Intelligent intestinal preparation scheme management method and system for colonoscope patient

Abstract

The invention discloses an intelligent intestinal preparation scheme management method and system for a colonoscope patient, and relates to the technical field of data processing, and the method comprises the steps: collecting the basic information of the patient and the failure experience of intestinal preparation, extracting personalized features, and carrying out the optimization of an intestinal preparation scheme based on the personalized features, generating a plurality of groups of alternative schemes and selecting an optimal scheme to execute intestinal tract cleaning; the method comprises the following steps: receiving a feedback image, dynamically generating anomaly compensation by using a time sequence anomaly recognition model, performing optimization iteration on an optimal scheme in combination with compensation information, and establishing an optimal intestinal preparation scheme. The technical problems of poor cleaning effect, frequent adverse reaction and low patient compliance caused by lack of individuation and insufficient real-time dynamic adjustment capability of intestinal preparation in the prior art are solved, and the effects of improving the intestinal cleaning effect, reducing the adverse reaction and improving the patient compliance are realized through the design of an individualized intestinal preparation scheme. The compliance of the patient and the examination efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to an intelligent intestinal preparation plan management method and system for colonoscopy patients. Background Art

[0002] Colorectal cancer is one of the most common malignant tumors globally, with both its incidence and mortality rates remaining high. As the gold standard for colorectal cancer screening and diagnosis, colonoscopy depends on adequate bowel preparation to ensure the clarity and accuracy of the examination. However, there are still many problems with the current bowel preparation methods. Traditional bowel preparation usually requires patients to take laxatives and adjust their diet. However, due to significant individual differences among patients (such as age, weight, health status, history of constipation, etc.) and the lack of personalized guidance, the bowel cleansing effects vary widely. Statistics show that up to 25% of patients have inadequate bowel preparation, which not only significantly increases the missed diagnosis rate of adenomas, but also prolongs the examination time and even requires repeated examinations, bringing additional economic burdens and physical discomforts to patients.

[0003] In addition, most existing bowel preparation plans lack intelligent management means and mainly rely on the experience guidance of medical staff and the subjective feedback of patients. During the bowel preparation process, problems such as low patient compliance, difficulty in real-time assessment of the cleansing effect, and frequent adverse reactions also significantly reduce the efficiency and effectiveness of the examination. At the same time, the traditional health education methods for bowel preparation are fragmented and insufficiently delivered. Especially for elderly patients and patients with low educational levels, it is difficult to understand and accurately execute the requirements for taking laxatives and the dietary adjustment suggestions. Summary of the Invention

[0004] The present application provides an intelligent intestinal preparation plan management method and system for colonoscopy patients, which are used to solve the technical problems in the prior art that bowel preparation lacks personalization and the ability of real-time dynamic adjustment, resulting in poor cleansing effects, frequent adverse reactions, and low patient compliance.

[0005] In the first aspect of the present application, a method for intelligent management of intestinal preparation plans for colonoscopy patients is provided. The method includes: collecting information of patients to establish a basic information set of patients, where the basic information set includes age, weight, health status, and history of constipation; extracting features based on the basic information set to establish a first feature set; obtaining the intestinal preparation failure experience of patients and establishing a second feature set according to the intestinal preparation failure experience; optimizing the intestinal preparation plan based on the first feature set and the second feature set to establish an optimization space, where the optimization space includes multiple groups of intestinal preparation plans; using the optimal intestinal preparation plan in the optimization space to perform intestinal cleaning on patients and receiving patient feedback images with time sequence identifiers, where the patient feedback images are patient stool images; inputting the patient feedback images into a time sequence anomaly recognition model to generate time sequence anomaly compensation; and optimizing the plan in the optimization space according to the time sequence anomaly compensation and the optimal intestinal preparation plan to establish an optimization result of the optimal intestinal preparation plan.

[0006] In the second aspect of the present application, a system for intelligent management of intestinal preparation plans for colonoscopy patients is provided. The system includes: an information collection module for collecting information of patients to establish a basic information set of patients, where the basic information set includes age, weight, health status, and history of constipation; a first feature extraction module for extracting features based on the basic information set to establish a first feature set; a second feature extraction module for obtaining the intestinal preparation failure experience of patients and establishing a second feature set according to the intestinal preparation failure experience; an optimization space establishment module for optimizing the intestinal preparation plan based on the first feature set and the second feature set to establish an optimization space, where the optimization space includes multiple groups of intestinal preparation plans; an intestinal cleaning feedback module for using the optimal intestinal preparation plan in the optimization space to perform intestinal cleaning on patients and receiving patient feedback images with time sequence identifiers, where the patient feedback images are patient stool images; a time sequence anomaly compensation module for inputting the patient feedback images into a time sequence anomaly recognition model to generate time sequence anomaly compensation; and a dynamic plan optimization module for optimizing the plan in the optimization space according to the time sequence anomaly compensation and the optimal intestinal preparation plan to establish an optimization result of the optimal intestinal preparation plan.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0008] An intelligent intestinal preparation plan management method and system for colonoscopy patients provided by the present application relate to the technical field of data processing. By collecting patients' basic information and intestinal preparation failure experiences, extracting key features to construct an optimization space for the plan, selecting the optimal plan to perform intestinal cleansing, receiving feedback images in real time and inputting them into a time series anomaly recognition model, dynamically generating anomaly compensation, and optimizing the intestinal preparation plan, it solves the technical problems in the prior art that intestinal preparation lacks personalization and the ability of real-time dynamic adjustment, resulting in poor cleansing effects, frequent adverse reactions, and low patient compliance. It realizes the technical effects of designing a personalized intestinal preparation plan through intelligent and dynamic optimization means, improving the intestinal cleansing effect, reducing adverse reactions, and enhancing patient compliance and examination efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0010] Figure 1 It is a schematic flowchart of an intelligent intestinal preparation plan management method for colonoscopy patients provided by an embodiment of the present application;

[0011] Figure 2 It is a schematic structural diagram of an intelligent intestinal preparation plan management system for colonoscopy patients provided by an embodiment of the present application.

[0012] Description of reference numerals: Information collection module 11, First feature extraction module 12, Second feature extraction module 13, Optimization space establishment module 14, Intestinal cleansing feedback module 15, Time series anomaly compensation module 16, Dynamic plan optimization module 17. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] The present application provides an intelligent intestinal preparation plan management method and system for colonoscopy patients, which are used to solve the technical problems in the prior art that intestinal preparation lacks personalization and the ability of real-time dynamic adjustment, resulting in poor cleansing effects, frequent adverse reactions, and low patient compliance.

