Hiromu Naraba
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Hiromu Naraba has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Kensuke Nakamura
Division of Disaster and Emergency Medicine, Department of Surgery Related, Kobe University Graduate School of Medicine, Kobe, Japan
6 shared publications - Tomohiro Sonoo
Department of Emergency and Critical Care Medicine, Hitachi General Hospital, Hitachi, Ibaraki 317-0077, Japan
6 shared publications - Tadahiro Goto
Department of Health Data Science, Graduate School of Data Science, Yokohama City University, Yokohama, Japan
5 shared publications - Kentaro Ogura
Faculty of Medicine The University of Tokyo Tokyo Japan
4 shared publications - Konan Hara
TXP Medical Co Ltd Tokyo Japan
4 shared publications - Ryo Fujimori
Faculty of Medicine The University of Tokyo Tokyo Japan
4 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 13 June 2022
- Published by JMIR Publications Inc. in JMIR Formative Research
- Vol. 6 (6) , e36501
Acceptance, Barriers, and Facilitators to Implementing Artificial Intelligence–Based Decision Support Systems in Emergency Departments: Quantitative and Qualitative Evaluation
Background: Despite the increasing availability of clinical decision support systems (CDSSs) and rising expectation for CDSSs based on artificial intelligence (AI), little is known about the acceptance of AI-based CDSS by physicians and its barriers and facilitators in emergency care settings. Objective: We aimed to evaluate the acceptance, barriers, and facilitators to implementing AI-based CDSSs in the emergency care setting through the opinions of physicians on our newly developed, real-time AI-based CDSS, which alerts ED physicians by predicting aortic dissection based on numeric and text information from medical charts, by using the Unified Theory of Acceptance and Use of Technology (UTAUT; for quantitative evaluation) and the Consolidated Framework for Implementation Research (CFIR; for qualitative evaluation) frameworks. Methods: This mixed methods study was performed from March to April 2021. Transitional year residents (n=6), emergency medicine residents (n=5), and emergency physicians (n=3) from two community, tertiary care hospitals in Japan were included. We first developed a real-time CDSS for predicting aortic dissection based on numeric and text information from medical charts (eg, chief complaints, medical history, vital signs) with natural language processing. This system was deployed on the internet, and the participants used the system with clinical vignettes of model cases. Participants were then involved in a mixed methods evaluation consisting of a UTAUT-based questionnaire with a 5-point Likert scale (quantitative) and a CFIR-based semistructured interview (qualitative). Cronbach α was calculated as a reliability estimate for UTAUT subconstructs. Interviews were sampled, transcribed, and analyzed using the MaxQDA software. The framework analysis approach was used during the study to determine the relevance of the CFIR constructs. Results: All 14 participants completed the questionnaires and interviews. Quantitative analysis revealed generally positive responses for user acceptance with all scores above the neutral score of 3.0. In addition, the mixed methods analysis identified two significant barriers (System Performance, Compatibility) and two major facilitators (Evidence Strength, Design Quality) for implementation of AI-based CDSSs in emergency care settings. Conclusions: Our mixed methods evaluation based on theoretically grounded frameworks revealed the acceptance, barriers, and facilitators of implementation of AI-based CDSS. Although the concern of system failure and overtrusting of the system could be barriers to implementation, the locality of the system and designing an intuitive user interface could likely facilitate the use of optimal AI-based CDSS. Alleviating and resolving these factors should be key to achieving good user acceptance of AI-based CDSS.
