Cancer patient home pain management decision-making system based on patient report outcome

Abstract

The invention relates to the technical field of nursing information management, and particularly discloses a cancer patient home pain management decision-making system based on a patient report outcome, which is used for solving the challenges of uneven data quality, slow update feedback speed, large and diverse data volume and the like in pain management of cancer patients in a home environment. The system comprises an intelligent sensing and data aggregation module, a cloud server, a distributed computing and data processing module, an intelligent decision support module and a user interaction feedback module. Noise and abnormal values can be removed through dynamic data cleaning and data standardization, data consistency is ensured through multi-source data fusion, real-time data is analyzed through large-scale data analysis and an LSTM model, future pain states and risks of cancer patients are predicted, personalized management suggestions are generated, and the efficiency and effect of home pain management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of nursing information management, and more specifically, to a cancer patient's home pain management decision-making system based on patient-reported outcomes. Background Art

[0002] Patient-Reported Outcomes (PROs) refer to any information directly reported by patients and related to their health status, rather than obtained through the interpretation and judgment of medical staff. PROs include information on patients' symptoms, functional status, quality of life, pain level, emotional state, etc. These data are usually collected through questionnaires, diaries, electronic devices, etc. The key to PROs is the subjectivity and real-time nature of their data sources, which can provide first-hand patient experience data. In a home environment, pain management for cancer patients involves the collection, transmission, storage and analysis of a large amount of data. Data collection in a home environment mainly relies on patients' active reports, which are affected by the patients' own understanding ability, equipment operation level and emotional state, resulting in uneven data quality. The lack of real-time data collection means slows down data updates and feedback. In addition, the amount of data in a home environment is large and diverse, involving structured and unstructured data. How to efficiently process and analyze these data and mine useful information has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art, the present invention provides a home pain management decision system for cancer patients based on patient-reported outcomes, which removes noise and outliers through dynamic data cleaning and data standardization, ensures data consistency by multi-source data fusion, and analyzes real-time data through large-scale data analysis and LSTM model to predict the future pain status and risk of cancer patients, generate personalized management suggestions, and improve the efficiency and effectiveness of home pain management.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A cancer patient home pain management decision system based on patient-reported outcomes includes an intelligent perception and data aggregation module, a cloud server, a distributed computing and data processing module, an intelligent decision support module, and a user interaction feedback module. The intelligent perception and data aggregation module is connected to the distributed computing and data processing module through a home network and a secure communication protocol, and directly communicates with the cloud server. The processing results of the distributed computing and data processing module are transmitted to the intelligent decision support module. The results generated by the intelligent decision support module are displayed to the patient's medical professionals through the user interaction feedback module. The user feedback data collected by the user interaction feedback module is transmitted back to the intelligent perception and data aggregation module and the distributed computing and data processing module. The intelligent perception and data aggregation module uses a dynamic data cleaning algorithm to remove The noise and outliers are eliminated, the data format is standardized, and after multi-source data fusion, the data is transmitted to the cloud server. The distributed computing and data processing module performs preliminary data analysis and data filtering on the edge device, responds in real time to data that requires timely feedback, transmits the processed data to the cloud server, and stores the data in the distributed database system. The feature extraction algorithm is used to extract key features, and the distributed computing framework is applied to perform large-scale data analysis, identify complex patterns and trends, and transmit the acquired features and preliminary analysis reports to the intelligent support module. The intelligent support module analyzes the current data through a pre-trained deep learning model, predicts the future pain status of cancer patients, and automatically generates personalized pain management recommendations for cancer patients at home based on the patient's historical data and prediction results.

[0006] As a further solution of the present invention, the intelligent perception and data aggregation module includes a data acquisition unit, a dynamic data cleaning unit and a multi-source data aggregation unit. The data acquisition unit includes a smart phone, a tablet computer, a home monitoring terminal, a wearable biosensor and an intelligent medical device. The collected data include pain score, emotional state score, nausea index, fatigue index, headache index, sleep quality score, daily activity ability score, heart rate, blood pressure, body temperature, blood oxygen level, blood sugar level, indoor PM2.5 concentration, and indoor carbon dioxide level.

