Personalized enteral nutrition infusion regulation and control device with multi-dimensional sign perception function

Abstract

The invention relates to the technical field of intelligent medical treatment and clinical nutrition, and discloses a multi-dimensional sign sensing personalized enteral nutrition infusion regulation and control device which comprises a multi-parameter sign sensing module, an intelligent analysis control module, an adjustable infusion execution module and a man-machine interaction and early warning module. The intelligent nutrition management platform is in communication connection with the modules; the intelligent nutrition management platform is composed of a nutrition screening and evaluation unit and a statistical analysis unit. According to the method, a full link from nutrition risk automatic screening, individualized scheme intelligent generation, bedside safety execution and self-adaptive regulation and control to whole-process data statistical analysis is formed; information integration and initial decision making are automatically completed, manual operation links and cognitive loads are reduced, scientificity and normalization of a treatment scheme based on newest clinical guidelines and patient data are ensured, and medical workers are liberated from tedious transactional work and focus on higher-value clinical judgment.

Description

Technical Field

[0001] This invention relates to the field of smart healthcare and clinical nutrition technology, specifically a personalized enteral nutrition infusion control device with multi-dimensional vital sign perception. Background Technology

[0002] Enteral nutrition support is a crucial component of basic clinical treatment and is widely used in critically ill patients, postoperative patients, and patients with chronic wasting diseases. However, current clinical practice faces systemic challenges throughout the entire nutritional management process, which limits treatment efficacy and safety. 1. Fragmented and inefficient management processes: In the existing model, nutritional screening, protocol development, bedside administration, and efficacy monitoring are scattered across different roles (doctors, nutritionists, nurses) and different information systems (electronic medical records, laboratory systems, prescription systems, and independent infusion pumps). Information transmission relies on manual transcription and communication, resulting in fragmented processes that are not only inefficient but also prone to information errors, execution delays, and increased burden on medical staff.

[0003] 2. Static infusion process, lacking physiological adaptability: Most enteral nutrition infusion pumps widely used in clinical practice are open-loop devices, capable of infusing at a preset fixed rate and temperature, unable to dynamically adjust according to the patient's real-time physiological state and metabolic feedback. Adjustments to the nutrition plan heavily rely on intermittent, experience-based assessments by healthcare professionals (such as assessing gastric residual volume every 4–6 hours). This "set and forget" approach cannot respond promptly to the patient's rapidly changing tolerance and needs, often leading to underfeeding, overfeeding, or gastrointestinal complications (such as diarrhea, bloating, reflux, and aspiration).

[0004] 3. Lack of data closure in quality management and decision optimization: Currently, there is a lack of systems capable of automatically integrating data across the entire chain from screening to implementation. Treatment effectiveness evaluation often relies on medical record review, making it difficult to conduct continuous and objective statistical analysis of key quality indicators such as the completion rate of nutrition plans, reasons for interruption, and incidence of complications for departments or patient groups. Clinical decision-making and process optimization are often based on experience rather than data evidence, hindering the systematic and continuous improvement of treatment quality. Summary of the Invention

[0005] The purpose of this invention is to provide a personalized enteral nutrition infusion control device with multi-dimensional vital sign perception to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A personalized enteral nutrition infusion control device with multi-dimensional vital sign perception includes: The system includes a multi-parameter vital sign sensing module, an intelligent analysis and control module, an adjustable infusion execution module, a human-computer interaction and early warning module; and an intelligent nutrition management platform that communicates with the above modules. The multi-parameter vital sign sensing module is used to collect multi-dimensional physiological, metabolic and gastrointestinal vital sign data of users in real time and to preprocess the data. The intelligent analysis and control module is used to perform fusion analysis of sensor data and nutrition plans based on machine learning algorithms, dynamically assess nutritional needs and tolerance, and generate individualized infusion rate, temperature and formula control strategies. The adjustable infusion execution module is used to precisely control the flow rate of the multi-channel infusion pump and the intelligent temperature control pipeline according to the control strategy, so as to achieve precise delivery and temperature regulation of nutrient solution, and integrate identity verification to ensure execution safety. The human-computer interaction and early warning module is used to provide users with parameter settings, real-time status visualization, and intelligent graded early warning when anomalies are detected. It is also responsible for bidirectional communication of all data.

[0007] The intelligent nutrition management platform is used to automatically acquire and integrate patient information, complete nutritional risk screening and assessment, generate individualized initial nutrition plans, and perform statistical analysis on data throughout the treatment process; it consists of a nutrition screening and assessment unit and a statistical analysis unit.

[0008] As a further aspect of the present invention: the multi-parameter vital sign sensing module includes: Physiological parameter sensing unit for monitoring heart rate, respiration, body temperature, blood pressure, and blood oxygen saturation via integrated or wirelessly connected wearable devices.

[0009] The metabolic and gastrointestinal sensing unit is used to monitor metabolic indicators and gastrointestinal status; among which... Metabolic indicator monitoring includes: acquiring blood glucose or electrolyte data using a continuous glucose monitoring sensor or a non-invasive electrolyte analysis patch; Gastrointestinal status monitoring includes: using abdominal bioelectrical impedance sensors to monitor gastric emptying and intestinal contents flow, or using a bowel sound acquisition array to identify bowel sound characteristics through AI to assess intestinal peristalsis status.

