Machine learning algorithm for predicting 30-day mortality in patients receiving rapid response system activation: A retrospective nationwide cohort study.

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.

Impact of dynamic parameter of trends in vital signs on the prediction of serious events in hospitalized patients -a retrospective observational study.

Aim: Although early detection of patients' deterioration may improve outcomes, most of the detection criteria use on-the-spot values of vital signs. We investigated whether adding trend values over time enhanced the ability to predict adverse events among hospitalized patients. Methods: Patients who experienced adverse events, such as unexpected cardiac arrest or unplanned ICU admission were enrolled in this retrospective study. The association between the events and the combination of vital signs was evaluated at the time of the worst vital signs 0-8 hours before events (near the event) and at 24-48 hours before events (baseline). Multivariable logistic analysis was performed, and the area under the receiver operating characteristic curve (AUC) was used to assess the prediction power for adverse events among various combinations of vital sign parameters. Results: Among 24,509 in-patients, 54 patients experienced adverse events(cases) and 3,116 control patients eligible for data analysis were included. At the timepoint near the event, systolic blood pressure (SBP) was lower, heart rate (HR) and respiratory rate (RR) were higher in the case group, and this tendency was also observed at baseline. The AUC for event occurrence with reference to SBP, HR, and RR was lower when evaluated at baseline than at the timepoint near the event (0.85 [95%CI: 0.79-0.92] vs. 0.93 [0.88-0.97]). When the trend in RR was added to the formula constructed of baseline values of SBP, HR, and RR, the AUC increased to 0.92 [0.87-0.97]. Conclusion: Trends in RR may enhance the accuracy of predicting adverse events in hospitalized patients.

Incidence and factors associated with newly implemented do-not-attempt-resuscitation orders among deteriorating patients after rapid response system activation: A retrospective observational study using a Japanese multicenter database.

Aim: The rapid response system (RRS) was initially aimed to improve patient outcomes. Recently, some studies have implicated that RRS might facilitate do-not-attempt-resuscitation (DNAR) orders among patients, their families, and healthcare providers. This study aimed to examine the incidence and factors independently associated with DNAR orders newly implemented after RRS activation among deteriorating patients. Methods: This observational study assessed patients who required RRS activation between 2012 and 2021 in Japan. We investigated patients' characteristics and the incidence of new DNAR orders after RRS activation. Furthermore, we used multivariable hierarchical logistic regression models to explore independent predictors of new DNAR orders. Results: We identified 7904 patients (median age, 72 years; 59% male) who required RRS activation at 29 facilities. Of the 7066 patients without pre-existing DNAR orders before RRS activation, 394 (5.6%) had new DNAR orders. Multivariable hierarchical logistic regression analyses revealed that new DNAR orders were associated with age category (adjusted odds ratio [aOR], 1.56; 95% confidence interval, 1.12-2.17 [65-74 years old reference to 20-64 years old], aOR, 2.56; 1.92-3.42 [75-89 years old], and aOR, 6.58; 4.17-10.4 [90 years old]), malignancy (aOR, 1.82; 1.42-2.32), postoperative status (aOR, 0.45; 0.30-0.71), and National Early Warning Score 2 (aOR, 1.07; 1.02-1.12 [per 1 score]). Conclusion: The incidence of new DNAR orders was one in 18 patients after RRS activation. The factors associated with new DNAR orders were age, malignancy, postoperative status, and National Early Warning Score 2.

Prevalence and predictors of direct discharge home following hospitalization of patients with serious adverse events managed by the rapid response system in Japan: a multicenter, retrospective, observational study.

