Implementation and Evaluation of a Real-Time Platform to Analyse and Display Antimicrobial Usage

Implementation and Evaluation of a Real-Time Platform to Analyse and Display Antimicrobial Usage at Great Ormond Street Hospital

Background:
Effective monitoring systems are the mainstay in the development of strategies aimed to rationalise and optimise antibiotic prescriptions (WHO, 2018; CDC, 2014). Point-prevalence surveys and/or sentinel-like schemes have been successfully used to inform the quality of antibiotic use in hospital-based programs. However, the requirement for regular data collection and centralised data input often results in delay in reports being available for stakeholders (Willemsen et al., 2007; Fridkin and Srinivasan, 2014; Mccormack et al., 2016). The aim of this PhD project is to design, develop and validate an automated surveillance tool for real-time analysis on antibiotic prescription that can identify early changes in patterns of prescription and quality of antibiotic use.

Aims/Objectives:
1) Receive a Formal Training in the Use of R and R-shiny programming;
2) Design and validate a platform that analyses live data from the new hospital Electronic Record Program and provides trend data and evaluate factors associated with antimicrobial consumption use;
3) Use ARIMA time series models to predict future use of Antimicrobials at the Trust.

Methods:
This project will see the development of a platform to integrate routinely collected data into an analytic dashboard that will provide summary of hospital antibiotic use at any given time, and a trend analysis to allow comparisons between expected vs. observed use. The platform will be piloted using routinely collected data on antibiotic prescription from Great Ormond Street Hospital (GOSH) and the UCL Great Ormond Street Institute of Child Health (ICH). In addition to descriptive summaries, time series analysis tools will be used to explore and analyse the dependence of adjacent observations in a time series (e.g. autocorrelations), overall trending behaviour of a series (e.g. trend extraction) and frequently re-occurring patterns (e.g. seasonality). Further, forecasting time series (e.g. ARIMA models) will allow for prediction of antibiotic use accounting for possible temporal patterns and confounders. The platform will be validated against current practise.

Timeline:

References:
1.  CDC. 2014. Core Elements of Hospital Antibiotic Stewardship Programs. [Online]. Atlanta, GA: US Department of Health and Human Services.
2.  WHO. 2018. Methodology for point prevalence survey on antibiotic use in hospitals. Geneva: World Health Organization. Available: [Accessed 8 April 2019].
3.  Willemsen, I., Groenhuijzen, A., Bogaers, D., Stuurman, A., Van Keulen, P. & Kluytmans, J. 2007. Appropriateness of antimicrobial therapy measured by repeated prevalence surveys. Antimicrob Agents Chemother, 51, 864-7.
4.  Fridkin, S. K. & Srinivasan, A. 2014. Implementing a strategy for monitoring inpatient antimicrobial use among hospitals in the United States. Clin Infect Dis, 58, 401-6.
5.  Mccormack, S et al 2016. Pre-exposure prophylaxis to prevent the acquisition of HIV-1 infection (PROUD): effectiveness results from the pilot phase of a pragmatic open-label randomised trial. Lancet, 387, 53-60.