WO2023106960A1 - Procédé de pronostic de survenance d'évènement médical sur la santé une personne - Google Patents
Procédé de pronostic de survenance d'évènement médical sur la santé une personne Download PDFInfo
- Publication number
- WO2023106960A1 WO2023106960A1 PCT/RU2021/000580 RU2021000580W WO2023106960A1 WO 2023106960 A1 WO2023106960 A1 WO 2023106960A1 RU 2021000580 W RU2021000580 W RU 2021000580W WO 2023106960 A1 WO2023106960 A1 WO 2023106960A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- health
- human health
- medical event
- parameters
- predicting
- Prior art date
Links
- 230000036541 health Effects 0.000 title claims abstract description 130
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000010801 machine learning Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 11
- 230000008859 change Effects 0.000 claims abstract description 9
- 238000011161 development Methods 0.000 claims description 13
- 201000010099 disease Diseases 0.000 claims description 9
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 9
- 230000007774 longterm Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 6
- 238000013528 artificial neural network Methods 0.000 description 12
- 230000036772 blood pressure Effects 0.000 description 9
- 210000004369 blood Anatomy 0.000 description 8
- 239000008280 blood Substances 0.000 description 8
- HVYWMOMLDIMFJA-DPAQBDIFSA-N cholesterol Chemical compound C1C=C2C[C@@H](O)CC[C@]2(C)[C@@H]2[C@@H]1[C@@H]1CC[C@H]([C@H](C)CCCC(C)C)[C@@]1(C)CC2 HVYWMOMLDIMFJA-DPAQBDIFSA-N 0.000 description 8
- 238000013459 approach Methods 0.000 description 7
- 230000002123 temporal effect Effects 0.000 description 6
- 230000007423 decrease Effects 0.000 description 5
- 230000000306 recurrent effect Effects 0.000 description 5
- 238000012549 training Methods 0.000 description 5
- 238000004820 blood count Methods 0.000 description 4
- DDRJAANPRJIHGJ-UHFFFAOYSA-N creatinine Chemical compound CN1CC(=O)NC1=N DDRJAANPRJIHGJ-UHFFFAOYSA-N 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 208000019693 Lung disease Diseases 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 235000012000 cholesterol Nutrition 0.000 description 3
- 208000037998 chronic venous disease Diseases 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 208000024172 Cardiovascular disease Diseases 0.000 description 2
- 206010053172 Fatal outcomes Diseases 0.000 description 2
- 102000017011 Glycated Hemoglobin A Human genes 0.000 description 2
- 230000002411 adverse Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 229940109239 creatinine Drugs 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 210000003743 erythrocyte Anatomy 0.000 description 2
- 108091005995 glycated hemoglobin Proteins 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 230000035485 pulse pressure Effects 0.000 description 2
- 230000000391 smoking effect Effects 0.000 description 2
- 230000035488 systolic blood pressure Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 208000001072 type 2 diabetes mellitus Diseases 0.000 description 2
- 238000002562 urinalysis Methods 0.000 description 2
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 description 1
- 206010019233 Headaches Diseases 0.000 description 1
- 206010019280 Heart failures Diseases 0.000 description 1
- 102000001554 Hemoglobins Human genes 0.000 description 1
- 108010054147 Hemoglobins Proteins 0.000 description 1
- 241001025261 Neoraja caerulea Species 0.000 description 1
- 208000031662 Noncommunicable disease Diseases 0.000 description 1
- 206010037596 Pyelonephritis Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000003276 anti-hypertensive effect Effects 0.000 description 1
- 239000002220 antihypertensive agent Substances 0.000 description 1
- 229940127088 antihypertensive drug Drugs 0.000 description 1
- 208000037849 arterial hypertension Diseases 0.000 description 1
- 238000009534 blood test Methods 0.000 description 1
- 230000036760 body temperature Effects 0.000 description 1
- 230000001684 chronic effect Effects 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 208000029078 coronary artery disease Diseases 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 206010012601 diabetes mellitus Diseases 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000035487 diastolic blood pressure Effects 0.000 description 1
- 230000006806 disease prevention Effects 0.000 description 1
- 230000002526 effect on cardiovascular system Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 239000008103 glucose Substances 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 208000035474 group of disease Diseases 0.000 description 1
- 231100000869 headache Toxicity 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000009533 lab test Methods 0.000 description 1
- 210000000265 leukocyte Anatomy 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000000491 multivariate analysis Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 229940126701 oral medication Drugs 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000036387 respiratory rate Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000004062 sedimentation Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
Definitions
- the invention relates to the field of information and communication technologies for processing medical data, in particular to a method for predicting the occurrence of a medical event in human health using machine learning, taking into account estimates of the variability of data on the state of human health over time.
