Introduction Antimicrobial resistance is a

Introduction Antimicrobial resistance is a growing health threat worldwide that has been associated with incorrect prescription and overuse of rki [1]. Considering that physicians play a central role in antibiotic usage, a better understanding of their prescribing behaviours and their knowledge of antimicrobial resistance is critical to design interventions that will enable implementation of effective antibiotic stewardship programmes (ASPs) [2], [3]. Despite some publications on antibiotic prescribing behaviour [4], [5], scarce data are available on prescribing practices in Portugal [6], where antibiotic consumption is still higher than the European average [7]. In addition, results from other countries may not be applicable since it has been proven that sociocultural factors impact antibiotic prescription [8]. The present study was developed under the scope of the HAITooL project (http://haitool.ihmt.unl.pt/). The main goals of the current study were to evaluate and understand physicians’ knowledge and prescription behaviours on antibiotics in two Portuguese hospitals and to identify difficulties and barriers in the antibiotic prescription process in order to allow the proper design and implementation of an ASP adapted to the Portuguese context.
Methods
Results
Discussion One-third of respondents did not recognise antibiotic resistance as a serious problem in their own hospital. This lack of awareness has been also described by several other authors [3], [5] and should be consider when ASPs are designed. In this study, the level of confidence of participating physicians was high (64.3%) when prescribing antibiotics, however it was less than reported in other studies [4]. Nevertheless, the confidence level in the microbiology laboratory should be highlighted: despite almost one-half of respondents identifying the (no) reporting of microbiological results/data (antibiotic susceptibility patterns and local epidemiological data) as a barrier to the antibiotic prescription process, the level of confidence in antibiotic prescription reached 100% when microbiology results were known. A positive association between antibiotic susceptibility reporting from microbiology laboratories and antibiotic prescribing has been described by other authors [10]. Based on that, good communication between physicians and the microbiology laboratory should be considered as a priority in any ASP [11]. Two major factors that influence antibiotic prescription were identified, namely ‘co-morbidities’ and ‘patient clinical situation’, as described in other studies [12]. However, it is important to notice that in contrast to other studies [13], participant physicians did not consider the influence of peers and senior colleagues as an important determinant in antibiotic prescription; it was even pointed out as a barrier. These results can be explained by the fact that 35.7% of participating physicians were senior doctors (>20 years after graduation) or by national cultural dimensions [14]. Despite some studies reporting no improvements following educational interventions (reviewed in [15]), education has been widely recognised as a successful strategy for rational antibiotic prescription [15], [16]. In the current study, education and training was pointed out as the most effective strategy to improve antibiotic prescription, as also described in England, France and Scotland [4], [5]. Thus, and according to the European guidelines [17], inclusion of lessons on appropriate antimicrobial use and antimicrobial resistance in medical school’s programmes, in parallel with regular training for practicing physicians, should be reinforced. Computerised decision-support systems have been described as an advantage in antibiotic prescription processes [18]. However, in the current study only one physician suggested it as an efficient strategy to improve antibiotic prescription, probably because physicians often describe information systems as obsolete (A.S. Simões et al., unpublished observations). To avoid this, the information system should be supported by end-user inputs [19], as happens in the HAITooL surveillance and decision-support system that has been co-designed and implemented with physicians in order to fulfil their needs and requirements for the antibiotic prescription process [20].