FormalPara Key Summary Points

Why carry out this study?

Infectious Diseases (ID) consultation is often unavailable at rural community hospitals.

An expansion of access to ID consultants from community hospitals may lead to improved patient outcomes, particularly through the improved use of antimicrobials.

What was learned from the study?

Many patients in rural community hospitals would have modifications to their treatment regimens if an ID consultant was available.

The availability of ID consultation in community hospitals could enhance patient care, primarily through modifying antimicrobial regimens to improve antimicrobial stewardship.

Introduction

Infectious diseases (ID) consultations have been demonstrated to improve patient outcomes in Staphylococcus aureus bacteremia [1,2,3,4,5,6,7,8,9], Pseudomonas bacteremia [9], candidemia [8, 10, 11], multidrug-resistant organisms [12], enterococcal bacteremia [13], and outpatient parenteral antimicrobial therapy [14, 15]. ID specialists generally practice in hospitals that serve urban population centers that can support robust inpatient practices. Little is known regarding the differences in the approach to treating infections in community hospitals that lack coverage from an ID specialist.

The improvements in patient care offered through ID consultation have led to calls to expand the workforce of ID physicians, particularly in community hospitals where ID physicians infrequently practice [16, 17]. ID consultation services are not available at many community hospitals. The long-term expansion of the ID workforce must be supported by demonstrating the need for ID consultants in rural communities. In the short-term, multiple strategies will be needed to improve the care of patients with infections that do not have access to an ID consultant, including measures such as pharmacy outreach programs, educational programs, and electronic medical record alerts. These efforts would improve the quality of care received in community hospitals. This multifaceted approach may also reduce the need to transfer some patients to tertiary facilities.

Telemedicine can expand the scope of ID specialists to community hospitals, as the role of ID telemedicine has expanded to include other common infections seen in the hospital or home care setting [18]. Telemedicine has also improved antimicrobial stewardship outcomes in community health systems [19, 22].

The utilization of telemedicine-ID was successfully demonstrated at the University of Pittsburgh Medical Center, which developed a service for outlying regional hospitals, thus expanding ID consultative services to 13 community hospitals [23].

Methods

Purpose

This project characterized patients seen in rural community hospitals of the ProMedica Health System that might benefit from access to an ID specialist.

Practice Setting:

The ProMedica Health System includes 12 hospitals throughout northwest Ohio and southeast Michigan. The Toledo Hospital (Toledo, OH) is a 794-bed hospital that serves as the region's tertiary-care center. At the outset of this investigation, only two ProMedica hospitals provided regular access to ID consultative services: the Toledo Hospital and Flower Hospital (Sylvania, OH). Patients admitted to eight rural community hospitals without access to an ID physician were evaluated: ProMedica Bay Park Hospital (Oregon, OH), ProMedica Monroe Regional Hospital (Monroe, Michigan), ProMedica Charles and Virginia Hickman Hospital (Adrian, Michigan), ProMedica Herrick Hospital (Tecumseh, Michigan), ProMedica Defiance Regional Hospital (Defiance, Ohio), ProMedica Fostoria Hospital (Fostoria, Ohio), ProMedica Memorial Hospital (Fremont, Ohio), and ProMedica Coldwater Hospital (Coldwater, Michigan). These hospitals range from 10 to 147 beds and are in areas with rural–urban commuting area codes between one and four (Table 1).

Table 1 Characteristics of included hospitals
Full size table

Patient population

Patients over 18 years of age that received parenteral or oral antimicrobials for at least 3 continuous days and were admitted between 15 May 2019 and 30 November 2019 were included for analysis and evaluated retrospectively. Different antimicrobial courses of more than 3 days were counted as separate instances. Patients requiring transfer to a hospital for ID services (The Toledo Hospital or ProMedica Flower Hospital) were identified. Patients requiring ID consultation within 24 h of transfer were classified as a transfer for ID services. Patients transferred for needed surgical procedures were excluded from the transfer analysis, unless it was specified in the documentation that ID consultation was among the reasons for the patients to be transferred.

Outcomes

The primary outcome was the need for hospital transfer to ProMedica Toledo Hospital or ProMedica Flower Hospital for inpatient ID consultative services. Transfer status was defined as either the stated reason for transfer or as implied by ID consultation within 24 h of transfer. The secondary outcome was the characterization of antimicrobial administration for any indication at the community hospitals.