[0010] As a further embodiment of the present invention: the intelligent analysis and control module includes: The data fusion and status assessment engine is used to fuse the initial nutrition plan with real-time data from the multi-parameter vital sign perception module, and dynamically outputs real-time nutritional requirement index, gastrointestinal tolerance level and absorption efficiency prediction through machine learning model. An adaptive control strategy generator is used to generate control commands that adjust the infusion rate, nutrient formulation ratio, and infusion temperature in real time based on the evaluation results of dynamic output.

[0011] As a further embodiment of the present invention: the adjustable infusion execution module includes: A multi-channel intelligent infusion pump is used to independently control the infusion rate of multiple nutrient components, enabling dynamic mixing of the formula ratio; Intelligent temperature control pipeline is used to precisely heat and maintain the temperature of the nutrient solution according to control commands. The identity security execution unit is used to perform dual identity verification before infusion by scanning the patient's and nutritional preparation's identification information.

[0012] As a further aspect of the present invention: the human-computer interaction and early warning module includes: The bedside interactive terminal is used to display nutrition plans, real-time vital signs curves, infusion parameters, and tolerance assessment results; The intelligent early warning system is used to issue tiered alarms and push treatment suggestions when signs of feeding intolerance, metabolic disorders or equipment malfunctions are detected. The data synchronization interface is used to upload all execution, control, and early warning data of the device to the intelligent nutrition management platform.

[0013] As a further aspect of the present invention, the specific method by which the nutrition screening and assessment unit performs nutrition screening and assessment is as follows: S11. Automatically obtain objective patient information from interconnected hospital information systems; S12. Based on objective information, call the preset standard nutritional risk screening model to automatically complete the nutritional risk score and classification of the patient; S13. Calculate the daily nutritional needs target based on the patient's disease status, metabolic level and anthropometric data; S14. Based on a medical knowledge base and logical rule model, and combined with daily nutritional needs targets, automatically generate an individualized initial nutrition plan.

[0014] As a further aspect of the present invention, the specific method by which the statistical analysis unit implements intelligent statistical analysis is as follows: S21. Integrate infusion execution data from the adjustable infusion execution module and nursing assessment data from electronic nursing records to form a dataset of the entire patient treatment process; S22. Based on the full-process dataset, automatically calculate predefined quality management indicators, including: nutrition plan completion rate, feeding interruption rate, and complication rate; S23. Dynamically display the calculated quality management indicators in the form of visual charts; the forms of visual charts include: dynamically updated bar charts, trend curves, pie charts, or dashboards, and support data drill-down and filtering by ward, disease diagnosis, and time dimension; S24. Based on the historical trend analysis of quality management indicators, when the indicator value deviates from the preset threshold or shows a deteriorating trend, a quality early warning prompt will be automatically generated.

[0015] A personalized enteral nutrition infusion regulation method based on multidimensional vital sign perception includes the following steps: S1. The intelligent nutrition management platform automatically acquires patient information, completes nutrition screening and assessment, and generates an individualized initial nutrition plan. S2. The initial nutrition plan is sent to the designated personalized enteral nutrition infusion control device, and the infusion is started after the identity is verified. S3. While the device is performing infusion, it collects data in real time through the multi-parameter vital sign sensing module, and the intelligent analysis and control module dynamically assesses the patient's status, generates and executes an adaptive regulation strategy, and adjusts the infusion parameters in real time. S4. The device will synchronize the infusion execution data, parameter adjustment records and early warning information back to the intelligent nutrition management platform in real time. S5, the intelligent nutrition management platform, performs statistical analysis on the received data throughout the entire process and generates quality management reports to optimize treatment plans or clinical decisions.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention establishes a complete chain from automatic nutritional risk screening, intelligent generation of individualized treatment plans, safe bedside execution and adaptive regulation to full-process data statistical analysis; it automatically completes information integration and initial decision-making, reduces manual operation and cognitive load, and ensures that treatment plans are based on the latest clinical guidelines and patient data in a scientific and standardized manner, freeing medical staff from tedious administrative work and allowing them to focus on higher-value clinical judgments.

[0017] By integrating sensors to continuously monitor metabolic indicators and gastrointestinal status, and combining machine learning algorithms for real-time analysis and evaluation, infusion parameters can be dynamically adjusted. This enables proactive prevention of hyperglycemia and electrolyte imbalances, and timely identification and intervention of feeding intolerances, thereby effectively reducing gastrointestinal complications and metabolic-related risks, and improving nutrient absorption efficiency and planned target achievement rate.

[0018] By automatically collecting and integrating all treatment process data, a complete digital treatment record is formed; its statistical analysis unit can automatically generate multi-dimensional quality management reports, helping managers to objectively and promptly identify common problems and improvement points in clinical practice, thereby driving evidence-based optimization of treatment pathways, nursing standards, or equipment parameters; and providing a powerful data infrastructure for hospital quality control, refined management, and clinical research. Attached Figure Description

[0019] Figure 1A schematic diagram of a personalized enteral nutrition infusion control device with multidimensional vital sign perception; Figure 2 This is a flowchart illustrating a personalized enteral nutrition infusion regulation method based on multidimensional vital sign perception. Detailed Implementation

