Optimization of Antibiotic Selection in the Emergency Department for Adult Skin and Soft Tissue Infections

Abstract

Objectives: Emergency department (ED) order sets that include skin and soft tissue infections (SSTI) stratification and antimicrobial selection criteria may improve clinical outcomes and appropriateness of initial antibiotic selection. The purpose of this study was to optimize and evaluate antimicrobial prescribing for SSTI in the ED by implementing an institution specific Infectious Diseases Society of America (IDSA) guideline directed computerized provider order entry (CPOE) order set. The primary outcome was the rate of appropriate antibiotic prescribing for SSTI in the ED before and after order set revision. Secondary outcomes were length of hospital stay, rate of continuity of antibiotics from the ED to hospital admission orders, and frequency of order set utilization. Methods: This was a single-centered, retrospective, cohort study. The ED SSTI order set revision reflected current IDSA guidelines, institution formulary, and institution antibiogram. Results: A total of 180 patients were included in the study. The rate of appropriate antibiotic prescribing was 74.4% and 78.9% (P = .60) in the pre-revision and post-revision groups, respectively. Length of hospital stay of admitted patients was 4.93 and 4.32 days (P = .61). Rate of antibiotics continued from the ED to admission was 62.1% and 59.4% (P = .99). Order set utilization was 17.8% and 24.4% (P = .36). A subgroup analysis found appropriateness increased with order set use in admitted patients (50% vs 88.2%; P = .0382) and total patients (50% vs 81.8%; P = .037). Conclusion: An order set reflective of current IDSA guidelines and institution specific antibiogram showed a similar rate of appropriate antibiotic selection compared to provider’s clinical judgment. Provider awareness of SSTI management could have been a limitation to the study.

Background

Patients with skin and soft tissue infections (SSTI) frequently present to the emergency department (ED). Annual visits to the ED in the United States (US) for SSTI have significantly increased from 3.55 million in 2007 to 4.21 million in 2010 and can present with a wide spectrum of disease severity, ranging from mild cellulitis to necrotizing infections. ¹ The most common pathogens associated with SSTI are Staphylococcus aureus and Streptococcus species with methicillin-resistant Staphylococcus aureus (MRSA) being a major cause of purulent SSTI in the United States. ² Gram-negative and anaerobic bacteria are more associated with surgical site infections, whereas polymicrobial infections occur where tissue vascular perfusion is compromised. ³

Initial antibiotic selection should provide adequate coverage for common pathogens using targeted initial antibiotic selection to minimize overuse of broad-spectrum antibiotics. Initial antibiotic selection may influence antibiotic prescribing when patients are admitted to the hospital. Furthermore, providing an antibiotic regimen based on current IDSA guidelines and institution-specific antibiogram data may improve appropriate antibiotic selection.

We hypothesized that ED treatment plans that include SSTI stratification and antibiotic selection criteria may improve clinical outcomes and minimize overuse of broad-spectrum antibiotics. The purpose of the study was to optimize and evaluate antibiotic prescribing for SSTI in the ED with implementation of an institution specific IDSA guideline directed computerized physician order entry (CPOE) order set.

Methods

This was a single-centered, retrospective, cohort study conducted in Jacksonville, FL. Patients were included if aged 18 years and older and initially diagnosed with a SSTI. Patients were excluded if diagnosed with diabetic foot infection or osteomyelitis, documented or suspected concomitant infection upon presentation, or repeat visit(s) to our institution’s ED within 72 hours of the same SSTI complaint. Patients identified with SSTI had a primary or secondary diagnosis using the International Classification of Diseases, 9th edition diagnosis codes. The Institutional Review Board at our institution approved the study.

The primary outcome was the rate of appropriate antibiotic prescribing for SSTI in the ED. Secondary outcomes included length of stay for admitted patients, rate of continuity of antibiotics from initial treatment in the ED to hospital admission orders, and frequency of order set utilization for antibiotic selection.

The pre-revision group was collected from January 2013 to December 2013. The order set revision occurred December 23, 2013. The post-revision group was collected from January 2014 to February 2014. Prior to revision, the antibiotic choices could be selected by a commercially available order set or the provider’s clinical on judgment. A current hospital recommended antibiotic therapy handbook existed but was not integrated into the commercially available order set.

