A laparoscopy postoperative shoulder pain syndrome prevention and control and comfort improvement method and device

Abstract

The present application relates to a kind of laparoscopic postoperative shoulder pain syndrome prevention and control and comfort degree promotion method and device, wherein the method comprises: obtaining the multi-modal medical data of patient including individual physiological characteristics, intraoperative real-time monitoring parameter and pain sensitivity related gene group composition;Multi-modal medical data is input into pre-trained timing-multiscale pain risk model, and the probability of moderate shoulder pain within 24 hours after operation and risk classification are output;The probability of shoulder pain and risk classification, patient real-time feedback and situational information are input into pre-trained individualized analgesia and comfort degree model, and the optimal individualized analgesia and comfort degree intervention scheme combination is output, for medical staff to carry out personalized intervention according to patient.The present application predicts pain risk according to timing-multiscale model and optimizes drug dose by pharmacokinetic model, significantly improves the postoperative experience and rehabilitation quality of patient in the prevention and control and comfort degree promotion of laparoscopic postoperative shoulder pain.

Description

Technical Field

[0001] This invention relates to the field of comprehensive prevention and control technology for shoulder pain after laparoscopic surgery, specifically to a method and device for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery, and computer equipment. Background Technology

[0002] Postoperative complications after laparoscopic surgery remain a significant concern during patient recovery. Postoperative shoulder pain (PLSP) is one of the most common and prevalent complications of laparoscopic surgery. PLSP typically manifests as unilateral or bilateral shoulder pain that can radiate to the neck and back, severely impacting postoperative recovery quality and comfort, potentially prolonging hospital stays and increasing the medical burden. Pain relief not only improves patients' mood and reduces the risk of anxiety and depression but also promotes early mobilization and reduces the incidence of complications such as deep vein thrombosis. Improving patient comfort helps shorten hospital stays, reduce medical costs, and enhance the quality of medical services and patient satisfaction. Therefore, developing methods that can accurately predict shoulder pain risk and provide personalized intervention plans has significant clinical value for modern surgical perioperative management.

[0003] Currently, the prevention and management of shoulder pain after laparoscopic surgery in clinical practice mainly includes intraoperative CO2 pressure management, postoperative analgesia, physical therapy, and postural adjustments. Among these, reducing pneumoperitoneum pressure is considered an effective way to reduce diaphragmatic stimulation, but excessively low pressure may affect the surgical field and operating space. Postoperative analgesia mainly includes nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, and local anesthetics. Dosing regimens usually follow standardized dosage guidelines, but individual patient differences (such as weight, age, liver and kidney function, and genetic polymorphisms) may lead to poor drug efficacy or adverse reactions. Physical therapy and postural adjustments, such as early postoperative mobilization, hot compresses, and postural drainage, are usually used as adjunctive methods and their implementation lacks systematic and individualized guidance.

[0004] To address the aforementioned issues, this invention proposes a comprehensive method for preventing and controlling shoulder pain and improving comfort after laparoscopic surgery, which significantly improves patients' postoperative experience and rehabilitation quality in terms of preventing and controlling shoulder pain and improving comfort after laparoscopic surgery. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method, device and computer equipment for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery, in order to overcome the shortcomings of the prior art.

[0006] According to one aspect of the present invention, a method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery is provided, comprising:

[0007] Acquire multimodal medical data from the patient, including individual physiological characteristics, intraoperative real-time monitoring parameters, and genomic data related to pain sensitivity; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... Pressure, perfusion rate, pneumoperitoneum duration, operation duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters;

[0008] The multimodal medical data is input into a pre-trained time-series-multiscale pain risk model, which outputs the probability and risk level of moderate to severe shoulder pain in patients within 24 hours after surgery.

[0009] The probability and risk grading of shoulder pain, real-time patient feedback, and contextual information are input into a pre-trained individualized analgesia and comfort model, which outputs the optimal combination of individualized analgesia and comfort intervention plans for medical staff to provide personalized interventions for the patients. The intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks.

[0010] In one alternative approach, the probability of the patient experiencing moderate to severe shoulder pain within 24 hours post-surgery is calculated using the following formula:

[0011]

[0012] in, For the intercept term; The weight coefficient for the j-th physiological characteristic; These are standardized physiological characteristic values; The number of physiological characteristics; This is the coefficient for operation time; This refers to the duration of the surgery. Genomic risk coefficient; Pressure cumulative exposure coefficient; Intraperitoneal space at time point t pressure; The safe pressure threshold; This refers to the duration of pneumoperitoneum maintenance. This refers to the anesthesia index coefficient. It is a comprehensive score for the type and dosage of anesthetic drugs.

[0013] In one alternative approach, the individualized analgesic intervention regimen optimizes the dosage based on a population pharmacokinetic model of flurbiprofen ester, wherein the optimized dosage extension formula is:

[0014]

[0015] in, To achieve the target blood drug concentration; It is the apparent distribution volume; To eliminate the rate constant; Bioavailability; , These are the covariate effect coefficients; This is a random effect between individuals; This is the proportional error term.

