Sepsis patient monitoring system based on medical information processing

Abstract

The application discloses a sepsis patient monitoring system based on medical information processing, comprising a data set construction module, a sepsis patient risk assessment model establishment module and a sepsis patient monitoring module. The application belongs to the field of medical information processing and specifically refers to a sepsis patient monitoring system based on medical information processing. The scheme defines a mixed distribution, defines a medical-oriented divergence distribution and enhances the early warning capability for extreme physiological states. The scheme introduces risk-sensitive weighted processing to prevent excessive alarm. The scheme further improves the sepsis patient monitoring accuracy. The normal reference value is taken as a demarcation point, an exponential saturation curve is used on both sides of the index high and low, the asymmetrical punishment of the risk degree is realized, the actual risk condition when the physiological index of the sepsis patient deviates from the normal range can be more accurately reflected, the dangerous signal lower than the normal value is captured in time to deteriorate the patient's condition, and the patient monitoring effect is improved.

Description

Technical Field

[0001] The present invention relates to the field of medical information processing, and in particular to a sepsis patient monitoring system based on medical information processing. Background Art

[0002] A sepsis patient monitoring system is a medical information system used for real-time, dynamic monitoring and assessment of sepsis patients, designed to help healthcare professionals stay informed of changes in their patients' conditions. However, typical sepsis patient monitoring systems suffer from weak risk capture capabilities, low early prediction accuracy, insufficient early warning for extreme physiological states, and high false alarm rates due to outlier monitoring values. These systems also fail to effectively address the asymmetric risk of indicators deviating from the normal range and provide insufficient monitoring for extreme conditions, resulting in poor patient monitoring outcomes. Summary of the Invention

[0003] In view of the above situation, in order to overcome the defects of the existing technology, the present invention provides a sepsis patient monitoring system based on medical information processing. In view of the problems of weak risk capture ability, low early prediction accuracy, insufficient early warning of extreme physiological states, and high monitoring false alarm rate caused by outlier monitoring values ​​in general sepsis patient monitoring systems, this solution defines a mixed distribution. While learning from historical monitoring data, it can make full use of medical prior knowledge and better adapt to the physiological changes of new patients; define a medical-oriented divergence distribution to enhance the early warning ability of extreme physiological states; introduce risk-sensitive weighted processing, automatically increase parameters when the current physiological state is not accurately grasped, and reduce parameters for stable patients , to prevent excessive alarms; thereby improving the accuracy of sepsis patient monitoring; in view of the problem that the general sepsis patient monitoring system cannot effectively handle the asymmetric risk of indicators deviating from the normal range, insufficient monitoring of extreme situations, and thus poor patient monitoring effect, this solution uses the normal reference value as the dividing point, and adopts exponential saturation curves for the high and low sides of the indicators respectively, and uses different saturation level parameters to achieve asymmetric punishment of risk, which can more accurately reflect the actual risk situation of sepsis patients when their physiological indicators deviate from the normal range; for dangerous signals below normal values, it can timely capture the deterioration of the patient's condition and improve the monitoring ability of extreme situations; thereby improving patient monitoring effect.

[0004] The technical solution adopted by the present invention is as follows: the sepsis patient monitoring system based on medical information processing provided by the present invention includes a data set construction module, a sepsis patient risk assessment model establishment module and a sepsis patient monitoring module;

[0005] The dataset construction module collects historical monitoring data and risk labels, and performs feature engineering to generate a sample set;

[0006] The sepsis patient risk assessment model establishment module is based on BiLSTM and mixed distribution; a risk-sensitive weighted loss function is introduced to establish a sepsis patient risk assessment model;

[0007] The sepsis patient monitoring module performs patient monitoring on real-time monitoring data based on a sepsis patient risk assessment model.

[0008] Furthermore, the dataset construction module collects historical sepsis patient monitoring data and risk assessment levels; uses the risk assessment levels as labels; and performs feature engineering on the sepsis patient monitoring data to obtain a patient sample set.

[0009] Furthermore, the sepsis patient risk assessment model establishment module divides the patient sample set into a training set and a test set to establish a sepsis patient risk assessment model; specifically, it includes the following contents:

[0010] Network architecture design unit; the front end uses a multi-layer bidirectional LSTM to extract temporal dependencies, expressed as: , the last layer output , and then take the attention summary as the fixed-length representation z; where, and are the bidirectional hidden states of the lth layer at time t and time t-1 respectively; It is the bidirectional LSTM of layer l; is the input vector of the lth layer at time t; H is the matrix of the hidden state of the top layer L at all times on the entire window, which is composed of time; and are the bidirectional hidden states of the Lth layer at time tT and time t respectively; the backend uses a fully connected layer to output the sepsis risk probability, which is expressed as: ;in, is the model's predicted probability of the patient's sepsis risk in the current window; is the activation function; and are the weights and biases of the fully connected layer respectively;

