Impact of interactive computerised decision support for hospital antibiotic use (COMPASS): an open-label, cluster-randomised trial in three Swiss hospitals.

BACKGROUND: Computerised decision-support systems (CDSSs) for antibiotic stewardship could help to assist physicians in the appropriate prescribing of antibiotics. However, high-quality evidence for their effect on the quantity and quality of antibiotic use remains scarce. The aim of our study was to assess whether a computerised decision support for antimicrobial stewardship combined with feedback on prescribing indicators can reduce antimicrobial prescriptions for adults admitted to hospital. METHODS: The Computerised Antibiotic Stewardship Study (COMPASS) was a multicentre, cluster-randomised, parallel-group, open-label superiority trial that aimed to assess whether a multimodal computerised antibiotic-stewardship intervention is effective in reducing antibiotic use for adults admitted to hospital. After pairwise matching, 24 wards in three Swiss tertiary-care and secondary-care hospitals were randomised (1:1) to the CDSS intervention or to standard antibiotic stewardship measures using an online random sequence generator. The multimodal intervention consisted of a CDSS providing support for choice, duration, and re-evaluation of antimicrobial therapy, and feedback on antimicrobial prescribing quality. The primary outcome was overall systemic antibiotic use measured in days of therapy per admission, using adjusted-hurdle negative-binomial mixed-effects models. The analysis was done by intention to treat and per protocol. The study was registered with ClinicalTrials.gov (identifier NCT03120975). FINDINGS: 24 clusters (16 at Geneva University Hospitals and eight at Ticino Regional Hospitals) were eligible and randomly assigned to control or intervention between Oct 1, 2018, and Dec 31, 2019. Overall, 4578 (40·2%) of 11 384 admissions received antibiotic therapy in the intervention group and 4142 (42·8%) of 9673 in the control group. The unadjusted overall mean days of therapy per admission was slightly lower in the intervention group than in the control group (3·2 days of therapy per admission, SD 6·2, vs 3·5 days of therapy per admission, SD 6·8; p<0·0001), and was similar among patients receiving antibiotics (7·9 days of therapy per admission, SD 7·6, vs 8·1 days of therapy per admission, SD 8·4; p=0·50). After adjusting for confounders, there was no statistically significant difference between groups for the odds of an admission receiving antibiotics (odds ratio [OR] for intervention vs control 1·12, 95% CI 0·94-1·33). For admissions with antibiotic exposure, days of therapy per admission were also similar (incidence rate ratio 0·98, 95% CI 0·90-1·07). Overall, the CDSS was used at least once in 3466 (75·7%) of 4578 admissions with any antibiotic prescription, but from the first day of antibiotic treatment for only 1602 (58·9%) of 2721 admissions in Geneva. For those for whom the CDSS was not used from the first day, mean time to use of CDSS was 8·9 days. Based on the manual review of 1195 randomly selected charts, transition from intravenous to oral therapy was significantly more frequent in the intervention group after adjusting for confounders (154 [76·6%] of 201 vs 187 [87%] of 215, +10·4%; OR 1·9, 95% CI 1·1-3·3). Consultations by infectious disease specialists were less frequent in the intervention group (388 [13·4%] of 2889) versus the control group (405 [16·9%] of 2390; OR 0·84, 95% CI 0·59-1·25). INTERPRETATION: An integrated multimodal computerised antibiotic stewardship intervention did not significantly reduce overall antibiotic use, the primary outcome of the study. Contributing factors were probably insufficient uptake, a setting with relatively low antibiotic use at baseline, and delays between ward admission and first CDSS use. FUNDING: Swiss National Science Foundation. TRANSLATIONS: For the French and Italian translations of the abstract see Supplementary Materials section.

Machine Learning for Automatic Encoding of French Electronic Medical Records: Is More Data Better?

The encoding of Electronic Medical Records is a complex and time-consuming task. We report on a machine learning model for proposing diagnoses and procedures codes, from a large realistic dataset of 245 000 electronic medical records at the University Hospitals of Geneva. Our study particularly focuses on the impact of training data quantity on the model's performances. We show that the performances of the models do not increase while encoded instances from previous years are exploited for learning data. Furthermore, supervised models are shown to be highly perishable: we show a potential drop in performances of around -10% per year. Consequently, great and constant care must be exercised for designing and updating the content of such knowledge bases exploited by machine learning.

How to Develop and Implement a Computerized Decision Support System Integrated for Antimicrobial Stewardship? Experiences From Two Swiss Hospital Systems.

