Considerations that are important for monitoring an OPAT program include: correctly identifying an infection and possible causative agents; determining appropriate empiric treatments; determining when to switch to a narrower spectrum treatment, to an oral option, or stopping antibiotic treatment; microbial factors that could affect antibiotic efficacy like infection location or susceptibility; and host factors such as immune status, age, or likelihood of compliance with medication. These factors make for difficult decisions, which are complicated by evolving antimicrobial resistance. For this reason, we predicted that ID specialist involvement resulted in fewer days patients spent in the OPAT program, mitigating burden on the healthcare system and resulting in less time spent at the hospital for patients.
We focused on patients with a diagnosis of cellulitis because it was by far the leading diagnosis in our cohort. Cellulitis is most commonly caused by beta-hemolytic Streptococcus species or Staphylococcus aureus [4], although in most cellulitis cases microbiologic diagnosis is not obtained, so treatment is largely empiric. Oral cephalexin has been shown to be an appropriate treatment method for mild to moderate cases of cellulitis. Moderate to severe cases respond well to IV cefazolin [4] and can be reassessed during treatment and switched to an oral antibiotic when symptoms improve [5]. Since the organisms that cause cellulitis can have varying levels of antibiotic resistance, they may require different antibiotic treatments. However, the duration of IV antibiotics before switching to oral antibiotics will vary depending on the supervising physician. Physicians presented with the same case information in a survey had large variability in the antibiotic treatment they prescribed [6].
Additionally, diagnosing cellulitis can be difficult. Cellulitis typically presents with cutaneous erythema, warmth, tenderness, and edema; leukocytosis and elevated ESR and/or CRP can also be present but are not required for diagnosis and are often not assessed [7,8,9]. The lower extremities are the site of infection in 70–80% of cases [10]. However, the common signs and symptoms of cellulitis are common findings; therefore, many diseases can masquerade as cellulitis. Common examples of diseases misdiagnosed as cellulitis include deep vein thrombosis, contact dermatitis, gouty arthritis, and cutaneous manifestations of peripheral vasculopathy such as venous or arterial insufficiency [11; 10]. A study by David et al. showed that 28% of cellulitis diagnoses made by EPs were determined to not be cellulitis when reassessed by dermatology or ID specialists [12].
Research has shown that ID specialists are very effective at managing OPAT programs. A multicentered ID-supervised OPAT study showed that 94% of patients were effectively treated [13]. However, this study did not compare OPAT programs supervised by ID specialists to those supervised by EPs. Directly comparing outcomes of patients treated by EPs to those treated by ID specialists at the same hospital can quantify this comparative benefit. Further, assessment of ID specialists’ involvement in the VGH OPAT program has not been formally done.
The purpose of this study was to assess the impact of ID specialist involvement in the program by analyzing patient data and comparing cases monitored by ID specialists to those seen by EPs. This was done by comparing patients that received IV antibiotics in the OPAT fast-track program of VGH for lower leg cellulitis while supervised by either an ID specialist or an EP. When comparing the mean number of days patients seen by EPs alone spent in the program with those seen by ID at least once there was no statistically significant difference. However, the layout of the OPAT program at the time of our study complicates analysis of the direct impact of ID supervision on the number of days patients spend in the program. There was often substantial patient supervision overlap as ID doctors only supervised at specific times and would not supervise over weekends. Further, more clinically complicated patients or those that were predicted to have prolonged IV antibiotic requirements could often be selected for early ID supervision, which would have inflated the average number of days in the program following ID involvement. Patients that were determined to have refractory infections after not responding to IV treatment would have been referred to ID physicians after multiple days in the program without clinical response. Therefore, even if the ID physician discharged patients from the program the first day they were assessed by switching them to an oral option or determining that the pathology was not infectious in nature, the mean days in OPAT would still be elevated due to circumstances that are not reflective of ID involvement.
To mitigate some of these confounding variables we calculated how many days patients spent in the OPAT program after their first assessment by an ID specialist. We compared this to the total number of OPAT days for those seen by EPs alone, which showed that after being seen by an ID physician, patients spent a mean of 2 days less in the program compared to the number of days in OPAT for patients seen by EPs alone.