[0007] As a further solution of the present invention, the specific implementation process of the multi-source data aggregation unit includes:

[0008] Step 11, data collection: the data collection unit collects data from smartphones, tablet computers, home monitoring terminals, wearable biosensors, and smart medical devices;

[0009] Step 12, data standardization and format conversion: unify the data formats and units of different data sources, convert the data collected by different devices and sensors into a standard preset format, unify the units of data collected by different devices, and synchronize the data collected by different devices according to a unified time standard;

[0010] Step 13, dynamic data cleaning: Dynamic data cleaning is performed on the pain score data of cancer patients at home based on a preset dynamic data cleaning strategy, and then the cleaned data and other data are transmitted to the multi-source data fusion unit through the home network and secure communication protocol;

[0011] Step 14, multi-source data fusion: align all data points in chronological order, assign weights to each data source according to its reliability and importance, and use the Bayesian network method to fuse the data. For different types of data, multi-dimensional splicing and fusion are performed to form a comprehensive pain management data set for cancer patients at home;

[0012] Step 15, data storage and transmission: Store the cleaned and integrated data in the distributed database system, and transmit the data to the cloud server and distributed computing and data processing module through a secure communication protocol.

[0013] As a further solution of the present invention, in the dynamic data cleaning unit, the pain score data of cancer patients at home are detected and processed using the dynamic interquartile range method, and the specific process includes:

[0014] Step 1, data preparation: collect and organize the pain score dataset of cancer patients at home, the dataset is S, the number of data points in the dataset S is n, sort the data in the pain score dataset S from small to large, and obtain the sorted dataset S sorted ;

[0015] Step 2, calculate the quartiles: take 25% of the data points in the sorted data set as the first quartile, take 75% of the data points in the sorted data set as the third quartile, take the middle data point of the data set as the median, the total number of data points n is an odd number, and the first quartile is equal to S sorted Middle The value of the data point, the third quartile is S sorted Middle The value of the data point, the total number of data points n is an even number, and the first quartile is equal to S sorted Middle The value of the data point is The mean of the values ​​of the data points, the third quartile is equal to S sorted Middle The value of the data point is The mean of the values ​​of the data points;

[0016] Step 3, calculate the interquartile range: the interquartile range is equal to the third quartile minus the first quartile;

[0017] Step 4: Determine the range of outliers: The lower bound of the pain score of cancer patients at home is max(0,Q 1 -ξ×k IQR ), the upper bound is min(10,Q 3 +ξ×k IQR ), where Q 1 , Q 3 are the first and third quartiles respectively, k IQR is the interquartile range, ξ is the sensitivity parameter of the pain score, which is set according to the actual situation;

[0018] Step 5, detect and process outliers: For each data point, if the data point is less than the lower limit or greater than the upper limit, it is marked as an outlier. The authenticity of the outlier is further verified by checking with the patient. For confirmed outliers, the mean of the adjacent data points is used for correction. Unconfirmed outliers are removed and the median of the adjacent data points is interpolated for correction.

[0019] As a further scheme of the present invention, the dynamic data cleaning unit obtains the pain score data of cancer patients at home, obtains the home pain score personality characteristic value through a preset pain score personality criterion number formula, obtains the home pain score characteristic category of the cancer patients based on the home pain score personality characteristic value according to a preset threshold range, obtains the sensitivity parameter of the cancer patients' home pain score according to the rules set for each category, and then performs outlier detection and processing on the patients' pain scores through the interquartile range method of the cancer patients' home pain scores.

[0020] As a further solution of the present invention, in the dynamic data cleaning unit, the preset pain score individual criterion number formula nonlinearly combines the volatility, discreteness and data trend of the pain score data in the patient's active report for the past month to obtain the pain score individual criterion number of the cancer patient at home. The preset pain score individual criterion number formula is:

[0021]

[0022] Where: C is the number of personality criteria for pain scores of cancer patients at home, k MAD is the median absolute deviation of the pain score data in the past month, μ recent is the average pain score in the past month, σ recent is the standard deviation of pain scores in the past month, k trend The data trend of pain scores in the past month was calculated by the moving average method.

