Advertisement

Do antimicrobial stewardship programs improve the quality of care in ICU patients diagnosed with infectious diseases following consultation? Experience in a tertiary care hospital

Open AccessPublished:December 06, 2021DOI:https://doi.org/10.1016/j.ijid.2021.12.315

      HIGHLIGHTS

      • The growing problem of antibiotic resistance is a major public health concern.
      • An antimicrobial stewardship study was conducted in intensive care units.
      • There was a significant increase in total quality-of-care scores.

      ABSTRACT

      Background

      One of the most important public health concerns is the ever-growing problem of antibiotic resistance. Importantly, the rate of introduction of new molecules into clinical practice has slowed down considerably. Moreover, the rapid emergence of resistance shortens the effective ‘lifespan’ of these molecules.

      Objective

      The quality of care before and after active intervention and feedback was evaluated in patients diagnosed with sepsis/septic shock or ventilator-associated pneumonia (VAP) in the ICUs of Hacettepe University Adult and Oncology Hospitals.

      Results

      There was a significant increase in total scores. Significant improvements were achieved in the management of these patients in terms of requests for necessary diagnostic tests, and the prolonged infusion of beta-lactam agents.

      Conclusion

      Implementation of an ASP in centers where antimicrobial management of ICU patients is largely controlled by infectious diseases specialists remains a feasible strategy that leads to better patient care.

      Keywords

      INTRODUCTION

      One of the most important public health concerns is the ever-growing problem of antibiotic resistance (
      • Aslam B
      • Wang W
      • Arshad MI
      • et al.
      Antibiotic resistance: a rundown of a global crisis.
      ). Importantly, the rate of introduction of new molecules into clinical practice has slowed down considerably. Moreover, the rapid emergence of resistance shortens the effective ‘lifespan’ of these molecules (
      • Conly J
      • Johnston B.
      Where are all the new antibiotics? The new antibiotic paradox.
      ). In our intensive care units (ICUs), 31.3% of Staphylococcus aureus strains isolated from blood and non-directed bronchial lavage (BL) cultures were methicillin resistant (MRSA), 51.3% of Enterobacterales isolated from BL cultures were extended-spectrum beta-lactamase (ESBL) producers, and 38.2% of Gram-negative bacilli isolated from BL cultures were carbapenem resistant in 2018 (data from the Hospital Infection Control Committee; not published).
      In our hospital, the use of third-generation cephalosporins, piperacillin-tazobactam, carbapenems, glycopeptides, tigecycline, parenteral fluoroquinolones, and aminoglycosides is restricted to the approval of infectious diseases (ID) specialists, as required by the Turkish Social Security Institution Regulation on Drug Reimbursement. The routine practice is consultation with the ID Department for patients — in ICUs or wards — suspected to be infected, and to plan management jointly. Thus, one would assume the baseline quality of care already to be very high, and anticipate minimal improvement following an antimicrobial stewardship program (ASP). However, a significant number of critically ill patients are usually consulted during night shifts or at weekends, when the initial evaluation of patients is completed by junior physicians, before being consulted by senior physicians on call. This approach leads to a potential delay in the administration of the first dose of the empirical antibiotic, which until now has has not been subject to a quality improvement program in our hospital. Until January 2019, there was no written local protocol for the diagnosis and treatment of sepsis/septic shock and ventilator-associated pneumonia; nor was there any formal feedback to the intensivists and ID specialists about the quality of antimicrobial therapy in ICU.
      Our study was conducted in patients diagnosed with sepsis/septic shock or ventilator-associated pneumonia (VAP), in the ICUs of Hacettepe University Adult and Oncology Hospitals, in order to assess the quality of care before and after implementation of active intervention and feedback.

      MATERIALS AND METHODS

      Study design

      This quasi-experimental study was conducted in patients diagnosed with sepsis/septic shock or VAP, who were hospitalized in seven ICUs (Internal Medicine, Oncology, Anesthesiology, Neurology, Stroke, Neurosurgery, and General Surgery), accounting for a total of 138 beds in a tertiary care university hospital. The routine practice was as follows: The consulting ID team was notified as soon as the primary physician suspected sepsis/septic shock or VAP in a patient. The ID resident evaluated the patient at the bedside, made necessary recommendations, and rounded daily with the consulting staff member until cessation of therapy, resolution of symptoms and signs, or discharge. Antimicrobial agents were ordered via the hospital intranet by the primary physician, stating with the route of administration, dose, dosing intervals, and special requirements if needed, such as the duration of infusion. The antibiotic was released for use from an automated medication dispensing system (the BD™ Pyxis MedStation™ ES) in the ICU following ID physician approval.
      The primary endpoint of the study was total quality-control scores for management of sepsis/septic shock and VAP, developed by the ID staff. Each parameter included in the quality-control scores was also evaluated separately.

