ABSTRACT
We evaluated the use of antimicrobials expressed as defined daily doses (DDDs) per 1,000 patient days and days of therapy (DOT) per 100 occupied bed-days in a intensive care unit (ICU) of a general hospital in Barcelona, Spain, before and after implementation of an antimicrobial stewardship (AMS) program (2007 to 2010 versus 2011 to 2015). The quarterly costs of antimicrobials used in the ICU and its weight in the overall hospital costs of antimicrobials were calculated. The effect of the applied AMS program on DDDs and DOT time series data was analyzed by means of intervention time series analysis. A total of 5,002 patients were included (1,971 for the first [before] period and 3,031 for the second [after] period). The percentage of patients treated with one or more antimicrobials decreased from 88.6 to 77.2% (P < 0.001). DDDs decreased from 246.8 to 192.3 (mean difference, −54.5; P = 0.001) and DOT from 66.7 to 54.6 (mean difference, −12.1; P = 0.066). The mean cost per trimester decreased from €115,543 to €73,477 (mean difference, −42,065.4 euros; P < 0.001), and the percentage of ICU antimicrobials cost with respect to the total cost of hospital antimicrobials decreased from 28.5 to 22.8% (mean difference, −5.59; P = 0.023). Implementation of an AMS program in the ICU was associated with a marked reduction in the use of antimicrobials, with cost savings close to one million euros since its implementation. An AMS program can have a significant impact on optimizing antimicrobial use in critical care practice.
KEYWORDS: antimicrobial stewardship, antimicrobials, intensive care, hospital costs, drug costs, ICU, antimicrobial agents, stewardship program
INTRODUCTION
Antimicrobial stewardship (AMS) programs have been developed for optimizing the treatment of infections, to reduce infection-related morbidity and mortality, to limit the appearance of multidrug-resistant organisms (MDROs), and to reduce unnecessary antimicrobial use. This practice ensures the optimal selection, dose, and duration of antimicrobials and leads to the best clinical outcome for the treatment or prevention of infection. In 2000, results of a survey on the status of programs to improve antimicrobial use in 47 U.S. hospitals were reported (1). All hospitals had some forms for the prescription of antimicrobials, and most centers combined some of the strategies recommended by the Society for Health Care Epidemiology of America (SHEA) and the Infectious Diseases Society of America (IDSA) in 1997, such as the development of clinical practice guidelines (70%), stop orders (60%), and restriction policies (40%) (2). Stewardship programs are extensively used in the United States and compliance with the various recommendations is monitored annually by the National Healthcare Safety Network Annual Hospital Survey (3, 4). In Spain, different scientific societies promoted the development of a consensus document for optimizing the use of antimicrobials in hospitals (PROA project) (5).
Patients admitted to intensive care units (ICUs) are commonly treated with one or more antimicrobial agents during their ICU stay, and different international and national registries of antimicrobial use in ICU patients have been developed. In a study of the extent and patterns of infections in 1,265 ICUs from 75 countries carried out in 2007, 71% of patients were receiving antibiotics on the day of the study (6). In Spain, the National ICU-Acquired Infection Surveillance Study (ENVIN-HELICS registry) showed that in 2015, 63.4% of patients were treated with antimicrobials at any point during their stay in the ICU (7).
Many hospitals have implemented AMS programs to improve prescribing practices for hospital inpatients. A Cochrane systematic review of the impact of interventions from the perspective of antimicrobial stewardship identified 96 interventions from 89 studies (8). Most (84%) of the interventions targeted the antibiotic prescribed (choice of antimicrobial, timing of first dose, and route of administration), whereas the remaining interventions aimed to change exposure of patients to antimicrobials by targeting the decision to treat or the duration of treatment (8). Also, a systematic review in the ICU setting suggested that antimicrobial stewardship is associated with improved antimicrobial utilization, with corresponding improvements in antimicrobial resistance and adverse events, and without compromise of short-term clinical outcomes (9).
This study reports the effectiveness of an AMS program implemented in a multimodal ICU of a general hospital in Barcelona, Spain. It was hypothesized that implementation of the program would contribute to reduce antimicrobial consumption in the unit. Therefore, the use of antimicrobials before and after implementation of the AMS program was evaluated.
RESULTS
A total of 5,002 patients were included in the study, 1,971 for the period before the intervention (2007 to 2010) and 3,031 for the period after the intervention (2011 to 2015), with 19,263 and 22,985 days of ICU stay, respectively. Demographic data, severity level on ICU admission, underlying illness, use of invasive techniques, and infection rates associated with invasive devices and/or MDROs in patients admitted between April and July of each of the years analyzed are shown in Table 1. Salient findings for the postintervention period including a significant decrease in elective surgical patients, the use of invasive procedures, the mean length of ICU stay, and the rate of devices-related infections, as well as a decrease in the number of ICU-acquired infections caused by MDROs.
TABLE 1.
