Critical patient enteral nutrition diagnosis and treatment method based on artificial intelligence, computer equipment and medium

Abstract

The invention discloses a severe patient enteral nutrition diagnosis and treatment method based on artificial intelligence, computer equipment and a medium. A severe patient enteral nutrition diagnosis and treatment method based on artificial intelligence comprises the steps that static variables and dynamic variables of a patient are obtained; obtaining inspection indexes, wherein the inspection indexes comprise a nutrition state index and a disease specificity index; on the basis of an intelligent prescription model formed after BP neural network pre-training, according to input patient static variables, dynamic variables, metabolic energy percentages and inspection indexes, automatically calculating enteral nutrition energy and nutrient proportions thereof; and according to the enteral nutrition energy and the nutrient ratio thereof, outputting an enteral nutrition prescription formed by combination of grams of specialty medical food and / or cooked food, and outputting a nutrition diagnosis and treatment prediction result at the same time. According to the invention, personalized nutrition prescriptions suitable for patients can be intelligently given, the intelligence of the nutrition prescriptions is greatly improved, and the system is more scientific and better meets the nutritional needs.

Description

Technical Field

[0001] The present invention relates to the technical field of nutrition prescription formation, and particularly to an enteral nutrition diagnosis and treatment method for critically ill patients based on artificial intelligence, a computer device, and a medium. Background Technique

[0002] Doing a good job in all the basic work that directly affects the curative effect of critically ill patients is the central task of medical institutions. The diagnosis and treatment ability of critically ill patients is often a sign reflecting the overall medical level of medical institutions, that is, there should be no short board in each basic diagnosis and treatment link of critically ill patients, including the nutrition diagnosis and treatment of critically ill patients. In medical institutions, the nutrition of critically ill patients is the symbol and core of clinical nutrition, because the condition of critically ill patients is highly dangerous and complex, and extremely high requirements are imposed on how to grasp the supply of nutrients and energy components.

[0003] There are problems in the current situation of nutrition diagnosis and treatment in domestic medical institutions, such as the lack of professional talents, the low level of work specialization, and the insufficient application of clinical nutrition. It is also an objective fact that the nutrition department has not entered the intensive care unit for a long time, and the clinical nutrition work in the intensive care unit is not its main business, which naturally forms the situation that the clinical nutrition of patients in the intensive care unit is in the "wooden barrel short board" state, that is, the nutrition department is unable to handle it, and the intensive care unit has no energy to handle clinical nutrition. Therefore, whether it is the nutrition department for talent cultivation or the intensive care unit for clinical nutrition training, as well as the integration of disciplines to change the existing clinical nutrition work mode, it will require a large amount of investment, policy support, and a long process.

[0004] For the enteral nutrition of patients, it is not only based on the supply of energy, but more importantly, it provides food for the intestinal flora to prevent it from changing in quantity, flora, and location and causing more serious consequences. As one of the most important human immune organs, the stability of the intestinal flora plays a key role in the diagnosis and treatment process of diseases. Therefore, providing scientific enteral nutrition that conforms to the laws of human body operation is of extremely important significance for the rehabilitation of patients, especially critically ill patients.

[0005] Artificial intelligence (AI), as the core of the Fourth Industrial Revolution, is flourishing in multiple fields. In medicine, AI applications cover various aspects, including medical image analysis, assisted diagnosis, drug development, health management, and telemedicine. Medical data, as the supporting driving force of AI, is of paramount importance. Recently, to encourage the development of medical AI, the government has gradually introduced a series of policies related to medical data, allowing for the trading of medical data under the premise of securely filtering personal information. Specifically, the trading of medical data in public hospitals has been implemented, setting a benchmark for the compliant application of medical data nationwide. Vast medical data is composed of multiple sub-fields, and AI is also composed of each sub-field or aspect. Regarding causal relationships in the medical diagnosis and treatment process, AI's autonomous learning algorithms analyze the data to discover the characteristics and patterns of causal relationships, providing precise and specific auxiliary references for clinical medical decision-making.