      Study periods

      Period 1 (from April 22, 2017 to October 24, 2018). This period served as baseline because there were no local hospital guidelines or interventions. Data were extracted from the hospital information technology (IT) system for a total of 50 patients with sepsis/septic shock or VAP.
      Period 2 (from April 4, 2019 to August 29, 2019). Local guidelines and algorithms were developed by the academic staff of the Department of Infectious Diseases for the diagnosis, treatment, and follow-up of the most common and/or challenging infectious diseases, and uploaded to the hospital server by the end of January 2019. Two educational meetings were held separately for the ID Department and the ICUs in March 2019. The residents were encouraged to use these local practice guidelines in departmental meetings, rounds, and morning report hours, but no intervention was introduced to increase compliance. The patients were followed and data were collected daily by the antimicrobial stewardship program (ASP) team, without any direct intervention.
      Period 3 (from February 3, 2020 to September 23, 2020). The patients were followed in the same manner as in period 2, but this time the ASP team intervened in the management of the patients via calls and notes in patient charts, as necessary.
      The ASP team consisted of an ID resident, an ID staff member, a clinical pharmacist, the chief of the Central Microbiology Lab, and an ID physician from the Infection Control Unit. Feedback was provided to all members of the ID Department at the ends of periods 2 and 3.

      Study population

      In total, 150 patients — 50 for each period — were included in the study. Patients who were at least 18 years of age, were hospitalized in the relevant ICUs, and started empirical antibiotics for sepsis/septic shock or VAP were included in the study. Patients with infective endocarditis, osteomyelitis, prosthetic device infection, fungal pneumonia, fungemia, and Legionnaires’ disease were excluded.

      Evaluation process

      The clinical conditions of the patients were evaluated on the first day of treatment, after 72 hours, at the end of the optimal treatment period, and on the final day — when antimicrobials were discontinued. The collected data were scored according to Quality Control Scores for Management of Sepsis/Septic Shock and Quality Control Scores for Management of VAP, developed by the ID staff (Tables 1 and 2).
      Table 1Time to the first dose of antibiotic in patients with sepsis/septic shock or VAP.
      SEPSIS/SEPTIC SHOCK
      Stages in clinical practicePeriod 1

      (n = 37), minutes
      Period 2

      (n = 39), minutes
      Period 3

      (n = 37), minutes
      p-value
      ID consultation requested to ID recommendations made, median (IQR)NA73 (105)42 (118)0.102
      ID recommendations made to antibiotic ordered, median (IQR)NA35 (197)27 (84)0.720
      Antibiotic ordered to ID approval, median (IQR)41 (114)16 (47)12 (47.5)0.009
      ID approval to first dose administered, median (IQR)66 (128)15 (36)25 (48)< 0.001
      Optimal time to first dose of antibiotic administered was accepted as ≤ 1 hour by ID staff.
      ID consultation requested to first dose administered, median (IQR)
      179 (359)205 (334)206 (342)0.502
      VENTILATOR-ASSOCIATED PNEUMONIA
      Stages in clinical practicePeriod 1

      (n = 13), minutes
      Period 2

      (n = 11), minutes
      Period 3

      (n = 13), minutes
      p-value
      ID consultation requested to ID recommendations made, median (IQR)NA90 (180)89 (120)1.000
      ID recommendations made to antibiotic ordered, median (IQR)NA40 (82)27 (68)0.776
      Antibiotic ordered to ID approval, median (IQR)160 (186)7 (19)113 (175.5)0.015
      ID approval to first dose administered, median (IQR)7 (74)8 (38)39 (84.5)0.323
      Optimal time to first dose of antibiotic administered was accepted as ≤ 3 hours by ID staff.
      ID consultation requested to first dose administered, median (IQR)
      294 (509)207 (270)157 (867)0.092
      low asterisk Optimal time to first dose of antibiotic administered was accepted as ≤ 1 hour by ID staff.
      low asterisklow asterisk Optimal time to first dose of antibiotic administered was accepted as ≤ 3 hours by ID staff.
      Table 2Quality-of-care parameters with significant improvements in patients with sepsis/septic shock and VAP.
      ParameterPeriod 1(n = 37)Period 2(n = 39)Period3(n = 37)p-value
      SEPSIS/SEPTIC SHOCK
      Calculation of SOFA score, n (%)1 (2.7)5 (12.8)14 (37.8)<0.001
      ≥ two sets of blood cultures, n (%)3 (8.1)10 (25.6)12 (32.4)0.034
      Cultures from all possible foci of infection12 (32.4)13 (33.3)27 (73.0)<0.001
      C-reactive protein and procalcitonin before antibiotics, n (%)9 (24.3)18 (46.2)31 (83.8)<0.001
      Appropriate antibiotic dosage, n (%)19 (51.4)23 (59.0)32 (86.5)0.004
      ≥ 2 hours of infusion for beta-lactam antibiotics, n (%)2 (5.4)2 (5.1)23 (62.2)<0.001
      VENTILATOR-ASSOCIATED PNEUMONIA
      ParameterPeriod 1(n = 13)Period 2(n = 11)Period 3(n = 13)p-value
      Presence of the radiological criteria for pneumonia, n (%)9 (69.2)9 (81.8)13 (100)0.030
      Gram staining and culture of non-directed bronchial lavage sample, n (%)10 (76.9)3 (27.3)10 (76.9)0.028
      Compliance of antibiotic dose and duration of infusion with the local guidelines, n (%)13 (100)8 (72.7)13 (100)0.021
      Clinical microbiology work-up: Bacteria isolated from any of the cultures that were drawn from the study patients were identified by matrix-assisted laser desorption/ionisation-time of flight (MALDI-TOF) mass spectrometry, and antibacterial susceptibility tests were performed according to