[0016] In one alternative approach, the time-series-multiscale pain risk model includes:

[0017] The static individual feature input layer is used to receive static medical data including individual physiological characteristics, operation duration, genomic data related to pain sensitivity, and types and dosages of anesthetic drugs. It performs preliminary feature mapping through a multilayer perceptron, a genome encoder, and an anesthesia encoder to generate high-dimensional static feature vectors.

[0018] The dynamic intraoperative parameter temporal input branch layer is used to receive data including intraperitoneal parameters. Multi-dimensional time-series data, including pressure, perfusion rate, specific operational intensity, body temperature, and hemodynamic parameters, are collected, and each dimension of the time-series data is equipped with an independent time-series encoder to capture its respective time dependence; wherein, the... The pressure timing encoder incorporates an attention unit to identify and weight data exceeding a safety threshold. Pressure exposure period calculated using differentiable accumulation layers Dynamic temporal feature vector of cumulative stress exposure;

[0019] A multi-scale feature fusion layer fuses the static feature vector with the dynamic temporal feature vector to generate a potential representation of pain risk;

[0020] The output layer maps the potential representation of the pain risk to Logit values ​​and outputs the probability of the patient experiencing moderate to severe shoulder pain within 24 hours post-surgery, along with the corresponding risk grading.

[0021] In one alternative approach, the individualized analgesia and comfort model includes:

[0022] The input layer is used to receive and encode the probability and risk grading of shoulder pain, real-time patient feedback, and contextual information.

[0023] The multimodal feature fusion layer includes a cross-modal attention fusion module, which extracts information from each modality and uses cross-attention to capture deep interactions and dependencies between different modalities. The fused features are then concatenated and dimensionality reduced to generate a unified comprehensive feature representation.

[0024] The intervention plan generation layer includes an intervention strategy graph neural network consisting of a measure selection head, a dose / intensity prediction head, a frequency / duration prediction head, an expected effect prediction head, a potential risk prediction head, and a sequence decision and proxy unit. This network is used to model multiple preset intervention measures as a graph structure, where the nodes of the graph structure represent intervention measures, and the edges of the graph structure represent synergistic, antagonistic, or temporal relationships between the nodes.

[0025] In one alternative approach, the intervention strategy graphical neural network further includes:

[0026] An intervention graph construction layer is used to model various potential analgesic and comfort interventions as nodes in the graph and initialize them as feature vectors through an embedding layer. At the same time, based on the knowledge of medical experts, the synergistic, antagonistic, temporal dependence and classification association edges between nodes are preset or dynamically learned.

[0027] A multi-layer graph attention network consists of stacked graph attention units. Each graph attention unit learns attention weights to selectively aggregate information about neighboring nodes in order to capture the complex contextual dependencies between intervention measures.

[0028] The comprehensive feature integration layer is used to perform attention fusion between the node features processed by the multi-layer graph attention network layer and the comprehensive feature representation output by the multimodal feature fusion layer, so as to ensure that the generation of intervention plans fully considers the intrinsic relationship between the patient's individualized multimodal data and the measures.

[0029] The system includes a multi-head output decoder and a sequence decision-making and proxy unit. The multi-head output decoder outputs in parallel the selection of measures, drug dosage, intervention frequency and duration, expected effects and potential risks. The sequence decision-making and proxy unit interacts with the simulated patient model in real time and uses the patient's pain relief, comfort improvement and side effect incidence as reward signals to dynamically learn and optimize the sequence of intervention strategies at different time steps.

[0030] In one alternative approach, the contextual information includes postoperative time, current medical environment, patient position, family accompaniment, patient's emotional state, history of pain, allergies, comorbidities, medication adherence, pain tolerance, preference for interventions, and financial affordability.

[0031] The real-time feedback information includes pain score, pain location, pain nature, pain frequency, nausea and vomiting, drowsiness, itching, respiratory depression, and the patient's subjective evaluation of their current comfort level.

[0032] In one alternative approach, the personalized intervention includes:

[0033] Drug intervention involves recommending the type, individualized dosage, and route of administration of nonsteroidal anti-inflammatory drugs, opioids, local anesthetics, or other adjunctive analgesics based on the patient's pain level, drug metabolism characteristics, comorbidities, and allergy history, and dynamically adjusting the frequency and duration of administration.

[0034] Physical therapy, depending on the location and nature of the pain, recommends hot compresses, cold compresses, body position adjustments, specific posture guidance, or gentle massage to relieve muscle tension and promote blood circulation.

[0035] Information support and education provide patients and their families with explanations of the causes of postoperative shoulder pain, the expected course of the disease, treatment options, and self-management methods to enhance patient compliance and participation and reduce anxiety caused by uncertainty;

[0036] Nutritional support includes recommending dietary recommendations based on the patient's individual needs to help reduce inflammation and promote tissue repair.

