富田 啓介

This study investigated the accuracy of a machine learning algorithm for predicting mortality in patients receiving rapid response system (RRS) activation. This retrospective cohort study used data from the In-Hospital Emergency Registry in Japan, which collects nationwide data on patients receiving RRS activation. The missing values in the dataset were replaced using multiple imputations (mode imputation, BayseRidge sklearn. linear model, and K-nearest neighbor model), and the enrolled patients were randomly assigned to the training and test cohorts. We established prediction models for 30-day mortality using the following four types of machine learning classifiers: Light Gradient Boosting Machine (LightGBM), eXtreme Gradient Boosting, random forest, and neural network. Fifty-two variables (patient characteristics, details of RRS activation, reasons for RRS initiation, and hospital capacity) were used to construct the prediction algorithm. The primary outcome was the accuracy of the prediction model for 30-day mortality. Overall, the data from 4,997 patients across 34 hospitals were analyzed. The machine learning algorithms using LightGBM demonstrated the highest predictive value for 30-day mortality (area under the receiver operating characteristic curve, 0.860 [95 % confidence interval, 0.825-0.895]). The SHapley Additive exPlanations summary plot indicated that hospital capacity, site of incidence, code status, and abnormal vital signs within 24 h were important variables in the prediction model for 30-day mortality.

To efficiently allocate medical resources at disaster sites, medical workers perform triage to prioritize medical treatments based on the severity of the wounded or sick. In such instances, evaluators often assess the severity status of the wounded or sick quickly, but their measurements are qualitative and rely on experience. Therefore, we developed a wearable device called Medic Hand in this study to extend the functionality of a medical worker's hand so as to measure multiple biometric indicators simultaneously without increasing the number of medical devices to be carried. Medic Hand was developed to quantitatively and efficiently evaluate "perfusion" during triage. Speed is essential during triage at disaster sites, where time and effort are often spared to attach medical devices to patients, so the use of Medic Hand as a biometric measurement device is more efficient for collecting biometric information. For Medic Hand to be handy during disasters, it is essential to understand and improve upon factors that facilitate its public acceptance. To this end, this paper reports on the usability evaluation of Medic Hand through a questionnaire survey of nonmedical workers.

Increased fluid overload (FO) is associated with poor outcomes in critically ill patients, especially in acute kidney injury (AKI). However, the exact timing from when FO influences outcomes remains unclear. We retrospectively screened intensive care unit (ICU) admitted patients with AKI between January 2011 and December 2015. Logistic or linear regression analyses were performed to determine when hourly %FO was significant on 90-day in-hospital mortality (primary outcome) or ventilator-free days (VFDs). In total, 1120 patients were enrolled in this study. Univariate analysis showed that a higher %FO was significantly associated with higher mortality from the first hour of ICU admission (odds ratio 1.34, 95% confidence interval 1.15-1.56, P < 0.001), whereas multivariate analysis adjusted with age, sex, APACHE II score, and sepsis etiology showed the association was significant from the 27th hour. Both univariate and multivariate analyses showed that a higher %FO was significantly associated with shorter VFDs from the 1st hour. The significant associations were retained during all following observation periods after they showed significance. In patients with AKI, a higher %FO was associated with higher mortality and shorter VFDs from the early phase after ICU admission. FO should be administered with a physiological target or goal in place from the initial phase of critical illness.

Predicting out-of-hospital cardiac arrest (OHCA) events might improve outcomes of OHCA patients. We hypothesized that machine learning algorithms using meteorological information would predict OHCA incidences. We used the Japanese population-based repository database of OHCA and weather information. The Tokyo data (2005-2012) was used as the training cohort and datasets of the top six populated prefectures (2013-2015) as the test. Eight various algorithms were evaluated to predict the high-incidence OHCA days, defined as the daily events exceeding 75% tile of our dataset, using meteorological and chronological values: temperature, humidity, air pressure, months, days, national holidays, the day before the holidays, the day after the holidays, and New Year's holidays. Additionally, we evaluated the contribution of each feature by Shapley Additive exPlanations (SHAP) values. The training cohort included 96,597 OHCA patients. The eXtreme Gradient Boosting (XGBoost) had the highest area under the receiver operating curve (AUROC) of 0.906 (95% confidence interval; 0.868-0.944). In the test cohorts, the XGBoost algorithms also had high AUROC (0.862-0.923). The SHAP values indicated that the "mean temperature on the previous day" impacted the most on the model. Algorithms using machine learning with meteorological and chronological information could predict OHCA events accurately.