[0020] Please see Figure 1 In this embodiment of the invention, a personalized enteral nutrition infusion control device with multi-dimensional vital sign perception includes: The system includes a multi-parameter vital sign sensing module, an intelligent analysis and control module, an adjustable infusion execution module, a human-computer interaction and early warning module; and an intelligent nutrition management platform that communicates with the above modules. The multi-parameter vital sign sensing module is used to collect multi-dimensional physiological, metabolic and gastrointestinal vital sign data of users in real time, and to preprocess the data; for example, to perform filtering, noise reduction and other preprocessing on the collected raw signal data. The intelligent analysis and control module is used to perform fusion analysis of sensor data and nutrition plans based on machine learning algorithms, dynamically assess nutritional needs and tolerance, and generate individualized infusion rate, temperature and formula control strategies. The adjustable infusion execution module is used to precisely control the flow rate of the multi-channel infusion pump and the intelligent temperature control pipeline according to the control strategy, so as to achieve precise delivery and temperature regulation of nutrient solution, and integrate identity verification to ensure execution safety. The human-computer interaction and early warning module is used to provide users with parameter settings, real-time status visualization, and intelligent graded early warning when anomalies are detected. It is also responsible for bidirectional communication of all data.

[0021] The intelligent nutrition management platform is used to automatically acquire and integrate patient information, complete nutritional risk screening and assessment, generate individualized initial nutrition plans, and perform statistical analysis on data throughout the treatment process; it consists of a nutrition screening and assessment unit and a statistical analysis unit.

[0022] Preferably, the multi-parameter vital sign sensing module includes: The physiological parameter sensing unit is used to monitor heart rate, respiration, body temperature, blood pressure, and blood oxygen saturation through integrated or wirelessly connected wearable devices. These data can be used to assess the patient's overall condition, metabolic consumption, and determine the presence of acute stress responses (such as infection or pain), providing a stable flow of basic vital signs for intelligent analysis.

[0023] For example, a postoperative patient uses the device; the physiological parameter sensing unit continuously monitors the patient's body temperature as it slowly rises from 36.8°C to 38.2°C via its connected wireless temperature patch, while the fingertip pulse oximeter shows the heart rate increasing from 72 beats / min to 95 beats / min; this real-time data is immediately uploaded, indicating that the patient may have an inflammatory response or an increased metabolic rate due to postoperative absorption heat.

[0024] The metabolic and gastrointestinal sensing unit is used to monitor metabolic indicators and gastrointestinal status; thus overcoming the limitations of traditional methods that rely on intermittent blood draws, it enables continuous bedside monitoring of the metabolic internal environment (blood glucose, electrolytes) and the functional status of digestive organs (gastric emptying, intestinal peristalsis), directly providing key evidence for assessing the "precision" and "tolerance" of nutritional support; among which, Metabolic indicator monitoring includes: acquiring blood glucose or electrolyte data using continuous glucose monitoring sensors or non-invasive electrolyte analysis patches; for example, using technologies such as continuous glucose monitoring (CGM) sensors and non-invasive electrolyte analysis patches to achieve minute-level continuous monitoring of blood glucose levels and key electrolytes (such as potassium and sodium) trends, which greatly improves the timeliness of detection and proactive intervention for metabolic disorders (such as hyperglycemia and electrolyte imbalance).

[0025] For example, during enteral nutrition support for a patient with severe diabetes, the CGM sensor in this unit showed that the patient's blood glucose level rose rapidly from 8 mmol / L to 15 mmol / L within 2 hours. The device immediately captured this trend, and the intelligent analysis module could dynamically adjust the proportion of carbohydrates in the nutrient solution accordingly, or the linkage warning module could suggest that medical staff adjust the insulin dosage to avoid the occurrence of severe hyperglycemia.

[0026] Gastrointestinal status monitoring includes: using abdominal bioelectrical impedance sensors to monitor gastric emptying and intestinal contents mobility, or using bowel sound acquisition arrays to identify bowel sound characteristics through AI to assess intestinal peristalsis. For example, integrating abdominal bioelectrical impedance (BIA) sensors and high-sensitivity bowel sound acquisition arrays can assess gastrointestinal function from both physical movement and physiological sound dimensions. BIA indirectly assesses gastric emptying rate and intestinal contents mobility by monitoring changes in resistance; bowel sound analysis uses AI models to identify the frequency, rhythm, and pitch characteristics of bowel sounds, objectively assessing the activity and regularity of intestinal peristalsis, as well as the presence of signs of obstruction or paralysis.

[0027] For example, when starting nasogastric feeding for a stroke patient with dysphagia, the abdominal BIA sensor showed a slow change in gastric impedance after initiation, indicating delayed gastric emptying. At the same time, the sound signals collected by the bowel sound array were analyzed by AI, showing that bowel sounds were sparse and weak. The device judged a "high risk of gastrointestinal intolerance". Therefore, the intelligent control module can immediately and significantly reduce the initial infusion rate, or even stop feeding, and suggest that the nursing staff assess the patient's abdominal signs, thereby effectively preventing complications such as reflux, aspiration or abdominal distension.

[0028] Preferably, the intelligent analysis and control module includes: The data fusion and status assessment engine integrates the initial nutrition plan with real-time data from the multi-parameter vital sign sensing module. Through a machine learning model, it dynamically outputs real-time nutritional requirement indices, gastrointestinal tolerance levels, and absorption efficiency predictions. Real-time nutritional requirements index: By combining the current metabolic rate (e.g., indirectly estimated through heart rate variability and carbon dioxide production), stress level (e.g., body temperature, CRP trend) and energy consumption model, it dynamically judges the patient's actual energy and protein needs at this moment and quantifies the deviation from the preset target. Gastrointestinal tolerance level: Based on gastric emptying data, bowel sound characteristics, changes in abdominal impedance, and previous infusion responses, tolerance was objectively classified into "good, vigilant, and intolerable" levels using a classification algorithm, and a confidence score was given. Absorption efficiency prediction: Based on the current gastrointestinal state (such as peristaltic activity and mucosal perfusion estimation) and nutritional formula characteristics (such as osmotic pressure and composition), the percentage of nutrients actually absorbed and utilized is predicted and used to correct the calculation of effective feeding amount.

