Antifungal stewardship in daily practice and health economic implications

Transcription

1 mycoses Diagnosis,Therapy and Prophylaxis of Fungal Diseases Supplement article Antifungal stewardship in daily practice and health economic implications Patricia Mu~noz, 1,2,3 Maricela Valerio, 1,3 Antonio Vena 1,2,3 and Emilio Bouza 1,2,3 1 Clinical Microbiology and Infectious Diseases Department, Hospital General Universitario Gregorio Mara~non, Madrid, Spain, 2 Department of Medicine, Complutense University of Madrid, Madrid, Spain and 3 Instituto de Investigacion Sanitaria del Hospital Gregorio Mara~non, Madrid, Spain Summary During recent years, inappropriate antifungal use has contributed to the global increase in antifungal resistance and has played a role in the shift in the aetiology of invasive fungal infections. Moreover, overuse of antifungals may also lead to higher toxicity associated with unnecessary medication exposure and to increased healthcare costs. Antifungal stewardship (AFS) programmes consist of multidisciplinary interventions, led by specialists in infectious disease, microbiology and pharmacy that cooperate and communicate with the major prescribing departments in order to optimise antifungal therapies evaluating the indication, dose, streamlining and duration. Herein, we review the available evidence for the use of AFS and their impact on health economics. We also describe our AFS program, the successive steps we followed and the main difficulties we found. Key words: Antifungal stewardship, candins, fluconazole, antifungal therapy, empiric antifungals, invasive candidiasis. Introduction More than 50% of antimicrobial use is inappropriate, adding considerable cost to patient care and leading to increased selection of resistant pathogens, morbidity and mortality and unjustified cost. 1 Antimicrobial stewardship (AMS) programmes are designed to ensure optimal antimicrobial selection, dose and length of therapy. Until very recently, most AMS efforts have focused on antibiotic stewardship due to the high frequency of bacterial infections and to the increasingly limited therapeutic options for multidrug-resistant bacterial infections. 2 Several groups have demonstrated that these programmes improve efficacy, reduce costs, Correspondence: P. Mu~noz, MD, PhD, Servicio de Microbiologıa Clınica y Enfermedades Infecciosas, Hospital General Universitario Gregorio Mara~non, Doctor Esquerdo 46, Madrid 28007, Spain. Tel.: Fax: pmunoz@hggm.es Submitted for publication 27 February 2015 Revised 31 March 2015 Accepted for publication 1 April 2015 minimise adverse events and limit the potential for emergence of antimicrobial resistance. Although less of a problem, it is now recognised that inappropriate antifungal use has contributed to the global increase in antifungal resistance, increases morbidity and mortality and has played a role in the shift in the aetiology of invasive fungal infections. Overuse of antifungals may also lead to toxicity associated with unnecessary medication exposure and significantly increase costs. Wide variability in patient complexity and in empiric therapy practices has precluded the establishment of widely accepted antifungal therapy standards. Data from some institutions range from 40 to 296 defined daily doses (DDDs) per 1000 patient-days. 3 7 In our centre, a tertiary hospital with very active transplantation, oncology and HIV programmes, a recent survey showed that 65.1 DDDs per 1000 patient-days were prescribed and that overall, 13% were considered unnecessary. 8 Figures clearly above these DDDs require a justification and a careful review of clinical practice. We will briefly analyse data to support the need for antifungal stewardship (AFS) and review the accumulating evidence for doi: /myc Mycoses, 2015, 58 (Suppl. 2), 14 25

2 Antifungal stewardship and economic implications its use and impact on health economics in daily practice. The need for AFS programmes The need for AFS programmes is based primarily on the fact that appropriate real-world use of antifungals is quite difficult, even for experienced physicians. Challenges include the unspecific clinical manifestations of invasive fungal infections, mainly in immunocompromised and critically ill patients, the poor sensitivity of culture-based microbiologic tests and the pressure to start treatment early due to the high morbidity and mortality of these infections. However, empiric antifungal therapy (EAT) can be both unnecessary and costly. In fact, empiric/preemptive therapy accounts for the majority of inpatient antifungal prescriptions in departments such as intensive care units (ICUs). At our centre, antifungal drugs were initially prescribed for empiric/preemptive therapy (62%), followed by targeted treatment (20%), prophylaxis (15%) or unclassifiable indication (3%). 8 Similar findings have been reported by other authors. 9 In a recent study from Texas, 64% of all micafungin courses and 62% of fluconazole courses were for empiric treatment, and the most frequent indications were suspicion of invasive candidiasis or suspected fungal infection of the urinary tract. 10 The new-generation antifungal drugs are better tolerated than older ones, which has facilitated their empiric use in a large number of patients, but the higher cost of the new drugs is one incentive for their targeted use. This is clearly an area for AFS programmes, probably with the aid of new microbiologic diagnostic methods. Appropriateness rates for antifungal drugs are low Traditionally, invasive fungal diseases were thought to occur primarily in patients with haematologic cancers, transplant recipients or ICU patients, and it was believed that the physicians of these departments had enough expertise, so they did not need the help of stewardship programmes. It is now clear that, independent of the level of expertise of the group of physicians, the adherence to recommendations on antifungal drug use is extremely poor. 11 We studied solid organ transplant teams in Spain and found wide variability in their diagnostic methods and antifungal management strategies, including those for prophylaxis and treatment. 12 Universal antifungal prophylaxis was still used by 26.7% and 15% of liver and heart transplant centres, respectively, and first-line combination therapy for invasive aspergillosis was used by 31.3% of the centres studied. Drugs, doses and durations of treatment also differed widely. The reasons for this variability and for the poor adherence of transplant physicians to current recommendations for antifungal prophylaxis and treatment are poorly understood and merit further research. 