AIM: The rapid response system (RRS) is an in-hospital medical safety system. To date, not much is known about patient disposition after RRS activation, especially discharge home. This study aimed to investigate the prevalence, characteristics, and outcomes of patients with adverse events who required RRS activation. METHODS: Retrospective data from the In-Hospital Emergency Registry in Japan collected from April 2016 to November 2020 were eligible for our analysis. We divided patients into Home Discharge, Transfer, and Death groups. The primary outcome was the prevalence of direct discharge home, and independently associated factors were determined using multivariable logistic regression. RESULTS: We enrolled 2,043 patients who met the inclusion criteria. The prevalence of discharge home was 45.7%; 934 patients were included in the Home Discharge group. Age (adjusted odds ratio [AOR] 0.96; 95% confidence interval [CI], 0.95-0.97), malignancy (AOR 0.69; 95% CI, 0.48-0.99), oxygen administration before RRS (AOR 0.49; 95% CI, 0.36-0.66), cerebral performance category score on admission (AOR 0.38; 95% CI, 0.26-0.56), do not attempt resuscitation order before RRS (AOR 0.17; 95% CI, 0.10-0.29), RRS call for respiratory failure (AOR 0.50; 95% CI, 0.34-0.72), RRS call for stroke (AOR 0.12; 95% CI, 0.03-0.37), and intubation (AOR 0.20; 95% CI, 0.12-0.34) were independently negative, and RRS call for anaphylaxis (AOR 15.3; 95% CI, 2.72-86.3) was positively associated with discharge home. CONCLUSION: Less than half of the in-hospital patients under RRS activation could discharge home. Patients' conditions before RRS activation, disorders requiring RRS activation, and intubation were factors that affected direct discharge home.

Validation of National Early Warning Score for predicting 30-day mortality after rapid response system activation in Japan.

AIM: Although rapid response systems (RRS) are used to prevent adverse events, Japan reportedly has low activation rates and high mortality rates. The National Early Warning Score (NEWS) could provide a solution, but it has not been validated in Japan. We aimed to validate NEWS for Japanese patients. METHODS: This retrospective observational study included data of 2,255 adult patients from 33 facilities registered in the In-Hospital Emergency Registry in Japan between January 2014 and March 2018. The primary evaluated outcome was mortality rate 30 days after RRS activation. Accuracy of NEWS was analyzed with the correlation coefficient and area under the receiver operating characteristic curve. Prediction weights of NEWS parameters were then analyzed using multiple logistic regression and a machine learning method, classification and regression trees. RESULTS: The correlation coefficient of NEWS for 30-day mortality rate was 0.95 (95% confidence interval [CI], 0.88-0.98) and the area under the receiver operating characteristic curve was 0.668 (95% CI, 0.642-0.693). Sensitivity and specificity values with a cut-off score of 7 were 89.8% and 45.1%, respectively. Regarding prediction values of each parameter, oxygen saturation showed the highest odds ratio of 1.36 (95% CI, 1.25-1.48), followed by altered mental status 1.23 (95% CI, 1.14-1.32), heart rate 1.21 (95% CI, 1.09-1.34), systolic blood pressure 1.12 (95% CI, 1.04-1.22), and respiratory rate 1.03 (95% CI, 1.05-1.26). Body temperature and oxygen supplementation were not significantly associated. Classification and regression trees showed oxygen saturation as the most heavily weighted parameter, followed by altered mental status and respiratory rate. CONCLUSIONS: National Early Warning Score could stratify 30-day mortality risk following RRS activation in Japanese patients.

The characteristics, types of intervention, and outcomes of postoperative patients who required rapid response system intervention: a nationwide database analysis.

PURPOSE: Improving the safety of general wards is a key to reducing serious adverse events in the postoperative period. We investigated the characteristics, treatment, and outcomes of postoperative patients managed by a rapid response system (RRS) in Japan to improve postoperative management. METHODS: This retrospective study analyzed cases requiring RRS intervention that were included in the In-Hospital Emergency Registry in Japan. We analyzed data reported by 34 Japanese hospitals between January 2014 and March 2018, mainly focusing on postoperative patients for whom the RRS was activated within 7 days of surgery. Non-postoperative patients, for whom the RRS was activated in all other settings, were used for comparison as necessary. RESULTS: There were 609 (12.7%) postoperative patients among the total patients in the registry. The major criteria were staff concerns (30.2%) and low oxygen saturation (29.7%). Hypotension, tachycardia, and inability to contact physicians were observed as triggers significantly more frequently in postoperative patients when compared with non-postoperative patients. Among RRS activations within 7 days of surgery, 68.9% of activations occurred within postoperative day 3. The ordering of tests (46.8%) and fluid bolus (34.6%) were major interventions that were performed significantly more frequently in postoperative patients when compared with non-postoperative patients. The rate of RRS activations resulting in ICU care was 32.8%. The mortality rate at 1 month was 16.2%. CONCLUSION: Approximately, 70% of the RRS activations occurred within postoperative day 3. Circulatory problems were a more frequent cause of RRS activation in the postoperative group than in the non-postoperative group.

Efficacy of prehospital National Early Warning Score to predict outpatient disposition at an emergency department of a Japanese tertiary hospital: a retrospective study.

