1
|
Han T, Xiong F, Sun B, Zhong L, Han Z, Lei M. Development and validation of an artificial intelligence mobile application for predicting 30-day mortality in critically ill patients with orthopaedic trauma. Int J Med Inform 2024; 184:105383. [PMID: 38387198 DOI: 10.1016/j.ijmedinf.2024.105383] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/21/2023] [Revised: 01/25/2024] [Accepted: 02/16/2024] [Indexed: 02/24/2024]
Abstract
BACKGROUND Given the intricate and grave nature of trauma-related injuries in ICU settings, it is imperative to develop and deploy reliable predictive tools that can aid in the early identification of high-risk patients who are at risk of early death. The objective of this study is to create and validate an artificial intelligence (AI) model that can accurately predict early mortality among critical fracture patients. METHODS A total of 2662 critically ill patients with orthopaedic trauma were included from the MIMIC III database. Early mortality was defined as death within 30 days in this study. The patients were randomly divided into a model training cohort and a model validation cohort. Various algorithms, including logistic regression (LR), extreme gradient boosting machine (eXGBM), decision tree (DT), support vector machine (SVM), random forest (RF), and neural network (NN), were employed. Evaluation metrics, including discrimination and calibration, were used to develop a comprehensive scoring system ranging from 0 to 60, with higher scores indicating better prediction performance. Furthermore, external validation was carried out using 131 patients. The optimal model was deployed as an internet-based AI tool. RESULTS Among all models, the eXGBM demonstrated the highest area under the curve (AUC) value (0.974, 95%CI: 0.959-0.983), followed by the RF model (0.951, 95%CI: 0.935-0.967) and the NN model (0.922, 95%CI: 0.905-0.941). Additionally, the eXGBM model outperformed other models in terms of accuracy (0.915), precision (0.906), recall (0.926), F1 score (0.916), Brier score (0.062), log loss (0.210), and discrimination slope (0.767). Based on the scoring system, the eXGBM model achieved the highest score (53), followed by RF (42) and NN (39). The LR, DT, and SVM models obtained scores of 28, 18, and 32, respectively. Decision curve analysis further confirmed the superior clinical net benefits of the eXGBM model. External validation of the model achieved an AUC value of 0.913 (95%CI: 0.878-0.948). Consequently, the model was deployed on the Internet at https://30-daymortalityincriticallyillpatients-fnfsynbpbp6rgineaspuim.streamlit.app/, allowing users to input patient features and obtain predicted risks of early mortality among critical fracture patients. Furthermore, the AI model successfully stratified patients into low or high risk of early mortality based on a predefined threshold and provided recommendations for appropriate therapeutic interventions. CONCLUSION This study successfully develops and validates an AI model, with the eXGBM algorithm demonstrating the highest predictive performance for early mortality in critical fracture patients. By deploying the model as a web-based AI application, healthcare professionals can easily access the tool, enabling them to predict 30-day mortality and aiding in the identification and management of high-risk patients among those critically ill with orthopedic trauma.
Collapse
Affiliation(s)
- Tao Han
- Department of Orthopedics, Hainan Hospital of PLA General Hospital, Hainan, China
| | - Fan Xiong
- Department of Orthopedic Surgery, People's Hospital of Macheng City, Huanggang, China
| | - Baisheng Sun
- Department of Critical Care Medicine, The First Medical Centre, PLA General Hospital, Beijing, China; Chinese PLA Medical School, Beijing, China
| | - Lixia Zhong
- Department of Intensive Care Medicine, Beijing Friendship Hospital, Capital Medical University, Beijing, China.
| | - Zhencan Han
- Xiangya School of Medicine, Center South University, Changsha, China.
| | - Mingxing Lei
- Department of Orthopedics, Hainan Hospital of PLA General Hospital, Hainan, China; Chinese PLA Medical School, Beijing, China; Department of Orthopedics, National Clinical Research Center for Orthopedics, Sports Medicine & Rehabilitation, PLA General Hospital, Beijing, China.