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0015] It should be noted that the terms "first", "second", etc. in the description of this application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0016] Embodiment 1, as Figure 1 shown, the present application provides an intelligent bowel preparation plan management method for colonoscopy patients, and the method includes:

[0017] P10: Collect information about the patient and establish a basic information set of the patient, where the basic information set includes age, weight, health status, and history of constipation.

[0018] First of all, in the intelligent bowel preparation plan management method of the present application, the first step is to comprehensively collect information about the patient and establish a basic information set, providing key data support for subsequent feature extraction and personalized plan optimization. The content of information collection includes age, weight, health status, and history of constipation, etc. These data directly affect the effect of bowel preparation and the accuracy of plan design.

[0019] Specifically, age is an important variable in the process of bowel preparation. There are significant differences in bowel function among patients of different age groups. For example, elderly patients may require a stronger laxative dose or a longer preparation time due to slower bowel motility or decreased metabolic capacity, while young patients have faster bowel motility and may be more likely to achieve a clean effect. By collecting the patient's age data and marking it as a feature, it can provide an important basis for subsequent plan adjustment.

[0020] Weight directly affects the determination of the laxative dose. Patients with a higher weight usually require a larger laxative dose to ensure bowel cleansing, while patients with a lower weight need to reduce the dose appropriately to avoid adverse reactions caused by drug overdose. By collecting weight data and combining it with a database of the laxative dose-weight relationship, more accurate dose recommendations can be provided for patients.

[0021] Health status is another key parameter, covering whether the patient has chronic diseases (such as diabetes, hypertension, heart disease, kidney disease, etc.). These diseases may affect the patient's tolerance to laxatives and the safety of bowel cleansing. For example, patients with diabetes may need to avoid certain types of laxatives to prevent blood sugar fluctuations, while patients with impaired renal function need to use electrolyte-containing drugs with caution. Therefore, collecting health status information and labeling it as a feature can significantly improve the safety of bowel preparation programs.

[0022] History of constipation is also one of the important factors affecting the effect of bowel preparation. Patients with long-term constipation usually have weaker intestinal motility and greater difficulty in bowel cleansing. Therefore, special adjustments need to be made in the type, dose, and administration time of laxatives. By collecting the data of the patient's history of constipation, the potential difficulties of bowel preparation can be predicted, and a more personalized preparation plan suitable for the patient can be designed in advance.

[0023] Moreover, the information collection is carried out by combining questionnaire surveys, electronic medical record reading, and patient interviews to ensure the comprehensiveness and accuracy of the data. After the collected basic information set is standardized, it is input into the intelligent system for subsequent feature extraction and optimization analysis. This process lays the foundation for the management of the entire bowel preparation program, ensuring that the subsequent steps can be dynamically adjusted around the personalized needs of the patient, improving the bowel cleansing effect and the patient experience.

[0024] P20: Feature extraction is performed based on the said basic information set to establish a first feature set.

[0025] Optionally, feature extraction is performed on the patient's basic information set to generate a first feature set that can accurately reflect the individual differences and bowel preparation needs of the patient. The core of this step is to extract, through data analysis techniques, the features that have an important impact on the bowel cleansing effect, providing an accurate basis for personalized plan optimization.

[0026] First of all, age is extracted as one of the important features. Age directly affects the intestinal motility function and drug metabolism ability of the patient. For example, the intestinal motility speed of elderly patients is usually slower, the onset time of laxatives is longer, and higher doses or longer preparation times may be required; while the intestinal function of young patients is relatively healthy, and they respond faster to laxatives. Therefore, during the feature extraction process, age data is standardized and grouped into multiple age groups (such as 18 - 40 years old, 40 - 60 years old, over 60 years old) to reflect the intestinal characteristics of different age groups.

[0027] Body weight is another important feature that directly affects the dosage selection of laxatives. The higher the body weight of a patient, the larger the dosage of laxatives required for bowel cleansing. However, too high a dosage may increase the risk of side effects. Through feature extraction, the system maps the patient's body weight data to a dosage-related coefficient and combines it with a scientific model of laxative dosage-body weight to help calculate the appropriate dosage range for each patient. This precise dosage feature extraction avoids insufficient cleansing effects or increased side effects caused by dosage mismatch.

[0028] The feature extraction of health status mainly focuses on the potential impact of chronic diseases on bowel preparation. For example, patients with diabetes need to avoid using laxatives that may cause blood sugar fluctuations, while patients with impaired renal function need to limit the use of drugs containing electrolytes. During the feature extraction process, the system generates specific restrictions on laxative selection by analyzing the medical history records of the patient's health status and marks them as features in the first feature set. This can ensure that the drug selection meets the bowel cleansing requirements while minimizing the patient's medication risk.

[0029] The history of constipation feature reflects the strength of the patient's intestinal peristalsis function and is an important factor affecting the efficiency of bowel cleansing. Patients with a long history of constipation may require more potent laxatives, a longer preparation time, or a combination of auxiliary dietary adjustments. Through the analysis of constipation history data, the system extracts the severity of constipation and generates relevant feature marks to provide guidance for the subsequent design of bowel preparation plans. For example, the system may recommend a segmented medication method for patients with "severe constipation" to ensure that the drug takes effect gradually and improves the cleansing effect.

[0030] Exemplarily, based on data mining and machine learning algorithms (such as decision tree or random forest models), variables in the basic information set that have the greatest impact on the bowel preparation effect can be identified, their weights can be quantified, and finally a personalized first feature set can be generated. This feature set not only includes the patient's physiological characteristics (such as age, body weight), but also covers the patient's health constraints (such as chronic disease restrictions) and historical behavior characteristics (such as the impact of constipation history), precisely depicting the personalized needs of the patient in bowel preparation, ensuring that the subsequent plan design can maximize the patient's cleansing goal while improving safety and compliance.

[0031] P30: Obtain the patient's experience of failed bowel preparation and establish a second feature set based on the experience of failed bowel preparation.

[0032] Furthermore, step P30 of the embodiment of the present application further includes:

[0033] P31: Establish a standardized questionnaire, send the standardized questionnaire to the patient, and receive the patient's feedback questionnaire; P32: Read the electronic medical records of the hospital, identify the objective data of the patient's failed bowel preparation through the electronic medical records, and establish an objective data set; P33: Use the patient's feedback questionnaire and the objective data set as the bowel preparation failure experience, perform feature extraction through a joint authentication network, and establish a second feature set. The feature extraction includes compliance feature extraction, cleaning effect feature extraction, and adverse reaction feature extraction, and the feature extraction is the joint authentication feature extraction of subjective data and objective data.