      The European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoint tables for interpretation of MICs and zone diameters. Version 10.0. 2020.

      criteria. Multidrug resistance was defined according to consensus criteria (
      • Magiorakos AP
      • Srinivasan A
      • Carey RB
      • Carmeli Y
      • Falagas ME
      • Giske CG
      • et al.
      Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance.
      ). Multidrug-resistant (MDR) Gram-negative bacilli were defined as being resistant to three or more antimicrobial classes (
      • Magiorakos AP
      • Srinivasan A
      • Carey RB
      • Carmeli Y
      • Falagas ME
      • Giske CG
      • et al.
      Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance.
      ).
      Follow-up parameters aside from the VAP and sepsis/septic shock management quality scores: The following data were recorded for all patients: crude mortality rates at 15 and 30 days after diagnosis of VAP or sepsis/septic shock; rates of isolation of Candida spp. or multidrug-resistant bacteria from any culture of the study patients within 1 month after cessation of antibacterial therapy; Clostridium difficile infection during and within 1 month following antibacterial therapy; and rediagnosis of VAP or sepsis/septic shock within 15 days following cessation of antibacterial therapy.

      Statistical analysis

      Results were analyzed using IBM SPSS for MacOS, version 27.0 (IBM Corp., Armonk, NY, USA). Descriptive statistical data were presented as mean, standard deviation, median (minimum–maximum), and percentage. The Pearson-chi-square test and Fisher-Freman-Halton test were used in the analysis of categorical variables; the Mann-Whitney U test and one-way analysis of variance (ANOVA) were used to analyze continuous variables.
      Multidrug-resistant Gram-negative bacilli infection rates were measured per 1000 patient-days per each study period using the hospital database. Meropenem, piperacillin-tazobactam, colistin, and glycopeptide (vancomycin and teicoplanin) consumption rates were measured in days of therapy (DOT) per 100 patient-days per each study period, using the hospital database. The mid-P exact test was used to calculate the 95% confidence interval for each period. OpenEpi version 3.01 was used for the analyses ().

      RESULTS

      Baseline patient characteristics

      In total, 113 patients diagnosed with sepsis/septic shock (37 in period 1, 39 in period 2, and 37 in period 3) were included in the study. The baseline demographics of patients were similar across the three periods for gender (51.4% male, 48.6% female; p = 0.103) and number of comorbidities (median: 3, range: 1–8; p = 0.092). The mean age of patients in period 1 was lower than those in periods 2 and 3 (62.9 years ± 16.2 years, 73.0 years ± 11.5 years, 71.6 years ± 15.2 years, respectively; p = 0.007).
      There were 37 patients with VAP in the study (13 in period 1, 11 in period 2, and 13 in period 3). These patients did not differ across the three periods in terms of gender (62.2% male, 37.8% female; p = 0.102), age (mean: 64.4 years ± 17.7 years; p = 0.67), or number of comorbidities (median: 3, range: 1–5; p = 0.25).