- the presented solution can be used, at least in clinical practice, by physicians and other medical professionals involved in the diagnosis, treatment and prevention of diseases, in predicting the onset of various medical events for a patient, such as hospital admission, death, the development of a specific complication, the development concomitant disease, when planning preventive measures, medical examination of patients, as well as when developing a set of predictive models that predict standard medical events (hospitalization, death, the development of a specific complication, the development of a concomitant disease, etc.).
- a model built on fixed values of input features does not allow one to analyze the state of health of an organism in the reverse perspective, taking into account all individual characteristics.
- the current forecasting model does not separate the above “scenarios” and gives different people a single forecast, which is inaccurate in at least 50% of cases.
- the model takes into account such a patient parameter as “I don’t smoke”. But the patient could quit smoking 10 years ago, or he could quit smoking yesterday - and his condition is due to the consequences of refusal or the accumulated effect.
- Models do not understand the dynamics of changes in the state of the body and therefore, although they can be used in a development environment (where there is no need to take into account unique or infrequent deviations), in real clinical practice it is not possible to achieve accuracy, and in the laboratory it is impossible to take into account all possible life scenarios that happen on practice.
- One method includes obtaining a first time series of health events, where the first time series includes relevant health data associated with a particular patient at each of the plurality of time steps; processing the first temporal sequence of health events with using the recurrent neural network to generate the output of the neural network for the first time sequence; and generating, from the output of the neural network for the first time series, health analysis data that is indicative of future health events that may occur after the last time step in the time series.
- US9652712B2 published May 16, 2017, discloses methods, systems, and devices, including computer programs encoded on computer media, for using recurrent neural networks to analyze health events.
- One of the methods includes: processing each of the plurality of initial health event time sequences to generate, for each of the initial time sequences, a corresponding internal state of the recurrent neural network network for each time step in the initial time sequence; storing, for each of the initial time sequences, one or more internal network states for the time steps in the time sequence in storage; obtaining the first time sequence; processing the first time sequence with the recurrent neural network to create an internal sequence state for the first time sequence; and selecting one or more initial time series likely to include health events that predict future health events in the first time series.
- US20200152333A1 discloses methods, systems and devices, including computer programs encoded on computer media, for predicting future adverse health events using neural networks.
- One method involves obtaining the patient's electronic health record data; generating, based on the data of the electronic medical record, an input sequence, including a corresponding representation of the feature at each of the plurality of time steps of the time window, including, for each time step of the time window: determining, for each of the possible numerical features, whether the numerical feature occurred during the time window; and generating, for each of the possible numerical features, one or more presence indicia that determine whether the numeric indicia occurred during the time window; and processing the input sequence with the neural network to generate a neural network output that is indicative of a predicted probability that an adverse health event will occur to the patient.
- the technical result of the claimed invention is to increase the accuracy of predicting the onset of a medical event in the patient's health, taking into account the dynamics of changes in the patient's health indicators over time in the forecasting process, improving the quality of the provided forecast using machine learning, taking into account estimates of the variability of data on the state of human health over time.
