Michelle R Ananda-Rajah1, Christoph Bergmeir1, François Petitjean1, Monica A Slavin1, Karin A Thursky1, Geoffrey I Webb1. 1. Michelle R. Ananda-Rajah, Alfred Health; Michelle R. Ananda-Rajah, Christoph Bergmeir, François Petitjean, and Geoffrey I. Webb, Monash University; and Monica A. Slavin and Karin A. Thursky, Peter Doherty Centre for Infection and Immunity; University of Melbourne, Melbourne, Victoria, Australia.
Abstract
PURPOSE: Prospective epidemiologic surveillance of invasive mold disease (IMD) in hematology patients is hampered by the absence of a reliable laboratory prompt. This study develops an expert system for electronic surveillance of IMD that combines probabilities using natural language processing (NLP) of computed tomography (CT) reports with microbiology and antifungal drug data to improve prediction of IMD. METHODS: Microbiology indicators and antifungal drug-dispensing data were extracted from hospital information systems at three tertiary hospitals for 123 hematology-oncology patients. Of this group, 64 case patients had 26 probable/proven IMD according to international definitions, and 59 patients were uninfected controls. Derived probabilities from NLP combined with medical expertise identified patients at high likelihood of IMD, with remaining patients processed by a machine-learning classifier trained on all available features. RESULTS: Compared with the baseline text classifier, the expert system that incorporated the best performing algorithm (naïve Bayes) improved specificity from 50.8% (95% CI, 37.5% to 64.1%) to 74.6% (95% CI, 61.6% to 85.0%), reducing false positives by 48% from 29 to 15; improved sensitivity slightly from 96.9% (95% CI, 89.2% to 99.6%) to 98.4% (95% CI, 91.6% to 100%); and improved receiver operating characteristic area from 73.9% (95% CI, 67.1% to 80.6%) to 92.8% (95% CI, 88% to 97.5%). CONCLUSION: An expert system that uses multiple sources of data (CT reports, microbiology, antifungal drug dispensing) is a promising approach to continuous prospective surveillance of IMD in the hospital, and demonstrates reduced false notifications (positives) compared with NLP of CT reports alone. Our expert system could provide decision support for IMD surveillance, which is critical to antifungal stewardship and improving supportive care in cancer.
PURPOSE: Prospective epidemiologic surveillance of invasive mold disease (IMD) in hematology patients is hampered by the absence of a reliable laboratory prompt. This study develops an expert system for electronic surveillance of IMD that combines probabilities using natural language processing (NLP) of computed tomography (CT) reports with microbiology and antifungal drug data to improve prediction of IMD. METHODS: Microbiology indicators and antifungal drug-dispensing data were extracted from hospital information systems at three tertiary hospitals for 123 hematology-oncology patients. Of this group, 64 case patients had 26 probable/proven IMD according to international definitions, and 59 patients were uninfected controls. Derived probabilities from NLP combined with medical expertise identified patients at high likelihood of IMD, with remaining patients processed by a machine-learning classifier trained on all available features. RESULTS: Compared with the baseline text classifier, the expert system that incorporated the best performing algorithm (naïve Bayes) improved specificity from 50.8% (95% CI, 37.5% to 64.1%) to 74.6% (95% CI, 61.6% to 85.0%), reducing false positives by 48% from 29 to 15; improved sensitivity slightly from 96.9% (95% CI, 89.2% to 99.6%) to 98.4% (95% CI, 91.6% to 100%); and improved receiver operating characteristic area from 73.9% (95% CI, 67.1% to 80.6%) to 92.8% (95% CI, 88% to 97.5%). CONCLUSION: An expert system that uses multiple sources of data (CT reports, microbiology, antifungal drug dispensing) is a promising approach to continuous prospective surveillance of IMD in the hospital, and demonstrates reduced false notifications (positives) compared with NLP of CT reports alone. Our expert system could provide decision support for IMD surveillance, which is critical to antifungal stewardship and improving supportive care in cancer.
Authors: D Neofytos; S Treadway; D Ostrander; C D Alonso; K L Dierberg; V Nussenblatt; C M Durand; C B Thompson; K A Marr Journal: Transpl Infect Dis Date: 2013-02-21 Impact factor: 2.228
Authors: O Lortholary; J-P Gangneux; K Sitbon; B Lebeau; F de Monbrison; Y Le Strat; B Coignard; F Dromer; S Bretagne Journal: Clin Microbiol Infect Date: 2011-06-10 Impact factor: 8.067
Authors: Douglas C Chang; Lauren A Burwell; G Marshall Lyon; Peter G Pappas; Tom M Chiller; Kathleen A Wannemuehler; Scott K Fridkin; Benjamin J Park Journal: Infect Control Hosp Epidemiol Date: 2008-01 Impact factor: 3.254
Authors: Gunnar Juliusson; Petar Antunovic; Asa Derolf; Sören Lehmann; Lars Möllgård; Dick Stockelberg; Ulf Tidefelt; Anders Wahlin; Martin Höglund Journal: Blood Date: 2008-11-13 Impact factor: 22.113
Authors: Sylvain DeLisle; Bernard Kim; Janaki Deepak; Tariq Siddiqui; Adi Gundlapalli; Matthew Samore; Leonard D'Avolio Journal: PLoS One Date: 2013-08-13 Impact factor: 3.240
Authors: H S C Paula; S B Santiago; L A Araújo; C F Pedroso; T A Marinho; I A J Gonçalves; T A P Santos; R S Pinheiro; G A Oliveira; K A Batista Journal: Braz J Med Biol Res Date: 2021-12-10 Impact factor: 2.590
Authors: Yousra El Alaoui; Adel Elomri; Marwa Qaraqe; Regina Padmanabhan; Ruba Yasin Taha; Halima El Omri; Abdelfatteh El Omri; Omar Aboumarzouk Journal: J Med Internet Res Date: 2022-07-12 Impact factor: 7.076