[0023] As a further solution of the present invention, in the dynamic data cleaning unit, based on the set classification threshold, the pain scores of cancer patients at home are divided into low-fluctuation pain class, medium-fluctuation pain class and high-fluctuation pain class according to the obtained value of the pain score personality criterion number, and the classification standards of low-fluctuation pain class, medium-fluctuation pain class and high-fluctuation pain class are C≤C L , C L <C≤C M 、C>C M , where C L and C M They are the set low category threshold and medium category threshold respectively.

[0024] As a further solution of the present invention, in the dynamic data cleaning unit, the pain score of the cancer patient at home is low-fluctuation pain, and the acquisition rule of the sensitive parameter of the pain score is:

[0025]

[0026] Where: 0 is the initial pain score sensitivity parameter, which is 1.5, α 1 It is the index adjustment parameter, which can be set according to the requirements.

[0027] As a further solution of the present invention, in the dynamic data cleaning unit, the pain score of the cancer patient at home is a medium fluctuation pain type, and the acquisition rule of the pain score sensitive parameter is:

[0028]

[0029] Where: β 1 , β 2 They are Item and The impact factor of the item is set according to the requirements.

[0030] As a further solution of the present invention, in the dynamic data cleaning unit, the pain score of the cancer patient at home is a high-fluctuation pain category, and the acquisition rule of the sensitive parameter of the pain score is:

[0031]

[0032] Where: γ 1 , γ 2 , γ 3 They are item, item, The impact factor of the item is set according to the requirements.

[0033] As a further solution of the present invention, the specific implementation process of the distributed computing and data processing module is as follows:

[0034] Step 21, preliminary data analysis and data filtering: Perform rapid preliminary screening based on statistical methods and rule engines, conduct real-time analysis of data that require timely feedback (acute pain changes), identify pain risk values, and filter out redundant data;

[0035] Step 22, feature extraction: according to the needs of pain management, heart rate, blood pressure, body temperature, and pain score are selected as features from multi-source data, the selected features are standardized and normalized, and the key features are extracted using the LDA (Linear Discriminant Analysis) method;

[0036] Step 23, large-scale data analysis: Utilize distributed computing framework and big data analysis tools to analyze complex patterns and trends in large-scale data through cluster analysis. Based on historical and current data, use time series analysis methods to predict future cancer patients’ pain trends and risks, and transmit the data to the intelligent decision support module.

[0037] As a further solution of the present invention, the specific implementation process of the intelligent decision support module includes:

[0038] Step 31, deep learning model analysis: select the LSTM model as the selected model based on data characteristics and analysis requirements, use historical data to train the model, optimize model parameters, and perform inference on real-time data to predict the future pain status and risk of cancer patients;

[0039] Step 32, generate personalized pain management recommendations: Based on the model prediction results and the patient's historical data, assess the pain risk and other risks faced by the patient, generate data on pain management drugs, frequency adjustments for physical pain management plans, frequency adjustments for doctor-patient communication, and adjustments for sleep.

[0040] Compared with the prior art, the present invention has the following technical effects for solving existing technical problems: the present invention effectively removes data noise and outliers and improves data quality by utilizing dynamic data cleaning algorithms and data standardization technologies; the multi-source data fusion method comprehensively processes data from different sources to ensure data integrity and consistency; through large-scale data analysis tools and distributed computing frameworks, it quickly identifies complex patterns and trends in data and extracts key health features; at the same time, it adopts an LSTM deep learning model to analyze and infer real-time data, predict future pain states and potential risks of cancer patients, and generate personalized pain management recommendations based on the patient's historical data and current analysis results; it can respond to patient needs in real time and accurately, and improve the efficiency and effectiveness of pain management in the home environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the present invention to clearly and completely describe the technical solutions in the present invention. Obviously, the described technical solutions are only part of the present invention but not all of it. Based on the content of the present invention, all other technical solutions obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0043] like Figure 1 As shown, the present invention proposes a cancer patient home pain management decision system based on patient-reported outcomes, including an intelligent perception and data aggregation module, a cloud server, a distributed computing and data processing module, an intelligent decision support module and a user interaction feedback module. The intelligent perception and data aggregation module is connected to the distributed computing and data processing module through a home network and a secure communication protocol, and directly communicates with the cloud server. The processing results of the distributed computing and data processing module are transmitted to the intelligent decision support module. The results generated by the intelligent decision support module are displayed to the patient's medical professionals through the user interaction feedback module. The user feedback data collected by the user interaction feedback module is transmitted back to the intelligent perception and data aggregation module and the distributed computing and data processing module. The intelligent perception and data aggregation module uses a dynamic data cleaning algorithm Remove noise and outliers from the data, standardize the data format, and perform multi-source data fusion before transmitting the data to the cloud server. The distributed computing and data processing module performs preliminary data analysis and data filtering on the edge device, responds in real time to data that requires timely feedback, transmits the processed data to the cloud server, and stores the data in a distributed database system. Feature extraction algorithms are used to extract key features, and a distributed computing framework is applied to perform large-scale data analysis to identify complex patterns and trends. The acquired features and preliminary analysis reports are transmitted to the intelligent support module. The intelligent support module analyzes the current data through a pre-trained deep learning model, predicts the future pain status of cancer patients, and automatically generates personalized pain management recommendations for cancer patients at home based on the patient's historical data and prediction results.