[0034] It should be understood that in order to further optimize the personalized plan, it is necessary to extract key information related to failed bowel preparation from the patient's subjective feedback and objective data to form a second feature set. First, by designing a standardized questionnaire, collect the patient's subjective responses during bowel preparation, including the time and dose of drug administration, whether the diet behavior conforms to the doctor's advice, adverse reactions (such as abdominal distension, nausea, vomiting, etc.), and the feeling of bowel cleaning effect. These subjective feedbacks can not only reflect the patient's compliance, but also reveal potential problems during the bowel preparation process. At the same time, the patient can provide more intuitive reference information for the evaluation of bowel cleaning by uploading specific data such as stool images. All questionnaire data will be collected and stored through an online system to provide a subjective data basis for subsequent feature extraction.

[0035] Next, obtain the objective data related to the patient's bowel preparation from the hospital's electronic medical record system. These data include bowel cleaning scores (such as BBPS score), adverse reactions recorded in previous bowel preparations (such as severe abdominal distension or hypotension), records of the time and dose of drug use, and historical records of examination failure or delay due to insufficient bowel cleaning. These objective data are directly derived from the medical system, can supplement the possible subjective biases in the patient's feedback, and provide a more accurate reference basis. Through the electronic medical record system, establish a standardized objective data set to ensure the accuracy and consistency of the data and lay a foundation for the feature extraction process.

[0036] Finally, the subjective feedback questionnaire data of the patient and the objective data in the electronic medical record are input into the joint authentication network, and multi-dimensional feature extraction is carried out using data fusion technology. This process includes three key aspects: First, compliance features are extracted. By analyzing the description of the laxative intake in the subjective feedback and comparing it with the actual medication time and dose in the medical record, specific features of the patient's compliance level are refined. Second, cleaning effect features are extracted. The subjective feedback images and descriptions uploaded by the patient are combined with the endoscopic cleaning score in the medical system, and machine learning algorithms are used to quantitatively evaluate the actual effect of bowel cleaning. Third, adverse reaction features are extracted. The patient's feelings (such as vomiting or abdominal distension) in the subjective feedback are combined with the incidence of adverse reactions recorded in the medical record to analyze the patient's tolerance during the bowel preparation process. Through the joint authentication technology, subjective and objective data can be mutually verified, errors from a single data source can be eliminated, and high-quality features closely related to the failure of bowel preparation can be extracted.

[0037] Finally, these extracted features are organized into a complete second feature set. This feature set comprehensively combines dimensions such as the patient's compliance, cleaning effect, and adverse reactions, and comprehensively reflects the key factors for the failure of bowel preparation. This not only provides accurate data support for the subsequent optimization of the bowel preparation plan but also lays a solid foundation for realizing intelligent personalized management.

[0038] P40: Optimize the bowel preparation plan based on the first feature set and the second feature set, and establish an optimization space, where the optimization space includes multiple groups of bowel preparation plans.

[0039] Furthermore, step P40 of the embodiment of the present application further includes:

[0040] P41: Analyze the first feature set and the second feature set to establish key features; P42: Use the key features to construct a set of bowel preparation plans; P43: Perform multi-objective optimization on the set of bowel preparation plans to establish the optimization space, where the optimization fitness features for multi-objective optimization include cleaning effect features, adverse reaction features, and compliance features.

[0041] Optionally, through the joint analysis of the first feature set and the second feature set, the bowel preparation plan is optimized and finally an optimization space is established. The optimization space consists of multiple groups of relatively optimal bowel preparation plans, which are the results with higher fitness selected from multiple alternative plans and can achieve the best balance among cleaning effect, adverse reactions, and patient compliance.

[0042] First, comprehensively analyze the patient's first feature set (such as age, weight, health status, history of constipation) and second feature set (such as compliance, adverse reactions, and cleaning effect in the experience of failed bowel preparation, etc.) to extract the key features that affect the bowel preparation effect. The key features refer to the variables that have a significant impact on the cleaning effect, adverse reactions, and the patient's execution ability. For example, the feature of constipation history affects the selection of the intensity of laxative dose; the compliance feature affects the complexity of the protocol design; the adverse reaction feature directly restricts the type and dose range of laxatives. These features can be screened through machine learning algorithms (such as feature importance ranking or recursive feature elimination) to generate high-weight variables for subsequent protocol design.

[0043] Next, based on the extracted key features, generate an initial set of bowel preparation protocols as the basis for optimization. The set of bowel preparation protocols consists of multiple groups of randomly generated alternative protocols, and each group of protocols includes different types of laxatives, doses, medication times, and dietary adjustment suggestions. For example, for patients with severe constipation, the protocol may preferentially select high-dose laxatives and combine with dietary guidance; for patients with low compliance, the protocol will tend to design simpler medication steps and shorter medication durations. This protocol generation process can be achieved through rule setting and random sampling techniques (such as the Latin hypercube sampling method) to ensure that the initial protocol set covers a variety of possibilities and provides an adequate solution space for subsequent optimization operations.

[0044] After the set of protocols is constructed, perform multi-objective optimization on these alternative protocols to screen out the protocols with higher fitness and form an optimization space. The core of multi-objective optimization lies in balancing the following three optimization objectives: the cleaning effect feature, which is used to ensure that the bowel cleanliness meets the expected standard, for example, the BBPS score reaches more than 3 points, and the adenoma missed diagnosis rate is minimized. The adverse reaction feature is used to reduce the side effects (such as nausea, vomiting, or abdominal distension) that patients experience during the preparation process and improve the safety and tolerance of laxative use. The compliance feature is used to optimize the execution difficulty of the protocol, making it easier for patients to understand and strictly follow the guidance, and avoiding insufficient compliance caused by high complexity.

[0045] Among them, genetic algorithms or particle swarm optimization algorithms can be used for multi-objective optimization. These intelligent optimization algorithms can quickly screen out the optimal solutions that meet multiple objectives in a vast protocol space. Specifically, the algorithm can perform iterative calculations based on the fitness of each protocol (the score that comprehensively evaluates the three features of cleaning effect, adverse reactions, and compliance), gradually eliminate the protocols with lower fitness, and retain and optimize the protocols with higher fitness. For example, if a certain protocol has excellent cleaning effect but strong adverse reactions, through iterative adjustment of the dose and medication time, the optimized protocol may achieve a better balance between the cleaning effect and adverse reactions.

[0046] Finally, the number of solutions in the optimization space is significantly reduced compared to the initial set of bowel preparation solutions, and only includes multiple groups of bowel preparation solutions with higher fitness. These solutions have undergone multi-dimensional optimization and screening, and can provide scientific and accurate bowel preparation guidance for the personalized needs of patients. For example, for patients with poor health conditions, the optimization space will preferentially recommend solutions with a better balance between cleaning effect and safety; while for patients with stronger execution ability, the system may provide solutions that focus more on the cleaning effect.

