Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study
Introduction: Early identification of sepsis may facilitate pre-alerting the emergency department (ED) enabling prompt initiation of antibiotics and source control. Whether sepsis can be reliably predicted using prehospital data recorded by paramedics remains uncertain. Study objectives: (1) To dete...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Journal Article |
| Published: |
Elsevier Inc
2016
|
| Online Access: | http://hdl.handle.net/20.500.11937/40080 |
| _version_ | 1848755768764923904 |
|---|---|
| author | Williams, Teresa Finn, Judith Fatovich, D. Tohira, Hideo Brink, D. Perkins, G. Ho, K. |
| author_facet | Williams, Teresa Finn, Judith Fatovich, D. Tohira, Hideo Brink, D. Perkins, G. Ho, K. |
| author_sort | Williams, Teresa |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | Introduction: Early identification of sepsis may facilitate pre-alerting the emergency department (ED) enabling prompt initiation of antibiotics and source control. Whether sepsis can be reliably predicted using prehospital data recorded by paramedics remains uncertain. Study objectives: (1) To determine the predictive ability of prehospital factors to predict a subsequent diagnosis of sepsis; (2) To examine the ability of the New Early Warning Score (NEWS) to predict a subsequent diagnosis of sepsis. Methods: This retrospective cohort study linked the prehospital data of patients aged 16 years and older, transported by the metropolitan St John Ambulance Service in Perth, between July 2012 and June 2014, with data from the ED Information System. Air transports, inter-facility transfers, patients with cardiac arrest, trauma and overdoses were excluded. Apart from the predictive ability of each individual demographic and vital physiological variable, the ability of the National Early Warning Score (NEWS) to predict a subsequent diagnosis of sepsis was also assessed. NEWS is a composite score: scores of 5+ are predictive of increased risk of death or ICU admission. Logistic regression and area under the operating-receiving-operating characteristic curves (AUROC) were used to identify prehospital predictors of sepsis and their discriminatory power, respectively.Results: Of 92,362 patients included in the study, 4565 (4.9%) had a diagnosis of sepsis subsequently in ED. Significant prehospital factors associated with sepsis: OR (95% CI) per 1 unit increment were age 1.08 (1.02–1.03), temperature 2.38 (2.00–2.51), systolic blood pressure 0.98 (0.97–0.98), respiratory rate 0.97 (0.94–0.99), heart rate 1.01 (1.01–1.02), and AVPU (alert = 1/verbal = 2/pain = 3/unresponsive = 4) 1.27 (1.13–1.42). The NEWS in the prehospital setting only had a moderate ability to differentiate between patients with and without sepsis (AUROC 0.74, 95% CI 0.72–0.77) and this predictive ability was no better than individual physiological parameters alone. Conclusion: Sepsis cannot be reliably predicted using prehospital data recorded by paramedics. |
| first_indexed | 2025-11-14T09:01:33Z |
| format | Journal Article |
| id | curtin-20.500.11937-40080 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T09:01:33Z |
| publishDate | 2016 |
| publisher | Elsevier Inc |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-400802017-09-13T15:07:14Z Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study Williams, Teresa Finn, Judith Fatovich, D. Tohira, Hideo Brink, D. Perkins, G. Ho, K. Introduction: Early identification of sepsis may facilitate pre-alerting the emergency department (ED) enabling prompt initiation of antibiotics and source control. Whether sepsis can be reliably predicted using prehospital data recorded by paramedics remains uncertain. Study objectives: (1) To determine the predictive ability of prehospital factors to predict a subsequent diagnosis of sepsis; (2) To examine the ability of the New Early Warning Score (NEWS) to predict a subsequent diagnosis of sepsis. Methods: This retrospective cohort study linked the prehospital data of patients aged 16 years and older, transported by the metropolitan St John Ambulance Service in Perth, between July 2012 and June 2014, with data from the ED Information System. Air transports, inter-facility transfers, patients with cardiac arrest, trauma and overdoses were excluded. Apart from the predictive ability of each individual demographic and vital physiological variable, the ability of the National Early Warning Score (NEWS) to predict a subsequent diagnosis of sepsis was also assessed. NEWS is a composite score: scores of 5+ are predictive of increased risk of death or ICU admission. Logistic regression and area under the operating-receiving-operating characteristic curves (AUROC) were used to identify prehospital predictors of sepsis and their discriminatory power, respectively.Results: Of 92,362 patients included in the study, 4565 (4.9%) had a diagnosis of sepsis subsequently in ED. Significant prehospital factors associated with sepsis: OR (95% CI) per 1 unit increment were age 1.08 (1.02–1.03), temperature 2.38 (2.00–2.51), systolic blood pressure 0.98 (0.97–0.98), respiratory rate 0.97 (0.94–0.99), heart rate 1.01 (1.01–1.02), and AVPU (alert = 1/verbal = 2/pain = 3/unresponsive = 4) 1.27 (1.13–1.42). The NEWS in the prehospital setting only had a moderate ability to differentiate between patients with and without sepsis (AUROC 0.74, 95% CI 0.72–0.77) and this predictive ability was no better than individual physiological parameters alone. Conclusion: Sepsis cannot be reliably predicted using prehospital data recorded by paramedics. 2016 Journal Article http://hdl.handle.net/20.500.11937/40080 10.1016/j.aucc.2015.12.024 Elsevier Inc restricted |
| spellingShingle | Williams, Teresa Finn, Judith Fatovich, D. Tohira, Hideo Brink, D. Perkins, G. Ho, K. Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study |
| title | Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study |
| title_full | Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study |
| title_fullStr | Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study |
| title_full_unstemmed | Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study |
| title_short | Predicting sepsis using prehospital data from the ambulance service: A linked data cohort study |
| title_sort | predicting sepsis using prehospital data from the ambulance service: a linked data cohort study |
| url | http://hdl.handle.net/20.500.11937/40080 |