Triage assistance system

Abstract

The purpose of the present invention is to provide a system for assisting triage before transporting a head trauma patient to a hospital. More specifically, the present invention provides a system for assisting triage before transporting a head trauma patient to a hospital, the system comprising: a head trauma patient information input unit for inputting information relating to the head trauma patient; a bleeding and surgery information computation unit for computing information relating to the necessity of head surgery for the head trauma patient on the basis of the input information relating to the head trauma patient; and a bleeding and surgery information output unit for outputting information relating to the presence or absence of bleeding in the head and the information relating to the necessity of head surgery that have been computed.

Description

Triage Support System

[0001] The present invention relates to a system for assisting in the triage of head-injured patients before transporting them to a hospital.The present invention relates to a method for triaging head-injured patients before transporting them to a hospital.The present invention relates to a device for use in a system for assisting in the triage of head-injured patients before transporting them to a hospital.

[0002] Head trauma is a very common condition in neurosurgery. While most injuries are mild and result in recovery without sequelae, some can be fatal or leave behind serious sequelae. In many cases of severe head trauma, the timing of treatment is delayed due to the time required between injury and hospital admission or transfer to a facility capable of providing treatment, resulting in death or, even if surgical treatment is successful, remaining with symptoms such as severe loss of consciousness or paralysis. If a method could be established to appropriately assess the severity of head trauma patients before they are transported to a hospital and appropriately transport patients to a hospital capable of providing treatment according to their severity, treatment delays could be avoided, reducing the likelihood of patients suffering from poor prognosis.

[0003] To date, screening methods such as the Canadian CT Head Rule (Non-Patent Document 1), the New Orleans Criteria (Non-Patent Document 2), and the NICE guidelines (Non-Patent Document 3) have been used to assess the severity of head injury in patients. These are tools for assessing the risk of intracranial hemorrhage due to trauma by using information such as the patient's post-injury course and neurological examination findings, and are considered to be highly sensitive and well-established methods. However, because these methods rely on information obtained through a physician's examination, they are completely ineffective in triaging severity at the scene of injury. Established triage methods such as the Canadian CT Head Rule and the New Orleans Criteria need to be established as systems that can be used by emergency medical teams at the scene of injury before transport to a hospital.

[0004] A method for assessing the severity of head trauma patients based on their condition before transport has been reported, detecting elevated intracranial pressure by combining the state of consciousness, pupil abnormalities, and Cushing's sign (Non-Patent Document 4). However, the sensitivity was only 36.8% and the ROC-AUC was 0.65, indicating a low predictive accuracy. The inventors previously developed a machine learning model that predicts the presence or absence of traumatic intracranial hemorrhage in head trauma patients transported by ambulance using only information obtained before transport. This model demonstrated a sensitivity of 74%, specificity of 75%, and ROC-AUC of 0.80, and was reported in JAMA Network Open (Non-Patent Document 5). This study was the first to report a highly accurate model using eXtreme Gradient Boosting (XGBoost) to determine the presence or absence of intracranial hemorrhage in head trauma patients before transport. Furthermore, we confirmed that this model achieved similar accuracy in both sensitivity and specificity to predictions obtained when a physician used the NICE guidelines for the same patients.

[0005] Stiell IG, Wells GA, Vandemheen K, et al. The Canadian CT Head Rule for patients with minor head injury. Lancet. 2001;357(9266):1391-1396.Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E, DeBlieux PM. Indications for computed tomography in patients with minor head injury. N Engl J Med. 2000;343(2):100-105.Shravat BP, Huseyin TS, Hynes KA. NICE guideline for the management of head injury: an audit demonstrating its impact on a district general hospital, with a cost analysis for England and Wales. Emerg Med J. 2006;23(2): 109-113.Ter Avest E, Taylor S, Wilson M, Lyon RL. Prehospital clinical signs are a poor predictor of raised intracranial pressure following traumatic brain injury. Emerg Med J. 2021;38(1):21-26.Abe D, Inaji M, Hase T, Takahashi S, Sakai R, Ayabe F, Tanaka Y, Otomo Y, Maehara T. A prehospital triage system to detect traumatic intracranial hemorrhage using machine learning algorithms. JAMA Netw. Open. 2022;5(6):e2216393. doi: 10.1001 / jamanetworkopen.2022.16393.

[0006] However, among head trauma patients, there are many cases where intracranial hemorrhage occurs but the condition improves with observation alone, without surgery. To truly improve the life and functional prognosis of head trauma patients, it is important not only to determine whether intracranial hemorrhage due to trauma is present, but also to determine whether the condition requires emergency treatment such as surgical treatment. However, this has not been analyzed to date, and this is thought to be an issue in practical use.

[0007] SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide a system for assisting in the triage of head injury patients before transporting them to a hospital.

[0008] As mentioned above, a paper published by the inventors in JAMA Network Open in 2022 demonstrated that traumatic intracranial hemorrhage in head trauma patients could be detected with a sensitivity and specificity of approximately 75% before transport. This was the world's first report of a method for stratifying hemorrhage in head trauma patients at the prehospital stage. However, while this model could detect the presence or absence of hemorrhage, it was unable to determine the severity of the hemorrhage, particularly for stratifying severely ill patients who required surgical treatment due to time constraints from injury to treatment. Therefore, it was not truly useful for triage of transport destinations. In this invention, we have added a predictive model that not only determines the presence or absence of intracranial hemorrhage in the same prehospital setting, but also determines whether early treatment is required. We constructed a predictive model using XGBoost to stratify cases requiring surgery or equivalent intensive care based solely on prehospital information, and were able to detect critically ill patients with an accuracy of 92% sensitivity and 98% specificity using test data that was not used for training. Applying this model to actual clinical settings will enable us to select patients who truly require urgent treatment and provide the basis for selecting hospitals that can immediately provide surgical treatment as a destination, potentially improving the prognosis of patients who would otherwise die or suffer serious after-effects due to delayed treatment.