General characteristics of the study patientsa
| Parameter | Before AMS program |
After AMS program |
Pb | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 2007 | 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | 2014 | 2015 | ||
| Patients analyzed (April to June) | 120 | 108 | 107 | 131 | 126 | 130 | 141 | 144 | 153 | |
| Men (%) | 70 | 72.2 | 60.8 | 63.4 | 65.9 | 66.2 | 63.8 | 63.2 | 60.1 | 0.353 |
| Mean age in yrs (SD) | 57.9 (17.5) | 59.4 (17.4) | 57.5 (19.8) | 61.7 (15.9) | 56.7 (17.9) | 59.4 (19.4) | 59.5 (19.7) | 60.6 (17.9) | 60.9 (16.2) | 0.366 |
| Mean APACHE II score on ICU admission (SD) | 16.3 (9.0) | 16.9 (9.5) | 16.7 (8.6) | 18.3 (9.0) | 17.1 (9.6) | 14.6 (8.0) | 16.3 (9.6) | 15.5 (9.8) | 16.9 (10.2) | 0.109 |
| Underlying illness (% patients) | 0.057 | |||||||||
| Medical | 73.9 | 78.6 | 83.3 | 87.8 | 84.1 | 82.3 | 83.0 | 85.4 | 79.7 | |
| Surgical | 15.3 | 11.2 | 9.3 | 5.7 | 9.5 | 6.2 | 4.3 | 3.5 | 8.5 | |
| Trauma | 10.8 | 10.2 | 8.3 | 6.5 | 6.4 | 11.5 | 12.8 | 11.1 | 11.8 | |
| Urgent surgery | 19.2 | 19.4 | 13.1 | 15.3 | 9.5 | 5.4 | 8.5 | 12.5 | 11.1 | <0.001 |
| Invasive procedures (% patients) | ||||||||||
| Mechanical ventilation | 54.2 | 64.8 | 53.3 | 53.4 | 45.2 | 50.0 | 46.1 | 48.6 | 43.8 | 0.002 |
| Central venous catheter | 83.3 | 87.0 | 74.8 | 70.1 | 65.1 | 64.6 | 72.3 | 68.1 | 64.1 | <0.001 |
| Urethral catheter | 84.2 | 95.4 | 89.7 | 87.0 | 75.4 | 80.8 | 85.1 | 72.2 | 75.2 | <0.001 |
| Extrarenal depuration techniques | 3.3 | 11.1 | 7.5 | 8.4 | 8.7 | 1.5 | 8.5 | 5.6 | 9.2 | 0.715 |
| Ventricular drainage catheter | 3.3 | 9.3 | 2.8 | 6.9 | 8.7 | 2.3 | 5.7 | 6.3 | 8.5 | 0.683 |
| Total parenteral nutrition | 10.0 | 8.3 | 9.4 | 9.9 | 7.9 | 3.1 | 7.8 | 9.7 | 7.2 | 0.207 |
| Health care-related infections (≥1) (% patients) | 11.0 | 12.0 | 8.4 | 9.4 | 6.2 | 3.7 | 3.6 | 4.3 | 4.3 | <0.001 |
| Noninvasive mechanical ventilation rate | 14.6 | 6.4 | 11.1 | 11.0 | 11.9 | 10.0 | 3.1 | 7.5 | 3.9 | 0.224 |
| Catheter-related bacteremia rate | 2.7 | 6.9 | 0.8 | 4.4 | 1.1 | 1.8 | 0 | 2.0 | 0 | 0.003 |
| Bladder catheter urinary infection rate | 5.8 | 3.1 | 2.3 | 3.8 | 8.1 | 0.8 | 5.9 | 4.8 | 2.4 | 0.920 |
| Antimicrobials in the ICU (% patients) | 70.8 | 78.0 | 69.9 | 77.8 | 71.4 | 59.2 | 66.7 | 68.1 | 69.9 | 0.012 |
| Previous antimicrobials (% patients) | 51.7 | 56.5 | 42.1 | 34.4 | 38.6 | 44.6 | 44.7 | 22.9 | 28.8 | <0.001 |
| Mean ICU stay in days (SD) | 11.5 (12.4) | 13.4 (14.3) | 11.8 (12.5) | 9.5 (9.4) | 8.7 (8.3) | 8.4 (10.8) | 8.7 (9.6) | 8.3 (11.3) | 8.0 (9.5) | <0.001 |
| ICU mortality (% patients) | 12.5 | 16.7 | 12.2 | 13.7 | 11.1 | 11.5 | 14.9 | 10.4 | 12.4 | 0.468 |
| ICU-acquired MDROs (% patients) | 15.8 | 32.4 | 27.1 | 22.1 | 12.7 | 14.6 | 18.4 | 2.8 | 3.3 | <0.001 |
AMS, antimicrobial stewardship; APACHE, Acute Physiology and Chronic Health Evaluation; MDROs, multidrug-resistant organisms.
Periods before and after implementation of the AMS program.
During the period before the intervention, 1,746 patients (88.6% of all ICU patients) received one or more antimicrobials, whereas during the second period, 2,339 patients (77.2% of all ICU patients) were treated with one or more agents. Defined daily doses (DDDs) per 1,000 patient days for each class of antimicrobials used during each study year are shown in Table 2. The DDDs of antimicrobials used in the ICU per 100 occupied beds before and after implementation of an antimicrobial stewardship program are shown in Fig. 1. The mean DDDs of all antimicrobials consumed in the ICU decreased from 246.8 to 192.3 per 1,000 patient days (mean difference, −54.5; P = 0.001). Quarterly changes in the consumption of the different families of antimicrobial agents are shown in Fig. 2. After implementation of the AMS program, there was a significant decrease in the use of cephalosporins (P < 0.001), aminoglycosides (P < 0.001), quinolones (P < 0.001), azoles (P = 0.030), lipid-associated amphotericin B (P = 0.001), and glycopeptides (P < 0.001). A significant reduction in the prescription of carbapenems was not achieved. Also, there was a nonsignificant decrease in the use of colistin, tigecycline, penicillins, echinocandins, metronidazole, and linezolid, as well as an increase in the consumption of macrolides and daptomycin.
TABLE 2.
