Effect of an intervention targeting inappropriate continued empirical parenteral vancomycin use: a quasi-experimental study in a region of high MRSA prevalence

Vancomycin use is a major risk factor for acquisition of vancomycin-resistant Enterococci (VRE) [1], and infectious diseases caused by VRE are associated with increased mortality and length of hospital stay [2]. Also, the use of unnecessary vancomycin exposes patients to the risk of nephrotoxicity [3]. In 1995 the Centers for Disease Control and Prevention (CDC) established guideline for the appropriate use of vancomycin to reduce vancomycin resistance [4]. Since that time, many clinical guidelines recommended that empirically initiated vancomycin should be stopped, even in immunocompromised hosts, such as those with neutropenia, if there is no evidence of beta-lactam-resistant gram-positive infection [5, 6]. Despite these guidelines, 20% to 70% of vancomycin use is inappropriate [7‐11].

In our institution, over the last 12 years, about 60% of S. aureus infections were methicillin-resistant (MRSA), and this proportion has not changed significantly [12], whereas vancomycin consumption has increased more than twofold, from 18 defined daily doses (DDDs) per 1000 patient-days to 40.0 DDDs per 1000 patient-days [13]. Our previous analysis showed that a quarter of total vancomycin use represented inappropriately continued empirical use [14]. We have been executing an antibiotics stewardship program including an educational program, a computer-assisted antibiotic prescribing program, antimicrobial formulary restriction, and a prior-approval program. Our prior approval program is not very restrictive: physicians who are going to prescribe vancomycin have to describe the indication on a vancomycin order form which is modified from the indications for vancomycin use developed by the CDC [4]. Prescribing physicians are recommended, but not obliged to obtain prior approval by an infectious diseases specialist. In March 2015, the Antimicrobial Control Team of Seoul National University Hospital established an intervention to improve the appropriateness of empirical vancomycin use. The intervention consisted of monitoring of appropriateness by a pharmacist and direct discussion between the prescribing physicians and infectious disease specialists when empirical vancomycin was continued inappropriately beyond 96 h. We report the details of the intervention program and its effects on empirical vancomycin use.

We have shown above that an intervention involving direct discussion between prescribing physicians and infectious disease clinicians increased the appropriateness of empirical vancomycin use. The intervention reduced the amount of inappropriate continued empirical vancomycin use from 8.0 to 5.8 DDDs/1000 patient-days, and the total amount of vancomycin prescribed represented by inappropriate continued empirical vancomycin use fell from 21.2% to 18.1%.

There have been several investigations of the effects of various interventions aimed at improving vancomycin use, including pharmacists’ interventions, automatic stop orders, antibiotic order forms, continuing education, computer-assisted antibiotic prescribing programs, antimicrobial formulary restrictions, and prior-approval programs. The effectiveness of pharmacists’ interventions was demonstrated in a previous study [16]; the initial interventions were performed by pharmacists, and if inappropriate vancomycin use continued, a consultation with an infectious disease clinician was offered. As a result, the accord with guidelines for empiric use of vancomycin improved from 47% in the pre-intervention period to 73% during the intervention [16]. Another pharmacists’ intervention consisting of contacting physicians and informing them of inappropriate vancomycin use significantly improved appropriate initiation of vancomycin [17]. These studies demonstrated that direct interventions of pharmacist and infectious disease clinicians with prescribers could be effective. An intervention consisting automatic 72 h stop orders also improved vancomycin prescribing [18]. However Bolon et al. reported that an antibiotic order form intervention did not improve or reduce vancomycin use [19], and the educational programs about vancomycin use failed to reduce inappropriate vancomycin prescribing in a prospective study [20].

An educational program, a computer-assisted antibiotic prescribing program, and a prior approval program have all been introduced in our institution to promote the prudent use of broad-spectrum antibiotics including vancomycin. Nevertheless, much of the vancomycin prescribed has been used inappropriately. In our opinion education and antibiotic order forms alone have little impact on the appropriateness of antibiotic use, and more intensive interventions such as direct discussions and compulsory stop orders would be more helpful for antimicrobial stewardship. Of course, such direct discussions with prescribing physician may not be long lasting, as it increases the demands on the time of infectious disease clinicians. Therefore, with this interventions, it is essential to establish local guidelines and policies for empirical prescribing. In Korea, to promote appropriate use of antibiotics, the Korea Centers for Disease Control and Prevention (KCDC) is carrying out a Policy Research Serving Project with related expert groups such as the Korean Society of Infectious Diseases, Korean Society for Chemotherapy since 2016 [21, 22].

Factors associated with the inappropriate use of vancomycin, such as critical clinical conditions, absence of documented causative organism, and suspected CNS infection, were identified in previous studies [7, 14]. In this study, patients with hematologic malignancies accounted for a significantly higher proportion of the patients in whom empirical vancomycin was continued inappropriately during the intervention period than during the pre-intervention period. This suggests that physicians treating patients with hematologic malignancies tend to be less compliant with intervention instructions.

Our study has several limitations. First, we only focused on reducing the amount of inappropriate continued empirical vancomycin use. Further interventions focused on prophylaxis, specific treatments, and the initial choice of empirical vancomycin would also be helpful in reducing the total amount of vancomycin prescribed. Second, this was a non-randomized, pre/post-intervention study and the lack of random assignment weakens its significance. Third, we did not investigate why many prescribers continued with inappropriate empirical vancomycin use after discussion with infectious diseases clinicians.

Interventions such as direct communication with prescribing physicians and infectious disease clinicians can help reduce the inappropriate continued use of vancomycin.