[0009] That is, the present invention provides the following. [Aspect 1] A system for supporting triage of a head-injured patient before transporting them to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about the head-injured patient regarding whether or not head surgery is required based on the input information about the head-injured patient; and a bleeding surgery information output unit for outputting the calculated information about whether or not head surgery is required. [Aspect 2] The system according to Aspect 1, in which the information about the head-injured patient is information about the head-injured patient that can be collected by an emergency team at the scene. [Aspect 3] The system according to Aspect 1, in which the information about the head-injured patient is one or more pieces of information selected from the group consisting of the head-injured patient's age, sex, vital signs, state of consciousness, neurological symptoms, high-energy trauma, and multiple trauma. [Aspect 4] The system according to Aspect 1, in which calculating information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient is performed using a trained machine learning model. [Aspect 5] The system according to Aspect 4, wherein the trained machine learning model receives information about head trauma patients as input data and information about whether head surgery is required or whether death occurred as output data. [Aspect 6] The system according to Aspect 4, wherein the trained machine learning model is created by inputting a dataset about past head trauma patients, including information about the head trauma patient and information about whether head surgery is required or whether death occurred, into the machine learning model. [Aspect 7] The system according to Aspect 4, wherein the machine learning algorithm in the trained machine learning model is eXtreme Gradient Boosting, Random Forest, Support Vector Machine, or Logistic Regression. [Aspect 8] The system according to Aspect 4, wherein the trained machine learning model is created without imputing missing values. [Aspect 9] The system according to Aspect 4, wherein learning of the trained machine learning model is performed by cross-validation.[Aspect 10] A system for supporting triage before transporting a head-injured patient to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; a transport hospital information calculation unit for calculating information about a hospital to transport the head-injured patient to based on the calculated information about whether or not head surgery is required; and a transport hospital information output unit for outputting the calculated information about a hospital to transport the head-injured patient to. [Aspect 11] A method for triaging a head-injured patient before transporting them to a hospital, the method comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; and a step of causing a bleeding surgery information output unit to output the calculated information about whether or not head surgery is required. [Aspect 12] A method for triaging a head-injured patient before transporting them to a hospital, the method comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; a step of causing a transport hospital information calculation unit to calculate information about a hospital to which the head-injured patient should be transported based on the calculated information about whether or not head surgery is required; and a step of causing a transport hospital information output unit to output the calculated information about the hospital to which the head-injured patient should be transported.[Aspect 13] A device for use in the system described in Aspect 1, including a head trauma patient information input unit for inputting information about a head trauma patient. [Aspect 14] A device for use in the system described in Aspect 1, including a bleeding surgery information calculation unit for calculating information about the head trauma patient regarding whether or not head surgery is required, based on the input information about the head trauma patient. [Aspect 15] A device for use in the system described in Aspect 1, including a bleeding surgery information output unit for outputting the calculated information about whether or not head surgery is required. [Aspect 16] A device for use in the system described in Aspect 10, including a head trauma patient information input unit for inputting information about the head trauma patient. [Aspect 17] A device for use in the system described in Aspect 10, including a bleeding surgery information calculation unit for calculating information about whether or not head surgery is required, based on the input information about the head trauma patient. [Aspect 18] An apparatus for use in the system described in Aspect 10, comprising a hospital transfer information calculation unit for calculating information regarding a hospital to which a head-injured patient should be transferred, based on calculated information regarding whether head surgery is required. [Aspect 19] An apparatus for use in the system described in Aspect 10, comprising a hospital transfer information output unit for outputting the calculated information regarding a hospital to which a head-injured patient should be transferred. [Aspect 20] A method for performing the method described in Aspect 11, comprising a step of inputting information regarding a head-injured patient into a head-injured patient information input unit for inputting information regarding the head-injured patient. [Aspect 21] A method for performing the method described in Aspect 11, comprising a step of causing a bleeding surgery information calculation unit to calculate information regarding whether head surgery is required for the head-injured patient, based on the input information regarding the head-injured patient.[Aspect 22] A method for performing the method described in Aspect 11, comprising the step of causing a bleeding surgery information output unit to output calculated information regarding the need for head surgery. [Aspect 23] A method for performing the method described in Aspect 12, comprising the step of inputting information regarding a head trauma patient into a head trauma patient information input unit for inputting information regarding a head trauma patient. [Aspect 24] A method for performing the method described in Aspect 12, comprising the step of causing a bleeding surgery information calculation unit to calculate information regarding the need for head surgery for the head trauma patient based on the input information regarding the head trauma patient. [Aspect 25] A method for performing the method described in Aspect 12, comprising the step of causing a hospital transfer information calculation unit to calculate information regarding a hospital to which the head trauma patient should be transferred based on the calculated information regarding the need for head surgery. [Aspect 26] A method for performing the method described in Aspect 12, comprising a step of causing a transport hospital information output unit, for outputting calculated information about a hospital to which a head-injured patient should be transported, to output information about a hospital to which the head-injured patient should be transported. [Aspect 27] A system for supporting triage of a head-injured patient before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient, a bleeding surgery information calculation unit for calculating, for the head-injured patient, information about the presence or absence of bleeding in the head and information about the need for head surgery based on the input information about the head-injured patient, and a bleeding surgery information output unit for outputting the calculated information about the presence or absence of bleeding in the head and information about the need for head surgery.[Aspect 28] A system for supporting triage before transporting a head-injured patient to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, information about the presence or absence of bleeding in the head of the head patient and information about whether head surgery is required; a transport hospital information calculation unit for calculating, based on the calculated information about the presence or absence of bleeding in the head and information about whether head surgery is required; and a transport hospital information output unit for outputting the calculated information about the hospital to which the head-injured patient is transported. [Aspect 29] A method for triaging a head-injured patient before transporting them to a hospital, the method comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about the head-injured patient regarding the presence or absence of bleeding in the head and information about whether head surgery is required, based on the input information about the head-injured patient, the information about the presence or absence of bleeding in the head and information about whether head surgery is required; and a step of causing a bleeding surgery information output unit to output the calculated information about the presence or absence of bleeding in the head and information about whether head surgery is required.[Aspect 30] A method for triaging a head-injured patient before transporting them to a hospital, the method comprising: a step of inputting information about the head trauma patient into a head trauma patient information input unit for inputting information about the head trauma patient; a step of causing a bleeding surgery information calculation unit to calculate information about the head trauma patient regarding the presence or absence of bleeding in the head and information about the need for head surgery based on the input information about the head trauma patient; a step of causing a transport hospital information calculation unit to calculate information about a hospital to transport the head trauma patient to based on the calculated information about the presence or absence of bleeding in the head and information about the need for head surgery; and a step of causing a transport hospital information output unit to output the calculated information about the hospital to transport the head trauma patient to. [Aspect 31] A system for supporting triage before transporting a head-injured patient to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; and a bleeding surgery information output unit for outputting the calculated information about whether or not head surgery is required, wherein the information about the head-injured patient is information about the head-injured patient that can be collected by an emergency team at the scene, and the information about the head-injured patient includes at least impaired consciousness, high-energy head injury, head trauma scar, E-GCS, and pupil abnormality, and calculating the information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient is performed using a trained machine learning model, and the trained machine learning model takes the information about the head-injured patient as input data and information about whether or not head surgery is required or whether or not the patient has died as output data.[Aspect 32] A method for triaging a head-injured patient before transporting them to a hospital, comprising: a step of inputting information about the head trauma patient into a head trauma patient information input unit for inputting information about the head trauma patient; a step of causing a surgery information calculation unit to calculate information about whether or not head surgery is required for the head trauma patient based on the input information about the head trauma patient; and a step of causing a bleeding surgery information output unit to output the calculated information about whether or not head surgery is required, wherein the information about the head trauma patient is information about the head trauma patient that can be collected by an emergency team at the scene, and the information about the head trauma patient includes at least impaired consciousness, high-energy head injury, head trauma scar, E-GCS, and pupil abnormality, and calculating information about whether or not head surgery is required for the head trauma patient based on the input information about the head trauma patient is performed using a trained machine learning model, A method in which a trained machine learning model takes information about head injury patients as input data and outputs information about whether head surgery is required or whether the patient has died.[Aspect 33] A method for triaging a head-injured patient before transporting the patient to a hospital, comprising: a step of inputting information about the head trauma patient into a head trauma patient information input unit for inputting information about the head trauma patient; a step of causing a bleeding surgery information calculation unit to calculate information about whether or not head surgery is required for the head trauma patient based on the input information about the head trauma patient; a step of causing a transport hospital information calculation unit to calculate information about a hospital to which the head trauma patient should be transported based on the calculated information about whether or not head surgery is required; and a step of causing a transport hospital information output unit to output the calculated information about the hospital to which the head trauma patient should be transported, wherein the information about the head trauma patient is information about the head trauma patient that can be collected by an ambulance at the scene, A method in which information about head trauma patients includes at least impaired consciousness, high-energy head trauma, head trauma scarring, E-GCS, and pupil abnormalities, and information regarding whether head surgery is required for the head trauma patient is calculated based on the input information about the head trauma patient using a trained machine learning model, wherein the trained machine learning model uses the information about the head trauma patient as input data and information regarding whether head surgery is required or whether the patient has died as output data, and the calculation of information regarding hospitals to which head trauma patients should be transported is performed by referring to a list that associates information regarding whether head surgery is required with information regarding hospitals to which head trauma patients should be transported.

[0010] According to the Fire and Disaster Management Agency database, there were approximately 6.19 million ambulance dispatches in fiscal year 2021, an increase of approximately 260,000 from the previous year. Breaking this down, 1.32 million trauma patients, including both general injuries and traffic injuries, accounted for 1.32 million. Analysis of the 2022 Trauma Data Bank revealed that head injuries accounted for 33% of all trauma cases. Combined with the aforementioned transport data, this estimate of approximately 440,000 emergency transports due to head trauma per year. Considering that transfers accounted for 8.4% of all emergency transports, a simple calculation suggests that the number of hospital transfers due to head trauma is approximately 35,000. The average cost of one ambulance dispatch is estimated at 45,000 yen. Simply reducing the number of hospital transfers for head trauma patients could potentially reduce medical costs by approximately 1.58 billion yen.

[0011] Furthermore, we consider the improvement of prognosis for head trauma patients. According to a database published by the Tokyo Fire Department, 85% of the reasons for transfer were "intractable." This figure does not only include cases requiring surgical treatment, but also includes cases undergoing follow-up observation at facilities with specialists. Based on the breakdown of 2,000 head trauma patients treated at the applicant's affiliated hospital, including 196 cases of traumatic intracranial hemorrhage and 59 surgical cases, we estimate that surgical treatment accounts for approximately 30% of head trauma emergency transfers. Given that the applicant's affiliated hospital is a tertiary emergency facility and the proportion of severely injured patients is high, we estimate that approximately 15-20% require surgical treatment. Therefore, assuming that approximately 15-20% of transfers are for surgical purposes, this translates into an estimated 70,000-90,000 cases per year. As mentioned above, surgical treatment for head injuries is a race against time, and if treatment is delayed, the life and functional prognosis will be significantly impaired even if surgical treatment is performed. By appropriately stratifying the severity of head injury patients before transport and transporting severely injured patients who require surgical treatment to facilities that can provide appropriate treatment early, it is expected that the prognosis of the 70,000 to 90,000 patients who require transfer to other hospitals due to severe head injuries will improve each year.

