Predicting patient-level antimicrobial choice and outcome (PLAMOS)

• Research type

  Research Study

• Full title

  Predicting patient-level empirical antibiotic treatment decisions using existing patient level data

• IRAS ID

  204543

• Contact name

  Ed Moran

• Contact email

  ed.moran@heartofengland.nhs.uk

• Sponsor organisation

  Birmingham Heartlands Hospital

• Duration of Study in the UK

  0 years, 11 months, 29 days

• Research summary

  The cornerstone of antibiotic stewardship is good empirical treatment guidance. Current guidance comes in the form of rigid protocols reviewed every few years as bacterial resistance patterns change. The increasingly wide use of electronic prescribing and other clinical information systems opens up the possibility of machine learning programmes that assess the information available on an individual patient and produce a personalised recommendation for empirical treatment. Such a system would be capable of recommending narrower agents than paper guidelines and would adapt automatically as resistance patterns changed. Using the wealth of electronic information available in our hospital we propose a “proof of principle” analysis to first describe the prescribing and patient factors associated with the isolation of resistant organisms, and secondly produce a set of rules that could potentially inform an “intelligent” personalised empirical antibiotic treatment system.

• REC name

  West Midlands - Solihull Research Ethics Committee

• REC reference

  17/WM/0406

• Date of REC Opinion

  24 Nov 2017

• REC opinion

  Favourable Opinion