Data Collection

Approval for the initiative was granted by the ProMedica Health System Institutional Review Board (IRB). Data was extracted from ProMedica’s electronic medical record. The data points collected were the number of patients transferred from community hospitals to tertiary medical centers, diagnoses associated with antimicrobial use, and antimicrobial use measured by days of therapy (DOT). Approval for the initiative was granted by the ProMedica Health System Institutional Review Board (IRB). Full IRB review was waived as this effort was not considered human subjects research. Therefore, a waiver for informed consent was not required from patients. This was performed in accordance with the Helsinki Declaration of 1964 and its later amendments.

Analysis

Two ID physicians evaluated the collected patient data to identify areas for improvement. The chart review consisted of an initial chart abstraction from a trained reviewer, followed by a review from one of two physicians with American Board of Internal Medicine certification in ID. Complete charts were available to the physicians if needed. Patients were given points as follows: 1 point for cases in which only antimicrobial stewardship interventions could be made and 2 points for each instance where potential intervention could have significantly changed the trajectory of the case due to increased risk for treatment failure or complications. These instances included inappropriate outpatient parenteral therapy, prolonged use of unnecessary antimicrobial spectrum antimicrobials, antimicrobial courses that were used significantly too long or too short for the indication, cases where a better antimicrobial was available for the indication, prolonged unnecessary fluoroquinolone use, inappropriate management of bacteremia, use of agents with inappropriate penetration or bioavailability, the use of an agent that did not offer therapy for the organism in question, culture data not being obtained to guide therapy, absence of a search for the source of infection, inappropriate double coverage of antimicrobials, inappropriate dosing of antimicrobials, inappropriate therapy for ampC-producing organisms, prolonged use of linezolid without monitoring, a contraindication to the antimicrobial given, inappropriate management of candidemia, double beta-lactam therapy, and inappropriate management of necrotizing fasciitis. Due to the high morbidity and mortality of S. aureus bacteremia, two points were also assigned to patients with inappropriate in-hospital management of S. aureus bacteremia and inappropriate outpatient management of S. aureus bacteremia, including discharge on oral therapy or no therapy when further treatment was indicated. More information is supplied in the accompanying appendix. Patients were grouped into three categories: 0 points (group 1), 1 point (group 2), and 2 or more points (group 3).

Indications for antimicrobial therapy were taken from the indication noted by the prescribing physician in the electronic medical record upon ordering the antimicrobials. If more than one antimicrobial was prescribed and multiple indications were noted, that patient fell under all the selected indications except for those prescribed antimicrobials for pneumonia. Although patients receiving antimicrobials for pneumonia were classified under more than one indication if necessary, they were not classified under more than one pneumonia classification. Instead, patients were placed into the undifferentiated pneumonia category with those that received antimicrobials for the indication of pneumonia if antimicrobials were ordered for more than one pneumonia indication, such as both community-acquired pneumonia and hospital-acquired pneumonia. Patients with confirmed positive blood cultures were placed in the bacteremia category, and patients with antimicrobials ordered for sepsis were placed in the sepsis category. Those with antimicrobials collected for sepsis and positive blood cultures were placed in both categories.

Compliance with Ethics Guidelines

This investigation was approved by the ProMedica Health System Investigational Review Board (IRB). Full IRB review was waived as this effort was not considered human subjects research. Therefore, a waiver of consent was not required from patients. The initiative was performed in accordance with the Helsinki Declaration of 1964 and its later amendments.

Results

A total of 3734 patients met the inclusion criteria. A total of 28 patients were excluded, including 11 patients seen by a local pediatric ID specialist and 17 whose antimicrobial courses were duplicated elsewhere. A total of 3706 patients were included in the final analysis. The baseline characteristics of the patients can be found in Table 2. Common diagnoses for which patients received antimicrobials included community-acquired pneumonia (20.6%), urinary tract infection (18.2%), skin and soft tissue infection (17.9%), intraabdominal infections (14.2%), and chronic obstructive pulmonary disease (COPD) (12.9%) (Fig. 1). Cases where the evaluating ID physician would have made a change (patients in group 2 or group 3) totaled 2538 patients (68.5%). Opportunities for improved antimicrobial use were found in 77.7% of community-acquired pneumonia, 68.9% of urinary tract infections, 62.2% of skin and soft tissue infection cases, 53.4% of intraabdominal infections, and 83.4% of cases of chronic obstructive pulmonary disease (Table 3). A total of 36 patients required transfer for ID consultation or received ID consultation within 24 h of transfer, comprising 0.01% of cases examined. Reasons for transfer included S. aureus bacteremia (5 cases), other bacteremia, graft infections, endocarditis, and complicated osteomyelitis.

