Clinical data-based colonoscope operation nursing scheme determination method and system

Abstract

The invention provides a colonoscope operation nursing scheme determination method and system based on clinical data, and relates to the technical field of colonoscope examination. The method comprises the steps of obtaining clinical data of a patient, obtaining a first feature vector and a second feature vector according to the clinical data, inputting the first feature vector into an improved CNN network, outputting a first nursing scheme, inputting the second feature vector into an improved LSTM network, and outputting a second nursing scheme. According to the method, accurate analysis vectors are constructed by fusing multi-dimensional patient data, a scientific basis of a personalized operation nursing scheme is realized, meanwhile, the effectiveness and pertinence of the scheme are improved by applying double models, a data processing flow is screened and optimized by correlation, and adaptive adjustment is performed on model parameters; and the scientificity, efficiency, patient acceptability and medical service quality of colonoscope operation nursing are comprehensively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of colonoscopy, and in particular to a method and system for determining a colonoscopy surgery nursing plan based on clinical data. Background Art

[0002] In the current colonoscopy nursing practice, the formulation of nursing plans mainly depends on the personal experience and professional knowledge of medical staff. Although it can meet the basic needs of patients to a certain extent, its subjectivity and uncertainty limit the optimization of nursing effects. In recent years, with the development of medical informatization, some medical institutions have begun to try to use mathematical models or algorithms to predict and formulate nursing plans in order to improve the accuracy and efficiency of nursing. However, the existing model-based nursing plan prediction methods have significant shortcomings.

[0003] First, these methods fail to fully utilize the rich clinical data resources when building models, such as the patient's medical history, physiological indicators, psychological status, etc., which are essential for developing personalized care plans. Therefore, models that lack clinical data support are greatly reduced in predictive accuracy and practicality.

[0004] Secondly, existing models ignore the characteristics of input and output data during design and application. Clinical data are often complex, diverse, and nonlinear, and existing models often fail to make adaptive adjustments and optimizations to these characteristics, resulting in poor performance of the models when processing actual data.

[0005] Finally, in terms of model training parameter setting, existing methods are also obviously inapplicable. Due to the lack of in-depth understanding of clinical data characteristics, the selection and adjustment of parameters in the model training process often rely on experience or default settings, which not only affects the training effect of the model, but also limits the generalization ability of the model in practical applications.

[0006] In summary, the existing colonoscopy nursing plan formulation methods, whether based on experience or model prediction, have obvious limitations and shortcomings. Therefore, it is urgent to develop a new colonoscopy nursing plan determination method and system that can make full use of clinical data, consider the characteristics of input and output data, and reasonably set model training parameters, so as to improve the accuracy and personalization of the nursing plan. Summary of the invention

[0007] In view of this, the present invention provides a method and system for determining a colonoscopy surgery nursing plan based on clinical data to solve the above-mentioned problems.

[0008] A method for determining a colonoscopy surgery nursing plan based on clinical data, specifically comprising the following steps: Step S1, obtaining clinical data of the patient to be tested, wherein the clinical data includes personal information, biochemical examination report and preoperative psychological monitoring index information; Step S2, obtaining a first feature vector according to the personal information and the biochemical examination report; Step S3, inputting the first feature vector into the trained first model, and outputting a first nursing plan; the learning rate of the first model is dynamically adjusted through a formula; Step S4, obtaining the patient's stress score according to the preoperative psychological monitoring index information; Step S5, obtaining a second eigenvector according to the first eigenvector and the patient's stress score; Step S6, input the second feature vector into the trained second model, and output a second nursing plan; the number of stacking layers of the second model is dynamically adjusted through a formula.

[0009] Furthermore, in step S2, a first feature vector is obtained according to the personal information and the biochemical examination report, specifically: Step S21, according to the correlation coefficient between the personal information and the key indicators of colonoscopy surgery, the personal information is screened to obtain a personal information vector; Step S22, screening the biochemical examination report according to the correlation coefficient between the biochemical examination report and the key indicators of colonoscopy surgery to obtain a biochemical indicator vector; Step S23, forming a first feature vector by using the personal information vector and the biochemical indicator vector.

[0010] Furthermore, in step S21, the personal information is screened according to the correlation coefficient between the personal information and the key indicators of colonoscopy surgery, specifically: Step S211, obtaining medical history information in the personal information; Step S212, calculating the correlation coefficient between each indicator in the past medical history information and the key indicators of colonoscopy surgery; Step S213, if the correlation coefficient is less than or equal to the first threshold, retain the corresponding indicator, otherwise delete the corresponding indicator; In step S22, the biochemical examination report is screened for indicators according to the correlation coefficient between the biochemical examination report and the key indicators of colonoscopy surgery, specifically: Step S221, obtaining abnormal items in the biochemical examination report; Step S222, calculating the correlation coefficient between each of the abnormal items and the key indicators of colonoscopy surgery; Step S223: if the correlation coefficient is greater than the second threshold, the corresponding indicator is retained; otherwise, the corresponding indicator is deleted.

[0011] Furthermore, the key indicators of colonoscopy surgery include: white blood cell count, red blood cell count, platelet count, prothrombin time, calprotectin, thrombin time, fibrinogen, alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin, urea nitrogen, creatinine, sodium ion, potassium ion, chloride ion, hepatitis B surface antigen, hepatitis C antibody, AIDS antibody and blood sugar.

