Ed Moran1, Esther Robinson2, Christopher Green3,4, Matt Keeling5, Benjamin Collyer5. 1. Southmead Hospital, North Bristol NHS Trust, Bristol BS10 5NB, UK. 2. Birmingham Public Health Laboratory, Public Health England, Birmingham Heartlands Hospital, Bordesley Green East, Birmingham B9 5SS, UK. 3. Birmingham Heartlands Hospital, University Hospitals Birmingham NHS Foundation Trust, Bordesley Green East, Birmingham B9 5SS, UK. 4. Institute of Microbiology and Infection, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK. 5. Zeeman Institute, Mathematics Institute and School of Life Sciences, University of Warwick, Gibbet Hill Road, Coventry CV4 7AL, UK.
Abstract
BACKGROUND: Electronic decision support systems could reduce the use of inappropriate or ineffective empirical antibiotics. We assessed the accuracy of an open-source machine-learning algorithm trained in predicting antibiotic resistance for three Gram-negative bacterial species isolated from patients' blood and urine within 48 h of hospital admission. METHODS: This retrospective, observational study used routine clinical information collected between January 2010 and October 2016 in Birmingham, UK. Patients from whose blood or urine cultures Escherichia coli, Klebsiella pneumoniae or Pseudomonas aeruginosa was isolated were identified. Their demographic, microbiology and prescribing data were used to train an open-source machine-learning algorithm-XGBoost-in predicting resistance to co-amoxiclav and piperacillin/tazobactam. Multivariate analysis was performed to identify predictors of resistance and create a point-scoring tool. The performance of both methods was compared with that of the original prescribers. RESULTS: There were 15 695 admissions. The AUC of the receiver operating characteristic curve for the point-scoring tools ranged from 0.61 to 0.67, and performed no better than medical staff in the selection of appropriate antibiotics. The machine-learning system performed statistically but marginally better (AUC 0.70) and could have reduced the use of unnecessary broad-spectrum antibiotics by as much as 40% among those given co-amoxiclav, piperacillin/tazobactam or carbapenems. A validation study is required. CONCLUSIONS: Machine-learning algorithms have the potential to help clinicians predict antimicrobial resistance in patients found to have a Gram-negative infection of blood or urine. Prospective studies are required to assess performance in an unselected patient cohort, understand the acceptability of such systems to clinicians and patients, and assess the impact on patient outcome.
BACKGROUND: Electronic decision support systems could reduce the use of inappropriate or ineffective empirical antibiotics. We assessed the accuracy of an open-source machine-learning algorithm trained in predicting antibiotic resistance for three Gram-negative bacterial species isolated from patients' blood and urine within 48 h of hospital admission. METHODS: This retrospective, observational study used routine clinical information collected between January 2010 and October 2016 in Birmingham, UK. Patients from whose blood or urine cultures Escherichia coli, Klebsiella pneumoniae or Pseudomonas aeruginosa was isolated were identified. Their demographic, microbiology and prescribing data were used to train an open-source machine-learning algorithm-XGBoost-in predicting resistance to co-amoxiclav and piperacillin/tazobactam. Multivariate analysis was performed to identify predictors of resistance and create a point-scoring tool. The performance of both methods was compared with that of the original prescribers. RESULTS: There were 15 695 admissions. The AUC of the receiver operating characteristic curve for the point-scoring tools ranged from 0.61 to 0.67, and performed no better than medical staff in the selection of appropriate antibiotics. The machine-learning system performed statistically but marginally better (AUC 0.70) and could have reduced the use of unnecessary broad-spectrum antibiotics by as much as 40% among those given co-amoxiclav, piperacillin/tazobactam or carbapenems. A validation study is required. CONCLUSIONS: Machine-learning algorithms have the potential to help clinicians predict antimicrobial resistance in patients found to have a Gram-negative infection of blood or urine. Prospective studies are required to assess performance in an unselected patient cohort, understand the acceptability of such systems to clinicians and patients, and assess the impact on patient outcome.
Authors: Shravan Kethireddy; Beliz Bilgili; Amanda Sees; H Lester Kirchner; Uchenna R Ofoma; R Bruce Light; Yazdan Mirzanejad; Dennis Maki; Aseem Kumar; A Joseph Layon; Joseph E Parrillo; Anand Kumar Journal: Crit Care Med Date: 2018-04 Impact factor: 7.598
Authors: E Weiss; J-R Zahar; P Lesprit; E Ruppe; M Leone; J Chastre; J-C Lucet; C Paugam-Burtz; C Brun-Buisson; J-F Timsit Journal: Clin Microbiol Infect Date: 2015-04-13 Impact factor: 8.067
Authors: J M Fitzpatrick; J S Biswas; J D Edgeworth; J Islam; N Jenkins; R Judge; A J Lavery; M Melzer; S Morris-Jones; E F Nsutebu; J Peters; D G Pillay; F Pink; J R Price; M Scarborough; G E Thwaites; R Tilley; A S Walker; M J Llewelyn Journal: Clin Microbiol Infect Date: 2015-11-11 Impact factor: 8.067
Authors: T M Rawson; L S P Moore; B Hernandez; E Charani; E Castro-Sanchez; P Herrero; B Hayhoe; W Hope; P Georgiou; A H Holmes Journal: Clin Microbiol Infect Date: 2017-03-06 Impact factor: 8.067
Authors: Mathupanee Oonsivilai; Yin Mo; Nantasit Luangasanatip; Yoel Lubell; Thyl Miliya; Pisey Tan; Lorn Loeuk; Paul Turner; Ben S Cooper Journal: Wellcome Open Res Date: 2018-10-10
Authors: Conor K Corbin; Lillian Sung; Arhana Chattopadhyay; Morteza Noshad; Amy Chang; Stanley Deresinksi; Michael Baiocchi; Jonathan H Chen Journal: Commun Med (Lond) Date: 2022-04-08