1. A machine learning approach to predict healthcare-associated infections at intensive care unit admission: findings from the SPIN-UTI project
- Author
-
Barchitta, Martina, Maugeri, Andrea, Favara, Giuliana, Riela, Paolo Marco, Gallo, Giovanni, Mura, Ida, Agodi, Antonella, Paola, Murgia, Maria Dolores Masia, Silvio, Brusaferro, Daniele, Celotto, Luca, Arnoldo, Emanuela, Bissolo, Alberto, Rigo, Stefano, Tardivo, Francesca, Moretti, Alberto, Carli, Diana, Pascu, Lorella, Tessari, Mara Olga Bernasconi, Marco, Brusaferro, Federico, Pappalardo, Francesco, Auxilia, Salesia, Fenaroli, Cesira, Pasquarella, Ennio, Sicoli, Maria Teresa Montagna, Giovanni, Egitto, Raffaele, Squeri, Salvatore, Tribastoni, Alessandro, Pulvirenti, Sebastiano, Catalano, Pietro, Battaglia, Patrizia, Bellocchi, Giacomo, Castiglione, Anna Rita Mattaliano, Marinella Astuto Marinella, LA CAMERA, Giuseppa, Anna Maria Longhitano, Giorgio, Scrofani, Maria Concetta Monea, Marina, Milazzo, Antonino, Giarratano, Giuseppe, Calamusa, Maria Valeria Torregrossa, Antonino Di Benedetto, Giuseppa Maria Gisella Rizzo, Giuseppe, Manta, Romano, Tetamo, Rosa, Mancuso, Laura Maria Mella, Ignazio, Dei, Irene, Pandiani, Antonino, Cannistrà, Paola, Piotti, Massimo, Girardis, Elena, Righi, Alberto, Barbieri, Patricia, Crollari, Albino, Borracino, Salvatore, Coniglio, Rosaria, Palermo, Sergio, Pintaudi, Daniela Di Stefano, Antonina, Romeo, Giovanna, Sticca, Massimo, Minerva, Leila, Fabiani, Alessandra, Gentile, Paolo, Stefanini, Marcello Mario D'Errico, Abele, Donati, Santa De Remigis, Federica, Venturoni, Manuela, Antoci, Riccardo, Pagliarulo, Aida, Bianco, Maria, Pavia, Marcello, Pasculli, Cesare, Vittori, Giovanni Battista Orsi, Cristina, Arrigoni, Maria Patrizia Olori, Massimo, Antonelli, Patrizia, Laurenti, Franco, Ingala, Carmela, Conte, Salvatore, Russo, Laura, Condorelli, Patrizia, Farruggia, Cristina Maria Luisa, and Italia, Galassi
- Subjects
Microbiology (medical) ,medicine.medical_specialty ,Psychological intervention ,030501 epidemiology ,Machine learning ,computer.software_genre ,intensive care unit ,law.invention ,Machine Learning ,03 medical and health sciences ,risk prediction ,law ,Intensive care ,medicine ,Infection control ,Humans ,Hospital Mortality ,Simplified Acute Physiology Score ,healthcare-associated infections ,machine learning ,0303 health sciences ,Cross Infection ,Receiver operating characteristic ,030306 microbiology ,business.industry ,Public health ,General Medicine ,Intensive care unit ,Intensive Care Units ,Infectious Diseases ,ROC Curve ,SAPS II ,Artificial intelligence ,0305 other medical science ,business ,computer ,Delivery of Health Care - Abstract
Identifying patients at higher risk of healthcare-associated infections (HAIs) in intensive care units (ICUs) represents a major challenge for public health. Machine learning could improve patient risk stratification and lead to targeted infection prevention and control interventions.To evaluate the performance of the Simplified Acute Physiology Score (SAPS) II for HAI risk prediction in ICUs, using both traditional statistical and machine learning approaches.Data for 7827 patients from the 'Italian Nosocomial Infections Surveillance in Intensive Care Units' project were used in this study. The Support Vector Machines (SVM) algorithm was applied to classify patients according to sex, patient origin, non-surgical treatment for acute coronary disease, surgical intervention, SAPS II at admission, presence of invasive devices, trauma, impaired immunity, and antibiotic therapy in 48 h preceding ICU admission.The performance of SAPS II for predicting HAI risk provides a receiver operating characteristic curve with an area under the curve of 0.612 (P0.001) and accuracy of 56%. Considering SAPS II along with other characteristics at ICU admission, the SVM classifier was found to have accuracy of 88% and an AUC of 0.90 (P0.001) for the test set. The predictive ability was lower when considering the same SVM model but with the SAPS II variable removed (accuracy 78%, AUC 0.66).This study suggested that the SVM model is a useful tool for early prediction of patients at higher risk of HAIs at ICU admission.
- Published
- 2020