[0047] By analyzing the first feature set and the second feature set, extracting key features to construct the initial solution set, and using multi-objective optimization technology to establish the optimization space, the transformation from random alternative solutions to precise optimized solutions is realized. The solutions in the optimization space have been optimized and screened, can better adapt to the personalized needs of patients, significantly improve the success rate of bowel cleaning, reduce the risk of adverse reactions, and enhance the compliance and experience of patients.

[0048] Furthermore, step P43 of the embodiment of the present application further includes:

[0049] P43-1: Establish the compensation weight of the specific group; P43-2: Conduct the specific group authentication for the patient and establish the authentication coefficient; P43-3: After correcting the compensation weight through the authentication coefficient, execute the multi-objective optimization constraint, and establish the optimization space according to the constraint result.

[0050] In a possible embodiment of the present application, the process of multi-objective optimization is further refined, and the compensation weight and authentication mechanism of the specific group are added, so that the optimization result is more precise and personalized, and finally a complete optimization space is constructed.

[0051] First, for the specific groups in the bowel preparation solutions, such as elderly patients, patients with chronic diseases, patients with long-term constipation, etc., through data analysis and modeling, establish the compensation weight of the specific group, that is, adjust the weight of the optimization objective according to the individual differences of the specific group to ensure the adaptability of the optimization result to this group. For example, for elderly patients, due to the slowdown of intestinal peristalsis, it may be necessary to increase the weight of the cleaning effect and reduce the sensitivity to adverse reactions; for patients with chronic diseases, it is necessary to reduce the risk of side effects caused by certain laxatives, so as to increase the binding force of the adverse reaction characteristics in the optimization process. The compensation weight can be obtained through historical cases and big data analysis, and is dynamically updated to adapt to the changes in group characteristics.

[0052] Next, perform specialized group authentication on the patient. By analyzing the patient's first feature set (such as age, health status) and second feature set (such as compliance, adverse reactions), the patient is classified into the most suitable specialized group, and an authentication coefficient is generated for the patient. The authentication coefficient is a numerical value calculated based on the matching degree between the patient's features and the specialized group, reflecting the proximity between the patient's individual features and the group features. For example, a patient with a long history of constipation and an older age may have an authentication coefficient more inclined towards the "elderly constipation group", while a patient with diabetes may be classified into the "chronic disease group". The authentication coefficient is generated through a clustering algorithm (such as K-Means clustering) or a deep learning model (such as Siamese network) to ensure the accuracy of authentication.

[0053] After obtaining the authentication coefficient, the system uses this coefficient to correct the compensation weight of the specialized group, thereby dynamically adjusting the optimization constraints in multi-objective optimization. The corrected compensation weight will be more in line with the patient's personalized needs. For example, if the patient's authentication coefficient shows that they highly conform to the "elderly constipation group", the weight of the cleaning effect feature will be further increased to ensure the best intestinal cleaning state, while appropriately reducing the sensitivity to laxative dosage to reduce adverse reactions.

[0054] Based on the corrected weight, multi-objective constraints are introduced in the optimization process. The core of multi-objective constraints is to balance the three optimization objectives of cleaning effect, adverse reactions, and compliance. Through an optimization algorithm (such as multi-objective genetic algorithm or particle swarm optimization algorithm), the optimal solution is searched in the solution space that meets the constraint conditions. For example, for patients prone to side effects, the constraint conditions for adverse reaction features will be significantly increased to ensure safety during laxative use; for patients with low compliance, the optimization will tend to select a more easily executable plan.

[0055] Finally, an optimization space is established based on multi-objective optimization. The optimization space is a multi-dimensional solution set, containing multiple groups of optimized intestinal preparation plans. Each group of plans is based on the balance of the three objectives of cleaning effect, adverse reactions, and compliance, and is generated in combination with the patient's specialized authentication results, providing diverse choices for the dynamic adjustment of the subsequent execution plan, and at the same time providing a scientific basis for customized plans for special patient groups (such as chronic disease patients or the elderly).

[0056] P50: Use the optimal intestinal preparation plan in the optimization space to perform intestinal cleaning on the patient, and receive the patient feedback image with a time sequence identifier, where the patient feedback image is the patient's stool image.

[0057] Furthermore, step P50 of the embodiment of the present application further includes:

[0058] P51: Conduct data supervision on the patient and establish an education reception data set; P52: Fit the education with the education reception data set and the optimal bowel preparation plan to establish an optimized education plan; P53: Use the optimized education plan to manage the patient's education.

[0059] It should be understood that the optimal bowel preparation plan with the maximum fitness is selected from multiple groups of bowel preparation plans in the optimization space to guide the patient to complete bowel cleansing, and the execution effect and optimization potential of the plan are evaluated in real time through the patient feedback data. Specifically, the system will guide the patient to perform the cleansing process such as taking laxatives and adjusting the diet according to the plan based on the patient's personalized characteristics and the optimal bowel preparation plan recommended by the optimization space, and at the same time collect the feedback data with time sequence marks (such as the patient's stool images). The feedback images are analyzed by intelligent recognition technologies (such as computer vision and image segmentation technologies) to analyze their cleansing effects, so as to provide real-time data support for plan optimization.

[0060] In this process, the time sequence mark is used to record the time point of the feedback data and the patient's current bowel preparation progress. For example, the stool cleansing situation 1 hour, 3 hours or 6 hours after taking the laxative. This time-sequenced data helps to analyze the dynamic changes of bowel cleansing and predict possible preparation problems (such as insufficient cleansing or over-cleaning). For example, the color, shape and residue ratio of the stool can be identified through the feedback images and compared with the bowel cleansing standard (such as the BBPS score) to monitor the cleansing effect in real time.

[0061] To ensure that the patient can correctly understand and execute the optimal bowel preparation plan, the system first conducts data supervision on the patient's education process and establishes an education reception data set. This data set evaluates the patient's understanding of the plan by recording the patient's learning behaviors during the education (such as whether they read the education materials, completed watching the education videos, participated in the Q&A interactions, etc.) and the feedback on the education content (such as questionnaire or test results). The purpose of data supervision is to identify possible knowledge blind spots or understanding deviations of the patient before executing the plan. For example, the confusion of elderly patients about the medication time and dose, or the misunderstanding of the diet adjustment requirements by patients with low educational levels.