[0012] FIG. 1 shows an exemplary schematic configuration of a system of the present invention. An exemplary computer (100) includes a control unit (101), a memory unit (102), a peripheral device I / F unit (103), an input unit (104), a display unit (105), a communication unit (106), and a bus (110). The computer (100) is connected to an external server (130) and a database (140) via a network (120). FIG. 2 shows an exemplary configuration of a system of the present invention. The computer (100a) includes a head trauma patient information input unit (210) and a bleeding surgery information output unit (230). The computer (100a) is connected to the external server (130) via the network (120). The external server (130) includes a bleeding surgery information calculation unit (220). FIG. 3 shows an exemplary configuration of a system of the present invention. The computer (100b) includes a head trauma patient information input unit (210). The computer (100b) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220). The external server (130) is connected to the computer (100c) via the network (120). The computer (100c) includes a bleeding surgery information output unit (230). FIG. 4 shows a schematic configuration of an exemplary system of the present invention. The computer (100d) includes a head injury patient information input unit (210) and a bleeding surgery information output unit (230). The computer (100d) is connected to the computer (100e) via a wired or wireless network. The computer (100e) includes the bleeding surgery information calculation unit (220). FIG. 5 shows a schematic configuration of an exemplary system of the present invention. The computer (100f) includes the head injury patient information input unit (210). The computer (100f) is connected to the computer (100g) via a wired or wireless network. The computer (100g) includes a bleeding surgery information calculation unit (220). The computer (100g) is connected to the computer (100h) via a wired or wireless network. The computer (100h) includes a bleeding surgery information output unit (230). Figure 6 shows a schematic configuration of an exemplary system of the present invention.The computer (100i) includes a head trauma patient information input unit (210), a bleeding surgery information calculation unit (220), and a bleeding surgery information output unit (230). FIG. 7 shows a schematic configuration of an exemplary system of the present invention. The computer (100j) includes a head trauma patient information input unit (210) and a transfer hospital information output unit (250). The computer (100j) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). FIG. 8 shows a schematic configuration of an exemplary system of the present invention. The computer (100k) includes a head trauma patient information input unit (210), a transfer hospital information calculation unit (240), and a transfer hospital information output unit (250). The computer (100k) is connected to the external server (130) via the network (120). The external server (130) includes a bleeding surgery information calculation unit (220). FIG. 9 shows a schematic configuration of an exemplary system of the present invention. A computer (100l) includes a head trauma patient information input unit (210). The computer (100l) is connected to the external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). The external server (130) is connected to a computer (100m) via the network (120). The computer (100m) includes a transfer hospital information output unit (250). FIG. 10 shows a schematic configuration of an exemplary system of the present invention. A computer (100n) includes the head trauma patient information input unit (210) and the transfer hospital information output unit (250). The computer (100n) is connected to the computer (100o) via a wired or wireless network. The computer (100o) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). Figure 11 shows a schematic configuration of an exemplary system of the present invention. The computer (100p) includes a head trauma patient information input unit (210), a transfer hospital information calculation unit (240), and a transfer hospital information output unit (250). The computer (100p) is connected to the computer (100q) via a wired or wireless network. The computer (100q) includes the bleeding surgery information calculation unit (220).FIG. 12 shows a schematic configuration of an exemplary system of the present invention. A computer (100r) includes a head injury patient information input unit (210). The computer (100r) is connected to a computer (100s) via a wired or wireless network. The computer (100s) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). The computer (100s) is connected to a computer (100t) via a wired or wireless network. The computer (100t) includes a transfer hospital information output unit (250). FIG. 13 shows a schematic configuration of an exemplary system of the present invention. The computer (100u) includes the head injury patient information input unit (210). The computer (100u) is connected to a computer (100v) via a wired or wireless network. The computer (100v) includes the bleeding surgery information calculation unit (220). The computer (100v) is connected to the computer (100w) via a wired or wireless network. The computer (100w) includes a hospital information calculation unit (240) and a hospital information output unit (250). Figure 14 shows a schematic configuration of an exemplary system of the present invention. The computer (100x) includes a head trauma patient information input unit (210), a bleeding surgery information calculation unit (220), a hospital information calculation unit (240), and a hospital information output unit (250). Figure 15 is a schematic diagram of an exemplary system consisting of terminals (left and right) that serve as both the head trauma patient information input unit and the bleeding surgery information output unit or the hospital information output unit, and a server corresponding to the bleeding surgery information calculation unit or the hospital information calculation unit (center). Figure 16 is a graph showing the ROC curve of an intracranial hemorrhage prediction model trained using data obtained from a single institution (Tokyo Medical and Dental University). Fig. 17 is a graph showing the PR curve of an intracranial hemorrhage prediction model trained using data obtained at a single institution (Tokyo Medical and Dental University). Fig. 18 is a graph showing the ROC curve when the trained intracranial hemorrhage prediction model is applied to data obtained at multiple institutions to make predictions. Fig. 19 is a graph showing the PR curve when the trained intracranial hemorrhage prediction model is applied to data obtained at multiple institutions to make predictions.Fig. 20 is a graph showing the ROC curve of a prediction model for cases requiring surgery that was trained using data obtained at a single facility (Tokyo Medical and Dental University). Fig. 21 is a graph showing the PR curve of a prediction model for cases requiring surgery that was trained using data obtained at a single facility (Tokyo Medical and Dental University). Fig. 22 is a graph showing the ROC curve when predictions are made by applying the trained prediction model for whether or not surgery is required to data obtained at multiple facilities. Fig. 23 is a graph showing the PR curve when predictions are made by applying the trained prediction model for whether or not surgery is required to data obtained at multiple facilities.

[0013] The present invention provides a system for assisting triage of head-injured patients before transporting them to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about the head-injured patient regarding the presence or absence of bleeding in the head and / or information regarding the need for head surgery based on the input information about the head-injured patient; and a bleeding surgery information output unit for outputting the calculated information about the presence or absence of bleeding in the head and / or information regarding the need for head surgery.

[0014] In a system to support the triage of head injury patients before they are transported to a hospital, triage means, for example, determining priorities based on the urgency and severity of the injury (whether or not surgery is required) in a situation where there are many injured or ill people.

[0015] In the head injury patient information input section, the information about the head injury patient is preferably information about the head injury patient that can be collected by an ambulance at the scene.

[0016] In the head trauma patient information input unit, the information about the head trauma patient is preferably one or more pieces of information selected from the group consisting of the head trauma patient's age, gender, vital signs, state of consciousness, neurological symptoms, high-energy trauma, and multiple trauma. For example, the information about the head trauma patient may preferably be impaired consciousness, high-energy head trauma, head trauma scarring, E-GCS, and pupil abnormalities. Therefore, the information about the head trauma patient may include at least one, two, three, four, or five of impaired consciousness, high-energy head trauma, head trauma scarring, E-GCS, and pupil abnormalities. Furthermore, the information about the head trauma patient is preferably one, two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, or eighteen pieces of information selected from the following eighteen pieces of information: 1. Patient age 2. Gender 3. Systolic blood pressure 4. Heart rate 5. Body temperature 6. Respiratory rate 7. State of consciousness-1: Ocular component of the Glasgow Coma Scale [E-GCS] 8. State of consciousness-2: Presence or absence of disorientation 9. Pupillary abnormalities (defined as ectopic or bilateral pupillary dilation and loss of light reflexes) 10. High-energy head injury (head injury with a dangerous mechanism as defined by the National Institute for Health and Care Excellence [NICE] guidelines) 11. Head trauma scarring 12. Multiple trauma (combination of head injury with other trauma to the body) 13. Post-traumatic seizures 14. Loss of consciousness 15. Vomiting 16. Alcohol use 17. Hemiplegia 18. Clinical deterioration known as "talk and deteriorate"

[0017] At emergency scenes, some information may not be collected, but the predictive model disclosed herein is highly robust, making it possible to make predictions even when information is missing.

[0018] In the head injury patient information input unit, information about head injury patients includes, for example, information expressed in numerical values, information expressed in symbols, etc. Information expressed in numerical values ​​includes, for example, information expressed in continuous values, information expressed in discrete values, etc. Information expressed in discrete values ​​includes, for example, information obtained by discretizing information expressed in continuous values.

[0019] In the bleeding surgery information calculation unit, information regarding the presence or absence of bleeding in the head of a head trauma patient may be, for example, information expressed in numerical values ​​or information expressed in symbols.

[0020] Here, examples of information expressed as a numerical value include information expressed as a numerical value indicating the possibility of head bleeding in a head trauma patient. The numerical value indicating the possibility of head bleeding in a head trauma patient may be, for example, but is not limited to, an integer or real number from 0 to 1, an integer or real number from 0 to 2, an integer or real number from 0 to 10, or an integer or real number from 0 to 100.

[0021] Here, examples of information represented by symbols include symbols that classify the likelihood of head bleeding in a head trauma patient. Examples of symbols that classify the likelihood of head bleeding in a head trauma patient include symbols that are assigned based on whether a numerical value indicating the likelihood of head bleeding in a head trauma patient falls within a predetermined range. Examples of predetermined ranges of numerical values ​​indicating the likelihood of head bleeding in a head trauma patient include 50 or more, 60 or more, 70 or more, 80 or more, and 90 or more.

[0022] In the bleeding surgery information calculation unit, information regarding the necessity of head surgery for a head injury patient may be, for example, information expressed in numerical values ​​or information expressed in symbols.

[0023] Here, examples of information expressed as a numerical value include information expressed as a numerical value indicating the possibility that a head injury patient will require head surgery. The numerical value indicating the possibility that a head injury patient will require head surgery may be, for example, but is not limited to, an integer or real number from 0 to 1, an integer or real number from 0 to 2, an integer or real number from 0 to 10, or an integer or real number from 0 to 100.