[0006] Clinical nutrition is less complex and risky than clinical medicine, but its tedious energy calculations are time-consuming. Based on this characteristic and the development of current technology, the introduction of artificial intelligence into the field of nutrition prescription can hopefully eliminate this "weak link" in clinical nutrition. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings and defects of existing technologies and provide an artificial intelligence-based method, computer equipment, and medium for the diagnosis and treatment of enteral nutrition in critically ill patients. The method of this invention selects a nonlinear correlation deep learning algorithm, which is trained with specific medical data input. This allows the data used by the artificial intelligence algorithm to continuously improve. Through model building, autonomous learning and training, and repeated iterations and improvements, the individualized characteristics and patterns of enteral nutrition in critically ill patients with improved conditions under different circumstances are ultimately obtained, providing a scientific platform for the formation of clinical nutrition prescriptions.

[0008] One objective of this invention is to provide an artificial intelligence-based method for the diagnosis and treatment of enteral nutrition in critically ill patients, comprising the following steps:

[0009] Obtain the patient's static and dynamic variables;

[0010] Obtain testing indicators, including nutritional status indicators and disease-specific indicators;

[0011] The intelligent prescription model, which is formed after pre-training of a BP neural network, automatically calculates enteral nutrition energy and its nutrient ratio based on the input static variables, dynamic variables, metabolic energy percentage, and test indicators of the patient.

[0012] Based on the enteral nutrition energy and its nutrient ratio, an enteral nutrition prescription consisting of a combination of grams of a special medical food and / or cooked food is output, along with nutritional diagnosis and treatment prediction results.

[0013] The static variables include the patient's age, gender, and height; the dynamic variables include the patient's weight.

[0014] The percentage of metabolic energy is determined based on the calculated basal metabolic rate multiplied by a preset percentage range according to the patient's condition. The basal metabolic rate is calculated based on the Harris-Benedict formula multiplied by a preset weight ratio.

[0015] The nutritional status indicators are used to assess the overall nutritional status of patients, including prealbumin, albumin, hemoglobin, and lymphocyte count. The disease-specific indicators are specific physiological indicators reflecting disease-related factors, including blood ammonia, serum creatinine, blood glucose, and C-reactive protein.

[0016] The test indicators are input into the intelligent prescription model after standardization. The standardization process involves unifying the data scale using the Z-score standardization formula to eliminate the magnitude differences between different indicators.

[0017] The BP neural network consists of an input layer, a hidden layer, and an output layer. The hidden layer uses the Sigmoid function to fit the nonlinear relationship. The output layer uses the Linear activation function to perform a regression task and outputs the predicted values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional diagnosis and treatment. The output layer uses the Softmax activation function to perform a classification task and outputs the probability distribution of the patient's nutritional risk level.

[0018] In the BP neural network, when performing regression tasks, the output layer uses mean squared error (MSE) to measure the deviation between the predicted and actual values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional treatment; when performing classification tasks, it uses cross-entropy loss function to measure the difference between the predicted probability distribution of patient nutritional risk level and the actual label.

[0019] During training, the BP neural network employs the Adam optimizer; the error of the output layer is propagated to the input layer layer by layer through the backpropagation algorithm, and the weights of each layer are updated using the gradient descent method to continuously reduce the loss function value and optimize the model parameters.

[0020] A second objective of this invention is to provide a computer device, including a memory and a processor, wherein the memory stores executable code; characterized in that, when the processor executes the executable code, it is used to implement the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients.

[0021] A third objective of this invention is to provide a computer storage medium on which a program is stored; when the program is executed by a processor, it is used to implement the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients.