[0037] According to another aspect of the present invention, a device for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery is provided, comprising:

[0038] The data acquisition module is used to acquire multimodal medical data of the patient, including individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... Pressure, perfusion rate, pneumoperitoneum duration, operation duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters;

[0039] The pain risk assessment module is used to input the multimodal medical data into a pre-trained time-series-multiscale pain risk model and output the probability and risk level of the patient experiencing moderate to severe shoulder pain within 24 hours after surgery.

[0040] The intervention plan generation module is used to input the probability and risk classification of the shoulder pain, real-time patient feedback, and contextual information into a pre-trained individualized analgesia and comfort model, and output the optimal combination of individualized analgesia and comfort intervention plans for medical staff to carry out personalized interventions for the patient; wherein, the intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks.

[0041] According to another aspect of the present invention, a computer device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0042] The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery.

[0043] According to the solution provided by the present invention, multimodal medical data of the patient is acquired, comprising individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... The multimodal medical data, including pressure, perfusion rate, pneumoperitoneum maintenance time, surgical duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters, are input into a pre-trained time-series-multiscale pain risk model. This model outputs the probability and risk stratification of moderate to severe shoulder pain within 24 hours post-surgery. The probability and risk stratification of shoulder pain, real-time patient feedback, and contextual information are then input into a pre-trained individualized analgesia and comfort model, which outputs the optimal combination of individualized analgesia and comfort interventions for healthcare professionals to implement personalized interventions for the patient. The intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks. This invention, based on a time-series-multiscale model to predict pain risk and a pharmacokinetic model to optimize drug dosage, significantly improves the postoperative experience and recovery quality of patients in terms of preventing and controlling shoulder pain and enhancing comfort after laparoscopic surgery.

[0044] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0045] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0046] Figure 1 A flowchart illustrating the method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to an embodiment of the present invention is shown.

[0047] Figure 2 A schematic diagram of the frame of the device for comprehensive prevention and control of shoulder pain after laparoscopic surgery and improvement of comfort according to an embodiment of the present invention is shown;

[0048] Figure 3 A schematic diagram of the structure of a computer device according to an embodiment of the present invention is shown. Detailed Implementation

[0049] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0050] Figure 1 This diagram illustrates a flowchart of a method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to an embodiment of the present invention. Specifically, as shown... Figure 1 As shown, it includes the following steps:

[0051] Step S101: Acquire multimodal medical data of the patient, including individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... Pressure, perfusion rate, duration of pneumoperitoneum, duration of operation, intensity of specific procedures, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters.

[0052] In this embodiment, intraoperative real-time monitoring parameters reflect the cumulative effect of stress on the body and the instantaneous stress response during surgery. Combining this dynamic information with static individual physiological characteristics and genomic data provides a more comprehensive understanding of the mechanisms of pain development and provides rich and valuable input for time-series-multiscale pain risk models. Genomic data can be obtained preoperatively and, combined with physiological characteristics, can be used to perform preliminary pain sensitivity stratification of patients. Intraoperative real-time monitoring data provides ongoing information, allowing the model to dynamically adjust risk assessments to achieve earlier and more forward-looking risk warnings. For example, Patient A: 35-year-old female, normal BMI, no underlying diseases. Intraoperative real-time monitoring parameters: prolonged pneumoperitoneum maintenance time, slightly large fluctuations in intraperitoneal CO2 pressure at a specific time period, high intensity of specific operations (such as traction), and low dose of anesthetic drugs (fentanyl). Genomic data related to pain sensitivity: Gene testing shows that her CYP2D6 gene (which affects the metabolism of certain opioids) is an ultra-rapid metabolizer, and there are SNPs (single nucleotide polymorphisms) associated with increased pain sensitivity.

[0053] Step S102: Input the multimodal medical data into the pre-trained time-series-multiscale pain risk model and output the probability and risk classification of the patient experiencing moderate to severe shoulder pain within 24 hours after surgery.

[0054] In this embodiment, the dynamic information of various intraoperative parameters changing over time is captured temporally, such as cumulative CO2 pressure exposure, rather than just static values. Multi-scale analysis at different time granularities (e.g., intraoperative data at the minute level and postoperative observation at the hour level) or different physiological levels (cells, organs, systems) can improve the accuracy of predictions. It can provide the probability of moderate to severe shoulder pain occurring within 24 hours postoperatively, enabling early warning and buying valuable intervention time for medical staff, transforming passive treatment into proactive prevention.

[0055] In one alternative approach, the probability of the patient experiencing moderate to severe shoulder pain within 24 hours post-surgery is calculated using the following formula:

[0056]

[0057] in, For the intercept term; The weight coefficient for the j-th physiological characteristic; These are standardized physiological characteristic values; The number of physiological characteristics; This is the coefficient for operation time; This refers to the duration of the surgery. Genomic risk coefficient; Pressure cumulative exposure coefficient; Intraperitoneal space at time point t pressure; The safe pressure threshold; This refers to the duration of pneumoperitoneum maintenance. This refers to the anesthesia index coefficient. It is a comprehensive score for the type and dosage of anesthetic drugs.

