Reassessment of intravenous antibiotic therapy using a reminder or direct counselling

Transcription

1 J Antimicrob Chemother 2010; 65: doi: /jac/dkq018 Advance publication 5 February 2010 Reassessment of intravenous antibiotic therapy using a reminder or direct counselling Philippe Lesprit 1 *, Caroline Landelle 1, Emmanuelle Girou 1 and Christian Brun-Buisson 2 1 Unité de Contrôle, Epidémiologie et Prévention de l Infection, Université Paris 12, Assistance Publique-Hôpitaux de Paris, Hôpital Henri Mondor, Créteil, France; 2 Service de Réanimation Médicale, Université Paris 12, Assistance Publique-Hôpitaux de Paris, Hôpital Henri Mondor, Créteil, France *Corresponding author. Tel: þ ; Fax: þ ; philippe.lesprit@hmn.aphp.fr Received 11 August 2009; returned 15 October 2009; revised 5 January 2010; accepted 11 January 2010 Objectives: Encouraging reassessment of intravenous antibiotic therapy at days 3 4 is an important step in the management of patients and may be done by delivering a questionnaire or through systematic infectious disease physician (IDP) advice to prescribers. Patients and methods: In this before-and-after study, prescriptions of 13 selected intravenous antibiotics from surgical or medical wards were screened from a computer-generated listing and prospectively included. Three strategies were compared over three consecutive 8 week periods: conventional management by the attending physician (control group); distribution of a questionnaire to the physician (questionnaire group); or distribution of the questionnaire followed by IDP advice (Q-IDP group). The primary outcome was the percentage of modifications of antibiotic therapy at day 4, including withdrawal of therapy, de-escalation, oral switch or reducing the planned duration of therapy. Results: Overall, 402 prescriptions were included. At day 4, 48.9% and 54.5% of prescriptions were modified in the control and questionnaire groups, respectively (P¼0.35). In contrast, more prescriptions (66.2%) were modified in the Q-IDP group as compared with the control group (P¼0.004). Stopping therapy in the absence of apparent bacterial infection occurred significantly more often in the Q-IDP group than in the control (P, ) or questionnaire groups (P¼0.002). Conclusions: This study shows a modest impact of only distributing a questionnaire aimed at reminding physicians to reassess therapy, whereas systematic IDP intervention improves the modification rate. Keywords: antibiotic review, infectious disease advice, intravenous to oral antibiotic switch Introduction Early reassessment of antibiotic therapy is an important step in the management of patients and has been promoted in recent guidelines as an effective measure to improve antimicrobial use in hospitals. 1 5 Between days 2 and 4 after the start of therapy, empirical therapy may be modified for several reasons. Microbiological results including pathogen identification and susceptibility are available at that time, allowing narrowing of the spectrum of drug(s) initially prescribed. Clinical reassessment of patients may also lead to stopping or modifying the planned duration of therapy. Finally, a switch from parenteral to oral therapy can be decided at this time. All these interventions have been promoted in antimicrobial stewardship programmes, to improve the quality of care, and help contain bacterial resistance and hospital antibiotic costs. 3 5 However, studies have shown that early reassessment of antibiotic therapy is often overlooked by the attending physicians. 6 Many factors could explain this finding, including time constraints, rapid rotation of physicians in charge of the patients, insufficient training or reluctance to modify an apparently efficacious empirical therapy. In order to encourage reassessment of intravenous antibiotic therapy in their hospital, Senn et al. 7 developed a short questionnaire mailed to the physicians in charge of the patient at days 3 4. This questionnaire emphasized key points for reassessment and its impact was evaluated in a randomized trial. As compared with the control group, time to modification of therapy was 14% shorter in the intervention group. In our hospital, we recently evaluated the impact of a computer-prompted post-prescription review on antibiotic use. After screening of prescriptions, this alert allowed direct counselling by an infectious disease physician (IDP) in one half of the # The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org 789