13 Other authors have assessed the performance of other expert antifungal prescribers, such as ICU, oncology and haematology physicians, in a large hospital in France. These departments accounted for 70% of antifungal consumption in that facility. Surprisingly, choice of antifungal agent was found to be in accordance with labelling or international guidelines in only 65% of cases. 14 The indications of other antifungals chosen were considered either debatable (13%) or fully inappropriate (22%). In a study of antifungal use at a tertiary care centre in Thailand, inappropriate use was identified in 70% (42/57) of cases. 15 Multivariate analysis found that isolation of Candida species from urine was associated with inappropriate antifungal use, whereas receipt of an infectious disease consultation was protective against inappropriate use. Inappropriate use of antifungal treatments may affect all hospital departments. We performed an evaluation of 100 patients receiving systemic antifungals at our institution. Overall, 43% of the prescriptions came from medical departments, 25% from haematology/oncology departments, 17% from ICUs and 12% from surgical departments. In other centres the proportion of antifungal use from medical departments is 30 32%. 14,16 Most importantly, patients outside critical care units may have higher mortality rates. Bassetti et al. showed that in a general hospital in Italy 38% of the candidaemias (133/348) occurred in internal medicine, and 30-day mortality was higher for candidaemias occurring in internal medicine than in other wards (51.1% vs. 38.2%, P < 0.02). The reason was at least partially related to a low level of clinical suspicion and delayed treatment (only 7% of therapy was initiated in the first 48 h after the first blood culture vs. 25% in other units). 17 Of course, if the AFS cannot cover the whole institution, the departments prescribing antifungals most frequently should be targeted initially. Prescribers do not always modify initial therapy according to microbiology results (streamlining) Only a small proportion of patients receiving EAT finally have a proven infection, 18 but, even in these Mycoses, 2015, 58 (Suppl. 2),

3 P. Mu~noz et al. patients, therapy is not always adjusted once the results of susceptibility testing are available. In a retrospective study of hospitalised patients with candidaemia, most patients were not given antifungals until Candida was detected in blood cultures. After microbiologic information was available, surprisingly, the majority of echinocandin-treated patients with fluconazole-susceptible isolates (61%) were not de-escalated to fluconazole. 19 In another prospective study, 96% of physicians felt that antifungal susceptibility testing results were helpful, but in real life it only led to a change in therapy in 12% of patients. 20 Possible explanations include that physicians were concerned with drug drug interactions, missed the results of the susceptibility testing or were just not comfortable de-escalating therapy. Education on the safety of early step-down to oral azole therapy when clinically feasible should be a part of most AFS programmes. 21 Minimum requirements to start an AFS programme Measures typically employed in AFS programmes are shown in Fig. 1, but each centre must establish its own targets, selecting the most cost-effective initiatives. Steps for implementing an AFS programme include the following: (1) Organising a collaborative group that includes infectious diseases specialists, clinical microbiologists with trained laboratory staff or laboratory physicians, pharmacists and specialists in the main prescribing departments (haematology, oncology, intensive care and transplant unit). Ideally AFS champions will be designated from these departments, at least from the ICU and haematology/oncology. The multidisciplinary group must have official approval, empowerment and support from the hospital managers. (Our group is also covered under the committee for antimicrobial policy and nosocomial infections umbrella.) Within the collaborative group, coordinators should try to achieve consensus on achievable goals and commitments among the most important prescribing departments. Key team members should be assigned part-time to this task. (2) Once human resources are ready, it is necessary to assess the magnitude of the problem at each centre. An audit of antifungal use is recommended to detect where, how and by whom antifungals are prescribed and which are the most important points of misuse. Figure 1 Effectiveness of the interventions should be measured using predefined indicators Share the information and every success of your intervention with all members of the team Bedside intervention Implementation of rapid serological and molecular diagnostic tests* Pharmacy alerts regarding new AF prescribed on a daily basis Local Guidelines and clinical flowcharts Educational programme to offer trainees knowledge in IFI diagnosis and management in clinical practice Pre AF Stewardship audit and identification of main AF prescribers Creation of a Collaborative Group on Mycosis and Antifungal treatment Figure 1 AF stewardship step-by-step. *Use of molecular diagnostics and/or serological biomarkers like galactomannan and beta-dglucan for early diagnosis. Implementation of TDM for AF plasma levels and susceptibility testing. IFI, invasive fungal infection; AF, antifungal. 16 Mycoses, 2015, 58 (Suppl. 2), 14 25

4 Antifungal stewardship and economic implications (We performed a prospective audit of knowledge and of 100 prescriptions, and results were used to target intervention and feedback.) (3) An educational programme with sessions available at least annually is highly desirable to increase the knowledge of antifungal prescribers in diagnosing and managing invasive fungal infections. Focused sessions involving the most significant prescriber departments may be useful for obtaining feedback or in the event that a major deviation in prescribing is detected. (4) As noted previously, universal guidelines are not always applicable to daily practice. Each centre should establish its own local guidelines based on the epidemiologic and demographic characteristics of the community it serves. Such guidelines should be implemented as part of the AFS programme to improve the everyday management of invasive fungal diseases. They should include diagnostic criteria, recommendations for antifungal prophylaxis and treatment, dose adjustments for hepatic and/or renal dysfunction and how to easily contact members of the AFS group. The guidelines should be drafted by and agreed on within the multidisciplinary group and updated at least every 2 years, incorporating suggestions of prescribers, microbiologists and pharmacists, as well as newly available knowledge. (At our centre, pocket leaflets containing the local guidelines are routinely offered to all physicians contacted by the AFS programme.) (5) The