[0058] In this embodiment, The cumulative stress exposure was accumulated through integration. The time-intensity effect of pressure exceeding the safety threshold. The longer the pressure remains high or fluctuates repeatedly beyond the safe range, the greater its contribution to the risk of shoulder pain, reflecting the impact of the dynamic process of intraoperative manipulation on postoperative outcomes.

[0059] In one alternative approach, the time-series-multiscale pain risk model includes:

[0060] The static individual feature input layer is used to receive static medical data including individual physiological characteristics, operation duration, genomic data related to pain sensitivity, and types and dosages of anesthetic drugs. It performs preliminary feature mapping through a multilayer perceptron, a genome encoder, and an anesthesia encoder to generate high-dimensional static feature vectors.

[0061] The dynamic intraoperative parameter temporal input branch layer is used to receive data including intraperitoneal parameters. Multi-dimensional time-series data, including pressure, perfusion rate, specific operational intensity, body temperature, and hemodynamic parameters, are collected, and each dimension of the time-series data is equipped with an independent time-series encoder to capture its respective time dependence; wherein, the... The pressure timing encoder incorporates an attention unit to identify and weight data exceeding a safety threshold. Pressure exposure period calculated using differentiable accumulation layers Dynamic temporal feature vector of cumulative stress exposure;

[0062] A multi-scale feature fusion layer fuses the static feature vector with the dynamic temporal feature vector to generate a potential representation of pain risk;

[0063] The output layer maps the potential representation of the pain risk to Logit values ​​and outputs the probability of the patient experiencing moderate to severe shoulder pain within 24 hours post-surgery, along with the corresponding risk grading.

[0064] In this embodiment, for The pressure timing encoder incorporates an attention unit and a differentiable accumulation layer, enabling it to intelligently identify and weight processes pressures exceeding safety thresholds. The period of stress exposure. It can not only monitor... Stress can also identify harmful stress exposure moments and take into account their cumulative effects, directly linking to the physiological mechanisms of shoulder pain after laparoscopic surgery (such as diaphragmatic stimulation). Data of different scales and properties are processed through a static individual feature input layer and a dynamic intraoperative parameter temporal input branch layer, and then fused in a multi-scale feature fusion layer to ensure that static, long-term patient characteristics and dynamic, short-term surgical process information are organically combined to jointly influence the prediction of pain risk.

[0065] Step S103: Input the probability and risk classification of shoulder pain, real-time patient feedback and contextual information into a pre-trained individualized analgesia and comfort model, and output the optimal combination of individualized analgesia and comfort intervention plans for medical staff to carry out personalized intervention for the patient; wherein, the intervention plan includes specific measures, dosage, frequency, expected effect and potential risks.

[0066] In this embodiment, the intervention plan not only specifies "what to do," but also clarifies "how to do it" (dosage, frequency), "what the possible effects are" (expected results), and "what precautions to take" (potential risks). This provides healthcare professionals with comprehensive decision support information, reducing the uncertainty of clinical judgment. By anticipating potential risks, healthcare professionals can take preventative measures in advance or choose lower-risk alternatives, thereby improving treatment safety.

[0067] For example, two patients, A and B, both underwent laparoscopic cholecystectomy and reported shoulder pain postoperatively.

[0068] In the traditional approach, both patients A and B scored 6 points (moderate pain) in pain assessment. Following departmental routine, patients A and B were given the same dose of nonsteroidal anti-inflammatory drugs (NSAIDs), such as flurbiprofen 100mg intravenously twice daily. In this application, based on multimodal medical data (physiological characteristics, intraoperative CO2 pressure, surgical duration, genomic data, anesthesia, etc.) for patients A and B, the following predictions were made: Patient A: 85% probability of experiencing moderate to severe shoulder pain within 24 hours post-surgery, risk level: high risk; Patient B: 40% probability of experiencing moderate to severe shoulder pain within 24 hours post-surgery, risk level: low to medium risk. Next, the prediction results, real-time patient feedback, and contextual information were input into an individualized analgesia and comfort model. In Patient A's personalized intervention plan (high risk, considering multiple factors), real-time feedback indicated a pain score of 7 points, dull pain in the left shoulder radiating to the neck, and mild nausea. The patient indicated sensitivity to drug side effects (such as stomach discomfort) but desired rapid pain relief. The patient, accompanied by family members, appears slightly anxious. Contextual information: 6 hours post-surgery, quiet medical environment, patient in a semi-recumbent position. History of mild gastritis, allergic to ibuprofen. Average pain tolerance, good financial capacity. Optimal treatment plan based on model output: Drug intervention: Based on the flurbiprofen ester population pharmacokinetic model combined with patient A's BMI and CYP2C9 genotype, the optimized dose is flurbiprofen ester 80mg intravenously every 10 hours (instead of the usual twice daily, to maintain a more stable blood drug concentration), and combined use of a gastric mucosal protectant is recommended. Considering the nature of the pain (radiation), it is recommended that after assessment 12 hours post-surgery, if pain is not relieved, topical anesthetic (such as lidocaine patch) or oral acetaminophen tablets 500mg may be considered, if necessary. Prophylactic antiemetic: ondansetron 4mg intravenously to relieve nausea. Expected outcome: Pain score is expected to decrease to 3-4 points within 1 hour, nausea relief. Physical therapy: It is recommended that healthcare professionals guide patients to adjust their posture (such as slightly tilting to the healthy side, elevating the head) to avoid maintaining the same position for extended periods. Gentle shoulder massage (non-surgical area) can be performed with the help of family members to promote local blood circulation. Information support and education: Healthcare professionals should explain to patients and their families in detail that shoulder pain is a common phenomenon caused by pneumoperitoneum and is not a surgical complication, usually subsiding within 24-48 hours. Inform them about the use of pain medication, side effects, and coping strategies, and encourage deep breathing relaxation exercises. Relieve anxiety. Nutritional support: It is recommended to drink plenty of water to promote blood circulation. Consuming fluids and eating a light diet, avoiding spicy and irritating foods, can help alleviate stomach discomfort.