[0012] Furthermore, in step S4, the patient's stress score is obtained according to the preoperative psychological monitoring index information, specifically: Step S41, obtaining the actual value, standard value, maximum normal value and minimum normal value of each psychological monitoring indicator according to the preoperative psychological monitoring indicator information; Step S42, calculating the patient's stress score based on the data of the psychological monitoring index; the specific formula is: ; Among them, T is the stress score, N is the number of psychological monitoring indicators, k is the influencing factor, T i实际值 is the actual measured value of the ith indicator, T i标准值 is the standard value of the ith indicator, T i最大值 is the maximum normal value of the ith indicator, T i最小值 is the minimum normal value of the ith indicator.

[0013] Furthermore, the first model is a convolutional neural network, and the learning rate adjustment formula is: ; in, α t is the learning rate of the tth iteration; α 0 is the initial learning rate; ω is the attenuation factor; t is the current iteration number; β is the learning rate adjustment factor; ||e t || is the norm of the output error in the current iteration step; ||Δx t ||² is the square norm of the difference between the current input feature and the input feature of the previous iteration; σx² is the square of the standard deviation of the input feature; ε is a positive constant.

[0014] Furthermore, the second model is an LSTM network, and the stacking layer number adjustment formula is: ; in: Lt is the number of stacking layers at the tth iteration; L base is the original number of stacked layers; γ is the stacking layer adjustment factor; P prev is the loss value of the previous iteration; P target is the preset target loss value; ||Δx|| is the change in input data x between the current iteration and the previous iteration; σx is the standard deviation of the input data x; θ is a positive constant.

[0015] Furthermore, the first nursing plan includes: How to popularize knowledge about colonoscopy, recommended diet, medication method, recommended amount of water intake, types of contraindications, normal reactions and adverse reactions after medication, if described 3.

[0016] Furthermore, the second nursing plan includes: Precautions during surgery, emotional soothing methods, breathing rhythm regulation, abdominal massage methods and types of music played.

[0017] A colonoscopy surgery nursing plan determination system based on clinical data, the system adopts the colonoscopy surgery nursing plan determination method based on clinical data as described in any of the above items, and specifically includes the following modules: A clinical data acquisition module is used to acquire clinical data of the patient to be tested, wherein the clinical data includes personal information, biochemical examination reports and preoperative psychological monitoring index information; A first feature vector acquisition module, connected to the clinical data acquisition module, for acquiring a first feature vector according to the personal information and the biochemical examination report; A first model is connected to the first feature vector acquisition module and is used to output a first nursing plan according to the first feature vector; the learning rate of the first model is dynamically adjusted through a formula; a second feature vector acquisition module, connected to the first feature vector acquisition module, for acquiring a patient's tension score according to the preoperative psychological monitoring index information; and acquiring a second feature vector according to the first feature vector and the patient's tension score; The second model is connected to the second feature vector acquisition module and is used to output a second nursing plan according to the second feature vector; the number of stacking layers of the second model is dynamically adjusted through a formula.

[0018] Compared with the prior art, the beneficial effects of the present invention are: First, by integrating the patient's personal information, biochemical test reports and psychological monitoring data, the present invention can construct an analysis vector that accurately reflects the patient's overall condition, which not only includes the patient's physiological condition but also reflects his or her psychological state, thereby providing a scientific basis for formulating personalized and accurate preoperative care plans. Compared with traditional care plans based on the experience and subjective judgment of medical staff, the present invention is more scientific and objective. Secondly, based on the acquired feature vectors, the present invention uses two models to output nursing plans for different periods, which not only improves the effectiveness and pertinence of the nursing plans, but also makes it easier for patients to accept and cooperate with nursing work, thereby improving the overall quality and efficiency of colonoscopy surgery nursing; Third, the present invention effectively reduces the amount of data by screening the correlation between personal information and biochemical test reports, which not only reduces the complexity and time cost of data processing, but also improves the accuracy and reliability of the analysis results. By optimizing the nursing plan determination process, the present invention can more quickly develop personalized surgical nursing plans for patients, thereby improving the efficiency and satisfaction of medical services; Fourthly, the present invention adaptively adjusts and optimizes the key parameters of the model in view of the complexity, diversity and nonlinearity of clinical data, which not only improves the model's ability to interpret clinical data, but also further enhances the practicality and reliability of the model in the formulation of colonoscopy surgery nursing plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 To identify a methodological flow chart for a colonoscopy care plan based on clinical data; Figure 2 The curve diagram of patients' acceptance of the generated nursing plan when different influencing factors are taken; Figure 3 This is a nursing plan interface diagram generated in one embodiment of the present application; Figure 4 Determine the system architecture for a clinical data-based colonoscopy care program. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0022] The present invention proposes a method and system for determining a colonoscopy nursing plan based on clinical data. In view of the defects of the prior art, the present invention obtains the patient's clinical data, obtains the first eigenvector and the second eigenvector according to the clinical data, inputs the first eigenvector into the improved CNN network, outputs the first nursing plan, and inputs the second eigenvector into the improved LSTM network to output the second nursing plan. The present invention constructs a precise analysis vector by integrating multi-dimensional patient data to realize the scientific basis of personalized surgical nursing plans. At the same time, the dual model is used to improve the effectiveness and pertinence of the plan, the correlation screening optimizes the data processing process, and the model parameters are adaptively adjusted to comprehensively improve the scientificity, efficiency, patient acceptance and medical service quality of colonoscopy nursing.