[0029] For example, for a burn patient with high metabolism, the initial protocol is set as a high-calorie target; the engine receives the following during operation: 1) CGM showed a rapid rise in blood sugar; 2) The metabolic monitoring patch showed a slight increase in blood lactate levels; 3) Heart rate and respiratory rate are consistently higher than baseline; After the data fusion and status assessment engine fused these data, the analysis concluded that the patient was in a state of severe stress and insulin resistance. Therefore, the dynamic output showed that the nutritional demand index was "high" (due to high metabolism), but the absorption efficiency was predicted to be "low" (due to insulin resistance and possible microcirculatory disorders). This seemingly contradictory assessment provided a key basis for subsequent precise regulation, namely, "the patient needs nutrition, but the current utilization capacity is limited."

[0030] An adaptive control strategy generator is used to generate control instructions for real-time adjustments to infusion rate, nutrient formulation ratio, and infusion temperature based on dynamically output evaluation results. For example, based on nutrient demand index and absorption efficiency prediction, it decides whether to accelerate to catch up with the target or slow down to avoid metabolic burden; based on gastrointestinal tolerance level, it decides whether to proceed as planned, pause and allow the intestines to rest, or reduce the rate and switch to a more easily absorbed formulation; and based on real-time metabolic indicators and core body temperature, it fine-tunes the proportion of carbohydrates in the formulation and the temperature of the infusion solution.

[0031] Assuming we continue with the example of the burn patient mentioned above; after receiving the engine's assessment of "high demand, low absorption efficiency, and rapid rise in blood sugar," the generator will not mechanically increase the infusion rate; instead, it will generate a refined, complex control instruction: 1) Rate regulation: To prevent further spikes in blood glucose and increased metabolic burden, temporarily maintain or slightly reduce the current infusion rate, rather than accelerate it; 2) Formula regulation: Dynamically adjust the ratio of multi-channel pumps, temporarily reduce the flow rate of carbohydrate components, correspondingly increase the proportion of fat emulsion (to provide non-carbohydrate energy), and may increase the infusion of specific amino acids (such as glutamine) to support intestinal mucosal repair. 3) Temperature control: In view of the possible fluctuations in the patient's peripheral circulation, the temperature of the nutrient solution will be precisely maintained at 37°C to optimize absorption and reduce intestinal irritation; 4) Planned Adjustment: Generate an "observation-increment" plan: After the current state is stable for 2 hours, if the blood sugar trend is stable, try to increase it at a slower step rate.

[0032] Preferably, the adjustable infusion execution module includes: The multi-channel intelligent infusion pump is used to independently control the infusion rate of multiple nutrient components and realize the dynamic mixing of the formula ratio. In the multi-channel intelligent infusion pump, a high-precision stepper motor drives multiple independent infusion channels to dynamically adjust the flow rate of each component and synthesize the target formula in real time in the mixer after the pump head.

[0033] For example, for a patient with severe pancreatitis and hyperglycemia, the intelligent analysis and control module determines, based on real-time blood glucose data, that the carbohydrate load needs to be reduced. It then issues a command to the multi-channel intelligent infusion pump: "Reduce the flow rate of the carbohydrate solution from 40 ml / hour to 25 ml / hour, while increasing the flow rate of the fat emulsion solution from 20 ml / hour to 35 ml / hour, to maintain the total energy supply unchanged." The multi-channel intelligent infusion pump responds immediately, precisely adjusting the motor speed of each channel, completing the mixing and infusion of the new formula within seconds, without requiring the nurse to manually change the entire nutrition bag, achieving a seamless and precise transition in treatment.

[0034] The intelligent temperature-controlled tubing is used to precisely heat and maintain the nutrient solution according to control commands. It consists of distributed heating elements and high-sensitivity temperature sensors built into the tubing, forming a closed-loop control. According to the target temperature set by the control commands (usually 37℃±0.5℃), the system monitors the temperature of the liquid in the tubing in real time and dynamically adjusts the heating power through a PID algorithm to achieve rapid heating and constant temperature maintenance from the nutrient bag outlet to the patient's nasogastric tube inlet, unaffected by ambient temperature or infusion rate.

[0035] For example, consider a patient recovering from shock who exhibits poor peripheral circulation and a low core body temperature. The system detects the patient's core body temperature at 35.5°C. The intelligent analysis and control module determines that administering fluid at 37°C might still cause relatively "cold stimulation" to the gastrointestinal tract. Therefore, it sends a command to the temperature control tubing: "Maintain a stable output fluid temperature of 38°C." The tubing heating system then activates, ensuring that every drop of nutrient solution is precisely heated to the set value before entering the patient's body, helping to maintain intestinal temperature, reducing the risk of intestinal spasms, and providing support for rewarming.