2 Lesprit et al. antibiotic courses and was well accepted by the majority of physicians. 8 Because of this encouraging result, we decided to evaluate the potential impact of a questionnaire similar to that proposed by Senn et al. 7 using the computer-generated alert system in place in our hospital. Our primary objective was to assess the impact of delivering this questionnaire to prescribing physicians and to compare the modification of prescriptions when the questionnaire was delivered alone or was followed by IDP counselling at day 4. Patients and methods Hospital setting This was a non-randomized open trial conducted over 6 months in a 950 bed general university hospital. The guidelines issued by the hospital s Anti-infective Drugs Committee are distributed hospital-wide in pocket format and are also available on the Intranet system. Every 6 months, all staff and junior physicians of each ward are offered educational sessions about antibiotic prescribing. There is one IDP who provides advice on an on-call basis in all hospital wards and performs post-prescription review of 15 selected antibiotics in surgical and medical wards. This review is based on a computer-generated alert system that has been described in detail elsewhere. 8 The same system was used for the purposes of the present study. However, the post-prescription review performed by the IDP was stopped 4 months before starting the study, at the time when new residents started on hospital wards. Overall, the medical and surgical wards included 150 physicians (half of which were residents) who were involved in the care of patients. Usually, antibiotic prescriptions were made by the residents. Inclusion and exclusion criteria All adult patients hospitalized in surgical or medical wards were screened 5 days a week by the IDP using the computer-generated alert. To be eligible, patients had to be treated with an intravenous antibiotic for 3 4 days. Only intravenous antibiotics included in the computer-generated alert were reviewed (including amoxicillin/clavulanate, piperacillin/tazobactam, cefotaxime, ceftriaxone, cefepime, ceftazidime, imipenem, gentamicin, ofloxacin, ciprofloxacin, levofloxacin, vancomycin and teicoplanin). Patients in whom modification of the antibiotic therapy occurred within the first 3 4 days of initiation of the intravenous antibiotic course were not included. Interventions Eligible patients were sequentially included in three consecutive groups each over a period of 8 weeks: first, conventional management by the attending physician without questionnaire delivery nor IDP postprescription review (control group); second, direct delivery of the questionnaire by a medical student to the physician in charge of the patients on the morning of day 4 (questionnaire group); and, third, direct delivery of the questionnaire on the morning of day 4 by a medical student, followed by systematic review of the prescription by the same IDP in the afternoon (Q-IDP group). This questionnaire was adapted from Senn et al. 7 and asked three questions regarding possible adaptation of antibiotic therapy on day 4 (Figure 1). Completed questionnaires were collected by the medical student 24 h later. If the physician had not yet completed it at that time, no further attempt was made. When the attending physician was absent, the questionnaire was left in the medical chart of the patient for 1 day. In the third period, IDP recommendations to modify the antibiotic regimen were provided orally to the attending physician when appropriate. When direct counselling to the physician could not be made, recommendations were written in the medical chart. These could be overridden and no further attempt was made if recommendations were not followed. However, compliance with recommendations was recorded. IDP advice solicited by ward physicians was allowed in the three periods of the study. There were two changes of the prescribing residents during the three periods: first, residents of surgical wards who made antibiotic prescriptions rotated between the first and second periods; and, second, new residents (80% of residents of medical and surgical wards) entered the study at the start of the third period. Data collection Data collected included the demographic characteristics of patients, the ward where the prescription was initiated, presumed indication for antibiotic therapy, microbiological results available at day 4, requested consultation by the IDP and information on prescriptions of the 13 antibiotics selected, including the type, modalities of administration, spectrum of antibacterial activity and planned duration of treatment. As described previously, the electronic prescription system in place in our hospital does not contain any pre-specified antibiotic duration and cannot influence the planned duration of therapy. 