JAC Advance Access originally published online on April 13, 2007
Journal of Antimicrobial Chemotherapy 2007 59(6):1148-1154; doi:10.1093/jac/dkm088
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Modified guidelines impact on antibiotic use and costs: duration of treatment for pneumonia in a neurosurgical ICU is reduced
1 Institute of Environmental Medicine and Hospital Epidemiology, Freiburg University Hospital, Hugstetter Str. 55, 79106 Freiburg, Germany 2 Department of Neurosurgery, Freiburg University Hospital, Freiburg, Germany 3 Department for Medical Microbiology and Hygiene, Freiburg University Hospital, Freiburg, Germany 4 Hospital Pharmacy, Freiburg University Hospital, Freiburg, Germany 5 Institute of Medical Microbiology and Hospital Epidemiology, Hannover School of Medicine, Germany 6 National Reference Centre for Surveillance of Nosocomial Infections, Germany 7 Institute of Hygiene and Environmental Medicine, Charité—University Medicine Berlin, Germany
* Corresponding author. Tel: +49-761-270-5487; Fax: +49-761-270-5485; E-mail: elisabeth.meyer{at}uniklinik-freiburg.de
Received 23 January 2007; returned 1 February 2007; revised 2 March 2007; accepted 4 March 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
Objectives: To evaluate the impact of an intervention to reduce the duration of antibiotic treatment for pneumonia in a neurosurgical intensive care unit (ICU). The usage of antibiotics and the resultant costs were examined using interrupted time series analysis while resistance and device-associated infection rates are also described.
Methods: In January 2004, revised guidelines for the use of antibiotics were implemented. As a consequence of this, the duration of antibiotic therapy for nosocomial pneumonia was reduced from 14 to 7 days, while for community-acquired pneumonia the period fell from 10 to 5 days. The effect on the antibiotic use density [AD; expressed as defined daily doses (DDD) per 1000 patient days (pd)] was calculated by segmented regression analysis of interrupted time series for the 24 months prior to (2002 and 2003) and after the intervention (2004 and 2005).
Results: The intervention was associated with a significant decrease in total AD from 949.8 to 626.7 DDD/1000 pd after the intervention. This was mainly due to reduced consumption of second-generation cephalosporins (–100.6 DDD/1000 pd), imidazoles (– 100.3 DDD/1000 pd), carbapenems (–33.3 DDD/1000 pd), penicillins with ß-lactamase inhibitor (–33.5 DDD/1000 pd) and glycopeptides (–30.2 DDD/1000 pd). Glycopeptide reduction might be associated with a significant decrease in the proportion of methicillin-resistant Staphylococcus aureus (8.4% before and 2.9% after the intervention). Similarly, total antibiotic costs/pd (
) showed a significant decrease from 13.16 /pd before to 7.31 /pd after the intervention. This is a saving of 5.85 /pd. The incidence of patients dying with pneumonia did not change significantly.
Conclusions: The most conservative estimate of segmented regression analysis over a 48 month period showed that halving the duration of treatment for pneumonia results in a reduction of over 30% in antibiotic consumption and costs. Because respiratory infections are most common in ICU patients, interventions targeting a reduction in the duration of treatment of pneumonia might be extremely worthwhile.
Keywords: intensive care units , duration of antibiotic treatment , costs , segmented regression analysis , interrupted time series
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
The overuse of antibiotics and poor compliance with infection control measures have been identified as the two major reasons for increasing antimicrobial resistance.1 Despite numerous guidelines from governmental and professional groups, there is broad evidence that antibiotics are prescribed inappropriately in up to 50% of cases.2 Goldmann et al. stated that ‘previous efforts have not worked because medical practice is locally driven, and national guidelines simply do not reflect or determine the systems of care and patterns of practice in individual hospitals’.1
Moreover, a Cochrane systematic review on interventions to improve antibiotic prescribing concluded that there is limited evidence of any improvement over time because of the fundamentally flawed methodology of the literature.3 The authors recommended the use of interrupted time series analysis as a robust and quasi-experimental approach to the evaluation of the longitudinal effects of interventions and segmented regression analysis for estimating the effects of those interventions.