[0024] Here, examples of information represented by symbols include symbols that classify the likelihood that a head-injured patient will require head surgery. Examples of symbols that classify the likelihood that a head-injured patient will require head surgery include symbols that are assigned based on whether a numerical value indicating the likelihood that a head-injured patient will require head surgery falls within a predetermined range. Examples of predetermined ranges of numerical values ​​indicating the likelihood that a head-injured patient will require head surgery include 50 or more, 60 or more, 70 or more, 80 or more, and 90 or more.

[0025] The bleeding surgery information calculation unit can calculate (or predict) one or more pieces of information selected from the group consisting of information on the presence or absence of bleeding in the head, information on the need for head surgery, information on the severity of bleeding, information on the need for early treatment, information on the need for intracranial pressure monitoring, and information on the risk of death for a head trauma patient. Furthermore, the bleeding surgery information calculation unit can calculate any combination of two or more, three or more, four or more, or five or more pieces of information selected from the group consisting of information on the presence or absence of bleeding in the head, information on the need for head surgery, information on the severity of bleeding, information on the need for early treatment, information on the need for intracranial pressure monitoring, and information on the risk of death for a head trauma patient. Furthermore, the bleeding surgery information calculation unit can calculate information on the presence or absence of bleeding in the head, information on the need for head surgery, information on the severity of bleeding, information on the need for early treatment, information on the need for intracranial pressure monitoring, and information on the risk of death for a head trauma patient.

[0026] Examples of cranial surgery include, but are not limited to, burr hole hematoma evacuation, craniotomy, decompressive craniectomy, ventricular drainage, and intracranial pressure sensor placement. In the context of this disclosure, the need for surgery may more specifically refer to the need for surgery or neurointensive care. Even in severely injured patients, surgery may not be indicated, and in such cases, an intracranial pressure sensor may be placed in the brain.

[0027] The bleeding surgery information calculation unit performs the calculation (or prediction) using, for example, a prediction model, such as a machine learning model, a statistical model, a simulation model, a rule-based model, a time series model, or a symbolic regression model.

[0028] In the bleeding surgery information calculation unit, the calculation is preferably performed using a trained machine learning model.

[0029] Here, the trained machine learning model preferably uses information about head trauma patients as input data. The trained machine learning model also preferably uses information about the presence or absence of bleeding in the head, the need for head surgery, or the presence or absence of death as output data. The trained machine learning model more preferably uses information about the presence or absence of bleeding in the head, the need for head surgery, and the presence or absence of death as output data.

[0030] Furthermore, here, the trained machine learning model is preferably created by inputting into the machine learning model a dataset about past head trauma patients, including information about the head trauma patient and information about whether or not there is bleeding in the head, whether or not head surgery is required, or whether or not there is death. More preferably, the trained machine learning model is created by inputting into the machine learning model a dataset about past head trauma patients, including information about the head trauma patient and information about whether or not there is bleeding in the head, whether or not there is head surgery required, and whether or not there is death.

[0031] Examples of the number of datasets for past head trauma patients include 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 200 or more, 300 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more, 2000 or more, 3000 or more, 4000 or more, 5000 or more, 6000 or more, 7000 or more, 8000 or more, 9000 or more, 10,000 or more, or 100,000 or more.

[0032] The number of types of explanatory variables (or features) included in the dataset of past head trauma patients can be, for example, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 200 or more, 300 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more, 2000 or more, 3000 or more, 4000 or more, 5000 or more, 6000 or more, 7000 or more, 8000 or more, 9000 or more, 10000 or more, or 100,000 or more, and examples thereof include values ​​of 1,000,000 or less, 100,000 or less, 10,000 or less, 9,000 or less, 8,000 or less, 7,000 or less, 6,000 or less, 5,000 or less, 4,000 or less, 3,000 or less, 2,000 or less, 1,000 or less, 900 or less, 800 or less, 700 or less, 600 or less, 500 or less, 400 or less, 300 or less, 200 or less, 100 or less, 90 or less, 80 or less, 70 or less, 60 or less, 50 or less, 40 or less, 30 or less, or 20 or less.

[0033] The above dataset is input into a machine learning model to create a trained machine learning model, for example, by the following steps: 1. Data collection and preprocessing: First, data appropriate to the problem is collected. This data is used to train the model. After collecting the data, preprocessing is performed, such as removing unnecessary parts, processing missing values, and normalizing the data. 2. Data division: The data is divided into three datasets: a training dataset, a validation dataset, and a test dataset. The training dataset is used to train the model, and the validation dataset is used to adjust the model parameters and hyperparameter tuning. The test dataset is used to evaluate the final model. 3. Model selection: An appropriate model is selected depending on the nature of the problem. For example, logistic regression, random forests, and neural networks are commonly used for classification problems. 4. Model construction: The selected model is implemented and fitted to the training dataset. In this process, the model parameters are adjusted to fit the data. 5. Model training: The model is trained on the training dataset. In this process, the model learns patterns and associations in the data. An optimization algorithm such as gradient descent is used for training. 6. Model evaluation: The model is evaluated on the validation dataset to evaluate its performance. Evaluation metrics include precision, recall, precision, F1 score, etc. 7. Model tuning: Adjust the model's hyperparameters and structure to improve performance on the validation dataset. This is expected to result in the model making better predictions. 8. Final evaluation: Evaluate the final model on the test dataset to confirm its performance. This step confirms how well the model can generalize to unknown data. 9. Deployment: Finally, the trained model is deployed in a production environment, ready to make predictions and classifications on new data.

[0034] The machine learning model may be selected from the group consisting of, for example, a decision tree, a random forest, a LightGBM, stacking (e.g., LR, RF, LGBM stacking), logistic regression, lasso regression, a support vector machine, a multilayer perceptron, a neural network, and combinations thereof.

[0035] The machine learning algorithm in the trained machine learning model is preferably eXtreme Gradient Boosting, Random Forest, Support Vector Machine, or Logistic Regression. The trained machine learning model is preferably created without imputing missing values. Training of the trained machine learning model is preferably performed by cross-validation. Cross-validation is performed, for example, once, twice, three times, four times, five times, six times, seven times, eight times, nine times, ten times, eleven times, twelve times, thirteen times, fourteen times, fifteen times, sixteen times, seventeen times, eighteen times, nineteen times, or twenty times. In the cross-validation process, for example, a grid search is performed to obtain optimal hyperparameters. To solve problems related to imbalanced datasets, for example, class weighting is introduced to modify the loss function. For example, positive cases (cases with tICH (traumatic intracranial hemorrhage)) are assigned a weight that is 2-100 times, 3-50 times, 4-40 times, or 5-15 times higher than negative cases (cases without tICH).

[0036] Furthermore, the present invention provides a system for assisting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, one or more pieces of information selected from the group consisting of information regarding the presence or absence of bleeding in the head, information regarding the need for head surgery, information regarding the severity of the bleeding, information regarding the need for early treatment, information regarding the need for intracranial pressure monitoring, and information regarding the risk of death; and a bleeding surgery information output unit for outputting the calculated information.

[0037] Furthermore, the present invention provides a system for assisting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, two or more, three or more, four or more, or five or more pieces of information about the head-injured patient selected from the group consisting of information regarding the presence or absence of bleeding in the head, information regarding the need for head surgery, information regarding the severity of the bleeding, information regarding the need for early treatment, information regarding the need for intracranial pressure monitoring, and information regarding the risk of death; and a bleeding surgery information output unit for outputting the calculated information.

[0038] Furthermore, the present invention provides a system for assisting triage of head-injured patients before transporting them to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, information regarding the presence or absence of bleeding in the head, information regarding the need for head surgery, information regarding the severity of the bleeding, information regarding the need for early treatment, information regarding the need for intracranial pressure monitoring, and information regarding the risk of death; and a bleeding surgery information output unit for outputting the calculated information.

[0039] The present invention provides a system for assisting triage of head-injured patients before they are transported to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about the head-injured patient regarding the presence or absence of bleeding in the head and / or information regarding the need for head surgery based on the input information about the head-injured patient; a transport hospital information calculation unit for calculating information about a hospital to transport the head-injured patient to based on the calculated information about the presence or absence of bleeding in the head and / or information regarding the need for head surgery; and a transport hospital information output unit for outputting the calculated information about a hospital to transport the head-injured patient to.

[0040] The present invention further provides a system for assisting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, one or more pieces of information selected from the group consisting of information regarding the presence or absence of bleeding in the head, information regarding the need for head surgery, information regarding the severity of the bleeding, information regarding the need for early treatment, information regarding the need for intracranial pressure monitoring, and information regarding the risk of death; a transport hospital information calculation unit for calculating, based on the calculated information, information regarding a hospital to which the head-injured patient should be transported; and a transport hospital information output unit for outputting the calculated information regarding a hospital to which the head-injured patient should be transported.