      Time to the first dose of antibiotic

      The median time from requesting ID consultation for suspected sepsis/septic shock to administration of the first dose of antibiotic was 179 minutes (IQR: 359 minutes) for period 1, 205 minutes (IQR: 334 minutes) for period 2, and 206 minutes (IQR: 342 minutes) for period 3 (p = 0.502) (Table 3). Delays were due to the time spent on different steps of routine clinical practice. The median time was shortened significantly for the steps from antibiotic ordered by the intensivist to ID approval (median 16 minutes in period 2 vs median 41 minutes in period 1; p = 0.009), and ID approval to the first dose administered (median 15 minutes in period 2 vs median 66 minutes in period 1; p < 0.001).
      Table 3Outcomes for patients with sepsis/septic shock and VAP.
      SEPSIS/SEPTIC SHOCK
      VariablePeriod 1(n = 37)Period 2(n = 39)Period 3(n = 37)p-value
      Mortality at day 14 from the start of treatment, n (%)15 (40.5)8 (20.5)9 (24.2)0.123
      Mortality at day 30 from the start of treatment, n (%)23 (62.2)16 (38.5)17 (45.9)0.159
      Candida isolation from patient within 1 month after treatment, n (%)9 (24.2)15 (39.5)16 (43.2)0.410
      Isolation of multidrug-resistant bacteria from patient within 1 month after treatment, n (%)6 (16.2)9 (23.1)8 (21.6)0.882
      C. difficile isolation from the patient within 1 month after antibiotics, n (%)1 (2.7)000.295
      Rediagnosis of sepsis within 15 days of completion of treatment period,
      Percentages calculated for the number patients alive in 15 days.
      n (%)
      7 (35.0)4 (13.8)5 (17.2)0.061
      Duration of ICU stay after ID consultation, days, median (range)16 (1–180)28 (1–305)28 (2–150)0.097
      Duration of hospital stay after ID consultation, days, median (range)21 (1–180)29 (1–310)30 (2–150)0.316
      VENTILATOR-ASSOCIATED PNEUMONIA
      VariablePeriod 1(n = 13)Period 2(n = 11)Period 3(n = 13)p-value
      Mortality at day 14 from the start of treatment, n (%)001 (7.6)1.000
      Mortality at day 30 from the start of treatment, n (%)1 (7.7)1 (9.1)3 (23.1)0.591
      Candida isolation from the patient within 1 month after antibiotics, n (%)2 (15.3)2 (18.2)1 (7.6)0.713
      Isolation of multidrug-resistant bacteria from the patient within 1 month after antibiotics, n (%)2 (15.3)1 (9.1)2 (15.3)1.000
      C. difficile enterocolitis within 1 month after antibiotics, n (%)0001.000
      Rediagnosis of VAP within 15 days after treatment,
      Percentages calculated for the number patients alive in 15 days.
      n (%)
      4 (30.8)2 (22.2)1 (8.3)0.360
      Duration of ICU stay after ID consultation, days, median (range)31 (9–245)56 (18–113)37 (5–78)0.080
      Duration of hospital stay after ID consultation, days, median (range)32 (12–245)58 (18–114)37 (9–78)0.058
      low asterisk Percentages calculated for the number patients alive in 15 days.
      The median time from requesting ID consultation for suspected VAP to administration of the first dose of the antibiotic was 294 minutes (IQR: 509 minutes) for period 1, 207 minutes (IQR: 270 minutes) for period 2, and 157 minutes (IQR: 867 minutes) for period 3 (p = 0.092) (Table 3). A significant shortening in median time needed from antibiotic ordered to ID approval was achieved in period 2 (7 minutes vs 160 minutes in period 1; p = 0.015), but this could not be maintained in period 3 (median 113 minutes).

      Quality control parameters

      The mean total quality-of-care scores for patients with sepsis/septic shock were 8.5 ± 3.76 in period 1, 9.91 ± 3.56 in period 2, and 14.94 ± 3.92 in period 3 (Table 4). There was a significant increase in total scores in period 3 (p < 0.001). Significant improvements were achieved in the management of these patients in period 3 in terms of requesting necessary diagnostic tests, and prolonged infusion of beta-lactam agents.
      Table 4Quality-of-care parameters with significant improvements in patients with sepsis/septic shock and VAP.
      VariablePeriod 1n (%)Period 2n (%)Period 3n (%)p-value
      Sepsis/septic shock
      Calculation of SOFA score1/37 (2.7)5/39 (12.8)14/37 (37.8)< 0.001
      ≥ two sets of blood cultures3/37 (8.1)10/39 (25.6)12/37 (32.4)0.034
      Urinalysis9/37 (24.3)14/39 (35.9)24/37 (64.9)0.001
      Urine culture11/37 (29.7)15/39 (38.5)25/37 (67.6)0.003
      Cultures from all possible foci of infection12/37 (32.4)13/39 (33.3)27/37 (73.0)< 0.001
      C-reactive protein and procalcitonin before antibiotics9/37 (24.3)18/39 (46.2)31/37 (83.8)< 0.001
      Appropriate antibiotic dosage19/37 (51.4)23/39 (59.0)32/37 (86.5)0.004
      ≥ 2 hours of infusion for beta-lactam antibiotics2/37 (5.4)2/39 (5.1)23/37 (62.2)< 0.001
      Ventilator-associated pneumonia
      Presence of radiological criteria for pneumonia9/13 (69.2)9/11 (81.8)13/13 (100)0.030
      Gram staining and culture of non-directed bronchial lavage sample10/13 (76.9)3/11 (27.3)10/13 (76.9)0.028
      Compliance of antibiotic dose and duration of infusion with local guidelines13/13 (100)8/11 (72.7)13/13 (100)0.021
      The median total quality-of-care scores for patients with VAP were 12 (range: 6–13) in period 1, 11 (range: 9–15) in period 2, and 15 (range: 12–16) in period 3 (Table 4). There was a significant increase in total scores in period 3 (p = 0.001). When each parameter was evaluated separately, significant improvement was achieved only for ordering microbiological tests in period 3.