[0062] Based on the missionary reception dataset, further perform fitting analysis on the learning ability and comprehension of patients, and design an optimized missionary plan in combination with the key execution nodes in the optimal bowel preparation plan (such as the time to take laxatives and dietary taboos). The optimized missionary plan fits the reception ability and needs of patients through machine learning models (such as decision trees or deep learning networks) to ensure that the missionary content is more precise and personalized. For example, for patients with poor comprehension ability, the system will use intuitive forms such as diagrams and video demonstrations instead of pure text content and simplify the step descriptions; for young patients, the dynamic monitoring and data feedback of the cleaning effect may be emphasized. The optimized missionary plan will also dynamically adjust the content according to the patient's feedback history, such as adding diet adjustment cases or reducing unnecessary technical terms.

[0063] Finally, use the optimized missionary plan for the comprehensive missionary management of patients. Specifically, the system transmits the optimized plan to patients through an online platform or a mobile application and real-time tracks the learning progress and implementation status of patients. For example, the system can regularly push reminders to take laxatives, dietary taboo precautions, and guidelines for uploading feedback images. Through the optimized missionary plan, the system can improve patients' understanding and compliance with the bowel preparation process, thereby reducing bowel preparation failures caused by information misunderstandings or execution errors. At the same time, the system continues to collect feedback data after each patient executes the plan, performs matching analysis with the optimized missionary plan, further improves the missionary content, and provides more efficient guidance for future patients.

[0064] Through the above steps, the dynamic execution and management of the bowel preparation plan are realized, and through optimized missionary work and real-time feedback of patient data, the quality of bowel cleaning and the overall experience of patients are greatly improved, providing technical support and practical basis for the continuous optimization of bowel preparation effects.

[0065] P60: Input the patient feedback image into the time series anomaly recognition model to generate a time series anomaly compensation.

[0066] Optionally, analyze the patient feedback image through the time series anomaly recognition model, dynamically evaluate the bowel cleaning effect, and generate a time series anomaly compensation for possible cleaning anomalies, thereby realizing the real-time optimization and adjustment of the bowel preparation plan.

[0067] First, the feedback images uploaded by the patient (such as stool images) are input into the temporal anomaly recognition model with temporal identifiers. The temporal identifiers record the time nodes when the feedback images are collected. For example, the excretion of stool 1 hour, 3 hours, or 6 hours after taking laxatives. This data with temporal identifiers provides a panoramic view of the dynamic process of bowel cleansing, enabling the model to analyze the trend of cleansing effect according to time changes. For example, the model will judge whether the patient's bowel has reached the expected cleansing standard through image features (such as color, particle distribution, residue ratio, etc.), and at the same time analyze whether there are abnormalities in the cleansing process in combination with the time node, such as insufficient cleansing or excessive cleansing.

[0068] Among them, the temporal anomaly recognition model can be constructed based on deep learning technology, using a convolutional neural network (CNN) to extract image features, and combining time series analysis algorithms (such as long short-term memory network LSTM or temporal convolutional network TCN) to evaluate the dynamic changes in the cleansing process. Specifically, the model first preprocesses the feedback images, including noise removal, image segmentation, and feature annotation. For example, the segmentation algorithm can accurately extract the key regions in the image (such as the morphology of excrement), providing clear input data for subsequent analysis. Then, the model uses a convolutional neural network to extract the feature patterns of the image. For example, the depth of color represents the degree of cleansing, and the size and distribution of particles reflect the removal of intestinal residues. Time series analysis is used to identify the time points of abnormal changes, such as abnormal excretion caused by too slow progress of bowel cleansing or too fast drug effect.

[0069] During the model analysis process, if an abnormality in the cleansing process is detected (such as a lag in cleansing effect or excessive cleansing caused by side effects), a temporal anomaly compensation is generated according to the abnormality. The temporal anomaly compensation is an optimization measure to dynamically adjust the bowel preparation plan according to the type and severity of the abnormality. For example:

[0070] Insufficient cleansing compensation. If the model detects that the cleansing effect of the patient 3 hours after taking laxatives does not meet the standard, it may indicate insufficient drug dosage or improper diet. The system will recommend that the patient increase the laxative dosage, extend the preparation time, or adjust the diet to improve the cleansing efficiency.

[0071] Excessive cleansing compensation. If the model recognizes that the patient has too fast a cleansing progress within 1 hour after taking laxatives, which may lead to excessive bowel cleansing and increase the risk of side effects. At this time, the system will recommend that the patient reduce the laxative dosage or delay taking the subsequent drugs.

[0072] Adverse reaction compensation. If the feedback image combined with the patient's description shows abnormal side effects (such as collapse caused by excessive diarrhea), the system will recommend that the patient suspend the use of laxatives and supplement electrolyte water according to the severity of the abnormality, and at the same time feedback the situation to the medical staff.

[0073] After generating the timing anomaly compensation, the system will notify the patient in real time and provide operation guidance. For example, send compensation suggestions through a mobile application and provide specific adjustment instructions, such as suggestions for laxative dosage, modification of taking intervals, or diet adjustment guidance. At the same time, these compensation suggestions will be recorded in the patient's personal file to provide reference for the subsequent optimization of the bowel preparation plan.

[0074] Through this process, dynamic monitoring and real-time intervention during the patient's bowel preparation are achieved, ensuring that the cleaning effect is always in the best state and avoiding cleaning failure or patient discomfort caused by abnormal situations. Through the timing anomaly recognition model, strong technical support is provided for the refined management of bowel cleaning, improving the quality of bowel preparation and patient satisfaction.

[0075] P70: Optimize the plan within the optimization space according to the timing anomaly compensation and the optimal bowel preparation plan to establish the optimization result of the optimal bowel preparation plan.

[0076] Specifically, use the timing anomaly compensation information to dynamically optimize the optimal bowel preparation plan in the optimization space to generate the final optimized plan that adapts to the individual needs of the patient. By combining the timing feedback data with the initial optimal plan, further adjustment of the bowel preparation plan is achieved to maximize the cleaning effect, while reducing adverse reactions and improving compliance.

[0077] First, the system receives the timing anomaly compensation data, which is generated by the patient feedback image and the timing anomaly recognition model, marking the abnormal conditions and recommended compensation measures that occur during the bowel preparation process. For example, if the feedback image shows that the cleaning effect is insufficient 3 hours after the patient takes the laxative, the system will increase the dose or adjust the taking time through the abnormal compensation suggestion; if excessive cleaning is identified, the compensation measure may suggest reducing the laxative dose or extending the taking interval. These compensation information are the core of dynamic adjustment, reflecting the individualized cleaning progress and adaptation status of the patient.