Pharmacy department is an essential part of the AFS team. The use of electronic prescription tools can help foster adherence to the indications agreed on in the local guidelines and provide alerts on possible drug drug interactions and toxicities. Pharmacy may also help transfer daily alerts on antifungal prescriptions to infectious diseases specialists in charge of bedside direct intervention. In some centres, a specialised pharmacist performs a systematic intervention, whereas in others, formulary restrictions require preauthorisation. In order to measure the impact of the AFS programme, costs should be monitored monthly by the pharmacy department. If possible, the overall and unit-level number of DDDs or prescribed DDDs per 1000 patient-days should be registered for benchmarking. (6) The clinical microbiology department or the laboratory medicine department also has an essential role. Implementation of innovative diagnostic tools and rapid turnover of laboratory results are mandatory to adjust antifungal therapy early. These measures can help to avoid unnecessary antifungal treatments and approve appropriate early antifungal therapy in septic patients. It is also important for clinicians who use new azoles to have access to the latest pharmacokinetic/pharmacodynamic data, at least for difficult cases. Dose optimisation through therapeutic drug monitoring is now widely recognised as an important tool to optimise clinical efficacy and minimise toxicity. The microbiology department or the laboratory medicine department may also contribute to the timeliness and comprehensiveness of diagnostic investigations when an invasive fungal infection is suspected. (7) Infectious diseases specialists, when available, may perform bedside interventions and advise and mentor other specialists in the management of invasive fungal infections. (In our experience a physician-to-physician intervention is better accepted by the prescribing clinician.) However, recommendations may be rejected if there is a perceived loss of autonomy in patient management, so involving prescribing physicians in the discussion is strategically crucial. 22 On the other hand, in large infectious diseases groups in which attention to different hospital departments is divided among specialists (solid transplant, haematology, ICU, etc.), the coordinator of the AFS programme should collaborate with the dedicated infectious diseases specialist, who is generally more influential in his/her area of practice. The best scenario is to involve as many members of the team as possible; ignoring a colleague may generate unnecessary divergences. However, the AFS coordinator remains accountable for the results of the programme. (8) Select the type of intervention. Generally it is accepted that passive methods that rely on educational activities and management guidelines are insufficient. (9) Select indicators for measuring the effectiveness of the interventions. Some indicators may be recorded monthly (number of candidaemias, cost of antifungals, percentage of patients treated with antifungals who were contacted by the AFS programme, percentage of acceptance of the recommendations, number of microbiological tests ordered by departments, etc.). Other indicators may be registered with wider periodicity (number of persons who were exposed to the educational activities, grade of satisfaction with the AFS programme, evolution of the knowledge score of prescribing physicians, etc.). The number of indicators should be low. It is important to emphasise that drug acquisition cost alone should never be the only indicator of an AFS programme s success, since 1 single case of candidaemia prevented through a multidisciplinary infection control policy is worth significantly more than many full courses of the most expensive antifungal. 23 (10) Finally, the information should flow in every direction and every success, even if small, should be Mycoses, 2015, 58 (Suppl. 2),

5 P. Mu~noz et al. shared with all the members of the multidisciplinary team and with the involved departments. Unit-specific feedback is necessary and should be provided periodically. Do not forget to share all successes with everybody and to give credit for them to the team. AFS programmes in practice There is much less experience with AFS programmes than with antibacterial initiatives (Table 1). However, published data on AFS programmes consistently report cost savings or at least containment of costs, with no significant changes in clinical outcomes and sometimes reduction in resistance and improvement in the quality of care. An excellent comprehensive review on this subject has been recently published. 24 Dedicated personnel Two articles demonstrate that at least part-time, dedicated physicians and pharmacists are necessary to maintain a successful AFS programme and that the impact of a key team member may go unrecognised until that person is no longer involved with the programme. The first study reported on an AMS programme that was fully implemented throughout 7 years in the USA. Essential to the programme was an antimicrobial monitoring team consisting of an infectious diseasestrained clinical pharmacist and a part-time infectious diseases physician who provided real-time monitoring of antimicrobial orders and active intervention and education when necessary. After 7 years, the programme was terminated in order to use the resources to increase infectious diseases consultations throughout the medical centre as an alternative mode of stewardship. 5 A descriptive cost analysis, before, during and after the programme found a 37% decrease in costs during the first 3 years. At baseline, antifungals accounted for almost half ($3.7 million) of the entire cost of antimicrobials, and the majority of the programme s cost savings occurred with antifungals, the cost of which was reduced by 60.7% throughout the first 3 years. DDDs per 1000 patient-days for antifungals were significantly reduced by 24% (P = 0.001) from 2004 to This was achieved with no significant changes in quality indicators, including length of stay, readmissions and mortality. However, after the programme was discontinued at the end of 2008, antimicrobial costs increased from $ to $ per 1000 patient-days, a 32.3% increase within 2 years that was equivalent to a $2 million increase in costs for the medical centre, mostly within the antibacterial category. 