5PSQ-033\u2005Effectiveness of computerised decision support system-based intervention in antimicrobial use: the HIGEA project

Abstract

Background Clinical decision support systems (CDSS) can play an important role in facilitating antimicrobial stewardship programmes (ASP). However, the effects of CDSS in improving antimicrobial therapy have been insufficiently studied. Purpose To evaluate the impact of an automated/integrated real-time CDSS called HIGEA for antimicrobial stewardship-related interventions. Material and methods This was a prospective descriptive study performed in a 1300-bed tertiary teaching hospital in Madrid. A CDSS was developed integrating microbiology data, laboratory data and the computerised prescription order system. The integration was performed using a standard language (HL7). The system generates alerts based on predefined clinical rules (CR) to select patients in whom antimicrobial therapy can be improved. Alerts are reviewed daily by an infectious disease pharmacist, who makes recommendations of the necessary changes on the treatment to the physician. Eight custom-built CR that promote stop/de-escalation of therapy were evaluated in the initial ASP review during 1 April 2017–31 August 2017. Data collection included total number of actionable alerts, recommendations provided and acceptance rates. For each CR, the positive predictive value (PPV) was calculated as the ratio of modifications in treatment to alerts reviewed. The severity of medication errors prevented and antimicrobial consumption were also analysed. Results In total, 701 alerts were reviewed during the study period (6.4 alerts per day). Overall, 419 (60%) alerts were actionable. The acceptance rate was 77% (321/419) and the PPV 0.46. The CR that induced the highest number of treatment changes was ‘treatment with penicillins/cephalosporins/quinolones>7 days’ (PPV=0.58), followed by ‘switch to oral therapy with quinolones/linezolid/azole’ (PPV=0.31), ‘Streptococcus/Enterococcus+carbapenem’ (PPV=0.70) and ‘candin +fluconazole\u2009sensitive Candida (PPV=0.82). Accordingly, the most common interventions were discontinuation of treatment (60%), switch to oral therapy (20%) and de-escalation (12%). Overall, 14% of errors intercepted were classified as being of moderate severity and 9.4% as serious. A significant reduction in the consumption of quinolones was achieved (from 15.0 to 12.6 defined daily doses/100 patient-days), with no significant change in the consumption of other antibiotics. Conclusion HIGEA has identified opportunities to optimize antimicrobial use. Future work must aim to incorporate new custom-built clinical rules, including those to alert the need for prompt initiation of antimicrobial therapy. References and/or acknowledgements Hosp Pharm 2017;52:679–84. No conflict of interest,