- 13 June 2022
- Published by JMIR Publications Inc. in JMIR Formative Research
- Vol. 6 (6) , e36501
Top Cited PapersOpen AccessAcceptance, Barriers, and Facilitators to Implementing Artificial Intelligence–Based Decision Support Systems in Emergency Departments: Quantitative and Qualitative Evaluation
Background: Despite the increasing availability of clinical decision support systems (CDSSs) and rising expectation for CDSSs based on artificial intelligence (AI), little is known about the acceptance of AI-based CDSS by physicians and its barriers and facilitators in emergency care settings. Objective: We aimed to evaluate the acceptance, barriers, and facilitators to implementing AI-based CDSSs in the emergency care setting through the opinions of physicians on our newly developed, real-time AI-based CDSS, which alerts ED physicians by predicting aortic dissection based on numeric and text information from medical charts, by using the Unified Theory of Acceptance and Use of Technology (UTAUT; for quantitative evaluation) and the Consolidated Framework for Implementation Research (CFIR; for qualitative evaluation) frameworks. Methods: This mixed methods study was performed from March to April 2021. Transitional year residents (n=6), emergency medicine residents (n=5), and emergency physicians (n=3) from two community, tertiary care hospitals in Japan were included. We first developed a real-time CDSS for predicting aortic dissection based on numeric and text information from medical charts (eg, chief complaints, medical history, vital signs) with natural language processing. This system was deployed on the internet, and the participants used the system with clinical vignettes of model cases. Participants were then involved in a mixed methods evaluation consisting of a UTAUT-based questionnaire with a 5-point Likert scale (quantitative) and a CFIR-based semistructured interview (qualitative). Cronbach α was calculated as a reliability estimate for UTAUT subconstructs. Interviews were sampled, transcribed, and analyzed using the MaxQDA software. The framework analysis approach was used during the study to determine the relevance of the CFIR constructs. Results: All 14 participants completed the questionnaires and interviews. Quantitative analysis revealed generally positive responses for user acceptance with all scores above the neutral score of 3.0. In addition, the mixed methods analysis identified two significant barriers (System Performance, Compatibility) and two major facilitators (Evidence Strength, Design Quality) for implementation of AI-based CDSSs in emergency care settings. Conclusions: Our mixed methods evaluation based on theoretically grounded frameworks revealed the acceptance, barriers, and facilitators of implementation of AI-based CDSS. Although the concern of system failure and overtrusting of the system could be barriers to implementation, the locality of the system and designing an intuitive user interface could likely facilitate the use of optimal AI-based CDSS. Alleviating and resolving these factors should be key to achieving good user acceptance of AI-based CDSS.
- Open Access
- 16 January 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Acceptance, Barriers, and Facilitators to Implementing Artificial Intelligence–Based Decision Support Systems in Emergency Departments: Quantitative and Qualitative Evaluation (Preprint)
BACKGROUND Despite the increasing availability of clinical decision support systems (CDSSs) and rising expectation for CDSSs based on artificial intelligence (AI), little is known about the acceptance of AI-based CDSS by physicians and its barriers and facilitators in emergency care settings. OBJECTIVE We aimed to evaluate the acceptance, barriers, and facilitators to implementing AI-based CDSSs in the emergency care setting through the opinions of physicians on our newly developed, real-time AI-based CDSS, which alerts ED physicians by predicting aortic dissection based on numeric and text information from medical charts, by using the Unified Theory of Acceptance and Use of Technology (UTAUT; for quantitative evaluation) and the Consolidated Framework for Implementation Research (CFIR; for qualitative evaluation) frameworks. METHODS This mixed methods study was performed from March to April 2021. Transitional year residents (n=6), emergency medicine residents (n=5), and emergency physicians (n=3) from two community, tertiary care hospitals in Japan were included. We first developed a real-time CDSS for predicting aortic dissection based on numeric and text information from medical charts (eg, chief complaints, medical history, vital signs) with natural language processing. This system was deployed on the internet, and the participants used the system with clinical vignettes of model cases. Participants were then involved in a mixed methods evaluation consisting of a UTAUT-based questionnaire with a 5-point Likert scale (quantitative) and a CFIR-based semistructured interview (qualitative). Cronbach α was calculated as a reliability estimate for UTAUT subconstructs. Interviews were sampled, transcribed, and analyzed using the MaxQDA software. The framework analysis approach was used during the study to determine the relevance of the CFIR constructs. RESULTS All 14 participants completed the questionnaires and interviews. Quantitative analysis revealed generally positive responses for user acceptance with all scores above the neutral score of 3.0. In addition, the mixed methods analysis identified two significant barriers (System Performance, Compatibility) and two major facilitators (Evidence Strength, Design Quality) for implementation of AI-based CDSSs in emergency care settings. CONCLUSIONS Our mixed methods evaluation based on theoretically grounded frameworks revealed the acceptance, barriers, and facilitators of implementation of AI-based CDSS. Although the concern of system failure and overtrusting of the system could be barriers to implementation, the locality of the system and designing an intuitive user interface could likely facilitate the use of optimal AI-based CDSS. Alleviating and resolving these factors should be key to achieving good user acceptance of AI-based CDSS.
