Patient preferences for treatment of acute bacterial skin and skin structure infections in the emergency department

Skin and skin structure infections (SSSI) are among the most commonly encountered infections in patients presenting to the emergency department (ED) and are responsible for an increasing number of hospital admissions [1‐4]. To identify those most likely to benefit from antibiotic therapy, a subset of SSSI that have lesions ≥75 cm² are classified as acute bacterial skin and skin structure infections (ABSSSI) by the Food and Drug Administration (FDA) [5]. The most frequently identified cause of ABSSSIs in EDs is Staphylococcus aureus, particularly those strains that are methicillin-resistant (MRSA) [1, 4]. Historically, MRSA was rarely identified as a cause of skin infection, but more recent literature after 2000 attributes upwards of 60% of culture positive cases presenting to the ED to MRSA [6, 7]. Importantly, guidelines from the Infectious Diseases Society of America recommend that an ideal agent for the treatment of purulent or severe, non-purulent ABSSSI should include activity against MRSA [8]. Several oral (e.g., sulfamethoxazole-trimethoprim, doxycycline, clindamycin, etc.) and intravenous (e.g., vancomycin, linezolid, daptomycin, ceftaroline, etc.) antibiotics with activity against MRSA are available for consideration. Also included among these are the newer long-acting, intravenous lipoglycopeptides, dalbavancin and oritavancin [9, 10]. The treatment of ABSSSI in the ED typically falls into one of the following care plans: treatment and discharge from the ED, admission to an observation unit for monitoring, or admission to the hospital for management; intravenous antibiotics, a staple of the latter two treatment strategies, are often administered in the ED before a disposition is decided [1, 11]. Finally, when oral antibiotic therapy is a less desirable option (e.g., larger infections, previous failure, stable sepsis, etc), the long-acting, intravenous lipoglycopeptides allow the entire duration of therapy to be administered as a single dose in the ED, facilitating discharge home for follow-up. While each treatment plan requires individual provider assessment for clinical success and feasibility with respect to the patient’s clinical presentation, providers should also consider patient-centric plans that align with patient preferences, when appropriate.

Although ABSSSI do not confer a significant mortality burden, they are responsible for an average hospital stay of 5.0 days, $9895 (US$) in hospital costs, and a substantial economic burden on hospitals, third-party payers, and society due to productivity losses [4, 12]. Given the fiscal challenges facing hospitals in recent years, novel strategies to reduce unnecessary cost burden should be an important consideration in the ED [1, 11]. A study using US discharge data from 520 hospitals and 610,867 ABSSSI patient encounters in 2012 observed that 60% of all admissions were among patients with no life-threatening conditions and low Charlson Comorbidity Index scores (0–1) [13]. In a separate study of adults with ABSSSI presenting to 12 EDs, the need for intravenous antibiotic therapy was the only reason listed for admission in 41.5% of patients [14]. Collectively, these data suggest that there may be opportunities for clinically stable patients presenting to the ED to be treated in the outpatient setting, thereby completing treatment in a cost-saving setting versus hospital admission.

Of interest, the patient’s preference in this decision process is not well described. To our knowledge, there are currently no data on patient treatment preferences in ABSSSI, and only a few studies have attempted to address patient satisfaction and general preferences on receipt of antibiotics in the treatment of traumatic wounds and uncomplicated infections [15, 16].

The primary objective of this study was to understand patient preferences for treatment of their ABSSSI and what characteristics might influence predilections for treatment in the hospital versus at home or with intravenous versus oral antibiotics.

SSSIs are an increasingly common reason for ED visits and hospital admissions in the last 15 years [1‐4]. Given that both severity and associated mortality are low for these infections, several studies make the case that many patients can be discharged home directly from the ED without requiring further management in the hospital [4, 13, 14]. This approach could help reduce the burden of SSSI infections to the hospital as well as the patient. Of note, patient preferences for treatment have not been explored for more complicated skin infections such as ABSSSI, and such preferences may inform the decision process as well as improve patient satisfaction. This multicenter prospective study aimed at identifying patient preferences and associated factors for treatment location (hospital versus home) and antibiotic treatment (oral versus multiple dose intravenous versus single dose intravenous) after presenting to the ED with ABSSSI. To our knowledge, this is the first study of prospectively identified ABSSSI patients in the ED where preference and satisfaction were described for various healthcare treatment settings and antibiotic treatment regimens.

