Half of patients in Intensive Care Units (ICU) suffer infection while 30% die secondary to it.1 Both ICU-acquired infection and antibiotic resistance are associated with an increased risk of death.1 It follows that 70% of ICU patients are receiving antimicrobial therapy (AMT) representing a burden to healthcare costs, predisposing to antimicrobial resistance (AMR) and worsening patients’ outcome.1,2 AMR is more prevalent in low- to middle-income countries (LMICs) than high-income countries (HICs) and feared to cause millions of deaths by 2050.2

Antimicrobial stewardship programme (ASP) is a mean of limiting AMR while improving patients’ outcome. It is defined as “the coordinated interventions designed to improve and measure the appropriate use of AMT by promoting the selection of the optimal drug regimen including dosing, duration of therapy, and route of administration‿.3 It is recommended as a core competency for intensivists and for all ICUs integrating clinicians, pharmacists, infectious disease (ID) physicians and clinical microbiologists (CM) into one multidisciplinary team.3,4

Understanding ICU ID practice remains essential for identifying gaps and opportunities, but is generally lacking. We conducted an international online survey (PRACT-INF-ICU) to explore from intensivists’ perspective; (1) the practice of ID in the ICU, (2) their satisfaction with ID training, and (3) the impact of defensive medicine on ICU ID practice. Variability in responses were explored regarding settings (HICs vs LMICs), background training and experience.

Our survey highlights shortfalls and heterogeneity in managing infectious diseases in the ICU. ASPs are variously implemented and intensivists reported a need for more ID education.

ASP aims to reduce AMT exposure and to improve patients’ outcomes. Despite valued by most intensivists, ASP implementation remains short.7 In 2012, it was applied in just 58% of hospitals worldwide, with significant variability in policies, strategies, priorities and programmes maturity.8 ASP can vary within the single hospital and while our survey focusses on ICU settings, it draws a similar picture to other hospital settings (1/3 of intensivists reported no ASP).8 Even in developed countries, less than half of French intensivists were aware of ASP, and only 74% of Canadian intensivists reported ASP in their institutions.7,9 A Spanish survey indifferently showed ASP to be implemented nationally in only 37% of ICUs.10

It is becoming crucial to study the barriers impeding ASP in the ICU. While some are shared across all healthcare settings (e.g., funds, staffing, clinicians’ resistance and lack of data technology), others remain ICU specific.7 Most importantly may be the lack of ICU specific ASP studies and guidance.4 Furthermore, conflicts can arise between intensivists’ goals (patient cure) and some perceived ASP targets (AMT use or cost). A local leadership within the ICU can mitigate most conflicts, compensate for staff shortage and provide the support needed for a successful, sustainable and effective ASP. ICU ID champions were shown also to be associated with better ASP implementation.9 In this context, it is noteworthy the high participation of Spanish intensivists in their units’ ASP programmes.10 Such training in addition to participation in audits and research can provide better insight and unify goals (e.g., drug toxicity, AMR and long-term impact).4 However, auditing and feedback are known to be among the least applied ASP measures.7 Last, tele-ASP can be the answer for smaller and remote units suffering staff or expertise shortage.

Heterogeneity in ICU ID/ASP practice is not new to report, and our survey is another confirmation.9,11 LMICs intensivists are less exposed to ASP, have less expert opinion and declared greater knowledge gap and need for training. This may reflect inadequate training or be institutional in origin (e.g., part-time work, smaller or private hospitals, less administrative awareness, lack of data technology and auditing culture).8 ASP was previously shown to be implemented in only 12% of African versus 77% of European hospitals.8 While limited access to AMT risks higher mortality, it also limits choices and proper usage.2 This should raise global concerns as ID and AMR are difficult to contain and a local substandard practice can progress easily across borders.

Despite the availability of expert opinion, intensivists preferred a protocolized over a personalized approach when initiating AMT. An independent AMT prescribers should be able to recognize the need for expertise and co-operate as appropriate.12 However, routine ID consultation is practiced in only 11% of Italian ICUs.13 In Spain, the reference expert is intensivist or anaesthesiologist in 91% of ICUs compared to ID specialist in only 7%.10 Lack of expertise risks prescribing higher tier and broader spectrum antibiotics due to diagnostic uncertainty and higher risk of deterioration in the critically ill.14