[0023] The specific implementation of the present invention is described below with reference to the accompanying drawings (tables).

[0024] Example 1 like Figure 1 As shown, the present invention proposes a method for determining a colonoscopy surgery nursing plan based on clinical data, which specifically includes steps S1-S6: Step S1, obtaining clinical data of the patient to be tested, wherein the clinical data includes personal information, biochemical examination report and preoperative psychological monitoring index information; Personal information, such as age, medical history, and allergy history, will affect the personalized customization and risk assessment of the nursing plan; the results of biochemical examination reports, such as blood sugar and electrolyte balance, will directly affect the specific measures such as drug use and dietary adjustment in the nursing plan; preoperative psychological monitoring index information is directly related to the patient's cooperation and anxiety level, which in turn affects the smoothness of the entire colonoscopy process and the ultimate success or failure. Therefore, comprehensive acquisition of this data is crucial for formulating a scientific and effective surgical nursing plan.

[0025] The existing technology analyzes this information based on doctors' experience, which may lead to subjective neglect of certain information or focus on certain information, resulting in low accuracy and strain on medical resources. This application uses this information as basic research data, which not only ensures the comprehensiveness of the data but also avoids subjective assumptions.

[0026] Step S2, obtaining a first feature vector according to the personal information and the biochemical examination report; specifically: Step S21, based on the correlation coefficient between the personal information and the key indicators of colonoscopy surgery, the personal information is screened to obtain a personal information vector; the purpose of screening the indicators based on the correlation coefficient between the personal information and the key indicators of colonoscopy surgery is to select factors that are directly related to the surgery and have an important impact on the formulation of nursing plans from a large amount of personal information, thereby constructing a concise and effective personal information vector to provide accurate data support for the subsequent formulation of nursing plans.

[0027] Step S22, based on the correlation coefficient between the biochemical examination report and the key indicators of colonoscopy surgery, the biochemical examination report is screened to obtain a biochemical indicator vector; by analyzing the correlation coefficient between each indicator in the biochemical examination report and the key indicators of colonoscopy surgery, the biochemical indicators that are directly related to the surgery and postoperative care and have a significant impact are screened out to form a biochemical indicator vector, so as to ensure that the formulation of the nursing plan can be based on the most relevant and most critical biochemical data, thereby improving the pertinence and effectiveness of nursing.

[0028] Step S23, forming a first feature vector by using the personal information vector and the biochemical indicator vector.

[0029] In step S21, the personal information is screened according to the correlation coefficient between the personal information and the key indicators of colonoscopy surgery, specifically: Step S211, obtaining medical history information in the personal information; Step S212, calculating the correlation coefficient between each indicator in the past medical history information and the key indicators of colonoscopy surgery; Step S213, if the correlation coefficient is less than or equal to the first threshold, retain the corresponding indicator, otherwise delete the corresponding indicator; In steps S211 to S213, the medical history information in the personal information is obtained, and the correlation coefficient between the medical history information and the key indicators of colonoscopy surgery is calculated. Then, the medical history information with a low correlation with the key surgical indicators is screened out and retained, while the highly correlated ones are deleted. In this way, the key indicators that are changed by the past medical history can be removed, the personal information vector can be streamlined, and redundant information can be removed to ensure the accuracy and pertinence of the nursing plan.

[0030] In step S22, the biochemical examination report is screened for indicators according to the correlation coefficient between the biochemical examination report and the key indicators of colonoscopy surgery, specifically: Step S221, obtaining abnormal items in the biochemical examination report; Step S222, calculating the correlation coefficient between each of the abnormal items and the key indicators of colonoscopy surgery; Step S223: if the correlation coefficient is greater than the second threshold, the corresponding indicator is retained; otherwise, the corresponding indicator is deleted.

[0031] In steps S221 to S223, by identifying abnormal items in the biochemical examination report and calculating the correlation coefficients between these abnormal items and the key indicators of colonoscopy surgery, abnormal biochemical indicators that are highly correlated with the key surgical indicators are retained, while indicators with low or insignificant correlation are eliminated, aiming to focus on key biochemical abnormalities and provide a strong basis for formulating targeted nursing plans.

[0032] The key indicators of colonoscopy surgery include: white blood cell count, red blood cell count, platelet count, prothrombin time, calprotectin, thrombin time, fibrinogen, alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin, urea nitrogen, creatinine, sodium ion, potassium ion, chloride ion, hepatitis B surface antigen, hepatitis C antibody, HIV antibody and blood sugar.

[0033] Common indices for intestinal disease examination were obtained, and the correlation coefficients between these indicators and the risk of colonoscopy were calculated. The obtained coefficient ranking is shown in Table 1: Table 1 Correlation coefficient ranking table

[0034] Indicators with correlation coefficients greater than a threshold are selected as key indicators for colonoscopy. In this embodiment, the threshold is 0.85, that is, indicators with correlation coefficients greater than or equal to 0.85 are selected as key indicators, while indicators less than 0.85 are selected as general indicators.