      Patient outcome

      The durations of ICU and hospital stay, mortality rates, superinfections with Candida spp. or Clostridium difficile, emergences of multidrug-resistant bacteria, and recurrences of sepsis/septic shock or VAP within 15 days were similar across the three study periods (Table 5).
      Table 5Outcomes for patients with sepsis/septic shock and VAP.
      VariablePeriod 1n (%)Period 2n (%)Period 3n (%)p-value
      Sepsis/septic shock
      Mortality at Day 14 from start of treatment15/37 (40.5)8/39 (20.5)9/37 (24.3)0.123
      Mortality at Day 30 from start of treatment23/37 (62.2)16/39 (41.0)17/37 (45.9)0.159
      Candida isolation from patient within 1 month after treatment9/37 (28.1)15/39 (39.5)16/37 (43.2)0.410
      Isolation of multidrug-resistant bacteria from patient within 1 month after treatment6/37 (18.8)9/39 (23.7)8/37 (21.6)0.882
      C. difficile isolation from patient within 1 month after antibiotics1/37 (3.2)000.295
      Rediagnosis of sepsis within 15 days of completion of treatment period7/20 (35.0)4/29 (13.8)5/29 (17.2)0.061
      Duration of ICU stay after ID consultation, days, median (range)16 (1–180)28 (1–305)28 (2–150)0.097
      Duration of hospital stay after ID consultation, days, median (range)21 (1–180)29 (1–310)30 (2–150)0.316
      Ventilator-associated pneumonia
      Mortality at Day 14 from start of treatment0/13 (0)0/11 (0)1/13 (7.7)1.000
      Mortality at Day 30 from start of treatment1/13 (7.7)1/11 (9.1)3/13 (23.1)0.591
      Candida isolation from patient within 1 month after antibiotics2/12 (16.7)2/10 (20.0)1/10 (10.0)0.713
      Isolation of multidrug-resistant bacteria from patient within 1 month after antibiotics2/12 (16.7)1/10 (10.0)2/10 (20.0)1.000
      C. difficile enterocolitis within 1 month after antibiotics0/12 (0)0/10 (0)0/10 (0)1.000
      Rediagnosis of VAP within 15 days after treatment4/13 (30.8)2/9 (22.2)1/12 (8.3)0.360
      Duration of ICU stay after ID consultation, days, median (range)31 (9–245)56 (18–113)37 (5–78)0.080
      Duration of hospital stay after ID consultation, days, median (range)32 (12–245)58 (18–114)37 (9–78)0.058

      Rates of antimicrobial consumption and MDR Gram-negative bacilli

      The rate of meropenem consumption was significantly increased in period 3 (16.94, 95% CI: 15.2–18.83) compared with period 1 (10.79, 95% CI: 9.354–12.38) and period 2 (9.796, 95% CI: 8.553–11.17). Piperacillin-tazobactam consumption decreased in period 2 (4.898, 95% CI: 4.037–5.890) and period 3 (1.916, 95% CI: 1.375–2.603) compared with period 1 (7.045, 95% CI: 5.901–8.347). The consumption rates of colistin and glycopeptides were stable across all periods. The rate of MDR Gram-negative bacilli infections was similar for all study periods: 7.71 (95% CI: 0.439–12.620) in period 1, 5.44 (95% CI: 2.949–9.252) in period 2, and 6.56 (95% CI: 3.646–10.93) in period 3.