5 After establishing that the dedicated monitoring team was extremely cost-effective and necessary to assure the functioning of the programme, the programme was reinitiated. A similar experience was reported for another American centre in which the AMS team comprised an infectious diseases physician, an infectious diseases-trained Table 1 Multifaceted aspects of AFS programmes. Intervention Comment References Educational Restrictive prescription Bedside ID advice Bundle of care Pharmaceutic advice New diagnostic strategy Evaluation of gaps in knowledge of antifungal prescribers in order to tailor AFS programmes ID consultant imposed the implementation of practice guidelines, provided approval of prescribed drugs or new diagnostic and therapeutic approaches for prescribing antifungal treatment Recommendations to change from IV to oral, change to fluconazole, cease antifungal treatment Antifungal order forms, educational and unit-specific feedback activities, expert infectious diseases bedside interventions; preauthorisation of treatment by antifungal team Recommendations from pharmacist to change or stop the controlled antimicrobial agents based on microbiological data and institutional criteria for antimicrobial use Application of PCR testing and serological markers for diagnosis of invasive fungal infection Use of molecular analysis for characterisation of clinical isolates Standiford et al. [5] Valerio et al. [11] Valerio et al. [35] Cook et al. [4] Swobada et al. [27] Aguilar-Guisado et al. [30] Lopez-Medrano et al. [7] Mondain et al. [3] Apisarnthanarak et al. [6] Antworth et al. [26] Guarascio et al. [28] Cook et al. [4] Cappelletty et al. [25] Guinea et al. [36] Escribano et al. [37] Marcos-Zambrano et al. [38] Escribano et al. [37] Martinez-Jimenez et al. [40] Barnes et al. [32] 18 Mycoses, 2015, 58 (Suppl. 2), 14 25

6 Antifungal stewardship and economic implications clinical pharmacist and a clinical microbiologist. The pharmacist prospectively identified patients receiving restricted-use antimicrobials and, based on culture results and institutional criteria for antimicrobial use, initiated the intervention within h, consulting with the infectious diseases physician as needed. After the pharmacist was no longer active in the group, the use of the studied candin (micafungin) increased 35%, the average duration of therapy was prolonged 3.2 days and the overall rate of appropriate use dropped from 84.6% to 50%. 25 Bedside interventions There is experience on the effectiveness of non-compulsory bedside interventions performed by experienced infectious diseases physicians. A non-compulsory 12-month AFS programme in Spain reduced antifungal costs by 12% (a saving of $ ) compared with the previous 12 months. 7 A recommendation to change treatment was made in 29% of cases, including a change from intravenous to oral treatment (15%), stopping antifungal treatment (8%) and a change to fluconazole (6%). There were important reductions in the prescribed doses of intravenous voriconazole and caspofungin (31.4% and 20.2%, respectively), and the use of oral voriconazole increased. The suggestions acceptance rate was 99%, and a mean of 4.5 treatment revisions per patient occurred in patients with confirmed infections. There were no significant changes in quality of care indicators. Bundled interventions Antifungal stewardship programmes may contain multiple simultaneous activities (bundled interventions) for a single disease, including education initiatives, dose adjustment tools, antifungal prescription forms and prescription-control strategies. In a hospital in Thailand with no haematology department, baseline inappropriate use of antifungals had reached 71%. A programme focused on the treatment of candidiasis was initiated. A large multidisciplinary group was created including two infectious diseases specialists, a clinical microbiologist, four pharmacists, two internists, a hospital epidemiologist, an infection control specialist and a computer systems analyst. Antifungal order forms became an obligatory pharmacy requirement, educational and unit-specific feedback activities were provided and expert infectious diseases bedside interventions put in place. With these measures the 18-month AFS programme resulted in a 59% reduction in antifungal prescriptions and a very important decrease in the treatment of candiduria. Inappropriate antifungal use decreased from 71% to 24%, and total cost savings reached $ during the 18-month postintervention period. Most importantly, the programme was associated with a decreased incidence of azole-resistant Candida species, Candida glabrata and Candida krusei, and an increased incidence of Candida albicans. 6 Improved quality of care Antifungal stewardship programmes do not only reduce inappropriate use of antifungals, cost and resistance, as we have seen; they also improve the quality of care of patients with proven invasive fungal diseases. This achievement was shown by Mondain et al., 3 who described a 6-year multifaceted AFS programme in a tertiary care hospital in France. Several actions were implemented successively, including the systematic evaluation of all costly antifungal prescriptions (echinocandins, lipid formulations of amphotericin B, posaconazole and voriconazole). Overall, 72% of the discussed prescriptions corresponded to the haematology department. Advice was fed back to the physician in charge of the patient in 54% of the cases, including the following changes in antifungal treatment: start therapy (4%), stop therapy (15%) or change therapy (30%); compliance rate with suggestions was 88%. Total antifungal prescriptions and their costs were contained between 2003 and 2010, despite the extension of indications for costly antifungals throughout the years. Most importantly, quality of care for patients with candidaemia and aspergillosis was improved. Optimal standard of care was achieved for galactomannan antigen testing, computed tomography (CT) scanning of the chest and therapeutic drug monitoring for invasive aspergillosis in patients receiving voriconazole, with no combination therapies used since Regarding candidaemia, optimal standard of care was achieved for the timing of antifungal therapy, recommended first-line therapy, duration of therapy and the removal of central venous catheters. 3 Antworth et al., 26 reported on another effective candidaemia bundle, which included preauthorisation of treatment by the antifungal team during a 6-month period at a trauma centre in the USA. Implementation of the bundle significantly improved selection of antifungal therapy and compliance with appropriate duration of therapy, with no significant differences in clinical outcomes observed. In both the Mondain et al. and Antworth et al. studies, patients with candidaemia Mycoses, 2015, 58 (Suppl. 2),