While the development of prehospital diagnosis scales has been reported in various regions, we have also developed a scale to predict stroke type using machine learning. In the present study, we aimed to assess for the first time a scale that predicts the need for surgical intervention across stroke types, including subarachnoid haemorrhage and intracerebral haemorrhage. A multicentre retrospective study was conducted within a secondary medical care area. Twenty-three items, including vitals and neurological symptoms, were analysed in adult patients suspected of having a stroke by paramedics. The primary outcome was a binary classification model for predicting surgical intervention based on eXtreme Gradient Boosting (XGBoost). Of the 1143 patients enrolled, 765 (70%) were used as the training cohort, and 378 (30%) were used as the test cohort. The XGBoost model predicted stroke requiring surgical intervention with high accuracy in the test cohort, with an area under the receiver operating characteristic curve of 0.802 (sensitivity 0.748, specificity 0.853). We found that simple survey items, such as the level of consciousness, vital signs, sudden headache, and speech abnormalities were the most significant variables for accurate prediction. This algorithm can be useful for prehospital stroke management, which is crucial for better patient outcomes.

INTRODUCTION: Capillary refill time (CRT) is an indicator of peripheral circulation and is recommended in the 2021 guidelines for treating and managing sepsis. STUDY OBJECTIVE: This study developed a portable device to realize objective CRT measurement. Assuming that peripheral blood flow obstruction by the artery occlusion test (AOT) or venous occlusion test (VOT) increases the CRT, the cut-off value for peripheral circulatory failure was studied by performing a comparative analysis with CRT with no occlusion test (No OT). METHODS: Fourteen (14) healthy adults (age: 20-26 years) participated in the study. For the vascular occlusion test, a sphygmomanometer was placed on the left upper arm of the participant in the supine position, and a pressure of 30mmHg higher than the systolic pressure was applied for AOT, a pressure of 60mmHg was applied for VOT, respectively, and no pressure was applied for No OT. The CRT was measured from the index finger of the participant's left hand. RESULTS: Experimental results revealed that CRT was significantly longer in the AOT and did not differ significantly in the VOT. The cut-off value for peripheral circulatory failure was found to be 2.88 seconds based on Youden's index by using receiver operating characteristic (ROC) analysis with AOT as positive and No OT as negative. CONCLUSION: Significant results were obtained in a previous study on the evaluation of septic shock patients when CRT > three seconds was considered abnormal, and the cut-off value for peripheral circulatory failure in the current study validated this.

Traumatic brain injury (TBI) in the form of diffuse axonal injury (DAI) is difficult to diagnose in the early phase of the injury. Early diagnosis of DAI may provide opportunity for developing treatment and management strategies. Tau protein has been demonstrated to increase in the early phase of TBI with high diagnostic accuracy in patients with DAI. We tested the biological plausibility of tau protein using a rat DAI model by evaluating the association between serum tau levels and the severity of brain injury. DAI was induced in animals using the Marmarou model. After a survival of 60 minutes, rats were anesthetized and sacrificed after obtaining blood samples (5ml) from the heart. Eighteen rats were employed in the present study and were randomly subjected to sham-operated control (n=4), mild DAI (n=7), and severe DAI (n=7). Of seven severe DAI rats, two rats that had focal injury caused by skull fracture were excluded in the measurement of tau protein level. The serum levels of tau protein in the rat DAI model were found to increase significantly and consistently according to the severity of the injury. Rats with DAI showed significantly higher serum levels of tau protein compared to sham rats; the severe DAI rats had higher levels of tau than moderate DAI and sham rats (sham vs. mild,  P=0.02; mild vs. severe,  P=0.02). In conclusion, serum tau protein levels may be useful as a biomarker for diagnosing and estimating the severity of DAI in the early phase.