Background: Computerized decision support systems (CDSS) provide new opportunities for automating antimicrobial stewardship (AMS) interventions and integrating them in routine healthcare. CDSS are recommended as part of AMS programs by international guidelines but few have been implemented so far. In the context of the publicly funded COMPuterized Antibiotic Stewardship Study (COMPASS), we developed and implemented two CDSSs for antimicrobial prescriptions integrated into the in-house electronic health records of two public hospitals in Switzerland. Developing and implementing such systems was a unique opportunity for learning during which we faced several challenges. In this narrative review we describe key lessons learned. Recommendations: (1) During the initial planning and development stage, start by drafting the CDSS as an algorithm and use a standardized format to communicate clearly the desired functionalities of the tool to all stakeholders. (2) Set up a multidisciplinary team bringing together Information Technologies (IT) specialists with development expertise, clinicians familiar with "real-life" processes in the wards and if possible, involve collaborators having knowledge in both areas. (3) When designing the CDSS, make the underlying decision-making process transparent for physicians and start simple and make sure to find the right balance between force and persuasion to ensure adoption by end-users. (4) Correctly assess the clinical and economic impact of your tool, therefore try to use standardized terminologies and limit the use of free text for analysis purpose. (5) At the implementation stage, plan usability testing early, develop an appropriate training plan suitable to end users' skills and time-constraints and think ahead of additional challenges related to the study design that may occur (such as a cluster randomized trial). Stay also tuned to react quickly during the intervention phase. (6) Finally, during the assessment stage plan ahead maintenance, adaptation and related financial challenges and stay connected with institutional partners to leverage potential synergies with other informatics projects.

Study protocol for a multicentre, cluster randomised, superiority trial evaluating the impact of computerised decision support, audit and feedback on antibiotic use: the COMPuterized Antibiotic Stewardship Study (COMPASS).

INTRODUCTION: Inappropriate use of antimicrobials in hospitals contributes to antimicrobial resistance. Antimicrobial stewardship (AMS) interventions aim to improve antimicrobial prescribing, but they are often resource and personnel intensive. Computerised decision supportsystems (CDSSs) seem a promising tool to improve antimicrobial prescribing but have been insufficiently studied in clinical trials. METHODS AND ANALYSIS: The COMPuterized Antibiotic Stewardship Study trial, is a publicly funded, open-label, cluster randomised, controlled superiority trial which aims to determine whether a multimodal CDSS intervention integrated in the electronic health record (EHR) reduces overall antibiotic exposure in adult patients hospitalised in wards of two secondary and one tertiary care centre in Switzerland compared with 'standard-of-care' AMS. Twenty-four hospital wards will be randomised 1:1 to either intervention or control, using a 'pair-matching' approach based on baseline antibiotic use, specialty and centre. The intervention will consist of (1) decision support for the choice of antimicrobial treatment and duration of treatment for selected indications (based on indication entry), (2) accountable justification for deviation from the local guidelines (with regard to the choice of molecules and duration), (3) alerts for self-guided re-evaluation of treatment on calendar day 4 of antimicrobial therapy and (4) monthly ward-level feedback of antimicrobial prescribing indicators. The primary outcome will be the difference in overall systemic antibiotic use measured in days of therapy per admission based on administration data recorded in the EHR over the whole intervention period (12 months), taking into account clustering. Secondary outcomes include qualitative and quantitative antimicrobial use indicators, economic outcomes and clinical, microbiological and patient safety indicators. ETHICS AND DISSEMINATION: Ethics approval was obtained for all participating sites (Comission Cantonale d'Éthique de la Recherche (CCER)2017-00454). The results of the trial will be submitted for publication in a peer-reviewed journal. Further dissemination activities will be presentations/posters at national and international conferences. TRIAL REGISTRATION NUMBER: NCT03120975; Pre-results.

The health and economic burden of bloodstream infections caused by antimicrobial-susceptible and non-susceptible Enterobacteriaceae and Staphylococcus aureus in European hospitals, 2010 and 2011: a multicentre retrospective cohort study.