[0035] The selection of the threshold will affect the accuracy of the model. In other embodiments, the threshold selects different values, and the accuracy of the model is different. This embodiment selects the threshold of 0.85 that makes the model accuracy the highest.

[0036] The screening of the above key indicators takes into account the indicators that are crucial to the nursing plan, greatly reduces the scope of indicators involved, reduces the amount of data, and improves calculation efficiency. It is crucial for evaluating patients' surgical tolerance, predicting surgical risks, and formulating postoperative care plans.

[0037] Step S3, inputting the first feature vector into the trained first model, and outputting a first nursing plan; The first model mainly processes the first feature vector composed of personal information vector and biochemical index vector. These feature vectors usually contain a large amount of numerical data, such as age, gender, biochemical indicators, etc. CNN performs well in processing numerical data and feature extraction, especially when the data has local correlation and spatial structure. In addition, the convolution layer and pooling layer of CNN can effectively extract features, reduce the amount of data, and improve computational efficiency.

[0038] The first model training method is as follows: Acquire historical data of the first feature vector, the historical data including multiple groups of sample data consisting of personal information vectors and biochemical indicator vectors and a first nursing plan corresponding to the sample data; The neural network model is trained using the historical data of the first feature vector until a training end condition is met, thereby obtaining a trained first model.

[0039] When analyzing the first model, the applicant found that the input and output data have the following characteristics: Input data: The first eigenvector, which is composed of the patient's personal information vector and biochemical index vector, including the patient's age, gender, medical history, key indicators in the biochemical examination report and other information.

[0040] Output data: The first nursing plan, which reflects the nursing plan for the patient determined based on the patient's physiological condition and general nursing needs.

[0041] Since the input data of the first model contains multiple types of clinical information, there may be complex correlations and nonlinear relationships between these data. Therefore, the model needs to continuously adjust the weights when learning these data to adapt to the complexity of the data. Therefore, the learning rate needs to be adjusted in real time to optimize model performance, improve training efficiency, and adapt to the complexity of the data.

[0042] An embodiment of the present invention is based on the commonly used Adam optimizer and combines the characteristics of the first model to provide a formula for adaptively adjusting the learning rate.

[0043] ; in, α t is the learning rate of the tth iteration; α 0 is the initial learning rate; it can be set to 0.001; ω is the decay factor; it is used to control the decay rate of the learning rate over time; t is the current iteration number; β is the learning rate adjustment factor; it is used to control the influence of output error on learning rate adjustment; ||e t || is the norm of the output error in the current iteration step; ||Δx t ||² is the square norm of the difference between the current input feature and the input feature of the previous iteration, reflecting the degree of change of the input feature; σx² is the square of the standard deviation of the input feature, which is used to normalize the degree of change of the input feature; ε is a positive constant used to prevent the denominator from being zero.

[0044] The learning rate adjustment formula dynamically adjusts the learning rate to optimize the training effect of the model by comprehensively considering multiple factors such as the current number of iterations, output error norm, input feature change, etc. This formula can improve the model's ability to interpret clinical data, enhance the model's practicality and reliability in the formulation of colonoscopy surgery nursing plans, thereby improving the scientificity, efficiency and patient acceptance of nursing plans.

[0045] The first nursing plan includes: Popularization methods of colonoscopy surgery knowledge, recommended diet, medication methods, recommended water intake, types of contraindications, normal reactions and adverse reactions after medication.

[0046] Popularization of colonoscopy knowledge: Explain in detail to patients the purpose, process, possible risks and benefits of colonoscopy, as well as precautions after surgery. This can be done through oral explanations, written materials, video demonstrations, and other methods to ensure that patients fully understand and actively cooperate with the surgery.

[0047] Recommended diet: Before surgery, patients need to follow specific dietary guidelines, such as prohibiting liquid diet or low-residue, semi-liquid diet 3 days before surgery, taking laxatives to clean the intestines 1 night before surgery or cleansing enema on the day of examination, and fasting for breakfast. After surgery, the principle is to eat light and easily digestible food, and avoid irritating foods, such as liquid or semi-liquid food (rice porridge, noodle soup, etc.), high-protein food (eggs, tofu, etc.) and vitamin-rich food (cooked vegetable puree, fruit puree, etc.).

[0048] Medication method: According to the surgical requirements and the patient's specific conditions, the doctor will prescribe appropriate medications, such as intestinal cleansers, anti-foaming agents, antibiotics, etc. The nurse should inform the patient in detail of the name, dosage, time and method of medication, and emphasize the precautions during medication, such as avoiding simultaneous intake with certain foods or drinks.

[0049] Recommended water intake: Before surgery, patients need to drink a lot of water to clean their intestines. During the drinking process, you can drink a small amount of water in several times, adjust the water temperature, change the taste, etc. to alleviate the difficulty of water intake. At the same time, nurses should pay close attention to the patient's drinking and ensure that the patient can complete the intestinal preparation on time.

[0050] Contraindication type: Colonoscopy is not suitable for all patients. There are certain contraindications, such as severe cardiopulmonary disease, pregnancy, suspected intestinal perforation, etc. Before the operation, the doctor should ask the patient's medical history and physical condition in detail to ensure that the patient meets the conditions for the operation and avoid surgical risks.