[0011] The mixture distribution defines the unit; the mixture distribution is expressed as: ;in, is a mixed distribution; is the variational approximate distribution, which means that under given parameters The model parameters under the condition distribution of is the prior distribution; is the blending weight; Reflects the network’s learning of historical monitoring data;

[0012] Medical-oriented divergence distribution definition unit; medical-oriented divergence distribution is expressed as: ;in, It is a medical-oriented divergence distribution; is the Rényi-β divergence; yes Divergence;

[0013] Loss function design unit; introduce risk-sensitive Weighted, the final loss function is designed as: ; ; Where L is the final loss function; is the loss weight; N is the total number of samples; i is the sample index; Is the input feature of the known i-th sample and model parameters , the sample label is probability; is the basic weight constant; is the sensitivity coefficient; is the variance of the overall forecast distribution;

[0014] Activation function design unit; take the normal reference value s as the dividing point, when x=s the output , risk-free increment; using exponential saturation curves on both sides of the indicator that are too high and too low, the activation function is: ; where a and c are saturation level parameters; b is the slope scale;

[0015] Model establishment; set the maximum number of training times, training convergence threshold, and prediction threshold; when the risk assessment of sepsis patients based on the training set loss converges or reaches the maximum number of training times, but the prediction of the test set does not meet the standard, adjust the parameters based on the particle swarm search algorithm and retrain; otherwise, the sepsis patient risk assessment model is established.

[0016] Furthermore, the sepsis patient monitoring module performs risk assessment on the real-time collected sepsis patient monitoring data based on the established sepsis patient risk assessment model, thereby realizing sepsis patient monitoring; when the output label is medium risk, the sampling frequency of the patient is increased, and when the output label is high risk, an early warning is issued.

[0017] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0018] (1) Aiming at the problems of weak risk capture capability, low early prediction accuracy, insufficient early warning for extreme physiological states, and high false alarm rate due to the influence of outlier monitoring values ​​in the general sepsis patient monitoring system, this scheme defines a mixed distribution. While learning from historical monitoring data, it can make full use of medical prior knowledge and better adapt to the physiological changes of new patients; define a medical-oriented divergence distribution to enhance the early warning capability for extreme physiological states; introduce risk-sensitive weighted processing, automatically increase parameters when the current physiological state is not accurately grasped, and reduce parameters for stable patients to prevent excessive alarms; thereby improving the accuracy of sepsis patient monitoring.

[0019] (2) In view of the problem that the general sepsis patient monitoring system cannot effectively handle the asymmetric risk of indicators deviating from the normal range, and insufficient monitoring of extreme situations, which leads to poor patient monitoring effects, this scheme uses the normal reference value as the demarcation point, and adopts exponential saturation curves for the high and low sides of the indicators respectively, and uses different saturation level parameters to achieve asymmetric punishment of risk, which can more accurately reflect the actual risk situation of sepsis patients when their physiological indicators deviate from the normal range; for dangerous signals below normal values, it can timely capture the deterioration of the patient's condition and improve the monitoring ability of extreme situations; thereby improving the patient monitoring effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic diagram of the process of the sepsis patient monitoring system based on medical information processing provided by the present invention;

[0021] Figure 2 Schematic diagram of the process for establishing modules for the risk assessment model for patients with sepsis.

[0022] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0024] In the description of the present application, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0025] Embodiment one, refer to Figure 1 The present application provides a sepsis patient monitoring system based on medical information processing, comprising a data set construction module, a sepsis patient risk assessment model establishment module and a sepsis patient monitoring module;

[0026] The data set construction module collects historical monitoring data and risk labels, and generates a sample set through feature engineering; and sends the data to the sepsis patient risk assessment model establishment module;

[0027] The sepsis patient risk assessment model establishment module is based on BiLSTM and mixed distribution; a risk-sensitive weighted loss function is introduced to establish a sepsis patient risk assessment model; and the data is sent to the sepsis patient monitoring module;

[0028] The sepsis patient monitoring module performs patient monitoring on real-time monitoring data based on the sepsis patient risk assessment model.

[0029] Embodiment two, refer to Figure 1 This embodiment is based on the above embodiment, the data set construction module collects historical sepsis patient monitoring data and risk assessment levels; the risk assessment levels are used as labels; the sepsis patient monitoring data is processed through feature engineering to obtain a patient sample set; the historical sepsis patient monitoring data includes blood pressure, heart rate, respiratory rate, blood oxygen saturation, body temperature, white blood cell count, lactic acid, creatinine, bilirubin, platelet count and C-reactive protein; the risk assessment levels include no risk, low risk, medium risk and high risk; a time window length T is taken, and sampling is performed once every hour; the feature engineering processing includes data conversion, data cleaning and standardization processing.