While adherence to guidelines is one of the ASP goals, AMR may spread fast and guidelines become outdated.15 Furthermore, guidelines are usually issued in HICs reflecting different resources, staffing, training, and AMR patterns.16 The American Thoracic Society (ATS) acknowledged better AMT optimization through individualization rather than rigid guidelines adherence.4 One observational work suggested adherence to guidelines may be associated with increased mortality.17 If local AMR is monitored, Individualized AMT can be preferred (e.g. hospital acquired infections).4 ICUs are sometimes exempt from AMT restriction (an ASP intervention) while surveillance cultures can further blur the boundaries between colonization and infection.4,8 Novel AMT methods may prove challenging for inexperienced prescribers (e.g., therapeutic drug monitoring [TDM], pharmacokinetics/pharmacodynamics [PK/PD] concept, continuous/extended infusion, nebulized AMT). Holding AMT in stable patients had also been suggested but needs expert support, monitoring and rapid diagnostic tests.18 For all the aforementioned reasons, and while guidelines can provide the minimal satisfactory standard, and unless ICU senior team has a high level of expertise, the role of pharmacist and ID specialists remains indispensable for a tailored management, however difficult to achieve.

Review and ADE usually follow 48–72h after AMT initiation. Such refinement is a precision decision: Individualizing AMT based on microbiological results, imaging and clinical course. In the present survey, intensivists sought non-ICU expertise more frequently on reviewing compared to initiation. However, the practice remained modest in best description (48/72h review, adjustment to culture and discontinuation in 32%, 39% and 21% respectively). The DIANA study already showed ADE to be an exceptional ICU practice (16% of cases).14 Feedback during AMT review represents an ASP learning strategy and we can point up here a lost educational opportunity. However, a meta-analysis demonstrated no impact of feedback on ICU mortality despite reduction of AMT use.19

Traditional microbiological results are usually available 12–72h after AMT initiation. Molecular diagnostic tests provide faster identification of the causative micro-organism and, possibly, antibiotic susceptibility (e.g., genetic testing). If applied along with ASP, such tests can improve mortality while reducing AMT use, costs, toxicity and resistance.20 Only two-thirds of respondents reported good knowledge about hospitals’ microbiological diagnostic capabilities, which was even lower in LMICs, and for the less experienced. Such inadequate understanding, may hinder a proper initial timely testing and make subsequent review harder. Our findings suggest the need for proper orientation of such new technology.4

Collaboration between ID specialists and intensivists may impact patients’ outcome.21–23 A previous survey highlighted the intensivists’ inclination for expert opinion especially where ASP is not in place. However, a considerable percentage of respondents in our work reported the absence of regular ID rounds and variable need for non-ICU expertise according to the settings, background training, and expertise. The involvement of ID specialist in managing Staphylococcus aureus bacteraemia24 and candidemia25 were previously shown to be associated with reduction in AMT duration and mortality. However, only one-third of respondents think blood stream or fungal infections require non-ICU expertise. Again, only a quarter considered consultation for viral infections. Taking into account the diagnostic and therapeutic complexity of those infections, we are not sure if the answers reflect an appropriate level of knowledge. To note that only 20% of UK ICUs had intensivists with special interest in fungal diseases.26 This situation may get worse with the emergence of new infections, increasingly complex AMR and the introduction of novel AMT and microbial tests.

AMR awareness and strengthening knowledge (through training and research) are 2 of 5 objectives of the 2015 WHO global action plan on AMR.27 Both AMR and ASP were recommended as CCM core training competencies.4,27 However, only 42% of intensivists consider their ID training and knowledge adequate. Multivariate analysis showed that LMICs, practice ≤10 years, and dual CCM training with anaesthesia or medicine to be independently associated with less training satisfaction. Nevertheless, the insight of a knowledge gap is positive considering previous reports showing absent acknowledgement of the problem.28,29

Competencies recommended by ICU societies30,31 are becoming short to meet the evolving threat of ID and AMR (e.g., COVID-19) and generally lack an updated detailed cover for ICU ID specificities. The ‘Competency-Based Training in Intensive Care Medicine in Europe’ (CoBaTrICE) comprises 4 ID/AMT competencies addressing microbiological sampling, managing sepsis, AMT and infection control.32 UK, USA and China issued also their national recommendations covering a variety of AMT, microbial testing, infection control and management.31,33,34 However, they remain generally missing in LMICs. Intensivists act usually as independent prescribers and should be able to fulfil the 35 competencies recommended by the European Society of Clinical Microbiology and Infectious Diseases (ESCMID).12 More recently, an ATS workshop recommended ASP as a core ICU competency.4 Pragmatically, 2 tiers of knowledge may be needed: Basic knowledge for every intensivist, and more advanced one for ID-subspecialised intensivists leading ASP, ID research and education. Alternatively, focused ASP training can be implemented when ID expertise is absent (e.g., LMICs). Integrating ID/ASP into daily ICU rounds can improve knowledge and awareness while setting ASP as a daily practice.4 Dual ID and ICM training was also proposed.35