The order set revision consisted of implementing the ED SSTI order set to reflect current IDSA guidelines, institution formulary, and institution antibiogram. SSTI were stratified to cellulitis, surgical site infections, necrotizing fasciitis, and animal bites. The Emergency, Pharmacy, and Infectious Disease Departments in addition to the Pharmacy and Therapeutics Committee approved the order set revision. The washout period between the implementation of order set revision and post-revision data collection period occurred from December 23, 2013 to January 1, 2014. During that week, education was provided to the ED staff. Supplemental Appendix shows the ED post-revision order set stratification used in this study. An inpatient order set was created separately and contained a complete dosing regimen and recommended duration of therapy.

Initial antibiotic selection was defined as appropriate following current IDSA guidelines, institution formulary, and institution antibiogram. An immunocompromised medical condition was defined as concomitant diagnosis of human immunodeficiency virus, transplant recipient, hepatitis, or malignancy.

Data collected included age, gender, weight, height, allergies, past medical history, admitting diagnosis, ED and admission antibiotic selection, antibiotic regimens, use of ED order set, length of stay, antibiotics received within 90 days before initial visit and indication, and documented wound description.

It was determined that in order to obtain a power of 80% and an alpha of 5%, a minimum sample size of 90 patients were needed in each group. Compliance rate was calculated at a rate of 70% with a ±10% margin of error. Statistical tests utilized were chi-squared test for categorical data and two-tailed t-test for continuous data.

Results

Out of the 405 patients reviewed, 90 patients met inclusion criteria for the pre-revision group and 90 patients in the post-revision group. Out of the 90 patients in each group, 29 pre-revision and 59 post-revision patients were admitted to the hospital. Patients were excluded due to concomitant infection (13.6%), diabetic foot infection (5.2%), osteomyelitis (4.7%), or arrived within 72 hours of a recent discharge with the same SSTI complaint (4.2%).

The baseline demographics are listed in Table 1. The average patient was male, 47.4 years old, and diagnosed with a non-purulent SSTI located on the lower extremity. The baseline demographics that were statistically significant between pre-revision and post-revision groups were age (42.3 years vs 52.5 years; P < .01) and admitted patients (32.2% vs 65.6%; P < .01). Table 2 reflects the primary and secondary outcomes. The rate of appropriate antibiotics in the pre-revision and post-revisions groups were 74.4% versus 78.9% (P = .60), respectively. Length of hospital stay of admitted patients was 4.93 versus 4.32 days (P = .61). ED antibiotics continued upon admission were 62.2% versus 60% (P = .99). Order set utilization was 17.8% versus 24.4% (P = .36).

Table 1.

Baseline Demographics.

	Pre-revision N (%)	Post-revision N (%)	P-value
Age (years)	42.3	52.5	<.01
Gender: Male %	52/90 (57.8%)	55/90 (61.1%)	.76
Obesity	27/90 (30.0%)	35/90 (38.9%)	.27
Hospital admission	29/90 (32.2%)	59/90 (65.6%)	<.01
Recurrent SSTI	16/90 (18%)	12/90 (13%)	.543
Received antibiotics within 90 days	18/90 (20.0%)	15/90 (16.7%)	.563
Past medical history
Diabetes mellitus	18/90 (20%)	17/90 (19%)	.851
Immunocompromised	4/90 (4.5%)	8/90 (9%)	.232
IV drug use	0/90 (0%)	3/90 (3%)	.246
SSTI type
Non-purulent	57/90 (63%)	57/90 (63%)	.104
Purulent	31/90 (34%)	21/90 (23%)
Necrotizing	1/90 (1%)	5/90 (6%)
Surgical	0/90 (0%)	5/90 (6%)
Bite	2/90 (2%)	3/90 (3%)
Anatomic location
Head and neck	10/90 (11%)	15/90 (17%)	.244
Upper extremity	33/90 (37%)	21/90 (23%)
Lower extremity	37/90 (41%)	42/90 (47%)
Trunk	10/90 (11%)	12/90 (13%)

Note. SSTI = skin and soft tissue infection; IV = intravenous.