7 P. Mu~noz et al. who died early were excluded from the primary analysis, as these patients did not have the opportunity to complete all elements of the care bundle. Programmes for ICUs Other programmes have been designed for determined units or hospital departments. Two studies of AFS programmes designed for ICUs are reviewed here. Swoboda et al., 27 investigated the impact of a standardised practice of antifungal treatment in the surgical ICU of a hospital in Germany throughout an 18-month period. The surgical ICU staff and an interdisciplinary team agreed on a local flow chart for antifungal therapy management. A clinical pharmacist was placed in charge of supervising every prescription of systemic antifungal drugs. After its approval, the practice guideline was implemented by the consultant responsible for antifungal strategies and the responsible person of the surgical ICU, and it became compulsory for all the team members. Implementation of the practice guidelines included off-label use of posaconazole for C. glabrata and C. krusei infections when enteral absorption was certain and the patient was stable. The intervention resulted in a 50% reduction in the costs of antifungal agents, with no changes in outcome measures. The second study reported on the implementation by the pharmacy department of an antifungal bundle for patients given caspofungin in the ICUs of a hospital in the USA. The bundle was assessed each morning during ICU rounds and included evaluation of initial diagnosis, indication, planned duration of therapy and subsequent potential for de-escalation or discontinuation of therapy and device removal. The intervention resulted in a significantly decreased median duration of therapy (4 vs. 2 days, P = 0.001) and was more effective in the medical ICU than in the surgical ICU. 28 The potential cost saving per patient was $1013. Restrictive programmes More restrictive approaches have also been reported. The results of an AMS programme including formulary restriction and a preauthorisation requirement at a tertiary care hospital in the USA were reported by Cook et al. 4 All systemic antifungals were classified as restricted drugs with the exception of intravenous fluconazole, which was classified as a controlled drug. The pharmacist, with input from the infectious diseases practitioner, made recommendations to change or stop the controlled antimicrobial agents. The recommendations were placed in the order section of the patient s chart, immediately in front of the most recent order. The attending physician had 24 h to either accept or reject the recommendation. If at the end of the 24-h time period there was no response to the recommendation, the pharmacist would then write the recommended changes as orders. Compared with the six quarters before the programme, total antifungal agent use decreased by 28% (144 DDD to 103 DDD per 1000 patient-days, P = 0.02). Lengths of stay and mortality rates were not significantly altered by the programme. This type of intervention, in which a pharmacist changes the prescription of critically ill patients, would not be easy to implement at all institutions, although in this study the acceptance of the programme was very high (over 90%). The particular case of high-risk neutropenic patients Persistent febrile neutropenia is a recognised clinical presentation of invasive fungal infection in haematologic patients. Together with the high morbidity and mortality of invasive fungal disease in these patients, this has led to the recommendation of EAT in every high-risk patient. 29 However, the scientific evidences supporting such recommendation are limited and this approach could lead to overuse of antifungal drugs, toxicity and excess of cost. During recent years, improvement in radiological (ultrasonography, computed tomography) and microbiological procedures (use of galactomannan), have prompted investigation of alternative approach utilising diagnosis-driven antifungal therapy (DDAT) based on clinical flow charts or new microbiological diagnostic methods with high negative predictive value. Aguilar-Guisado et al., 30 investigated use of a DDAT approach based on clinical criteria and risk profile to determine patients with persistent febrile neutropenia who require antifungal therapy in a Spanish tertiary centre. The population included very high-risk neutropenic patients (98% with prolonged or profound neutropenia, and 41.6% receiving allogeneic haematopoietic stem cell transplantation or with relapsed leukaemia). Febrile patients were clinically classified as having severe sepsis/shock, pneumonia, rhinosinusitis, central nervous system abscess, abdominal focus, skin lesions or fever of unknown origin. The choice of antifungal drug was guided in each case by the most probable fungal aetiology, depending on clinical criteria and/or results of the diagnostic workup. Patients without severe sepsis and/or clinical focus and with 20 Mycoses, 2015, 58 (Suppl. 2), 14 25

8 Antifungal stewardship and economic implications negative CT scan and negative galactomannan were recommended to not receive antifungal drugs, despite the fever. Antifungal therapy was not indicated in 33 episodes (38.8%), and the overall incidence of proven and probable invasive fungal infection was 14.1%. All patients who presented an invasive fungal infection had been recommended appropriate empiric therapy; the invasive fungal infection cure rate was 66.7% (8/12) and related mortality was 2.7%. The high sensitivity and negative predictive value of this clinical pathway (both 100%) indicated that this approach could be useful for identifying febrile neutropenic patients who are not likely to have fungal infections and do not, therefore, require antifungal therapy. Interestingly, this study showed that the most frequent cause of persistent fever was a non-fungal infection that responded well to antibacterial therapy, whereas probable or proven invasive fungal infection accounted for only 14.1% of the episodes. Overall, even with this tailored systematic approach, it was calculated that 76.9% of patients who received antifungal therapy did not have either proven or probable fungal infection. A cost-effectiveness analysis comparing this DDAT approach with the standard approach of EATS showed that the economic advantage for the former per persistent febrile neutropenic episode was $5879 (33%). 31 The average cost-effectiveness per episode was $ for the DDAT approach compared with $ for the EAT approach. Another study showed that a combination of antigen testing [galactomannan enzyme immunoassay (GM EIA)] and Aspergillus polymerase chain reaction (PCR) allowed exclusion of a diagnosis of invasive aspergillosis and thus for EAT to be safely withheld in febrile patients. 