Table 2 Baseline characteristics of included patients
Full size table
Fig. 1
figure 1

Indication for antibiotic therapy at community hospitals in percentage. CAP community-acquired pneumonia, UTI urinary tract infection, SSTI skin and soft tissue infection, intraabdominal intraabdominal infection, COPD chronic obstructive pulmonary disease, pneumonia undifferentiated pneumonia, HAP hospital-acquired pneumonia

Full size image
Table 3 Opportunities for improvement categorization by diagnosis
Full size table

Patients were divided into three categories based on different areas for improvement: 0 points (group 1), 1 point (group 2), and 2 or more points (group 3). A total of 1166 patients fit into group 1 (31.5%), 2184 patients fit into group 2 (58.9%), and 356 patients fit into group 3 (9.6%) (Table 3). Overall, 1166 patients were assigned 1 point, 144 patients were assigned 2 points, 129 patients were assigned 4 points, 51 patients were assigned 6 points, 19 patients were assigned 8 points, 5 patients were assigned 10 points, and 2 patients were assigned 12 points.

Examples of opportunities for improvement in group 2 included treatment of chronic obstructive pulmonary disease (COPD) exacerbation with broad-spectrum antimicrobials, inappropriate broad-spectrum therapy for skin and soft tissue infection (SSTI), treatment of non-fulminant C. difficile colitis with both oral vancomycin and metronidazole, inappropriate use of antimicrobials for non-infectious intra-abdominal indications, and prolonged therapy for both hospital and community-acquired pneumonia. Opportunities for improvement in group 3 included patients discharged on outpatient parenteral antimicrobial therapy (OPAT) without necessary plans for laboratory monitoring and follow-up, prolonged use of an unjustifiably broad antimicrobial spectrum of activity, inappropriate antimicrobial duration, unnecessary empiric antimicrobials, deviation from antimicrobial of choice when there was no contraindication, mismatched application of antimicrobials against the underlying pathogen, and improper management of bacteremia.

Opportunities for improvement in bacteremia management included failure to search for the bacterial source, absence of echocardiography when appropriate, choice of suboptimal therapy, and inappropriate discharge plans. S. aureus bacteremia management was further evaluated, given its known high mortality. Fifteen patients received improper antimicrobials, eight received antimicrobial courses deemed too short, and one received a course that was too long. Two patients were given prolonged courses of linezolid with no laboratory monitoring, and four patients did not receive proper echocardiography. Additionally, the source of the bacteremia was not pursued in four patients, eight patients were discharged on OPAT without recommended labs, four patients received oral therapy at discharge (i.e., trimethoprim-sulfamethoxazole, cephalexin, and doxycycline), three patients received no therapy on discharge, and two patients were noted with infected central venous access catheters that were not exchanged during the hospitalization. Other findings included inappropriate therapy for ampC-beta-lactamase-producing organisms, inappropriate double beta-lactam therapy, prolonged use of linezolid without laboratory monitoring, and empiric treatment of osteomyelitis without cultures being obtained (Fig. 2).

Fig. 2
figure 2

Common themes in patients categorized in group 3 by evaluating physicians. The reasons above are not mutually exclusive, and many patients fell into multiple categories.

Full size image

There were three cases in group 3 in which improved management of antimicrobials could have possibly prevented patient harm: a patient with Candida albicans candidemia with normal renal function that was treated inappropriately with 200 mg of fluconazole daily, a patient with Pseudomonas aeruginosa pneumonia that was inappropriately treated with ertapenem, and the same patient that was treated for methicillin-resistant S. aureus (MRSA) pneumonia, without reculturing or broadening of therapy despite a change in clinical status.

Total antimicrobial days of therapy were tabulated (Fig. 3). There were 22,807 days of antimicrobial therapy, 662 of which were carbapenem therapy. The most common antimicrobials used were ceftriaxone, vancomycin, and piperacillin-tazobactam.

Fig. 3
figure 3

Antibiotic days of therapy per antibiotic prescribed at community hospitals

Full size image

Discussion

Our results highlight the need for improved access to ID physicians. In 2017, the national average density of ID physicians in the US was 1.76 per 100,000 persons. However, ID physicians are geographically skewed toward urban population centers, leaving 2499 US counties (79.5%) without ID services [16]. Although community hospitals regularly treat infections, our data demonstrates the potential added value of a dedicated ID service. Including ID physicians at community hospitals can facilitate opportunities for improved antimicrobial stewardship, enhanced patient care via increased conformity to guideline-based care, increased safety when patients require outpatient parenteral antimicrobial therapy, and decreased length of hospitalization due to quicker recognition of ID problems. ID physicians are trained to work with every medical specialty and subspecialty and improve care in many arenas; therefore, increased access to these physicians would improve care across many medical fields.