[0022] The present invention provides an artificial intelligence-based enteral nutrition treatment method for critically ill patients. This method processes input data through an intelligent prescription model trained on a neural network. Based on the analysis of individualized characteristics and patterns of enteral nutrition in critically ill patients under different conditions as their condition improves, including the theoretical and actual energy ratio, the parenteral-to-enteral ratio, the proportion of the three major enteral nutrients, whether and how much of the usual food is added, the duration of individualized nutritional treatment, and the correlation between changes in disease and nutritional indicators and enteral nutrition, the method intelligently provides nutritional prescriptions adapted to different patients. This significantly improves the intelligence of nutritional prescriptions, making them more scientific and better suited to nutritional needs. Attached Figure Description

[0023] Figure 1 This is a flowchart of an artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to an embodiment of the present invention.

[0024] Figure 2 This is a schematic diagram of the network model of the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to an embodiment of the present invention.

[0025] Figure 3 This is a schematic diagram of the computer device according to an embodiment of the present invention. Detailed Implementation

[0026] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0027] An exemplary embodiment of this application utilizes acquired clinical nutrition diagnosis and treatment data, including general patient information, nutritional status indicators, and disease-specific indicators. A model is established using a backpropagation (BP) neural network, and nonlinear supervised learning and training are performed on the clinical nutrition diagnosis and treatment data. Based on the training, the model generates correlation data between individualized clinical diagnosis and treatment and nutritional prescriptions for different diseases, disease stages, ages, and genders. Based on the input data, intelligent computational processing is used to generate nutritional prescriptions for clinical enteral nutrition diagnosis and treatment and predictions of changes in relevant indicators over n days of using the enteral (EN) nutrition prescription. This provides an important and practical reference for efficient clinical nutrition diagnosis and treatment.

[0028] See Figure 1 As shown in the exemplary embodiment of this application, the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients includes the following steps:

[0029] S1. Obtain the patient's static and dynamic variables;

[0030] S2. Obtain the test indicators, which include nutritional status indicators and disease-specific indicators;

[0031] S3. The intelligent prescription model, based on the pre-trained BP (Back Propagation) neural network, automatically calculates enteral nutrition energy and its nutrient ratio according to the input static variables, dynamic variables, basal metabolic energy percentage and test indicators of the patient.

[0032] S4. Based on the enteral nutrition energy and its nutrient ratio, output an enteral nutrition prescription consisting of a combination of grams of special medical foods and / or cooked foods, and simultaneously output nutritional diagnosis and treatment prediction results.

[0033] In this embodiment, a BP neural network is used to establish an intelligent prescription model, which is then subjected to nonlinear supervised learning and training. The obtained medical data is used to train the model to obtain data on the correlation between individualized clinical treatment plans and nutritional prescriptions for different diseases, disease stages, ages, and genders. This provides an important and practical reference for efficient clinical nutritional diagnosis and treatment.

[0034] In this embodiment of the application, the nutritional diagnosis and treatment prediction results can be given as a standard based on the median of the normal values ​​of nutritional status indicators, disease-specific indicators, and changes in weight, in order to measure the effectiveness or effect of the output enteral nutrition diagnosis and treatment prescription.

[0035] In this embodiment, the percentage of basal metabolism can be dynamically adjusted by clinicians and nutritionists based on the patient's real-time status or by an intelligent prescription model based on the input basal metabolic value, that is, to determine the ratio of the theoretical energy value to the actual energy value of basal metabolism; generally, the actual energy value of basal metabolism is usually 80%-120% of its theoretical value.

[0036] In this embodiment of the application, the intelligent prescription model automatically processes the input data, and based on the model algorithm, outputs the energy required for enteral nutrition (EN) and the proportion of its nutrients according to the determined basal metabolic rate.

[0037] In this embodiment of the application, the static variables include unchanging physiological parameters such as the patient's age, gender, and height (taking adults as an example, excluding minors whose height changes); the dynamic variables include the patient's weight, which changes dynamically depending on the situation.