[0036] The identity security execution unit is used to perform dual identity verification before infusion by scanning the patient's and nutritional preparation's identification information. It integrates a QR code or RFID scanner. Before starting the infusion, the system will require the nurse to scan the identification tag on the patient's wristband and the preparation tag on the nutrition bag / pump. The unit will automatically compare the scanned information with the medical order information issued by the intelligent management platform. Only when the "patient identity-nutritional preparation-treatment order" are completely matched will the infusion pump be unlocked and start-up be allowed. If any of them do not match, the pump will be hardware locked and a clear warning will be issued.

[0037] Preferably, the human-computer interaction and early warning module includes: The bedside interactive terminal displays nutrition plans (such as target amount and pathway), real-time vital signs curves (such as blood glucose and bowel sounds), infusion parameters (such as infusion rate, temperature, and amount already infused), and tolerability assessment results; enabling medical staff to quickly grasp the overall situation at the bedside, understand the system's decision-making logic, and build confidence in treatment.

[0038] For example, a newly admitted nurse assigned to care for a patient receiving enteral nutrition can simply walk to the bedside and look at the terminal screen to see the following on a single interface: On the left is a summary of the patient's nutrition prescription (jejunal feeding, target 1500 kcal / day); in the middle is a real-time trend area showing that blood glucose (green curve) has remained stable within the ideal range over the past 4 hours, while bowel motility index (blue curve) is steadily rising; on the right is the execution status area, showing the current infusion rate of 65 ml / hour, fluid temperature of 37°C, and 820 ml infused today; at the bottom of the screen is a prominent green "well tolerated" status bar. The nurse completes a comprehensive assessment within 10 seconds without needing to consult documents or view multiple device screens.

[0039] The intelligent early warning system is used to issue graded alarms and push handling suggestions when it detects signs of feeding intolerance, metabolic disorders, or equipment malfunctions. The system divides alarms into multiple levels (such as "Alert / Yellow", "Warning / Orange", "Crisis / Red") based on the severity and urgency of the risk. While issuing audible and visual alarms, it pushes specific and contextualized handling suggestions on the screen, escalating the alarm from "notifying the problem" to "initiating a solution dialogue".

[0040] For example, in a patient who has undergone abdominal surgery, the system monitored: 1) Abdominal impedance data indicated significantly slowed gastric emptying; 2) Bowel sound characteristic analysis showed that peristalsis was almost absent; 3) The patient experienced a mild cough.

[0041] The intelligent early warning system immediately integrates this information, determines it as "high risk of reflux and aspiration," and triggers a "warning / orange" level alert; the pop-up window not only displays the alert but also clearly lists: suspected signs of gastric retention and paralytic ileus; recommendations: 1) Immediately suspend enteral nutrition infusion; 2) Assess the patient's breath sounds and abdominal signs; 3) Consider performing gastric residue testing.

[0042] The data synchronization interface is used to upload all execution, control, and early warning data of the device to the intelligent nutrition management platform; For example, for the patient mentioned above who triggered the "gastric retention warning," the nurse followed the advice to suspend feeding and conducted an assessment; the nurse recorded the assessment results of "300ml of gastric residue" and "weak bowel sounds" through the bedside terminal. The data synchronization interface immediately packaged and uploaded this complete event sequence—including the warning trigger time, system recommendation, infusion suspension instruction, and nurse's assessment results—to the intelligent nutrition management platform; The platform's statistical analysis unit then updates the patient's "feeding interruption record," while the algorithm model may use reinforcement learning to reinforce the association pattern of "slowed gastric emptying + weak bowel sounds → high risk of aspiration," providing earlier and more accurate warnings for similar patients in the future.

[0043] Preferably, the specific method for implementing nutrition screening and assessment in the nutrition screening and assessment unit is as follows: S11. Automatically obtain the patient's objective information from the interconnected hospital information system. The objective information includes: nursing assessment results, disease diagnosis and medical records, laboratory test results, anthropometric data and feeding history. Suppose a patient (Mr. Zhang, 45 years old) is admitted to the hospital with severe acute pancreatitis. After the initial medical record entry is completed upon admission, the nutrition screening and assessment unit automatically retrieves the following information about Mr. Zhang from the hospital's HIS, LIS, and nursing systems: Diagnosis: Severe acute pancreatitis (medical record describes signs of organ failure). Laboratory tests: Albumin 28 g / L (low), C-reactive protein 180 mg / L (significantly high); Anthropometric measurements: Height 175cm, weight 70kg (at admission), BMI 22.9; Feeding history: No special history, and the patient was kept from eating or drinking after admission; S12. Based on objective information, call the preset standard nutritional risk screening model to automatically complete the nutritional risk score and classification of the patient; the standard nutritional risk screening model includes NRS-2002 or NUTRIC scoring tools. Following the example in step S11; the system calls the NRS-2002 scoring model to automatically calculate the above information; due to the severity of the patient's disease (pancreatitis with organ failure, score 3), impaired nutritional status (decreased albumin, score 3), and age not exceeding 70 years (score 0), the system automatically calculates a total score of 6, determines "nutritional risk exists", marks the patient as a high-risk patient, and triggers the priority treatment process; S13. Based on the patient's disease status, metabolic level and anthropometric data, calculate their daily nutritional needs target, which includes: daily target energy intake and daily target protein intake. Following the example in step S12; based on the patient's disease status (high metabolic stress), weight (70kg), and activity level (bedridden), the system uses an energy calculation formula suitable for critically ill patients (such as an estimation formula based on the principle of indirect calorimetry) to calculate the daily target energy intake as 25–30 kcal / kg, or approximately 1750–2100 kcal; at the same time, according to the guidelines, the daily target protein intake is calculated to be 1.5–2.0 g / kg, or approximately 105–140 g; S14. Based on the medical knowledge base and logical rule model, and combined with the daily nutritional needs target, an individualized initial nutrition plan is automatically generated. The medical knowledge base and logical rule model are configured to automatically match nutrition plan rules according to specific disease states, including: matching a high-protein, high-calorie nutrition plan for patients with severe infections; and matching a low-protein or specific amino acid ratio nutrition plan for patients with renal dysfunction. Individualized initial nutrition plans are generated in the form of structured medical orders and automatically pushed to the doctor's workstation for confirmation or modification; the content includes: the type of nutritional preparation, the route of infusion, the initial infusion rate, and related nursing requirements; Following the example in step S13; the system accesses the medical knowledge base and matches the nutritional support rules for "severe acute pancreatitis": "jejunal feeding" is preferred to avoid stimulating the pancreas; the formula should be a "short peptide or amino acid" enteral nutrition preparation for easy absorption; the initial feeding rate should be slow; combined with the calculated nutritional goals, the system automatically generates a structured initial nutritional plan, as follows: Infusion route: via nasojejunal tube; Nutritional preparation: Short peptide enteral nutrition suspension (energy density 1.0 kcal / mL). Infusion plan: Start with a rate of 20 mL / h, and increase by 10–20 mL / h every 8–12 hours according to tolerance. The target infusion volume is dynamically adjusted according to the total energy target. Nursing requirements: Elevate the head of the bed ≥30°; assess abdominal signs and gastric / intestinal residue every 4 hours (if feasible).