[0030] Embodiment three, refer to Figure 1 and Figure 2 This embodiment is based on the above embodiment, the sepsis patient risk assessment model establishment module divides the patient sample set into a training set and a test set to establish a sepsis patient risk assessment model; specifically including the following contents:

[0031] Network architecture design unit; the front end extracts time sequence dependence by multiple layers of bidirectional LSTM, represented as: , the output of the last layer is , and then take the attention summary as the fixed-length representation z; where, and are the bidirectional hidden states of the lth layer at time t and time t-1 respectively; It is the bidirectional LSTM of layer l; is the input vector of the lth layer at time t; H is the matrix of the hidden state of the top layer L at all times on the entire window, which is composed of time; and are the bidirectional hidden states of the Lth layer at time tT and time t respectively; the backend uses a fully connected layer to output the sepsis risk probability, which is expressed as: ;in, is the model's predicted probability of the patient's sepsis risk in the current window; is the activation function; and are the weights and biases of the fully connected layer respectively;

[0032] The mixture distribution defines the unit; the mixture distribution is expressed as: ;in, is a mixed distribution; is the variational approximate distribution, which means that under given parameters The model parameters under the condition distribution of is the prior distribution; is the mixing weight, ranging from [0.1, 0.9]; Reflects the network's learning of historical monitoring data; Adjustment, which preserves medical priors while allowing the model to flexibly adapt to the physiological changes of new patients;

[0033] Medical-oriented divergence distribution definition unit; medical-oriented divergence distribution is expressed as: ;in, It is a medical-oriented divergence distribution; is the Rényi-β divergence, β is used to adjust the sensitivity to the tail probability, and its value range is [0.5, 2.0]; yes Divergence emphasizes the matching of the prior and the low probability of the mixed distribution in the tail; it captures tail risk through the Rényi-β divergence, is extremely sensitive to the deviation of q in the low probability area of ​​the tail, and can quickly amplify the early warning signals in extreme physiological states such as a sharp drop in blood pressure and an inflammatory storm; and Divergence in When the denominator is very small, the penalty strength is increased to ensure that when the patient's indicators have outrageous values, the posterior will not be pushed into a dangerous area that deviates from the medical prior due to data noise;

[0034] Loss function design unit; introduce risk-sensitive Weighted, the final loss function is designed as: ; ; Where L is the final loss function; is the loss weight; N is the total number of samples; i is the sample index; Is the input feature of the known i-th sample and model parameters , the sample label is probability; is the basic weight constant, ranging from [0.1, 1.0]; is the sensitivity coefficient, ranging from [0.1, 1.0]; is the variance of the overall predictive distribution.

[0035] By performing the above operations, we can address the problems of weak risk capture capability, low early prediction accuracy, insufficient early warning for extreme physiological states, and high false alarm rate caused by outlier monitoring values ​​in general sepsis patient monitoring systems. By defining a mixed distribution, this solution can fully utilize medical prior knowledge while learning from historical monitoring data to better adapt to the physiological changes of new patients; define a medically guided divergence distribution to enhance the early warning capability for extreme physiological states; introduce risk-sensitive weighted processing, automatically increase parameters when the current physiological state is not accurately grasped, and reduce parameters for stable patients to prevent excessive alarms, thereby improving the accuracy of sepsis patient monitoring.

[0036] Example 4, see Figure 1 and Figure 2 This embodiment is based on the above embodiment, and the sepsis patient risk assessment model establishment module further includes:

[0037] Activation function design unit; take the normal reference value s as the dividing point, when x=s the output , risk-free increment; using exponential saturation curves on both sides of the indicator that are too high and too low, the activation function is: Among them, a and c are saturation level parameters with a value range of [0.5, 3.0], which realize asymmetric penalty of risk; b is the slope scale with a value range of [0.1, 10.0], which determines how much deviation is considered a significant abnormality; Gradient vanishing mitigation: Compared with the traditional ReLU activation function, the negative interval is not always 0, but has a negative output. When the patient's indicators drop to extremely low values ​​in a short period of time during the shock period, effective gradients can still be obtained instead of completely losing their voice, and dangerous signals below normal values ​​are not missed, which helps to quickly converge to the correct high-risk prediction; the shift parameter s aligns the maximum slope with the true clinical threshold, making the inflection point of the activation function closer to the key feature interval, thereby retaining stronger gradient sensitivity at the disease classification boundary, which is conducive to capturing minority sample features in unbalanced scenarios;

[0038] Model establishment; set the maximum number of training times, the value range is [50,200], the training convergence threshold, the value range is [10 -4 ,10 -2 ] and prediction threshold, with a value range of [0.8, 0.95]; when the risk assessment of sepsis patients based on the training set loss converges or reaches the maximum number of training times, but the prediction of the test set does not meet the standard, the particle swarm search algorithm is used to adjust the parameters and retrain; otherwise, the risk assessment model for sepsis patients is established.