It was in 2003 that Chastre et al. published their landmark paper on 8 vs 15 days of antibiotic therapy for ventilator-associated pneumonia.4 They found no clinical advantage in extending antimicrobial therapy to 15 days compared with 8 days in intensive care unit (ICU) patients who had developed proven ventilator-associated pneumonia.
This gave us a reason to revise the local guidelines in our neurosurgical ICU and to implement a shortened duration of treatment for pneumonia.
The aim of this study was to use appropriate methods, i.e. segmented regression analysis as proposed by Ramsay et al., to measure the impact of our intervention on antibiotic use and costs and to describe resistance and device-associated infection rates in the ICU for the 24 months prior to and after the intervention.3
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
Setting
The neurosurgical ICU monitored is a 12 bed unit for patients with intracranial pathologies, i.e. mainly head injuries, intracranial haemorrhages and intracranial tumours, but also for all post-operative patients following major intracranial surgery. Furthermore, if capacity allows, treatment is given to poly-trauma patients, patients who may have required resuscitation or patients who have suffered myocardial infarction or a lung embolism. Of the approximately 1300 patients in the year 2005, 1030 were ventilated with a mean ventilation length of 26 h. Length of stay ranged from 1 to 50 days. The nurse/patient ratio is 1:2. During the day, the patients were cared for by three physicians and by one physician at night. The anaesthetist in charge of the ICU is responsible for training the rotating neurosurgeons. All patients over 60 years of age are screened for methicillin-resistant Staphylococcus aureus (MRSA) (nose and groin swabs). Otherwise, no routine screening cultures are performed. Microbiology samples are taken when infection is suspected.
The written guidelines on empirical antibiotic treatment in the ICU were revised in December 2003 upon publication of the study by Chastre et al. and with respect to the local resistance situation.4 This change of empirical therapy was performed by a multidisciplinary team consisting of the intensive care specialist responsible for the ward and an infection control physician, and occasionally included also a microbiologist and a pharmacist.
In the ICU under study, before the intervention, second-generation cephalosporins were used for first-line empirical therapy of community-acquired and early-onset nosocomial pneumonia. Tazobactam or a carbapenem in combination with ciprofloxacin or an aminoglycoside was recommended for treatment of late-onset nosocomial pneumonia, and second-generation cephalosporins and metronidazole were recommended for pneumonia with suspected aspiration. The revision of the guidelines focused on reducing the duration of antibiotic treatment. In January 2004, antibiotic therapy for early- and late-onset nosocomial pneumonia was reduced from 14 days to 7 days and the duration of therapy for community-acquired pneumonia including pneumonia with suspected aspiration was reduced from 10–14 days to 5 days. Carbapenems were removed from the recommendations for late-onset nosocomial pneumonia because of the ICU's good susceptibility data for piperacillin against Pseudomonas aeruginosa. During the development and discussion in the revision process, it became clear that the prescribing practice and indication for imidazoles was too broadly applied in the ICU under study before the intervention.
The dissemination and implementation of the revised guidelines in January 2004 was straightforward enough because the experienced intensive care physician who was responsible for educating the rotating neurosurgeons provided clear leadership.
Monthly data on antimicrobial usage and costs of antibiotics were obtained from the computerized pharmacy database. Consumption, i.e. antimicrobial usage density (AD), was expressed as daily defined doses (DDD) and normalized per 1000 patient days (pd). One DDD is the standard adult daily dose of an antimicrobial agent for 1 day of treatment defined by the WHO (ATC/DDD index 2006, www.whocc.no).
Monthly resistance data were collected from the microbiology laboratory. The data were specified as resistant by the clinical laboratory using interpretive criteria recommended by the German Industrial standard (DIN).5 Copy strains—defined as an isolate of the same species showing the same susceptibility pattern throughout the period of one month in the same patient, no matter what the site of isolation—were excluded. The proportion of resistant isolates was calculated by dividing the number of resistant isolates by the total number of the isolates of this species tested against this antibiotic multiplied by 100.