8 Modifications of the initial antibiotic regimen prescribed and occurring from day 4 until the end of therapy were retrieved from the computerized prescription system. The total duration of antibiotic therapy included both intravenous and oral courses, including any switch of antibiotic because of toxicity for a given episode of infection. Replacement by another antibiotic because of a new infection was not considered. Outcomes The primary outcome of the study was modification of the initial therapy at day 4, defined as one or more of the following modifications: discontinuation of antibiotic therapy; switch to oral therapy; de-escalating therapy according to the clinical situation or by targeting documented pathogens; and reduction of the planned duration of therapy. 9 Secondary outcomes included any modification of initial therapy (including oral switch, de-escalation, withdrawal of therapy or reducing its duration) from day 4 and median duration of intravenous therapy. Statistical analysis The sample size was estimated on the basis of pre-study observations at our hospital that 50% of intravenous antibiotic prescriptions were modified by the physicians at day 4. 8 One hundred and three prescriptions were required in each group to reach 80% power of demonstrating an increase in the day 4 modification rate to 70% in one of the two intervention groups. This anticipated improvement of 20% was based on the results of the effect of providing usage feedback to prescribing physicians. 10 An 8 week period was the estimated time necessary to achieve the calculated sample size in each group. Categorical variables were compared using the Pearson x 2 test or the Fisher exact test, as appropriate. Continuous variables were expressed as the median with interquartile range (IQR) and were compared using the Mann Whitney U-test. Physicians responses and modifications of therapy ordered were compared using the non-parametric McNemar matched pairs test. All statistical tests were two-tailed and statistical significance was set at Analyses were performed by SPSS software, version 11.5 (SPSS). Results A total of 402 prescriptions ordered for 342 patients were assigned to the observational or intervention groups. Patients 790

3 Reassessment of antibiotic therapy JAC Figure 1. Questionnaire used in the study (adapted from Senn et al. 7 ). characteristics were similar in the three groups (Table 1). Prescriptions analysed originated from medical (60%) or surgical (40%) wards. The most common infections treated were pneumonia, intra-abdominal infections and urinary tract infections. Microbiologically documented infections accounted for 54% of the prescriptions. A majority of prescriptions were of broad-spectrum penicillins with inhibitors or third-generation cephalosporins. In the control group, IDP advice was rarely requested by the physicians in charge of the patients. Although the questionnaire did not recommend requesting an IDP advice to consider modifying prescriptions, its distribution was associated with a significant 2-fold rise of IDP calls in the questionnaire group in comparison with the control group (35.6% versus 18.2%, P¼0.001). IDP counselling was performed by only one physician throughout the study and was mainly based on reviewing patient s charts, often with the attending physician. Physical examination of patients was performed in 30% of cases. Overall, IDP counselling needed a mean of 10 min for each prescription review. Comparison of the modifications of antibiotic prescriptions at day 4 In the control group, one half of the prescriptions were modified by the physician in charge, mostly via a reduction of planned 791