- a computer-implemented method for predicting the occurrence of a medical event in human health using machine learning taking into account estimates of the variability of data on the state of human health over time, in which: data on the state of human health is extracted containing human health parameters by processing medical documents characterizing the state of health person; on the basis of the extracted data on the state of human health, time sequences of data on the state of human health are formed; wherein each time sequence is obtained for values of one health parameter; determining a set of health parameters for predicting the occurrence of a medical event in human health; moreover, the specified set of health parameters depends on the predicted medical event in human health; for each health parameter from the set of parameters for predicting the occurrence of a medical event in human health, converting the generated time sequence of health parameter values into an estimated indicator of the dynamics of changes in health parameter values; and receive a set of estimated indicators of the dynamics of changes in the values of health parameters over time to predict the occurrence of a medical event in human health; the resulting set of estimated indicators of the dynamics of change in time of health
- medical events in human health can be at least the following: hospitalization, death, development of a specific complication, development of a concomitant disease.
- the health parameters for predicting the occurrence of a medical event in human health can be at least the following: physiological parameters, laboratory test data.
- the evaluation indicator may be an index of long-term variability of a human health parameter.
- each time sequence can be obtained for one health parameter for a given period of time.
- Fig. 1 illustrates a flowchart variant of a method for predicting the occurrence of a medical event in human health, taking into account estimates of the variability of human health data over time.
- Fig. 2 illustrates a flowchart variant of the data conversion process in the present invention.
- Fig. 3 illustrates an example of a data cycle in the present invention.
- Fig. 4 illustrates a general diagram of a computing device for implementing the present invention.
- the input of the predictive model is not the absolute values of the signs that characterize the state of human health, but an analytical assessment of the dynamics of changes in these signs over time (for example, a numerical characteristic, index). That is, for each sign, not an absolute number is obtained, but a value depending on the retrospective analysis of this sign of human health over time.
- the claimed solution work is carried out with sequences of time-discrete numerical values of specific patient health parameters (objective data, laboratory parameters, etc.), which allows taking into account the dynamics of the parameter and favorably affects the accuracy and specificity of the forecast.
- the claimed solution can be used in predictive models for various purposes and the class of the problem being solved (classification, regression), provided that they are trained on real medical data that reflect the state of human health, for example, on electronic medical records (EMR) data.
- EMR electronic medical records
- This solution is used for ready-made, already formed sequences and aims to improve the accuracy of the forecast based on the use of data from these sequences. That is, the formation of the sequence occurs at the stage of processing medical documents by a complex of NLP models (Neuro-Linguistic Programming, abbr.
- the solution works separately for each time sequence of one trait (for example, a sequence of values for a human health parameter such as blood pressure or blood sugar or glycated hemoglobin, etc.) obtained for one patient during a certain period of time.
- the purpose of transforming the time sequence of values into a numerical index is to provide the input of the predictive model not with a single record, but with an assessment of the dynamics of change in time of the values of a sign (parameter). This approach makes it possible to take into account the dynamics of the patient's health indicator in the process of training predictive models and improve the quality of the forecast they provide (Fig. 1).
- An index of long-term variability of a feature can be used as such an estimate (estimated indicator).
- an index may be an assessment of the inter-visit or temporal variability of a feature.
- inter-visit variability of blood pressure is widely used, defined as deviations of blood pressure from the average level calculated for a specific period (a year or several).
- an increase in visit-to-visit systolic pressure variability is considered a predictor of an increased risk of cardiovascular outcomes in heart failure and arterial hypertension.
- standard deviation indicators or the coefficient of variation are used as indicators of variability. In this case, the standard deviation is a less preferred choice due to its dependence on the average level of the parameter.
- VIM Variability Independent of
- the next step is a correlation analysis of the relationship of each transformed indicator with the target parameter (feature importance). This approach will allow us to determine the optimal method for calculating indices for each feature.
- a transition to boolean variables such as “positive gain in magnitude” or “negative gain in magnitude” can also be used, with such an estimate given for different time periods, for example, “positive gain in body mass index over the past 12 months” or “decrease in creatinine after oral administration.
