Sustained impact of a computer-assisted antimicrobial stewardship intervention on antimicrobial use and length of stay.

Objectives: : Prospective audit and feedback interventions are the core components of an antimicrobial stewardship programme. Herein, we describe the sustained impact of an antimicrobial stewardship programme, based on a novel clinical decision-support system (Antimicrobial Prescription Surveillance System; APSS), on antimicrobial use and costs, hospital length of stay (LOS) in days and the proportion of inappropriate antimicrobial prescriptions. Methods: A quasi-experimental, retrospective study was conducted using interrupted time series between 2008 and 2013. Data on all hospitalized adults receiving antimicrobials were extracted from the data warehouse of a 677 bed academic centre. The intervention started in August 2010. Prospective audit and feedback interventions, led by a pharmacist, were triggered by APSS based on deviations from published and local guidelines. Changes in outcomes before and after the intervention were compared using segmented regression analysis. Results: APSS reviewed 40 605 hospitalizations for 35 778 patients who received antimicrobials. The intervention was associated with a decrease in the average LOS (level change -0.92, P < 0.01; trend -0.08, P < 0.01; intercept 11.4 days), antimicrobial consumption in DDDs/1000 inpatient days (level change -32.4, P < 0.01; trend -1.12, P < 0.02; intercept 243 DDDs per 1000 days of hospitalization), antimicrobial spending in Canadian dollars (level change -19 649, P = 0.01; trend -1881, P < 0.01; intercept $74 683) and proportion of non-concordance with local guidelines for prescribing antimicrobials (level change -2.3, P = 0.04; intercept 41%). Conclusions: The implementation of the APSS-initiated strategy was associated with a positive impact on antimicrobial use and spending, LOS and inappropriate prescriptions. The high rate of accepted interventions may have contributed to these results.

Impact of guideline-consistent therapy on outcome of patients with healthcare-associated and community-acquired pneumonia.

BACKGROUND: A new category of healthcare-associated pneumonia (HCAP) has been added in the most recent American Thoracic Society/Infectious Diseases Society of America (ATS/IDSA) guidelines, since multidrug-resistant (MDR) pathogens are more common in patients with HCAP than in those with community-acquired pneumonia (CAP). The optimal empirical management of patients with HCAP remains controversial and adherence to guidelines is inconsistent. METHODS: A retrospective cohort study of 3295 adults admitted for pneumonia in an academic centre of Canada, between 1997 and 2008. RESULTS: MDR pathogens were more common among patients with HCAP than in those with CAP, but less so than in other studies. Compared with patients with CAP, those with HCAP had a higher all-cause 30 day mortality [68/563 (12%) versus 201/2732 (7%); P < 0.001] and more frequent need for mechanical ventilation [78/563 (14%) versus 276/2732 (10%); P = 0.01]. In patients with CAP, mortality was lower when treatment was concordant with guidelines [86/1557 (6%) versus 109/1097 (10%) if discordant; adjusted odds ratio 0.6 (0.4-0.8); P < 0.001]. In HCAP, mortality was similar whether or not empirical treatment was concordant with guidelines [6/35 (17%) versus 18/148 (12%) if discordant; P = 0.4]. However, 30 day mortality tended to be higher when the empirical treatment was microbiologically ineffective [4/22 (18%) versus 17/187 (9%) when effective; P = 0.3]. CONCLUSIONS: HCAP is associated with worse outcomes than CAP. MDR pathogens were implicated in only a small fraction of HCAP cases. In our study, unlike CAP, non-respect of current HCAP guidelines had no adverse effect on the ultimate outcome. Strategies for the empirical management of HCAP should be tailored to the local epidemiological context.

Clinical impact of accepting or rejecting a recommendation from a clinical decision support system-assisted antibiotic stewardship program.

Background: Outcomes associated with physician responses to recommendations from an antimicrobial stewardship program (ASP) at an individual patient level have not yet been assessed. We aimed to compare clinical characteristics and mortality risk among patients for whom recommendations from an ASP were accepted or refused. Methods: A prospective cohort study was performed with hospitalized adults who received intravenous or oral antimicrobials at a 677-bed academic centre in Canada in 2014-2017. We included patients with an alert produced by a clinical decision support system (CDSS) for whom a recommendation was made by the pharmacist to the attending physician. The outcome was 90-day in-hospital all-cause mortality. Results: We identified 3,197 recommendations throughout the study period, of which 2,885 (90.2%) were accepted. The median length of antimicrobial treatment was significantly shorter when a recommendation was accepted (0.26 versus 1.78 d; p < 0.001). Refusal of a recommendation was not associated with mortality (odds ratio 1.32; 95% confidence interval, 0.93 to 1.89; p = 0.12). The independent risk factors associated with in-hospital mortality were age, Charlson Comorbidity Index score, admission to a critical care unit, duration between admission and recommendation, and issuance of a recommendation on a carbapenem. Conclusions: The duration of antimicrobial treatment was significantly shorter when a recommendation originating from a CDSS-assisted ASP program was accepted. Future prospective studies including potential residual confounding variables, such as the source of infection or physiological derangement, might help in understanding whether CDSS-assisted ASP will have a direct impact on patient mortality.