[0051] Normal and adverse reactions after medication: During medication, patients may experience some normal reactions, such as abdominal distension, abdominal pain, and increased bowel movements. Nurses should explain the causes and countermeasures of these reactions to patients to reduce their anxiety and restlessness. At the same time, nurses should also pay close attention to patients' medication and promptly detect and deal with adverse reactions, such as allergic reactions and severe abdominal pain.

[0052] The first nursing plan is a comprehensive and meticulous nursing plan that aims to provide patients with comprehensive preoperative preparation and postoperative guidance. By popularizing surgical knowledge, recommending diet and medication, guiding water intake, clarifying contraindications, and monitoring medication reactions, it can ensure that patients can complete preoperative bowel cleansing safely and smoothly, be fully prepared for colonoscopy, and promote postoperative recovery.

[0053] Step S4, obtaining the patient's tension score based on the preoperative psychological monitoring index information; in medical procedures such as colonoscopy, the patient's psychological state, especially tension, is an important consideration. Tension not only reflects the patient's anxiety and fear of surgery, but may also directly affect the surgical effect and postoperative recovery. High tension may cause the patient's heart rate to increase, blood pressure to rise, increase surgical risks, and is not conducive to postoperative physical recovery. Therefore, accurately assessing the patient's tension and taking appropriate nursing measures to intervene are of great significance to ensure the smooth progress of the operation and the patient's early recovery.

[0054] Specifically: Step S41, obtaining the actual value, standard value, maximum normal value and minimum normal value of each psychological monitoring indicator according to the preoperative psychological monitoring indicator information; Step S42, calculating the patient's stress score based on the data of the psychological monitoring index; the specific formula is: ; Among them, T is the tension score, N is the number of psychological monitoring indicators, and k is the influencing factor, which is used to control the growth rate of the score with the degree of deviation. i实际值 is the actual measured value of the ith indicator, T i标准值 is the standard value of the ith indicator, T i最大值 is the maximum normal value of the ith indicator, T i最小值 is the minimum normal value of the ith indicator.

[0055] The psychological monitoring index information in this application can select indicators related to the psychological state, reflecting the multi-dimensional data of the patient's preoperative psychological state, including physiological indicators and biochemical indicators.

[0056] Physiological indicators include heart rate, skin conductance level (SCL), blood pressure, respiratory rate, heart rate variability (HRV), etc. Biochemical indicators include cortisol (saliva / serum), epinephrine / norepinephrine (plasma), serotonin (serum), C-reactive protein (CRP), blood glucose (fasting), etc.

[0057] In one embodiment, it is verified through multiple groups of clinical experiments that different values ​​of k have different effects on the stress score, thereby affecting the output result of the model.

[0058] When the k value is small, such as k=5, the score is less sensitive to indicator deviation, a small deviation from the standard value has a weaker impact on the total score, and the score distribution is flatter; the generated nursing plan is more conservative, and only intervention measures are taken for tension states with significant deviations, which is suitable for patients with less fluctuations in psychological state.

[0059] When the k value is large, such as k=8, the sensitivity is extremely high, and a small deviation can lead to a large jump in the score. The scores of extreme values ​​are quickly saturated, which may amplify the impact of noise or short-term fluctuations. The nursing plan may be too radical and needs to be verified in combination with other indicators to avoid misjudgment. It is suitable for close monitoring scenarios of high-risk patients.

[0060] In this implementation, through multiple clinical experiments, it was verified that when k=6.7, the score can effectively capture the patient's anxiety characteristics and maintain a high degree of stability. When k=6.7, the patient acceptance rate of the generated nursing plan reached 93%. The test results are as follows Figure 2 As shown in Table 2: Table 2 Patients' acceptance of nursing plans under different influencing factor values

[0061] From the above table and Figure 2 It can be seen that by combining the dynamic characteristics of clinical data with patient feedback, using cross-validation and ROC curve analysis, k=6.7 was finally determined to be the optimal solution for balancing sensitivity and stability. This value not only significantly improves the personalization level of the nursing plan, but also ensures the robustness of the model in complex clinical scenarios.

[0062] In this embodiment, the stress scoring formula fully considers multiple psychological monitoring indicators and is easy to obtain. It has the advantages of comprehensiveness, objectivity, sensitivity, comparability and ease of operation. It can provide accurate and reliable evaluation results for medical staff, thereby guiding them to take appropriate nursing measures for intervention.

[0063] Step S5, obtaining a second eigenvector according to the first eigenvector and the patient's stress score; By introducing the patient stress score, the meaning and usage of medical terms in the patient's specific psychological state can be more accurately understood and processed, thereby improving the accuracy of medical term standardization. Fusion of the first eigenvector with the patient's psychological state information can form a more comprehensive data set, providing more valuable information for subsequent data analysis and mining.

[0064] Step S6, inputting the second feature vector into the trained second model, and outputting a second nursing plan; The second model not only processes the first eigenvector, but also considers the patient’s stress score, which involves time series data (because the psychological state may change over time). LSTM performs well in processing time series data and can capture temporal and long-term dependencies in the data.