We performed a multicentre retrospective cohort study including 606,649 acute inpatient episodes at 10 European hospitals in 2010 and 2011 to estimate the impact of antimicrobial resistance on hospital mortality, excess length of stay (LOS) and cost. Bloodstream infections (BSI) caused by third-generation cephalosporin-resistant Enterobacteriaceae (3GCRE), meticillin-susceptible (MSSA) and -resistant Staphylococcus aureus (MRSA) increased the daily risk of hospital death (adjusted hazard ratio (HR) = 1.80; 95% confidence interval (CI): 1.34-2.42, HR = 1.81; 95% CI: 1.49-2.20 and HR = 2.42; 95% CI: 1.66-3.51, respectively) and prolonged LOS (9.3 days; 95% CI: 9.2-9.4, 11.5 days; 95% CI: 11.5-11.6 and 13.3 days; 95% CI: 13.2-13.4, respectively). BSI with third-generation cephalosporin-susceptible Enterobacteriaceae (3GCSE) significantly increased LOS (5.9 days; 95% CI: 5.8-5.9) but not hazard of death (1.16; 95% CI: 0.98-1.36). 3GCRE significantly increased the hazard of death (1.63; 95% CI: 1.13-2.35), excess LOS (4.9 days; 95% CI: 1.1-8.7) and cost compared with susceptible strains, whereas meticillin resistance did not. The annual cost of 3GCRE BSI was higher than of MRSA BSI. While BSI with S. aureus had greater impact on mortality, excess LOS and cost than Enterobacteriaceae per infection, the impact of antimicrobial resistance was greater for Enterobacteriaceae.

Healthcare-associated infections are associated with insufficient dietary intake: an observational cross-sectional study.

BACKGROUND: Indicators to predict healthcare-associated infections (HCAI) are scarce. Malnutrition is known to be associated with adverse outcomes in healthcare but its identification is time-consuming and rarely done in daily practice. This cross-sectional study assessed the association between dietary intake, nutritional risk, and the prevalence of HCAI, in a general hospital population. METHODS AND FINDINGS: Dietary intake was assessed by dedicated dieticians on one day for all hospitalized patients receiving three meals per day. Nutritional risk was assessed using Nutritional Risk Screening (NRS)-2002, and defined as a NRS score ≥ 3. Energy needs were calculated using 110% of Harris-Benedict formula. HCAIs were diagnosed based on the Center for Disease Control criteria and their association with nutritional risk and measured energy intake was done using a multivariate logistic regression analysis. From 1689 hospitalised patients, 1024 and 1091 were eligible for the measurement of energy intake and nutritional risk, respectively. The prevalence of HCAI was 6.8%, and 30.1% of patients were at nutritional risk. Patients with HCAI were more likely identified with decreased energy intake (i.e. ≤ 70% of predicted energy needs) (30.3% vs. 14.5%, P = 0.002). The proportion of patients at nutritional risk was not significantly different between patients with and without HCAI (35.6% vs.29.7%, P = 0.28), respectively. Measured energy intake ≤ 70% of predicted energy needs (odds ratio: 2.26; 95% CI: 1.24 to 4.11, P = 0.008) and moderate severity of the disease (odds ratio: 3.38; 95% CI: 1.49 to 7.68, P = 0.004) were associated with HCAI in the multivariate analysis. CONCLUSION: Measured energy intake ≤ 70% of predicted energy needs is associated with HCAI in hospitalised patients. This suggests that insufficient dietary intake could be a risk factor of HCAI, without excluding reverse causality. Randomized trials are needed to assess whether improving energy intake in patients identified with decreased dietary intake could be a novel strategy for HCAI prevention.

Assessing the relationships between hospital resources and activities: a systematic review.

Face the challenge of minimizing their resource utilization without reducing the quality of healthcare. Achieving this aim requires precise analysis and optimization of various inputs and outputs. This paper presents a systematic review of the relationships between hospital resources (considered productivity inputs) and financial and activity outcomes (considered productivity outputs). Several electronic bibliographic databases and the Internet were searched for articles published between January 1990 and December 2013 that examined the relationships between hospital resources and financial and activity outcomes. We assessed the quality of the study design, the nature of the sample, input and output indicators, and the statistical methods used for each selected study. Thirty-eight original papers were selected. Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA) were the most statistical methods used. Based on our analysis, we retained 18 input and 19 output indicators that could constitute the basis for hospital productivity benchmarking. Selecting a small set of shared economic and activity indicators is relevant for assessing the productivity of a hospital, measuring trends and performing national or international benchmarking. Such indicators should be combined with quality measures for a comprehensive evaluation approach.

The CASSANDRE Project: automated alerts for optimal coding of diagnosis and interventions.