[0038] In this embodiment, the metabolic energy percentage is determined based on the calculated basal metabolic rate (BMR) multiplied by a preset percentage range according to the patient's condition. The BMR is calculated using the Harris-Benedict formula and multiplied by a preset weighting ratio. This preset weighting ratio can be designed based on experience or suited to the needs of Chinese people, such as being set to 87.5%. According to the Harris-Benedict formula, based on the patient's name, height, weight, and age, the calculated BMR for males is 66.4730 + 13.751 (kg) + 5.0033 (cm) - 6.7550 (years) = kcal. The calculated BMR for females is 655.0955 + 9.463 (kg) + 1.8496 (cm) - 4.6755 (years) = kcal. Considering racial differences, a preset weighting of 87.5% is selected, which can be used as the BMR for Chinese people. The percentage of metabolic energy can be adjusted by the doctor based on the patient's or the patient's individual physical condition, within the range of 80%-120% of the basal metabolic value. The remaining EN energy value, which is calculated by subtracting the energy value from intravenous infusion and snacks from the basal metabolic value, is determined by the doctor and nutritionist, and then input into the intelligent prescription model. The intelligent prescription model processes the data according to this ratio and outputs the processing result.

[0039] In this embodiment, the nutritional status indicators are indicators used to assess the overall nutritional status of the patient, and may include prealbumin, albumin, hemoglobin, lymphocyte count, total white blood cell count, platelet count, etc. The specific implementation is not limited to these, and may include more indicators. The disease-specific indicators are specific physiological indicators reflecting disease-related factors, including blood ammonia, serum creatinine, blood urea nitrogen (BUN), blood potassium, blood sodium, blood glucose, blood lipids, C-reactive protein, alanine aminotransferase (ALT), total bilirubin (Bil), prothrombin time (PT), alkaline phosphatase (AKP), cholinesterase (Che), 24-hour urine protein quantification, creatine kinase (CK), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), etc. The specific implementation is not limited to these, and may include more indicators.

[0040] In this embodiment of the application, the test index is input into the intelligent prescription model after standardization. The standardization process is to unify the data scale through the Z-score standardization formula to eliminate the difference in magnitude between different indicators.

[0041] The Z-score standardization formula is as follows:

[0042] z = x - μ / σ, where μ is the median, i.e. the median of the normal reference range of the indicator, and σ is the standard deviation of the reference range.

[0043] See Figure 2 As shown in the embodiment of this application, the BP neural network consists of an input layer, a hidden layer, and an output layer; the hidden layer uses the Sigmoid function to fit the nonlinear relationship; the number of neurons in the hidden layer is determined by cross-validation, and it is recommended that the initial value be 1.5 times the number of nodes in the input layer (e.g., 38).

[0044] In the BP neural network, the output layer uses a Linear activation function to perform a regression task, outputting predicted values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional treatment. The output layer uses a Softmax activation function to perform a classification task, outputting the probability distribution of the patient's nutritional risk level. The predicted values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional treatment, along with the probability distribution of the patient's nutritional risk level, constitute the nutritional treatment prediction result. This prediction result can be the result ≥7 days after the patient completes clinical nutritional treatment; this number of days can be set as needed.

[0045] In this embodiment of the application, when performing regression tasks, the output layer of the BP neural network can use mean squared error (MSE) to measure the deviation between the predicted and actual values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional diagnosis and treatment; when performing classification tasks, it can use the cross-entropy loss function to measure the difference between the predicted probability distribution of the patient's nutritional risk level and the actual label.

[0046] For a dataset containing n samples, let the true value be y. i The corresponding predicted value is (where i = 1, 2, ..., n), then the formula for calculating MSE is:

[0047]

[0048] The cross-entropy loss function is as follows:

[0049] n represents the total number of categories, y i It is a unique hot encoding of the real label. The model's predicted probability for class i.