To further mitigate bias the mean number of days in the OPAT program for patients was compared with patients grouped by the day first seen by ID (Table 1). This showed that the mean number of days in the program and the day of initial ID assessment were closely aligned (except for day 1), which implies that on average patients were typically discharged from the program after initially seeing an ID doctor. This finding gives substantial support to our hypothesis that ID supervision decreases the number of days patients spend in the OPAT program. The exception to this is patients that were seen by ID on day one, but this was exceedingly rare as patients are typically seen by an EP the first time they receive IV antibiotics before being assigned to the OPAT program. Therefore, the sample size for OPAT patients that saw ID on day one was only two patients, who were likely complicated patients warranting immediate ID assessment. It is therefore likely that they would have had longer stays in OPAT for full treatment.
As a balancing measure for early discharge, we assessed repeat visits to ED within a month. Overall, being seen by ID while on OPAT was associated with a significant 14.9% reduction in returning to the hospital within one month of discharge compared to those seen by EP patients alone, showing that the lower mean days of treatment of ID supervised OPAT patients was not the result of insufficiently treated infections. There is the possibility that patients would follow up for recurrent infections at a different hospital, but this would presumably be the same for both ID and EP supervised patients.
Contemporaneous comparison of cohorts separated by the date of increased ID involvement in the OPAT program showed that patients spent an average of 0.5 days less in the OPAT program after increased ID supervision. This finding shows that ID involvement in the OPAT program did decrease the mean number of days patients spent in the OPAT program baring any significant reductions in infection severity or antibiotic resistance over time, which is unlikely. Further, when we defined a successful OPAT intervention as being switched to oral antibiotics within 5 days of IV antibiotics, the initial requirement for being referred to OPAT, then being seen after increased ID involvement was significantly associated with success (p = 0.0263).
Secondary findings of our study were that ID physicians had greater variety in the antibiotics they prescribed, the final diagnosis patients received, and the organisms cultured. The larger variety in antibiotic prescription could be due to increased comfort with a wider assortment of antibiotics, which would allow more assurance with narrower spectrum antibiotic treatments. The antibiotic choices reflected more narrow spectrum antibiotic agents to target specific pathogens and less reliance on empiric treatment with cefazolin. Patients seen by EPs alone were not prescribed ertapenem during this period. This is likely because any complex cases that may require ertapenem would have prompted an ID referral.
ID physicians were more likely to not prescribe antibiotics and to discharge people from the program without any continuing oral antibiotics. Typically, patients being discharged from the program would be the only patients that would not receive IV antibiotics. This could be a reflection of when patients are assessed, i.e. if patients on OPAT are seen by a physician before they are given antibiotics or after. ID physicians would routinely do an assessment before any antibiotics were ordered. In comparison, EPs were also responsible for seeing emergency patients and therefore OPAT patients were commonly given IV treatments on arrival to the OPAT program while they waited to be assessed by an EP each day. However, this could also be because ID doctors felt more confident determining cases that would not amenable to further antibiotic treatment, be that because of an alternative diagnosis or a chronic ailment that would continue to have some baseline superficial infection like a diabetic wound.
EP physicians had fewer cultures taken, less imaging done and fewer consults ordered. This could be due to more comfort making clinical acumen decisions based on the increased volume of patients that EPs see, as well as the increased mindfulness of medical costs that is more paramount with increased patient volume. However, it is also possible that these tests were ordered less frequently because ID would typically be involved at this point and the decisions could be left to them to make. Further, it should be noted that the increased cultures ID ordered likely contributed to being able to use a greater variety of antibiotics.
In addition, the higher frequency of consultation of internal medicine by EPs could be an artifact of the admission process in our hospital, since patients referred for admission to the internal medicine Clinical Teaching Units would require an internal medicine consult. It is thus possible that the sickest patients— such as those failing initial outpatient treatment or with positive blood cultures at their first visit–were selected for hospital admission by the EPs in lieu of continued OPAT visits and subsequent ID assessment. In contrast, the increased consultation of vascular and plastic surgery by ID could reflect the more complex nature of diabetic foot infections (typically managed by vascular surgery or plastic surgery) which could be more likely to have been referred for ID involvement and/or to require longer antibiotic courses.