- 16 January 2022
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessAcceptance, Barriers, and Facilitators to Implementing Artificial Intelligence–Based Decision Support Systems in Emergency Departments: Quantitative and Qualitative Evaluation (Preprint)
BACKGROUND Despite the increasing availability of clinical decision support systems (CDSSs) and rising expectation for CDSSs based on artificial intelligence (AI), little is known about the acceptance of AI-based CDSS by physicians and its barriers and facilitators in emergency care settings. OBJECTIVE We aimed to evaluate the acceptance, barriers, and facilitators to implementing AI-based CDSSs in the emergency care setting through the opinions of physicians on our newly developed, real-time AI-based CDSS, which alerts ED physicians by predicting aortic dissection based on numeric and text information from medical charts, by using the Unified Theory of Acceptance and Use of Technology (UTAUT; for quantitative evaluation) and the Consolidated Framework for Implementation Research (CFIR; for qualitative evaluation) frameworks. METHODS This mixed methods study was performed from March to April 2021. Transitional year residents (n=6), emergency medicine residents (n=5), and emergency physicians (n=3) from two community, tertiary care hospitals in Japan were included. We first developed a real-time CDSS for predicting aortic dissection based on numeric and text information from medical charts (eg, chief complaints, medical history, vital signs) with natural language processing. This system was deployed on the internet, and the participants used the system with clinical vignettes of model cases. Participants were then involved in a mixed methods evaluation consisting of a UTAUT-based questionnaire with a 5-point Likert scale (quantitative) and a CFIR-based semistructured interview (qualitative). Cronbach α was calculated as a reliability estimate for UTAUT subconstructs. Interviews were sampled, transcribed, and analyzed using the MaxQDA software. The framework analysis approach was used during the study to determine the relevance of the CFIR constructs. RESULTS All 14 participants completed the questionnaires and interviews. Quantitative analysis revealed generally positive responses for user acceptance with all scores above the neutral score of 3.0. In addition, the mixed methods analysis identified two significant barriers (System Performance, Compatibility) and two major facilitators (Evidence Strength, Design Quality) for implementation of AI-based CDSSs in emergency care settings. CONCLUSIONS Our mixed methods evaluation based on theoretically grounded frameworks revealed the acceptance, barriers, and facilitators of implementation of AI-based CDSS. Although the concern of system failure and overtrusting of the system could be barriers to implementation, the locality of the system and designing an intuitive user interface could likely facilitate the use of optimal AI-based CDSS. Alleviating and resolving these factors should be key to achieving good user acceptance of AI-based CDSS.
- 11 June 2021
- Published by SAGE Publications in Journal of Diabetes Science and Technology
- Vol. 16 (5) , 1128-1135
Accuracy and Stability of a Subcutaneous Flash Glucose Monitoring System in Critically Ill Patients
Flash glucose monitoring (FGM) systems can reduce glycemic variability and facilitate blood glucose management within the target range. However, in critically ill patients, only small (n < 30) studies have examined the accuracy of FGM and none have assessed the stability of FGM accuracy. We evaluated the accuracy and stability of FGM in critically ill patients. This was a single-center, retrospective observational study. We included a total of 116 critically ill patients who underwent FGM for glycemic control. The accuracy of FGM was assessed as follows using blood gas glucose values as a reference: (1) numerical accuracy using the mean absolute relative difference, (2) clinical accuracy using consensus error grid analysis, and (3) stability of accuracy assessing 14-day trends in consensus error grid distribution. FGM sensors remained in situ for a median of 6 [4, 11] days. We compared 2014 pairs of measurements between the sensor and blood gas analysis. Glucose values from the sensor were consistently lower, with a mean absolute relative difference of 13.8% (±16.0%), than those from blood gas analysis. Consensus error grid analysis demonstrated 99.4% of the readings to be in a clinically acceptable accuracy zone. The accuracy of FGM was stable across the 14 days after device insertion. FGM had acceptable reliability and accuracy to arterial blood gas analysis in critically ill patients. In addition, the accuracy of FGM persisted for at least 14 days. Our study promotes the potential usefulness of FGM for glycemic monitoring in critically ill patients.