Defined daily doses (DDD) per 1,000 patient days before and after implementation of an AMS program in the ICUa
| Antimicrobial family classes | DDD per 1,000 patient days |
Pb | ||
|---|---|---|---|---|
| Overall (2007–2015) | Before AMS program (2007–2010) | After AMS program (2011–2015) | ||
| Penicillins | 60.66 | 64.09 | 57.22 | 0.110 |
| Cloxacillin | 7.60 | 8.66 | 6.53 | 0.142 |
| Amoxicillin | 3.52 | 4.17 | 2.86 | 0.013 |
| Ampicillin | 9.16 | 10.18 | 8.14 | 0.368 |
| Amoxicillin-clavulanate | 22.63 | 20.08 | 25.17 | 0.001 |
| Piperacillin-tazobactam | 14.82 | 17.06 | 12.57 | 0.015 |
| Cephalosporins | 20.41 | 24.86 | 15.95 | 0.004 |
| First generation | 0.68 | 0.71 | 0.64 | 0.543 |
| Second generation | 0.14 | 0.13 | 0.15 | NE |
| Third generation | 17.87 | 21.80 | 13.94 | <0.001 |
| Fourth generation | 1.72 | 2.23 | 1.21 | 0.308 |
| Fifth generation | 0 | 0 | 0.01 | NE |
| Aminoglycosides | 5.24 | 8.00 | 2.47 | <0.001 |
| Amikacin | 4.41 | 7.12 | 1.70 | <0.001 |
| Glycylcyclines (tigecycline) | 4.37 | 5.61 | 3.13 | 0.722 |
| Macrolides, lincosamides | 10.94 | 10.65 | 11.23 | 0.463 |
| Azithromycin | 3.27 | 2.97 | 3.57 | 0.351 |
| Glycopeptides | 6.63 | 10.83 | 2.43 | <0.001 |
| Vancomycin | 6.52 | 10.66 | 2.38 | <0.001 |
| Daptomycin | 6.57 | 5.76 | 7.38 | 0.263 |
| Metronidazole | 2.17 | 1.81 | 2.52 | 0.294 |
| Amphotericin | 1.43 | 2.17 | 0.69 | 0.001 |
| Azoles | 17.21 | 19.80 | 14.62 | 0.031 |
| Echinocandins | 5.56 | 5.96 | 5.16 | 0.365 |
| Fluoroquinolones | 22.22 | 27.56 | 16.88 | <0.001 |
| Ciprofloxacin | 11.25 | 13.60 | 8.90 | 0.005 |
| Levofloxacin | 10.83 | 13.69 | 7.97 | <0.001 |
| Colistin | 10.19 | 12.65 | 7.72 | 0.301 |
| Carbapenems | 33.79 | 34.64 | 32.93 | 0.858 |
| Ertapenem | 0.85 | 0.60 | 1.09 | 0.102 |
| Imipenem | 14.67 | 14.95 | 14.36 | 0.965 |
| Meropenem | 17.75 | 18.12 | 17.38 | 0.971 |
| Linezolid | 12.97 | 13.18 | 12.76 | 0.290 |
| Fosfomycin | 0.70 | 0.42 | 0.98 | 0.020 |
AMS, antimicrobial stewardship; NE, not estimable; DDD, defined daily dose.
Statistical significance obtained from the intervention time series analysis (ITSA) model estimated in quarterly DDD data.
FIG 1.
Defined daily doses (DDDs) of antimicrobials used in the ICU per 100 occupied beds before and after implementation of an antimicrobial stewardship program in 2011 (quarterly data).
FIG 2.
Time trends (quarterly data) of DDDs of antimicrobial families used in the ICU before and after implementation of an AMS program in 2011 (vertical line). (a) Natural penicillins (T1-MA0, −6.197; SE 3.875; P = 0.110); (b) cephalosporins (T1-MA0, −2.196; SE 0.832; P = 0.008); (c) aminoglycosides (T1-MA0, −10.388; SE 2.505; P < 0.001); (d) quinolones (T1-MA0, −10.388; SE 2.505; P < 0.001); (e) carbapenems (T1-MA0, −0.555; SE 3.103; P = 0.858); (f) glycopeptides (T1-MA0, −9.340; SE 1.655; P < 0.001), daptomycin (T1-MA0, 1.919; SE 1.716; P = 0.263), and linezolid (T1-MA0, −1.925; SE 1.820; P = 0.290); (g) other active antimicrobials against multidrug-resistant organisms, such as tigecycline (T1-MA0, −0.871; SE 2.452; P = 0.722) and colistin (T1-MA0, −3.402; SE 3.298; P = 0.301); and 2 h: antifungals, such as azoles (T1-MA0, −5.993; SE 2.769; P = 0.030), amphotericin (T1-MA0, −1.510; SE 0.472; P = 0.001) or candins (T1-MA0, −1.154; SE 1.273; P = 0.365).
The cumulative number of treatment days with the different antimicrobials was 34,456 days during the preintervention period and 34,167 during the postintervention period, with treatment days/100 days of ICU stay ratios of 178.9 and 148.6, respectively (P < 0.001). The days of therapy (DOT) per 100 occupied bed-days between the two periods decreased from 66.7 to 54.6 (mean difference −12.1; P = 0.066). The DOT decreased significantly for the consumption of cephalosporins (P = 0.008), aminoglycosides (P < 0.001), glycopeptides (P = 0.007), lipid-associated amphotericin B (P = 0.001), and azoles (P = 0.005). The DOT decreases for quinolones, carbapenems, colistin, tigecycline, and candins were less relevant. In contrast, there was a nonsignificant increase in DOT of penicillins (P = 0.076), macrolides, metronidazole, linezolid, and daptomycin (Table 3).