Historique: Les résultats liés aux réponses des médecins aux recommandations du programme de gestion antimicrobienne (PGA) n'ont pas encore été évalués à l'égard de chaque patient. Les chercheurs ont visé à comparer les caractéristiques cliniques et le risque de mortalité chez les patients dont les recommandations provenant d'une PGA ont été acceptées ou refusées. Méthodologie: Les chercheurs ont procédé à une étude de cohorte prospective auprès d'adultes hospitalisés qui avaient reçu des antimicrobiens par voie intraveineuse ou orale à un centre universitaire de 667 lits composé de deux établissements du Canada entre 2014 et 2017. Ils ont inclus les patients pour qui s'était déclenchée une alerte produite par un système d'aide à la décision clinique (SADC) et pour qui le pharmacien avait fait une recommandation au médecin traitant. Le résultat était la mortalité toutes causes confondues après un séjour hospitalier de 90 jours. Résultats: Les chercheurs ont extrait 3 197 recommandations tout au long de l'étude, dont 2 885 (90,2 %) ont été acceptées. La durée médiane du traitement antimicrobien était considérablement plus courte lorsqu'une recommandation était acceptée (0,26 par rapport à 1,78 jour; p < 0,001). Le refus d'une recommandation n'était pas associé à la mortalité (rapport de cotes de 1,32; IC de 95 %, 0,93 à 1,89; p = 0,12). Les facteurs de risque indépendants associés à la mortalité en milieu hospitalier étaient l'âge, l'indice de Charlson, l'admission dans une unité de soins intensifs, la période entre l'admission et la recommandation, et la formulation d'une recommandation sur un carbapénem. Conclusion: La durée du traitement antimicrobien était beaucoup plus courte lorsque la recommandation d'un PGA assisté par un SADC était acceptée. De futures études prospectives incluant de potentielles variables confusionnelles résiduelles, telles que la source de l'infection ou le dérangement physiologique, pourraient contribuer à établir si un PGA assisté par un SADC aura des conséquences directes sur la mortalité des patients.

Antimicrobial stewardship in acute care centres: A survey of 68 hospitals in Quebec.

BACKGROUND: Antimicrobial stewardship programs (ASPs) and quantitative monitoring of antimicrobial use are required to ensure that antimicrobials are used appropriately in the acute care setting, and have the potential to reduce costs and limit the spread of antimicrobial-resistant organisms and Clostridium difficile. Currently, it is not known what proportion of Quebec hospitals have an ASP and/or monitor antimicrobial use. OBJECTIVES: To determine what proportion of Quebec hospitals have an ASP, and what is the nature of such a program. METHODS: A detailed questionnaire was sent to the pharmacy directors of all acute care hospitals in the province of Quebec. Information was collected on antimicrobial surveillance; antimicrobial stewardship and resource allocation to these areas were assessed. RESULTS: Questionnaires were completed for 68 of 81 (84%) hospitals contacted. ASPs were identified at 50 (74%) hospitals, but only 20 (29%) of hospitals had a quantitative antimicrobial surveillance program (QASP) in 2006. Academic centres (P=0.03) and hospitals with over 200 beds (P=0.02) were more likely to have a QASP. Even among hospitals with an ASP, 18% had less than one full-time pharmacist for a QASP. CONCLUSIONS: Over one-quarter of Quebec hospitals do not have an ASP, and few hospitals in Quebec are currently evaluating their use of antimicrobials on a quantitative basis. In some cases, the lack of a QASP may be due to the allocation of insufficient pharmaceutical resources to antimicrobial stewardship (ie, less than one full-time pharmacist).

Evaluation of a machine learning capability for a clinical decision support system to enhance antimicrobial stewardship programs.

OBJECTIVE: Antimicrobial stewardship programs have been shown to limit the inappropriate use of antimicrobials. Hospitals are increasingly relying on clinical decision support systems to assist in the demanding prescription reviewing process. In previous work, we have reported on an emerging clinical decision support system for antimicrobial stewardship that can learn new rules supervised by user feedback. In this paper, we report on the evaluation of this system. METHODS: The evaluated system uses a knowledge base coupled with a supervised learning module that extracts classification rules for inappropriate antimicrobial prescriptions using past recommendations for dose and dosing frequency adjustments, discontinuation of therapy, early switch from intravenous to oral therapy, and redundant antimicrobial spectrum. Over five weeks, the learning module was deployed alongside the baseline system to prospectively evaluate its ability to discover rules that complement the existing knowledge base for identifying inappropriate prescriptions of piperacillin-tazobactam, a frequently used antimicrobial. RESULTS: The antimicrobial stewardship pharmacists reviewed 374 prescriptions, of which 209 (56% of 374) were identified as inappropriate leading to 43 recommendations to optimize prescriptions. The baseline system combined with the learning module triggered alerts in 270 prescriptions with a positive predictive value of identifying inappropriate prescriptions of 74%. Of these, 240 reviewed prescriptions were identified by the alerts of the baseline system with a positive predictive value of 82% and 105 reviewed prescriptions were identified by the alerts of the learning module with a positive predictive value of 62%. The combined system triggered alerts for all 43 recommendations, resulting in a rate of actionable alerts of 16% (43 recommendations of 270 reviewed alerts); the baseline system triggered alerts for 38 interventions, resulting in a rate of actionable alerts of 16% (38 of 240 reviewed alerts); and the learning module triggered alerts for 17 interventions, resulting in a rate of actionable alerts of 16% (17 of 105 reviewed alerts). The learning module triggered alerts for every inappropriate prescription missed by the knowledge base of the baseline system (n=5). CONCLUSIONS: The learning module was able to extract clinically relevant rules for multiple types of antimicrobial alerts. The learned rules were shown to extend the knowledge base of the baseline system by identifying pharmacist interventions that were missed by the baseline system. The learned rules identified inappropriate prescribing practices that were not supported by local experts and were missing from its knowledge base. However, combining the baseline system and the learning module increased the number of false positives.