[0044] The present invention utilizes dynamic data cleaning algorithms and data standardization technology to effectively remove data noise and outliers and improve data quality. The multi-source data fusion method comprehensively processes data from different sources to ensure data integrity and consistency. Through large-scale data analysis tools and distributed computing frameworks, it quickly identifies complex patterns and trends in the data and extracts key health features. At the same time, the LSTM deep learning model is used to analyze and infer real-time data to predict the future pain status and potential risks of cancer patients. Combined with the patient's historical data and current analysis results, personalized pain management recommendations are generated, which can respond to patient needs in real time and accurately, and improve the efficiency and effectiveness of pain management in the home environment.

[0045] It should be noted that the intelligent perception and data aggregation module includes a data acquisition unit, a dynamic data cleaning unit and a multi-source data aggregation unit. The data acquisition units include smart phones, tablet computers, home monitoring terminals, wearable biosensors and smart medical devices. The collected data include pain score, emotional state score, nausea index, fatigue index, headache index, sleep quality score, daily activity ability score, heart rate, blood pressure, body temperature, blood oxygen level, blood sugar level, indoor PM2.5 concentration, and indoor carbon dioxide level.

[0046] The intelligent sensing and data aggregation module integrates multiple devices (smartphones, tablets, home monitoring terminals, wearable biosensors, and smart medical devices) through the data acquisition unit, and can collect rich physiological and environmental data, including pain scores, emotional state scores, nausea index, fatigue index, headache index, sleep quality scores, daily activity ability scores, heart rate, blood pressure, body temperature, blood oxygen level, blood sugar level, indoor PM2.5 concentration, and indoor carbon dioxide level. Through comprehensive data collection, the system can monitor the patient's health status and home environment in real time, and provide a comprehensive understanding of the patient's pain and related factors. Pain scores and emotional state scores help assess the patient's pain level and psychological state, physiological indicators such as heart rate, blood pressure, and body temperature can monitor changes in vital signs, and environmental data (indoor air quality) can help identify external factors that affect the patient's comfort and health. Such comprehensive data analysis can not only provide medical professionals with scientific decision-making basis, but also develop personalized pain management plans for patients, significantly improving the effectiveness and safety of pain management for cancer patients at home.

[0047] It should be noted that the specific implementation process of the multi-source data aggregation unit includes:

[0048] Step 11, data collection: the data collection unit collects data from smartphones, tablet computers, home monitoring terminals, wearable biosensors, and smart medical devices;

[0049] Step 12, data standardization and format conversion: unify the data formats and units of different data sources, convert the data collected by different devices and sensors into a standard preset format, unify the units of data collected by different devices, and synchronize the data collected by different devices according to a unified time standard;

[0050] Step 13, dynamic data cleaning: Dynamic data cleaning is performed on the pain score data of cancer patients at home based on a preset dynamic data cleaning strategy, and then the cleaned data and other data are transmitted to the multi-source data fusion unit through the home network and secure communication protocol;

[0051] Step 14, multi-source data fusion: align all data points in chronological order, assign weights to each data source according to its reliability and importance, and use the Bayesian network method to fuse the data. For different types of data, multi-dimensional splicing and fusion are performed to form a comprehensive pain management data set for cancer patients at home;

[0052] Step 15, data storage and transmission: Store the cleaned and integrated data in the distributed database system, and transmit the data to the cloud server and distributed computing and data processing module through a secure communication protocol.