[0069] In one alternative approach, the individualized analgesic intervention regimen optimizes the dosage based on a population pharmacokinetic model of flurbiprofen ester, wherein the optimized dosage extension formula is:

[0070]

[0071] in, To achieve the target blood drug concentration; It is the apparent distribution volume; To eliminate the rate constant; Bioavailability; , These are the covariate effect coefficients; This is a random effect between individuals; This is the proportional error term.

[0072] In this embodiment, CYP2C9 is an important drug-metabolizing enzyme, and its gene polymorphism significantly affects the metabolic rate of flurbiprofen ester. Body mass index (BMI) is associated with the drug's volume of distribution and clearance. By using CYP2C9 genotype and BMI as covariates, the pharmacokinetic behavior of the drug in specific patients can be predicted more accurately, thus achieving truly personalized dosing and avoiding standardized "one-size-fits-all" dosing. Even after considering known covariates, differences in pharmacokinetic parameters may still exist between individuals. Inter-individual random effects represent inter-individual variability not fully explained by the model, allowing the model to better adapt to individual differences and further improve the accuracy of dose prediction.

[0073] In one alternative approach, the individualized analgesia and comfort model includes:

[0074] The input layer is used to receive and encode the probability and risk grading of shoulder pain, real-time patient feedback, and contextual information.

[0075] The multimodal feature fusion layer includes a cross-modal attention fusion module, which extracts information from each modality and uses cross-attention to capture deep interactions and dependencies between different modalities. The fused features are then concatenated and dimensionality reduced to generate a unified comprehensive feature representation.

[0076] The intervention plan generation layer includes an intervention strategy graph neural network consisting of a measure selection head, a dose / intensity prediction head, a frequency / duration prediction head, an expected effect prediction head, a potential risk prediction head, and a sequence decision and proxy unit. This network is used to model multiple preset intervention measures as a graph structure, where the nodes of the graph structure represent intervention measures, and the edges of the graph structure represent synergistic, antagonistic, or temporal relationships between the nodes.

[0077] In this embodiment, the intervention plan generation layer employs an intervention strategy graph neural network structure, modeling multiple pre-defined intervention measures (such as medication, physical therapy, and psychological intervention) as nodes in the graph. This allows for flexible representation of the "synergistic," "antagonistic," or "temporal" relationships between measures. For example, some medications may have synergistic effects, while others may be antagonistic, and some physical therapies need to be administered after medication intervention. Therefore, this aligns better with the concept of multimodal analgesia in clinical practice, enabling the layer to learn and utilize the complex relationships between measures to recommend optimal combinations (e.g., in specific situations, administering small doses of two synergistic drugs simultaneously may be more effective and have fewer side effects than administering a large dose of a single drug). Simultaneously, it identifies and avoids recommending antagonistic combinations of measures, reducing potential medical risks.

[0078] In one alternative approach, the intervention strategy graphical neural network further includes:

[0079] An intervention graph construction layer is used to model various potential analgesic and comfort interventions as nodes in the graph and initialize them as feature vectors through an embedding layer. At the same time, based on the knowledge of medical experts, the synergistic, antagonistic, temporal dependence and classification association edges between nodes are preset or dynamically learned.

[0080] A multi-layer graph attention network consists of stacked graph attention units. Each graph attention unit learns attention weights to selectively aggregate information about neighboring nodes in order to capture the complex contextual dependencies between intervention measures.

[0081] The comprehensive feature integration layer is used to perform attention fusion between the node features processed by the multi-layer graph attention network layer and the comprehensive feature representation output by the multimodal feature fusion layer, so as to ensure that the generation of intervention plans fully considers the intrinsic relationship between the patient's individualized multimodal data and the measures.