| |
Collapse
|
2
|
Newey C, Skaar JR, O'Hara M, Miao B, Post A, Kelly T. Systematic Literature Review of the Association of Fever and Elevated Temperature with Outcomes in Critically Ill Adult Patients. Ther Hypothermia Temp Manag 2024; 14:10-23. [PMID: 37158862 DOI: 10.1089/ther.2023.0004] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 05/10/2023] Open
Abstract
Although most commonly associated with infection, elevated temperature and fever also occur in a variety of critically ill populations. Prior studies have suggested that fever and elevated temperature may be detrimental to critically ill patients and can lead to poor outcomes, but the evidence surrounding the association of fever with outcomes is rapidly evolving. To broadly assess potential associations of elevated temperature and fever with outcomes in critically ill adult patients, we performed a systematic literature review focusing on traumatic brain injury, stroke (ischemic and hemorrhagic), cardiac arrest, sepsis, and general intensive care unit (ICU) patients. Searches were conducted in Embase® and PubMed® from 2016 to 2021, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, including dual-screening of abstracts, full texts, and extracted data. In total, 60 studies assessing traumatic brain injury and stroke (24), cardiac arrest (8), sepsis (22), and general ICU (6) patients were included. Mortality, functional, or neurological status and length of stay were the most frequently reported outcomes. Elevated temperature and fever were associated with poor clinical outcomes in patients with traumatic brain injury, stroke, and cardiac arrest but not in patients with sepsis. Although a causal relationship between elevated temperature and poor outcomes cannot be definitively established, the association observed in this systematic literature review supports the concept that management of elevated temperature may factor in avoidance of detrimental outcomes in multiple critically ill populations. The analysis also highlights gaps in our understanding of fever and elevated temperature in critically ill adult patients.
Collapse
Affiliation(s)
- Christopher Newey
- Department of Neurocritical Care, Sanford USD Medical Center, Sioux Falls, South Dakota, USA
| | | | | | | | - Andrew Post
- Trinity Life Sciences, Waltham, Massachusetts, USA
| | - Tim Kelly
- Becton Dickinson, Franklin Lakes, New Jersey, USA
| |
Collapse
|
3
|
Delaplain PT, Santos J, Dvorak J, Mele TS, Gelbard RB, Guidry CA, Barie PS, Schubl SD. An Exploratory and Qualitative Analysis of Self-Reported Evaluations for Fever. Surg Infect (Larchmt) 2024; 25:116-124. [PMID: 38324100 DOI: 10.1089/sur.2023.294] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 02/08/2024] Open
Abstract
Background: Despite the high prevalence of post-operative fever, a variety of approaches are taken as to the components of a fever evaluation, when it should be undertaken, and when empiric antibiotic agents should be started. Hypothesis: There is a lack of consensus surrounding many common components of a post-operative fever evaluation. Patients and Methods: The Surgical Infection Society membership was surveyed to determine practices surrounding evaluation of post-operative fever. Eight scenarios were posed in febrile (38.5°C), post-operative general surgery or trauma patients, with 19 possible components of work-up (physical examination, complete blood count [CBC], fungal biomarkers, lactate and procalcitonin [PCT] concentrations, cultures, imaging) and management (antibiotic agents). Each scenario was then re-considered for intensive care unit (ICU) patients (intubated/unstable hemodynamics). Agreement on a parameter (<1/4 or >3/4 of respondents) achieved consensus, positive or negative. Parameters between had equipoise; α was set at 0.05. Results: Among the examined scenarios, only CBC and physical examination received positive consensus across most scenarios. Blood/urine cultures, imaging, lactate, inflammatory biomarkers, and the empiric administration of antibiotic agents did not reach consensus; support was variable depending on the clinical scenario, illness severity, and the individual preferences of the answering clinician. The qualitative portion of the survey identified "fever threshold and duration," "clinical suspicion," and "physiologic manifestation" as the most important factors for deciding about the initiation of a fever evaluation and the potential empiric administration of antibiotic agents. Conclusions: There is consensus only for physical and examination routine laboratory work when initiating the evaluation of febrile post-operative patients. However, there are multiple components of a fever evaluation that individual respondents would select depending on the clinical scenario and severity of illness. Parameters demonstrating equipoise are potential candidates for formal guidance or pragmatic prospective trials.