Data on device-associated nosocomial infection rates (pneumonia, bloodstream infection and urinary tract infection) were obtained twice a week by trained medical staff from the infection control department. Device-associated nosocomial infections according to CDC definitions were reported to the ICU module of the Krankenhaus Infektions Surveillance System (KISS).6
A list of all patients who had died in the ICU was obtained from the hospital administration database and checked for whether pneumonia was diagnosed by the treating physician. Diagnosis of pneumonia was documented according to the International Classification of Diseases (ICD) for all patients and differentiated into ICU-acquired and non-ICU-acquired. If pneumonia was diagnosed on day 1 and day 2 after admission to the ICU, it was considered to be non-ICU-acquired. If pneumonia was diagnosed on day 3 or more after admission to the ICU, it was considered to be ICU-acquired.
We used segmented regression analysis of interrupted time series—a robust modelling technique that allows the analyst to estimate dynamic changes in various outcomes—to assess the changes of antibiotic use and costs before and after the implementation of the revised guidelines.
Level and slope are the two parameters which define each segment of a time series. The level is the value of the series at the beginning of a given time interval, the slope is the rate of change of a measure during a segment. An abrupt intervention effect constitutes a drop or jump in the level of the outcome after the intervention. A change in slope is defined by an increase or decrease in the slope of the segment after the intervention as compared with the segment preceding the intervention. It represents a gradual change of the outcome parameter during the segment. The method is described in greater detail by Wagner et al.7 and Ansari et al.8
For the statistical analysis of monthly antibiotic consumption and cost data, we looked at 24 time points before and 24 after the intervention. All antibiotic groups with an AD > 30 were considered in the regression analysis. Costs were not normalized for inflation. The analysis of the monthly use and costs of antibiotics was performed stepwise: use and cost variables were tested for normal distribution by Shapiro-Wilk test and for autocorrelation by Durbin Watson test.
The full segmented regression model included the baseline level and all level and trend changes; slope before intervention, change in level (at the moment of intervention) and change in slope after intervention were calculated. Therefore, non-significant variables were removed stepwise.
Differences in resistance rates, device-associated infection rates and incidence of patients who had died in the ICU (2002–03 versus 2004–05) were tested by the appropriate tests, i.e. by Fisher's exact test and/or by incidence density test. For these data, we could not use segmented regression analysis of interrupted time series because there were too few isolates and nosocomial infections to be reasonably calculated on a monthly basis. Therefore, resistance rates and device-associated infection rates are not an outcome parameter and neither are incidences of patients who had died in the ICU.
The significance level was P < 0.05 and all analyses were performed using SPSS 12.0 and EpiInfo 6.04.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
Antibiotic use
The intervention was associated with a significant decrease in total AD from 949.8 before to 626.7 DDD/1000 pd after the intervention (Figure 1 and Table 1). Applying the conservative and most parsimonious model, this results in a decrease in level of 323.1 DDD/1000 pd. If the full segmented regression model is used, the reduction is as much as 386 DDD/1000 pd. This decrease was due to a significant reduction in the use of second-generation cephalosporins (–100.6 DDD/1000 pd), imidazoles (–100.3 DDD/1000 pd), penicillins with ß-lactamase inhibitor and glycopeptides. Carbapenem use was not normally distributed and peaked in 4 out of 48 months. Calculation without the four wild points (AD > 100 DDD/1000 pd) showed a decrease in carbapenem AD after the intervention of 33.3 DDD/1000 pd. No significant change in level was seen in the consumption of quinolones and third-generation cephalosporins (Table 1).
|
|
Total antibiotic use density in all the antibiotic groups with an AD > 30 DDD/1000 showed no significant month-to-month change (slope) before and after the intervention.