      DISCUSSION

      Implementation of an ASP has proven to be beneficial in several facets of patient care, leading to reductions in unnecessary use of broad-spectrum agents (
      • Buising KL
      • Thursky KA
      • Robertson MB
      • Black JF
      • Street AC
      • Richards MJ
      • et al.
      Electronic antibiotic stewardship — reduced consumption of broad-spectrum antibiotics using a computerized antimicrobial approval system in a hospital setting.
      ), emergence of resistance (
      • Rice LB.
      Antimicrobial stewardship and antimicrobial resistance.
      ), superinfections with Candida spp. (
      • Molina J
      • Peñalva G
      • Gil-Navarro MV
      • Praena J
      • Lepe JA
      • Pérez-Moreno MA
      • et al.
      Long-term impact of an educational antimicrobial stewardship program on hospital-acquired candidemia and multidrug-resistant bloodstream infections: a quasi-experimental study of interrupted time-series analysis.
      ), C. difficile-associated diarrhea (
      • Aldeyab MA
      • Kearney MP
      • Scott MG
      • Aldiab MA
      • Alahmadi YM
      • Darwish Elhajji FW
      • et al.
      An evaluation of the impact of antibiotic stewardship on reducing the use of high-risk antibiotics and its effect on the incidence of Clostridium difficile infection in hospital settings.
      ) and, finally, antibiotic costs (
      • Karanika S
      • Paudel S
      • Grigoras C
      • Kalbasi A
      • Mylonakis E.
      Systematic review and meta-analysis of clinical and economic outcomes from the implementation of hospital-based antimicrobial stewardship programs.
      ). An ASP can be implemented without an ID specialist if conditions necessitate it, but the success of the program is greater when an ID specialist is included (
      • Ostrowsky B
      • Banerjee R
      • Bonomo RA
      • Cosgrove SE
      • Davidson L
      • Doron S
      • et al.
      Infectious diseases physicians: leading the way in antimicrobial stewardship.
      ).
      Our study was unique because almost all patients with a suspected infection in our hospital routinely underwent consultation by the Department of Infectious Diseases, with the ID team following the patient actively on a daily basis until the infectious problem was resolved. Therefore, it was unclear whether an ID physician-directed active intervention could improve the quality of management of these patients.
      Our study observed a significant increase in quality-of-care scores during the intervention period (period 3) in patients with sepsis/septic shock as well as those with VAP. In general, the majority of significant improvements were achieved in ordering diagnostic tests and procedures. With regard to treatment, many of the parameters were satisfactory at baseline, with the exception of prolonged (> 2 hours) infusion of beta-lactam agents. This was not surprising, because the ID team decided on and recommended therapy approaches, such as antibiotic choice, dosage, and duration of treatment during daily practice in the ICU.
      There is no definitive diagnostic test for sepsis. The signs and symptoms of infection, host response, and organ dysfunction are important components for diagnosis. Thus, diagnostic workup includes a search for the presence and potential focus of infection (at least two sets of blood smears and cultures from potential sites of infection), an inflammatory response from the patient (increased levels of c-reactive protein, procalcitonin), as well as signs of organ failure (renal insufficiency, hepatic dysfunction, disseminated intravascular coagulation, acute respiratory distress syndrome) and increased lactate levels. These tests should be ordered early and the results interpreted correctly for a timely diagnosis. In our study, significant improvements were achieved in ordering the necessary diagnostic tests. However, at least two sets of blood cultures were obtained in only 32.4% of patients during active intervention. This was significantly higher compared with the retrospective period (8.1%; p = 0.034), but far from being optimal. Blood culture is important in detecting the causative agent, and obtaining the samples before antibiotic treatment begins increases culture positivity by 23% compared with that after the first antibiotic dose is given (
      • Scheer CS
      • Fuchs C
      • Gründling M
      • Vollmer M
      • Bast J
      • Bohnert JA
      • et al.
      Impact of antibiotic administration on blood culture positivity at the beginning of sepsis: a prospective clinical cohort study.
      ). Obtaining at least two sets of blood cultures instead of a single set increases the possibility of isolating the microorganism (
      • Lee A
      • Mirrett S
      • Reller LB
      • Weinstein MP.
      Detection of bloodstream infections in adults: how many blood cultures are needed?.
      ). SOFA scores were also missing in patient charts for almost all patients with suspected sepsis in period 1 (only one patient out of 37 had a SOFA score calculated prior to ID consultation). During active intervention, 37.8% had SOFA scores documented in their charts. It is clear that we need to develop and implement new methods in our hospital to achieve further improvements in these two parameters. This may include, but is not not limited to, reminders on electronic patient data in the hospital server, frequent educational meetings, and continued feedback to the primary physician.
      Diagnosis of VAP is hampered by the presence of comorbidities, such as cardiac failure, pulmonary thromboembolism, hypervolemia, or mucous plugs, as well as by the bad quality of portable chest X-rays, and the impracticability of transferring the unstable patient to the radiology department for CT scans. Deterioration of respiratory status, as evidenced by desaturation, increased need for oxygen, increased amount and change in consistency of respiratory secretions, and, finally, a new infiltrate in pulmonary imaging is usually sufficient to diagnose VAP. In our study, more patients had radiological evidence of pneumonia in period 3 compared with those in period 1 (13 out of 13 patients, 100% vs 9 out of 13 patients, 69.2%; p = 0.03). This could not be explained by changes in radiological practice — i.e. CT scans were performed in 23% of patients in both periods 1 and 3, while preliminary reporting of CT by the Radiology Department, and evaluation of portable chest-X rays by the ICU and ID teams, did not change throughout the study. Tests for microbiological documentation of the etiological agent in VAP were ordered in a similar manner in periods 1 and 3. The sharp decrease in period 2 could have been related to a temporary shortage of aspirate suction sets.