- 11 June 2021
- Published by SAGE Publications in Journal of Diabetes Science and Technology
- Vol. 16 (5) , 1128-1135
Accuracy and Stability of a Subcutaneous Flash Glucose Monitoring System in Critically Ill Patients
Flash glucose monitoring (FGM) systems can reduce glycemic variability and facilitate blood glucose management within the target range. However, in critically ill patients, only small (n < 30) studies have examined the accuracy of FGM and none have assessed the stability of FGM accuracy. We evaluated the accuracy and stability of FGM in critically ill patients. This was a single-center, retrospective observational study. We included a total of 116 critically ill patients who underwent FGM for glycemic control. The accuracy of FGM was assessed as follows using blood gas glucose values as a reference: (1) numerical accuracy using the mean absolute relative difference, (2) clinical accuracy using consensus error grid analysis, and (3) stability of accuracy assessing 14-day trends in consensus error grid distribution. FGM sensors remained in situ for a median of 6 [4, 11] days. We compared 2014 pairs of measurements between the sensor and blood gas analysis. Glucose values from the sensor were consistently lower, with a mean absolute relative difference of 13.8% (±16.0%), than those from blood gas analysis. Consensus error grid analysis demonstrated 99.4% of the readings to be in a clinically acceptable accuracy zone. The accuracy of FGM was stable across the 14 days after device insertion. FGM had acceptable reliability and accuracy to arterial blood gas analysis in critically ill patients. In addition, the accuracy of FGM persisted for at least 14 days. Our study promotes the potential usefulness of FGM for glycemic monitoring in critically ill patients.
- Open Access
- 27 May 2021
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 16 (5) , e0252158
Time in blood glucose range 70 to 180 mg/dL and survival rate in critically ill patients: A retrospective cohort study
Background: While time in targeted blood glucose range (TIR) 70–140 mg/dL is a known factor associated with mortality in critically ill patients, it remains unclear whether TIR is associated with 28-day mortality under the glycemic control with a less tight target glucose range of 70–180 mg/dL. We aimed to examine whether TIR 70–180 mg/dL was associated with 28-day mortality.Methods: This is a retrospective cohort study using data from a tertiary care center in Japan collected from January 2016 through October 2019. We included adult patients (aged ≥20 years) admitted to the ICU. We excluded patients 1) with diabetic ketoacidosis patients, 2) discharged within 48 hours, 3) with repeated ICU admissions. We calculated TIR 70–180 mg/dL using the measured blood glucose values (≥3 times per day). The primary outcome was 28-day mortality. We examined the association between TIR and 28-day mortality using a logistic regression and Cox proportional hazard models with a stratification by glycosylated hemoglobin (HbA1c) level of 6.5%. Additionally, we repeated the analyses using the TIR category to assess the optimal TIR. For the sensitivity analysis, we repeated the primary analysis using TIR during the first three days from ICU admission.Results: Of 1,230 patients, the median age was 72 years, 65% were male, and 250 patients (20%) had HbA1c ≥6.5% on admission. In patients with HbA1c <6.5%, TIR <80% was associated with an increased risk of 28-day mortality, with an adjusted odds ratio (OR) of 1.88 (95%CI: 1.36–2.61). Likewise, when using 10% incremental TIR as a categorical variable, lower TIR was associated with a worse 28-day mortality compared with TIR ≥90% (e.g., adjusted OR of TIR <60%, 3.62 [95%CI 2.36–5.53]). Similar associations were found in the analyses using Cox proportional hazards model and using TIR during the first three days. By contrast, in patients with HbA1c ≥6.5%, there was no consistent association of TIR with 28-day mortality.Conclusions: We found that lower TIR 70–180 mg/dL was associated with a higher 28-day mortality in critically ill patients with HbA1c <6.5%, whereas there was no consistent association in patients with HbA1c ≥6.5%.
