectrum beta- lactamase, methicillin-resistant Staphylococcus aureus, and vancomycin resistant enterococci cultures in 2019 and 2021. The total number of such isolates in 2019 was 5715 and in 2021 was 4137, suggesting that the raw number of MDR isolates at our institution (not ED specific) was not increasing during the time separating the pre- and post- implementation periods. Chart review to determine the presence or ab- sence of a true infection is not without subjectivity. We attempted to minimize bias by creating strict criteria for infection vs. not, having mul- tiple researchers determine this outcome for all patients, and blinding some (at least one for each patient) to the study hypothesis. Agreement for this outcome was substantial. Mortality data was based on hospital mortality. Thus, if a patient was sent home or transferred (unusual for this tertiary referral center) and died shortly thereafter, this would not

Table 3

Secondary outcomes among patients receiving broad spectrum antibiotics in the emergency department before and after the sepsis quality improvement initiative.

Secondary Outcomes	May 2019 (n = 265)	May 2021 (n = 389)	P-value
Duration of antibiotic therapy, n (%) <48 h	109 (41%)	150 (38.5%)	0.46
48 h – 7 days	106 (40%)	150 (38.5%)
>7 days	50 (19%)	89 (23%)
Incidence of clostridium difficile within two months of ED visit, n (% of BS antibiotic treated patients)	2 (0.75%)	6 (1.5%)	0.40
Incidence of clostridium difficile within two months of ED visit, n (% of total ED population)	2 (0.03%)	6 (0.08%)	0.07
Multi-drug resistant cultures three months post-ED visit, n (% of BS antibiotic treated patients)	28 (10.6%)	53 (13.6%)	0.25
Multi-drug resistant cultures three months post-ED visit, n (% of total ED population)	28 (0.35%)	53 (0.72%)	0.0009
Not infected, n (% of BS antibiotic treated patients)	64 (24.1%)	107 (27.5%)	0.34
In-hospital mortality, n (% of BS antibiotic treated patients)	14 (5.3)	13 (3.3)	0.22

ED = Emergency Department; BS = Broad Spectrum.

have been captured. Culture data relied on searches of multiple EMR da- tabases, but some patients could have been seen and diagnosed with MDR or CDiff infections outside the state, which would not have been captured. We chose to evaluate one month time periods separated by 2 years, which could introduce confounders in practice changes outside of only the sepsis initiative. This was in order to minimize the impact of COVID-19 on our post-implementation period while minimizing sea- sonal bias by using the same month of the year in each time period. Lastly, this study was performed at a single academic institution and may not be generalizable to other hospitals or settings.

1. Conclusions

We found that a QI sepsis initiative was associated with an increase in the proportion of patients who received BS antibiotics in the ED, and a small absolute increase in associated subsequent MDR infections, with no apparent effect on mortality in all ED patients or the subset treated with BS antibiotics. Further research is needed to assess the impact on all patients affected by aggressive sepsis protocols and initiatives, rather than only those with sepsis.

Support

This work was not supported by any grants or other financial support.

CRediT authorship contribution statement

Tara Flack: Writing – review & editing, Project administration, Methodology, Formal analysis, Data curation, Conceptualization. Derrick M. Oaxaca: Writing – review & editing, Project administration, Investigation, Data curation, Conceptualization. Chris M. Olson: Writing – review & editing, Project administration, Data curation, Conceptuali- zation. Carl Pafford: Writing – review & editing, Investigation, Data curation. Christian C. Strachan: Writing – review & editing, Investiga- tion, Data curation. Daniel W. Epperson: Writing – review & editing, In- vestigation, Data curation. Jessica Reyes: Writing – review & editing, Investigation, Data curation. Demilade Akinrotimi: Writing – review & editing, Investigation, Data curation. Luke Ho: Writing – review & editing, Investigation, Data curation. Benton R. Hunter: Writing – re- view & editing, Writing – original draft, Supervision, Project administra- tion, Methodology, Formal analysis, Data curation, Conceptualization.

Declaration of Competing Interest

None.