4 Lesprit et al. Table 1. Characteristics of the patients P value d C group a (n¼137) Q group b (n¼132) Q-IDP group c (n¼133) Q vs C Q-IDP vs C Q-IDP vs Q Median age (IQR) 65 ( ) 62 (52 78) 61 (50 76) Ward category, n (%) medical 76 (55.5) 83 (62.9) 82 (61.7) surgical 61 (44.5) 49 (37.1) 51 (38.3) Infection, n (%) pneumonia 22 (16.1) 26 (19.7) 26 (19.5) gastrointestinal 25 (18.2) 17 (12.9) 18 (13.5) urinary tract infection 21 (15.3) 23 (17.4) 15 (11.3) surgical site infection 16 (11.7) 16 (12.1) 16 (12.0) bacteraemia 12 (8.8) 13 (9.8) 17 (12.8) fever or sepsis 15 (10.9) 7 (5.3) 16 (12.0) skin and soft tissue 12 (8.8) 12 (9.1) 7 (5.3) others 14 (10.2) 18 (13.6) 18 (13.5) Pathogen identified, n (%) 71 (51.8) 77 (58.3) 68 (51.1) Staphylococcus sp. 15 (21.1) 9 (11.7) 9 (13.2) Streptococcus sp. 7 (9.9) 6 (7.8) 4 (5.9) Enterobacteriaceae 28 (39.4) 35 (45.5) 22 (32.4) polymicrobial infection 16 (22.5) 19 (24.7) 21 (30.9) others 5 (7.0) 8 (10.4) 12 (17.6) Antibiotic initially prescribed, n (%) amoxicillin/clavulanate 48 (35.0) 46 (34.8) 38 (28.6) piperacillin/tazobactam 16 (11.7) 17 (12.9) 22 (16.5) third-generation cephalosporins 39 (28.5) 36 (27.3) 33 (24.8) imipenem 7 (5.1) 10 (7.6) 6 (4.5) fluoroquinolones 7 (5.1) 4 (3.0) 10 (7.5) vancomycin 9 (6.6) 7 (5.3) 17 (12.8) gentamicin 11 (8.0) 12 (9.1) 7 (5.3) Median days of planned duration of therapy (IQR) 7 (7 8) 7 (7 10) 7 (7 10) a Control group. b Questionnaire group. c Questionnaire plus IDP advice group. d P value for comparison of distribution of wards, infections, pathogens or antibiotics between the three groups. duration of therapy (Table 2). Distribution of the questionnaire alone did not significantly increase the modification rate in comparison with the control group. In the Q-IDP group, most of the modifications (n¼61, 69.3%) occurred following IDP advice. Prescriptions were more often modified in the Q-IDP group in comparison with the control group (P¼0.004). Therapy was stopped (in the absence of apparent bacterial infection) significantly more often in the Q-IDP group than in the control group (P,0.0001). As compared with the questionnaire group, the modification rate after IDP counselling was higher but not significantly so (P¼0.053). A discontinuation of therapy was more often suggested by the IDP (P¼0.002); as a result, a reduction of planned duration was less often suggested (P¼0.019). Similar results were obtained when the analyses were restricted to the subgroup of physicians who answered the questionnaire (119 physicians in the questionnaire group and 95 physicians in the Q-IDP group; Table 3). Comparison of the modifications of antibiotic prescriptions until the end of therapy The overall rate of modifications slightly increased to 60.6% in the control group, suggesting that few physicians decided to modify their prescriptions later than day 4. A similar rate of overall modifications was observed in the period when physicians received the questionnaire. However, these rates remained significantly lower than when prescriptions were systematically reviewed by the IDP (P¼0.001) (Table 2). Comparison of median duration of therapy The planned duration initially ordered by physicians was similar in the three groups (Table 1). There was a non-significant decrease in the overall duration of therapy in the Q-IDP group, as compared with the control group. However, no reduction of 792

5 Reassessment of antibiotic therapy JAC Table 2. Modifications of initial antibiotic therapy from day 4 onward P value Antibiotic course C group a (n¼137) Q group b (n¼132) Q-IDP group c (n¼133) Q vs C Q-IDP vs C Q-IDP vs Q Modification at day 4 d, n (%) 67 (48.9) 72 (54.5) 88 (66.2) discontinuation 7 (5.1) 11 (8.3) 29 (21.8) 0.29 < switch to oral therapy 27 (19.7) 23 (17.4) 34 (25.6) de-escalating 6 (4.4) 10 (7.6) 14 (10.5) reduction of planned duration 45 (32.8) 47 (35.6) 30 (22.6) Modification after day 4 e, n (%) 16 (11.7) 9 (6.8) 18 (13.5) Modification from day 4 e, n (%) 83 (60.6) 81 (61.4) 106 (79.7) Median days of total duration (IQR) 8 (6 11) 7 (5 11) 7 (4 11) Median days of intravenous duration (IQR) 6 (4 8) 5 (4 8) 4 (4 7.5) Switch to oral therapy from day 4 until the end of therapy, n (%) 56 (40.9) 40 (30.3) 52 (39.1) a Control group. b Questionnaire group. c Questionnaire plus IDP advice group. d Including any of the following events: discontinuation; switch to oral therapy; de-escalating therapy according to the clinical situation or by targeting documented pathogens; and reduction of the planned duration of therapy. e Including any of the following events: switch to oral therapy; de-escalating therapy according to the clinical situation or by targeting documented pathogens; withdrawal of therapy; and reduction of