[0041] The present invention further provides a system for assisting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, two or more, three or more, four or more, or five or more pieces of information about the head-injured patient selected from the group consisting of information regarding the presence or absence of bleeding in the head, information regarding the need for head surgery, information regarding the severity of the bleeding, information regarding the need for early treatment, information regarding the need for intracranial pressure monitoring, and information regarding the risk of death; a transport hospital information calculation unit for calculating, based on the calculated information, information regarding a hospital to transport the head-injured patient; and a transport hospital information output unit for outputting the calculated information about a hospital to transport the head-injured patient.

[0042] The present invention further provides a system for assisting triage of head-injured patients before transporting them to a hospital, the system including: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating, based on the input information about the head-injured patient, information about the presence or absence of bleeding in the head, information about the need for head surgery, information about the severity of the bleeding, information about the need for early treatment, information about the need for intracranial pressure monitoring, and information about the risk of death for the head-injured patient; a transport hospital information calculation unit for calculating, based on the calculated information, information about hospitals to transport the head-injured patient; and a transport hospital information output unit for outputting the calculated information about hospitals to transport the head-injured patient.

[0043] In the hospital transfer information calculation unit, information relating to the hospital to which the head-injured patient is transferred includes, for example, information for identifying the hospital to which the head-injured patient is transferred, such as the name, address, telephone number, expected arrival time, name of the doctor, working status of the doctor, bed availability status, operating status of surgical equipment, etc. of the hospital to which the head-injured patient is transferred.

[0044] The hospital transfer information calculation unit performs the calculation by referring to, for example, a list associating information regarding the presence or absence of bleeding in the head with information regarding hospitals to which head-injured patients should be transferred, or a list associating information regarding the need for head surgery with information regarding hospitals to which head-injured patients should be transferred. That is, for example, using a list associating a certain numerical value or greater possibility of bleeding in the head with a specific hospital name, information indicating the name of the hospital is output based on information indicating that the certain numerical value or greater possibility of bleeding in the head is present. As a result, for example, if the risk of bleeding is low, information indicating the name of a hospital without a neurosurgeon can be output; if bleeding is present but surgery is not required, information indicating the name of a hospital with a neurosurgeon but that does not perform emergency surgery can be output; and if bleeding is present and surgery is required, information indicating the name of a hospital with multiple neurosurgeons that can perform emergency surgery can be output.

[0045] The following describes in more detail the embodiments of the present invention. The following embodiments are preferred examples of the present invention, and the present invention is not limited to these examples.

[0046] Fig. 1 is a diagram showing an exemplary schematic configuration of a system of the present invention. In Fig. 1, reference numeral 100 denotes a computer, which includes a control unit (101), a storage unit (102), a peripheral device I / F unit (103), an input unit (104), a display unit (105), and a communication unit (106), all of which are connected by a bus (110). Note that this configuration is merely an example, and various other configurations can be adopted as appropriate.

[0047] The control unit (101) is composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The CPU loads programs stored in the memory unit (102), ROM, recording medium, etc. into a work memory area on the RAM and executes them, driving and controlling each device connected via the bus (110) to realize the processing performed by the computer. The ROM is a non-volatile memory that stores programs, data, etc. The RAM is a volatile memory that temporarily stores programs, data, etc. loaded from the memory unit (102), ROM, recording medium, etc., and also has a work area used by the control unit (101) when performing various processing. The memory unit (102) is, for example, a hard disk drive (HDD) that stores the programs executed by the control unit (101) and various other data.

[0048] The peripheral device I / F (interface) unit (103) is a port for connecting the computer (100) to peripheral devices. The peripheral device I / F unit (103) is configured using a USB, IEEE 1394, RS-232C, etc. The connection to the peripheral device may be wired or wireless. The input unit (104) has input devices such as a keyboard, a pointing device such as a mouse, and a numeric keypad, and issues operational instructions, operation instructions, data input, etc. to the computer (100). The display unit (105) is a logic circuit or device driver for displaying videos, images, etc. on a display device such as a liquid crystal panel. The input unit (104) and the display unit (105) can also be configured integrally as a touch display.

[0049] The communication unit (106) has a communication control device, a communication port, etc., and is a wired or wireless communication interface that mediates communication with the network (120). The bus (110) is a communication path that mediates the exchange of control signals, data signals, etc. between each device. The network (120) can further be connected to an external server (130) or a database (or net storage) (140).

[0050] FIG. 2 is a diagram showing the schematic configuration of an exemplary system of the present invention. The computer (100a) includes a head trauma patient information input unit (210) and a bleeding surgery information output unit (230). The computer (100a) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The bleeding surgery information output unit (230) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The bleeding surgery information output unit (230) is, for example, composed of the display unit (105). The computer (100a) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220). The external server (130) includes, for example, a processor such as a central processing unit (CPU) and a control unit including memories such as ROM and RAM. The external server (130) may be a cloud server distributed over a network. Information stored on the external server (130) is updated, for example, daily, weekly, or monthly. The memory unit of the external server (130) may be, for example, a storage device such as a hard disk, SSD, or flash memory. The memory unit of the external server (130) may be, for example, cloud storage. The bleeding surgery information calculation unit (220) is composed of, for example, a control unit and a memory unit of the external server (130). Note that FIG. 2 illustrates a configuration in which the computer (100a) is connected to the external server (130) via a network, but the system may be configured so that the computer (100a) can operate offline by itself.

[0051] FIG. 3 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100b) includes a head trauma patient information input unit (210). The computer (100b) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The computer (100b) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220). The external server (130) includes, for example, a processor such as a central processing unit (CPU) and a control unit including memories such as ROM and RAM. The external server (130) may also be a cloud server distributed on a network. The information stored on the external server (130) is updated, for example, daily, weekly, or monthly. The memory unit of the external server (130) may be, for example, a storage device such as a hard disk, SSD, or flash memory. The memory unit of the external server (130) may be, for example, cloud storage. The bleeding surgery information calculation unit (220) is, for example, composed of a control unit and a memory unit of the external server (130). The external server (130) is connected to a computer (100c) via a network (120). The computer (100c) includes a bleeding surgery information output unit (230). The computer (100c) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The bleeding surgery information output unit (230) is, for example, composed of a communication unit (106), a control unit (101), and a display unit (105). The bleeding surgery information output unit (230) is configured by, for example, the display unit (105).

[0052] FIG. 4 is a diagram showing the schematic configuration of an exemplary system of the present invention. The computer (100d) includes a head trauma patient information input unit (210) and a bleeding surgery information output unit (230). The computer (100d) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The bleeding surgery information output unit (230) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The bleeding surgery information output unit (230) is, for example, composed of the display unit (105). The computer (100d) is connected to the computer (100e) via a wired or wireless network. The computer (100e) includes a bleeding surgery information calculation unit (220). The computer (100e) is a device such as a smartphone, tablet, smartwatch, notebook PC, or desktop PC. The bleeding surgery information calculation unit (220) is composed of, for example, a control unit (101) and a memory unit (102).

[0053] FIG. 5 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100f) includes a head trauma patient information input unit (210). The computer (100f) is a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is configured, for example, by an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is also configured, for example, by the input unit (104). The computer (100f) is connected to the computer (100g) via a wired or wireless network. The computer (100g) includes a bleeding surgery information calculation unit (220). The computer (100g) is a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The bleeding surgery information calculation unit (220) is configured, for example, by a control unit (101) and a memory unit (102). The computer (100g) is connected to the computer (100h) via a wired or wireless network. The computer (100h) includes a bleeding surgery information output unit (230). The computer (100h) is, for example, a device such as a smartphone, tablet, smartwatch, notebook PC, or desktop PC. The bleeding surgery information output unit (230) is, for example, composed of a communication unit (106), a control unit (101), and a display unit (105). The bleeding surgery information output unit (230) is, for example, composed of the display unit (105).

[0054] FIG. 6 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100i) includes a head trauma patient information input unit (210), a bleeding surgery information calculation unit (220), and a bleeding surgery information output unit (230). The computer (100i) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The bleeding surgery information calculation unit (220) is, for example, composed of the control unit (101) and a memory unit (102). The bleeding surgery information output unit (230) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The bleeding surgery information output unit (230) is configured by, for example, the display unit (105).

[0055] FIG. 7 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100j) includes a head trauma patient information input unit (210) and a transfer hospital information output unit (250). The computer (100j) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The transfer hospital information output unit (250) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The transfer hospital information output unit (250) is, for example, composed of the display unit (105). The computer (100j) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). The external server (130) includes, for example, a processor such as a central processing unit (CPU) and a control unit including memories such as ROM and RAM. The external server (130) may be a cloud server distributed over a network. Information stored on the external server (130) is updated, for example, daily, weekly, or monthly. The memory unit of the external server (130) may be, for example, a storage device such as a hard disk, SSD, or flash memory. The memory unit of the external server (130) may be, for example, cloud storage. The bleeding surgery information calculation unit (220) is, for example, composed of a control unit and a memory unit of the external server (130). The transfer hospital information calculation unit (240) is, for example, composed of a control unit and a memory unit of the external server (130).