TABLE 3.
Days of therapy before and after implementation of an AMS program in the ICUa
| Antimicrobial family classes | DOT per 100 occupied bed-days |
Pb | ||
|---|---|---|---|---|
| Overall (2007–2015) | Before AMS program (2007–2010) | After AMS program (2011–2015) | ||
| Penicillins | 11.77 | 11.02 | 12.52 | 0.076 |
| Cloxacillin | 0.59 | 0.59 | 0.58 | 0.949 |
| Amoxicillin | 0.21 | 0.26 | 0.15 | 0.191 |
| Ampicillin | 0.49 | 0.53 | 0.45 | 0.632 |
| Amoxicillin-clavulanate | 4.78 | 3.61 | 5.95 | <0.001 |
| Piperacillin-tazobactam | 5.56 | 5.94 | 5.18 | 0.230 |
| Cephalosporins | 5.94 | 6.92 | 4.95 | 0.008 |
| First generation | 0.19 | 0.20 | 0.18 | 0.696 |
| Third generation | 5.44 | 6.39 | 4.48 | 0.009 |
| Fourth generation | 0.31 | 0.33 | 0.29 | 0.839 |
| Aminoglycosides | 2.07 | 3.25 | 0.89 | <0.001 |
| Amikacin | 1.81 | 2.86 | 0.75 | <0.001 |
| Glycylcyclines (tigecycline) | 1.83 | 2.83 | 0.82 | 0.069 |
| Macrolides, lincosamides | 1.87 | 1.71 | 2.03 | 0.388 |
| Azithromycin | 0.30 | 0.25 | 0.34 | 0.545 |
| Glycopeptides | 2.37 | 3.41 | 1.33 | 0.007 |
| Vancomycin | 2.34 | 3.41 | 1.26 | 0.006 |
| Daptomycin | 1.55 | 1.48 | 1.62 | 0.797 |
| Metronidazole | 1.21 | 0.98 | 1.43 | 0.429 |
| Amphotericin | 0.89 | 1.35 | 0.43 | 0.001 |
| Azoles | 4.99 | 6.19 | 3.80 | 0.005 |
| Echinocandins | 2.53 | 3.10 | 1.95 | 0.157 |
| Fluoroquinolones | 5.10 | 5.91 | 4.29 | 0.097 |
| Ciprofloxacin | 2.31 | 2.92 | 1.70 | 0.079 |
| Levofloxacin | 2.79 | 2.99 | 2.59 | 0.385 |
| Colistin | 2.73 | 3.63 | 1.82 | 0.771 |
| Carbapenems | 10.44 | 10.58 | 10.30 | 0.819 |
| Ertapenem | 0.20 | 0.12 | 0.28 | 0.333 |
| Imipenem | 6.33 | 6.53 | 6.12 | 0.667 |
| Meropenem | 3.70 | 3.59 | 3.81 | 0.901 |
| Linezolid | 4.78 | 4.40 | 5.15 | 0.740 |
| Fosfomycin | 0.03 | 0.02 | 0.03 | 0.875 |
DOT, days of therapy; AMS, antimicrobial stewardship.
Statistical significance obtained from the intervention time series analysis (ITSA) model estimated in quarterly DOT data.
The mean cost per trimester decreased from €115,543 to €73,477 (mean difference, −42,065.4 euros; P < 0.001) (Fig. 3). Also, the percentage of costs of antimicrobials consumed in the ICU with respect to the total costs of antimicrobials consumed in the hospital decreased from 28.5 to 22.8% (mean difference, −5.59, P = 0.023) (Fig. 4) between both study periods.
FIG 3.
Time trends (quarterly data) of consumption of antimicrobials in the ICU expressed in euros (in 2015) before and after implementation of an antimicrobial stewardship program.
FIG 4.
Percentage of hospital consumption of antimicrobials corresponding to the ICU expressed in euros (in 2015) before and after implementation of an antimicrobial stewardship program.
DISCUSSION
This study shows that implementation of an AMS program in the ICU of a general hospital led by intensivists, with special expertise in infections in critically ill patients, was associated with an important reduction of antimicrobial use. The DDDs and DOT decreased significantly in the postintervention period, with a reduction of overall costs of antimicrobial agents close to one million euros during the first 5 years of implementation of the program.
Our data confirm previous studies demonstrating a decrease of DDDs after implementation of programs for optimizing the use of antimicrobials in critical patients (9–18), although there are notable differences across studies in the assessment of the impact of this decrease. Slain et al. (10) showed an important decline in the ICU use of ciprofloxacin and ceftazidime but a relationship with a decrease of Pseudomonas aeruginosa resistance rate was not found. However, Elligsen et al. (11) reported a decrease in the incidence of Clostridium difficile infections associated with a reduction of broad-spectrum antibiotic use. In the study of Katsios et al. (12), a decrease in DDDs was related to a reduced use of antimicrobials for the treatment of bacteria recovered from nonsterile site cultures. DiazGranados (13) concluded that antimicrobial audit and feedback had an influence on antimicrobial prescription pattern in the ICU with a favorable impact on the emergence of resistance. After implementation of an AMS program, Amer et al. (14) reported an increase in the prescribing appropriateness rate of the empirical antibiotic therapy. Also, Rimawi et al. (15) showed that regular interaction of an infection disease fellow with the medical ICU team resulted in a significant decreases of extended-spectrum penicillins, carbapenems, vancomycin, and metronidazole, with reductions in mechanical ventilation days, length of ICU stay, and hospital mortality. In contrast, Taggart et al. (16) assessed the impact of an AMS program in two ICUs with inconsistent results, including a reduction of antibiotic use in the trauma and neurosurgery ICU and an increase in a medical-surgical ICU, and no effect on ICU stay, mortality, or changes in the resistance patterns of Escherichia coli and P. aeruginosa in either intervention unit. Finally, Chen et al. (17) evaluated the effect of an online AMS program and showed a decline of DDDs, shorter duration of administration of all classes of antibiotics, and lower incidences of health care-associated infections (HAIs). The present findings are consistent with the results of these studies, although reductions in HAIs were also due to the application of recommendations of safety projects developed in our ICU during implementation of the antimicrobial optimizing program. A main difference in relation to previous studies is that our program has been led in the ICU by an intensivist physician who was an expert in infectious diseases in critically ill patients, with a daily intervention at two different times (at first hour in the morning and at first hour in the afternoon at shift changes), and in which there was also a differentiation of the impact of the program on the ICU and on the remaining hospital.