32 In this study, a cohort of 549 highrisk haematologic and stem cell transplant recipients was followed during a 5-year period. During this time patients were managed within an integrative care pathway whereby a diagnostically driven approach replaced EAT during neutropenic fever. Patients received prophylaxis with itraconazole, with serum levels measured weekly and considered therapeutic if above 0.5 mg l 1. Fungal diagnostic tests (GM EIA and PCR) were performed twice weekly in stem cell transplant patients and during fever in other patients. EAT was not used in patients receiving adequate prophylaxis. High-resolution CT scan was requested in accordance with standards of care. When both diagnostic tests were used together, the sensitivity was 98.1%, the negative predictive value was 99.6% and the negative likelihood ratio was 0.04 for proven probable disease. Antifungal expenditure decreased in the first 6 months of the care pathway with savings of approximately double the cost of implementing the biomarker screening. Antifungal expenditure continued to show a downward trend throughout the 5-year period whilst the costs of screening remained stable. Improved microbiologic diagnostic methods are also AFS Although out of the scope of this review, it is clear that new non-culture-based microbiologic diagnostic methods will prove to be an essential part of AFS programmes. These methods can be addressed to obtain a more rapid diagnosis, to determine the infecting species and its antifungal susceptibility, to help determine the origin of the fungaemia or to exclude the presence of invasive fungal infection. Simple methods such as including e-test strips in all blood cultures in which a yeast is detected in order to provide earlier antifungal information or more rapid identification of the isolates with matrix-assisted laser desorption/ionisation-timeof-flight mass spectrometry (MALDI-TOF MS) or fluorescence in situ hybridisation using peptide nucleic acid probes (PNA-FISH) may be useful for these programmes. However, the real impact of these techniques in terms of antifungal use or mortality needs more research. As mentioned before, the main objective of an AFS programme is not only to reduce pharmacy expenditure in antifungals, but to improve the care of patients with fungal disease. These new microbiologic tests will clearly help with this task, for example, by helping attending physicians provide more accurate empiric therapy. 10 Despite widespread use of empiric antifungals, the truth is that many studies show that most patients with candidaemia do not have the drug started until the diagnosis from the microbiology department is established. A recent study of patients with candidaemia showed that the average time to start antifungal therapy was days. By means of a Monte-Carlo simulation, it was calculated that Candida PCR would reduce average time to initiation of antifungal therapy to days, while with PNA-FISH and MALDI-TOF MS it would be started at and days, respectively, after blood culture collection. Other interesting upcoming technology for rapid diagnosis of candidaemia is the T2 magnetic resonance (T2MR), it directly analyses whole blood specimens to identify species without the need for prior isolation of Candida species and represents a breakthrough shift into a new era of molecular diagnostics. It has a high negative predictive value Mycoses, 2015, 58 (Suppl. 2),

9 P. Mu~noz et al. ( %) providing an identification in a mean time of h. 33 Considering that it has been reported that anticipation of therapy by just 1 day supposes a reduction in cost of $ per episode of candidaemia, the impact of these simple interventions are surely cost-effective. Our experience We designed an AFS programme that was implemented in a stepwise manner and that now works as a bundled or multi-faceted programme. We recently reported the results from the first 2 years. 34 During the first year we performed a baseline intervention (Steps 1 5) that consisted of the following tasks: (1) Creation of a multidisciplinary collaborative group on mycology (COMIC study group) in which prescribing physicians of different specialities, clinical microbiologists, infectious diseases specialists and pharmacists developed the programme and reached a consensus on general goals. The group was assigned the responsibility of assessing the local epidemiology of fungal infections, suggesting preventive and therapeutic strategies and monitoring use of antifungal drugs. (2) The group developed and distributed local guidelines for the diagnosis and treatment of invasive candidiasis and aspergillosis that were approved by the antimicrobial subcommittee of the pharmacy and therapeutics committee. They were based on international guidelines but were adapted according to local epidemiology and available diagnostic methods. (3) The pharmacy department incorporated a computerised physician order entry (CPOE). The CPOE system requires the prescribing physician to fill out the reason for each new antifungal prescription. The CPOE tool also includes pop-out information about diagnosis and treatment of invasive fungal infections, reminders for voriconazole and posaconazole pharmacokinetics determinations and drug drug interactions and comparative cost of antifungal treatments. (4) In order to evaluate the prescribing etiquette and use of antifungal drugs, we performed an audit of 100 antifungal courses to answer the following questions: how many antifungal drugs were prescribed, by whom, for which type of patients and conditions, and which were the most common mistakes? The adequacy of antifungal use was evaluated using a point score ranging from 0 to 10. We assigned more impact (0 2 points) to mistakes that could imply a major risk for the patient (prescription of a not-needed antifungal agent) or to aspects that were clear intervention targets (lack of adjustment following receipt of microbiologic information or excessive duration of treatment). Less detrimental mistakes, such as incorrect dosage or lack of switching to an oral form, were given a smaller impact (0 or 1 point) in the global score. In the case of drug selection, a non-efficacious drug was a major mistake (0 points), an effective but non-optimal selection was a minor mistake (1 point) and a perfect selection was awarded 2 points. 