[0056] FIG. 8 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100k) includes a head trauma patient information input unit (210), a hospital destination information calculation unit (240), and a hospital destination information output unit (250). The computer (100k) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The hospital destination information calculation unit (240) is, for example, composed of the control unit (101) and a memory unit (102). The hospital destination information output unit (250) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The transfer hospital information output unit (250) is configured, for example, by a display unit (105). The computer (100k) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220). The external server (130) includes, for example, a processor such as a central processing unit (CPU) and a control unit including memories such as ROM and RAM. The external server (130) may be a cloud server distributed on the network. The information stored on the external server (130) is updated, for example, daily, weekly, or monthly. The memory unit of the external server (130) may be, for example, a storage device such as a hard disk, SSD, or flash memory. The memory unit of the external server (130) may be, for example, cloud storage. The bleeding surgery information calculation unit (220) is configured, for example, by the control unit and memory unit of the external server (130).

[0057] FIG. 9 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (1001) includes a head trauma patient information input unit (210). The computer (1001) is a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is composed of, for example, an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is also composed of, for example, the input unit (104). The computer (1001) is connected to an external server (130) via a network (120). The external server (130) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). The external server (130) includes, for example, a processor such as a central processing unit (CPU) and a control unit including memories such as ROM and RAM. The external server (130) may also be a cloud server distributed over a network. The information stored on the external server (130) is updated, for example, daily, weekly, or monthly. The memory unit of the external server (130) may be a storage device such as a hard disk, SSD, or flash memory. The memory unit of the external server (130) may be, for example, cloud storage. The bleeding surgery information calculation unit (220) is, for example, composed of a control unit and a memory unit of the external server (130). The transfer hospital information calculation unit (240) is, for example, composed of a control unit and a memory unit of the external server (130). The external server (130) is connected to a computer (100m) via a network (120). The computer (100m) includes a transfer hospital information output unit (250). The computer (100m) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The transfer hospital information output unit (250) is configured, for example, by the communication unit (106), the control unit (101), and the display unit (105). The transfer hospital information output unit (250) is also configured, for example, by the display unit (105).

[0058] FIG. 10 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100n) includes a head trauma patient information input unit (210) and a transfer hospital information output unit (250). The computer (100n) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The transfer hospital information output unit (250) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The transfer hospital information output unit (250) is, for example, composed of the display unit (105). The computer (100n) is connected to the computer (100o) via a wired or wireless network. The computer (100o) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). The computer (100o) is, for example, a device such as a smartphone, tablet, smartwatch, notebook PC, or desktop PC. The bleeding surgery information calculation unit (220) is, for example, composed of a control unit (101) and a memory unit (102). The transfer hospital information calculation unit (240) is, for example, composed of the control unit (101) and the memory unit (102).

[0059] FIG. 11 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100p) includes a head trauma patient information input unit (210), a hospital destination information calculation unit (240), and a hospital destination information output unit (250). The computer (100p) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The hospital destination information calculation unit (240) is, for example, composed of the control unit (101) and a memory unit (102). The hospital destination information output unit (250) is, for example, composed of the communication unit (106), the control unit (101), and a display unit (105). The transfer hospital information output unit (250) is configured, for example, by a display unit (105). The computer (100p) is connected to the computer (100q) via a wired or wireless network. The computer (100q) includes a bleeding surgery information calculation unit (220). The computer (100q) is a device such as a smartphone, tablet, smartwatch, notebook PC, or desktop PC. The bleeding surgery information calculation unit (220) is configured, for example, by a control unit (101) and a memory unit (102).

[0060] FIG. 12 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100r) includes a head trauma patient information input unit (210). The computer (100r) is a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is composed of, for example, an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is also composed of, for example, the input unit (104). The computer (100r) is connected to the computer (100s) via a wired or wireless network. The computer (100s) includes a bleeding surgery information calculation unit (220) and a transfer hospital information calculation unit (240). The computer (100s) is a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The bleeding surgery information calculation unit (220) is composed of, for example, a control unit (101) and a memory unit (102). The transport hospital information calculation unit (240) is composed of, for example, a control unit (101) and a memory unit (102). The computer (100s) is connected to the computer (100t) via a wired or wireless network. The computer (100t) includes a transport hospital information output unit (250). The computer (100t) is, for example, a device such as a smartphone, tablet, smartwatch, notebook PC, or desktop PC. The transport hospital information output unit (250) is composed of, for example, a communication unit (106), a control unit (101), and a display unit (105). Furthermore, the transport hospital information output unit (250) is composed of, for example, the display unit (105).

[0061] FIG. 13 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100u) includes a head trauma patient information input unit (210). The computer (100u) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The computer (100u) is connected to the computer (100v) via a wired or wireless network. The computer (100v) includes a bleeding surgery information calculation unit (220). The computer (100v) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The bleeding surgery information calculation unit (220) is, for example, composed of a control unit (101) and a memory unit (102). The computer (100v) is connected to the computer (100w) via a wired or wireless network. The computer (100w) includes a transfer hospital information calculation unit (240) and a transfer hospital information output unit (250). The computer (100w) is, for example, a device such as a smartphone, tablet, smartwatch, notebook PC, or desktop PC. The transfer hospital information calculation unit (240) is, for example, composed of a control unit (101) and a memory unit (102). The transfer hospital information output unit (250) is, for example, composed of a communication unit (106), the control unit (101), and a display unit (105). The transfer hospital information output unit (250) is, for example, composed of the display unit (105).

[0062] FIG. 14 is a diagram showing a schematic configuration of an exemplary system of the present invention. The computer (100x) includes a head trauma patient information input unit (210), a bleeding surgery information calculation unit (220), a transfer hospital information calculation unit (240), and a transfer hospital information output unit (250). The computer (100x) is, for example, a device such as a smartphone, tablet, smartwatch, laptop PC, or desktop PC. The head trauma patient information input unit (210) is, for example, composed of an input unit (104), a control unit (101), and a communication unit (106). The head trauma patient information input unit (210) is, for example, composed of the input unit (104). The bleeding surgery information calculation unit (220) is, for example, composed of a control unit (101) and a memory unit (102). The transfer hospital information calculation unit (240) is, for example, composed of the control unit (101) and the memory unit (102). The transfer hospital information output unit (250) is configured, for example, by the communication unit (106), the control unit (101), and the display unit (105). The transfer hospital information output unit (250) is also configured, for example, by the display unit (105).

[0063] FIG. 15 is a schematic diagram of an exemplary system including computer terminals (at the left and right ends) that serve as a head injury patient information input unit and a bleeding surgery information output unit or a transfer hospital information output unit, and a server that serves as a bleeding surgery information calculation unit or a transfer hospital information calculation unit (center). The computer terminals may be, for example, smartphones or tablets equipped with mobile communication capabilities. The server may be, for example, a cloud server connected to the Internet. However, the server's calculation functions may be incorporated into the computer terminals to enable offline operation.

[0064] The present invention also relates to an apparatus, a computer-implemented method, a program, and a non-transitory recording medium having the program recorded thereon, for assisting in the triage of a head-injured patient before transporting the patient to a hospital. The details described herein for the system for assisting in the triage of a head-injured patient before transporting the patient to a hospital equally apply to these apparatus, the computer-implemented method, the program, and the non-transitory recording medium having the program recorded thereon.

[0065] Methods: Study population: We retrospectively reviewed the electronic medical records of patients transported with head trauma to Tokyo Medical and Dental University Hospital (Tokyo, Japan) between April 1, 2018, and March 31, 2021. Based on this retrospective review, we enrolled 2,258 patients with head trauma who underwent head computed tomography (CT). We excluded patients under the age of 16 who were in cardiac arrest at the time of contact by emergency medical personnel (n = 83) or patients with missing data on two or more features (n = 26), leaving 2,123 patients for analysis.

[0066] Outcome: The predicted outcome was the presence or absence of tICH on the initial head CT scan, where tICH was defined as traumatic subarachnoid hemorrhage, acute subdural hematoma, acute epidural hematoma, or cerebral contusion at presentation.Radiological findings were double-checked by a radiologist and a neurosurgeon or by a radiologist and an emergency physician.

[0067] Features for the Prediction Model Data was obtained from the electronic medical records of patients at Tokyo Medical and Dental University Hospital. The following 18 features, which are routinely collected by Japanese paramedics, were selected as explanatory variables: 1. Patient age 2. Gender 3. Systolic blood pressure 4. Heart rate 5. Body temperature 6. Respiratory rate 7. State of consciousness-1 (Ocular component of the Glasgow Coma Scale [E-GCS]) 8. State of consciousness-2 (presence or absence of disorientation) 9. Pupillary abnormalities (defined as ectopic or bilateral pupillary dilation and loss of light reflex) 10. High-energy head injury (head injury due to a dangerous mechanism as defined by the National Institute for Health and Care Excellence [NICE] guidelines) 11. Head trauma scarring 12. Multiple trauma (combination of head injury and other trauma to the body) 13. Post-traumatic seizures 14. Loss of consciousness 15. Vomiting 16. Alcohol use 17. Hemiplegia 18. Clinical deterioration, also known as "talk and deteriorate"

[0068] Data preprocessing Missing data for body temperature and respiratory rate were filled with the overall mean value for each category. Patients with missing data other than body temperature and respiratory rate were excluded. For support vector machine and logistic regression, the values ​​of each feature were normalized.