The general characteristics of ICU patients included in the study were similar before and after the intervention, without substantial changes in mean age or severity level on ICU admission, although there was a decrease in the number of urgent surgery patients due to organizational changes in the care of surgical ICU patients. A reduction in the mean length of stay and ICU-related mortality can be attributed to various concomitant factors, including development of projects for reducing the rates of catheter-related bloodstream infection and ventilator-associated pneumonia, as well as the introduction of new technologies and policies, which have contributed to improve the quality of care to reduce morbidity.
The AMS program implemented in this study did not include restrictive measures on prescription. It was focused on making available therapeutic protocols for the main community-acquired and nosocomial infections and on the control of 10 antimicrobial agents with well-known influence on clinical outcome and hospital costs (18, 19). To this purpose, a working group of specialists with recognized experience in the field of infectious disease assessed appropriateness of indications of these 10 antimicrobial agents in the next 48 to 72 h after their prescription. Evaluations made were registered in the hospital information system. In the ICU, the AMS program was implemented to the routine clinical practice of the unit, so that the need of antimicrobials, doses, regimens, conditions that could modify concentrations at the infection site (renal function, volume of distribution, and diffusion to infected tissues), and decisions taken during the patient's clinical course were systematically recorded. The support of the Pharmacy Service for the measurement of plasma and body fluid concentrations of several antimicrobials during the second period of the study allowed timely modification of the dose and/or the interval of administration of some agents (20).
The total percentage of patients treated with antimicrobials and the overall DDDs decreased in the postintervention period. This was due, among other reasons, to a reduction in HAIs associated with active participation of the ICU in projects to reduce catheter-related bacteremia and pneumonia, as well as to an increase in short stays of neurological patients undergoing successful intervention procedures. In addition, there was a decrease in the DOT as a result of the implementation of the antimicrobial optimizing program, in which daily evaluation of the possibility of stopping antimicrobial administration was included among the actions of the program.
The reduction of antimicrobials consumption was mainly associated with agents of the families of cephalosporins, quinolones, aminoglycosides, glycopeptides, and lipid-associated amphotericin B, with persistence or increases of carbapenems and natural penicillins at the expense of a higher use of amoxicillin-clavulanate and piperacillin-tazobactam. Also, in the postintervention period there was a decrease in the use of colistin and tigecycline in relation to a reduction of MDROs, particularly multiresistant P. aeruginosa and the disappearance of multiresistant Acinetobacter baumannii or carbapenemase-producing Gram-negative bacteria, in this last case due to implementation of recommendations of an ongoing “antibiotic resistance zero” project. The concomitant and progressive implementation of “bacteremia zero” and “pneumonia zero” programs during the postintervention period also affected the consumption of antimicrobials.
An important consequence of the AMS program has been a significant reduction in antimicrobial-related costs which has independently of price changes, since calculations were based on prices for 2015 to avoid bias of generic formulations following patent expiration of some drugs (e.g., meropenem). Also, in the postintervention period in 2013, selective digestive decontamination was adopted for preventing mechanical ventilation-associated pneumonia, but the administration of nonabsorbable antimicrobial combinations did not involve additional expenses.
Although the AMS program has been implemented in the whole hospital, the impact has been greater in the ICU than in the remaining services as shown by a decrease in the fraction of antimicrobial consumption attributable to the ICU. This may be due to the particular application of the program in the ICU setting, with an intensivist directly responsible for the antimicrobial policy and compliance with the program. Daily assessment of each specific indication and clinical response, evaluation of treatment withdrawal and possible morbidities, and daily adjustment of doses and dosing intervals of some agents according to glomerular filtration rate and pharmacokinetic data have been key features for success.
Limitations of the study include (i) the potential influence of changes in the hospital and ICU antimicrobial policies favoring the use of some agents after being included in therapeutic protocols (linezolid, daptomycin, and candins); (ii) the fact that ICU participation in safety projects has been associated with a >50% decrease in HAIs, which contributed to reduce the use of antimicrobials; and (iii) the incorporation of the recommendations of the “antibiotic resistance zero” project and the introduction of selective digestive decontamination, which have been associated with a decrease of MDROs, with a subsequent decline in the use of rescue antimicrobials, such as colistin and tigecycline. In addition, internal organization changes in the hospital that have modified the flow of patients (reduction of patients requiring urgent surgery) may have contributed to modify the use of antimicrobials. Also, the single-center characteristics of the study may limit generalizability of results to different critical care settings.
Conclusions.
After 5 years of implementation of an AMS program in the ICU, significant reductions of antimicrobial consumption with cost savings close to one million euros were documented. The particular integration of the program in the routine daily practice of the unit and the support of the hospital Pharmacy Service have been key elements for success. Appropriate and judicious antimicrobial use guided by an AMS program is associated with remarkable benefits in critically patients.