We evaluated the accuracy and precision of a novel non-invasive monitoring device in comparison with conventional monitoring methods used in intensive care units (ICU). The study device was developed to measure blood pressure, pulse rate, respiratory rate, and oxygen saturation, continuously with a single sensor using the photoplethysmographic technique. Patients who were monitored with arterial pressure lines in the ICU were enrolled. Systolic and diastolic blood pressure, pulse rate, respiratory rate, and arterial oxygen saturation were measured continuously for 30 min at 5-min intervals using the conventional methods and the study device. The primary outcome variable was blood pressure. Blood pressure measured by the study device highly correlated with the arterial pressure line values (correlation coefficients > 0.95). Percent errors for systolic, diastolic and mean blood pressures were 2.4% and 6.7% and 6.5%, respectively. Percent errors for pulse rate, respiratory rate and oxygen saturation were 3.4%, 5.6% and 1.4%, respectively. The non-invasive, continuous, multi-parameter monitoring device presented high level of agreement with the invasive arterial blood pressure monitoring, along with sufficient accuracy and precision in the measurements of pulse rate, respiratory rate, and oxygen saturation.

BACKGROUND: Diffuse axonal injury (DAI) is difficult to identify in the early phase of traumatic brain injury (TBI) using common diagnostic methods. Tau protein is localized specifically in nerve axons. We hypothesized that serum level of tau can be a useful biomarker to diagnose DAI in the early phase of TBI. METHODS & RESULTS: We measured serum tau levels in 40 TBI patients who were suspected of DAI within 6 hours after TBI to evaluate the accuracy of the tau level as a diagnostic marker for DAI. Diagnosis of DAI was confirmed according to magnetic resonance imaging (MRI) findings. The serum tau level in the DAI group (n = 13) was significantly higher than that in the non-DAI group (n = 27) (DAI vs. non-DAI, 25.3 [0 to 99.1] pg/mL vs. 0 [0 to 44.4] pg/mL, P = 0.03)). A receiver-operating characteristic curve to evaluate the diagnostic ability of serum tau level within 6 hours for DAI showed an area under the curve of 0.690 with 74.1% for sensitivity and 69.2% for specificity. Serum tau level was not significantly higher in unfavorable outcome group (Glasgow Outcome scale [GOS] score = 1-3 at hospital discharge) compared with favorable outcome group (GOS score = 4-5) (P = 0.19). CONCLUSIONS: Tau protein may be a useful biomarker for diagnosis of DAI in the early phase of TBI.

BACKGROUND: Resuscitative endovascular balloon occlusion of the aorta (REBOA) performed by emergency physicians has been gaining acceptance as a less invasive technique than resuscitative thoracotomy. OBJECTIVE: To evaluate access-related complications and duration of occlusions during REBOA. METHODS: Patients with haemorrhagic shock requiring REBOA, from 18 hospitals in Japan, included in the DIRECT-IABO Registry were studied. REBOA-related characteristics were compared between non-survivors and survivors at 24 hours. 24-Hour survivors were categorised into groups with small (≤8 Fr), large (≥9 Fr) or unusual sheaths (oversized or multiple) to assess the relationship between the sheath size and complications. Haemodynamic response, occlusion duration and outcomes were compared between groups with partial and complete REBOA. RESULTS: Between August 2011 and December 2015, 142 adults undergoing REBOA were analysed. REBOA procedures were predominantly (94%) performed by emergency medicine (EM) physicians. The median duration of the small sheath (n=53) was 19 hours compared with 7.5 hours for the larger sheaths (P=0.025). Smaller sheaths were more likely to be removed using external manual compression (96% vs 45%, P<0.001). One case of a common femoral artery thrombus (large group) and two cases of amputation (unusual group) were identified. Partial REBOA was carried out in more cases (n=78) and resulted in a better haemodynamic response than complete REBOA (improvement in haemodynamics, 92% vs 70%, P=0.004; achievement of stability, 78% vs 51%, P=0.007) and allowed longer occlusion duration (median 58 vs 33 min, P=0.041). No statistically significant difference in 24-hour or 30-day survival was found between partial and complete REBOA. CONCLUSION: In Japan, EM physicians undertake the majority of REBOA procedures. Smaller sheaths appear to have fewer complications despite relatively prolonged placement and require external compression on removal. Although REBOA is a rarely performed procedure, partial REBOA, which may extend the occlusion duration without a reduction in survival, is used more commonly in Japan.