OBJECTIVES: The National Early Warning Score (NEWS) was originally developed to assess hospitalised patients in the UK. We examined whether the NEWS could be applied to patients transported by ambulance in Japan. DESIGN: This retrospective study assessed patients and calculated the NEWS from paramedic records. Emergency department (ED) disposition data were categorised into the following groups: discharged from the ED, admitted to the ward, admitted to the intensive care unit (ICU) or died in the ED. The predictive performance of NEWS for patient disposition was assessed using receiver operating characteristic curve analysis. Patient dispositions were compared among NEWS-based categories after adjusting for age, sex and presence of traumatic injury. SETTING: A tertiary hospital in Japan. PARTICIPANTS: Overall, 2847 patients transported by ambulance between April 2017 and March 2018 were included. RESULTS: The mean (±SD) NEWS differed significantly among patients discharged from the ED (n=1330, 3.7±2.9), admitted to the ward (n=1263, 60.3±3.8), admitted to the ICU (n=232, 9.4±4.0) and died in the ED (n=22, 110.7±2.9) (p<0.001). The prehospital NEWS C-statistics (95% CI) for admission to the ward, admission to the ICU or death in the ED; admission to the ICU or death in the ED; and death in the ED were 0.73 (0.72-0.75), 0.81 (0.78-0.83) and 0.90 (0.87-0.93), respectively. After adjusting for age, sex and trauma, the OR (95% CI) of admission to the ICU or death in the ED for the high-risk (NEWS ≥7) and medium-risk (NEWS 5-6) categories was 13.8 (8.9-21.6) and 4.2 (2.5-7.1), respectively. CONCLUSION: The findings from this Japanese tertiary hospital setting showed that prehospital NEWS could be used to identify patients at a risk of adverse outcomes. NEWS stratification was strongly correlated with patient disposition.

Independent lung ventilation with use of a double-lumen endotracheal tube for refractory hypoxemia and shock complicating severe unilateral pneumonia: A case report.

BACKGROUND: The indications for independent lung ventilation (ILV) in critical care settings have not been fully clarified, especially because extracorporeal membrane oxygenation (ECMO) is being used increasingly in cases of severe respiratory failure. CASE REPORT: A 90-year-old man presented with severe unilateral pneumonia, and despite conventional mechanical ventilation management with use of a single lumen endotracheal tube and high positive endo-expiratory pressure (PEEP), oxygenation and hemodynamics deteriorated. We then performed ILV using a double-lumen endotracheal tube (DLT) and two ventilators, each set at a different respiratory mode. With continuous administration of a neuromuscular blocking agent, the ventilator for the left lung (non-affected lung) was set to pressure-controlled ventilation (PCV) mode, whereas the ventilator for the right lung (affected lung) was set to bi-level mode, 1 breath/min, and high PEEP. ILV and the high PEEP applied to the affected lung prevented hyperinflation of the non-affected lung and increased pulmonary blood perfusion on the non-affected side. Thus, ILV immediately improved oxygenation and hemodynamics by correcting ventilation/perfusion mismatch. DISCUSSION: Although ECMO is a valid treatment option for patients with severe respiratory failure, it is highly invasive intervention. ILV performed with use of a DLT is less invasive and more useful than ECMO. Thus, ILV should be kept in mind as a treatment option, especially in cases of refractory respiratory failure and circulatory failure in which the pathophysiology of the left and right lungs differs markedly.

How to facilitate a rapid response system in Japan: a promotion course based on TeamSTEPPS.

AIM: In Japan, the number of facilities introducing a rapid response system (RRS) has been increasing. However, many institutions have had unsuccessful implementations. In order to implement RRS smoothly, a plan that meets the needs of each hospital is needed. METHODS: Rapid response system teams from each hospital, including a physician and staff in charge of medical safety, from the RRS online registry were invited to attend a workshop. The workshop aimed to develop and implement RRS. The course curriculum was based on the Team Strategies and Tools to Enhance Performance and Patient Safety (TeamSTEPPS) developed in the USA. Participating facilities were required to formulate an RRS introduction plan referring to Kotter's 8-step change model to overcome barriers in the implementation of RRS. The change in medical emergency team activations comparing the intervention and control group hospitals was compared. RESULTS: Sixteen institutions were eligible for this study. After participating in the workshop, there was a tendency toward more frequent activation of medical emergency teams in the intervention group (P = 0.075). According to a self-evaluation from each facility, there is great difficulty in overcoming the 5th step of Kotter's model (empower people to act the vision). CONCLUSION: This step-by-step evaluation clearly identified a problem with implementation and provided measures for resolution corresponding to each facility. There was a major barrier to overcome the 5th step of Kotter's model in leading change, which represents the attitude toward implementing RRS in institutions.