[0044] The plan, along with its calculation basis, was automatically pushed to Mr. Zhang's attending physician's workstation computer. After the doctor reviewed and confirmed the plan, the system immediately sent it to the personalized enteral nutrition infusion control device at the bedside, initiating the intelligent execution and control process.

[0045] Preferably, the specific method for implementing intelligent statistical analysis in the statistical analysis unit is as follows: S21. Integrate infusion execution data from the adjustable infusion execution module and nursing assessment data from electronic nursing records to form a dataset of the entire patient treatment process; Suppose that a hospital's ICU wants to evaluate and improve the quality of enteral nutrition implementation over the past month; the statistical analysis unit runs automatically; it extracts detailed data logs from all personalized enteral nutrition infusion control devices at the ICU bedside throughout the hospital for the past 30 days, including: each patient's daily "target feeding amount", "actual infusion amount", "infusion rate curve", "interruption event records (such as pauses due to examination or intolerance)" and "equipment warning records (such as high gastric residual risk alarms)".

[0046] At the same time, relevant "nursing records" are retrieved from the electronic nursing system, such as daily records of "stool frequency and characteristics (for judging diarrhea)," "abdominal distension assessment," and "aspiration risk assessment." All these data are aligned and integrated according to patients and timelines to form a structured database. S22. Based on the full-process dataset, automatically calculate predefined quality management indicators, including: nutrition plan completion rate, feeding interruption rate, and complication rate; among which, Nutrition plan completion rate is calculated as the percentage of total energy intake actually consumed by the patient within the statistical period relative to the target total energy intake. Feeding interruption rate is calculated as the percentage of patients or events that experienced unplanned feeding interruptions within a statistical period relative to the total number of feedings or the total feeding time. The incidence of complications is calculated as the percentage of patients who develop specific gastrointestinal complications (including diarrhea, abdominal distension, vomiting, reflux, and aspiration) within the total number of monitored cases during the statistical period. Following the example in step S21; based on the above dataset, the system automatically calculates the core metrics: Nutrition plan completion rate: For all patients in the ICU who received enteral nutrition for more than 72 hours, the mean of their "actual total energy intake / target total energy intake" was calculated. For example, the calculation results showed that the overall average completion rate in the ICU was 68%. Feeding interruption rate: All unplanned interruption events were recorded. For example, 45 interruptions were found within 30 days, mainly attributed to "outpatient visits" (20 times) and "suspected intolerance" (15 times). Complication rate: Based on the criteria for "diarrhea (more than 3 watery stools per day)" in the nursing records, the system identified a total of 8 patients who developed diarrhea, accounting for 20% of the total number of patients fed during the same period (40 patients). S23. The calculated quality management indicators are dynamically displayed in the form of visual charts; the forms of visual charts include: dynamically updated bar charts, trend curves, pie charts or dashboards, and support data drill-down and filtering by ward, disease diagnosis, and time dimension. Following the example in step S22; the system automatically generates and updates the following charts on the nursing manager's data dashboard: A trend dashboard: a real-time dashboard showing the current overall nutrition program completion rate in the ICU (68%); A set of comparative bar charts: showing the average plan completion rate of patients under the care of different nursing teams (Group A, Group B, Group C) (e.g., Group A 75%, Group B 65%, Group C 64%), visually revealing the differences; A pie chart clearly shows the composition of reasons for feeding interruption (e.g., out-of-town examination 44%, suspected intolerance 33%, other 23%). A complication trend curve: showing the weekly trend of diarrhea incidence over the past 30 days; Interactive features: When an administrator clicks on the bar chart for "Group B 65%", they can drill down to see which specific patients in Group B have low completion rates, and further view the daily infusion curve and interruption records of one of these patients.