The National Hospital Ambulatory Medical Care Survey (NHAMCS) in 2010 reported that 4.2% of visits to the ED were caused by a SSSI [17]. These estimates were largely in agreement with the 3.4% of visits coded for a skin infection (ICD-9680–686) in our study’s EDs during the enrollment time period (data not shown). After further applying inclusion and exclusion criteria, we estimate that our study population accounts for approximately 6.9% of the eligible patients with ABSSSI. Nonetheless, the included study population appears to represent a more severely ill ABSSSI population then the most common patient presenting to an ED with a SSSI. The study eligibility criteria were designed to enroll a population that was most likely to require antibiotic therapy for their skin infection based on lesion size, yet not exclude those with history of failure or multiple infections, factors which may influence the decision to admit patients to the hospital. Indeed, we observed the majority (73.4%) of participants were admitted to the hospital. Although no reason for admission was collected in this study, observed lesion size was quite large in our population, 31.9% had diabetes, and 63.8% had a previous ABSSSI prior to the enrolling episode in this study, with nearly half presenting to the ED with a relapse or failure of an ongoing infection. This is in contrast to data by Talan and colleagues who observed admission in 15.2% of 619 patients across 12 US EDs who presented with a SSSI [14]. Notably, the lesion size was significantly smaller in that cohort; only 10.9% of patients had a lesion area ≥ 78.5 cm², and only 12% had diabetes. Additionally, their patient population was slightly younger than observed in our study. Consistent with the reduction in patient/infection severity, the need for intravenous antibiotics was the sole reason for admission in 41.5% of patients in the Talan study. We have also previously observed an admission rate equal to 29.7% of all SSSI presentations (excluding necrotizing fasciitis and osteomyelitis) to one of the six EDs that participated herein [18]. In the current study, 26% of participants presented with leukocytosis and 9.6% had fever, rates that are similar to, if not slightly lower, than those for patients enrolled in recent ABSSSI trials with dalbavancin and oritavancin [9, 10]. The median (25th, 75th percentile) length of stay for admitted patients in our study was 4 (2, 7) days, which was similar to reports from the US Healthcare Cost and Utilization Project National Inpatient Sample (5 days in 2011), but shorter than observed from our hospital in our previous study (7.3 days in 2015) [4, 18].

Despite the large lesion size and frequent history of ABSSSI observed, 21.7% of admitted patients had low Charlson Co-morbidity Index scores (i.e., 0 or 1) and presented with no fever or leukocytosis. It has been suggested that patients with limited comorbidities and no acute symptoms suggestive of sepsis may be candidates for treatment in the outpatient setting [10, 13, 19]. Since many of the participants (11 of 15) with limited comorbidities had a history of a prior ABSSSI, it is possible this history in some part spurred the decision to admit. This is in line with a previous study that reported higher admission rates in patients with cellulitis recurrence [20]. Indeed, failure of prior treatment for the same skin infection was significantly associated with hospital admission in the Talan study, but was not listed as a reason for admission provided by physicians [14].

Importantly, we observed that 40% of study participants preferred to be treated at home, which was discordant with the high admission rate. Only 26% of participants in our study preferred to be treated in the hospital. These rates were similar between baseline and follow-up surveys (Fig. 1). According to Marra and colleagues, the majority of patients in an OPAT program preferred to be treated at home, and their preference was independent of their income or actual treatment location [21]. In our study, a higher education, defined by completion of a 4 year college degree or greater, was the only patient factor independently associated with a 82% lower odds of preferring hospitalization; a higher education was also numerically associated with preference for treatment at home, but did not obtain statistical significance (p = 0.055). Higher education levels could suggest a better understanding of health status, the goals of ABSSSI treatment, or could reflect a better home environment to provide self-care. Additionally, a higher education level could be associated with increased employment responsibilities and therefore, less opportunity for missed time at work, which was quite significant in this population (85% of employed participants reported missing work).