MATERIALS AND METHODS
Study design and setting.
We conducted a prospective, interventional, cohort (before-and-after) study in the ICU of a 400-bed acute-care teaching hospital in Barcelona, Spain. Institutional Review Board approval was obtained for the study. The hospital belongs to the Catalan Health Care System (CatSalut), which is a public and free-access service that provides health care coverage for all citizens resident in the Autonomous Community of Catalonia. The hospital has a 14-bed multimodal ICU for the care of critically ill patients, excluding coronary patients. The unit presents a circular structure with independent rooms that can be isolated by clear glass walls. Nursing staff for each shift includes one registered nurse for each two beds and one nurse assistant for each six beds. Written therapeutic protocols for the main community-acquired and hospital-acquired infections are also available. These protocols, developed and approved by the Infection Committee, which is responsible for the antibiotics policy in the hospital, are updated every 2 years and are included in the computerized hospital information system. The ICU participates annually in the National ICU-Acquired Infection Surveillance Study (ENVIN-HELICS registry) and has adhered to nationwide “bacteremia zero,” “pneumonia zero,” and “resistance zero” projects sponsored by the Spanish Ministry of Health, Social Services, and Equality and led by the Working Group of Infectious Diseases and Sepsis of the Spanish Society of Intensive Care Medicine and Coronary Units (SEMICYUC) (21).
Antimicrobial stewardship program.
Between 1 January and 31 March 2011, a program for optimizing the use of antimicrobials was implemented in the hospital. Prior to the application of the AMS program, each attending physician or medical team decided on which antimicrobials had to be administered to their patients and the duration of treatment. No restriction programs for the use of antimicrobials have been implemented before 2011. In the hospital and in the ICU, there were some guidelines for the use of antibiotics, but these recommendations have not been subjected to audits or assessments of adherence. The AMS program included the following characteristics: (i) establishment of a multidisciplinary working group with specialists in infectious diseases, pharmacy, microbiology, and intensive care medicine who were responsible for the design and implementation of the AMS program; (ii) development of a computer application for the specific prescription of antimicrobials, which was added to the patient's computerized medical record; and (iii) selection of 10 antimicrobial agents, which due to their greater environmental and economic impact, underwent special control measures, including the need to justify their indications in writing through the application form of the program, compulsory detailed information of the duration of treatment, immediate information of the cost of prescription, and automatic discontinuation of drug administration on the day set by the prescriber physician, with reassessment of indications during the next 24 to 72 h by a member of the working group. These antimicrobials were carbapenems (imipenem and meropenem), tigecycline, linezolid, voriconazole, candins (caspofungin, anidulafungin, and micafungin), and lipid-associated amphotericin B (liposomal, lipid complex). In the ICU setting, an expert intensivist in infectious diseases was responsible for implementation of the program. Actions included daily review of antibiotics regimens of all patients during the shift change (8:00 a.m., 3:00 p.m., and 9:00 p.m.), obligation to include the number of days of antimicrobial administration in the computerized clinical course record, and reasonable proposals of dose adjustment, deescalation, or withdrawal in the daily clinical sessions of the ICU staff (2:00 to 3:30 p.m.).
The AMS program was well accepted in the ICU setting, with no rejections by the ICU personnel. Details of the implementation of the recommendations were discussed at daily sessions and final decisions were taken by consensus of the ICU team. In case of disagreement, the opinion of the expert who had received empowerment by the medical director of the hospital and the chief of the ICU prevailed.
Outcome measures.
The primary outcome of the study was the change of antimicrobials consumption in the ICU before and after implementation of the AMS program, that is, between 2007 and 2010 (preintervention or first period) and between 2011 and 2015 (postintervention or second period). The secondary outcome was cost savings attributable to antimicrobial expenses derived from implementation of the program.
Characteristics of the study population.
Data collected from patients admitted to the ICU between 1 April and 30 June each year and recorded in the ENVIN-ICU registry (21) were retrieved for the study. Data included demographics (age and sex), underlying illness, need for urgent surgery, severity level (APACHE II score within the first 24 h of ICU admission), invasive procedures, use of antimicrobials during ICU stay, ICU length of stay, and ICU-related mortality. Also, infection rates related to invasive devices and the numbers of patients with infections caused by MDROs were recorded. Definitions of invasive device-associated infections and MDROs were included in the ENVIN-HELICS registry manual (21). Patients were classified according to diagnoses—trauma, surgical, or medical—at the time of ICU admission. Trauma patients were patients for whom the reason for admission included lesions related to an acute trauma injury. Surgical patients were admitted to the ICU for postoperative control of an elective surgical procedure. Medical patients included patients for whom the reason for admission was none of the above and included surgical patients admitted to the ICU because of medical complications. Coronary patients were excluded.
Assessment criteria.