8 Fluconazole predominated for all indications (urinary tract infections, oral candidiasis, skin and soft tissue infections, catheter related candidaemia, oesophageal candidiasis); candins were widely used as preemptive therapy in critically ill patients and patients with candidaemia, while liposomal amphotericin B was given most frequently for the tailored treatment of Leishmania infections (2 of 4 total prescriptions). Overall, 16% of the antifungals were judged to be fully unnecessary and the prevalence of inappropriate antifungal use was 57%. The mean point score for antifungal use in the study patients was A score of <5 points was recorded for 14% of prescriptions. As expected, scores were lower for empiric (6.6) than for tailored (9.5) or prophylactic (9.1) therapies. Although difference in mean score among hospital departments was not significant, ICUs clearly accounted for the highest inadequate costs (frequent use of candins for empiric therapy). Inadequate empiric therapy was responsible of 78.6% of total potential cost savings. The most common inadequacies were no adjustment after microbiologic results (35%), drug selection (31%), length of therapy (27%), no switch from intravenous to oral administration (20%) and dosing (16%). An infectious diseases consultation was requested for 42% of patients (54.8% from medical wards, 28.6% from ICUs, 14.3% from surgical wards and 2.4% from oncology and haematology wards). The proportion of inappropriate prescriptions was higher in patients for whom an infectious diseases consultation was not requested (74.1% vs. 33.3%, P < 0.001). The main reason for the unnecessary prescription of an antifungal drug was colonisation by Candida species. As for suboptimal antifungal selection, the most common error was prescribing echinocandins to patients with azole-susceptible Candida infections. Other common mistakes were not stopping treatment when invasive fungal disease risk factors disappear or when sepsis was shown to be caused by another type of infection. Regarding cost, we calculated that if 100% of DDDs would have been optimal, the cost of acquisition would have been reduced by 23%. 8 In our opinion 22 Mycoses, 2015, 58 (Suppl. 2), 14 25

10 Antifungal stewardship and economic implications this simple score may be used to identify opportunities for improvement, provide baseline information and to assess the results of interventions. (5) Our last baseline intervention was a survey to assess knowledge on diagnosis and management of invasive candidiasis and invasive aspergillosis. 35 Attending physicians from the departments that prescribe most antifungals were invited to complete a questionnaire based on current guidelines on diagnosis and therapy of invasive candidiasis and invasive aspergillosis. The survey was completed by 60.8% (200/ 329) of the physicians who were invited to participate: medical (60%), paediatric (19%), intensive care (15.5%) and surgical departments (5.5%). The mean (standard deviation) score of correct responses was In the case of candidiasis, only 55% of the physicians clearly distinguished between colonisation and infection, and 17.5% knew the local rate of fluconazole resistance. Thirty-three per cent knew the accepted indications for antifungal prophylaxis and 23% the indications for empiric therapy. However, most physicians knew which antifungals to choose when starting empiric therapy (73.5%). For aspergillosis, most physicians (67%) could differentiate between colonisation and infection, and 34.5% knew the diagnostic value of galactomannan testing. The radiologic features of invasive aspergillosis were well recognised by 64%, but only 31.5% were aware of the first line of treatment. Only 36% of the prescribers knew that the guidelines recommend that treatment of invasive pulmonary aspergillosis should be continued for a minimum of 6 12 weeks. The usefulness of antifungal levels was known by 67%. This simple, easily completed questionnaire enabled us to identify which areas of our training strategy could be improved. After those initial steps, we started more active interventions (Steps 6 8) during the second year of the programme. (6) We performed diverse educational activities, including an interactive annual course based in clinical cases and preceptorships in antifungal use for critical care physicians. We also performed local education meetings in the most important departments and convened joint meetings of the COMIC group every 2 months to discuss general issues and present updates on the programme. We have repeated the knowledge questionnaire before and after our annual educational activities, with encouraging results. (7) Newly available diagnostic procedures implemented by the clinical microbiology department include local determination of azole drugs levels (therapeutic drug monitoring), molecular analysis of strains causing fungaemia to detect horizontal transmission, 36,37 close follow-up of resistance in clinical isolates 38,39 and diverse serologic diagnostic techniques [C. albicans germ tube antibodies (CAGTA), Platelia Candida antigen and beta-d glucan]. 40 (8) Bedside intervention. Ours is a bedside non-compulsory intervention that includes a pharmacist s alert to infectious diseases specialists on every patient receiving candins, liposomal amphotericin B, voriconazole or posaconazole. The infectious diseases specialist visits the patients to advise attending physicians on diagnostic and therapeutic issues. We recommended de-escalation to fluconazole when an azole-susceptible species was isolated, organ dysfunctions were resolved and drug drug interactions were unlikely. Staff attitudes towards the programme recommendations were also monitored. We calculated the economic impact of our intervention, comparing cost ( ) and DDDs per 1000 admissions before and during the AFS programme. Antifungal consumption per department was also calculated. The first year bedside intervention included 453 patients. Main prescribing departments were haematology (35%), medical departments (23%) and ICUs (20%). Indications for antifungal therapy were targeted (36%), prophylaxis (32%), empiric (20%) and preemptive (12%). At the initial visit, diagnostic procedures were suggested in 40.4% of cases and change of therapy was considered necessary in 32%. The most common recommendations were to de-escalate the antifungal drug (17.4%) or to stop it (7.1%). A fungal infection was demonstrated in 28.9% of patients. Overall, a therapeutic inadequacy was found in 45% of patients. The most common problems were inadequate duration of therapy (28.5%), inadequate selection of antifungal drug (24.1%) and lack of adjustment after microbiology results (18.5%). Compared with the 12 months prior to programme implementation, during bedside intervention there was a 17.5% reduction in the number of consumed DDDs (from 66.4 to 54.8 DDDs per 1000 patient-days). Expenditure on antifungals was reduced by $ (26.2%), mainly due to a reduction in the consumption of candins and voriconazole. Acknowledgments This study was sponsored by Pfizer Inc. Editorial support was provided by Annette Smith, PhD, of Complete Medical Communications and sponsored by Pfizer Inc. Conflict of interest Patricia Mu~noz has received travel grants, given talks or participated in advisory boards with Astellas, Mycoses, 2015, 58 (Suppl. 2),