Patients in this study were predominantly treated with intravenous antibiotics (i.e., beta-lactams and vancomycin) during both their ED stay and in the hospital. Given Gram-positive bacteria accounted for the majority of ABSSSI, the need for empiric broad spectrum beta-lactams with Gram-negative or anaerobic activity could be debated and is an opportunity for antimicrobial stewardship interventions in the ED [19]. Of note, patients who were discharged directly from the ED received a shorter total duration of antibiotic therapy compared with those discharged from the hospital. Also of interest was the greater frequency of patients discharged from the hospital on OPAT compared with those from the ED, which also suggests increased severity of infection in these patients, which was not otherwise captured on our data collection tool.

Despite differences in actual treatment received versus their preference, these participants were generally satisfied with their care in the ED, at home, in the hospital, and with their antibiotic regimen overall. Median satisfaction scores were 8 out of 10. Additionally, we observed no differences in satisfaction scores when comparing patients discharged directly from the ED compared with those who were admitted. While a median rank of 8 suggests good satisfaction, there is room for improvement. Currently, hospitals receive reimbursement for inpatients from the Center for Medicare and Medicaid Services (CMS) based, in part, on patient satisfaction scores [22]. Previous satisfaction studies focused around antibiotics have generally observed increased satisfaction with more informed patients [23]. However, we found that only few participants reported being involved in the treatment decisions around their ABSSSI. We hypothesize that increased communication with patients with ABSSSI in the ED and offering educated choices based on their preferences would lead to increased satisfaction, but this would require confirmation in a future prospective study.

Noted limitations associated with this study include the high frequency of larger lesions and the high admission rate of this population. These patients may represent a more severe infection than is routinely seen in EDs; therefore, our results may not represent the preferences of patients with more simple SSSIs. That said, 17% of participants did have smaller lesion sizes, and their preferences were not different from those with larger lesions that met the definition of ABSSSI; there was also no association with lesion size and preferences for disposition or route of antibiotic therapy in the univariate and multivariate analyses. Second, the evaluable population was small and only represented about 6.9% of eligible patients presenting with SSSI in these EDs. Third, Spanish only speaking patients were excluded due to the requirement for a follow-up survey to be conducted via telephone by non-Spanish speaking individuals. Larger studies with inclusion of patients of more varying ethnicities are needed to confirm the patient characteristics associated with certain preferences observed herein. Notably, the length of time between the baseline and follow up surveys was long at 30–40 days. This duration may have caused recall bias when responding to the follow up telephone survey, which may explain some differences in responses between both surveys. However, all follow up surveys were delivered by one of two individuals trained to address perceived participant confusion. As a part of survey study design, loss at follow-up occurs. Nonetheless, overall results from the baseline and follow-up survey were similar when compared. Additionally, the option for selecting “No Preference” on certain survey questions may have allowed some participants to opt out on providing their true preference. No clinical outcome data were collected in this study, and such responses (i.e., successful versus poor) could be a factor contributing to changes in patient preference and their overall satisfaction with care. Finally, there were not sufficient patient numbers at any single participating center to analyze the data by site to see if regional differences might affect preferences for treatment location and antibiotic choice.

Our observations suggest that opportunities remain for improving treatment decision processes and satisfaction of patients presenting to the ED with ABSSSI. We identified certain patient characteristics including higher level of education, younger age, and current employment in these patients with more severe ABSSSI that were associated with preferences for treatment outside of the hospital and with route of antibiotic therapy. We conclude that ED providers should engage patients on their treatment preferences and involve them in the decisions around treatment plans. Further studies that offer treatment choices for patients presenting with ABSSSI are needed to determine if such an approach leads to improved outcomes and satisfaction.