      The quality-of-care parameters for sepsis/septic shock treatment phases were mostly satisfactory across the three periods, including compliance with the local guidelines relating to antibiotic choice (64.9%, 66.7%, and 78.4%, respectively; p = 0.386), administration at appropriate dosing intervals (78.4%, 76.9%, and 91.9%, respectively; p = 0.174), and the monitoring of potential adverse events (100% for all three periods) as well as drug–drug interactions (100% for all three periods). These parameters were similar for all patients with VAP. This could be explained by the close collaboration between the ID and ICU teams during the management of patients with infection.
      There were no differences in the rates of colonization/infection with multidrug-resistant bacteria, C. difficile, or Candida spp. across the three periods. This may have been due to the relatively small sample size, but duration of treatment could also have played a role. When possible, duration of treatment was 7–10 days in more than 60% of patients with sepsis/septic shock or VAP. It has also been suggested that ASPs with successful reductions in targeted antibiotics may not see changes in rates of infection with resistant organisms for 2–6 years (
      • Tartof SY
      • Chen LH
      • Tian Y
      • Wei R
      • Im T
      • Yu K
      • Rieg G
      • Bider-Canfield Z
      • Wong F
      • Takhar HS
      • Qian L.
      Do inpatient antimicrobial stewardship programs help us in the battle against antimicrobial resistance?.
      ).
      Surviving Sepsis Campaign (SSC) guidelines recommend starting appropriate antibiotic treatment within the first hour to reduce mortality in sepsis (
      • Dellinger RP
      • Levy MM
      • Rhodes A
      • Annane D
      • Gerlach H
      • Opal SM
      • et al.
      Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012.
      ). In spite of the active intervention in period 3, the time to the first antibiotic dose in these patients (median 206 minutes) could not be shortened, with delays observed in each of the steps of this process. The median duration from ID consultation requested to ID recommendations made was 42 minutes in period 3. This included bedside evaluation by the ID resident, and consultation with the patient by an ID fellow or staff member. Recommendations were usually made verbally in the ICU or via calls, with a formal note written afterwards. The primary physician ordered the antibiotic after ID recommendation. This latter step took a median of 27 minutes. Approval of the antibiotic ordered was relatively fast (median: 12 minutes in period 3), but administration of the first dose after approval took 25 minutes in the ICU. Novel approaches could be helpful in preventing delays. For instance, a system in which the primary physician orders the first dose of the antibiotic without the need for ID approval could shorten the time required by a median of 81 minutes. In addition, the first dose of the antibiotic may be administered earlier by making the necessary arrangements within the internal dynamics of the ICU.
      Time needed from ID consultation to antibiotic administration in VAP patients was 420 minutes in period 3. Although there is no consensus on the optimal time to start antibiotics in VAP, early empirical treatment improves the outcome (
      • Ramirez P
      • Lopez-Ferraz C
      • Gordon M
      • Gimeno A
      • Villarreal E
      • Ruiz J
      • et al.
      From starting mechanical ventilation to ventilator-associated pneumonia, choosing the right moment to start antibiotic treatment.
      ). In our opinion, it should be possible to administer the appropriate antibiotic earlier in patients with VAP, using further interventions to shorten each step to administration of the first dose of the antibiotic, in a similar manner to the treatment of patients with sepsis/septic shock.
      There were no differences in 14-day or 30-day mortality in patients with sepsis or VAP. The impact of ASP on mortality remains unclear. Lindsay et al. reported no effect (
      • Lindsay PJ
      • Rohailla S
      • Taggart LR
      • Lightfoot D
      • Havey T
      • Daneman N
      • et al.
      Antimicrobial stewardship and intensive care unit mortality: a systematic review.
      ), whereas Okumura et al. showed a reduction in mortality (
      • Okumura LM
      • Silva MM
      • Veroneze I.
      Effects of a bundled antimicrobial stewardship program on mortality: a cohort study.
      ). It is difficult to assess the effect of a particular treatment modality on overall mortality in ICU patients because of a multitude of factors, such as the severity of the patient's condition, the presence of comorbidities, irreversible changes caused by infection, the resistance pattern of the infecting organism, or conditions in the ICU itself. The relatively small number of cases in our study also precluded any further interpretation.
      One of the main goals of an ASP is to limit the use of broad-spectrum antibiotics. In our study, the consumption rate of meropenem increased, while that of piperacillin-tazobactam decreased significantly, with the implementation of the ASP program in the ICUs. MDR Gram-negative bacilli infections are of concern in our ICUs. This might explain the initiation of broad-spectrum empirical antibacterial therapy in critically ill patients. One would expect higher rates of de-escalation after 72 hours and discontinuation of treatment after 7–10 days as a result of a significant improvement in the rate of requesting appropriate diagnostic tests, including cultures, in period 3, but our results failed to show any influence. This could be related to several factors, including the small number of patients in the study, and failure to obtain a clinical response in unstable patients in spite of appropriate antimicrobial treatment.
      Our study had several limitations. First, the number of patients included in the study was small. This was, in part, due to the slow turnover of ward beds, which prevented the timely transfer of patients from the ICU, the inclusion of patients only on weekdays, and failure to include patients with no documented timestamps for each of the steps to the start of empirical treatment, because of only verbal communication between the ID and ICU teams. Second, educational meetings were held only once with ICU personnel — at the beginning of period 2. It is unclear whether a more intense educational program might have had an extra impact. Finally, period 3 coincided with the COVID-19 pandemic. It was not possible to assess the impact of the pandemic on patient care, but negative consequences on the psychological status of the ICU personnel (
      • Kok N
      • van Gurp J
      • Teerenstra S
      • van der Hoeven H
      • Fuchs M
      • Hoedemaekers C
      • et al.
      Coronavirus disease 2019 immediately increases burnout symptoms in ICU professionals: a longitudinal cohort study.
      ), the recruitment of inexperienced healthcare workers in ICUs to cope with the the increased demand (which may have led to reluctance to order diagnostic tests), as well delays in patient evaluation, recommendations, and antibiotic approval by the ID consulting physician were all possible factors.