Collapse
Affiliation(s)
- Patrick T Delaplain
- Boston Children's Hospital, Harvard Medical System, Boston, Massachusetts, USA
| | - Jeffrey Santos
- Department of Surgery, University of California-Irvine, Orange, California, USA
| | - Justin Dvorak
- Division of Trauma, Critical Care, Burns, and Acute Care Surgery, Department of Surgery, Case Western Reserve University School of Medicine, MetroHealth Medical Center, Cleveland, Ohio, USA
| | - Tina S Mele
- Divisions of General Surgery and Critical Care, Department of Surgery, Schulich School of Medicine and Dentistry, University of Western Ontario, London, Ontario, Canada
| | - Rondi B Gelbard
- Department of Surgery, Heersink School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama, USA
| | - Christopher A Guidry
- Division of Trauma/Critical and Acute Care Surgery, Department of Surgery, University of Kansas Medical Center, Kansas City, Kansas
| | - Philip S Barie
- Division of Trauma, Burns, Acute and Critical Care, Department of Surgery, Weill Cornell Medicine, New York, New York, USA
| | - Sebastian D Schubl
- Department of Surgery, University of California-Irvine, Orange, California, USA
| |
Collapse
|
4
|
O'Grady NP, Alexander E, Alhazzani W, Alshamsi F, Cuellar-Rodriguez J, Jefferson BK, Kalil AC, Pastores SM, Patel R, van Duin D, Weber DJ, Deresinski S. Society of Critical Care Medicine and the Infectious Diseases Society of America Guidelines for Evaluating New Fever in Adult Patients in the ICU. Crit Care Med 2023; 51:1570-1586. [PMID: 37902340 DOI: 10.1097/ccm.0000000000006022] [Citation(s) in RCA: 6] [Impact Index Per Article: 6.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/31/2023]
Abstract
RATIONALE Fever is frequently an early indicator of infection and often requires rigorous diagnostic evaluation. OBJECTIVES This is an update of the 2008 Infectious Diseases Society of America and Society (IDSA) and Society of Critical Care Medicine (SCCM) guideline for the evaluation of new-onset fever in adult ICU patients without severe immunocompromise, now using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) methodology. PANEL DESIGN The SCCM and IDSA convened a taskforce to update the 2008 version of the guideline for the evaluation of new fever in critically ill adult patients, which included expert clinicians as well as methodologists from the Guidelines in Intensive Care, Development and Evaluation Group. The guidelines committee consisted of 12 experts in critical care, infectious diseases, clinical microbiology, organ transplantation, public health, clinical research, and health policy and administration. All task force members followed all conflict-of-interest procedures as documented in the American College of Critical Care Medicine/SCCM Standard Operating Procedures Manual and the IDSA. There was no industry input or funding to produce this guideline. METHODS We conducted a systematic review for each population, intervention, comparison, and outcomes question to identify the best available evidence, statistically summarized the evidence, and then assessed the quality of evidence using the GRADE approach. We used the evidence-to-decision framework to formulate recommendations as strong or weak or as best-practice statements. RESULTS The panel issued 12 recommendations and 9 best practice statements. The panel recommended using central temperature monitoring methods, including thermistors for pulmonary artery catheters, bladder catheters, or esophageal balloon thermistors when these devices are in place or accurate temperature measurements are critical for diagnosis and management. For patients without these devices in place, oral or rectal temperatures over other temperature measurement methods that are less reliable such as axillary or tympanic membrane temperatures, noninvasive temporal artery thermometers, or chemical dot thermometers were recommended. Imaging studies including ultrasonography were recommended in addition to microbiological evaluation using rapid diagnostic testing strategies. Biomarkers were recommended to assist in guiding the discontinuation of antimicrobial therapy. All recommendations issued were weak based on the quality of data. CONCLUSIONS The guidelines panel was able to formulate several recommendations for the evaluation of new fever in a critically ill adult patient, acknowledging that most recommendations were based on weak evidence. This highlights the need for the rapid advancement of research in all aspects of this issue-including better noninvasive methods to measure core body temperature, the use of diagnostic imaging, advances in microbiology including molecular testing, and the use of biomarkers.
Collapse
Affiliation(s)
- Naomi P O'Grady
- Internal Medicine Services, National Institutes of Health Clinical Center, Bethesda, MD
| | - Earnest Alexander
- Clinical Pharmacy Services, Department of Pharmacy, Tampa General Hospital, Tampa, FL
| | - Waleed Alhazzani
- Department of Medicine, McMaster University, Hamilton, ON, Canada
| | - Fayez Alshamsi
- Department of Internal Medicine, College of Medicine and Health Sciences, United Arab Emirates University, Al Ain, United Arab Emirates
| | - Jennifer Cuellar-Rodriguez
- Laboratory of Clinical Immunology and Microbiology, National Institutes of Allergy and Infectious Diseases, Bethesda, MD
| | - Brian K Jefferson
- Division of Hepatobiliary and Pancreatic Surgery, Department of Internal Medicine-Critical Care Services, Atrium Health Cabarrus, Concord, NC
| | - Andre C Kalil
- Infectious Diseases Division, University of Nebraska Medical Center, Omaha, NE
| | - Stephen M Pastores
- Department of Anesthesiology and Critical Care Medicine, Memorial Sloan Kettering Cancer Center, New York, NY
| | - Robin Patel
- Division of Clinical Microbiology, Department of Laboratory Medicine and Pathology, Rochester, MN
- Division of Infectious Diseases, Department of Medicine, Mayo Clinic, Rochester, MN
| | - David van Duin
- Division of Infectious Diseases, University of North Carolina, Chapel Hill, NC
| | - David J Weber
- Division of Infectious Diseases, University of North Carolina, Chapel Hill, NC
| | - Stanley Deresinski
- Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, Stanford, CA
| |
Collapse
|