The coronavirus disease (COVID-19) poses an urgent threat to global public health and is characterized by rapid disease progression even in mild cases. In this study, we investigated whether machine learning can be used to predict which patients will have a deteriorated condition and require oxygenation in asymptomatic or mild cases of COVID-19.

This single-center, retrospective, observational study included COVID-19 patients admitted to the hospital from February 1, 2020, to May 31, 2020, and who were either asymptomatic or presented with mild symptoms and did not require oxygen support on admission. Data on patient characteristics and vital signs were collected upon admission. We used seven machine learning algorithms, assessed their capability to predict exacerbation, and analyzed important influencing features using the best algorithm.

In total, 210 patients were included in the study. Among them, 43 (19%) required oxygen therapy. Of all the models, the logistic regression model had the highest accuracy and precision. Logistic regression analysis showed that the model had an accuracy of 0.900, precision of 0.893, and recall of 0.605. The most important parameter for predictive capability was SpO2, followed by age, respiratory rate, and systolic blood pressure.

In this study, we developed a machine learning model that can be used as a triage tool by clinicians to detect high-risk patients and disease progression earlier. Prospective validation studies are needed to verify the application of the tool in clinical practice.

In Tohoku, the northeastern part of the main island of Japan, students entered medical school following the Great East Japan earthquake that occurred on March 11, 2011. Such students wished to volunteer at the time of disaster, however, the undergraduate medical curriculum was inadequate to enable the practice of disaster medicine. Thus, the Tohoku Disaster Medical Assistance Student (DMAS) holds workshops for undergraduate students to acquire disaster medicine knowledge.

Tohoku DMAS offers Peer Learning Education. In the DMAS course, students learned disaster medicine through lectures and simulations under the supervision of disaster medicine experts. The workshops vary in length between 3–8 hours. Tohoku DMAS’s goal is to support disaster management headquarters and shelters. Students are expected to provide logistical support that includes recounting the chronology of events at disaster management headquarters and helping with managing evacuation shelters.

According to the activity reports and roster of the course, there were only three students initially when the course was formed in 2018, however, the group continued to grow, and 165 students currently belong to the Tohoku DMAS. Those students include medical students, nursing students, and paramedics students at various universities and colleges. The DMAS has held 30 training sessions since 2018. The total number of training participants was 1,308. The DMAS has held tabletop simulation exercises and lectures on various topics such as shelter management, disaster triage, and nuclear disasters. Furthermore, some members have participated in emergency drills for each prefecture. The current challenge of the program was obtaining adequate insurance coverage for students and financial support during the activity at the disaster scene.

The DMAS plays a role in disaster medicine education for undergraduate medical students in the Tohoku region. The program continues to grow and faces opportunities and challenges.

Foreign body airway obstruction (FBAO) is a life-threatening emergency, and the prognosis of patients with FBAO is greatly affected by the prehospital process. There are only a few large-scale studies analyzing prehospital process databases of the fire department.

The aim of this study was to investigate whether characteristics of patients with FBAO were associated with prehospital factors and outcomes.

In this retrospective observational study, patients transferred to the hospital by the Tokyo, Japan Fire Department for FBAO from 2017 through 2019 were included. The association between neurologically favorable survival among the characteristics of patients with FBAO and prehospital factors affecting the outcomes was evaluated.

Of the 2,429,175 patients, 3,807 (0.2%) patients had FBAO. The highest number of FBAO cases was 99 (2.6%), which occurred on January 1 (New Year’s Day), followed by 40 cases (1.1%) on January 2, and 28 cases (0.7%) on January 3. The number of patients who experienced out-of-hospital cardiac arrest (OHCA) caused by FBAO was 1,644 (43.2%). Comparing the OHCA and non-OHCA groups, there were significant differences in age, sex, time spent at the site, and distance between the site and hospital. Cardiac arrest was significantly lower in infants after FBAO (P < .001). In total, 98.2% of patients who did not have return of spontaneous circulation (ROSC) before hospital arrival died within 30 days, a significantly higher mortality rate than that in patients who had ROSC (98.2% versus 65.8%; P < .001).