[0082] The system includes a multi-head output decoder and a sequence decision-making and proxy unit. The multi-head output decoder outputs in parallel the selection of measures, drug dosage, intervention frequency and duration, expected effects and potential risks. The sequence decision-making and proxy unit interacts with the simulated patient model in real time and uses the patient's pain relief, comfort improvement and side effect incidence as reward signals to dynamically learn and optimize the sequence of intervention strategies at different time steps.

[0083] In this embodiment, the multi-layer graph attention network layer learns attention weights to selectively aggregate information from neighboring nodes. When considering a certain intervention, the model dynamically assigns different weights based on the current patient's specific condition and the characteristics of neighboring interventions, thereby more accurately assessing its impact. The multi-head output decoder outputs in parallel the intervention selection, drug dosage, intervention frequency and duration, expected effects, and potential risks, providing a comprehensive and actionable intervention plan. The sequence decision-making and proxy unit, through real-time interaction with the simulated patient model and reward signals (pain relief, comfort improvement, and side effect incidence), achieves dynamic learning and sequence optimization of the intervention strategy, continuously learning from experience and adjusting its long-term intervention strategy to achieve the best results.

[0084] In one alternative approach, the contextual information includes postoperative time, current medical environment, patient position, family accompaniment, patient's emotional state, history of pain, allergies, comorbidities, medication adherence, pain tolerance, preference for interventions, and financial affordability.

[0085] The real-time feedback information includes pain score, pain location, pain nature, pain frequency, nausea and vomiting, drowsiness, itching, respiratory depression, and the patient's subjective evaluation of their current comfort level.

[0086] In this embodiment, by considering patients' preferences for interventions and their economic affordability, the recommended plan is made more feasible, making patients more willing to accept and adhere to it. By incorporating patients' emotional state, family accompaniment, and subjective evaluations of comfort, the humanistic care for patients is demonstrated, significantly improving their treatment experience and satisfaction.

[0087] In one alternative approach, the personalized intervention includes:

[0088] Drug intervention involves recommending the type, individualized dosage, and route of administration of nonsteroidal anti-inflammatory drugs, opioids, local anesthetics, or other adjunctive analgesics based on the patient's pain level, drug metabolism characteristics, comorbidities, and allergy history, and dynamically adjusting the frequency and duration of administration.

[0089] Physical therapy, depending on the location and nature of the pain, recommends hot compresses, cold compresses, body position adjustments, specific posture guidance, or gentle massage to relieve muscle tension and promote blood circulation.

[0090] Information support and education provide patients and their families with explanations of the causes of postoperative shoulder pain, the expected course of the disease, treatment options, and self-management methods to enhance patient compliance and participation and reduce anxiety caused by uncertainty;

[0091] Nutritional support includes recommending dietary recommendations based on the patient's individual needs to help reduce inflammation and promote tissue repair.

[0092] In this embodiment, multiple approaches such as drug intervention, physical therapy, information support and education, and nutritional support are integrated to comprehensively improve patient comfort from multiple levels, including physiological, psychological, and behavioral aspects. Each intervention is customized based on the individual's unique physiological and pathological state (such as drug metabolism characteristics, comorbidities, and allergy history), psychological state (such as pain tolerance, preference for interventions, and anxiety level), and socioeconomic background (such as economic affordability).

[0093] According to the solution provided by the present invention, multimodal medical data of the patient is acquired, comprising individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... The multimodal medical data, including pressure, perfusion rate, pneumoperitoneum maintenance time, surgical duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters, are input into a pre-trained time-series-multiscale pain risk model. This model outputs the probability and risk stratification of moderate to severe shoulder pain within 24 hours post-surgery. The probability and risk stratification of shoulder pain, real-time patient feedback, and contextual information are then input into a pre-trained individualized analgesia and comfort model, which outputs the optimal combination of individualized analgesia and comfort interventions for healthcare professionals to implement personalized interventions for the patient. The intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks. This invention, based on a time-series-multiscale model to predict pain risk and a pharmacokinetic model to optimize drug dosage, significantly improves the postoperative experience and recovery quality of patients in terms of preventing and controlling shoulder pain and enhancing comfort after laparoscopic surgery.

[0094] Figure 2 A schematic diagram of the frame of the comprehensive prevention and comfort improvement device for laparoscopic postoperative shoulder pain according to an embodiment of the present invention is shown. The comprehensive prevention and comfort improvement device for laparoscopic postoperative shoulder pain includes:

[0095] Data acquisition module 210 is used to acquire multimodal medical data of the patient, including individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... Pressure, perfusion rate, pneumoperitoneum duration, operation duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters;

[0096] The pain risk assessment module 220 is used to input the multimodal medical data into a pre-trained time-series-multiscale pain risk model and output the probability and risk level of the patient experiencing moderate to severe shoulder pain within 24 hours after surgery.