[0065] The second model training method is specifically as follows: Acquire historical data of the second feature vector, the historical data including multiple groups of sample data consisting of the first feature vector and the tension score and a second nursing plan corresponding to the sample data; The neural network model is trained using the historical data of the second feature vector until a training end condition is met to obtain a trained second model.

[0066] When analyzing the second model, the applicant found that the input and output data have the following characteristics: Characteristics of input data: The input of the second model includes the first eigenvector (personal information and biochemical test reports) and the patient's tension score, which is a multidimensional eigenvector; and the patient's mental state and tension will change over time, especially during preoperative preparation.

[0067] Output data characteristics: The output is a personalized care plan determined based on the second eigenvector, which requires the model to accurately capture subtle differences in the input data and make corresponding output adjustments.

[0068] The number of LSTM stacking layers has the following characteristics: stacking multiple LSTM layers can increase the complexity of the model, enabling it to learn more complex feature representations and data relationships; by adjusting the number of LSTM stacking layers in real time, the complexity and generalization ability of the model can be balanced during the training process; by adjusting the number of LSTM stacking layers, computing efficiency and resource utilization can be optimized while ensuring model performance.

[0069] Therefore, combined with the data characteristics of the input and output of the second model, calculating the number of LSTM stacking layers can improve the complexity of the model, avoid overfitting and underfitting, and optimize performance and efficiency, which helps the model to more accurately capture the features and information in the input data, thereby outputting more personalized care plans.

[0070] The second model is a stacking layer adjustment formula: ; in: L t is the number of stacked layers of LSTM at the tth iteration or evaluation; L base is the number of basic layers, indicating the default number of stacked layers of LSTM without performance feedback or under specific conditions; γ is the stacking layer adjustment factor, which is used to control the speed and amplitude of the layer adjustment and can be fine-tuned according to actual needs; P prev To represent the performance indicators of the model at the last iteration or evaluation (such as accuracy, loss value, etc.); P target is the preset target performance indicator, indicating the performance level that the model is expected to achieve; ||Δx|| is the change in input data x between the current iteration and the previous iteration, reflecting the dynamic nature of the input data.

[0071] σx is the standard deviation of the input data x or some stability measure, which is used to normalize the variation of the input data.

[0072] θ is a small positive number used to prevent the denominator from being zero and increase the stability of the formula.

[0073] The stacking layer adjustment formula dynamically adjusts the number of stacking layers of the LSTM network by combining the loss value of the previous iteration, the preset target loss value, the input data change and standard deviation and other parameters, thereby optimizing the complexity and generalization ability of the model, effectively improving the model's adaptability to clinical data, and being able to output more personalized colonoscopy surgery care plans, thereby improving the overall quality and efficiency of care.

[0074] The second nursing plan includes precautions during surgery, emotional soothing methods, breathing rhythm adjustment, abdominal massage methods, and music types: Precautions during surgery: During the operation, the patient should maintain a stable body posture and avoid moving around. Emotional comfort, breathing rhythm adjustment, music playing, etc. can be used to assist the patient in successfully completing the colonoscopy surgery.

[0075] Emotional soothing methods: Soothe the patient's emotions through gentle words and patient explanations, so that the patient can relax and reduce tension and fear; use professional psychological intervention techniques, such as deep breathing guidance and mindfulness meditation, to help patients relax and improve surgical tolerance; Breathing rhythm adjustment: During the colonoscopy, guide the patient to do deep breathing exercises, and relax the body and reduce tension by adjusting the breathing rhythm. This helps to lower physiological indicators such as heart rate and blood pressure, and improve the safety of the operation; Type of music played: During colonoscopy or postoperative recovery, playing soft, soothing music can help reduce patients' anxiety, increase pain threshold, and promote physical and mental relaxation. Medical staff can choose appropriate music to play according to the nursing plan. For example, some light music, classical music, or natural sounds can be selected, which usually have a soothing and relaxing effect.

[0076] Postoperative abdominal massage method: After colonoscopy, patients may experience abdominal distension, abdominal pain and other discomfort symptoms. At this time, medical staff guide patients to perform appropriate abdominal massage according to the plan output by the model to promote intestinal peristalsis and relieve abdominal distension, constipation and other discomfort symptoms. Patients and their families can also perform massage according to the plan output by the model, giving appropriate techniques and moderate strength to avoid causing additional pain or discomfort to patients due to inappropriate massage methods.

[0077] The application of the second nursing plan in colonoscopy surgery aims to provide patients with a comprehensive and meticulous intraoperative and postoperative nursing experience. By paying attention to intraoperative precautions, emotional comfort, breathing regulation, abdominal massage, and music therapy, it can promote the patient's physical and mental recovery and improve the safety and comfort of the operation.

[0078] In one embodiment, the patient information is as follows: Gender: male, age: 62 years old, primary school graduate, patient tension score: T=0.82 (significant anxiety); Biochemical abnormalities: high blood sugar (fasting blood sugar 8.5mmol / L), low platelet count (120×10 9 / L); Past medical history: hypertension, diabetes.

[0079] The first and second nursing plans output according to the patient's condition are as follows: The first nursing plan is: 1. Popularization of surgical knowledge: Use easy-to-understand pictures or animated videos (no text or a small amount of large text) to explain the colonoscopy process, focusing on "painless", "safe", and "doctor's presence throughout the process".