First report based on the online registry of a Japanese multicenter rapid response system: a descriptive study of 35 institutions in Japan.

AIM: Although the concept of a rapid response system (RRS) has been gradually accepted in Japan, detailed information on the Japanese RRS is not well known. We provide the first report of the RRS epidemiological situation based on 4 years of RRS online registry data. METHODS: This is a prospective observational study. All patients registered between January 2014 and March 2018 were eligible for this study. Data related to RRS including physiological measurements were recorded. The mortality rates after rapid response team/medical emergency team (RRT/MET) intervention and after 30 days were recorded as outcomes. RESULTS: In total, 6,784 cases were registered at 35 facilities. Cancer (23.1%) was the most common existing comorbidity. Limitation of medical treatment was identified in 12.7% of the cases. The respiratory category was most frequently activated in 41.3% of the cases. Only two institutions had received more than 15 calls per 1,000 admissions. During RRT/MET intervention, death occurred in 3.6% and transfers to intensive care units occurred in 28.2% of the cases. After 30 days, the mortality rate was significantly higher in the night than in the day shift (30.7% versus 20.4%, respectively, P < 0.01). CONCLUSIONS: We report the first epidemiological study of RRS in Japan. Japanese facilities had a very low rate of RRT/MET calls and a higher mortality rate in the night than in the day shift. Further promotion to increase the number of calls and implementation of a 24-h RRT/MET is required.

Impact of increased calls to rapid response systems on unplanned ICU admission.

BACKGROUND: Whether hospital bed number and rapid response system (RRS) call rate is associated with the clinical outcomes of patients who have RRS activations is unknown. We test a hypothesis that hospital volume and RRS call rates are associated with the clinical outcomes of patients with RRSs. METHODS: This is a retrospective chart analysis of an existing dataset associated with In-Hospital Emergency Registry in Japan. In the present study, 4818 patients in 24 hospitals from April 2014 to March 2018 were analyzed. Primary outcome variable was an unplanned intensive care unit (ICU) admission after RRS activation. RESULTS: In the primary analysis of the study using a multivariate analysis adjusting potential confounding factors, higher RRS call rate was significantly associated with decreased unplanned ICU admissions (P < 0.0001, Odds ratio [OR] 0.95, 95% confidence interval [CI] 0.92-0.98), but there was no significant association of hospital volume with unplanned ICU admissions (P = 0.44). In the secondary analysis of the study, there was a non-significant trend of increased cardiac arrest on arrival at the location of the RRS provider at large-volume hospitals (P = 0.084, OR 1.16, 95% CI 0.98-1.38). Large-volume hospitals had a significantly higher 1-month mortality rate (P = 0.0040, OR 1.10, 95% CI 1.03-1.18). CONCLUSION: Hospitals with increased RRS call rates had significantly decreased unplanned ICU admission in patients who had RRS activations. Patients who had RRS activations at large-volume hospitals had an increased 1-month mortality rate.

Increase in intra-abdominal pressure during airway suctioning-induced cough after a successful spontaneous breathing trial is associated with extubation outcome.

BACKGROUND: A patient's ability to clear secretions and protect the airway with an effective cough is an important part of the pre-extubation evaluation. An increase in intra-abdominal pressure (IAP) is important in generating the flow rate necessary for a cough. This study investigated whether an increase from baseline in IAP during a coughing episode induced by routine pre-extubation airway suctioning is associated with extubation outcome after a successful spontaneous breathing trial (SBT). METHODS: Three hundred thirty-five (335) mechanically ventilated patients who passed an SBT were enrolled. Baseline IAP and peak IAP during successive suctioning-induced coughs were measured with a fluid column connected to a Foley catheter. RESULTS: Extubation was unsuccessful in 24 patients (7.2%). Unsuccessful extubation was 3.40 times as likely for patients with a delta IAP (ΔIAP) of ≤ 30 cm H2O than for those with a ΔIAP > 30 cm H2O, after adjusting for APACHE II score (95% CI, 1.39-8.26; p = .007). CONCLUSION: ΔIAP during a coughing episode induced by routine pre-extubation airway suctioning is significantly associated with extubation outcome in patients with a successful SBT. Trial registration UMIN-CTR Clinical Trial, UMIN000017762. Registered 1 June 2015.