- an analysis of a parameter can be carried out for which a significant amplitude of the indicator for a certain period of time is revealed, for example, a significant amplitude of the parameters of a biochemical blood test for the previous 12 months is revealed.
- accompanying events are analyzed - diseases, medical recommendations (executed or not), variability of other health indicators.
- the models use the gradient boosting method to determine the numerical value of the probability of a fatal outcome in patients from a certain sample (inclusion criteria - the presence of type 2 diabetes in the list of the patient's main diagnoses, age in the range of 18-99 years).
- inclusion criteria the presence of type 2 diabetes in the list of the patient's main diagnoses, age in the range of 18-99 years.
- gradient boosting predictions are made based on sequential ensembles of algorithms, in which each successive model is trained using the error data of previous models to further reduce errors.
- the result of the operation of each algorithm is the resulting value issued by the ensemble - this is the sum of the results individual regression models.
- the algorithms included in the ensemble work in turn, while each subsequent one is trained, including on the data errors of the previous one.
- Model input parameters include: age, gender, antihypertensive therapy use, weight, blood hemoglobin, headache, diastolic blood pressure, systolic blood pressure, blood creatinine, white blood cell count, total cholesterol, oral drug use, height, sedimentation rate red blood cells (ESR), body temperature, platelet count, fatigue, respiratory rate, red blood cell count.
- ESR sedimentation rate red blood cells
- Table 1 shows the prior art - an exemplary view of the data set for training a model of predicting a fatal outcome, for the formation of which fixed values of each parameter were taken.
- This approach can be used to predict standard medical events (for example, hospital admission, death, development of a specific complication, development of a concomitant disease, etc.) using predictive machine learning models.
- the patient's electronic medical record contains a set of medical documents from which, using NLP models (Neuro-Linguistic Programming, abbr. NLP) or mapping, specific health parameters can be extracted in a formalized form (Fig. 1).
- NLP models Neuro-Linguistic Programming, abbr. NLP
- mapping specific health parameters can be extracted in a formalized form (Fig. 1).
- FIG. 3 shows an example of a cycle of working with data in the claimed solution. From formalized medical data that have undergone preprocessing, new features are formed (estimates of parameter variability). The storage of formalized medical data can be replenished with new features that are used to train various predictive models, as well as for their additional training.
- the peculiarity of the data contained in the EHR is that all medical information is collected at the time of contacting the health facility, and the composition of this data may differ depending on the reason for the visit, and the frequency of entering the same parameter into the EHR depends on the treatment plan and condition patient.
- Table 3 shows an example of a set of medical information obtained for the same patient on different dates and during different visits. So, when calling a doctor at home or going to a polyclinic, one set of information is formed, which includes the patient's complaints and objective examination data, after which a recommendation is given for additional examinations (laboratory or instrumental), as a result of which a fundamentally new set of data is formed.
- Table 3 An example of values of health parameters extracted from an electronic medical record.
- any data set will include physiological parameters (eg, blood pressure, pulse pressure, body mass index, heart rate, etc.), laboratory data (complete blood count, blood biochemistry, urinalysis, antibody tests, and etc.) and information about the history (for example, a history of chronic non-communicable diseases, such as diabetes, coronary heart disease, pyelonephritis, etc.; the total number of major diagnoses; the number of newly diagnosed diagnoses in the last year; the number of hospitalizations in the hospital, changes in conditions, occurrence of complications).
- physiological parameters eg, blood pressure, pulse pressure, body mass index, heart rate, etc.
- laboratory data complete blood count, blood biochemistry, urinalysis, antibody tests, and etc.
- information about the history for example, a history of chronic non-communicable diseases, such as diabetes, coronary heart disease, pyelonephritis, etc.; the total number of major diagnoses; the number of newly diagnosed diagnoses in the last year; the number of hospitalizations in the hospital
- Indexing can be carried out by methods of assessing the inter-visit variability of the indicator.