[0078] After obtaining the abnormal compensation, the system starts the dynamic optimization process based on the optimal bowel preparation plan and combines other alternative plans in the optimization space. The key to dynamic optimization is to adjust the parameters of the current plan, such as laxative type, dose, taking interval, and diet guidance, and find a better solution in the optimization space through global optimization technology. Global optimization is a technology based on multi-objective optimization that can simultaneously balance the following three optimization goals among all possible plans: cleaning effect, ensuring that the bowel cleaning level meets the examination requirements, such as evaluating the cleaning adequacy through the BBPS score. Adverse reactions, minimizing the side effects of the patient during the bowel preparation process (such as abdominal distension, nausea). Compliance, optimizing the operation complexity of the plan to make it easier for the patient to execute, so as to avoid failure caused by insufficient compliance.

[0079] During the optimization process, the effectiveness of the adjusted solution can be evaluated based on a multi-objective fitness function. For example, if a solution scores high in terms of cleaning effectiveness but causes strong adverse reactions, the system will make adjustments by reducing the dose of laxatives or extending the interval; if a solution has poor compliance (such as high medication frequency making it difficult for patients to execute), the number of medication administrations will be reduced or the steps will be simplified. Optimization algorithms (such as genetic algorithms or particle swarm optimization algorithms) gradually improve the solution parameters in multiple iterations, continuously screening and enhancing the solution fitness.

[0080] Finally, after multiple optimization iterations, the optimized results of the optimal bowel preparation solution generated by the system can fully integrate the patient's real-time feedback and personalized needs. For example, for patients with insufficient cleaning effectiveness, the optimized solution may include a higher dose of laxatives combined with dietary assistance; for patients with excessive cleaning, it may be adjusted to a phased medication plan to reduce unnecessary side effects. This optimized result maintains the balance of patient compliance and safety while dynamically adjusting.

[0081] Through the above steps, the evolutionary process of the bowel preparation solution from static recommendation to dynamic adjustment is realized, making the final optimized solution more in line with the actual needs of patients, improving the success rate of bowel cleaning, reducing the risk of examination failure, optimizing the patient experience at the same time, and providing strong guarantee for the efficient and safe implementation of colonoscopy.

[0082] Furthermore, the embodiment of the present application further includes step P80, and step P80 further includes:

[0083] P81: Use the client to read the real-time acquisition information of the colonoscope device, where the real-time acquisition information includes the acquired images and operation time, and upload the real-time acquisition information to the server;

[0084] P82: Use the convolutional neural network of the server to perform image recognition on the real-time acquisition information and establish an image feedback score; P83: Perform nursing management for the patient according to the image feedback score.

[0085] In a possible embodiment of the present application, the bowel cleaning effect of the patient can be dynamically evaluated by real-time collecting the examination information of the colonoscope device, and the nursing management process can be optimized in combination with the scoring results.

[0086] Specifically, first use the client to read the real-time acquisition information of the colonoscope device, including key data such as the acquired intestinal images and operation time. These real-time information are uploaded to the server through the client, and the operation time is used to mark the specific time point of image acquisition, forming a data chain with time sequence identification, providing an accurate time dimension for subsequent cleaning effect analysis. The real-time acquired information not only reflects the key nodes of colonoscopy operation, but also records the cleaning status of different regions of the intestine, laying a data foundation for the dynamic tracking of the examination.

[0087] Furthermore, the server uses a convolutional neural network (CNN) to perform intelligent analysis and scoring on the real-time collected image information. Exemplarily, through image preprocessing and feature extraction, key cleaning features in the intestinal tract image are identified, such as the distribution of residues, the degree of mucus coverage, and abnormal signs such as bleeding. The CNN model combines a standardized intestinal tract cleaning scoring system (such as the BBPS score) to generate a quantitative cleaning score for each image and forms a time-series report of the cleaning effect based on the operation time. This scoring process realizes the accurate quantification of the cleaning status through a deep learning model, can quickly detect uncleaned or over-cleaned areas in the intestinal tract, and provides instant feedback.

[0088] Next, based on the image feedback score, the patient's care management is further optimized, and the score result is pushed to the nursing staff in real time for guiding personalized nursing decisions. For example, when the score shows that some intestinal areas are insufficiently cleaned, the system will recommend taking supplementary cleaning measures or extending the examination time to improve the examination accuracy; if abnormal conditions (such as severe residues or mucosal bleeding) are found in the score, a nursing reminder will be issued, recommending that medical staff conduct additional interventions on the patient. In addition, the score result is also used for the formulation of postoperative care. For example, an enhanced preparation plan is designed for patients with poor cleaning effects and guides them to avoid similar problems in subsequent examinations. Through this closed-loop process, the real-time monitoring, intelligent analysis, and dynamic care management of the intestinal tract cleaning effect are realized, providing data support for optimizing the efficiency and quality of colonoscopy.

[0089] Furthermore, the embodiment of the present application further includes step P90, and step P90 further includes:

[0090] P91: When the colonoscopy device performs data collection, activate the linkage acquisition sensor to perform synchronous linkage information collection of the patient, and establish a linkage data set, where the linkage data set includes vital sign data and pain score data; P92: Perform early warning identification according to the vital sign data and the pain score data, establish an early warning signal, and report the early warning through the early warning signal.

[0091] Optionally, through the linkage collection of the patient's vital sign and pain score data, the real-time early warning identification and reporting of potential risks are realized to ensure the safety of the patient during the examination.

[0092] Specifically, while the colonoscopy device performs data acquisition, the system synchronously acquires the patient's vital signs and pain score data by activating the linkage acquisition sensor, and establishes a linkage dataset. The linkage acquisition sensor works in coordination with the colonoscopy device to capture the patient's vital signs data and pain score data in real time. The vital signs data includes the patient's heart rate, blood pressure, respiratory rate, blood oxygen saturation, etc. These parameters can reflect the patient's physiological state during the examination. For example, an abnormal increase in heart rate may indicate that the patient is experiencing pain or stress, while a decrease in blood oxygen saturation may suggest that the patient is at risk of hypoxia. The pain score data is collected through the patient's self-report or by using facial expression recognition technology to quantify the patient's subjective pain perception. These data are integrated into the linkage dataset and synchronously associated with the examination data collected by the colonoscopy device, thus providing multi-modal data support with consistent timing for subsequent early warning identification.

[0093] Next, using the vital signs data and pain score data in the linkage dataset, the real-time state of the patient is analyzed through the early warning identification algorithm. The early warning identification is based on multi-dimensional data fusion technology. The patient's physiological parameters and pain score are input into an intelligent algorithm model (such as time series analysis models like random forest, LSTM, etc.) for risk assessment. For example, analyze the dynamic change trend of the vital signs, identify abnormal patterns (such as continuous increase in heart rate or decrease in blood pressure), and combine with the fluctuation of the pain score to determine whether the patient is in a high-risk state. Furthermore, based on the early warning signals generated from these risk assessment results, the signals are graded according to the risk level (such as low risk, medium risk, high risk), and the signals are reported in real time through the early warning module.