Appendix A. Definitions of infected vs not infected

POSITIVE for INFECTION

positive blood cultures (2 or more out of 4) and treated for bacteremia for at least 5 days

Positive Urine culture with completed treatment (minimum 3 days) Documented and diagnosed soft tissue or skin infection (cellulitis, skin abscess) Pneumonia on imaging and treated for pneumonia (minimum 5 days)

Any culture positive or gram stain positive CSF CT or MRI imaging read by radiologist as infectious in nature [any abscess, cholecystitis,

bowel perforation, appendicitis, diverticulitis, colitis (if treated with abx for at least 5 days), pyelonephritis]

PROBABLY vs probably NOT infected

Underwent full course of abx but no positive cultures or definitive infectious source identified – differentiate based on Clinical Notes, ID notes to take precedence over hospitalist or other notes

Pneumonia – will likely be cases of CXR reads like “infiltrate vs atelectasis” – if both CT and CXR done – CT results to be ruled as “truth”- if no pneumonia identified by radiology but patient treated for pneumonia by admitting, this should usually be “probably not”, especially if CT imaging negative.

Urine with positive culture and given full treatment but clinical picture not c/w urinary infection – count as “probably” as no objective way to say otherwise, although it may have been asymptomatic bacteruria

NEGATIVE for infection

Patients who did fine with an abbreviated course of abx and clinical notes stating not thought to be infected

Cultures all negative (1/4 blood cultures thought to be contaminant and not treated as bacteremia counts as negative cultures)