[0069] Machine Learning Algorithms To build the classification models, five representative machine learning algorithms (e.g., XGBoost (eXtreme Gradient Boosting), random forest, support vector machine [linear kernel and radial basis function kernel], and logistic regression) were used as algorithms based on the prediction model.

[0070] The XGBoost algorithm is an efficient implementation of gradient tree boosting, a representative ensemble learning algorithm based on decision trees. Furthermore, by adding a regularization term to the loss function, it can produce a strong learner without overfitting.

[0071] Random forest is another ensemble classifier based on decision trees. In this algorithm, a large number of weak learners are trained using randomly bootstrapped data. Each weak learner votes for its prediction, and the most voted class is adopted as the prediction of all learners.

[0072] A support vector machine is a supervised machine learning algorithm in which a binary classification boundary is set to maximize the margin from each data sample.

[0073] Logistic regression is one of the most widely used classification algorithms. It uses a sigmoid function to calculate the probability of occurrence of a binary outcome.

[0074] To build classification models based on these five machine learning algorithms, we used Python version 3.6 (Python Software Foundation) and several Python modules (numpy, sklearn, matplotlib, pandas, and xgboost). We performed five-fold cross-validation to validate the classification models and optimize the model parameters.

[0075] Training, Cross-Validation, and Testing: Eighty percent of the data was used to train classification models through five-fold stratified cross-validation. The remaining 20% ​​of the data was reserved for testing the predictive ability of the resulting classification models. In the cross-validation process, grid search to obtain optimal hyperparameters was performed using scikit-learn, version 0.23.2 (scikit-learn Developers). Hyperparameters were adjusted to maximize the area under the receiver operating characteristic curve (ROC-AUC). The trained classification models calculated two class probabilities for each patient: the probability of a class with tICH and the probability of a class without tICH. The class probabilities of each classification model were calibrated using Platt calibration.

[0076] Due to the relative rarity of tICH in head trauma patients, the dataset used was imbalanced. Therefore, to address the issues associated with the imbalanced dataset, we introduced class weighting to modify the loss function: positive cases (cases with tICH) were assigned a weight 10 times higher than negative cases (cases without tICH).

[0077] Performance Evaluation To evaluate the predictive performance of classification models, we calculated six representative performance evaluation indices: sensitivity, specificity, positive predictive value (PPV), negative predictive value, positive likelihood ratio, and negative likelihood ratio. We also calculated the ROC-AUC and the area under the precision-recall curve (PR-AUC). Recall corresponds to sensitivity, and precision corresponds to PPV. The PR curve is often used together with the ROC curve to evaluate model performance, especially in imbalanced datasets.

[0078] Statistical analysis Statistical analysis was performed using R version 4.0.3 (R Group for Statistical Computing). The Mann-Whitney test was used to analyze non-normally distributed variables, and the χ test was used to analyze categorical data. 2 Tests were selected. The McNemar test was used to compare sensitivity and specificity between models. All P values ​​were two-sided, and results were considered statistically significant at P < 0.05. Bonferroni correction was applied for multiple comparison tests.

[0079] Predicting severely ill patients (whether or not surgery is required) Non-patent document 5 (Abe et al.) created a model to predict whether or not there will be intracranial hemorrhage due to trauma. In this analysis, a model was created to predict which patients among those transported by ambulance will be seriously ill (surgery patients, patients requiring intracranial pressure monitoring, and patients who will die within seven days of head trauma) before they are transported, making it possible to triage not only those with intracranial hemorrhage but also those with severe hemorrhage.

[0080] Using data from 2,123 patients who were rushed to Tokyo Medical and Dental University Hospital due to head trauma between April 2018 and March 2021, we used the machine learning algorithm eXtreme Gradient Boosting (XGBoost) and patient information that emergency responders could collect on-site (more specifically, including the 18 features listed above; similar to Non-Patent Document 5 (Abe D et al.) except that diastolic blood pressure was added and missing values ​​were not imputed). We trained a model using cross-validation to predict (1) the presence or absence of traumatic intracranial hemorrhage and (2) the need for intensive care such as surgery or death. To verify the accuracy of the trained model, we tested the predictive accuracy using data from 1,100 patients who were rushed to Tokyo Medical and Dental University and affiliated facilities (Tsuchiura Kyodo Hospital, Musashino Red Cross Hospital, Disaster Medical Center, and Ome City General Hospital) between April 2023 and March 2024.

[0081] Results: (1) Detection of traumatic intracranial hemorrhage: In cross-validation, the ROC-AUC (area under the receiver operating characteristic curve) was 0.77, and the PR-AUC (area under the precision recall curve) was 0.44. When prediction was performed using multi-center prospective data, the ROC-AUC was 0.83, and the PR-AUC was 0.58. (2) Detection of severe cases: In cross-validation, the ROC-AUC was 0.91, and the PR-AUC was 0.58. When prediction was performed using multi-center prospective data, the ROC-AUC was 0.92, and the PR-AUC was 0.64.

[0082] Interpretation: (1) Detection of traumatic intracranial hemorrhage. In Non-Patent Document 5 (Abe D et al.), data from patients transferred to Tokyo Medical and Dental University were used for model training and validation. The ROC-AUC and PR-AUC for the test data were 0.80 and 0.51, respectively. In this analysis, predictions were performed on a patient population including data from other institutions to validate the model accuracy. The results showed that both the ROC-AUC and PR-AUC did not decrease, and were slightly higher (Figures 16-19). These results support the general applicability of the prediction model developed, not only for the patient population collected at Tokyo Medical and Dental University, but also for patients in other regions. Furthermore, as indicators of performance evaluation of machine learning prediction models, ROC-AUCs of 0.8 and PR-AUCs of 0.5 are considered numerical standards for objectively determining the accuracy of a predictive model in imbalanced data with few positive cases, such as the present dataset. The prediction results for the present test data met these standards. (2) Detection of Severe Cases Detection of severe cases, including those requiring surgery, was not included in the previous paper (Abe D et al.). Here too, the ROC-AUC and PR-AUC values ​​for multi-institutional data were compared with the cross-validation values, and the predictive accuracy was comparable to or slightly higher than that of cross-validation (Figures 20-23). ​​Setting the cutoff value using the Youden Index enabled prediction with a sensitivity of 86% and a specificity of 85%. Detection of severe cases showed higher ROU-AUC and PR-AUC values ​​than for intracranial hemorrhage, indicating higher prediction accuracy. Thus, the machine learning model disclosed herein can accurately predict the severity of a patient's condition, more specifically, the need for surgery, including neurointensive care.

[0083] Use by Paramedics As an example of the use of the system disclosed herein by paramedics, we will explain the response of a middle-aged man who suffered a head injury at the scene of a traffic accident. When paramedics arrive at the scene, they first quickly collect information such as the patient's age, gender, level of consciousness (GCS score), blood pressure, heart rate, presence or absence of high-energy trauma, and presence or absence of multiple trauma. This information is entered into a tablet or smartphone in the ambulance and sent to a prediction model built on the cloud. Based on the information received, the model automatically determines the severity of the injury (whether or not surgery is required) and returns a diagnosis such as "high probability of requiring surgery." In response to this result, the system activates the transport hospital information calculation unit, which recommends appropriate hospitals to transport the patient to based on factors such as distance from the current location, estimated arrival time, and whether the medical institution has the capacity to treat severe head injuries. This prediction result is displayed on the terminal along with several candidate hospitals, their estimated arrival time, whether they can perform surgery, and their real-time acceptance status. The paramedic selects the hospital that can arrive the fastest, is capable of performing surgery, and is available for acceptance, and then transmits his or her intention to transport the patient via the system. If the hospital agrees to accept the patient, this information is displayed on the terminal and the ambulance begins transporting the patient to the selected hospital. During transport, the patient's symptoms, vital signs, and accident details are sent to the hospital in advance, allowing the hospital to prepare for the patient's arrival. In this way, the system disclosed herein assists emergency medical personnel in making decisions and enables rapid and accurate hospital selection and transport, thereby contributing to improved patient prognosis.

[0084] System Use in Southeast Asia Emergency medical care in Southeast Asia, such as Thailand, differs significantly from that in Japan. Patients are sometimes transported to hospitals by volunteers or by ambulance crews. When ambulance crews transport patients, the system is expected to be used in the form of a web application at the scene of the injury, similar to the approach expected in Japan. Furthermore, unlike Japan, a problem with head injury treatment in Thailand is the limited number of neurosurgeons, making it difficult for doctors to assess the severity of head injuries. Even when trauma patients are transported to the nearest hospital, they are often transferred to other hospitals, resulting in many patients concentrating in a few large hospitals. In response to such situations, the system disclosed herein could be used to assess the severity of a patient's injury when they are transported to the nearest hospital.