The only antimicrobials analyzed (doses and days of use) were those administered to patients during their stay in the ICU, regardless of whether the patients had been treated with these agents before or after discharge from the ICU. The consumption and costs of antimicrobials were calculated independently by the Pharmacy Service, using a computer tool for estimating DDDs, DOT, and costs. DDD is the assumed average maintenance dose per day for a drug used for its main indication in adults (22). DOT represents the number of days that a patient is on an antimicrobial, regardless of dose. Costs of antimicrobials were based on 2015 reference prices and were calculated automatically for the hospital as a whole and for each of the hospital services. The following indicators were calculated on a quarterly basis: (i) overall DDDs of antimicrobials consumed in the ICU; (ii) DDDs for the families of natural penicillins (penicillin G, cloxacillin, amoxicillin, ampicillin, amoxicillin-clavulanic acid, and piperacillin-tazobactam), cephalosporins (classified into generations), aminoglycosides (gentamicin, tobramycin, and amikacin), quinolones (ciprofloxacin and levofloxacin), carbapenems (ertapenem, imipenem, and doripenem), glycopeptides and other active antimicrobials against multiresistant Gram-positive cocci (teicoplanin, vancomycin, daptomycin, and linezolid), other antimicrobials (tigecycline, colistin, phosphomycin, and azithromycin), and antifungals (amphotericins, azoles, and candins); (iii) costs of antimicrobials used in the whole hospital and in the ICU based on the acquisition price for 2015 and expressed in euros (€); (iv) cumulative treatment days expressed in DOT (days of treatment of each agent were added regardless of whether several agents were administered on the same day); and (v) the percentage of hospitals costs for antimicrobials corresponding to the ICU. For the second period of the study, real-time measurements of linezolid, colistin, and micafungin concentrations in plasma and biological fluids were available, which facilitated dosing and regimen adjustments.
Statistical analysis.
Categorical variables are expressed as frequencies and percentages, and continuous variables are expressed as means and standard errors (SE) or medians and interquartile ranges (25th to 75th percentiles) according to the normal or nonnormal distribution of the variables. Data recorded for the preintervention (2007 to 2010) and postintervention (2011 to 2015) periods were compared using the chi-square (χ2) test or the Fisher exact test for categorical variables or the Student t test or the Mann-Whitney U test for quantitative variables according to the conditions of application. Rates were analyzed by applying Poisson regression to assess differences between periods. Intervention time series analysis was used to assess the effect of the AMS program on DDDs and DOT time series data. First, the previous to the intervention application data was used to define the ARIMA model for the series; this model was transferred to forecast values after the intervention in order to estimate the intervention effect. This analysis allowed estimation of changes of DDDs and DOT between preintervention and postintervention phases, while accounting for both sudden changes and the change trends of the outcome of interest. Analyses were performed using R version 3.1.2 language for statistical analysis (23).
ACKNOWLEDGMENTS
We thank all ICU professionals for their dedication and efforts to implement the AMS program and to complete data sheets for the patients. We also thank the pharmacists and technicians of the Pharmacy Service for their contribution to the measurement of plasma drug levels and valuable recommendations on dosing adjustments, Marta Gras for data collection, and Marta Pulido for editing the manuscript and editorial assistance.
The authors declare no conflict of interest.
No financial support was received for this study.
REFERENCES
- 1.Lawton RM, Fridkin SK, Gaynes RP, McGowan JE Jr. 2000. Practices to improve antimicrobial use at 47 US hospitals: the status of the 1997 SHEA/IDSA position paper recommendations. Society for Healthcare Epidemiology of America/Infectious Diseases Society of America. Infect Control Hosp Epidemiol 21:256–259. doi: 10.1086/501754. [DOI] [PubMed] [Google Scholar]
- 2.Shlaes DM, Gerding DN, John JF Jr, Craig WA, Bornstein DL, Duncan RA, Eckman MR, Farrer WE, Greene WH, Lorian V, Levy S, McGowan JE Jr, Paul SM, Ruskin J, Tenover FC, Watanakunakorn C. 1977. Society for Healthcare Epidemiology of America and Infectious Diseases Society of America Joint Committee on the prevention of antimicrobial resistance: guidelines for the prevention of antimicrobial resistance in hospitals. Infect Control Hosp Epidemiol 18:275–291. doi: 10.1086/647610. [DOI] [PubMed] [Google Scholar]
- 3.Pollack LA, van Santen KL, Weiner LM, Dudeck MA, Edwards JR, Srinivasan A. 2016. Antibiotic stewardship programs in U.S. acute care hospitals: findings from the 2014 National Healthcare Safety Network Annual Hospital Survey. Clin Infect Dis 63:443–449. doi: 10.1093/cid/ciw323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.O'Leary EN, van Santen KL, Webb AK, Pollock DA, Edwards JR, Srinivasan A. 2017. Uptake of antibiotic stewardship programs in US acute care hospitals: findings from the 2015 National Healthcare Safety Network Annual Hospital Survey. Clin Infect Dis 65:1748–1750. doi: 10.1093/cid/cix651. [DOI] [PubMed] [Google Scholar]