- the optimal approach is simple exponential smoothing, in which the weight of the parameter value decreases as you go deeper into the historical data. Yes, for predicting a target medical event (for example, a hospital admission), all available observations are important, but the last records of the time sequence will have a decisive weight.
- the dynamics of the patient's body mass index is estimated by the logical variable "growth_bmi", where the variable takes the value True if the difference between the last and first retrieved is positive and False if this value is negative.
- An assessment of the dynamics of laboratory parameters is obtained as follows - STD, ARV and VIM are calculated for each informative parameter, this set is used in experimental training with at least 5 algorithms (gradient boosting, adaptive boosting, multilayer neural network, logistic regression, decision trees), after which the feature selection method determines the final set of input parameters and, based on the metrics, selects the final machine learning algorithm.
- the model gives the probability of the target event occurring (hospitalization).
- human health data can be extracted from human EHR using NLP models or mapping.
- Human health parameters are, for example, physiological parameters (for example, blood pressure, pulse pressure, body mass index, heart rate, etc.), laboratory data (complete blood count, blood biochemistry, urinalysis, antibody tests, etc.).
- time sequences of data on the state of human health are formed; wherein each time sequence is obtained for the values of one health parameter.
- a set of health parameters is determined to predict the occurrence of a medical event in human health. At the same time, the specified set of health parameters depends on the predicted medical event in human health.
- the set of parameters for predicting the hospitalization of patients with cardiovascular diseases will differ from the set of parameters for predicting the hospitalization of patients with pulmonary diseases (Table 4).
- the generated temporal sequence of health parameter values is converted into an estimated indicator of the dynamics of changes in health parameter values, for example, the index of long-term variability of a human health parameter.
- a set of estimated indicators of the dynamics of changes in the values of health parameters over time is obtained to predict the onset of a medical event in human health.
- the resulting set of estimated indicators of the dynamics of changes over time in the values of health parameters is fed to the input of the predictive machine learning model to predict the occurrence of a medical event in human health, and at the output of the predictive machine learning model, the probability of the occurrence of a medical event in human health is obtained, for example, the probability of hospitalization in a hospital, the probability death, the likelihood of developing a specific complication, the likelihood of developing a concomitant disease.
- FIG. 4 shows a general diagram of a computing device (400) that provides the data processing necessary to implement the claimed solution.
- the device (400) contains components such as: one or more processors (401), at least one memory (402), data storage (403), input/output interfaces (404), I/O ( 405), networking tools (406).
- the processor (401) of the device performs the basic computing operations necessary for the operation of the device (400) or the functionality of one or more of its components.
- the processor (401) executes the necessary machine-readable instructions contained in the main memory (402).
- the memory (402) is typically in the form of RAM and contains the necessary software logic to provide the desired functionality.
- the data storage means (403) can be in the form of HDD, SSD disks, raid array, network storage, flash memory, optical information storage devices (CD, DVD, MD, Blue-Ray disks), etc. Means (403) allows you to perform long-term storage of various types of information.
- Interfaces (404) are standard means for connecting and working with the server part, for example, USB, RS232, RJ45, LPT, COM, HDMI, PS/2, Lightning, FireWire, etc.
- interfaces (404) depends on the specific implementation of the device (400), which can be a personal computer, mainframe, server cluster, thin client, smartphone, laptop, and the like.
- the keyboard should be used as the data I/O (405) in any embodiment of the system.
- the keyboard hardware can be any known: it can be either a built-in keyboard used on a laptop or netbook, or a separate device connected to a desktop computer, server, or other computer device.
- the connection can be either wired, in which the keyboard connection cable is connected to the PS / 2 or USB port located on the system unit of the desktop computer, or wireless, in which the keyboard exchanges data via a wireless communication channel, for example, a radio channel, with base station, which, in turn, is directly connected to the system unit, for example, to one of the USB ports.