[0053] The implementation process of the above-mentioned multi-source data aggregation unit effectively solves the problems of uneven data quality, large data volume and diversity in home pain management for cancer patients. Through data collection, standardization and format conversion, the consistency and comparability of data from different devices and sensors are ensured, and the unified data format and unit make data processing more efficient. The dynamic data cleaning step removes noise and outliers, and improves the reliability of the data. Multi-source data fusion uses Bayesian network method and multi-dimensional splicing to integrate information from different data sources to form a comprehensive patient health status data set, which helps to identify complex patterns and trends. The data storage and transmission steps ensure the security and availability of the data, and provide a solid data foundation for subsequent analysis and personalized pain management. The overall process improves the accuracy and timeliness of data processing, enabling the system to respond to patient needs in real time and accurately, and improve the effectiveness of home pain management.

[0054] It should be noted that in the dynamic data cleaning unit, the dynamic interquartile range method is used to detect and process outliers in the pain score data of cancer patients at home. The specific process includes:

[0055] Step 1, data preparation: collect and organize the pain score dataset of cancer patients at home, the dataset is S, the number of data points in the dataset S is n, sort the data in the pain score dataset S from small to large, and obtain the sorted dataset S sorted ;

[0056] Step 2, calculate the quartiles: take 25% of the data points in the sorted data set as the first quartile, take 75% of the data points in the sorted data set as the third quartile, take the middle data point of the data set as the median, the total number of data points n is an odd number, and the first quartile is equal to S sorted Middle The value of the data point, the third quartile is S sorted Middle The value of the data point, the total number of data points n is an even number, and the first quartile is equal to S sorted Middle The value of the data point is The mean of the values ​​of the data points, the third quartile is equal to S sorted Middle The value of the data point is The mean of the values ​​of the data points;

[0057] Step 3, calculate the interquartile range: the interquartile range is equal to the third quartile minus the first quartile;

[0058] Step 4: Determine the range of outliers: The lower bound of the pain score of cancer patients at home is max(0,Q 1 -ξ×k IQR ), the upper bound is min(10,Q 3 +ξ×k IQR ), where Q 1 , Q 3 are the first and third quartiles respectively, k IQR is the interquartile range, ξ is the sensitivity parameter of the pain score, which is set according to the actual situation;

[0059] Step 5, detect and process outliers: For each data point, if the data point is less than the lower limit or greater than the upper limit, it is marked as an outlier. The authenticity of the outlier is further verified by checking with the patient. For confirmed outliers, the mean of the adjacent data points is used for correction. Unconfirmed outliers are removed and the median of the adjacent data points is interpolated for correction.

[0060] The application of the above-mentioned dynamic interquartile range method in the dynamic data cleaning unit effectively solves the problem of outlier detection and processing in the pain score data of cancer patients at home. The specific process uses strict statistical methods to calculate quartiles and interquartile ranges, clarify the range of outliers, and accurately identify abnormal fluctuations in the data. Compared with the prior art, this method has several significant advantages. First, through the dynamically adjusted sensitivity parameters, the system can flexibly set the detection threshold according to the actual situation, adapt to the individual differences and data characteristics of different patients, and thus improve the accuracy of outlier detection. Secondly, the strategy of segmented processing of outliers is adopted. By verifying with the patient and using the mean or median of the adjacent data points to correct the data, the accuracy and rationality of the data correction are guaranteed, and the data deviation caused by outliers is avoided. Third, this method can handle extreme values ​​and noise in the data set, ensure the overall quality and reliability of the data, and provide reliable basic data support for subsequent data analysis and personalized pain management recommendations. Finally, this refined dynamic data cleaning strategy improves the real-time and accuracy of data processing, enables the system to better respond to changes in the patient's health status, provide more personalized and accurate pain management solutions, and significantly improve the patient's home care effect.

[0061] It should be noted that the dynamic data cleaning unit obtains the pain score data of cancer patients at home, obtains the home pain score personality characteristic value through the preset pain score personality criterion number formula, obtains the home pain score characteristic categories of cancer patients based on the home pain score personality characteristic value according to the preset threshold range, and obtains the sensitivity parameters of the cancer patients' home pain score according to the rules set for each category, and then detects and processes the patients' pain scores for outliers through the cancer patients' home pain score interquartile range method.