the planned duration of therapy. Table 3. Physicians responses to the questionnaire and modifications actually ordered at day 4 Questionnaire group (n¼119) Q-IDP group a (n¼95) Question response yes b, n (%) modification ordered c, n (%) P value response yes b, n (%) modification ordered d, n (%) P value One or more modification e 59 (49.6) 66 (55.5) (61.1) 62 (65.3) 0.56 Discontinuation 13 (10.9) 11 (9.2) (13.7) 24 (25.3) Switch to oral therapy 30 (25.2) 21 (17.6) (29.5) 28 (29.5) 0.99 De-escalating therapy 13 (10.9) 10 (8.4) (17.9) 12 (12.6) 0.38 Reduction of planned duration of therapy 11 (9.2) 41 (34.5) < (6.3) 15 (15.8) a Questionnaire plus IDP advice group. b Positive response given by the physician to the questionnaire. c Modification of the prescription ordered by the physician on day 4, after responding to the questionnaire. d Modification of the prescription ordered by the physician on day 4, after responding to the questionnaire and receiving IDP advice. e Including any of the following events: discontinuation; switch to oral therapy; de-escalating therapy according to the clinical situation or by targeting documented pathogens; and reduction of the planned duration of therapy. duration of intravenous therapy was observed in any intervention group (Table 2). Physicians responses to the questionnaire Among 265 questionnaires distributed to the physicians in charge of the patients in the questionnaire or Q-IDP groups, 231 (87.2%) were collected and 214 (80.8%) were completed by the attending physician. Among the reasons provided for continuing therapy, physicians answered that a total of 94 (43.9%) prescriptions were microbiologically documented. However, analysis of the available data by the IDP the same day found that a pathogen had been significantly more often identified, corresponding to 109 prescriptions (50.9%) (P¼0.03). Analysis of physicians responses at day 4 showed that they were frequently reluctant to consider the following modifications: stopping therapy; de-escalating; or reducing the planned duration (Table 3). In the questionnaire group, comparison of physicians responses to modifications ordered soon after responding to the questionnaire showed no improvement in the rate of discontinuation, de-escalation or switching to oral therapy. In contrast, distribution of the questionnaire seemed to encourage physicians to reduce the planned duration of therapy. Stopping therapy or reducing the planned duration was rarely considered by physicians in the Q-IDP group, but IDP intervention increased the rate of these modifications. 793

6 Lesprit et al. Discussion This study evaluated two different ways to promote early modification of intravenous antibiotic therapy, i.e. simply distributing a questionnaire to remind prescribers of the opportunity for reassessing therapy or distributing a questionnaire followed by IDP counselling. We found that only distributing the questionnaire to physicians had a limited impact on modification of the therapy actually ordered, except for a non-significant reduction of the median duration of therapy by 1 day. In contrast, the modification rate was significantly higher when the questionnaire distribution was followed by recommendations by an IDP. This evaluation took place as part of the antimicrobial stewardship programme developed in our hospital, using a prospective audit with intervention and feedback. This core strategy has been recommended in guidelines because it allows a review of antimicrobial prescriptions in terms of indication, drug selection, dose, route and duration. 2 4 Intravenous antibiotics prescriptions are an ideal target to optimize antibiotic use, because interventional studies aiming at influencing their use have proved cost effective. 11 In a randomized study where an intervention by an infectious disease fellow and a clinical pharmacist was compared with a control group, 50% of antibiotic courses were found to justify intervention after 3 days of therapy. 