- the following I/O devices can also be used: joystick, display (touchscreen), projector, touchpad, mouse, trackball, light pen, speakers, microphone, etc.
- Means of networking are selected from devices that provide network reception and transmission of data, for example, an Ethernet card, WLAN/Wi-Fi module, Bluetooth module, BLE module, NFC module, IrDa, RFID module, GSM modem, etc.
- the organization of data exchange over a wired or wireless data transmission channel is provided, for example, WAN, PAN, LAN (LAN), Intranet, Internet, WLAN, WMAN or GSM, 3G, 4G, 5G.
- the components of the device (400) are coupled via a common data bus (407).
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
L'invention concerne un procédé de pronostic de survenance d'évènement médical sur la santé une personne utilisant l'apprentissage machine, et tenant compte des estimations de variabilité des données sur l'état de santé d'une personne dans le temps. Le résultat technique de la présente invention consiste en une augmentation de la précision et de la qualité de pronostic de la survenance d'un évènement médical sur la santé d'un patient, et la prise en compte de la dynamique de changement des indicateurs de santé dans le temps lors du processus de pronostic. Ce procédé consiste à: extraire des données sur l'état de santé d'une personne, comprenant des paramètres de santé de la personne, ceci en traitant des documents médicaux caractérisant l'état de santé de la personne; sur la base des données extraites sur l'état de santé de la personne, générer des séquences temporelles de données sur l'état de santé de la personne; chaque séquence temporelle est obtenue pour des valeurs d'un paramètre de santé; déterminer un ensemble de paramètres de santé pour le pronostic de survenance d'évènement médical sur la santé une personne; ledit ensemble de paramètres de santé dépend de l'évènement médical à pronostiquer sur l'état de santé de la personne; pour chaque paramètre de santé de l'ensemble de paramètres pour le pronostic de survenance d'évènement médical sur la santé une personne, convertir la séquence temporelle générée de valeurs de paramètre de santé en un indicateur estimatif de la dynamique de changement des valeurs du paramètre de santé; obtenir un ensemble d'indicateurs estimatifs de la dynamique de changement dans le temps des valeurs de paramètres de santé pour le pronostic de survenance d'évènement médical sur la santé une personne; envoyer à l'entrée d'un modèle de pronostic d'apprentissage machine l'ensemble obtenu d'indicateurs estimatifs de la dynamique de changement dans le temps des valeurs de paramètres de santé pour le pronostic de survenance d'évènement médical sur la santé une personne; et à la sortie du modèle de pronostic d'apprentissage machine, obtenir une probabilité de survenance d'évènement médical sur la santé une personne.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
RU2021136205A RU2021136205A (ru) | 2021-12-08 | Способ прогнозирования наступления медицинского события в здоровье человека с применением машинного обучения с учетом оценок изменчивости данных о состоянии здоровья человека во времени | |
RU2021136205 | 2021-12-08 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023106960A1 true WO2023106960A1 (fr) | 2023-06-15 |
Family
ID=86730874
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/RU2021/000580 WO2023106960A1 (fr) | 2021-12-08 | 2021-12-16 | Procédé de pronostic de survenance d'évènement médical sur la santé une personne |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2023106960A1 (fr) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116933046A (zh) * | 2023-09-19 | 2023-10-24 | 山东大学 | 基于深度学习的多模态健康管理方案生成方法和系统 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180301221A1 (en) * | 2017-02-28 | 2018-10-18 | Rothman Steven I | Adverse Event Prediction and Detection Using Wearable Sensors and Adaptive Health Score Analytics |