[0080] I repeatedly emphasized verbally: "The examination is like taking a nap, it doesn't hurt, the doctor will do it gently, and you can go home and rest after it's done." 2. Recommended diet: 3 days before surgery: Breakfast: white porridge + steamed egg (without oil or salt); Lunch: soft noodles (no vegetables, light soy sauce for flavor); Dinner: rice porridge (rice crushed and cooked into paste).

[0081] Prohibited: Rice, vegetables, fruits, beans.

[0082] 1 day before surgery: Drink only clear liquids: light salt water, filtered broth (without residue), sugar-free jelly.

[0083] On the day of surgery: Completely fast and stop drinking water 4 hours before surgery.

[0084] 3. How to take the medicine: Intestinal cleansers: Drink the "big bottle of medicine" in 4 times, half a cup (about 250ml) each time, with an interval of 15 minutes. You can hold a sugar cube in your mouth after drinking to prevent nausea.

[0085] Blood sugar control: The dose of insulin injected the night before surgery was adjusted by the physician to ensure that fasting blood glucose was ≤7.0 mmol / L.

[0086] Figure 3 This is the generated nursing plan interface diagram.

[0087] 4. Please use popular taboo tips, such as: If the platelet count is low, the patient needs to be given platelet-replenishing medication before undergoing examination to avoid bleeding.

[0088] Measure your blood pressure before the examination. If your systolic value is over 160, you need to take medicine to lower your blood pressure.

[0089] 5. Warning of adverse reactions: It is normal to feel bloated and have diarrhea after taking medicine. Only after you have a complete bowel movement can you have a thorough check.

[0090] If you feel dizzy or have cold sweats, take a candy immediately and call a nurse.

[0091] The second nursing plan is: 1. Precautions during the operation: During the examination, lie on your left side with your knees hugging your chest. The doctor will pat your shoulders and remind you not to move.

[0092] 2. Emotional soothing methods: Preoperative: Arrange family members to accompany the patient, and the nurse repeatedly comforted them in dialect: "Don't be afraid, it will be fine soon, the doctor's skills are very good." During the operation: The nurse held the patient's hand and said every 2 minutes: "Well done, it will be over soon!" 3. Breathing to adjust the rhythm: Teach the patient to "breathe in slowly through the nose (count 1-2-3), and breathe out slowly through the mouth (count 1-2-3-4)" and practice following the nurse's gestures.

[0093] 4. Music type: Choose classic songs from the 60s and 70s (such as Teresa Teng's "Sweet Honey"), light music (piano piece "Autumn Whispers"), and adjust the volume to a soft level.

[0094] 5. Postoperative abdominal massage: Family members are advised to "rub the area around the belly button clockwise, as gently as petting a kitten", three times a day, each time for five minutes.

[0095] like Figure 4 As shown, the embodiment of the present invention further proposes a colonoscopy surgery nursing plan determination system based on clinical data, using the colonoscopy surgery nursing plan determination method based on clinical data as described in any of the above items, including the following modules: A clinical data acquisition module is used to acquire clinical data of the patient to be tested, wherein the clinical data includes personal information, biochemical examination reports and preoperative psychological monitoring index information; A first feature vector acquisition module, connected to the clinical data acquisition module, for acquiring a first feature vector according to the personal information and the biochemical examination report; A first model is connected to the first feature vector acquisition module and is used to output a first nursing plan according to the first feature vector; the learning rate of the first model is dynamically adjusted through a formula; a second feature vector acquisition module, connected to the first feature vector acquisition module, for acquiring a patient's tension score according to the preoperative psychological monitoring index information; and acquiring a second feature vector according to the first feature vector and the patient's tension score; The second model is connected to the second feature vector acquisition module and is used to output a second nursing plan according to the second feature vector; the number of stacking layers of the second model is dynamically adjusted through a formula.

[0096] An embodiment of the present invention further provides an electronic device, the electronic device comprising: Processor and memory; The processor is used to execute the steps of the method for determining a colonoscopy surgery care plan based on clinical data as described in any of the above items by calling the program or instruction stored in the memory.

[0097] An embodiment of the present invention further provides a computer-readable storage medium, which includes computer program instructions, and the computer program instructions enable a computer to execute the steps of a method for determining a colonoscopy surgery care plan based on clinical data as described in any one of the above items.

[0098] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for determining a colonoscopy nursing plan based on clinical data, characterized in that: The steps include: Step S1, obtaining clinical data of the patient to be tested, wherein the clinical data includes personal information, biochemical examination report and preoperative psychological monitoring index information; Step S2, obtaining a first feature vector according to the personal information and the biochemical examination report; Step S3, inputting the first feature vector into the trained first model, and outputting a first nursing plan; the learning rate of the first model is dynamically adjusted through a formula; Step S4, obtaining the patient's stress score according to the preoperative psychological monitoring index information; Step S5, obtaining a second eigenvector according to the first eigenvector and the patient's stress score; Step S6, input the second feature vector into the trained second model, and output a second nursing plan; the number of stacking layers of the second model is dynamically adjusted through a formula.