Wet Beriberi Associated with Hikikomori Syndrome.

Wet beriberi, characterized by high cardiac output with predominantly right-sided heart failure and lactic acidosis, is a disease caused by thiamine deficiency, and is rarely seen in modern society. However, patients with social withdrawal syndrome, also known as hikikomori syndrome, may be a new population at risk of thiamine deficiency. Hikikomori syndrome, first recognized in Japan, is becoming a worldwide issue. A 39-year-old Japanese patient was brought to our hospital, with a 3-week history of progressive shortness of breath and generalized edema. The patient had right-sided high-output heart failure, lactic acidosis, and Wernicke-Korsakoff syndrome. Because of his history of social isolation, we diagnosed hikikomori syndrome according to the Japanese government's definition, which is as follows: lifestyle centered at home; no interest or willingness to attend school or work; persistence of symptoms beyond 6 months; and exclusion of other psychiatric and developmental disorders. Considering his diagnosis of hikikomori syndrome and social isolation, we suspected malnutrition, particularly thiamine deficiency, and successfully treated him. Clinicians should be aware of the potential risk of thiamine deficiency associated with hikikomori syndrome and initiate thiamine replacement in cases of high-output heart failure associated with lactic acidosis.

Impact of physician education and availability of parameters regarding esophageal pressure and transpulmonary pressure on clinical decisions involving ventilator management.

PURPOSE: This study investigated the effects of physician education and the availability of Peso and PL data on physicians' decisions regarding ventilator management during specific simulated clinical conditions. MATERIALS AND METHODS: The study was a prospective, before-after study using a case scenario-based questionnaire and a case simulator device comprising an Avea ventilator and an artificial lung and esophagus, which was connected to a Series 1101 Electronic Breathing Simulator. The 99 physicians participating in the study were provided with five simulated cases with on-time ventilator graphics without Peso and PL and completed a questionnaire on decisions they would make regarding ventilator management of the cases. Then, after receiving instruction on Peso and PL, they were given the same cases along with ventilator graphics that included Peso and PL. RESULTS: After receiving instruction and data on Peso and PL, statistically significant numbers of physicians changed their answers regarding ventilator management decisions in all five cases. CONCLUSIONS: Providing education and data for Peso and PL had a significant effect on physician decisions regarding ventilator management in simulated cases. The use of case scenario-based education with simulator devices for physicians may hasten worldwide understanding and clinical application of Peso and PL.

What is the true definition of a "Do-Not-Resuscitate" order? A Japanese perspective.

BACKGROUND: Japan has no official guidelines for do-not-resuscitate (DNR) orders. Therefore, we investigated the effect of DNR orders on physician decision making in relation to performing noncardiopulmonary resuscitation (CPR) and CPR procedures. METHODS: A case-scenario-based questionnaire that included a case of advanced cancer, a case of advanced dementia, and a case of nonadvanced heart failure was administered to physicians. The questions determined whether physicians would perform different non-CPR procedures and CPR procedures in the presence or absence of DNR orders. The number of non-CPR procedures each physician would perform and the number of physicians who would perform each non-CPR and CPR procedure in the absence and presence of DNR ocrders were compared. Physicians from three Japanese municipal acute care hospitals participated. RESULTS: We analyzed 111 of 161 (69%) questionnaires. Physicians would perform significantly fewer non-CPR procedures in the presence of DNR orders than in the absence of DNR orders for all three case scenarios (median [interquartile range] percentages: Case 1: 72% [45%-90%] vs 100% [90%-100%]; Case 2: 55% [36%-72%] vs 91% [63%-100%]; Case 3: 78% [55%-88%] vs 100% [88%-100%]). Fewer physicians would perform non-CPR and CPR procedures in the presence of DNR orders than in the absence of DNR orders. However, considerable numbers of physicians would perform electric shock treatment for ventricular fibrillation in the presence of DNR orders (Case 1: 26%; Case 2: 16%; Case 3: 20%). CONCLUSION: DNR orders affect physician decision making about performing non-CPR procedures. Although some physicians would perform CPR for ventricular fibrillation in the presence of DNR orders, others would not. Therefore, a consensus definition for DNR orders should be developed in Japan, otherwise DNR orders may cause harm.