- 5.Rodríguez-Baño J, Paño-Pardo JR, Alvarez-Rocha L, Asensio A, Calbo E, Cercenado E, Cisneros JM, Cobo J, Delgado O, Garnacho-Montero J, Grau S, Horcajada JP, Hornero A, Murillas-Angoiti J, Oliver A, Padilla B, Pasquau J, Pujol M, Ruiz-Garbajosa P, San Juan R, Sierra R; Grupo de Estudio de la Infección Hospitalaria-Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica; Sociedad Española de Farmacia Hospitalaria; Sociedad Española de Medicina Preventiva Salud Pública e Higiene. 2012. Programs for optimizing the use of antibiotics (PROA) in Spanish hospitals: GEIH-SEIMC, SEFH and SEMPSPH consensus document. Enferm Infecc Microbiol Clin 30:22.e1–22.e23. (In Spanish). doi: 10.1016/j.eimc.2011.09.018. [DOI] [PubMed] [Google Scholar]
- 6.Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD, Moreno R, Lipman J, Gomersall C, Sakr Y, Reinhart K; EPIC II Group of Investigators. 2009. International study of the prevalence and outcomes of infection in intensive care units. JAMA 302:2323–2329. doi: 10.1001/jama.2009.1754. [DOI] [PubMed] [Google Scholar]
- 7.Spanish Society of Intensive Care Medicine and Coronary Units. 2017. Study Group of Infectious Diseases and Sepsis: National Surveillance Study of Nosocomial Infection—2015 annual report. SEMICYUC, Madrid, Spain: http://hws.vhebron.net/envin-helics. [Google Scholar]
- 8.Davey P, Marwick CA, Scott CL, Charani E, McNeil K, Brown E, Gould IM, Ramsay CR, Michie S. 2017. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev 2:CD003543. doi: 10.1002/14651858.CD003543.pub4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kaki R, Elligsen M, Walker S, Simor A, Palmay L, Daneman N. 2011. Impact of antimicrobial stewardship in critical care: a systematic review. J Antimicrob Chemother 66:1223–1230. doi: 10.1093/jac/dkr137. [DOI] [PubMed] [Google Scholar]
- 10.Slain D, Sarwari AR, Petros KO, McKnight RL, Sager RB, Mullett CJ, Wilson A, Thomas JG, Moffett K, Palmer HC, Dedhia HV. 2011. Impact of a multimodal antimicrobial stewardship program on Pseudomonas aeruginosa susceptibility and antimicrobial use in the intensive care unit setting. Crit Care Res Pract 2011:416426. doi: 10.1155/2011/416426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Elligsen M, Walker SA, Pinto R, Simor A, Mubareka S, Rachlis A, Allen V, Daneman N. 2012. Audit and feedback to reduce broad-spectrum antibiotic use among intensive care unit patients: a controlled interrupted time series analysis. Infect Control Hosp Epidemiol 33:354–361. doi: 10.1086/664757. [DOI] [PubMed] [Google Scholar]
- 12.Katsios CM, Burry L, Nelson S, Jivraj T, Lapinsky SE, Wax RS, Christian M, Mehta S, Bell CM, Morris AM. 2012. An antimicrobial stewardship program improves antimicrobial treatment by culture site and the quality of antimicrobial prescribing in critically ill patients. Crit Care 16:R216. doi: 10.1186/cc11854. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.DiazGranados CA. 2012. Prospective audit for antimicrobial stewardship in intensive care: impact on resistance and clinical outcomes. Am J Infect Control 40:526–529. doi: 10.1016/j.ajic.2011.07.011. [DOI] [PubMed] [Google Scholar]
- 14.Amer MR, Akhras NS, Mahmood WA, Al-Jazairi AS. 2013. Antimicrobial stewardship program implementation in a medical intensive care unit at a tertiary care hospital in Saudi Arabia. Ann Saudi Med 33:547–554. doi: 10.5144/0256-4947.2013.547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Rimawi RH, Mazer MA, Siraj DS, Gooch M, Cook PP. 2013. Impact of regular collaboration between infectious diseases and critical care practitioners on antimicrobial utilization and patient outcome. Crit Care Med 41:2099–2107. doi: 10.1097/CCM.0b013e31828e9863. [DOI] [PubMed] [Google Scholar]
- 16.Taggart LR, Leung E, Muller MP, Matukas LM, Daneman N. 2015. Differential outcome of an antimicrobial stewardship audit and feedback program in two intensive care units: a controlled interrupted time series study. BMC Infect Dis 15:480. doi: 10.1186/s12879-015-1223-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chen IL, Lee CH, Su LH, Wang YL, Liu JW. 2016. Effects of implementation of an online comprehensive antimicrobial-stewardship program in ICUs: a longitudinal study. J Microbiol Immunol Infect pii:S1684-1182(16)30068-8. doi: 10.1016/j.jmii.2016.06.007. [DOI] [PubMed] [Google Scholar]
- 18.Karanika S, Paudel S, Grigoras C, Kalbasi A, Mylonakis E. 2016. Systematic review and meta-analysis of clinical and economic outcomes from the implementation of hospital-based antimicrobial stewardship programs. Antimicrob Agents Chemother 60:4840–4852. doi: 10.1128/AAC.00825-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Paladino JA. 2004. Economics of antibiotic use policies. Pharmacotherapy 24:232S–238S. doi: 10.1592/phco.24.18.232S.52234. [DOI] [PubMed] [Google Scholar]
- 20.Álvarez-Lerma F, Muñoz-Bermúdez R, Samper-Sánchez MA, Gracia Arnilla MP, Grau S, Luque S. 2017. Successful treatment of Panton-Valentine leukocidin-positive methicillin-resistant Staphylococcus aureus pneumonia with high doses of linezolid administered in continuous infusion. Med Intensiva 41:56–59. (In Spanish.) doi: 10.1016/j.medin.2016.04.006. [DOI] [PubMed] [Google Scholar]
- 21.Spanish Society of Intensive Care Medicine and Coronary Units 2017. Working Group of Infectious Diseases and Sepsis (GTEIS). National Surveillance Study of Nosocomial Infection in the ICU (ENVIN-HELICS). Bacteremia Zero project Pneumonia Zero project, Resistance Zero project. SEMICYUC, Madrid, Spain: http://hws.vhebron.net/envin-helics. [Google Scholar]
- 22.World Health Organization Collaborating Centre for Drugs Statistics Methodology 2017. DDD: Definition and general considerations. World Health Organization Collaborating Centre for Drugs Statistics Methodology, Geneva, Switzerland: https://www.whocc.no/ddd/definition_and_general_considera/. [Google Scholar]
- 23.Core Team R 2017. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria: http://www.R-project.org. [Google Scholar]