2. A method for determining a colonoscopy surgery nursing plan based on clinical data according to claim 1, characterized in that: In step S2, a first feature vector is obtained according to the personal information and the biochemical examination report, specifically: Step S21, according to the correlation coefficient between the personal information and the key indicators of colonoscopy surgery, the personal information is screened to obtain a personal information vector; Step S22, screening the biochemical examination report according to the correlation coefficient between the biochemical examination report and the key indicators of colonoscopy surgery to obtain a biochemical indicator vector; Step S23, forming a first feature vector by using the personal information vector and the biochemical indicator vector.

3. A method for determining a colonoscopy nursing plan based on clinical data according to claim 2, characterized in that: In step S21, the personal information is screened according to the correlation coefficient between the personal information and the key indicators of colonoscopy surgery, specifically: Step S211, obtaining medical history information in the personal information; Step S212, calculating the correlation coefficient between each indicator in the past medical history information and the key indicators of colonoscopy surgery; Step S213, if the correlation coefficient is less than or equal to the first threshold, retain the corresponding indicator, otherwise delete the corresponding indicator; In step S22, the biochemical examination report is screened for indicators according to the correlation coefficient between the biochemical examination report and the key indicators of colonoscopy surgery, specifically: Step S221, obtaining abnormal items in the biochemical examination report; Step S222, calculating the correlation coefficient between each of the abnormal items and the key indicators of colonoscopy surgery; Step S223: if the correlation coefficient is greater than the second threshold, the corresponding indicator is retained; otherwise, the corresponding indicator is deleted.

4. A method for determining a colonoscopy surgery nursing plan based on clinical data according to claim 3, characterized in that: The key indicators of colonoscopy surgery include: white blood cell count, red blood cell count, platelet count, prothrombin time, calprotectin, thrombin time, fibrinogen, alanine aminotransferase, aspartate aminotransferase, total bilirubin, direct bilirubin, urea nitrogen, creatinine, sodium ion, potassium ion, chloride ion, hepatitis B surface antigen, hepatitis C antibody, AIDS antibody and blood sugar.

5. The method for determining a colonoscopy surgery nursing plan based on clinical data according to claim 1, characterized in that: In step S4, the patient's stress score is obtained according to the preoperative psychological monitoring index information, specifically: Step S41, obtaining the actual value, standard value, maximum normal value and minimum normal value of each psychological monitoring indicator according to the preoperative psychological monitoring indicator information; Step S42, calculating the patient's stress score based on the data of the psychological monitoring index; the specific formula is: ； Among them, T is the stress score, N is the number of psychological monitoring indicators, k is the influencing factor, T i实际值 is the actual measured value of the ith indicator, T i标准值 is the standard value of the ith indicator, T i最大值 is the maximum normal value of the ith indicator, T i最小值 is the minimum normal value of the ith indicator.

6. The method for determining a colonoscopy nursing plan based on clinical data according to claim 1, characterized in that: The first model is a convolutional neural network, and the adjustment formula of the learning rate is: ； in, α t is the learning rate of the tth iteration; α0 is the initial learning rate; ω is the attenuation factor; t is the current iteration number; β is the learning rate adjustment factor; ||e t || is the norm of the output error in the current iteration step; ||Δx t ||² is the square norm of the difference between the current input feature and the input feature of the previous iteration; σx² is the square of the standard deviation of the input feature; ε is a positive constant.

7. The method for determining a colonoscopy nursing plan based on clinical data according to claim 1, characterized in that: The second model is an LSTM network, and the adjustment formula for the number of stacked layers is: ； in: L t is the number of stacking layers at the tth iteration; L base is the original number of stacked layers; γ is the stacking layer adjustment factor; P prev is the loss value of the previous iteration; P target is the preset target loss value; ||Δx|| is the change in input data x between the current iteration and the previous iteration; σx is the standard deviation of the input data x; θ is a positive constant.

8. The method for determining a colonoscopy surgery nursing plan based on clinical data according to claim 1, characterized in that: The first nursing plan includes: Popularization methods of colonoscopy surgery knowledge, recommended diet, medication methods, recommended water intake, types of contraindications, normal reactions and adverse reactions after medication.

9. The method for determining a colonoscopy nursing plan based on clinical data according to claim 1, characterized in that: The second nursing plan includes: Precautions during surgery, emotional soothing methods, breathing rhythm regulation, abdominal massage methods and types of music played.

10. A colonoscopy surgery nursing plan determination system based on clinical data, characterized in that: The system adopts a method for determining a colonoscopy surgery nursing plan based on clinical data as described in any one of claims 1 to 9, and specifically includes the following modules: A clinical data acquisition module is used to acquire clinical data of the patient to be tested, wherein the clinical data includes personal information, biochemical examination reports and preoperative psychological monitoring index information; A first feature vector acquisition module, connected to the clinical data acquisition module, for acquiring a first feature vector according to the personal information and the biochemical examination report; A first model is connected to the first feature vector acquisition module and is used to output a first nursing plan according to the first feature vector; the learning rate of the first model is dynamically adjusted through a formula; a second feature vector acquisition module, connected to the first feature vector acquisition module, for acquiring a patient's tension score according to the preoperative psychological monitoring index information; and acquiring a second feature vector according to the first feature vector and the patient's tension score; The second model is connected to the second feature vector acquisition module and is used to output a second nursing plan according to the second feature vector; the number of stacking layers of the second model is dynamically adjusted through a formula.

Images