[0094] The early warning reporting mechanism aims to remind medical staff to pay attention to the patient's abnormal state in the first time. For example, when the system detects that the patient's vital signs are abnormal (such as a sudden increase in heart rate) and the pain score increases significantly, a high-risk early warning signal will be immediately triggered. This signal notifies the medical staff to take intervention measures, such as pausing the examination, adjusting the operation intensity, or giving appropriate analgesic treatment, through the display screen, sound, or mobile device. The timeliness of early warning reporting can not only prevent the further development of potential risks, but also improve the safety and patient experience of the examination process.

[0095] Through the above steps, the real-time monitoring and dynamic assessment of the patient's physiological and subjective states during the colonoscopy examination are realized. Based on linkage data acquisition and intelligent early warning, it not only enhances the medical staff's real-time control ability of the patient's state, but also can greatly improve the safety and accuracy of the examination.

[0096] Furthermore, the embodiment of the present application further includes a bedside intelligent interaction system, which can provide detailed bowel preparation guidance to the patient through various methods such as touch screen and voice prompts, and record the patient's feedback and bowel preparation situation in real time.

[0097] Specifically, this bedside intelligent interaction system runs through the entire process of the patient's bowel preparation. As the key interaction interface between the patient and the system, it provides personalized and intuitive operation guidance through the touch screen and voice prompts, ensuring that the patient can accurately understand and execute the requirements of the plan during the preparation process. First, the system combines the patient's basic information (such as age, weight, health status, history of constipation, etc.) and the optimized optimal bowel preparation plan to generate a detailed preparation plan, and gradually guides the patient to complete key steps such as taking medicine, drinking water, and diet adjustment through graphic display, voice explanation, etc.

[0098] During the execution of bowel preparation, the bedside intelligent interaction system collects the patient's feedback information in real time, including the time of taking medicine, drug reactions, changes in bowel movements, etc. The patient can manually input data through the touch screen or directly report relevant situations through voice interaction, such as "took the first dose of laxative" "feeling slightly bloated", etc. These feedback data are recorded in real time with time sequence identification and transmitted to the server system for dynamic analysis and optimization.

[0099] In addition, the bedside intelligent interaction system also integrates warning and reminder functions. For example, when the patient fails to complete a certain step on time (such as delaying taking the laxative or insufficient water intake), the system will remind the patient to complete the operation in time through voice prompts or screen pop-ups; when it is recognized that the patient may have adverse reactions (such as feedback of severe abdominal distension), the system will recommend that the patient suspend certain operations and contact the medical staff. Through this closed-loop management, the intelligent interaction function effectively improves the patient's compliance and reduces the risk of bowel cleansing failure caused by operation deviation or adverse reactions.

[0100] The bedside intelligent interaction system is not only an auxiliary tool for the patient's operation, but also an important node for data collection and transmission. The patient feedback data recorded in real time can be seamlessly integrated with the time sequence anomaly recognition model, providing accurate data support for dynamically optimizing the bowel preparation plan, so as to achieve intelligent management of the whole process and significant improvement of the patient experience.

[0101] In summary, the embodiments of the present application have at least the following technical effects:

[0102] This application collects the patient's basic information and the experience of bowel preparation failure, extracts key features to construct a scheme optimization space, selects the optimal scheme to perform bowel cleansing, receives feedback images in real time and inputs them into the time sequence anomaly recognition model, dynamically generates anomaly compensation, optimizes the bowel preparation plan, and finally realizes personalized and efficient bowel preparation management.

[0103] It achieves the technical effects of realizing the design of personalized bowel preparation plans, improving the bowel cleansing effect, reducing adverse reactions, improving the patient's compliance and examination efficiency through intelligent and dynamic optimization means.

[0104] Example 2. Based on the same inventive concept as the intelligent intestinal preparation plan management method for colonoscopy patients in the foregoing example, as Figure 2 shown, the present application provides an intelligent intestinal preparation plan management system for colonoscopy patients. The system and method embodiments in the present application are based on the same inventive concept. Among them, the system includes:

[0105] An information collection module 11, which is used to collect information of patients and establish a basic information set of patients. The basic information set includes age, weight, health status, and history of constipation.

[0106] A first feature extraction module 12, which is used to extract features based on the basic information set and establish a first feature set.

[0107] A second feature extraction module 13, which is used to obtain the intestinal preparation failure experience of patients and establish a second feature set according to the intestinal preparation failure experience.

[0108] An optimization space establishment module 14, which is used to optimize the intestinal preparation plan based on the first feature set and the second feature set, and establish an optimization space. The optimization space includes multiple groups of intestinal preparation plans.

[0109] An intestinal cleansing feedback module 15, which is used to perform intestinal cleansing of patients using the optimal intestinal preparation plan in the optimization space, and receive patient feedback images with time sequence identifiers. The patient feedback images are patient stool images.

[0110] A time sequence anomaly compensation module 16, which is used to input the patient feedback images into a time sequence anomaly recognition model to generate time sequence anomaly compensation.

[0111] A dynamic plan optimization module 17, which is used to optimize the plan in the optimization space according to the time sequence anomaly compensation and the optimal intestinal preparation plan, and establish an optimization result of the optimal intestinal preparation plan.

[0112] Furthermore, the second feature extraction module 13 is further used to perform the following steps:

[0113] Establish a standardized questionnaire, send the standardized questionnaire to the patient, and receive the patient's feedback questionnaire; read the electronic medical records of the hospital, identify the objective data of the patient's failed bowel preparation through the electronic medical records, and establish an objective data set; use the patient's feedback questionnaire and the objective data set as the experience of failed bowel preparation, perform feature extraction through a joint authentication network, and establish a second feature set. The feature extraction includes compliance feature extraction, cleaning effect feature extraction, and adverse reaction feature extraction, and the feature extraction is the joint authentication feature extraction of subjective data and objective data.

[0114] Further, the optimization space establishment module 14 is further configured to perform the following steps:

[0115] Analyze the first feature set and the second feature set to establish key features; use the key features to construct a set of bowel preparation plans; perform multi-objective optimization on the set of bowel preparation plans to establish the optimization space, where the optimization fitness features of the multi-objective optimization include cleaning effect features, adverse reaction features, and compliance features.

[0116] Further, the optimization space establishment module 14 is further configured to perform the following steps:

[0117] Establish a compensation weight for the specific population; perform authentication of the specific population on the patient to establish an authentication coefficient; after correcting the compensation weight through the authentication coefficient, perform multi-objective optimization constraints, and establish the optimization space according to the constraint results.