[0097] The intervention plan generation module 230 is used to input the probability and risk classification of the shoulder pain, real-time patient feedback and contextual information into a pre-trained individualized analgesia and comfort model, and output the optimal combination of individualized analgesia and comfort intervention plans for medical staff to carry out personalized intervention for the patient; wherein, the intervention plan includes specific measures, dosage, frequency, expected effect and potential risks.

[0098] Figure 3 The diagram shows a structural schematic of an embodiment of the computer device of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the computer device.

[0099] like Figure 3 As shown, the computer device may include: a processor 302, a communications interface 304, a memory 306, and a communications bus 308.

[0100] The processor 302, communication interface 304, and memory 306 communicate with each other via communication bus 308. Communication interface 304 is used to communicate with other network elements, such as clients or other servers. The processor 302 executes program 310, specifically performing the relevant steps in the above-described embodiment of the comprehensive prevention and comfort improvement method for shoulder pain after laparoscopic surgery.

[0101] Specifically, program 310 may include program code that includes computer operation instructions.

[0102] Processor 302 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The computer device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0103] Memory 306 is used to store program 310. Memory 306 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0104] According to the solution provided by the present invention, multimodal medical data of the patient is acquired, comprising individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... The multimodal medical data, including pressure, perfusion rate, pneumoperitoneum maintenance time, surgical duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters, are input into a pre-trained time-series-multiscale pain risk model. This model outputs the probability and risk stratification of moderate to severe shoulder pain within 24 hours post-surgery. The probability and risk stratification of shoulder pain, real-time patient feedback, and contextual information are then input into a pre-trained individualized analgesia and comfort model, which outputs the optimal combination of individualized analgesia and comfort interventions for healthcare professionals to implement personalized interventions for the patient. The intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks. This invention, based on a time-series-multiscale model to predict pain risk and a pharmacokinetic model to optimize drug dosage, significantly improves the postoperative experience and recovery quality of patients in terms of preventing and controlling shoulder pain and enhancing comfort after laparoscopic surgery.

[0105] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination of all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed can be employed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose. Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims listing several devices, several of these devices may be embodied by the same hardware item. Unless otherwise specified, the steps in the above embodiments should not be construed as limiting the order of execution.

Claims

1. A method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery, characterized in that, include: The patient's multimodal medical data is acquired, including individual physiological characteristics, intraoperative real-time monitoring parameters, and genomic data related to pain sensitivity. The intraoperative real-time monitoring parameters include intraperitoneal CO2 pressure, perfusion rate, pneumoperitoneum maintenance time, operation duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters. The multimodal medical data is input into a pre-trained time-series-multiscale pain risk model, which outputs the probability and risk level of moderate to severe shoulder pain in patients within 24 hours after surgery. The probability and risk grading of shoulder pain, real-time patient feedback, and contextual information are input into a pre-trained individualized analgesia and comfort model, which outputs the optimal combination of individualized analgesia and comfort intervention plans for medical staff to provide personalized interventions for the patients. The intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks.

2. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 1, characterized in that, The formula for calculating the probability of a patient experiencing moderate to severe shoulder pain within 24 hours post-surgery is as follows: ； in, For the intercept term; The weight coefficient for the j-th physiological characteristic; These are standardized physiological characteristic values; The number of physiological characteristics; This is the coefficient for operation time; This refers to the duration of the surgery. Genomic risk coefficient; for Pressure cumulative exposure coefficient; Intraperitoneal space at time point t pressure; The safe pressure threshold; This refers to the duration of pneumoperitoneum maintenance. This refers to the anesthesia index coefficient. It is a comprehensive score for the type and dosage of anesthetic drugs.

3. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 1, characterized in that, The individualized analgesic intervention regimen optimizes the dosage based on a population pharmacokinetic model of flurbiprofen ester. The optimized extended formula for the dosage is: ； in, To achieve the target blood drug concentration; It is the apparent distribution volume; To eliminate the rate constant; Bioavailability; , These are the covariate effect coefficients; This is a random effect between individuals; This is the proportional error term.

4. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 1, characterized in that, The time-series-multiscale pain risk model includes: The static individual feature input layer is used to receive static medical data including individual physiological characteristics, operation duration, genomic data related to pain sensitivity, and types and dosages of anesthetic drugs. It performs preliminary feature mapping through a multilayer perceptron, a genome encoder, and an anesthesia encoder to generate high-dimensional static feature vectors. The dynamic intraoperative parameter temporal input branch layer is used to receive data including intraperitoneal parameters. Multi-dimensional time-series data, including pressure, perfusion rate, specific operational intensity, body temperature, and hemodynamic parameters, are collected, and each dimension of the time-series data is equipped with an independent time-series encoder to capture its respective time dependence; wherein, the... The pressure timing encoder incorporates an attention unit to identify and weight data exceeding a safety threshold. Pressure exposure period calculated using differentiable accumulation layers Dynamic temporal feature vector of cumulative stress exposure; A multi-scale feature fusion layer fuses the static feature vector with the dynamic temporal feature vector to generate a potential representation of pain risk; The output layer maps the potential representation of the pain risk to Logit values ​​and outputs the probability of the patient experiencing moderate to severe shoulder pain within 24 hours post-surgery, along with the corresponding risk grading.

5. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 1, characterized in that, The individualized analgesia and comfort model includes: The input layer is used to receive and encode the probability and risk grading of shoulder pain, real-time patient feedback, and contextual information. The multimodal feature fusion layer includes a cross-modal attention fusion module, which extracts information from each modality and uses cross-attention to capture deep interactions and dependencies between different modalities. The fused features are then concatenated and dimensionality reduced to generate a unified comprehensive feature representation. The intervention plan generation layer includes an intervention strategy graph neural network consisting of a measure selection head, a dose / intensity prediction head, a frequency / duration prediction head, an expected effect prediction head, a potential risk prediction head, and a sequence decision and proxy unit. This network is used to model multiple preset intervention measures as a graph structure, where the nodes of the graph structure represent intervention measures, and the edges of the graph structure represent synergistic, antagonistic, or temporal relationships between the nodes.

6. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 5, characterized in that, The intervention strategy graph neural network also includes: An intervention graph construction layer is used to model various potential analgesic and comfort interventions as nodes in the graph and initialize them as feature vectors through an embedding layer. At the same time, based on the knowledge of medical experts, the synergistic, antagonistic, temporal dependence and classification association edges between nodes are preset or dynamically learned. A multi-layer graph attention network consists of stacked graph attention units. Each graph attention unit learns attention weights to selectively aggregate information about neighboring nodes in order to capture the complex contextual dependencies between intervention measures. The comprehensive feature integration layer is used to perform attention fusion between the node features processed by the multi-layer graph attention network layer and the comprehensive feature representation output by the multimodal feature fusion layer, so as to ensure that the generation of intervention plans fully considers the intrinsic relationship between the patient's individualized multimodal data and the measures. The system includes a multi-head output decoder and a sequence decision-making and proxy unit. The multi-head output decoder outputs in parallel the selection of measures, drug dosage, intervention frequency and duration, expected effects and potential risks. The sequence decision-making and proxy unit interacts with the simulated patient model in real time and uses the patient's pain relief, comfort improvement and side effect incidence as reward signals to dynamically learn and optimize the sequence of intervention strategies at different time steps.

7. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 1, characterized in that, The contextual information includes postoperative time, current medical environment, patient position, family accompaniment, patient's emotional state, history of pain, allergies, comorbidities, medication adherence, pain tolerance, preference for interventions, and financial affordability. The real-time feedback information includes pain score, pain location, pain nature, pain frequency, nausea and vomiting, drowsiness, itching, respiratory depression, and the patient's subjective evaluation of their current comfort level.

8. The method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery according to claim 1, characterized in that, The personalized intervention includes: Drug intervention involves recommending the type, individualized dosage, and route of administration of nonsteroidal anti-inflammatory drugs, opioids, local anesthetics, or other adjunctive analgesics based on the patient's pain level, drug metabolism characteristics, comorbidities, and allergy history, and dynamically adjusting the frequency and duration of administration. Physical therapy, depending on the location and nature of the pain, recommends hot compresses, cold compresses, body position adjustments, specific posture guidance, or gentle massage to relieve muscle tension and promote blood circulation. Information support and education provide patients and their families with explanations of the causes of postoperative shoulder pain, the expected course of the disease, treatment options, and self-management methods to enhance patient compliance and participation and reduce anxiety caused by uncertainty; Nutritional support includes recommending dietary recommendations based on the patient's individual needs to help reduce inflammation and promote tissue repair.

9. A device for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery, characterized in that, include: The data acquisition module is used to acquire multimodal medical data of the patient, including individual physiological characteristics, intraoperative real-time monitoring parameters, and pain sensitivity-related genomic data; wherein, the intraoperative real-time monitoring parameters include intra-abdominal... Pressure, perfusion rate, pneumoperitoneum duration, operation duration, specific operative intensity, type and dosage of anesthetic drugs, body temperature, and hemodynamic parameters; The pain risk assessment module is used to input the multimodal medical data into a pre-trained time-series-multiscale pain risk model and output the probability and risk level of the patient experiencing moderate to severe shoulder pain within 24 hours after surgery. The intervention plan generation module is used to input the probability and risk classification of the shoulder pain, real-time patient feedback, and contextual information into a pre-trained individualized analgesia and comfort model, and output the optimal combination of individualized analgesia and comfort intervention plans for medical staff to carry out personalized interventions for the patient; wherein, the intervention plan includes specific measures, dosage, frequency, expected effects, and potential risks.

10. A computer device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the method for comprehensive prevention and control of shoulder pain and improvement of comfort after laparoscopic surgery as described in any one of claims 1-8.

Images