Impact of an Innovative Smartphone “App” on Intensive Care Unit (ICU) Antimicrobial Utilization Across a Large Metropolitan Health Region in Canada: An Interrupted Time Series Analysis

Abstract

Background. Spectrum MD is a customizable stewardship smartphone app that offers algorithmic decision support based on clinical guidelines and local microbiology data. The app was introduced to all ICU prescribers across a large metropolitan health region followed by an interrupted time series analysis (ITSA) to evaluate the impact on ICU antimicrobial utilization (AMU). Methods. Spectrum MD was launched 11 January 2013 to ICU prescribers with monthly educational sessions. Penetration of the app was tracked with Flurry Analytics, a post-intervention survey, and via session attendance. An uncontrolled beforeafter intervention ITSA (SPSS v19.0) was performed to compare AMU in the 12 months pre- and post-intervention (allowing a 3 month wash-in period) using daily defined doses/100 patient days. Patient demographics, central line associated bloodstream infection (CLABSI) and influenza rates were analyzed for both time periods. Student's t-test was used for continuous variables and two-tailed χ2 or Fishers exact as appropriate for categorical variables. Results. Patient characteristics were similar between the pre- (n = 2867) and postintervention (n = 2717) groups, including age (x 56.9 versus 56.3 years), admission APACHE (19.7 versus 19.6), and immunosuppression (15% versus 16%), as were rates of CLABSI and influenza. Post wash-in there were 634 active Spectrum MD users within the health region, which increased to 1145 during the post-intervention period, and 76% (54/71; response rate 34%) of survey respondents reported using the app during their ICU rotation. There was no statistically significant change in overall AMU between the time periods. Agent-specific analysis revealed a trend toward decreased anti-pseudomonal utilization (slope change parameter -0.983; p = 0.07). Of 64 respondents, 70% agreed this app changed their antimicrobial prescribing practices while 89% would recommend it to colleagues. Conclusion. Spectrum MD was a widely accepted intervention that yielded positive user feedback. Though there was no significant reduction in overall AMU, a trend toward decreased anti-pseudomonal agent use may represent more appropriate antimicrobial prescribing. Further longitudinal analysis is required to assess the impact of this promising app.