[0062] By obtaining the pain score data of cancer patients at home through the dynamic data cleaning unit and using the preset pain score personality criterion number formula, the system can accurately calculate the personality characteristic values ​​of each patient. These personality characteristic values ​​help to divide patients into different pain score characteristic categories, and set corresponding sensitivity parameters according to the rules of each category. In this way, the system can dynamically adjust the sensitivity of outlier detection according to the specific situation of each patient, identify and process abnormal data more accurately, and reflect the patient's pain experience more personalized and accurately, avoiding false positives or omissions of abnormal values, ensuring data quality, and providing a scientific basis for subsequent personalized pain management plans. Ultimately, this method helps to provide more refined and personalized home care recommendations, improving the quality of life and care effectiveness of patients.

[0063] It should be noted that in the dynamic data cleaning unit, the preset pain score individual criterion number formula nonlinearly combines the volatility, discreteness and data trend of the pain score data in the patient's active report for the past month to obtain the pain score individual criterion number of home-based cancer patients. The preset pain score individual criterion number formula is:

[0064]

[0065] Where: C is the number of personality criteria for pain scores of cancer patients at home, k MAD is the median absolute deviation of the pain score data in the past month, μ recent is the average pain score in the past month, σ recent is the standard deviation of pain scores in the past month, k trend The data trend of pain scores in the past month was calculated by the moving average method.

[0066] The preset formula for the individual criteria number of pain scores can fully reflect the individual differences in patients' pain experience by nonlinearly combining the volatility, median absolute deviation, discreteness, and data trends of the patients' pain score data over the past month. This formula enables the system to quantify the patient's pain characteristics and accurately obtain the individual criteria number of cancer patients at home, providing a scientific basis for personalized pain management. By carefully analyzing volatility, discreteness, and trends, the system can more accurately classify the patient's pain type and dynamically adjust the sensitivity parameters of outlier detection to ensure the accuracy of data processing and the effectiveness of personalized management, which helps to provide more accurate and personalized pain management solutions and improve patients' care experience and quality of life.

[0067] It should be noted that in the dynamic data cleaning unit, based on the set classification threshold, the pain scores of cancer patients at home are divided into low-fluctuation pain, medium-fluctuation pain and high-fluctuation pain according to the values ​​of the pain score personality criteria. The classification criteria for low-fluctuation pain, medium-fluctuation pain and high-fluctuation pain are C≤C L , C L <C≤C M 、C>C M , where C L and C M They are the set low category threshold and medium category threshold respectively.

[0068] In the dynamic data cleaning unit, the pain scores of cancer patients at home are divided into low-volatility, medium-volatility, and high-volatility pain categories according to the personality criterion number C through the set classification threshold. This classification method can accurately divide the patient's pain type according to their pain characteristics, and can provide more accurate management strategies and outlier detection methods for patients with different fluctuation types. For example, patients with low-volatility pain may need higher sensitivity to detect small changes, while patients with high-volatility pain need to pay attention to abnormal situations with large fluctuations. Classification standard C L and C M The settings enable the system to personalize the management plan to ensure that each patient receives the most appropriate pain management advice, thereby improving the effectiveness of home care and the patient's quality of life.

[0069] It should be noted that in the dynamic data cleaning unit, the pain score of cancer patients at home is low-fluctuation pain, and the rules for obtaining sensitive parameters of the pain score are:

[0070]

[0071] Where: 0 is the initial pain score sensitivity parameter, which is 1.5, α 1 It is the index adjustment parameter, which can be set according to the requirements.

[0072] Set the initial pain score sensitivity parameter ξ 0 The value is 1.5, which is a sensitivity multiple commonly used in the interquartile range method. This value is often used in statistics to set the threshold for outlier detection, which can effectively identify significant deviations in the data. 1 Further sensitivity adjustment enables the system to dynamically adjust the monitoring threshold according to individual patient characteristics, especially in the low-fluctuation pain category, the exponential function The impact of small changes is amplified, ensuring that even subtle changes in pain scores can be identified and addressed. The technical effect of this is to increase the sensitivity and accuracy of the system, enabling it to better detect subtle pain changes in patients with low fluctuations, providing a more precise basis for personalized pain management.