[0039] By performing the above operations, we address the problem that general sepsis patient monitoring systems are unable to effectively handle the asymmetric risk of indicators deviating from the normal range, insufficiently monitor extreme situations, and thus lead to poor patient monitoring results. This solution uses the normal reference value as the demarcation point, and adopts exponential saturation curves for both the high and low sides of the indicators, using different saturation level parameters to achieve asymmetric penalty for risk. This can more accurately reflect the actual risk situation of sepsis patients when their physiological indicators deviate from the normal range; for dangerous signals below normal values, it can promptly capture the deterioration of the patient's condition, improve the monitoring ability for extreme situations, and thus improve patient monitoring results.

[0040] Example 6, see Figure 1 This embodiment is based on the above embodiment. The sepsis patient monitoring module performs risk assessment on the real-time collected sepsis patient monitoring data based on the established sepsis patient risk assessment model, thereby realizing sepsis patient monitoring. When the output label is medium risk, the sampling frequency of the patient is increased, and when the output label is high risk, an early warning is issued.

[0041] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0042] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0043] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. A sepsis patient monitoring system based on medical information processing, characterized by: The system includes a data set construction module, a sepsis patient risk assessment model establishment module, and a sepsis patient monitoring module; The dataset construction module collects historical monitoring data and risk labels, and performs feature engineering to generate a sample set; The sepsis patient risk assessment model establishment module is based on BiLSTM and mixed distribution; a risk-sensitive weighted loss function is introduced to establish a sepsis patient risk assessment model; The sepsis patient monitoring module performs patient monitoring on real-time monitoring data based on a sepsis patient risk assessment model; The sepsis patient risk assessment model establishment module divides the patient sample set into a training set and a test set to establish a sepsis patient risk assessment model; specifically, it includes the following contents: Network architecture design unit; the front end uses a multi-layer bidirectional LSTM to extract temporal dependencies, expressed as: , the last layer output , and then take the attention summary as the fixed-length representation z; where, and are the bidirectional hidden states of the lth layer at time t and time t-1 respectively; It is the bidirectional LSTM of layer l; is the input vector of the lth layer at time t; H is the matrix of the hidden state of the top layer L at all times on the entire window, which is composed of time; and are the bidirectional hidden states of the Lth layer at time tT and time t respectively; the backend uses a fully connected layer to output the sepsis risk probability, which is expressed as: ;in, is the model's predicted probability of the patient's sepsis risk in the current window; is the activation function; and are the weights and biases of the fully connected layer respectively; The mixture distribution defines the unit; the mixture distribution is expressed as: ;in, is a mixed distribution; is the variational approximate distribution, which means that under given parameters The model parameters under the condition distribution of is the prior distribution; is the blending weight; Reflects the network’s learning of historical monitoring data; Medical-oriented divergence distribution definition unit; medical-oriented divergence distribution is expressed as: ;in, It is a medical-oriented divergence distribution; is the Rényi-β divergence; yes Divergence; Loss function design unit; introduce risk-sensitive Weighted, the final loss function is designed as: ; ; Where L is the final loss function; is the loss weight; N is the total number of samples; i is the sample index; Is the input feature of the known i-th sample and model parameters , the sample label is probability; is the basic weight constant; is the sensitivity coefficient; is the variance of the overall forecast distribution; Activation function design unit; take the normal reference value s as the dividing point, when x=s the output , risk-free increment; using exponential saturation curves on both sides of the indicator that are too high and too low, the activation function is: ; where a and c are saturation level parameters; b is the slope scale.

2. The sepsis patient monitoring system based on medical information processing according to claim 1, characterized in that: The data set construction module collects historical sepsis patient monitoring data and risk assessment levels; uses the risk assessment levels as labels; and performs feature engineering on the sepsis patient monitoring data to obtain a patient sample set.

3. The sepsis patient monitoring system based on medical information processing according to claim 2, characterized in that: The sepsis patient monitoring module performs risk assessment on the real-time collected sepsis patient monitoring data based on the established sepsis patient risk assessment model, thereby realizing sepsis patient monitoring; when the output label is medium risk, the sampling frequency of the patient is increased, and when the output label is high risk, an early warning is issued.

Images