[0047] S24. Based on the historical trend analysis of quality management indicators, when the indicator value deviates from the preset threshold or shows a deteriorating trend, a quality early warning prompt will be automatically generated. Following the example in step S23; the system sets thresholds based on historical data, and automatically sends a quality warning to the nursing department when it finds that "the overall plan completion rate is below 70% for 3 consecutive days" or "the incidence of diarrhea exceeds 15% in a single week"; By combining visualization charts, managers can quickly pinpoint problems; for example, they may find that low completion rates are mainly related to interruptions caused by "out-of-town examinations," thereby promoting process optimization with departments such as radiology and operating rooms, and developing standardized "enteral nutrition patient transport and feeding programs"; or they may find that high rates of diarrhea are related to a certain batch of nutritional preparations or excessively rapid growth, thereby revising the initial feeding program.

[0048] Please see Figure 2 In this embodiment of the invention, a method for regulating personalized enteral nutrition infusion based on multi-dimensional vital sign perception includes the following steps: S1. The intelligent nutrition management platform automatically obtains patient information, completes nutritional screening and assessment, and generates an individualized initial nutrition plan. For example, for a patient (Mr. Wang) who was admitted to the ICU after surgery for "severe acute pancreatitis", the intelligent nutrition management platform automatically obtains his information from HIS and LIS after admission: diagnosis of "severe acute pancreatitis, acute respiratory distress syndrome", and laboratory results showing "albumin 28g / L, C-reactive protein 200mg / L". The platform automatically calls the NUTRIC score and determines that it is at high nutritional risk. Based on the disease rules (pancreatitis requires jejunal feeding and the use of easily absorbed formula) and metabolic calculations, it automatically generates an initial plan: "Infuse short peptide nutritional formula via nasojejunal tube, starting rate 20ml / hour, target energy 1500kcal / day". In this way, a personalized, structured nutrition prescription can be generated within minutes and sent to the attending physician for review and confirmation; this can effectively avoid the delays and errors caused by manually reviewing data, calculating and writing prescriptions in the traditional model. S2. The initial nutrition plan is sent to the designated personalized enteral nutrition infusion control device, and the infusion is started after the identity is verified. Following the example in step S1; after the doctor's order is confirmed, the plan is sent to the infusion control device next to Mr. Wang's bedside; when the nurse is ready to operate, the device forces a "double scan": first scan the patient's wristband, and then scan the barcode on the nutritional preparation bag; Once the information is successfully matched, the device unlocks and allows startup; this fundamentally eliminates serious errors such as "giving to the wrong patient" or "using the wrong nutrient solution"; the infusion begins according to the initial protocol (20ml / hour); S3. While the device is performing infusion, it collects data in real time through the multi-parameter vital sign sensing module, and the intelligent analysis and control module dynamically assesses the patient's status, generates and executes an adaptive regulation strategy, and adjusts the infusion parameters in real time. Following the example in step S2; after the infusion begins, the device continues to operate: 1) Sensing: Continuous glucose monitoring (CGM) showed blood glucose rising to 12 mmol / L; abdominal impedance sensor indicated slow bowel movements; 2) Assessment: The intelligent analysis and control module integrates the data and determines the status as "high blood sugar, gastrointestinal intolerance alert"; 3) Adjustment: The module automatically generates instructions: ① Reduce the infusion rate from 20ml / hour to 15ml / hour; ② Temporarily reduce the proportion of carbohydrate components in the formula.

[0049] Then, the adjustable infusion execution module responds immediately, adjusting the pump speed and formula; a few hours later, blood glucose drops back to 9 mmol / L, gastrointestinal data improves, and the system automatically and gradually increases the speed; S4. The device will synchronize the infusion execution data, parameter adjustment records and early warning information back to the intelligent nutrition management platform in real time. Following the example in step S3; all the above events (initial infusion, blood glucose rise alarm, rate adjustment, formula fine-tuning, parameter recovery) are fully recorded with millisecond-level timestamps; These high-fidelity execution and response data streams are transmitted back to the central intelligent nutrition management platform in real time through data interfaces, forming a complete and traceable "treatment digital twin" that provides a solid foundation for analysis and optimization. S5. The intelligent nutrition management platform performs statistical analysis on the received data throughout the entire process, generating a quality management report for optimizing treatment plans or clinical decisions; following the example in step S4; at the end of the month, the ICU head nurse reviews the quality report through the platform's statistical analysis unit; she finds: Overall trend: The department's average "nutrition plan completion rate" was 72%, which fell short of the target of 80%. Drill-down analysis: Clicking on this indicator revealed a significantly low completion rate for the "pancreatitis patients" group (average only 65%). Further investigation showed the main reason was a high "feeding interruption rate," with the primary reason for interruption labeled as "suspected intolerance." Reviewing the case: She retrieved detailed records of several pancreatitis patients, including Mr. Wang, and found that the system repeatedly slowed down or paused due to high blood sugar or poor gastrointestinal condition.

[0050] Based on this data insight, the department organized a discussion and revised the clinical pathway: for patients with severe pancreatitis, the initial infusion rate was adjusted from a uniform 20ml / hour to a more conservative 15ml / hour, and more refined blood glucose control rules were set in advance.