Table 2.

Outcomes.

Outcome	Pre-revision N = 90	Post-revision N = 90	P-value
Rate of appropriate antibiotic prescribing for SSTI in the ED	67/90 (74.4%)	71/90 (78.9%)	.60
Length of hospital stay (days)	4.93	4.32	.61
Rate of ED antibiotics continued upon admission	56/90 (62.2%)	54/90 (60%)	.99
Frequency of order set use	16/90 (17.8%)	22/90 (24.4%)	.36

Note. SSTI = skin and soft tissue infection; ED = emergency department.

A subgroup analysis was performed to account for the difference in baseline demographics and the percentage of admitted patients. Table 3 represents this subgroup analysis showing the rate of appropriateness of antibiotic prescribing in conjunction with the order set use stratified by admitted and non-admitted patients. The order set was used 8/16 (50%) in the pre-revision group versus 18/22 (81.8%) (P = .037) in the post-revision group. A comparison of the appropriateness using the order set showed the pre-revision versus post-revision for admitted patients was 6/12 (50%) versus 15/17 (88.2%); P = .0382, respectively.

Table 3.

Subgroup Analysis.

Group		Appropriate prescribing with order set used N (%)	Appropriate prescribing with order set not used N (%)
Pre-revision	Admitted	6/12 (50%)	12/17 (70.5%)
	Not admitted	2/4 (50%)	47/57 (82.4%)
	Total	8/16 (50%)	59/74 (79.7%)
Post-revision	Admitted	15/17 (88.2%)	31/42 (73.8%)
	Not admitted	3/5 (60%)	22/26 (84.6%)
	Total	18/22 (81.8%)	53/68 (77.9%)

Discussion

Our retrospective analysis of 180 patients in an adult population institution demonstrated similar rates of appropriate antibiotic selection between the pre-revision and post-revision groups. This similarity may be attributed to the availability of an antibiotic therapy handbook recommended by the institution, but not integrated into the commercially available order set. Additionally, the use of antibiotics used within 90 days or allergy history could have led providers not to use the order set as alternative agents may have been warranted.

The differences in baseline demographics may have led to the increase in admission rate in the post-revision group. These differences may be attributed to increased age (42.3 vs 52.5 years; P < .01), increased obesity (30% vs 38.9%; P = .27), and increased surgical and necrotizing infections (1% vs 6%; P = .104) between pre-revision and post-revision groups, respectively. Halilovic et al. found independent risk factors associated with clinical failure in hospitalized patients with cellulitis including weight ≥100 kg, body mass index ≥40 kg/m², recent antimicrobial therapy, and inadequate empiric therapy. ⁴ Lipsky et al. assessed risk factors associated with inappropriate initial antibiotic treatment in complicated SSTI in 525 adults and found 22.5% of initial antibiotic treatment inappropriate, which was similar to our study. ³ Also, Lipsky et al. found that inappropriate initial antibiotic treatment led to a longer length of stay, rate of subsequent unscheduled visits, ED visits, and hospital admissions. ⁵ Black and Schrock associated SSTI hospital admissions with age, infection size, intravenous drug use, and recent antibiotics. ⁶ While, SSTI treatment failure was associated with age, recent surgery at infection site, intravenous antibiotics, and hospital admission. ⁶

Antibiotic failure can be multifactorial but subsequent treatment is crucial for SSTI management, whether that includes additional incision and drainage, alternative antibiotics, or longer treatment duration. Hurley et al. evaluated cellulitis in the ambulatory care setting in 292 cases and found avoidable antibiotic exposure occurred in 46% including broad gram-negative activity (4%), combination therapy (12%), and treatment duration ≥10 days (42%). ⁷ Clinical failure was attributed to 14% of avoidable antibiotic exposure. ⁷ Our study facilitated SSTI management providing a guideline driven order set for antibiotic selection criteria that was comparable to provider’s judgment.