- 27 May 2021
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 16 (5) , e0252158
Open AccessTime in blood glucose range 70 to 180 mg/dL and survival rate in critically ill patients: A retrospective cohort study
Background: While time in targeted blood glucose range (TIR) 70–140 mg/dL is a known factor associated with mortality in critically ill patients, it remains unclear whether TIR is associated with 28-day mortality under the glycemic control with a less tight target glucose range of 70–180 mg/dL. We aimed to examine whether TIR 70–180 mg/dL was associated with 28-day mortality.Methods: This is a retrospective cohort study using data from a tertiary care center in Japan collected from January 2016 through October 2019. We included adult patients (aged ≥20 years) admitted to the ICU. We excluded patients 1) with diabetic ketoacidosis patients, 2) discharged within 48 hours, 3) with repeated ICU admissions. We calculated TIR 70–180 mg/dL using the measured blood glucose values (≥3 times per day). The primary outcome was 28-day mortality. We examined the association between TIR and 28-day mortality using a logistic regression and Cox proportional hazard models with a stratification by glycosylated hemoglobin (HbA1c) level of 6.5%. Additionally, we repeated the analyses using the TIR category to assess the optimal TIR. For the sensitivity analysis, we repeated the primary analysis using TIR during the first three days from ICU admission.Results: Of 1,230 patients, the median age was 72 years, 65% were male, and 250 patients (20%) had HbA1c ≥6.5% on admission. In patients with HbA1c <6.5%, TIR <80% was associated with an increased risk of 28-day mortality, with an adjusted odds ratio (OR) of 1.88 (95%CI: 1.36–2.61). Likewise, when using 10% incremental TIR as a categorical variable, lower TIR was associated with a worse 28-day mortality compared with TIR ≥90% (e.g., adjusted OR of TIR <60%, 3.62 [95%CI 2.36–5.53]). Similar associations were found in the analyses using Cox proportional hazards model and using TIR during the first three days. By contrast, in patients with HbA1c ≥6.5%, there was no consistent association of TIR with 28-day mortality.Conclusions: We found that lower TIR 70–180 mg/dL was associated with a higher 28-day mortality in critically ill patients with HbA1c <6.5%, whereas there was no consistent association in patients with HbA1c ≥6.5%.
- 15 February 2021
- Published by Oxford University Press (OUP) in American Journal of Respiratory and Critical Care Medicine
- Vol. 203 (4) , 515-518
Urine Titin N-Fragment as a Biomarker of Muscle Injury for Critical Illness Myopathy
- 15 February 2021
- Published by Oxford University Press (OUP) in American Journal of Respiratory and Critical Care Medicine
- Vol. 203 (4) , 515-518
Urine Titin N-Fragment as a Biomarker of Muscle Injury for Critical Illness Myopathy
- Open Access
- 27 October 2020
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 8 (10) , e20324
Institution-Specific Machine Learning Models for Prehospital Assessment to Predict Hospital Admission: Prediction Model Development Study
Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Although multiple prediction models have been developed to predict hospital admission to emergency departments (EDs) to address overcrowding and patient safety, few studies have examined prediction models for prehospital use. Development of institution-specific prediction models is feasible in this age of data science, provided that predictor-related information is readily collectable. Objective: We aimed at developing a hospital admission prediction model based on patient information that is commonly available during ambulance transport before hospitalization. Methods: Patients transported by ambulance to our ED from April 2018 through March 2019 were enrolled. Candidate predictors were age, sex, chief complaint, vital signs, and patient medical history, all of which were recorded by emergency medical teams during ambulance transport. Patients were divided into two cohorts for derivation (70%) and validation (30%). For statistical models, logistic regression, logistic lasso, random forest, and gradient boosting machine were used. Prediction models were developed in the derivation cohort. Model performance was assessed by area under the receiver–operator characteristic curve (ROC–AUC) and association measures in the validation cohort. Results: Of 5145 patients transported by ambulance, including deaths in the ED and hospital transfers, 2699 (52.5%) required hospital admission. Prediction performance was higher according to the addition of predictive factors, attaining the best performance of an ROC–AUC of 0.818 (95% confidence interval 0.792–0.839) with a machine learning model and predictive factors of age, sex, chief complaint, and vital signs. Sensitivity and specificity of this model were, respectively, 0.744 (0.716–0.773) and 0.745 (0.709–0.776). Conclusions: For patients transferred to EDs, we developed a well-performing hospital admission prediction model based on prehospital information including chief complaints.
- 27 October 2020
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 8 (10) , e20324
Open AccessInstitution-Specific Machine Learning Models for Prehospital Assessment to Predict Hospital Admission: Prediction Model Development Study
Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Although multiple prediction models have been developed to predict hospital admission to emergency departments (EDs) to address overcrowding and patient safety, few studies have examined prediction models for prehospital use. Development of institution-specific prediction models is feasible in this age of data science, provided that predictor-related information is readily collectable. Objective: We aimed at developing a hospital admission prediction model based on patient information that is commonly available during ambulance transport before hospitalization. Methods: Patients transported by ambulance to our ED from April 2018 through March 2019 were enrolled. Candidate predictors were age, sex, chief complaint, vital signs, and patient medical history, all of which were recorded by emergency medical teams during ambulance transport. Patients were divided into two cohorts for derivation (70%) and validation (30%). For statistical models, logistic regression, logistic lasso, random forest, and gradient boosting machine were used. Prediction models were developed in the derivation cohort. Model performance was assessed by area under the receiver–operator characteristic curve (ROC–AUC) and association measures in the validation cohort. Results: Of 5145 patients transported by ambulance, including deaths in the ED and hospital transfers, 2699 (52.5%) required hospital admission. Prediction performance was higher according to the addition of predictive factors, attaining the best performance of an ROC–AUC of 0.818 (95% confidence interval 0.792–0.839) with a machine learning model and predictive factors of age, sex, chief complaint, and vital signs. Sensitivity and specificity of this model were, respectively, 0.744 (0.716–0.773) and 0.745 (0.709–0.776). Conclusions: For patients transferred to EDs, we developed a well-performing hospital admission prediction model based on prehospital information including chief complaints.