12 Interventions were associated with a reduction in antibiotic charges without altering the clinical and microbiological responses rates. Reassessment of therapy by the attending physician may be encouraged in different ways. It can be made directly by a member of the antimicrobial stewardship programme as unsolicited feedback, but this requires a continuous effort to be effective and many hospitals may lack the resources to implement it. Another way is to implement questionnaires or computergenerated reminders designed to foster reassessment of therapy by physicians. These tools have shown some efficacy for specific objectives, such as encouraging an early switch to oral therapy, or improving intraoperative antibiotic prophylaxis or vancomycin use. 7,13 15 In the study by Senn et al., 7 a computer-generated questionnaire was sent to prescribers at 3 days of intravenous therapy to remind them of the opportunity for modification of therapy. The benefit of this intervention was rather modest. In fact, the ratio for modification with distribution of the questionnaire was only close to statistical significance, resulting in a 14% reduction in the time elapsed until a modification of therapy occurred. In this study, there was no beneficial effect of an intervention limited to distributing to physicians a questionnaire encouraging reassessment of intravenous antibiotic therapy at day 4; indeed, the rates of modifications ordered at that time were comparable in the control and questionnaire groups. However, some effect of distributing the questionnaire was observed on the reduction of total duration of therapy, but this result should be interpreted with caution because of the before-and-after design of the study. Although not reaching statistical significance, distribution of the questionnaire may have had a progressive and latent effect on the third period by contributing to educate physicians to reassess and streamline therapy. However, this effect must have been limited, because of the rotation of residents (mostly to other hospitals) just before the third period. Many factors could explain our findings. First, one half of prescriptions were modified by the physicians in the control period, suggesting that routine reassessment of prescription was already quite common in our hospital. Moreover, analysis of the rate of modification after day 4 showed that modification ordered by the physicians increased to 60.6%. Second, it is of interest that distribution of the questionnaire was associated with a 2-fold increase in the requests for IDP advice, suggesting that physicians had some difficulties and needed counselling to reassess ongoing therapies. Lastly, prescription review by the IDP was associated with a significantly higher rate of modifications at day 4 or at the end of follow-up. Of note, the main modification following IDP intervention was an early discontinuation of antibiotic courses that were considered no longer necessary. The impact obtained through IDP counselling was therefore of interest, because reduction of the duration of antibiotic therapy is one of the most important steps to improve antimicrobial use in hospital. As previously described, IDP counselling performed by only one physician using the prescription review in place in our hospital used limited resources and was associated with an acceptable working time. 8 Although the modification rate from day 4 until the end of therapy was significantly higher during IDP review and rose up to 80%, it also suggests that 20% of the prescriptions analysed were either fully adequate or not easily amenable to modification. The main limitation of our study was its non-randomized before-and-after design. Because we did not perform a crossover study with washout periods, we could not exclude that the previous two periods may have gradually increased the compliance to IDP intervention to a higher level in the third period. However, analysis of physicians responses to the questionnaire between the second and third periods did not show any positive change. This suggests that their behaviour towards antibiotic reassessment was not influenced by the previous period. Another limitation was the short period of the intervention, so we could not exclude that the effect of the questionnaire may have been more pronounced had the intervention period been longer. Comparison of physicians responses to the questionnaire with the modification of therapy actually ordered provided some interesting observations. First, it is noteworthy that the response rate to the distributed questionnaire was 80% in our study, a rate higher than the 70% rate recorded by Senn et al. 7 when using a mailed questionnaire. Second, an unexpected finding was that physicians less often had knowledge of bacterial documentation than the IDP. The reason for this is not clear, because IDP evaluation