This study showed that an SSTI order set can improve initial antibiotic appropriateness in the ED but it is more important that the providers can prescribe appropriately independent of this guidance. The rate of antibiotics continued from the ED to admission was similar in our study between pre-revision versus post-revision groups, 62.1% versus 59.4% (P = .99). This rate of continuity reflects agreement with antibiotic treatment, clinical judgment, and prescribing habits. The lack of agreement could be due to alternative treatment utilized, alternative formulation preference, or increased time prescribed in the order set. Beatty et al. showed 34.4% of the physicians and medical students were familiar with SSTI diagnosis and management as described in the IDSA guidelines. ⁸ Kessler et al. found 15.7% of physicians aware of SSTI MRSA prevalence and 69.7% would use an antibiotic with MRSA coverage. ⁹ Szczypinska et al. demonstrated antibiotic selection was based on strong physician preference of one antibiotic regimen over another. ¹⁰

The order set was utilized similarly in the pre-revision group compared to the post-revision group. However, Table 3 indicates an increased trend of appropriateness with order set use when comparing pre-revision and post-revision groups in admitted patients (50% vs 88.2%; P = .0382) and total patients (50% vs 81.8%; P = .037). When the order set was not used, our providers were able to select appropriate antibiotics in 59/74 (79.7%) in the pre-revision group versus 53/68 (77.9%) indicating awareness of SSTI management.

Our institution obtained 405 total wound cultures in 385 patients from January 2013 through December 2013. Of the 405 total cultures, 214 were specific for SSTI isolates while the rest pertained to surgical sites and ulcers. No growth occurred in 54 (25.2%) of the cultures. From the 160 positive growth cultures, the bacterial etiologies were as follows: MRSA 60 (37.5%), Methicillin-sensitive Staphylococcus aureus 40 (25.0%), and Streptococcus species 18 (11.3%). The remaining 26.2% of cultures consisted of anaerobes, E. coli, polymicrobial, and skin flora. From the positive growth cultures, 118/160 (73.8%) were gram positive pathogens. At our institution, the most frequent SSTI pathogen was MRSA, therefore continued empiric antibiotic selection with anti-MRSA agents should be continued.

Ray et al. reported 648,699 SSTI episodes between 2006 and 2009 of which 6% required hospital admission, 14% a pathogen was identified, and 63% were caused by a cellulitis and abscess. ¹¹ Of the cellulitis and abscess group, 71% had a culture-confirmed pathogen, 45% caused by MRSA, 36% caused by Methicillin-sensitive Staphylococcus aureus, and 9% beta-hemolytic streptococci. Lipsky et al. found that diabetic patients with SSTI had similar culture isolates, including an increasing rate of MRSA in both foot and non-foot infections. ¹²

One study assessed length of stay evaluating initial SSTI antibiotic treatment in patients who were hospitalized at least 2 days with vancomycin, linezolid, daptomycin, ceftaroline, or tigecycline. ¹³ Compared to the other agents, vancomycin had a shorter length of stay but similar length of therapy. ¹³ Szczypinska et al. also found no difference on length of stay between vancomycin, linezolid, or daptomycin in SSTI. ¹⁰ At our institution, vancomycin was used for SSTI treatment without antibiotic restriction, thus supporting continued use among the newer agents.

Antibiograms assist in antibiotic selection; therefore order sets incorporating an institution specific antibiogram may lead to improved antibiotic clinical effectiveness and antimicrobial stewardship. At our institution, the antibiogram was created from a hospital-wide cumulative reporting. Draper et al. compared antibiograms between the ED and hospital-wide and found similar frequencies of pathogens but varying susceptibility profiles. ¹⁴

Limitations of this study include the retrospective study design, difference in baseline demographics, proportion of admitted patients, and duration of washout period between implementation and post-revision collection. This study did not assess appropriateness of prescribing antibiotics after incision and drainage, appropriateness of antibiotic regimens, appropriateness of ordering cultures for purulent infections, or differentiating hospital-acquired versus community-acquired MRSA risk.

Conclusion

An order set reflective of current IDSA guidelines and institution specific antibiogram showed a similar rate of appropriate antibiotic selection compared to provider’s clinical judgment. Provider awareness of SSTI management could have been a limitation to the study. Furthermore, rate of appropriateness increased with order set use suggesting an increased order set use is warranted. Despite the low rate of order set use, the rate of appropriateness was sufficient.