[0085] 100 Computer 100a to 100x Computer 101 Control unit 102 Memory unit 103 Peripheral device I / F unit 104 Input unit 105 Display unit 106 Communication unit 110 Bus 120 Network 130 External server 140 Database 210 Head injury patient information input unit 220 Bleeding surgery information calculation unit 230 Bleeding surgery information output unit 240 Transfer hospital information calculation unit 250 Transfer hospital information output unit

Claims

1. A system for supporting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; and a bleeding surgery information output unit for outputting the calculated information about whether or not head surgery is required; the information about the head-injured patient is information about the head-injured patient that can be collected by an emergency team at the scene, and the information about the head-injured patient includes at least impaired consciousness, high-energy head injury, head trauma scarring, E-GCS, and pupil abnormalities; the calculation of information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient is performed using a trained machine learning model, and the trained machine learning model takes the information about the head-injured patient as input data and outputs information about whether or not head surgery is required or whether or not the patient has died as output data.

2. A method for triaging a head-injured patient before transporting them to a hospital, comprising: a step of inputting information about the head trauma patient into a head trauma patient information input unit for inputting information about the head trauma patient; a step of causing a surgery information calculation unit to calculate information about whether or not head surgery is required for the head trauma patient based on the input information about the head trauma patient; and a step of causing a bleeding surgery information output unit to output the calculated information about whether or not head surgery is required, wherein the information about the head trauma patient is information about the head trauma patient that can be collected by an emergency team at the scene, and the information about the head trauma patient includes at least impaired consciousness, high-energy head injury, head trauma scar, E-GCS, and pupil abnormality, and calculating information about whether or not head surgery is required for the head trauma patient based on the input information about the head trauma patient is performed using a trained machine learning model, A method in which a trained machine learning model takes information about head injury patients as input data and outputs information about whether head surgery is required or whether the patient has died.

3. A method for triaging a head-injured patient before transporting the patient to a hospital, comprising: a step of inputting information about the head trauma patient into a head trauma patient information input unit for inputting information about the head trauma patient; a step of causing a bleeding surgery information calculation unit to calculate information about whether or not head surgery is required for the head trauma patient based on the input information about the head trauma patient; a step of causing a transport hospital information calculation unit to calculate information about a hospital to which the head trauma patient should be transported based on the calculated information about whether or not head surgery is required; and a step of causing a transport hospital information output unit to output information about a hospital to which the head trauma patient should be transported, A method in which, based on input information about a head-injured patient, information regarding whether head surgery is required for the head-injured patient is calculated using a trained machine learning model, wherein the trained machine learning model uses the information about the head-injured patient as input data and information regarding whether head surgery is required or whether the patient has died as output data, and the calculation of information regarding hospitals to which head-injured patients should be transported is performed by referring to a list that associates information regarding whether head surgery is required with information regarding hospitals to which head-injured patients should be transported.

4. A system for supporting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; and a bleeding surgery information output unit for outputting the calculated information about whether or not head surgery is required.

5. The system according to claim 4, wherein the information about the head injury patient is information about the head injury patient that can be collected by an emergency team at the scene.

6. The system of claim 4, wherein the information about the head trauma patient is one or more pieces of information selected from the group consisting of the head trauma patient's age, sex, vital signs, state of consciousness, neurological symptoms, high-energy trauma, and multiple trauma.

7. The system of claim 4, wherein the trained machine learning model is used to calculate information regarding whether or not a head injury patient requires head surgery based on input information regarding the head injury patient.

8. The system described in claim 4, wherein the trained machine learning model takes information about head trauma patients as input data and information about whether head surgery is required or whether the patient has died as output data.

9. The system of claim 4, wherein the trained machine learning model is created by inputting a dataset about past head trauma patients, including information about the head trauma patient and information about whether head surgery was required or whether the patient died, into the machine learning model.

10. The system of claim 4, wherein the machine learning algorithm in the trained machine learning model is eXtreme Gradient Boosting, Random Forest, Support Vector Machine, or Logistic Regression.

11. The system of claim 4, wherein the trained machine learning model is created without imputing missing values.

12. The system of claim 4, wherein training in the trained machine learning model is performed using cross-validation.

13. A system for supporting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; a transport hospital information calculation unit for calculating information about a hospital to which the head-injured patient should be transported based on the calculated information about whether or not head surgery is required; and a transport hospital information output unit for outputting the calculated information about a hospital to which the head-injured patient should be transported.

14. A method for triaging a head-injured patient before transporting them to a hospital, comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; and a step of causing a bleeding surgery information output unit to output the calculated information about whether or not head surgery is required.

15. A method for triaging a head-injured patient before transporting them to a hospital, comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about whether or not head surgery is required for the head-injured patient based on the input information about the head-injured patient; a step of causing a transport hospital information calculation unit to calculate information about a hospital to which the head-injured patient should be transported based on the calculated information about whether or not head surgery is required; and a step of causing a transport hospital information output unit to output the calculated information about a hospital to which the head-injured patient should be transported.

16. An apparatus for use in the system of claim 4, comprising a head injury patient information input section for inputting information about a head injury patient.

17. A device for use in the system described in claim 4, including a bleeding surgery information calculation unit for calculating information regarding the necessity of head surgery for a head trauma patient based on input information about the head trauma patient.

18. A device for use in the system according to claim 4, comprising a bleeding surgery information output unit for outputting calculated information on the necessity of head surgery.

19. An apparatus for use in the system of claim 13, comprising a head injury patient information input section for inputting information about a head injury patient.

20. A device for use in the system described in claim 13, including a bleeding surgery information calculation unit for calculating information regarding the necessity of head surgery for a head trauma patient based on input information about the head trauma patient.

21. An apparatus for use in the system of claim 13, including a hospital transfer information calculation unit for calculating information on the hospital to which a head injury patient should be transferred based on the calculated information on whether or not head surgery is required.

22. An apparatus for use in the system according to claim 13, comprising a transport hospital information output unit for outputting calculated information regarding a hospital to which a head injury patient should be transported.

23. A method for carrying out the method of claim 14, comprising the step of inputting information about a head trauma patient to a head trauma patient information input unit for inputting information about the head trauma patient.

24. A method for carrying out the method described in claim 14, comprising a step of causing a bleeding surgery information calculation unit to calculate information regarding the need for head surgery for a head trauma patient based on input information regarding the head trauma patient, to calculate information regarding the need for head surgery for the head trauma patient.

25. A method for carrying out the method according to claim 14, comprising a step of causing a bleeding surgery information output unit, which outputs calculated information on whether head surgery is required, to output information on whether head surgery is required.

26. A method for carrying out the method of claim 15, comprising the step of inputting information about a head trauma patient to a head trauma patient information input unit for inputting information about the head trauma patient.

27. A method for carrying out the method described in claim 15, comprising a step of causing a bleeding surgery information calculation unit to calculate information regarding the need for head surgery for a head trauma patient based on input information regarding the head trauma patient, to calculate information regarding the need for head surgery for the head trauma patient.

28. A method for carrying out the method according to claim 15, comprising a step of causing a hospital transfer information calculation unit to calculate information about a hospital to which a head-injured patient should be transferred, based on the calculated information about whether or not head surgery is required.

29. A method for carrying out the method described in claim 15, comprising a step of causing a transport hospital information output unit for outputting calculated information about hospitals to which head-injured patients should be transported to output information about hospitals to which head-injured patients should be transported.

30. A system for supporting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about the head-injured patient regarding the presence or absence of bleeding in the head and information about whether head surgery is required based on the input information about the head-injured patient; and a bleeding surgery information output unit for outputting the calculated information about the presence or absence of bleeding in the head and information about whether head surgery is required.

31. A system for supporting triage of head-injured patients before transporting them to a hospital, comprising: a head-injured patient information input unit for inputting information about the head-injured patient; a bleeding surgery information calculation unit for calculating information about the head-injured patient regarding the presence or absence of bleeding in the head and information regarding the need for head surgery based on the input information about the head-injured patient; a hospital transfer information calculation unit for calculating information about the hospital to which the head-injured patient should be transported based on the calculated information about the presence or absence of bleeding in the head and information regarding the need for head surgery; and a hospital transfer information output unit for outputting the calculated information about the hospital to which the head-injured patient should be transported.

32. A method for triaging a head-injured patient before transporting them to a hospital, comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about the head-injured patient regarding the presence or absence of bleeding in the head and information about whether head surgery is required, based on the input information about the head-injured patient; and a step of causing a bleeding surgery information output unit to output the calculated information about the presence or absence of bleeding in the head and information about whether head surgery is required.

33. A method for triaging a head-injured patient before transporting them to a hospital, comprising: a step of inputting information about the head-injured patient into a head-injured patient information input unit for inputting information about the head-injured patient; a step of causing a bleeding surgery information calculation unit to calculate information about the head-injured patient regarding the presence or absence of bleeding in the head and information about whether head surgery is required, based on the input information about the head-injured patient; a step of causing a transport hospital information calculation unit to calculate information about a hospital to which the head-injured patient should be transported, based on the calculated information about the presence or absence of bleeding in the head and information about whether head surgery is required; and a step of causing a transport hospital information output unit to output the calculated information about the hospital to which the head-injured patient should be transported.

Images