A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice
Aim: The aim of this study was to develop machine learning (ML) models to mitigate the inappropriate request of Procalcitonin (PCT) in clinical wards. Material and methods: We built six different ML models based on both demographical data, i.e., sex and age, and laboratory parameters, i.e., cell blo...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-03-01
|
| Series: | Heliyon |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405844024025878 |
| _version_ | 1827316520951218176 |
|---|---|
| author | Luisa Agnello Matteo Vidali Anna Maria Ciaccio Bruna Lo Sasso Alessandro Iacona Giuseppe Biundo Concetta Scazzone Caterina Maria Gambino Marcello Ciaccio |
| author_facet | Luisa Agnello Matteo Vidali Anna Maria Ciaccio Bruna Lo Sasso Alessandro Iacona Giuseppe Biundo Concetta Scazzone Caterina Maria Gambino Marcello Ciaccio |
| author_sort | Luisa Agnello |
| collection | DOAJ |
| description | Aim: The aim of this study was to develop machine learning (ML) models to mitigate the inappropriate request of Procalcitonin (PCT) in clinical wards. Material and methods: We built six different ML models based on both demographical data, i.e., sex and age, and laboratory parameters, i.e., cell blood count (CBC) parameters, inclusive of monocyte distribution width (MDW), and C-reactive protein (CRP). The dataset included 1667 PCT measurements of different patients. Based on a PCT cut-off of 0.50 ng/mL, we found 1090 negative (65.4%) and 577 positive (34.6%) results. We performed a 70:15:15 train:validation:test splitting based on the outcome. Results: Random Forest, Support Vector Machine and eXtreme Gradient Boosting showed optimal performances for predicting PCT positivity, with an area under the curve ranging from 0.88 to 0.89. Conclusions: The ML models developed could represent a useful tool to predict PCT positivity, avoiding unusefulness PCT requests. ML models are based on laboratory tests commonly ordered together with PCT but have the great advantage to be easy to measure and low-cost. |
| first_indexed | 2024-03-07T19:09:09Z |
| format | Article |
| id | doaj.art-3deb69eaf58b4a719726cd477f705a99 |
| institution | Directory Open Access Journal |
| issn | 2405-8440 |
| language | English |
| last_indexed | 2024-04-24T23:15:58Z |
| publishDate | 2024-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Heliyon |
| spelling | doaj.art-3deb69eaf58b4a719726cd477f705a992024-03-17T07:55:44ZengElsevierHeliyon2405-84402024-03-01105e26556A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practiceLuisa Agnello0Matteo Vidali1Anna Maria Ciaccio2Bruna Lo Sasso3Alessandro Iacona4Giuseppe Biundo5Concetta Scazzone6Caterina Maria Gambino7Marcello Ciaccio8Institute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, Department of Biomedicine, Neurosciences and Advanced Diagnostics, University of Palermo, 90127, Palermo, ItalyFoundation IRCCS Ca’ Granda Ospedale Maggiore Policlinico, 20122, Milan, ItalyInternal Medicine and Medical Specialties “G. D'Alessandro”, Department of Health Promotion, Maternal and Infant Care, University of Palermo, 90127, Palermo, ItalyInstitute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, Department of Biomedicine, Neurosciences and Advanced Diagnostics, University of Palermo, 90127, Palermo, Italy; Department of Laboratory Medicine, AOUP “P. Giaccone”, Palermo, ItalyDepartment of Laboratory Medicine, AOUP “P. Giaccone”, Palermo, ItalyDepartment of Laboratory Medicine, AOUP “P. Giaccone”, Palermo, ItalyInstitute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, Department of Biomedicine, Neurosciences and Advanced Diagnostics, University of Palermo, 90127, Palermo, ItalyInstitute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, Department of Biomedicine, Neurosciences and Advanced Diagnostics, University of Palermo, 90127, Palermo, Italy; Department of Laboratory Medicine, AOUP “P. Giaccone”, Palermo, ItalyInstitute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, Department of Biomedicine, Neurosciences and Advanced Diagnostics, University of Palermo, 90127, Palermo, Italy; Department of Laboratory Medicine, AOUP “P. Giaccone”, Palermo, Italy; Corresponding author. Institute of Clinical Biochemistry, Clinical Molecular Medicine and Clinical Laboratory Medicine, Department of Biomedicine, Neurosciences and Advanced Diagnostics, University of Palermo, Via del Vespro, 129, CAP, 90127, Palermo, Sicily, Italy.Aim: The aim of this study was to develop machine learning (ML) models to mitigate the inappropriate request of Procalcitonin (PCT) in clinical wards. Material and methods: We built six different ML models based on both demographical data, i.e., sex and age, and laboratory parameters, i.e., cell blood count (CBC) parameters, inclusive of monocyte distribution width (MDW), and C-reactive protein (CRP). The dataset included 1667 PCT measurements of different patients. Based on a PCT cut-off of 0.50 ng/mL, we found 1090 negative (65.4%) and 577 positive (34.6%) results. We performed a 70:15:15 train:validation:test splitting based on the outcome. Results: Random Forest, Support Vector Machine and eXtreme Gradient Boosting showed optimal performances for predicting PCT positivity, with an area under the curve ranging from 0.88 to 0.89. Conclusions: The ML models developed could represent a useful tool to predict PCT positivity, avoiding unusefulness PCT requests. ML models are based on laboratory tests commonly ordered together with PCT but have the great advantage to be easy to measure and low-cost.http://www.sciencedirect.com/science/article/pii/S2405844024025878Laboratory medicineSepsisProcalcitoninMDWCRPArtificial intelligence |
| spellingShingle | Luisa Agnello Matteo Vidali Anna Maria Ciaccio Bruna Lo Sasso Alessandro Iacona Giuseppe Biundo Concetta Scazzone Caterina Maria Gambino Marcello Ciaccio A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice Heliyon Laboratory medicine Sepsis Procalcitonin MDW CRP Artificial intelligence |
| title | A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice |
| title_full | A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice |
| title_fullStr | A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice |
| title_full_unstemmed | A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice |
| title_short | A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice |
| title_sort | machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice |
| topic | Laboratory medicine Sepsis Procalcitonin MDW CRP Artificial intelligence |
| url | http://www.sciencedirect.com/science/article/pii/S2405844024025878 |
| work_keys_str_mv | AT luisaagnello amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT matteovidali amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT annamariaciaccio amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT brunalosasso amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT alessandroiacona amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT giuseppebiundo amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT concettascazzone amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT caterinamariagambino amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT marcellociaccio amachinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT luisaagnello machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT matteovidali machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT annamariaciaccio machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT brunalosasso machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT alessandroiacona machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT giuseppebiundo machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT concettascazzone machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT caterinamariagambino machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice AT marcellociaccio machinelearningstrategytomitigatetheinappropriatenessofprocalcitoninrequestinclinicalpractice |