Similarly to total antibiotic use, total antibiotic costs/pd () showed a significant change in level (– 5.86 /pd) after the intervention (Table 1). Before and after the intervention there was no significant change in slope of total antibiotic costs/pd. Application of the full segmented regression model revealed a non-significant slope of 0.15 /pd before the intervention, a significant estimated change in level of – 8.96 /pd and a non-significant change in slope of 0.04 /pd after the intervention (Figure 2). Yearly antibiotic costs/pd are depicted in Figure 3.
|
|
Resistance
Two-yearly resistance proportions of selected pathogens showed a significant decrease in the MRSA proportion after the intervention by Fisher's exact test: of 167 S. aureus isolates 8.4% were resistant in 2002–03, and of 208 S. aureus isolates only 2.9% were resistant in 2004–05 (Table 2). Although imipenem-resistant P. aeruginosa increased from 5.8% before the intervention to 15.2% after the intervention, this change was not statistically significant.
|
Device-associated infections and other parameters
Device-associated infection rates did not change significantly in the 2 years before or after the intervention (Table 3). Benchmarking data (median) from neurosurgical ICUs participating in the KISS system are for catheter-related urinary tract infections 5.0, for catheter-related bloodstream infection 1.4 and for ventilator-associated pneumonias 13.7 (8.9 in 2005). Therefore, the study ICU lay below the median of device-associated bloodstream infections and pneumonia, but above with respect to device-associated urinary tract infections. No second pneumonia was reported in patients with ventilator-associated pneumonia. The case mix index was 2.78 in 2003 (not available for 2002) and increased to 3.833 in 2004–2005. The utilization rate of ventilation, central venous catheters and urinary catheters increased significantly (P < 0.001). The absolute number and the percentage of patients who had died in the ICU increased likewise after the intervention (P < 0.001). However, the incidence (number of patients who had died with a diagnosis of ‘pneumonia’ per 100 patients) did not become statistically significant before or after the intervention (P = 0.262).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
The most conservative estimate of the segmented regression analysis over a 48 month period showed that changed antibiotic guidelines recommending a reduction in the duration of treatment of pneumonia are associated with a significant reduction in total antibiotic use and total antibiotic costs.
ICUs adopt many strategies to control antibiotic use and the spread of bacterial resistance and numerous publications emphasize the need for a multidisciplinary approach along with the implementation of antibiotic guidelines.1,2,9 Kollef et al. summarized antibiotic management strategies, and among other things listed formal protocols and guidelines, the use of narrow-spectrum antibiotics and shorter courses of antibiotic treatment as being effective measures.10 Generally, there is a consensus that all strategies aimed at reducing antibiotic resistance should be tailored to the needs of individual wards or ICUs, since antibiotic prescribing practice and the resistance situation may differ considerably even within a single hospital.
Respiratory infections in ICU patients are the most common infections, and countering them accounts for ~50% of all the antibiotics used.11 Therefore, in the ICU setting, preventing respiratory tract infections is the most effective method of reducing antibiotic consumption. The recommended duration of therapy for hospital-acquired pneumonia, including VAP, is traditionally long: a minimum of 7–10 days for susceptible Haemophilus or staphylococcal infections and 14–21 days for more typical cases.12 The approaches to limiting the duration of treatment that have been evaluated have included the employment of a 3 day short course of therapy compared with standard treatment for low-risk patients and the halving of the standard 2 week course of treatment.13,14 Although the study ICU did not face the problem of increasing antibiotic consumption or deteriorating resistance situation, the landmark paper by Chastre et al. provided the stimulus for revising existing antibiotic guidelines, shortening the duration of antibiotic treatment for pneumonia (both for community-acquired and nosocomial) and also provided the motivation to save costs and reduce the selection pressure of antibiotics. The most prominent reduction after the intervention was seen in the use of second-generation cephalosporins and imidazoles. In contrast, third-generation cephalosporins which were not used for pneumonia therapy in the ICU under study did not change significantly. They can therefore be taken as a control group to separate the effects of this intervention from other effects that may occur at the same time. However, glycopeptide use also decreased after the