      CONCLUSION

      The implementation of ASPs in centers where antimicrobial management of ICU patients is largely controlled by infectious diseases specialists remains a feasible strategy for improving patient care.

      Funding

      This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

      Conflicts of interest statement

      All authors have no conflicts of interest to disclose.

      Ethical approval

      Informed consent was obtained for experimentation on human subjects. The study was approved by the local ethics committee (Feb 12, 2019; 16969557-356).

      References

        • Aldeyab MA
        • Kearney MP
        • Scott MG
        • Aldiab MA
        • Alahmadi YM
        • Darwish Elhajji FW
        • et al.
        An evaluation of the impact of antibiotic stewardship on reducing the use of high-risk antibiotics and its effect on the incidence of Clostridium difficile infection in hospital settings.
        J Antimicrob Chemother. 2012; 67: 2988-2996
        • Aslam B
        • Wang W
        • Arshad MI
        • et al.
        Antibiotic resistance: a rundown of a global crisis.
        Infect Drug Resist. 2018; 11: 1645-1658
        • Buising KL
        • Thursky KA
        • Robertson MB
        • Black JF
        • Street AC
        • Richards MJ
        • et al.
        Electronic antibiotic stewardship — reduced consumption of broad-spectrum antibiotics using a computerized antimicrobial approval system in a hospital setting.
        J Antimicrob Chemother. 2008; 62: 608-616
        • Conly J
        • Johnston B.
        Where are all the new antibiotics? The new antibiotic paradox.
        Can J Infect Dis Med Microbiol. 2005; 16: 159-160
        • Dellinger RP
        • Levy MM
        • Rhodes A
        • Annane D
        • Gerlach H
        • Opal SM
        • et al.
        Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012.
        Crit Care Med. 2013; 41: 580-637
      1. https://www.openepi.com

        • Karanika S
        • Paudel S
        • Grigoras C
        • Kalbasi A
        • Mylonakis E.
        Systematic review and meta-analysis of clinical and economic outcomes from the implementation of hospital-based antimicrobial stewardship programs.
        Antimicrob Agents Chemother. 2016; 60: 4840-4852
        • Kok N
        • van Gurp J
        • Teerenstra S
        • van der Hoeven H
        • Fuchs M
        • Hoedemaekers C
        • et al.
        Coronavirus disease 2019 immediately increases burnout symptoms in ICU professionals: a longitudinal cohort study.
        Critical Care Medicine. 2021; 49: 419-427
        • Lee A
        • Mirrett S
        • Reller LB
        • Weinstein MP.
        Detection of bloodstream infections in adults: how many blood cultures are needed?.
        Journal of Clinical Microbiology. 2007; 45: 3546-3548
        • Lindsay PJ
        • Rohailla S
        • Taggart LR
        • Lightfoot D
        • Havey T
        • Daneman N
        • et al.
        Antimicrobial stewardship and intensive care unit mortality: a systematic review.
        Clin Infect Dis. 2019; 68: 748-756
        • Magiorakos AP
        • Srinivasan A
        • Carey RB
        • Carmeli Y
        • Falagas ME
        • Giske CG
        • et al.
        Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance.
        Clin Microbiol Infect. 2012; 18: 268-281
        • Molina J
        • Peñalva G
        • Gil-Navarro MV
        • Praena J
        • Lepe JA
        • Pérez-Moreno MA
        • et al.
        Long-term impact of an educational antimicrobial stewardship program on hospital-acquired candidemia and multidrug-resistant bloodstream infections: a quasi-experimental study of interrupted time-series analysis.
        Clin Infect Dis. 2017; 65: 1992-1999
        • Okumura LM
        • Silva MM
        • Veroneze I.
        Effects of a bundled antimicrobial stewardship program on mortality: a cohort study.
        Braz J Infect Dis. 2015; 19: 246-252
        • Ostrowsky B
        • Banerjee R
        • Bonomo RA
        • Cosgrove SE
        • Davidson L
        • Doron S
        • et al.
        Infectious diseases physicians: leading the way in antimicrobial stewardship.
        Clin Infect Dis. 2018; 66: 995-1003
        • Ramirez P
        • Lopez-Ferraz C
        • Gordon M
        • Gimeno A
        • Villarreal E
        • Ruiz J
        • et al.
        From starting mechanical ventilation to ventilator-associated pneumonia, choosing the right moment to start antibiotic treatment.
        Critical Care (London, England). 2016; 20 (): 169
        • Rice LB.
        Antimicrobial stewardship and antimicrobial resistance.
        Med Clin North Am. 2018; 102: 805-818
        • Scheer CS
        • Fuchs C
        • Gründling M
        • Vollmer M
        • Bast J
        • Bohnert JA
        • et al.
        Impact of antibiotic administration on blood culture positivity at the beginning of sepsis: a prospective clinical cohort study.
        Clin Microbiol Infect. 2019; 25: 326-331
        • Tartof SY
        • Chen LH
        • Tian Y
        • Wei R
        • Im T
        • Yu K
        • Rieg G
        • Bider-Canfield Z
        • Wong F
        • Takhar HS
        • Qian L.
        Do inpatient antimicrobial stewardship programs help us in the battle against antimicrobial resistance?.
        Clin Infect Dis. 2020 Jul 15; : ciaa1004
      2. The European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoint tables for interpretation of MICs and zone diameters. Version 10.0. 2020.