11 P. Mu~noz et al. Gilead, Merck, and Pfizer Inc. Maricela Valerio recently received a travel grant from Merck and Astellas. Antonio Vena has received a travel grant from Astellas. Emilio Bouza has received travel grants, given talks or participated in advisory boards with Astellas, Gilead, Merck and Pfizer Inc. References 1 Dellit TH, Owens RC, McGowan JE Jr 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: Boucher HW. Challenges in anti-infective development in the era of bad bugs, no drugs: a regulatory perspective using the example of bloodstream infection as an indication. Clin Infect Dis 2010; 50 (Suppl. 1): S Mondain V, Lieutier F, Hasseine L et al. A 6-year antifungal stewardship programme in a teaching hospital. Infection 2013; 41: Cook PP, Catrou PG, Christie JD, Young PD, Polk RE. Reduction in broad-spectrum antimicrobial use associated with no improvement in hospital antibiogram. J Antimicrob Chemother 2004; 53: Standiford HC, Chan S, Tripoli M, Weekes E, Forrest GN. Antimicrobial stewardship at a large tertiary care academic medical center: cost analysis before, during, and after a 7-year program. Infect Control Hosp Epidemiol 2012; 33: Apisarnthanarak A, Yatrasert A, Mundy LM. Impact of education and an antifungal stewardship program for candidiasis at a Thai tertiary care center. Infect Control Hosp Epidemiol 2010; 31: Lopez-Medrano F, San Juan R, Lizasoain M et al. A non-compulsory stewardship programme for the management of antifungals in a university-affiliated hospital. Clin Microbiol Infect 2013; 19: Valerio M, Rodriguez-Gonzalez CG, Munoz P, Caliz B, Sanjurjo M, Bouza E. Evaluation of antifungal use in a tertiary care institution: antifungal stewardship urgently needed. J Antimicrob Chemother 2014; 69: des Champs-Bro B, Leroy-Cotteau A, Mazingue F et al. Invasive fungal infections: epidemiology and analysis of antifungal prescriptions in onco-haematology. J Clin Pharm Ther 2011; 36: Aitken SL, Beyda ND, Shah DN et al. Clinical practice patterns in hospitalized patients at risk for invasive candidiasis: role of antifungal stewardship programs in an era of rapid diagnostics. Ann Pharmacother 2014; 48: Valerio M, Vena A, Bouza E et al. How much European prescribing physicians know about invasive fungal infections management? BMC Infect Dis 2015; 15: Munoz P, Rojas L, Cervera C et al. Poor compliance with antifungal drug use guidelines by transplant physicians: a framework for educational guidelines and an international consensus on patient safety. Clin Transplant 2012; 26: Singh N, Wagener MM, Cacciarelli TV, Levitsky J. Antifungal management practices in liver transplant recipients. Am J Transplant 2008; 8: Nivoix Y, Launoy A, Lutun P et al. Adherence to recommendations for the use of antifungal agents in a tertiary care hospital. J Antimicrob Chemother 2012; 67: Sutepvarnon A, Apisarnthanarak A, Camins B, Mondy K, Fraser VJ. Inappropriate use of antifungal medications in a tertiary care center in Thailand: a prospective study. Infect Control Hosp Epidemiol 2008; 29: Pavese P, Ouachi Z, Vittoz JP et al. Adequacy of new systemic antifungal agents prescriptions in a teaching hospital. Medecine et Maladies Infectieuses 2007; 37(Suppl. 3): S Bassetti M, Molinari MP, Mussap M, Viscoli C, Righi E. Candidaemia in internal medicine departments: the burden of a rising problem. Clin Microbiol Infect 2013; 19: E Gonzalez de Molina FJ, Leon C, Ruiz-Santana S, Saavedra P. Assessment of candidemia-attributable mortality in critically ill patients using propensity score matching analysis. Crit Care 2012; 16: R Shah DN, Yau R, Weston J et al. Evaluation of antifungal therapy in patients with candidaemia based on susceptibility testing results: implications for antimicrobial stewardship programmes. J Antimicrob Chemother 2011; 66: Baddley JW, Patel M, Jones M, Cloud G, Smith AC, Moser SA. Utility of real-time antifungal susceptibility testing for fluconazole in the treatment of candidemia. Diagn Microbiol Infect Dis 2004; 50: Vazquez J, Reboli AC, Pappas PG et al. Evaluation of an early stepdown strategy from intravenous anidulafungin to oral azole therapy for the treatment of candidemia and other forms of invasive candidiasis: results from an open-label trial. BMC Infect Dis 2014; 14: Charani E, Castro-Sanchez E, Sevdalis N et al. Understanding the determinants of antimicrobial prescribing within hospitals: the role of prescribing etiquette. Clin Infect Dis 2013; 57: Dodds Ashley E, Drew R, Johnson M et al. Cost of invasive fungal infections in the era of new diagnostics and expanded treatment options. Pharmacother 2012; 32: Ananda-Rajah MR, Slavin MA, Thursky KT. The case for antifungal stewardship. Curr Opin Infect Dis 2012; 25: Cappelletty D, Jacobs D. Evaluating the impact of a pharmacist s absence from an antimicrobial stewardship team. Am J Health Syst Pharm 2013; 70: Antworth A, Collins CD, Kunapuli A et al. Impact of an antimicrobial stewardship program comprehensive care bundle on management of candidemia. Pharmacother 2013; 33: Swoboda S, Lichtenstern C, Ober MC et al. Implementation of practice guidelines for antifungal therapy in a surgical intensive care unit and its impact on use and costs. Chemother 2009; 55: Guarascio AJ, Slain D, McKnight R et al. A matched-control evaluation of an antifungal bundle in the intensive care unit at a university teaching hospital. Int J Clin Pharm 2013; 35: Freifeld AG, Bow EJ, Sepkowitz KA et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the Infectious Diseases Society of America. Clin Infect Dis 2011; 52: e Aguilar-Guisado M, Martin-Pena A, Espigado I et al. Universal antifungal therapy is not needed in persistent febrile neutropenia: a tailored diagnostic and therapeutic approach. Haematologica 2012; 97: Martin-Pena A, Gil-Navarro MV, Aguilar-Guisado M et al. Cost-effectiveness analysis comparing two approaches for empirical antifungal therapy in hematological patients with persistent febrile neutropenia. Antimicrob Agents Chemother 2013; 57: Barnes RA, Stocking K, Bowden S, Poynton MH, White PL. Prevention and diagnosis of invasive fungal disease in high-risk patients within an integrative care pathway. J Infect 2013; 67: Mylonakis E, Clancy CJ, Ostrosky-Zeichner L et al. T2 magnetic resonance assay for the rapid diagnosis of candidemia in whole blood: a clinical trial. Clin Infect Dis 2015; 60: Valerio M, Mu~noz P, Rodriguez CG et al. Two years of an antifungal stewardship program. Clin Microbiol Infect In press. 35 Valerio M, Munoz P, Rodriguez-Gonzalez C, Sanjurjo M, Guinea J, Bouza E. Training should be the first step toward an antifungal stewardship program. Enferm Infecc Microbiol Clin 2014; 33: Guinea J, Zaragoza O, Escribano P et al. Molecular identification and antifungal susceptibility of yeast isolates causing fungemia collected in a population-based study in Spain in 2010 and Antimicrob Agents Chemother 2014; 58: Escribano P, Rodriguez-Creixems M, Sanchez-Carrillo C, Munoz P, Bouza E, Guinea J. Endemic genotypes of Candida albicans causing 24 Mycoses, 2015, 58 (Suppl. 2), 14 25