[0051] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A personalized enteral nutrition infusion control device based on multi-dimensional vital sign perception, characterized in that, include: Multi-parameter vital sign sensing module, intelligent analysis and control module, adjustable infusion execution module, human-computer interaction and early warning module; and an intelligent nutrition management platform that communicates with the above modules; wherein, The multi-parameter vital sign sensing module is used to collect users' multi-dimensional physiological, metabolic, and gastrointestinal vital sign data in real time and to preprocess the data; specifically, it includes: A physiological parameter sensing unit is used to monitor heart rate, respiration, body temperature, blood pressure, and blood oxygen saturation via an integrated or wirelessly connected wearable device. The metabolic and gastrointestinal sensing unit is used to monitor metabolic indicators and gastrointestinal status; among which... Metabolic indicator monitoring includes: acquiring blood glucose or electrolyte data using a continuous glucose monitoring sensor or a non-invasive electrolyte analysis patch; Gastrointestinal status monitoring includes: using abdominal bioelectrical impedance sensors to monitor gastric emptying and intestinal contents flow, or using a bowel sound acquisition array to identify bowel sound characteristics through AI to assess intestinal peristalsis status; The intelligent analysis and control module is used to fuse and analyze sensory data and nutritional plans based on machine learning algorithms, dynamically assess nutritional needs and tolerance, and generate individualized infusion rate, temperature, and formulation control strategies; specifically, it includes: The data fusion and status assessment engine is used to fuse the initial nutrition plan with real-time data from the multi-parameter vital sign perception module, and dynamically outputs real-time nutritional requirement index, gastrointestinal tolerance level and absorption efficiency prediction through machine learning model. An adaptive control strategy generator is used to generate control commands for real-time adjustment of infusion rate, nutrient formulation ratio and infusion temperature based on the evaluation results of dynamic output. An adjustable infusion execution module is used to precisely control the flow rate of the multi-channel infusion pump and the intelligent temperature control pipeline according to the regulation strategy, so as to achieve precise delivery and temperature regulation of nutrient solution, and integrates identity verification to ensure execution safety; specifically, it includes: A multi-channel intelligent infusion pump is used to independently control the infusion rate of multiple nutrient components, enabling dynamic mixing of the formula ratio; Intelligent temperature control pipeline is used to precisely heat and maintain the temperature of the nutrient solution according to control commands. The identity security execution unit is used to perform dual identity verification before infusion by scanning the patient's and nutritional preparation's identification information; The human-computer interaction and early warning module provides users with parameter settings, real-time status visualization, and intelligent tiered early warnings when anomalies are detected. It is also responsible for bidirectional communication of all data. Specifically, it includes: The bedside interactive terminal is used to display nutrition plans, real-time vital signs curves, infusion parameters, and tolerance assessment results; The intelligent early warning system is used to issue tiered alarms and push treatment suggestions when signs of feeding intolerance, metabolic disorders or equipment malfunctions are detected. The data synchronization interface is used to upload all execution, control, and early warning data of the device to the intelligent nutrition management platform; The intelligent nutrition management platform is used to automatically acquire and integrate patient information, complete nutritional risk screening and assessment, generate individualized initial nutrition plans, and perform statistical analysis on data throughout the treatment process; it consists of a nutrition screening and assessment unit and a statistical analysis unit.

2. A personalized enteral nutrition infusion control device based on multi-dimensional vital sign perception according to claim 1, characterized in that, The specific methods by which the nutrition screening and assessment unit performs nutrition screening and assessment are as follows: S11. Automatically obtain objective patient information from interconnected hospital information systems; S12. Based on objective information, call the preset standard nutritional risk screening model to automatically complete the nutritional risk score and classification of the patient; S13. Calculate the daily nutritional needs target based on the patient's disease status, metabolic level and anthropometric data; S14. Based on a medical knowledge base and logical rule model, and combined with daily nutritional needs targets, automatically generate an individualized initial nutrition plan.

3. A personalized enteral nutrition infusion control device based on multi-dimensional vital sign perception according to claim 1, characterized in that, The specific method by which the statistical analysis unit implements intelligent statistical analysis is as follows: S21. Integrate infusion execution data from the adjustable infusion execution module and nursing assessment data from electronic nursing records to form a dataset of the entire patient treatment process; S22. Based on the full-process dataset, automatically calculate predefined quality management indicators, including: nutrition plan completion rate, feeding interruption rate, and complication rate; S23. Dynamically display the calculated quality management indicators in the form of visual charts; the forms of visual charts include: dynamically updated bar charts, trend curves, pie charts, or dashboards, and support data drill-down and filtering by ward, disease diagnosis, and time dimension; S24. Based on the historical trend analysis of quality management indicators, when the indicator value deviates from the preset threshold or shows a deteriorating trend, a quality early warning prompt will be automatically generated.

4. A method for controlling a personalized enteral nutrition infusion control device for multi-dimensional vital sign perception as described in any one of claims 1-3, characterized in that, Includes the following steps: S1. The intelligent nutrition management platform automatically acquires patient information, completes nutrition screening and assessment, and generates an individualized initial nutrition plan. S2. The initial nutrition plan is sent to the designated personalized enteral nutrition infusion control device, and the infusion is started after the identity is verified. S3. While the device is performing infusion, it collects data in real time through the multi-parameter vital sign sensing module, and the intelligent analysis and control module dynamically assesses the patient's status, generates and executes an adaptive regulation strategy, and adjusts the infusion parameters in real time. S4. The device will synchronize the infusion execution data, parameter adjustment records and early warning information back to the intelligent nutrition management platform in real time. S5, the intelligent nutrition management platform, performs statistical analysis on the received data throughout the entire process and generates quality management reports to optimize treatment plans or clinical decisions.

Images