- Open Access
Time in blood glucose range 70 to 180 mg/dL and survival rate in critically ill patients: A retrospective observational study
Objective: Time in targeted blood glucose range (TIR) 70-140 mg/dL has been associated with an increased risk of mortality in critically ill patients. Nevertheless, it remains unclear whether TIR is associated with 28-day mortality in critically ill patients under glycemic control with a less tight target glucose range of 70-180 mg/dL. We aimed to assess whether TIR 70-180 mg/dL was associated with 28-day mortality and to identify the optimal TIR. Design: A retrospective observational study. Setting: Data from a tertiary care centre in Japan, from 1 January 2016 through 31 October 2019. Participants: 1,230 adult patients admitted to the intensive care unit for more than three days. Outcome measure: The primary outcome was 28-day mortality. Results: Of 1,230 patients, patients with HbA1c ≥6.5% had a higher 28-day mortality than those with <6.5% (32.0% vs. 22.7%; p=0.003). In the multivariate logistic regression, TIR <80% was associated with an increased risk of 28-day mortality in patients with HbA1c <6.5% with an adjusted odds ratio (OR) of 1.88 (95% confidence interval [CI]: 1.36-2.61). When using 10% incremental TIR as a categorical variable, lower TIR was associated with worse 28-day mortality compared to TIR ≥90% in patients with HbA1c <6.5% (e.g., adjusted OR of TIR <60%, 3.62 [95%CI 2.36-5.53]). Similar associations were found in the analyses using the COX proportional hazards model. In addition, sensitivity analyses using TIR of the first three days showed that the overall associations were consistent with primary analyses. Conclusions: Our study demonstrated that lower TIR 70-180 mg/dL was associated with higher 28-day mortality in nondiabetic critically ill patients. Article Summary: Strengths and limitations of this study: This is the first study to evaluate the association between less tight TIR (70-180 mg/dl) and mortality in critically ill patients. We found consistent results in the analyses using logistic regression model and Cox proportional hazard model. The primary findings were also observed in the analysis using TIR in the first three days. We did not use a continuous glucose monitoring device to calculate TIR. Due to the nature of the single-centre study design, the generalizability of the findings might be limited.
Open AccessTime in blood glucose range 70 to 180 mg/dL and survival rate in critically ill patients: A retrospective observational study
Objective: Time in targeted blood glucose range (TIR) 70-140 mg/dL has been associated with an increased risk of mortality in critically ill patients. Nevertheless, it remains unclear whether TIR is associated with 28-day mortality in critically ill patients under glycemic control with a less tight target glucose range of 70-180 mg/dL. We aimed to assess whether TIR 70-180 mg/dL was associated with 28-day mortality and to identify the optimal TIR. Design: A retrospective observational study. Setting: Data from a tertiary care centre in Japan, from 1 January 2016 through 31 October 2019. Participants: 1,230 adult patients admitted to the intensive care unit for more than three days. Outcome measure: The primary outcome was 28-day mortality. Results: Of 1,230 patients, patients with HbA1c ≥6.5% had a higher 28-day mortality than those with <6.5% (32.0% vs. 22.7%; p=0.003). In the multivariate logistic regression, TIR <80% was associated with an increased risk of 28-day mortality in patients with HbA1c <6.5% with an adjusted odds ratio (OR) of 1.88 (95% confidence interval [CI]: 1.36-2.61). When using 10% incremental TIR as a categorical variable, lower TIR was associated with worse 28-day mortality compared to TIR ≥90% in patients with HbA1c <6.5% (e.g., adjusted OR of TIR <60%, 3.62 [95%CI 2.36-5.53]). Similar associations were found in the analyses using the COX proportional hazards model. In addition, sensitivity analyses using TIR of the first three days showed that the overall associations were consistent with primary analyses. Conclusions: Our study demonstrated that lower TIR 70-180 mg/dL was associated with higher 28-day mortality in nondiabetic critically ill patients. Article Summary: Strengths and limitations of this study: This is the first study to evaluate the association between less tight TIR (70-180 mg/dl) and mortality in critically ill patients. We found consistent results in the analyses using logistic regression model and Cox proportional hazard model. The primary findings were also observed in the analysis using TIR in the first three days. We did not use a continuous glucose monitoring device to calculate TIR. Due to the nature of the single-centre study design, the generalizability of the findings might be limited.