US20190216350A1 (en) * | 2014-11-14 | 2019-07-18 | Zoll Medical Corporation | Medical premonitory event estimation |
WO2021066918A1 (fr) * | 2019-09-30 | 2021-04-08 | DawnLight Technologies Inc. | Systèmes et procédés de détermination de fréquence cardiaque et de fréquence respiratoire à partir d'un signal radar à l'aide de procédés d'apprentissage automatique |
WO2021133935A1 (fr) * | 2019-12-26 | 2021-07-01 | Kpn Innovations, Llc | Procédés et systèmes de personnalisation de traitements |
-
2021
- 2021-12-16 WO PCT/RU2021/000580 patent/WO2023106960A1/fr active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190216350A1 (en) * | 2014-11-14 | 2019-07-18 | Zoll Medical Corporation | Medical premonitory event estimation |
US20180301221A1 (en) * | 2017-02-28 | 2018-10-18 | Rothman Steven I | Adverse Event Prediction and Detection Using Wearable Sensors and Adaptive Health Score Analytics |
WO2021066918A1 (fr) * | 2019-09-30 | 2021-04-08 | DawnLight Technologies Inc. | Systèmes et procédés de détermination de fréquence cardiaque et de fréquence respiratoire à partir d'un signal radar à l'aide de procédés d'apprentissage automatique |
WO2021133935A1 (fr) * | 2019-12-26 | 2021-07-01 | Kpn Innovations, Llc | Procédés et systèmes de personnalisation de traitements |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116933046A (zh) * | 2023-09-19 | 2023-10-24 | 山东大学 | 基于深度学习的多模态健康管理方案生成方法和系统 |
CN116933046B (zh) * | 2023-09-19 | 2023-11-24 | 山东大学 | 基于深度学习的多模态健康管理方案生成方法和系统 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Shickel et al. | DeepSOFA: a continuous acuity score for critically ill patients using clinically interpretable deep learning | |
Fu et al. | Development and validation of early warning score system: A systematic literature review | |
US8554580B2 (en) | Automated management of medical data using expert knowledge and applied complexity science for risk assessment and diagnoses | |
US10039485B2 (en) | Method and system for assessing mental state | |
Sk et al. | Coronary Heart Disease Prediction and Classification using Hybrid Machine Learning Algorithms | |
US8595159B2 (en) | Predicting near-term deterioration of hospital patients | |
US10638965B2 (en) | Method and system for monitoring stress conditions | |
RU2619644C2 (ru) | Клиническая система поддержки принятия решений для прогнозирующего планирования выписки | |
JP2006507875A (ja) | 患者の健康を自動診断するシステムおよび方法 | |
CA2702042A1 (fr) | Cotation de gravite multi-automatisee | |
WO2014033681A2 (fr) | Techniques de modélisation pour prédire la mortalité dans les unités de soins intensifs | |
WO2018096310A1 (fr) | Dispositif de surveillance de l'état d'un patient et procédé de surveillance de l'état d'un patient | |
US20100217094A1 (en) | Point-of-care enactive medical system and method | |
CN114724716A (zh) | 进展为2型糖尿病的风险预测的方法、模型训练及装置 | |
Wang et al. | Predictive classification of ICU readmission using weight decay random forest | |
JP2023548253A (ja) | 医学検査値分析を実行するためのコンピュータ実装された方法、及び、装置 | |
CN116864139A (zh) | 疾病风险评估方法、装置、计算机设备及可读存储介质 | |
WO2023106960A1 (fr) | Procédé de pronostic de survenance d'évènement médical sur la santé une personne | |
Chen et al. | PREDICTIVE MODELING OF POSTOPERATIVE OUTCOMES | |
CN113782197B (zh) | 基于可解释性机器学习算法的新冠肺炎患者转归预测方法 | |
CN112820368A (zh) | 重症患者数据集的构建方法、系统、设备和存储介质 | |
Zhao et al. | Development and application of an intensive care medical data set for deep learning | |
Zhao et al. | External validation of a deep learning prediction model for in-hospital mortality among ICU patients | |
JP7333549B2 (ja) | 疾病リスク評価方法、疾病リスク評価システム、及び健康情報処理装置 | |
WO2023121503A1 (fr) | Procédé de pronostic de maladies non infectieuses chroniques dans des organismes biologiques |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21967396 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202393055 Country of ref document: EA |