intervention even though it was not used for treatment of pneumonia. This might be associated with the significantly reduced MRSA proportion and isolates after the intervention. However, there was no change in infection control measures or screening policy over the 4 year period. Neither did quinolone use drop. The reduction in carbapenem consumption was not accompanied by a rise in the imipenem susceptibility of P. aeruginosa. This could have been expected based on reports of fairly rapid development of resistance and an association between imipenem use and the spread of resistance.15–17 Imipenem-resistant P. aeruginosa increased from 5.8% (before the intervention) to 15.2% (after the intervention) in the ICU under study. This is still lower than the mean resistance proportions of 21.4% reported from other German ICUs and might reflect the input of resistant pathogens from other hospital areas.18 Carbapenems are more widely used in German ICUs than they are in US ICUs, and this is also true for non-ICU areas.19 ICUs are not only subject to the export but also to the import of resistance. However, Cook et al. reported that even a hospital-wide reduction of 28% in the use of broad spectrum antibiotics did not lead to an improvement in the hospital antibiogram.20
With respect to device-associated infection rates and the incidence of patients who had died in the ICU with the ICD diagnosis ‘pneumonia’, the decrease in antibiotic consumption was not accompanied by any significant change in these rates. However, the crude mortality, i.e. the number of patients who had died in the ICU, increased significantly. Utilization rates for ventilator and other devices increased likewise. It might be hypothesized that sicker patients were treated in 2004 and 2005 if device utilization rates are taken as indirect parameters for the severity of illness.
The study has some limitations which must be taken into account. First, quantitative data per se cannot provide information about quality, i.e. about the appropriateness of antibiotic treatment given. This belongs to the domain of audits. Secondly, over the course of a longitudinal study, the composition of the study population may change. In the neurosurgical ICU mean length of stay did not change, whereas the number of patient days increased as well as the case mix index. This indicates that more and sicker patients were being treated and that the ICU had a higher turnover. Thirdly, the definition for device-associated nosocomial pneumonia changed in 2005 and calculation therefore had to be done by testing data before the intervention versus those for 2004. Fourthly, resistance data and device-associated infections could not be analysed by segmented regression analysis because there were too few isolates or infections per month. Therefore, only descriptive analysis could be provided. Fifthly, the calculation of the attributable mortality of pneumonia was beyond the scope of this study. Therefore, crude ICU mortality rates and the incidence of patients who had died with the diagnosis ‘pneumonia’ are adversely affected by the failure to adjust for severity of illness and the failure to distinguish between those patients who died as a result of the infection and those who died from other causes although they had the infection.
In conclusion, halving the duration of treatment for pneumonia resulted in a reduction of more than 30% in antibiotic consumption and costs. Antibiotic use and costs remained at a low level in the neurosurgical ICU after the intervention. Because respiratory infections are most common in ICU patients, interventions targeting a reduction in the duration of treatment of pneumonia might be most worthwhile.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
None to declare.
|
Acknowledgements |
|---|
We would like to thank S. Probst from the Hospital Pharmacy at Freiburg University Hospital for her helpful assistance with pharmacy data from the ICU under study and Deborah Lawrie-Blum for her help in preparing the manuscript. The ICU participated in SARI (Surveillance of Antimicrobial use and antimicrobial Resistance in German Intensive Care Units), a project that is supported by a grant from the Federal Ministry of Education and Research (01Kl 9907).
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionTransparency declarationsReferences |
|---|
1 Goldmann DA, Weinstein RA, Wenzel RP, et al. Strategies to prevent and control the emergence and spread of antimicrobial-resistant microorganisms in hospitals. A challenge to hospital leadership. JAMA (1996) 275:234–40.[Abstract]
2 John JF Jr, Fishman NO. Programmatic role of the infectious diseases physician in controlling antimicrobial costs in the hospital. Clin Infect Dis (1997) 24:471–85.[ISI][Medline]
3
Ramsay C, Brown E, Hartman G, et al. Room for improvement: a systematic review of the quality of evaluations of interventions to improve hospital antibiotic prescribing. J Antimicrob Chemother (2003) 52:764–71.