- Open Access
- 4 June 2020
- Published by JMIR Publications Inc. in JMIR Preprints
Institution-Specific Machine Learning Models for Prehospital Assessment to Predict Hospital Admission: Prediction Model Development Study (Preprint)
BACKGROUND Although multiple prediction models have been developed to predict hospital admission to emergency departments (EDs) to address overcrowding and patient safety, only a few studies have examined prediction models for prehospital use. Development of institution-specific prediction models is feasible in this age of data science, provided that predictor-related information is readily collectable. OBJECTIVE We aimed to develop a hospital admission prediction model based on patient information that is commonly available during ambulance transport before hospitalization. METHODS Patients transported by ambulance to our ED from April 2018 through March 2019 were enrolled. Candidate predictors were age, sex, chief complaint, vital signs, and patient medical history, all of which were recorded by emergency medical teams during ambulance transport. Patients were divided into two cohorts for derivation (3601/5145, 70.0%) and validation (1544/5145, 30.0%). For statistical models, logistic regression, logistic lasso, random forest, and gradient boosting machine were used. Prediction models were developed in the derivation cohort. Model performance was assessed by area under the receiver operating characteristic curve (AUROC) and association measures in the validation cohort. RESULTS Of 5145 patients transported by ambulance, including deaths in the ED and hospital transfers, 2699 (52.5%) required hospital admission. Prediction performance was higher with the addition of predictive factors, attaining the best performance with an AUROC of 0.818 (95% CI 0.792-0.839) with a machine learning model and predictive factors of age, sex, chief complaint, and vital signs. Sensitivity and specificity of this model were 0.744 (95% CI 0.716-0.773) and 0.745 (95% CI 0.709-0.776), respectively. CONCLUSIONS For patients transferred to EDs, we developed a well-performing hospital admission prediction model based on routinely collected prehospital information including chief complaints.
- 4 June 2020
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessInstitution-Specific Machine Learning Models for Prehospital Assessment to Predict Hospital Admission: Prediction Model Development Study (Preprint)
BACKGROUND Although multiple prediction models have been developed to predict hospital admission to emergency departments (EDs) to address overcrowding and patient safety, only a few studies have examined prediction models for prehospital use. Development of institution-specific prediction models is feasible in this age of data science, provided that predictor-related information is readily collectable. OBJECTIVE We aimed to develop a hospital admission prediction model based on patient information that is commonly available during ambulance transport before hospitalization. METHODS Patients transported by ambulance to our ED from April 2018 through March 2019 were enrolled. Candidate predictors were age, sex, chief complaint, vital signs, and patient medical history, all of which were recorded by emergency medical teams during ambulance transport. Patients were divided into two cohorts for derivation (3601/5145, 70.0%) and validation (1544/5145, 30.0%). For statistical models, logistic regression, logistic lasso, random forest, and gradient boosting machine were used. Prediction models were developed in the derivation cohort. Model performance was assessed by area under the receiver operating characteristic curve (AUROC) and association measures in the validation cohort. RESULTS Of 5145 patients transported by ambulance, including deaths in the ED and hospital transfers, 2699 (52.5%) required hospital admission. Prediction performance was higher with the addition of predictive factors, attaining the best performance with an AUROC of 0.818 (95% CI 0.792-0.839) with a machine learning model and predictive factors of age, sex, chief complaint, and vital signs. Sensitivity and specificity of this model were 0.744 (95% CI 0.716-0.773) and 0.745 (95% CI 0.709-0.776), respectively. CONCLUSIONS For patients transferred to EDs, we developed a well-performing hospital admission prediction model based on routinely collected prehospital information including chief complaints.