[0050] In this embodiment, the BP neural network can employ the Adam optimizer during training with an initial learning rate of 0.001. The Adam optimizer combines momentum and adaptive learning rate adjustment strategies, resulting in fast convergence and high stability. The error from the output layer is propagated layer by layer to the input layer via backpropagation, and the weights of each layer are updated using gradient descent to continuously reduce the loss function value, thereby optimizing the model parameters. Training stops when the error drops to a threshold or the maximum number of iterations is reached, yielding the desired intelligent prescription model. For example, training stops when the training set loss function value drops below 0.001 or the number of iterations reaches a preset maximum value (e.g., 1000 times) to avoid overfitting or undertraining.

[0051] In the training phase of the BP neural network, the error signal, representing the difference between the expected and actual output information, is propagated back through the output layer and through the hidden layers, adjusting the connection weights layer by layer. This process, known as "learning convergence," involves repeated "pattern forward propagation" and "error back propagation" to achieve network "memory training," causing the overall network error to approach a minimum. During training, each weight is adjusted to minimize the training set error rate until a specific condition—the optimal clinical nutrition prescription—is met. The termination condition can be the minimum total error, a specific time frame, or the maximum number of repetitions. This cyclical process of forward and backward propagation of information is a continuous adjustment of the weights at each layer, and also a learning and training process for the neural network. This process continues until the network output error is reduced to an acceptable level or the predetermined number of learning iterations is reached.

[0052] During the training of the BP neural network, the collected data can be divided into training set, test set, and validation set proportionally, such as 70% training set, 15% validation set, and 15% test set. The training set is used for learning model parameters, the validation set is used to adjust the model's hyperparameters (such as the number of hidden layer neurons), and the test set is used to evaluate the model's final generalization ability.

[0053] When training the BP neural network, data preprocessing is a crucial step in ensuring the accuracy of the algorithm. Specifically, this involves transforming the raw data into a numerical format suitable for model training through operations such as standardization and normalization. For example, normalization maps the numerical data to the [0,1] interval, eliminating the influence of differences in units (such as units or magnitudes) between different data points, thus ensuring a balanced weighting of each data point during model training. Normalization formula: X norm =XX min / X max -X minFor example, if the reference range for an indicator is 20-100, and a patient's test value is 60, then the normalized value is (60-20) / (100-20) = 0.5. For percentages in the data, the percentage can be directly divided by 100 to convert it to a decimal (e.g., 50% is converted to 0.5), avoiding interference from differences in magnitude (e.g., 50% vs. 0.5) in percentage-form numerical values, thus unifying the data format.

[0054] The data processing process also includes data cleaning, checking data integrity, and removing missing or outlier values ​​(such as test values ​​that significantly exceed the clinically reasonable range); ensuring that all data are in numerical format to avoid calculation errors caused by text format. After the data is processed, a random sample of data is manually verified to ensure the accuracy of the standardized and normalized results.

[0055] A BP (Back Propagation) neural network is a multi-layer feedforward neural network trained using an error backpropagation algorithm. It learns nonlinear relationships by simulating the connection patterns of neurons in the human brain. The BP neural network design in this embodiment is as follows:

[0056] Its input layer includes multiple selected dynamic indicators, including body weight, nutritional status indicators (prealbumin, albumin, etc.), disease-specific indicators (blood ammonia, blood glucose, etc.), and metabolic percentage (the ratio of theoretical energy to actual energy metabolized).

[0057] In this embodiment, since static variables have little impact on the dynamic adjustment of the nutrition plan, static variables such as age and gender are excluded in the input layer. The focus is mainly on dynamic indicators that can reflect changes in the patient's condition in real time, thereby improving the model's sensitivity to changes in the condition.

[0058] The optimal number of neurons in the hidden layer was determined using a 5-fold cross-validation method. The initial suggested value was 1.5 times the number of input layer nodes (i.e., 25 × 1.5 = 38, where 25 represents the number of input metrics; this can be adjusted as needed). During cross-validation, the configuration with the smallest error was selected by comparing the validation set error of the model with different numbers of neurons. The activation function used was the Sigmoid function, with the formula f(x) = 1 / (1 + e^(-x)). -x The Sigmoid function maps input values ​​to the (0,1) interval, enabling the fitting of nonlinear relationships and enhancing the model's ability to learn associations of complex nutritional data.

[0059] When performing regression tasks, its output layer uses the Linear activation function to directly predict nutritional status indicators (such as albumin and hemoglobin), disease-specific indicators, and predicted weight values ​​to assess the trend of nutritional regimens improving patients' nutritional status. When performing classification tasks, it uses the Softmax activation function to output the probability distribution of patients' nutritional risk levels (such as low risk, medium risk, and high risk) to assist in clinical decision-making.

[0060] The inventors of this application acknowledge the significant lag of clinical nutrition behind clinical medicine, even impacting its development. This invention is the first to apply artificial intelligence (AI) technology to clinical nutrition diagnosis and treatment, expanding the application scope and scenarios of AI in the medical field. It allows AI to fully leverage its efficiency, precision, and speed in clinical nutrition diagnosis and treatment, fundamentally solving the prevalent problems of talent shortage, insufficient personnel, low skill levels, and long-term losses in the clinical nutrition workforce. With the continuous accumulation and sharing of data, AI's machine learning continues to deepen and evolve, enabling AI to replace manual labor throughout the entire clinical nutrition diagnosis and treatment process. This makes AI-based nutrition diagnosis and treatment based on causal relationships in clinical nutrition diagnosis and treatment feasible, including AI's highly auxiliary functions and even its ability to automatically generate and administer nutritional prescriptions, making it possible.

[0061] The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients in this application utilizes tradable clinical medical big data, selects a nonlinear correlation artificial neural network algorithm, and analyzes the causal relationship and patterns between changes in the clinical and nutritional objective indicators of critically ill patients and the specific proportion of their individualized enteral and parenteral nutrition supply, as well as the composition of their enteral nutrition prescriptions, through model building and autonomous learning and training. This provides more scientific nutritional prescriptions and prediction results, paving the way for the comprehensive artificial intelligence of clinical nutrition, fundamentally changing the unbalanced development of clinical medicine and clinical nutrition, and enabling rational clinical nutrition to universally benefit critically ill patients.

[0062] The AI-based enteral nutrition treatment method for critically ill patients in this application is the first of its kind to create a novel nutritional treatment technology model that universally meets the individualized clinical nutritional needs of critically ill patients. It enables precise calculation and application of AI-based individualized enteral nutrition treatment. This AI-based enteral nutrition treatment method for critically ill patients addresses the awkward situation where insufficient capacity in nutrition departments and insufficient manpower in critical care medicine departments have made individualized enteral nutrition for critically ill patients a weak link in clinical medicine. AI-based individualized enteral nutrition treatment technology can fully accomplish the nutritional treatment work that previously required a large amount of manpower in a universal, accurate, rapid, and low-cost manner. With the help of massive amounts of domestic clinical nutrition data, the AI-based enteral nutrition treatment method for critically ill patients in this application makes it possible to formulate relevant clinical nutrition standards for Chinese patients, breaking the current situation where all nutrition-related standards are formulated by foreign countries.

[0063] In exemplary embodiments, a computer device is also provided, intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The computer device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0064] In an exemplary embodiment, reference is made to Figure 3 As shown, a computer device includes a processor and memory. The processor may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor can be implemented using at least one of the following hardware forms: DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array).

[0065] In some embodiments, the processor may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake-up state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state.

[0066] In some embodiments, the processor may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content that the display screen needs to show. In some embodiments, the processor may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0067] In some embodiments, the readable storage medium in the memory may be tangible and non-transitory. The memory may also include high-speed random access memory and non-volatile memory, such as one or more disk storage terminals or flash memory terminals. In some embodiments, the non-transitory computer-readable storage medium in the memory is used to store at least one instruction, which is executed by a processor to implement the AI-based enteral nutrition diagnosis and treatment method for critically ill patients provided in this application.

[0068] In some embodiments, multiple components of the computer device are connected to an I / O interface, including: input units such as a keyboard and mouse; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication units allow the electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0069] In an exemplary embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients as provided in all embodiments of this application.

[0070] Any combination of one or more computer-readable media can be used. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example—but not limited to—electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, and portable compact disk read-only memory. Optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0071] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0072] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0073] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0074] In an exemplary embodiment, an application product is also provided, including one or more instructions that can be executed by the processor of the aforementioned device to complete the aforementioned artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients.

[0075] Various embodiments of the systems and technologies described above in this application can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0076] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0077] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, and portable compact disk read-only memory. Optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0078] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer device having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0079] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0080] Computing systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. This is achieved by having clients running on corresponding computers and interacting with each other. Computer programs that establish server-client relationships use these relationships to create connections between clients and servers. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service ecosystem. It addresses the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0081] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0082] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention.

[0083] Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of the equivalents of the claims be included within the invention.

[0084] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients, characterized in that, Includes the following steps: Obtain the patient's static and dynamic variables; Obtain testing indicators, including nutritional status indicators and disease-specific indicators; The intelligent prescription model, which is formed after pre-training of a BP neural network, automatically calculates enteral nutrition energy and its nutrient ratio based on the input static variables, dynamic variables, metabolic energy percentage, and test indicators of the patient. Based on the enteral nutrition energy and its nutrient ratio, an enteral nutrition prescription consisting of a combination of grams of a special medical food and / or cooked food is output, along with nutritional diagnosis and treatment prediction results.

2. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 1, characterized in that, The static variables include the patient's age, gender, and height; the dynamic variables include the patient's weight.

3. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 1, characterized in that, The percentage of metabolic energy is determined based on the calculated basal metabolic rate multiplied by a preset percentage range according to the patient's condition. The basal metabolic rate is calculated based on the Harris-Benedict formula multiplied by a preset weighting ratio.

4. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 1, characterized in that, The nutritional status indicators are used to assess the patient's overall nutritional status, including prealbumin, albumin, hemoglobin, and lymphocyte count. The disease-specific indicators are specific physiological indicators reflecting disease-related factors, including blood ammonia, serum creatinine, blood glucose, and C-reactive protein.

5. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 1, characterized in that, The test indicators are input into the intelligent prescription model after standardization. The standardization process involves unifying the data scale using the Z-score standardization formula to eliminate the magnitude differences between different indicators.

6. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 1, characterized in that, The BP neural network consists of an input layer, a hidden layer, and an output layer. The hidden layer uses the Sigmoid function to fit the nonlinear relationship. The output layer uses the Linear activation function to perform a regression task, outputting the predicted values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional diagnosis and treatment. The output layer uses the Softmax activation function to perform a classification task, outputting the probability distribution of the patient's nutritional risk level.

7. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 6, characterized in that, When performing regression tasks, the output layer of the BP neural network uses mean squared error (MSE) to measure the deviation between the predicted and actual values ​​of nutritional status indicators, disease-specific indicators, and weight after nutritional treatment; when performing classification tasks, it uses cross-entropy loss function to measure the difference between the predicted probability distribution of patient nutritional risk level and the actual label.

8. The artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients according to claim 1, characterized in that, During the training of the BP neural network, the Adam optimizer is used; the error of the output layer is passed to the input layer layer by layer through the backpropagation algorithm, and the weights of each layer are updated by the gradient descent method to continuously reduce the loss function value and optimize the model parameters.

9. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores executable code; characterized in that, when the processor executes the executable code, it is used to implement the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients as described in any one of claims 1 to 8.

10. A computer storage medium having a program stored thereon; characterized in that, When the program is executed by the processor, it is used to implement the artificial intelligence-based enteral nutrition diagnosis and treatment method for critically ill patients as described in any one of claims 1 to 8.

Images