Among patients who did not have ROSC following FBAO upon arrival at the hospital, 98.2% died within 30 days. Thus, it is important to remove foreign bodies promptly and provide sufficient ventilation to the patient at the scene to increase the potential for ROSC. Further, more precautions should be exercised to prevent FBAO at the beginning of the year.

The Minimum Data Set (MDS) has allowed governments of disaster-affected countries to collect, examine, and evaluate standardized medical data from Emergency Medical Teams in real-time. However, little study has been conducted on the use of MDS data to predict health care needs.

This research proposes an outlook on the use of machine learning and MDS data to predict the need for medical care in disaster-affected areas.

The characteristics of the data collected by MDS and the optimal machine learning model were discussed.

The primary causes of disease after disasters are trauma (MDS Nos. 4–8), which frequently occurs immediately after a disaster, and infectious diseases (MDS Nos. 9–18), which can increase due to decreasing hygiene conditions. Furthermore, certain infectious diseases can spread quickly because of living in congested evacuation centers, and early detection is crucial.

Therefore, predicting the need for medical care in a disaster area is complicated and requires a combination of many machine-learning models. Data-driven methods are mostly linear approaches and cannot capture the dynamics of infectious disease transmission. Additionally, statistical models depend heavily on assumptions, making real-time infection prediction challenging. Thus, deep learning is employed to model without losing the temporal component.

Real-time prediction of health care needs using machine learning and MDS can be useful to policymakers by enabling them to better deploy and allocate health care resources, which is useful to patients and front-line health care providers. More detailed predictions for regions and diseases are also anticipated.

We conducted a systematic review to determine the prevalence and characteristics of earthquake-associated head injuries for better disaster preparedness and management.

We searched for all publications related to head injuries and earthquakes from 1985 to 2018 in MEDLINE and other major databases. A search was conducted using “earthquakes,” “wounds and injuries,” and “cranio-cerebral trauma” as a medical subject headings.

Included in the analysis were 34 articles. With regard to the commonly occurring injuries, earthquake-related head injury ranks third among patients with earthquake-related injuries. The most common trauma is lower extremity (36.2%) followed by upper extremity (19.9%), head (16.6%), spine (13.1%), chest (11.3%), and abdomen (3.8%). The most common earthquake-related head injury was laceration or contusion (59.1%), while epidural hematoma was the most common among inpatients with intracranial hemorrhage (9.5%) followed by intracerebral hematoma (7.0%), and subdural hematoma (6.8%). Mortality rate was 5.6%.

Head injuries were found to be a commonly occurring trauma along with extremity injuries. This knowledge is important for determining the demands for neurosurgery and for adequately managing patients, especially in resource-limited conditions.

Earthquakes have killed around 800,000 people globally in the past 20 years, with head injury being the main cause of mortality and morbidity.

To conduct a systematic review to determine the characteristics of head injuries after earthquakes for better disaster preparedness and management.

All publications related to head injuries and earthquakes were searched using Pubmed, Web of Science, the Cochrane Library, and Ichushi.

Thirty-six articles were included in the analysis. Head injury was the third most common cause of injury among survivors of earthquakes. The most common injury after an earthquake occurred was in the lower extremities (36.2%), followed by the upper extremities (19.9%), head (16.6%), spine (13.3%), chest (11.3%), and abdomen (3.8%). Earthquake-related head injuries were predominantly caused by a blunt strike (79%), and were more frequently associated with soft tissue injury compared to non-earthquake-related head injuries and less frequently with intracranial hemorrhage. The mean age of patients with earthquake-related head injuries was 32.6 years, and 55.1% of sufferers were male. The most common earthquake-related head injury was laceration or contusion (59.2%) while epidural hematoma was most common among inpatients with intracranial hemorrhage after an earthquake (9.5%). Early wound irrigation and debridement and antibiotics administration are needed to decrease the risk of infection. Mortality due to earthquake-related head injuries was 5.6%.

Head injury was the main cause of mortality and morbidity after an earthquake. The characteristics of earthquake-related head injuries differed from those of non-earthquake-related head injuries, including the frequency of multiple injuries, and occurrence of contaminated soft tissue injury and epidural hematoma. This knowledge is important for determining demands for neurosurgery and for adequate management of patients, especially in resource-limited conditions.