CCM is a multi-disciplinary specialty where diversity and inclusiveness are considered points of strength. Even in developed and closely related countries (e.g., European Union), the training pathways for intensivists are not unified.36 Classically, intensivists are initially trained as anaesthetists or internists (or in any medical subspecialty) before acquiring CCM as super-specialty. Less frequently, CCM is a standalone training. Our data show that intensivists trained solely in CCM are seeking generally less non-ICU expertise and follow more personalized approach on initiating AMT. They are also more confident to manage multi-resistant bacterial and viral infections and slightly more satisfied with their ID training. These findings may reflect more profound practical and theoretical ID training for doctors solely trained as intensivists. In this context, Vidal-Cortés et al. showed that CCM trainees in Spain to have more scheduled time in ID units compared to anaesthetic trainees.10 The different training pathways recently raised concerns about the flexibility of intensivists’ movement across Europe in view of the COVID-19 pandemic.36 We hope our survey by shedding lights on the ICU ID practice can help reforming and unifying ICU ID training.

The fear of missing a causative pathogen leading to adverse outcome is a drive for excessive AMT rather than knowledge deficit.4 Such fear from legal liabilities (i.e., defensive medicine) is emerging to impact practice by the mean of unnecessary referrals, unnecessary tests/prescriptions and avoiding high risk procedures.37 In the AntibioLegalMap survey, 75% of ID and CM specialists reported engagement in some form of defensive behaviour.38 71% of intensivists in our survey admitted a defensive attitude by lowering the threshold to start AMT, or by prescribing longer courses or combinations of AMT (55.2%, 43.8% and 42.7% respectively). Certain defensive behaviours (e.g., over prescription or unnecessary invasive procedures) impose risks to patients, while others (e.g., referral to more specialized physicians) might improve quality of care.37 However, only 21% of respondents think defensive behaviour would lead to seeking non-ICU expertise. Furthermore, AMT may wrongly become a substitute for source control interventions reflecting inadequate awareness of the AMT associated harms.

We admit many limitations of our work. Surveys are subject to “response bias‿ with possible discrepancies between what is self-reported and actual practice. Despite our aim was to explore the intensivists’ insight and opinion, we appreciate microbiologists, ID doctors or pharmacists may have a different point of view and probably better insight for some aspects of the questionnaire. The survey was focused on the practice of ID in the ICU; however, we did not revisit more complex principles like therapeutic drug monitoring and pharmacodynamics/pharmacokinetics which may be of interest as standalone survey.

A point of strength is the number of survey responders; however, the target population remains much bigger. Furthermore, there may be “a participation bias‿ where ID-interested intensivists more likely to participate. Most respondents were from either Europe or Asia and the results might be more representative of these regions. Furthermore, the survey targeted intensivists as persons rather than being institutional, so some respondents may be practicing in the same unit and answers were skewed. Despite these limitations, the diversity of respondents’ countries, settings, experience and training are positive points and we believe our data are clearly highlighting a practice variability and gap of knowledge when managing ID in the ICU worldwide.

Infectious diseases practice is heterogeneous across ICUs underscoring the need for a common set of minimal standards. Intensivists feel the need to improve their ID training and knowledge. Such need is more evident in certain settings such in LMICs, but also for part time and less experienced intensivists. In contrast, doctors with CCM as sole training were more satisfied with ID training. This raises the point for a minimal standard set for ID CCM training. There is also a need to issue ICU specific ASP guidance while encouraging a more precise ID approach.

AR and AE designed and disseminated the survey questionnaire, AS reviewed the questionnaire for intellectual content and is responsible for the statistical analysis. AR and AS wrote the first draft. All authors critically revised the manuscript. All authors read and approved the final manuscript.

Ashraf Roshdy declares owning shares in Astra Zeneca, Pfizer and Merk pharmaceutical companies. He gave also unpaid lecture for Pfizer Company. Ahmad Sabry Saleh and Ahmad Samy Elsayed declares no conflict of interest.

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