4
Chastre J, Wolff M, Fagon JY, et al. Comparison of 8 versus 15 days of antibiotic therapy for ventilator-associated pneumonia in adults: a randomized trial. JAMA (2003) 290:2588–98.
5 DIN 58940–4. Medical Microbiology—Methods for the Determination of Susceptibility of Pathogens (Except Mycobacteria) to Antimicrobial Agents. Part 4: Evaluation Classes of the Minimum Inhibitory Concentration (1998) Berlin: Beuth Verlag. 1–11.
6 Gastmeier P, Geffers C, Sohr D, et al. Five years working with the German nosocomial infection surveillance system (Krankenhaus Infektions Surveillance System). Am J Infect Control (2003) 31:316–21.[CrossRef][ISI][Medline]
7
Ansari F, Gray K, Nathwani D, et al. Outcomes of an intervention to improve hospital antibiotic prescribing: interrupted time series with segmented regression analysis. J Antimicrob Chemother (2003) 52:842–8.
8 Wagner AK, Soumerai SB, Zhang F, et al. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther (2002) 27:299–309.[CrossRef][ISI][Medline]
9 Cornaglia G, Hryniewicz W, Jarlier V, et al. European recommendations for antimicrobial resistance surveillance. Clin Microbiol Infect (2004) 10:349–83.[CrossRef][ISI][Medline]
10 Kollef MH, Micek ST. Strategies to prevent antimicrobial resistance in the intensive care unit. Crit Care Med (2005) 33:1845–53.[CrossRef][ISI][Medline]
11
Bergmans DC, Bonten MJ, Gaillard CA, et al. Indications for antibiotic use in ICU patients: a one-year prospective surveillance. J Antimicrob Chemother (1997) 39:527–35.
12 Anon. Hospital-acquired pneumonia in adults: diagnosis, assessment of severity, initial antimicrobial therapy, and preventive strategies. A consensus statement, American Thoracic Society, November 1995. Am J Respir Crit Care Med (1996) 153:1711–25.[ISI][Medline]
13
Fagon JY, Chastre J, Wolff M, et al. Invasive and noninvasive strategies for management of suspected ventilator-associated pneumonia. A randomized trial. Ann Intern Med (2000) 132:621–30.
14
Singh N, Rogers P, Atwood CW, et al. Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit. A proposed solution for indiscriminate antibiotic prescription. Am J Respir Crit Care Med (2000) 162:505–11.
15
Lepper PM, Grusa E, Reichl H, et al. Consumption of imipenem correlates with ß-lactam resistance in Pseudomonas aeruginosa. Antimicrob Agents Chemother (2002) 46:2920–5.
16 Livermore DM. The impact of carbapenemases on antimicrobial development and therapy. Curr Opin Investig Drugs (2002) 3:218–24.[Medline]
17 Livermore DM. Multiple mechanisms of antimicrobial resistance in Pseudomonas aeruginosa: our worst nightmare? Clin Infect Dis (2002) 34:634–40.[CrossRef][ISI][Medline]
18 Meyer E, Schwab F, Jonas D, et al. Temporal changes in bacterial resistance in German intensive care units, 2001–2003: data from the SARI (surveillance of antimicrobial use and antimicrobial resistance in intensive care units) project. J Hosp Infect (2005) 60:348–52.[CrossRef][ISI][Medline]
19 Meyer E, Schwab F, Jonas D, et al. Surveillance of antimicrobial use and antimicrobial resistance in intensive care units (SARI): 1. Antimicrobial use in German intensive care units. Intensive Care Med (2004) 30:1089–96.[CrossRef][ISI][Medline]
20
Cook PP, Catrou PG, Christie JD, et al. Reduction in broad-spectrum antimicrobial use associated with no improvement in hospital antibiogram. J Antimicrob Chemother (2004) 53:853–9.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Vernaz, H. Sax, D. Pittet, P. Bonnabry, J. Schrenzel, and S. Harbarth Temporal effects of antibiotic use and hand rub consumption on the incidence of MRSA and Clostridium difficile J. Antimicrob. Chemother., September 1, 2008; 62(3): 601 - 607. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||