The coronavirus disease 2019 (COVID-19) pandemic has further highlighted the need to expand the ID workforce. The pandemic has added stress to community hospitals, with increases in central-line-associated bloodstream infections, ventilator-associated events, and C. difficile infections reported at community hospitals throughout the pandemic [17]. In an environment where antimicrobial utilization is increased, access to ID physicians becomes more critical [24]. Our data demonstrate the potential value of ID involvement in the community setting prepandemic. This potential impact would only increase in pandemic situations.

A total of 36 patients required transfer to our tertiary care centers without the sole need for surgical intervention and received ID consultation during the period examined, leading to a transfer rate to a tertiary care center of only 0.01% of patients for the indication of an ID consultation. Possible explanations for the low utilization of transfer may include the reticence of patients to transfer to tertiary care facilities due to inconvenience and cost, lack of awareness of physicians practicing in community hospitals of the value that ID physicians provide, and an informal reliance on “curbside” phone calls from ID physicians as a substitute for transfer. Increasing awareness in community hospitals regarding the role of ID consultants should be prioritized so that resources such as patient transfer for ID consultation or ID-telemedicine services will be utilized. Improved access via telemedicine could save patients and hospital systems money, while potentially improving patient satisfaction with their medical care. More importantly, access to an ID consultant offers the possibility of improved antibiotic selection, improved management of many infections, and permitted access to an OPAT program for safer intravenous antimicrobial receipt if needed on discharge. These data demonstrate the potential for improvement in antimicrobial management in patients hospitalized in community hospitals. Several strategies could be employed to improve the care of patients with infections in community hospitals. Expanding the ID workforce to increase coverage to more community hospitals must remain a public health priority [16], and the expansion of telemedicine offers one potential opportunity to expand coverage to address this unmet need. Even with an expansion of coverage of ID consultants in community hospital settings, not all patients with an infection require ID consultation. Other modalities might render care improvements for these patients, such as a decision support system in the electronic medical record and pharmacy outreach with an ID physician. Antimicrobial stewardship programs result in lowered costs to patients and the hospital system via cessation of unneeded therapy [25]. Expanding antimicrobial stewardship programs to community hospitals that do not have access to an ID specialist could improve patient management.

The involvement of ID specialists leads to less unnecessary antimicrobial use, reduced cost, and decreased mortality when involved with critically ill patients [26]. Our data demonstrate that there would be significant changes in antimicrobial use for many patients, had the reviewing ID physicians been consulted. Additionally, further necessary interventions would have been recommended in several patients that potentially could have had significant positive implications for patients. Furthermore, including ID physicians at community hospitals would potentially give access to other services that an ID physician provides, such as infection prevention. Involvement of an ID specialist in the care of patients with ID diagnoses is associated with better outcomes, such as fewer days of intensive care stay, lower 30-day mortality rate, lower cost of care, decreased C. difficile rates, and improved susceptibilities of some microorganisms due to improved antimicrobial stewardship. [27]

Limitations of our work include its retrospective nature. Medical records were thoroughly examined, but there may have been indications for antimicrobial therapy that were not documented, which could have altered results. Although the need for ID consultation is strongly implied by obtaining ID consultation within 24 h of transfer, some of these instances may have been due to the need for another consultative service that was unavailable at the initial hospital. Additionally, although our hospitals require new consultations to be seen within 24 h, it is possible that a small number of patients transferred for ID consultation were not seen within this timeframe and may not have been captured in our analysis. Finally, this project was conducted in northwest Ohio and southeastern Michigan, and thus cannot be fully generalized to other areas of the country due to endemic infections in other geographic regions; however, one could argue that this would only increase the need for access to ID physicians.

Conclusion

Community hospitals routinely treat conditions that require prolonged antimicrobial courses. ID consultation offers the opportunity to improve antibiotic utilization. Patients are rarely transferred to tertiary care centers for their ID-related conditions, and most do not gain access to an ID specialist. ID physician-led stewardship programs should expand to include community hospitals. Efforts to expand the ID workforce to include coverage at community hospitals will improve antibiotic utilization.