[0118] Further, the bowel cleaning feedback module 15 is further configured to perform the following steps:

[0119] Perform data supervision on the patient to establish an education reception data set; perform education fitting with the education reception data set and the optimal bowel preparation plan to establish an optimized education plan; use the optimized education plan to perform education management on the patient.

[0120] Further, the system further includes:

[0121] A real-time information acquisition module, which is used to use the client to read the real-time acquisition information of the colonoscopy device. The real-time acquisition information includes the acquired image and the operation time, and upload the real-time acquisition information to the server; an image feedback score establishment module, which is used to use the convolutional neural network of the server to perform image recognition on the real-time acquisition information and establish an image feedback score; a nursing management module, which is used to perform nursing management on the patient according to the image feedback score.

[0122] Further, the system further includes:

[0123] A linkage information acquisition module, which is used to activate a linkage acquisition sensor to perform synchronous linkage information acquisition of a patient and establish a linkage data set when the colonoscopy device performs data acquisition. The linkage data set includes vital sign data and pain score data; a warning identification module, which is used to perform warning identification according to the vital sign data and the pain score data, establish a warning signal, and give a warning through the warning signal.

[0124] It should be noted that the above-mentioned order of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above-mentioned specific embodiments of this specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0125] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included within the protection scope of the present application.

[0126] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for managing an intelligent intestinal preparation program for patients undergoing colonoscopy, characterized in that: The method comprises: Collecting information from patients and establishing a basic information set for the patients, wherein the basic information set includes age, weight, health status, and constipation history; Perform feature extraction based on the basic information set to establish a first feature set; Acquire the patient's intestinal preparation failure experience, and establish a second feature set according to the intestinal preparation failure experience; Optimizing the intestinal preparation scheme based on the first feature set and the second feature set to establish an optimization space, wherein the optimization space includes multiple groups of intestinal preparation schemes; Using the optimal intestinal preparation scheme in the optimization space to clean the patient's intestines, and receiving a patient feedback image with a time sequence mark, wherein the patient feedback image is a patient's stool image; Inputting the patient feedback image into a timing abnormality recognition model to generate timing abnormality compensation; According to the timing anomaly compensation and the optimal intestinal preparation scheme, scheme optimization is performed in the optimization space to establish an optimization result of the optimal intestinal preparation scheme.

2. The method for managing an intelligent intestinal preparation program for colonoscopy patients according to claim 1, characterized in that: The method further comprises: Using the client to read the real-time acquisition information of the colonoscopy device, the real-time acquisition information includes the acquisition image and the operation time, and uploading the real-time acquisition information to the server; Using the convolutional neural network of the server to perform image recognition on the real-time collected information and establish an image feedback score; The patient's nursing management is performed based on the image feedback score.

3. The intelligent intestinal preparation program management method for colonoscopy patients according to claim 2, characterized in that: The method further comprises: When the colonoscopy device performs data collection, the linkage collection sensor is activated to perform synchronous linkage information collection of the patient and establish a linkage data set, wherein the linkage data set includes vital sign data and pain score data; Early warning identification is performed based on the vital sign data and the pain score data, an early warning signal is established, and an early warning alarm is issued through the early warning signal.

4. The method for managing an intelligent intestinal preparation program for colonoscopy patients according to claim 1, characterized in that: The step of obtaining the patient's intestinal preparation failure experience and establishing a second feature set according to the intestinal preparation failure experience includes: Establishing a standardized questionnaire, sending the standardized questionnaire to patients, and receiving patient feedback questionnaires; Reading the electronic medical records of the hospital, identifying the patient's intestinal preparation failure objectively through the electronic medical records, and establishing an objective data set; The patient feedback questionnaire and the objective data set are used as intestinal preparation failure experience, and feature extraction is performed through a joint certification network to establish a second feature set. The feature extraction includes compliance feature extraction, cleaning effect feature extraction, and adverse reaction feature extraction, and the feature extraction is a joint certification feature extraction of subjective data and objective data.

5. The method for managing an intelligent intestinal preparation program for colonoscopy patients according to claim 1, characterized in that: The step of optimizing the intestinal preparation scheme based on the first feature set and the second feature set and establishing an optimization space includes: Analyze the first feature set and the second feature set to establish key features; constructing a bowel preparation program set using the key features; A multi-objective optimization is performed on the intestinal preparation scheme set to establish the optimization space, wherein the optimization fitness characteristics of the multi-objective optimization include cleaning effect characteristics, adverse reaction characteristics, and compliance characteristics.

6. The method for managing an intelligent intestinal preparation program for colonoscopy patients according to claim 5, characterized in that: The performing multi-objective optimization on the intestinal preparation scheme set to establish the optimization space includes: Establish compensation weights for specialized groups; Performing the specialized group certification on the patients and establishing a certification coefficient; After the compensation weight is corrected by the authentication coefficient, multi-objective optimization constraints are executed, and the optimization space is established according to the constraint results.

7. The method for managing an intelligent intestinal preparation program for colonoscopy patients according to claim 1, characterized in that: The method of using the optimal intestinal preparation scheme in the optimization space to clean the patient's intestines also includes: Conduct data supervision on the patients and establish a data set for education reception; Performing education fitting based on the education reception data set and the optimal intestinal preparation plan to establish an optimized education plan; The optimized education program is used to carry out education and management of patients.

8. An intelligent bowel preparation program management system for colonoscopy patients, characterized in that: The system comprises: An information collection module, which is used to collect information about the patient and establish a basic information set of the patient, wherein the basic information set includes age, weight, health status, and constipation history; A first feature extraction module, the first feature extraction module is used to extract features based on the basic information set to establish a first feature set; A second feature extraction module, the second feature extraction module is used to obtain the patient's experience of failed bowel preparation and establish a second feature set according to the failed bowel preparation experience; An optimization space establishment module, the optimization space establishment module is used to optimize the intestinal preparation scheme based on the first feature set and the second feature set, and establish an optimization space, wherein the optimization space includes multiple groups of intestinal preparation schemes; An intestinal cleaning feedback module, the intestinal cleaning feedback module is used to perform intestinal cleaning of the patient using the optimal intestinal preparation scheme in the optimization space, and receive a patient feedback image with a time sequence mark, the patient feedback image being a patient stool image; A timing anomaly compensation module, the timing anomaly compensation module is used to input the patient feedback image into the timing anomaly recognition model to generate timing anomaly compensation; A dynamic scheme optimization module is used to optimize the scheme in the optimization space according to the timing anomaly compensation and the optimal intestinal preparation scheme, and establish an optimization result of the optimal intestinal preparation scheme.

Images