was performed the same day and used the same tool. However, this discrepancy could explain in part the low rate of de-escalating therapy recorded in the control and questionnaire groups. Similarly, physicians often appeared reluctant to perform early switch to oral therapy. This could be targeted by a specific intervention within a computerized prescribing system, by implementing a mandatory response to the questionnaire at the time of antibiotic order renewal, similar to that described by Fischer et al. 13 On the other hand, a difference was observed with regard to the planned duration of the therapy. As shown by comparison of physicians responses and the modifications they ordered, prescribers were rarely prone to stop therapy. With regard to these possible modifications, IDP interventions were clearly more effective than when only distributing the questionnaire. 16,17 To conclude, an intervention based on only distributing a questionnaire reminder was not associated with a significant 794

7 Reassessment of antibiotic therapy JAC increase of the rate of early modifications of intravenous therapy and was less effective than direct counselling by an IDP. However, reminding prescribers of the early opportunity for modification of therapy at day 4 via a questionnaire may be an effective component of an antimicrobial stewardship programme, because of its educational role. Implementation of the questionnaire as a mandatory step within a computerized order renewal process after 3 days of intravenous therapy might enhance the effectiveness of this approach. Funding No specific funding was received for this study. Transparency declarations None to declare. References 1 Zillich AJ, Sutherland JM, Wilson SJ et al. Antimicrobial use control measures to prevent and control antimicrobial resistance in US hospitals. Infect Control Hosp Epidemiol 2006; 27: Dellit TH, Owens RC, McGowan JE et al. Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis 2007; 44: Owens RC. Antimicrobial stewardship: concepts and strategies in the 21st century. Diagn Microbiol Infect Dis 2008; 61: MacDougall C, Polk RE. Antimicrobial stewardship programs in health care systems. Clin Microbiol Rev 2005; 18: CDC s campaign to prevent antimicrobial resistance in health-care settings. MMWR Morb Mortal Wkly Rep 2002; 51: Thuong M, Schortgen F, Zazempa V et al. Appropriate use of restricted antimicrobial agents in hospitals: the importance of empirical therapy and assisted re-evaluation. J Antimicrob Chemother 2000; 46: Senn L, Burnand B, Francioli P et al. Improving appropriateness of antibiotic therapy: randomized trial of an intervention to foster reassessment of prescription after 3 days. J Antimicrob Chemother 2004; 53: Lesprit P, Duong T, Girou E et al. Impact of a computer-generated alert system prompting review of antibiotic use in hospitals. J Antimicrob Chemother 2009; 63: van der Meer JW, Gyssens IC. Quality of antimicrobial drug prescription in hospital. Clin Microbiol Infect 2001; 7: Arnold FW, McDonald LC, Smith S et al. Improving antimicrobial use in the hospital setting by providing usage feedback to prescribing physicians. Infect Control Hosp Epidemiol 2006; 27: Carling PC, Fung T, Coldiron JS. Parenteral antibiotic use in acute-care hospitals: a standardized analysis of fourteen institutions. Clin Infect Dis 1999; 29: Fraser GL, Stogsdill P, Dickens JD Jr et al. Antibiotic optimization. An evaluation of patient safety and economic outcomes. Arch Intern Med 1997; 157: Fischer MA, Solomon DH, Teich JM et al. Conversion from intravenous to oral medications. Arch Intern Med 2003; 163: Zanetti G, Flanagan HL, Cohn LH et al. Improvement of intraoperative antibiotic prophylaxis in prolonged cardiac surgery by automated alerts in the operating room. Infect Control Hosp Epidemiol 2003; 24: Shojania KG, Yokoe D, Platt R et al. Reducing vancomycin use utilizing a computer guideline: results of a randomized controlled trial. J Am Med Inform Assoc 1998; 5: Cosgrove SE, Patel A, Song X et al. Impact of different methods of feedback to clinicians after postprescription antimicrobial review based on the Centers for Disease Control and Prevention s 12 steps to prevent antimicrobial resistance among hospitalized adults. Infect Control Hosp Epidemiol 2007; 28: LaRocco A. Concurrent antibiotic review programs a role for infectious disease specialists at small community hospitals. Clin Infect Dis 2003; 37: