A nasogastric feeding care system and method based on a recommendation model

Abstract

This invention relates to the field of nasogastric feeding technology, specifically disclosing a nasogastric feeding system and method based on a recommendation model. This system addresses the challenge of identifying the current feeding goals based on the real-time status of the food during the liquid feeding process, and then collaboratively determining the crushing, venting, and output strategies. This achieves coordinated flow control and degassing, reducing the risks of tube blockage, bloating, and gastroesophageal reflux, and assisting inexperienced caregivers. The system includes a housing, end caps, drive components, pressure plates, mesh plates, venting components, output components, status acquisition components, display components, a memory, and a processor. This invention introduces a recommendation mechanism based on nasogastric food status data to identify feeding goals and coordinate the crushing, venting, and output. This transforms the separate processing of liquid food refinement and air removal into a coordinated process, improving the uniformity and stability of the liquid feeding, and reducing the risks of small tube blockage, gastric bloating, and gastroesophageal reflux.

Description

Technical Field

[0001] This invention relates to the field of nasogastric feeding care technology, and in particular to a nasogastric feeding care system and method based on a recommendation model. Background Technology

[0002] Nasogastric feeding is a method used when patients are unable to swallow independently or are not getting enough food orally. A gastric tube is placed through the nasal cavity into the esophagus and stomach, and water and food are then injected into the patient's stomach in liquid form through the gastric tube to maintain metabolism, weight, and nutrition. Although thin gastric tubes cause less damage, they are more prone to problems such as poor injection, fluctuating flow, and bloating and reflux in the patient's stomach due to larger liquid particles, uneven viscosity, or too many air bubbles. Chinese invention patent CN116725876B discloses a nasogastric feeding device for gastrointestinal care that prevents backflow. It uses a pressure plate, mesh plate, and drive rod to pulverize food, and then uses an air pump and four-way tube to remove air from the casing before feeding. While this improves the liquid food processing and feeding process to some extent, the pulverization of food and the removal of excess air are still separate processes. It lacks a mechanism to identify the priority feeding goals based on the real-time state of the food during pulverization, and to determine pulverization, air removal, and output conditions based on these goals to assist caregivers lacking nasogastric feeding experience. Furthermore, in actual nasogastric feeding, food is repeatedly squeezed, sheared, and processed between the pressure plate and mesh plate. During the refining process, the particle size, slurry viscosity, and entrained air bubble state of the food are constantly changing dynamically. While the food is being refined into a liquid suitable for nasogastric feeding, air is continuously being entrained and new air bubbles are being released. If the crushing and degassing are set as separate steps, it is difficult to reflect the constantly changing nursing risk status during the crushing process in a timely manner. This can easily lead to the premature output of refined but still high-air-content liquid food, increasing the risk of gastric bloating and gastroesophageal reflux. Alternatively, liquid food that has been degassed but is not sufficiently refined may enter the narrow gastric tube, making it difficult to achieve the nursing goal of smooth feeding. Therefore, the existing technology cannot simultaneously identify the current nasogastric feeding nursing goal information during the process of food crushing to form a liquid food, and accordingly achieve the coordinated and simultaneous crushing and degassing of excess air to assist inexperienced nasogastric feeding caregivers. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a nasogastric feeding system and method based on a recommendation model. By introducing a recommendation mechanism that identifies nasogastric feeding care goals based on nasogastric food status data and links and controls crushing, air expulsion and output, the process of refining liquid food and removing air is changed from separate processing to coordinated processing, thereby improving the uniformity and stability of liquid food supply, reducing the risk of blockage of the nasogastric tube, gastric bloating and gastroesophageal reflux, and improving the operational accuracy of nursing staff.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] In a first aspect, the present invention provides a nasogastric feeding system based on a recommendation model, comprising a housing, an end cap, a drive assembly, a pressure plate, a mesh plate, an exhaust assembly, an output assembly, a status acquisition assembly, a display assembly, a memory, and a processor; a food processing chamber is formed within the housing, the end cap is detachably disposed at the upper end of the housing, the pressure plate and the mesh plate are disposed within the food processing chamber and move relative to each other under the action of the drive assembly to perform squeezing, shearing, and mesh-crushing processing on the food, the exhaust assembly is connected to the food processing chamber, and the output assembly is connected to the food processing chamber; the status acquisition assembly is used to acquire nasogastric feeding food status data characterizing the food during the crushing process, and the memory stores nasogastric feeding food status data, nasogastric feeding care data, and other related information. The processor, which processes target information, expert control data, and collaborative execution history, executes programs stored in memory to input nasogastric feeding food status data into a first recommendation model, output target nasogastric feeding care information, input the target nasogastric feeding care target information into a second recommendation model, output target expert control data, and input the nasogastric feeding food status data and target expert control data into a third recommendation model, outputting collaborative execution results. The processor also controls the drive component and exhaust component to operate collaboratively and simultaneously during the food crushing process to form liquid food, and controls the output component to output liquid food when safe feeding conditions are met, while simultaneously controlling the display component to display nasogastric feeding care prompts.

[0006] As a further aspect of the present invention, the nasogastric feeding food status data includes compression resistance data between the pressure plate and the mesh plate, air pressure data inside the food processing chamber, exhaust channel flow data, residual air content data of the slurry, and liquid food refinement data.

[0007] As a further aspect of the present invention, the nasogastric feeding nursing target information is used to characterize one of the following nursing targets during the current nasogastric feeding process: smooth tube feeding, reduced gastric distension, reflux prevention, or safe feeding conditions.

[0008] As a further aspect of the present invention, the expert control data includes one or more of the following: pressure plate movement template, mesh plate movement template, exhaust intensity template, pressure holding template, and output permission template, which correspond to different nasogastric feeding care target information.

[0009] As a further embodiment of the present invention, the first recommendation model is used to output nasogastric feeding care target information based on the nasogastric feeding food status data, the second recommendation model is used to output expert control data based on the nasogastric feeding care target information, and the third recommendation model is used to output collaborative execution results based on the nasogastric feeding food status data and the expert control data.

[0010] As a further aspect of the present invention, the processor controls the drive assembly to increase the rolling frequency, rolling stroke, or mesh refinement intensity of the pressure plate and the mesh plate, so as to improve the refinement of the liquid food.

[0011] As a further aspect of the present invention, when the collaborative execution result is the nursing goal of reducing gastric bloating, the processor controls the venting component to increase the venting intensity and controls the drive component to reduce the crushing disturbance intensity or maintain the turning state, so as to reduce the residual gas content in the liquid food.

[0012] As a further aspect of the present invention, when the result of the collaborative execution is the goal of reflux prevention care, the processor controls the drive component and the exhaust component to operate synchronously and enhancedly, wherein the drive component is used to continue to improve the fineness of the liquid diet, and the exhaust component is used to simultaneously reduce the residual gas content of the liquid diet.

[0013] As a further aspect of the present invention, when the collaborative execution result indicates that the nursing goal of safe feeding conditions has been achieved, the processor controls the output component to open the discharge passage to output the liquid food in the food processing chamber to the nasogastric tube and display nasogastric feeding care prompt information. The safe feeding conditions include at least the liquid food fineness reaching the preset condition and the residual gas content being lower than the preset threshold. The nasogastric feeding care prompt information includes the current nasogastric food status information, nasogastric feeding care goal information, equipment control execution information, and abnormal warning information.

[0014] Secondly, this invention proposes a nasogastric feeding method based on a recommendation model, applying the aforementioned nasogastric feeding system based on a recommendation model, the method comprising:

[0015] Acquire nasogastric feeding food status data during the crushing process; determine current nasogastric feeding care goals based on the nasogastric feeding food status data; determine expert control data based on the nasogastric feeding care goals; determine collaborative execution results based on the nasogastric feeding food status data and expert control data; control the pressing plate and mesh plate to crush and flow the food according to the collaborative execution results, and simultaneously control the exhaust component to remove excess air; when the fineness of the liquid food reaches the preset conditions and the residual air content is lower than the preset threshold, control the output component to output the liquid food to the nasogastric feeding tube and display nasogastric feeding care prompts.

[0016] Compared with the prior art, the technical effects of the present invention are as follows:

[0017] This invention constructs a first recommendation model driven by nasogastric feeding food status data, a second recommendation model driven by nasogastric feeding care goal information, and a third recommendation model driven by nasogastric feeding food status data and expert control data. Throughout the entire process of food being squeezed, sheared, and refined through a mesh between a pressure plate and a mesh to form a liquid diet, it simultaneously identifies nasogastric feeding care goals that should be prioritized, such as smooth tube feeding, reduced gastric distension, reflux prevention, or safe feeding. It also determines the crushing frequency, crushing stroke, gas expulsion intensity, pressure holding method, and permissible output conditions, ensuring that crushing and expelling excess air occur simultaneously. This improves the uniformity of liquid diet refinement, reduces residual gas content and flow fluctuations, and decreases the risk of tube blockage, gastric distension, and gastroesophageal reflux. Furthermore, by displaying nasogastric feeding care prompts, it enhances the accuracy of judgment, operational safety, and nursing efficiency for inexperienced nurses performing nasogastric feeding. Attached Figure Description

[0018] Figure 1 This is a system block diagram of the present invention;

[0019] Figure 2 This is a flowchart of the method of the present invention;

[0020] Figure 3 This is a flowchart illustrating the implementation of the online operation phases of the three recommendation models of this invention; Figure 4 This is a specific example diagram illustrating the display of nasogastric feeding care prompts by the display component of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1

[0023] like Figure 1As shown, this embodiment provides a nasogastric feeding and nursing system based on a recommendation model, including a housing, end cap, drive assembly, pressure plate, mesh plate, exhaust assembly, output assembly, status acquisition assembly, display assembly, memory, and processor. The housing forms a food processing chamber, and an end cap is detachably mounted on the upper part of the housing. A pressure plate is fixedly mounted on the lower side of the end cap, and a mesh plate is located below the pressure plate and connected to the drive assembly. The drive assembly drives the mesh plate to perform reciprocating approach and separation movements relative to the pressure plate, and rotates the mesh plate during at least part of its stroke, so that the food is subjected to squeezing, shearing, turning, and mesh crushing between the pressure plate and the mesh plate. An exhaust assembly is connected to the food processing chamber and is used to simultaneously exhaust excess air from the food processing chamber and the slurry during the process of crushing the food into a liquid diet. An output assembly is connected to the food processing chamber and is used to output the processed liquid diet to the nasogastric feeding tube when safe feeding conditions are met. A display assembly is used to display nasogastric feeding care prompts. A memory is used to store nasogastric feeding food status data, nasogastric feeding care target information, expert control data, and collaborative execution history. A processor is used to execute the program stored in the memory to achieve coordinated control of the drive assembly, exhaust assembly, output assembly, and display assembly.

[0024] In this embodiment, the nurse added 150g of cooked rice porridge, 80g of pumpkin puree, 70g of chicken breast puree, and 220mL of warm water to the food processing chamber, for a total loading weight of 520g. The target gastric tube size was 8Fr. The initial crushing frequency of the drive component was set to 32 times / min, the single crushing stroke was set to 20mm, the initial rotation speed of the mesh plate was set to 90rpm, and the initial exhaust intensity of the exhaust component was set to −14kPa. The status acquisition component collected nasogastric feeding food status data every 2s. The nasogastric feeding food status data included the compression resistance data between the pressure plate and the mesh plate, the air pressure data in the food processing chamber, the exhaust channel flow rate data, the residual gas content data of the slurry, and the fineness data of the liquid food. In practice, the compression resistance data between the pressure plate and the mesh plate can be obtained through a torque sensor, strain gauge force sensor, or drive motor current detection unit installed on the drive assembly; the gas pressure data inside the food processing chamber can be obtained through a pressure sensor installed at the pressure tap at the top of the food processing chamber; the flow rate data of the exhaust channel can be obtained through a gas flow sensor installed in the exhaust channel; the residual gas content data of the slurry can be obtained through an image acquisition module installed at the transparent observation window combined with a bubble recognition algorithm, or calculated through a weight sensor, a liquid level detection module, and a density inversion model; the fineness data of the liquid food can be obtained through particle recognition by an image acquisition module installed downstream of the mesh plate, or calculated jointly through the pressure difference between the upper and lower parts of the mesh plate, the resistance to passing through the mesh, and rheological parameters.

[0025] Twenty seconds after system startup, the status data of the nasogastric feeding food collected by the status acquisition component were as follows: the compression resistance between the pressure plate and the mesh plate was 42.6 N, the pressure fluctuation amplitude in the food processing chamber was ±3.8 kPa, the exhaust channel flow rate was 1.1 L / min, the residual gas content of the slurry was 4.2%, and the fineness of the liquid food was 69%. The processor input the nasogastric feeding food status data into the first recommendation model, which output the target nasogastric feeding care goal information as "smooth tube feeding care goal" with a confidence level of 0.82. Subsequently, the processor input the target nasogastric feeding care goal information into the second recommendation model, which output the target expert control data as "enhanced compaction template T1". Then, the nasogastric feeding food status data and the target expert control data were input into the third recommendation model, which output the collaborative execution result as "continue to enhance the compaction flow, maintain basic exhaust, and prohibit the output of liquid food in the current stage".

[0026] In this embodiment, the enhanced compaction template T1 includes: increasing the compaction frequency from 32 times / min to 46 times / min, increasing the compaction stroke from 20mm to 28mm, increasing the mesh rotation speed from 90rpm to 130rpm, and maintaining the exhaust assembly within the exhaust intensity range of -14kPa to -16kPa. The processor controls the drive assembly and exhaust assembly to operate according to the above parameters based on the coordinated execution results, prioritizing the refinement of the liquid diet and reducing the risk of blockage in the thin gastric tube.

[0027] While performing the above controls, the display component outputs nasogastric feeding care prompts. The current nasogastric feeding food status information is displayed as "liquid food refinement degree 69%, residual air content 4.2%, insufficient refinement"; the nasogastric feeding care goal information is displayed as "current care goal: smooth supply of nutrients through the tube"; the equipment control execution information is displayed as "the crushing frequency has been increased to 46 times / min, the mesh refinement intensity has been increased, the flow is currently being controlled, and no output is being output for the time being"; the abnormal warning information is displayed as "insufficient refinement warning, the conditions for safe supply of nutrients through an 8 Fr gastric tube have not been met".

[0028] After the system ran continuously for 180 seconds, it collected nasogastric feeding data again, obtaining the following results: compressibility resistance decreased to 19.4 N, pressure fluctuation amplitude in the food processing chamber decreased to ±1.9 kPa, exhaust flow rate was 1.5 L / min, residual gas content in the slurry decreased to 3.3%, and the fineness of the liquid diet improved to 96%. The processor input the updated nasogastric feeding data back into the first recommendation model. The first recommendation model output the target nasogastric feeding care goal information as "Safe feeding conditions are met", with a confidence level of 0.91; the second recommendation model output the target expert control data as the output permission template T4; and the third recommendation model output the collaborative execution result as "Supplementary liquid diet allowed".

[0029] In this embodiment, the safe feeding conditions set for the 8Fr gastric tube are: the fineness of the liquid food reaches 95% or more, the residual gas content of the slurry is less than 3.5%, and the pressure fluctuation amplitude in the processing cavity is no greater than ±1.5 kPa. After the processor determines that the fineness of the liquid food reaches 96%, the residual gas content of the slurry reaches 3.3%, and the pressure fluctuation in the processing cavity is close to stable, the control output component opens the discharge valve and delivers the liquid food into the patient's stomach through the nasogastric tube at an output flow rate of 16 mL / min.

[0030] During the output phase, the display component outputs nasogastric feeding care prompts. The current nasogastric feeding food status is displayed as "96% liquid food refinement, 3.3% residual gas content, status meets standards"; the nasogastric feeding care goal information is displayed as "Current care goal: Safe feeding conditions are met"; the equipment control execution information is displayed as "Discharge path is open, currently discharging liquid food at 16 mL / min"; and the abnormal warning information is displayed as "None". Through the above process, this embodiment fully discloses the specific implementation methods of the system structure, status acquisition, model reasoning, collaborative control, output permission, and prompt display provided by the present invention.

[0031] Example 2

[0032] Example 2 has the same system structure, module composition and control process as Example 1. The difference is that this example is applicable to scenarios where the degree of fineness of liquid food is basically up to standard but the residual gas content of the liquid is too high. The priority nursing goal for nasogastric feeding is "the nursing goal of reducing gastric bloating".

[0033] In this embodiment, the raw materials added to the food processing chamber are 180g of nutritional powder preparation solution, 100g of mashed potatoes, 60g of steamed egg custard, and 260mL of warm water, with a total loading weight of 600g. The target gastric tube specification is 10Fr. The initial crushing frequency of the drive component is set to 34 times / min, the single crushing stroke is set to 22mm, the initial rotation speed of the mesh plate is set to 100rpm, and the initial exhaust intensity of the exhaust component is set to −15kPa.

[0034] Thirty seconds after system startup, the status data of the nasogastric feeding food collected by the status acquisition component were as follows: compression resistance 21.2 N, pressure fluctuation amplitude in the food processing chamber ±4.6 kPa, exhaust flow rate 0.8 L / min, residual gas content in the slurry 7.8%, and liquid food refinement degree 93%. The processor input the nasogastric feeding food status data into the first recommendation model, which outputs the target nasogastric feeding care goal information as "reducing gastric distension care goal" with a confidence level of 0.86; the second recommendation model outputs the target expert control data as enhancing exhaust template T2; and the third recommendation model outputs the collaborative execution result as "enhancing exhaust, reducing crushing disturbance, and delaying the output of liquid food".

[0035] In this embodiment, the enhanced venting template T2 includes: increasing the venting intensity of the venting component from −15kPa to −28kPa, reducing the crushing frequency from 34 times / min to 18 times / min, reducing the crushing stroke from 22mm to 12mm, and reducing the mesh rotation speed from 100rpm to 45rpm to maintain a low-disturbance turning state. The display component outputs corresponding nasogastric feeding care prompts, including: the current nasogastric feeding food status information is displayed as "liquid food refinement degree 93%, residual gas content 7.8%, venting efficiency is low"; the nasogastric feeding care goal information is displayed as "current care goal: relief of gas in the stomach"; the equipment control execution information is displayed as "venting intensity has been increased to −28kPa, crushing disturbance has been reduced, currently maintaining low-disturbance turning and continuous venting"; and the abnormal warning information is displayed as "warning of excessive gas content, please suspend the output of liquid food".

[0036] After the system continued running for 160 seconds, the data collected by the status acquisition component was updated as follows: compression resistance 17.8 N, pressure fluctuation amplitude in the food processing chamber ±1.4 kPa, exhaust channel flow rate 1.9 L / min, residual gas content in the slurry 3.1%, and liquid food refinement degree 95%. The safe feeding conditions set for the 10Fr gastric tube are: liquid food refinement degree above 92%, residual gas content in the slurry below 4.0%, and pressure fluctuation amplitude in the processing chamber not exceeding ±2.0 kPa. Based on this, the processor determined that the safe feeding conditions were met and controlled the output component to open the discharge channel at 15 mL / min.

[0037] In the output stage of this embodiment, the display component outputs nasogastric feeding care prompts. The current nasogastric feeding food status information is displayed as "95% liquid food refinement, 3.1% residual gas content, status meets standards"; the nasogastric feeding care target information is displayed as "Current care target: Safe feeding conditions are met"; the equipment control execution information is displayed as "Discharge path opened, currently discharging liquid food at 15mL / min"; and the abnormal warning information is displayed as "None". Compared with Embodiment 1, the difference in this embodiment is that the current nasogastric feeding food status information is "Refinement basically meets standards but gas content is high", the nasogastric feeding care target information changes to "Reduced gas in the stomach", the equipment control execution information is reflected as "Enhanced gas expulsion, reduced crushing disturbance", and the abnormal warning information is reflected as "Warning of excessively high gas content".

[0038] Example 3

[0039] Example 3 has the same system structure, module composition and control process as Example 1. The difference is that this example is applicable to scenarios where the liquid diet is not refined enough and the residual gas content of the liquid is high, and there is an increased risk of gastroesophageal reflux. The priority goal of nasogastric feeding is "reflux prevention nursing goal".

[0040] In this embodiment, the raw materials added to the food processing chamber are 120g vegetable puree, 90g chicken puree, 90g thick rice paste and 260mL warm water, with a total loading mass of 560g. The target gastric tube specification is 8Fr. The initial crushing frequency of the drive component is set to 30 times / min, the single crushing stroke is set to 18mm, the initial rotation speed of the mesh plate is set to 80rpm, and the initial exhaust intensity of the exhaust component is set to −16kPa.

[0041] 25 seconds after system startup, the status data of the nasogastric feeding food collected by the status acquisition component are as follows: compression resistance 36.8N, pressure fluctuation amplitude in the food processing chamber ±5.4kPa, exhaust channel flow rate 0.9L / min, residual gas content in the slurry 7.2%, and liquid food refinement degree 78%. The processor inputs the nasogastric feeding food status data into the first recommendation model. The first recommendation model outputs the target nasogastric feeding care goal information as "reflux prevention care goal" with a confidence level of 0.79; the second recommendation model outputs the target expert control data as the synchronous enhancement template T3; and the third recommendation model outputs the collaborative execution result as "synchronous enhancement of crushing and exhaust, continuous delayed output of liquid food".

[0042] In this embodiment, the synchronous enhancement template T3 includes: increasing the crushing frequency from 30 times / min to 44 times / min, increasing the crushing stroke from 18mm to 28mm, increasing the mesh rotation speed from 80rpm to 120rpm, and increasing the exhaust intensity of the exhaust component from −16kPa to −30kPa, while keeping the output component closed for at least 180s. The display component outputs nasogastric feeding care prompts, including: the current nasogastric feeding food status information is displayed as "liquid food refinement degree 78%, residual gas content 7.2%, supply risk status increased"; the nasogastric feeding care goal information is displayed as "current care goal: reflux prevention"; the equipment control execution information is displayed as "synchronous enhancement of crushing flow and exhaust treatment has been implemented, current liquid food output is delayed"; and the abnormal warning information is displayed as "gastroesophageal reflux risk warning, liquid food output is prohibited at the current stage".

[0043] After the system continued running for 180 seconds, the data collected by the status acquisition component was updated as follows: compression resistance 22.7 N, pressure fluctuation amplitude in the food processing chamber ±2.3 kPa, exhaust channel flow rate 1.8 L / min, residual gas content in the slurry 4.6%, and liquid food refinement degree 93%. Since the safe feeding conditions for the 8Fr gastric tube were not yet fully met, the first recommendation model continued to output the "reflux prevention nursing target," while the third recommendation model continued to maintain the coordinated execution result of "synchronously enhancing crushing and exhaust, delaying output." After the system ran for another 70 seconds, the data collected by the status acquisition component was as follows: compression resistance 17.5 N, pressure fluctuation amplitude in the food processing chamber ±1.2 kPa, exhaust channel flow rate 2.1 L / min, residual gas content in the slurry 2.9%, and liquid food refinement degree 96%. Based on this, the processor determined that the safe feeding conditions for the 8Fr gastric tube were met and controlled the output component to open the discharge channel at 12 mL / min.

[0044] In the output stage of this embodiment, the display component outputs nasogastric feeding care prompts. The current nasogastric feeding food status information is displayed as "96% liquid food refinement, 2.9% residual gas content, status meets standards"; the nasogastric feeding care target information is displayed as "Current care target: Safe feeding conditions are met"; the equipment control execution information is displayed as "Discharge pathway opened, currently discharging liquid food at 12mL / min"; and the abnormal warning information is displayed as "None". Compared with Embodiment 1, the difference in this embodiment is that the current nasogastric feeding food status information is "insufficient refinement and high gas content coexist," the nasogastric feeding care target information is changed to "reflux prevention," the equipment control execution information is reflected as "Simultaneously enhance crushing and degassing while delaying output," and the abnormal warning information is reflected as "Gastroesophageal reflux risk warning."

[0045] In this invention, the first recommendation model, the second recommendation model, and the third recommendation model are used to complete three different tasks: state recognition, template matching, and collaborative control decision-making, respectively. Therefore, the three can adopt different types of models.

[0046] The first recommended model is used to output nasogastric feeding care target information based on nasogastric feeding food status data. Since its input data includes compression resistance data, food processing chamber air pressure data, exhaust channel flow rate data, residual gas content data of slurry data, and liquid food refinement data, and these data change continuously over time, the first recommended model is preferably a time-series classification model.

[0047] In a preferred embodiment, the first recommended model employs: a one-dimensional convolutional neural network (1D-CNN) + a gated recurrent unit network (GRU) + a softmax classification layer. The one-dimensional convolutional neural network is used to extract local variation features of the nasogastric feeding food status data over time; the gated recurrent unit network is used to extract temporal evolution features during food crushing and gas expulsion; and the softmax classification layer is used to output the current nasogastric feeding care goal information. This goal information includes one of the following: smooth tube feeding, relief of gastric distension, reflux prevention, and the achievement of safe feeding conditions.

[0048] In other alternative implementations, the first recommended model may also be a random forest model, an XGBoost model, a LightGBM model, a multilayer perceptron model (MLP), or a long short-term memory network model (LSTM).

[0049] The second recommendation model is used to output target expert control data based on the target nasogastric feeding care target information.

[0050] Since the second recommendation model essentially performs the task of matching "target to template", it is more suitable to use template matching model, ranking model or rule constraint model.

[0051] In a preferred embodiment, the second recommendation model employs a target label embedding network and a fully connected matching network. The target label embedding network maps different nasogastric feeding care goals into low-dimensional target vectors, while the fully connected matching network outputs the target expert control data that best matches the current care goal based on the similarity between the target vector and pre-stored expert control data. The expert control data includes one or more of the following: pressure plate movement template, mesh plate movement template, exhaust intensity template, pressure holding template, output permission template, and abnormal warning template.

[0052] In other alternative implementations, the second recommendation model may also employ a decision tree model, a rule engine model, a K-nearest neighbor matching model, a support vector machine model, or a vector retrieval ranking model.

[0053] The third recommendation model outputs collaborative execution results based on nasogastric feeding status data and target expert control data. Since the third recommendation model essentially performs a decision-making task of state + template → control action, it is more suitable to use a sequential decision model or a control decision model.

[0054] In a preferred embodiment, the third recommendation model employs a Long Short-Term Memory (LSTM) network combined with a fully connected decision layer. The LSTM network is used to extract the coupling relationship between nasogastric feeding food status data and target expert control data during continuous processing, while the fully connected decision layer outputs the collaborative execution results. These collaborative execution results include one or more of the following: increasing compaction frequency, increasing compaction stroke, increasing netting refinement intensity, increasing venting and extraction intensity, reducing compaction disturbance intensity, maintaining the turning state, delaying the output of liquid feed, and allowing the output of liquid feed.

[0055] In other alternative implementations, the third recommendation model may also employ a deep Q-network (DQN), a policy network model, a combination of state machine and neural network model, an XGBoost decision model, or a rule-constrained classification model.

[0056] The implementation process of this invention is preferably divided into two stages: the first stage: the learning and construction stage; and the second stage: the online operation stage.

[0057] Phase 1: Learning the specific implementation methods of the construction phase

[0058] In the first phase, the main tasks were to construct the correspondence between nasogastric feeding food status data, nasogastric feeding care goal information, expert control data, and collaborative execution results, as well as to train three recommendation models.

[0059] In practice, nurses, equipment technicians, or process engineers with experience in nasogastric feeding can take multiple samples and record the feeding process under different food combinations, different target gastric tube specifications, and different treatment conditions to form a training sample library.

[0060] In one specific implementation, such as Figure 2 As shown, the training samples include the following:

[0061] (1) Input status data sample: The status data sample includes the compression resistance data between the pressure plate and the mesh plate, the air pressure data in the food processing chamber, the flow rate data of the exhaust channel, the residual air content data of the slurry, and the fineness data of the liquid food; each sample can be collected at 2s intervals, with 10 sampling points continuously, forming a status window with a length of 20s;

[0062] (2) Nursing goal labeling sample: Experienced nurses label the nasogastric feeding nursing goals that should be prioritized at each status window according to the requirements of nursing safety;

[0063] For example:

[0064] If the degree of refinement of liquid food is less than 90% and the size of the thin gastric tube is 8Fr, it is marked as "smooth feeding and nursing goal of thin tube";

[0065] If the degree of refinement of liquid food is higher than 92%, but the residual gas content of the liquid is higher than 5.0%, it is marked as "nursing goal of reducing gastric bloating";

[0066] If the degree of refinement of the liquid diet is less than 92% and the residual gas content is higher than 5.5%, it is marked as "reflux prevention nursing target";

[0067] If the liquid diet is more than 95% refined and the residual gas content is less than 3.5%, it is marked as "the nursing goal of safe feeding conditions has been met";

[0068] (3) Expert control data annotation samples: Write the experience control strategy templates corresponding to different nursing goals into the expert control database;

[0069] For example:

[0070] Template T1: The rolling frequency is increased to 46 times / min, the rolling stroke is increased to 28mm, and the screen rotation speed is increased to 130rpm;

[0071] Template T2: Exhaust intensity increased to −28kPa, rolling frequency reduced to 18 times / min, and mesh plate speed reduced to 45rpm;

[0072] Template T3: The rolling frequency is increased to 44 times / min, the exhaust intensity is increased to −30kPa, and the delayed output is at least 180s;

[0073] Template T4: Open the discharge valve and output the liquid feed at a rate of 12mL / min to 16mL / min;

[0074] (4) Sample of collaborative execution results: Based on status data and expert control data, experienced personnel mark the actual execution actions, such as "enhanced compaction", "enhanced exhaust", "synchronized enhanced compaction and exhaust", "delayed output", and "allowed output".

[0075] In the first stage, a first recommendation model is trained using state window samples and nursing goal annotation samples to establish a mapping relationship between nasogastric feeding food status data and nasogastric feeding nursing goal information. Then, a second recommendation model is trained using nursing goal information and expert control data samples to establish a mapping relationship between nasogastric feeding nursing goal information and expert control data. Finally, a third recommendation model is trained using nasogastric feeding food status data, expert control data, and collaborative execution result samples to establish a mapping relationship between nasogastric feeding food status data + expert control data and collaborative execution results.

[0076] Preferably, after the model training is completed, the trained model parameters are written into the memory for use in the second stage.

[0077] Phase Two: Specific Implementation Methods for the Online Operation Phase

[0078] In the second stage, the system enters the actual nasogastric feeding operation state, and the processor calls the three trained recommendation models in real time to continuously identify and control the food processing process.

[0079] In specific implementation, such as Figure 3 As shown, the second stage includes the following steps:

[0080] Step S1: Real-time acquisition of nasogastric feeding food status data: The status acquisition component continuously acquires compression resistance data, food processing chamber air pressure data, exhaust channel flow data, residual gas content data of slurry, and liquid food refinement data according to a preset sampling cycle, and inputs them into the processor;

[0081] Step S2, determine the current nasogastric feeding care target information: The processor inputs the real-time collected nasogastric feeding food status data into the first recommendation model to obtain the current nasogastric feeding care target information;

[0082] Step S3, Matching target expert control data: The processor inputs the current nasogastric feeding care target information output in step S2 into the second recommendation model to obtain the target expert control data that best matches the current care target;

[0083] Step S4, Generate collaborative execution results: The processor inputs the nasogastric feeding food status data collected in step S1 and the target expert control data obtained in step S3 into the third recommendation model to generate the collaborative execution results for the current stage.

[0084] Step S5, execute linkage control: Based on the collaborative execution result generated in step S4, the processor controls the drive component, exhaust component and output component to perform corresponding actions, and controls the display component to display nasogastric feeding care prompt information.

[0085] In summary, in Examples 1, 2, and 3, the first recommendation model is preferably used to identify the current nasogastric feeding care target information based on the nasogastric feeding food status data; the second recommendation model is preferably used to match the target expert control data based on the nasogastric feeding care target information; and the third recommendation model is preferably used to generate collaborative execution results based on the nasogastric feeding food status data and the target expert control data. The entire implementation process is preferably divided into a learning and construction phase and an online operation phase. The learning and construction phase is used to establish the correspondence between the nasogastric feeding food status data, the nasogastric feeding care target information, the teacher control data, and the collaborative execution results, and to train the three recommendation models. The online operation phase is used to call the three trained recommendation models to identify, control, and display prompts in real time during the nasogastric feeding process.

[0086] like Figure 4As shown, the display component is used to display nasogastric feeding care prompts, which include current nasogastric feeding food status information, nasogastric feeding care goal information, equipment control execution information, and abnormal warning information. In a feasible implementation scheme, the display component adopts a partitioned display method, including a food status display area, a nursing goal display area, a device execution display area, and an abnormal warning display area. The food status display area displays the current nasogastric feeding food status information collected by the status acquisition component and calculated by the processor. The current nasogastric feeding food status information includes one or more of the following: compression resistance, intracavitary pressure, gas exhaust flow rate, and liquid food refinement level. The nursing goal display area displays the current nasogastric feeding nursing goal information output by the first recommendation model, such as smooth tube feeding, reduced gastric distension, reflux prevention, or the achievement of safe feeding conditions. The device execution display area displays the device control execution information output by the third recommendation model and executed by the processor, including one or more of the following: crushing frequency, crushing stroke, gas exhaust intensity, whether to continue flow control, whether to delay output, or whether output is allowed. The abnormal warning display area displays abnormal warning information corresponding to the current status, including low liquid food refinement level, high residual gas content, increased risk of gastroesophageal reflux, and temporarily disallowed output. Furthermore, the display component can be equipped with an operating status indicator area and a manual intervention indicator area at the bottom of the display interface. The operating status indicator area is used to display the current status of the device, such as flow control, exhaust, waiting to output, or outputting. The manual intervention indicator area is used to prompt the nursing staff whether it is necessary to pause the output, continue observation, or perform manual verification. Preferably, the food status display area, nursing goal display area, device execution display area, and abnormal warning display area are distinguished by different colors or different icons, so that inexperienced nasogastric feeding nurses can quickly identify the current nasogastric feeding status, current priority nursing goals, device execution actions, and abnormal risks, thereby improving the readability of information, ease of operation, and nursing safety during the nasogastric feeding process.

[0087] It should be noted that, in Figure 4 In the example shown, the food status display area displays "compression resistance 42.6N, intracavitary pressure ±3.8kPa, exhaust flow rate 1.1L / min, fineness 69%", the nursing goal display area displays "smooth feeding through the tube", the equipment execution display area displays "crushing frequency 46 times / minute, crushing stroke 28mm, exhaust intensity −26kPa", and the abnormal warning display area displays "warning: low fineness of liquid food, do not output". This indicates that the current processing stage has not yet met the safe feeding conditions. The system has automatically increased the crushing intensity and limited the output according to the recommended model to help nurses avoid delivering insufficiently fined liquid food to the nasogastric tube.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0089] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A nasogastric feeding care system based on a recommendation model, characterized by, The system includes a housing, end caps, a drive assembly, a pressure plate, a mesh plate, an exhaust assembly, an output assembly, a status acquisition assembly, a display assembly, a memory, and a processor. A food processing chamber is formed within the housing. The end cap is located at the upper end of the housing. The pressure plate and mesh plate are positioned within the food processing chamber and move relative to each other under the action of the drive assembly to perform squeezing, shearing, and mesh-crushing processing on the food. The exhaust assembly and the output assembly are connected to the food processing chamber. The status acquisition assembly acquires data characterizing the nasogastric feeding food status during the crushing process. The memory stores nasogastric feeding food status data, nasogastric feeding care target information, expert control data, and collaborative execution history. The processor executes the program stored in the memory to input the nasogastric feeding food status data into a first recommendation model, outputting target nasogastric feeding care target information; inputting the target nasogastric feeding care target information into a second recommendation model, outputting target expert control data; and inputting the nasogastric feeding food status data and target expert control data into a third recommendation model, outputting the collaborative execution result. The processor is also used to control the drive component and the exhaust component to operate simultaneously and collaboratively during the process of food crushing to form liquid food, based on the results of the coordinated execution, and to control the output component to output liquid food when the safe feeding conditions are met, while controlling the display component to display nasogastric feeding care prompts.

2. The feeding system based on recommendation model according to claim 1, wherein, Nasogastric feeding data includes data on the compression resistance between the pressure plate and the mesh plate, data on the air pressure inside the food processing chamber, data on the flow rate of the exhaust channel, data on the residual air content of the slurry, and data on the degree of refinement of the liquid diet. 3.The nasal feeding care system based on the recommendation model of claim 1, wherein, Nasogastric feeding nursing goal information is used to characterize one of the following nursing goals during the current nasogastric feeding process: smooth tube feeding, relief of gastric distension, prevention of reflux, or the achievement of safe feeding conditions. 4.The nasogastric feeding care system based on a recommendation model of claim 1, wherein, Expert control data includes one or more of the following templates corresponding to different nasogastric feeding care goals: pressure plate movement template, mesh plate movement template, exhaust intensity template, pressure holding template, and output permission template. 5.The nasogastric feeding care system based on a recommendation model of claim 1, wherein, The first recommendation model is used to output nasogastric feeding care target information based on nasogastric feeding food status data. The second recommendation model is used to output expert control data based on nasogastric feeding care target information. The third recommendation model is used to output collaborative execution results based on nasogastric feeding food status data and expert control data. 6.The nasogastric feeding care system based on a recommendation model of claim 1, wherein, The processor controls the drive components to increase the rolling frequency, rolling stroke, or screen refining intensity of the pressure plate and screen, thereby improving the fineness of the liquid food. 7.The nasogastric feeding care system based on a recommendation model according to claim 1, wherein, When the collaborative execution result is to reduce gastric bloating, the processor controls the degassing assembly to increase the degassing intensity and controls the drive assembly to reduce the crushing disturbance intensity or maintain the turning state to reduce the residual gas content in the liquid diet. 8.The nasogastric feeding care system based on a recommendation model of claim 1, wherein, When the collaborative execution result is the reflux prevention care goal, the processor controls the drive component and the exhaust component to synchronously enhance operation, wherein the drive component is used to continue to improve the fineness of the liquid diet, and the exhaust component is used to synchronously reduce the residual gas content of the liquid diet.

9. A nasogastric feeding and nursing system based on a recommendation model according to claim 1, characterized in that, When the collaborative execution result indicates that the nursing goal of safe feeding conditions has been achieved, the processor control output component opens the discharge passage to output the liquid food in the food processing chamber to the nasogastric tube and displays nasogastric feeding care prompts. The safe feeding conditions include at least the liquid food fineness reaching the preset conditions and the residual gas content being lower than the preset threshold. The nasogastric feeding care prompts include the current nasogastric food status information, nasogastric feeding care goal information, equipment control execution information, and abnormal warning information.

10. A nasogastric feeding and nursing method based on a recommendation model, employing a nasogastric feeding and nursing system based on a recommendation model as described in any one of claims 1 to 9, characterized in that, The method includes: Acquire nasogastric feeding food status data during the crushing process; determine current nasogastric feeding care goals based on the nasogastric feeding food status data; determine expert control data based on the nasogastric feeding care goals; determine collaborative execution results based on the nasogastric feeding food status data and expert control data; control the pressing plate and mesh plate to crush and flow the food according to the collaborative execution results, and simultaneously control the exhaust component to remove excess air; when the fineness of the liquid food reaches the preset conditions and the residual air content is lower than the preset threshold, control the output component to output the liquid food to the nasogastric feeding tube and display nasogastric feeding care prompts.

Images