Blood purification therapy using hemofilters with high adsorbing capabilities has been reported to remove excessive humoral mediators from the blood of patients with sepsis. However, there are insufficient studies of the adsorbates bound to hemofilter membranes. We hypothesized that these adsorbates in acute kidney injury (AKI) patients with sepsis were different from those in patients without sepsis and that proteome analysis of the adsorbates would identify novel substances of sepsis. This study included 20 patients who had AKI upon admission to intensive care units (ICUs) and who received continuous renal replacement therapy using polymethyl methacrylate hemofilters. We isolated adsorbates from the hemofilters after use and performed comprehensive proteome analysis. A total of 429 proteins were identified in these adsorbates. Adsorbates from the hemofilters of patients with sepsis had significantly increased frequency of proteins associated with "immune system process" and "biological adhesion" functions compared to those of non-sepsis patients (P < 0.05). Of 429 proteins, 197 were identified only in sepsis adsorbates. Of these, 3 proteins including carbonic anhydrase 1 (CA1) and leucine-rich alpha-2-glycoprotein (LRG1) were identified in all samples from sepsis patients and have not been previously reported in sepsis patients. Validation analysis of patient serum revealed that patients with sepsis had increased serum levels of CA1 and LRG1 compared to patients without sepsis (P < 0.05). To conclude, there were significant differences in the characteristics of the adsorbates from sepsis and non-sepsis patients. CA1 and LRG1 appear to be novel substances associated with sepsis.

BACKGROUND: A sudden shortage of physician resources due to overwhelming patient needs can affect the quality of care in the emergency department (ED). Developing effective response strategies remains a challenging research area. We created a novel system using information and communication technology (ICT) to respond to a sudden shortage, and tested the system to determine whether it would compensate for a shortage. METHODS: Patients (n = 4890) transferred to a level I trauma center in Japan during 2012-2015 were studied. We assessed whether the system secured the necessary physicians without using other means such as phone or pager, and calculated fulfillment rate by the system as a primary outcome variable. We tested for the difference in probability of multiple casualties among total casualties transferred to the ED as an indicator of ability to respond to excessive patient needs, in a secondary analysis before and after system introduction. RESULTS: The system was activated 24 times (stand-by request [n = 12], attendance request [n = 12]) in 24 months, and secured the necessary physicians without using other means; fulfillment rate was 100%. There was no significant difference in the probability of multiple casualties during daytime weekdays hours before and after system introduction, while the probability of multiple casualties during night or weekend hours after system introduction significantly increased compared to before system introduction (4.8% vs. 12.9%, P < 0.0001). On the whole, the probability of multiple casualties increased more than 2 times after system introduction 6.2% vs. 13.6%, P < 0.0001). DISCUSSION: After introducing the system, probability of multiple casualties increased. Thus the system may contribute to improvement in the ability to respond to sudden excessive patient needs in multiple causalities. CONCLUSIONS: A novel system using ICT successfully secured immediate responses from needed physicians outside the hospital without increasing user workload, and increased the ability to respond to excessive patient needs. The system appears to be able to compensate for a shortage of physician in the ED due to excessive patient transfers, particularly during off-hours.

CASE: A 19-year-old man with a hemangioma that extended from the left arm to the axillary region had deteriorated due to shock, and no improvement was observed after fluid resuscitation. His status on arrival led to complications of hemorrhagic and septic shock with his left arm swollen and deep purple in color. Left arm amputation to control the source of bleeding and infection was thought to be indicated, however, the coagulation system had collapsed. Therefore, initial reduction of the blood flow to the hemangioma with angiographic endovascular treatment was carried out. Thereafter, continuous hemodiafiltration using a polymethylmethacrylate membrane hemofilter against hypercytokinemia was introduced in the intensive care unit for damage control, which resulted in success. OUTCOME: We completed the arm amputation, and subsequently carried out a latissimus dorsi muscle flap transfer on the amputated stump. The patient achieved an ambulatory discharge. CONCLUSIONS: We successfully treated the very rare case of massive venous malformation with shock due to hemorrhage and infection by performing damage control.