[0073] It should be noted that in the dynamic data cleaning unit, the pain score of cancer patients at home is classified as moderately volatile pain, and the rules for obtaining sensitive parameters of the pain score are as follows:

[0074]

[0075] Where: β 1 , β 2 They are Item and The impact factor of the item is set according to the requirements.

[0076] In the dynamic data cleaning unit, the scoring data of patients with moderate fluctuation pain is analyzed using the formula

[0077] To obtain the sensitive parameter ξ. The technical effect of this is that the system's sensitivity to data fluctuations and discreteness can be finely adjusted by combining the logarithmic function and the hyperbolic tangent function. The logarithmic function is sensitive to smaller changes and is suitable for processing moderately fluctuating data, while the hyperbolic tangent function limits the output range, making the system's response to extreme discreteness smoother. By setting an appropriate influencing factor β 1 and β 2 , the system is able to balance the impact of volatility and discreteness on detection sensitivity and accurately identify abnormal data with moderate fluctuations. This approach ensures that the system neither misses important outliers nor over-responds to normal data, thereby providing more accurate pain management support.

[0078] It should be noted that in the dynamic data cleaning unit, the pain score of cancer patients at home is a high-volatility pain category, and the rules for obtaining sensitive parameters of the pain score are:

[0079]

[0080] Where: γ 1 , γ 2, γ 3 They are item, item, The impact factor of the item is set according to the requirements.

[0081] In the dynamic data cleaning unit, for patients with high fluctuation pain, the formula To determine the sensitive parameter ξ. The advantage of this is that the hyperbolic cosine function The sensitivity to trend changes is amplified, allowing the system to pay more attention to abnormal changes in long-term trends. In addition, by combining the weighted combination of multi-dimensional features such as median absolute deviation, standard deviation and trend, the system can accurately capture complex anomalies in highly volatile data. Set an appropriate impact factor γ 1 , γ 2 , γ 3 , can effectively balance the impact of different features on detection sensitivity, thereby improving the system's responsiveness and anomaly detection accuracy in high-volatility environments. This approach helps provide more refined pain management and intervention measures to ensure the quality of home care for patients.

[0082] It should be noted that the specific implementation process of the distributed computing and data processing module is as follows:

[0083] Step 21, preliminary data analysis and data filtering: Perform rapid preliminary screening based on statistical methods and rule engines, conduct real-time analysis of data that require timely feedback (acute pain changes), identify pain risk values, and filter out redundant data;

[0084] Step 22, feature extraction: according to the needs of pain management, heart rate, blood pressure, body temperature, and pain score are selected as features from multi-source data, the selected features are standardized and normalized, and the key features are extracted using the LDA (Linear Discriminant Analysis) method;

[0085] Step 23, large-scale data analysis: Utilize distributed computing framework and big data analysis tools to analyze complex patterns and trends in large-scale data through cluster analysis. Based on historical and current data, use time series analysis methods to predict future cancer patients’ pain trends and risks, and transmit the data to the intelligent decision support module.

[0086] The process of the above-mentioned distributed computing and data processing modules can significantly improve data processing efficiency and analysis accuracy. Through preliminary data analysis and data filtering, the system can identify acute pain risks in real time and remove redundant data, thereby improving data quality and processing speed. In the feature extraction stage, key physiological indicators such as heart rate and blood pressure are selected, and standardization and normalization are performed to ensure data consistency and comparability; the LDA method is used to extract key features to enhance classification and prediction capabilities. Finally, through large-scale data analysis through a distributed computing framework, the system can deeply explore complex patterns and trends, combine historical and real-time data, and accurately predict future pain trends and risks. This comprehensive data processing process provides a high-quality data foundation for the intelligent decision support module, assists in the formulation of personalized pain management plans, and improves patients' quality of care and life experience.

[0087] It should be noted that the specific implementation process of the smart decision support module includes:

[0088] Step 31, deep learning model analysis: select the LSTM model as the selected model based on data characteristics and analysis requirements, use historical data to train the model, optimize model parameters, and perform inference on real-time data to predict the future pain status and risk of cancer patients;

[0089] Step 32, generate personalized pain management recommendations: Based on the model prediction results and the patient's historical data, assess the pain risk and other risks faced by the patient, generate data on pain management drugs, frequency adjustments for physical pain management plans, frequency adjustments for doctor-patient communication, and adjustments for sleep.