As of 1 January 2012,all Swiss hospitals have had to charge acute somatic care hospitalization according to the Swiss disease related group (DRG) System. In this system, hospital bills are based on the discharge summaries. Coders analyze these in order to identify diagnostic and interventional codes. These codes are used by the system grouper to determine a specific DRG code and cost-weight. The amount to be charged per episode is based on this cost-weight. Since acute care billing relies on discharge summaries and knowing that these are incomplete, our aim was to inprove the completeness of these documents by automatically detecting pathologies that should have been coded and charged. We also aimed to help improve the selection of the main diagnosis. We have implemented algorithms for the automatic detection of pathologies that directly inform the coders whilst by-passing the physician. Final validation of the new pathologies remains with the physician. Our results are very encouraging from a financial point of view.

Interoperability in hospital information systems: a return-on-investment study comparing CPOE with and without laboratory integration.

Despite its many advantages, using a computerized patient record is still considered as a time consuming activity for care providers. In numerous situations, time is wasted because of the lack of interoperability between systems. In this study, we aim to assess the time gains that nursing teams could achieve with a tightly integrated computerized order entry system. Using a time-motion method, we compared expected versus effective time spent managing laboratory orders for two different computerized systems: one integrated, the other not integrated. Our results tend to show that nurses will complete their task an average of five times faster than their expected performance (p<0.001). We also showed that a tightly integrated system provides a threefold speed gain for nurses compared to a non-integrated CPOE with the laboratory information system (p<0.001). We evaluated the economic benefit of this gain, therefore arguing for a strong interoperability of systems, in addition to patient safety benefits.

Optimal asymmetrical SVM using pattern search. A health care application.

This paper considers the model selection problem for Support Vector Machines. A well-known derivative Pattern Search method, which aims to tune hyperparameter values using an empirical error estimate as a steering criterion, is proposed. This approach is experimentally evaluated on a health care problem which involves discriminating nosocomially infected patients from non-infected patients. The Hooke and Jeeves Pattern Search (HJPS) method is shown to improve the results achieved by Grid Search (GS) in terms of solution quality and computational efficiency. Unlike most other parameter tuning techniques, our approach does not require supplementary effort such as computation of derivatives, making them well suited for practical purposes. This method produces encouraging results: it exhibits good performance and convergence properties.

Measuring the effectiveness of hospital-acquired infection prevention.

This article deals with data on nosocomial infections acquired in the Geneva University Hospitals. Goal of the work is to derive a model from a hospital-acquired infection (HAI) prevalence survey of year Y and apply them to a prevalence survey of years Y+1, Y+2. This analysis permits to evaluate the effectiveness of preventive measures taken after the prevalence survey in year Y. It also analyzes the robustness of the SVM algorithm on time-variable attributes. The model build on the dataset of year Y gives better results than in a previous study. The application of the model on the Y+1 and Y+2 prevalence surveys shows simultaneously improvements and deteriorations of 5 performance measures. This highlights the effectiveness of prevention and reduces the risk of HAI after the prevalence survey of year Y. We introduce a new method to detect redundancy in a dataset with the SVM algorithm.

@neurIST: infrastructure for advanced disease management through integration of heterogeneous data, computing, and complex processing services.

The increasing volume of data describing human disease processes and the growing complexity of understanding, managing, and sharing such data presents a huge challenge for clinicians and medical researchers. This paper presents the @neurIST system, which provides an infrastructure for biomedical research while aiding clinical care, by bringing together heterogeneous data and complex processing and computing services. Although @neurIST targets the investigation and treatment of cerebral aneurysms, the system's architecture is generic enough that it could be adapted to the treatment of other diseases. Innovations in @neurIST include confining the patient data pertaining to aneurysms inside a single environment that offers clinicians the tools to analyze and interpret patient data and make use of knowledge-based guidance in planning their treatment. Medical researchers gain access to a critical mass of aneurysm related data due to the system's ability to federate distributed information sources. A semantically mediated grid infrastructure ensures that both clinicians and researchers are able to seamlessly access and work on data that is distributed across multiple sites in a secure way in addition to providing computing resources on demand for performing computationally intensive simulations for treatment planning and research.

Choosing the right amount of healthcare information technologies investments.