References

1. Rhee C, Dantes R, Epstein L, et al. Incidence and trends of Sepsis in US hospitals using clinical vs Claims data, 2009-2014. Jama. Oct 3 2017;318(13):1241-9. https://doi. org/10.1001/jama.2017.13836.
2. Fleischmann C, Scherag A, Adhikari NK, et al. Assessment of global incidence and mor- tality of hospital-treated sepsis. current estimates and limitations. Am J Respir Crit Care Med. Feb 1 2016;193(3):259-72. https://doi.org/10.1164/rccm.201504-0781OC.
3. Evans L, Rhodes A, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. Nov 2021;47(11):1181-247. https://doi.org/10.1007/s00134-021-06506-y.
4. Sterling SA, Miller WR, Pryor J, Puskarich MA, Jones AE. The impact of timing of an- tibiotics on outcomes in severe Sepsis and septic shock: a systematic review and Meta-analysis. Crit Care Med. Sep 2015;43(9):1907-15. https://doi.org/10.1097/ ccm.0000000000001142.
5. Wang J, Strich JR, Applefeld WN, et al. Driving blind: instituting SEP-1 without high quality outcomes data. J Thorac Dis. Feb 2020;12(Suppl. 1):S22-36. https://doi.org/ 10.21037/jtd.2019.12.100.
6. Pepper DJ, Jaswal D, Sun J, Welsh J, Natanson C, Eichacker PQ. Evidence underpin- ning the centers for medicare & medicaid Services’ severe sepsis and septic shock management bundle (SEP-1): a systematic review. Ann Intern Med. Apr 17 2018; 168(8):558-68. https://doi.org/10.7326/m17-2947.
7. Rothrock SG, Cassidy DD, Barneck M, et al. Outcome of immediate versus early anti- biotics in severe Sepsis and septic shock: a systematic review and Meta-analysis. Ann Emerg Med. Oct 2020;76(4):427-41. https://doi.org/10.1016/j.annemergmed. 2020.04.042.
8. Damiani E, Donati A, Serafini G, et al. Effect of performance improvement programs on compliance with Sepsis Bundles and mortality: a systematic review and meta- analysis of observational studies. PloS One. 2015;10(5):e0125827. https://doi.org/ 10.1371/journal.pone.0125827.
9. Levy MM, Rhodes A, Phillips GS, et al. Surviving Sepsis campaign: association be- tween Performance metrics and outcomes in a 7.5-year study. Crit Care Med. Jan 2015;43(1):3-12. https://doi.org/10.1097/ccm.0000000000000723.
10. McColl T, Gatien M, Calder L, et al. Implementation of an emergency department Sepsis bundle and system redesign: a process improvement initiative. Cjem. Mar 2017;19(2):112-21. https://doi.org/10.1017/cem.2016.351.
11. Hiensch R, Poeran J, Saunders-Hao P, et al. Impact of an electronic sepsis initiative on antibiotic use and health care facility-onset Clostridium difficile infection rates. Am J In- fect Control. Oct 1 2017;45(10):1091-100. https://doi.org/10.1016/j.ajic.2017.04.005.
12. Minderhoud TC, Spruyt C, Huisman S, Oskam E, Schuit SCE, Levin MD. Microbiolog- ical outcomes and antibiotic overuse in emergency department patients with suspected sepsis. Neth J Med. Jun 2017;75(5):196-203.
13. Curran J, Lo J, Leung V, et al. Estimating daily antibiotic harms: an umbrella review with individual study meta-analysis. Clin Microbiol Infect. Apr 2022;28(4):479-90. https://doi.org/10.1016/j.cmi.2021.10.022.
14. LaFave J, Levy D, Gekle R, Bramante R. Incidence of Clostridium difficile infection after Sepsis protocol antibiotics. West J Emerg Med. Oct 24 2019;20(6):977-81. https://doi.org/10.5811/westjem.2019.10.42070.
15. Pepper DJ, Sun J, Cui X, Welsh J, Natanson C, Eichacker PQ. Antibiotic- and fluid- focused bundles potentially improve sepsis management, but high-quality evidence is lacking for the specificity required in the centers for Medicare and Medicaid Ser- vice’s Sepsis bundle (SEP-1). Crit Care Med. Oct 2019;47(10):1290-300. https://doi. org/10.1097/ccm.0000000000003892.
16. Septimus EJ, Coopersmith CM, Whittle J, Hale CP, Fishman NO, Kim TJ. Sepsis Na- tional Hospital Inpatient Quality Measure (SEP-1): Multistakeholder work group recommendations for appropriate antibiotics for the treatment of sepsis. Clin Infect Dis. Oct 16 2017;65(9):1565-9. https://doi.org/10.1093/cid/cix603.
17. Brown KA, Fisman DN, Moineddin R, Daneman N. The magnitude and duration of Clos- tridium difficile Infection risk associated with antibiotic therapy: a hospital cohort study. PloS One. 2014;9(8):e105454. https://doi.org/10.1371/journal.pone.0105454.
18. Hensgens MP, Goorhuis A, Dekkers OM, Kuijper EJ. Time interval of increased risk for Clostridium difficile infection after exposure to antibiotics. J Antimicrob Chemother. Mar 2012;67(3):742-8. https://doi.org/10.1093/jac/dkr508.
19. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. Nov 8 2001;345(19):1368-77. https:// doi.org/10.1056/NEJMoa010307.
20. Angus DC, Barnato AE, Bell D, et al. A systematic review and meta-analysis of early goal-directed therapy for septic shock: the ARISE, ProCESS and ProMISe investiga- tors. Intensive Care Med. Sep 2015;41(9):1549-60. https://doi.org/10.1007/ s00134-015-3822-1.
21. Xantus G, Allen P, Norman S, Kanizsai P. Antibiotics administered within 1 hour to adult emergency department patients screened positive for sepsis: a systematic

review. Eur J Emerg Med. Aug 2020;27(4):260-7. https://doi.org/10.1097/mej. 0000000000000654.

1. Yoshida H, Motohashi T, De Bus L, et al. Use of broad-spectrum antimicrobials for more than 72 h and the detection of multidrug-resistant bacteria in Japanese inten- sive care units: a multicenter retrospective cohort study. Antimicrob Resist Infect Control. Sep 29 2022;11(1):119. https://doi.org/10.1186/s13756-022-01146-3.
2. Ho AMH, Phelan R, Mizubuti GB, et al. Bias in before-after studies: narrative over- view for anesthesiologists. Anesth Analg. May 2018;126(5):1755-62. https://doi. org/10.1213/ane.0000000000002705.