OBJECTIVES: Choosing and justifying the right amount of investment in healthcare information technologies (HITECH or HIT) in hospitals is an ever increasing challenge. Our objectives are to assess the financial impact of HIT on hospital outcome, and propose decision-helping tools that could be used to rationalize the distribution of hospital finances. DESIGN: We used a production function and microeconomic tools on data of 21 Paris university hospitals recorded from 1998 to 2006 to compute the elasticity coefficients of HIT versus non-HIT capital and labor as regards to hospital financial outcome and optimize the distribution of investments according to the productivity associated with each input. RESULTS: HIT inputs and non-HIT inputs both have a positive and significant impact on hospital production (elasticity coefficients respectively of 0.106 and 0.893; R(2) of 0.92). We forecast 2006 results from the 1998 to 2005 dataset with an accuracy of +0.61%. With the model used, the best proportion of HIT investments was estimated to be 10.6% of total input and this was predicted to lead to a total saving of 388 million Euros for the 2006 dataset. CONCLUSION: Considering HIT investment from the point of view of a global portfolio and applying econometric and microeconomic tools allow the required confidence level to be attained for choosing the right amount of HIT investments. It could also allow hospitals using these tools to make substantial savings, and help them forecast their choices for the following year for better HITECH governance in the current stimulation context.

Assessing the capital efficiency of healthcare information technologies investments: an econometric perspective.

OBJECTIVES: To examine the different methods that can be used in the quantification of the added value of information technologies (IT) in the health care sector. This quantification represents a major issue for decision-makers and health care professionals when they have to plan an IT investment. METHODS: Articles were chosen via Medline, internet and the University of Geneva bibliographic portal. Some of the papers were obtained directly from their authors. We examine the most current methods used to evaluate IT return on investment (ROI) in the general business and in the health care sector, drawing attention on methods traditionally used in macroeconomic studies that could reveal themselves disruptive for IT ROI impact evaluation in hospitals. RESULTS: Financial and accounting methods can provide interesting data on a specific IT project but are usually incomplete for revealing the global IT investment influence. Econometric methods tend to demonstrate the positive impact of health care IT (HIT) on hospital production and productivity. Hospitals having higher levels of IT investment tend to deliver a higher level of clinical quality and show improved hospital cost performances. CONCLUSIONS: Information technologies are so intermingled with people and processes that the identification of specific IT benefit remains questionable. Using macroeconomic tools could be the best way to analyze and compute IT ROI in health care. Econometric tools take into account all types investments (inputs) and all the returns (outputs) enabling the precise measurement of IT investments impact, breakeven points, and possible threshold levels, thus providing helpful intelligence to reach the higher levels of IT governance in hospitals.

Impact of health care information technology on hospital productivity growth: a survey in 17 acute university hospitals.

The quantification of the added value of information technologies (IT) in the health sector is a major issue for decision-makers and health care professionals. This paper relates the application of an economic production function in hospitals with different integration levels of their clinical information systems (CIS). The study concerns 17 university hospitals within the Assistance Publique Hôpitaux de Paris group that were followed from 1998 to 2005. Using an extended Cobb-Douglas production function, yearly incomes (Y) were correlated with three inputs: capital (K), labor (L) and IT expenses (T). The results indicate that incomes are significantly and positively associated with the three input variables with elasticity coefficients: alpha, beta and gamma of 0.81, 0.17, and 0.09 that appear to be in the range of values found in secondary and tertiary sectors. The IT elasticity coefficient (gamma) is higher in the subgroup of 6 hospitals that integrate, or started to integrate, a complete CIS within the study period than in the 11 reference hospitals. In a general production function, hospital costs appear to be positively connected to the level of IT expenses, capital and labor. Calculations in two subgroups of AP-HP hospitals divided according to the importance of the IT integration level seem to indicate that the more the clinical information system is integrated, the more its influence is positive in hospital production. The results of this first survey are sufficiently encouraging to try to refine them (better granularity) and to spread them in time (over a longer period) and space (to other hospital structures).

Nasal septal perforations must and can be closed / Perfuraçöes septais devem e podem ser cerradas

According to the experience of the main author extending over more than 20 years (more than 100 cases) in the treatm,ent of septal perforations, we have the deep conviction that all techniques using local flaps should actually be considered to be obsolete, as the blood supply is insufficient. Perforations with diameter up to 4cm can be closed in a one stage procedure with extensive disection of the mucoperichondrium and mucoperiosteum of the septum, the vault and the vestibular and cavity floor. The hole in the quadrangular plate can be reduced by a push back or push down manoeuver of part of the cartilage. For closure of perforations larger than 4 cm in diameter, we use a 2-3 stage procedure using a composite bucal flap with three layers (mucosa0cartilage and mucosa) and a gingivolabial pedicle. Thus all kinds and size of perfrations can be closed surgically avoiding the use of silicone obturatirs which only enlarge the hole