[0090] The process of the intelligent decision support module uses the deep learning model LSTM to analyze the patient's historical and real-time data to accurately predict the future pain status and risks of cancer patients. The LSTM model can effectively capture the complex dependencies in time series data and improve the accuracy and sensitivity of predictions. Based on these prediction results, the system can provide each patient with personalized pain management recommendations, including appropriate drug dosage, adjustment of physical therapy frequency, optimization of doctor-patient communication frequency, and sleep management. This personalized advice can not only help patients manage pain more effectively and improve their quality of life, but also provide scientific decision-making support for medical professionals. Through this intelligent and personalized management model, the system significantly improves the effectiveness of patients' home care.

[0091] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0092] Finally: The above description is only the technical solution of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A cancer patient home pain management decision system based on patient-reported outcomes, characterized in that: It includes an intelligent perception and data aggregation module, a cloud server, a distributed computing and data processing module, an intelligent decision support module and a user interaction feedback module. The intelligent perception and data aggregation module is connected to the distributed computing and data processing module through a home network and a secure communication protocol, and directly communicates with the cloud server. The processing results of the distributed computing and data processing module are transmitted to the intelligent decision support module. The results generated by the intelligent decision support module are displayed to the patient's medical professionals through the user interaction feedback module. The user feedback data collected by the user interaction feedback module is transmitted back to the intelligent perception and data aggregation module and the distributed computing and data processing module. The intelligent perception and data aggregation module uses a dynamic data cleaning algorithm to remove noise and outliers in the data and standardize the data. The distributed computing and data processing module performs preliminary data analysis and data filtering on the edge device, responds to data that requires timely feedback in real time, transmits the processed data to the cloud server, and stores the data in the distributed database system. The feature extraction algorithm is used to extract key features, and the distributed computing framework is applied to perform large-scale data analysis to identify complex patterns and trends. The acquired features and preliminary analysis reports are transmitted to the intelligent support module. The intelligent support module analyzes the current data through a pre-trained deep learning model, predicts the future pain status of cancer patients, and automatically generates personalized pain management recommendations for cancer patients at home based on the patient's historical data and prediction results.

2. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 1, characterized in that: The intelligent perception and data aggregation module includes a data acquisition unit, a dynamic data cleaning unit and a multi-source data aggregation unit. The data acquisition units include smartphones, tablets, home monitoring terminals, wearable biosensors and smart medical devices. The collected data include pain score, emotional state score, nausea index, fatigue index, headache index, sleep quality score, daily activity ability score, heart rate, blood pressure, body temperature, blood oxygen level, blood sugar level, indoor PM2.5 concentration, and indoor carbon dioxide level.

3. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 2, characterized in that: In the dynamic data cleaning unit, the dynamic interquartile range method is used to detect and process outliers in the pain score data of cancer patients at home. The specific process includes: Step 1, data preparation: collect and organize the pain score dataset of cancer patients at home, the dataset is S, the number of data points in the dataset S is n, sort the data in the pain score dataset S from small to large, and obtain the sorted dataset S sorted ; Step 2, calculate the quartiles: take 25% of the data points in the sorted data set as the first quartile, take 75% of the data points in the sorted data set as the third quartile, take the middle data point of the data set as the median, the total number of data points n is an odd number, and the first quartile is equal to S sorted Middle The value of the data point, the third quartile is S sorted Middle The value of the data point, the total number of data points n is an even number, and the first quartile is equal to S sorted Middle The value of the data point is The mean of the values ​​of the data points, the third quartile is equal to S sorted Middle The value of the data point is The mean of the values ​​of the data points; Step 3, calculate the interquartile range: the interquartile range is equal to the third quartile minus the first quartile; Step 4: Determine the range of outliers: The lower bound of the pain score of cancer patients at home is max(0,Q1-ξ×k IQR ), the upper bound is min(10,Q3+ξ×k IQR ), where Q1 and Q3 are the first and third quartiles respectively, k IQR is the interquartile range, ξ is the sensitivity parameter of the pain score, which is set according to the actual situation; Step 5, detect and process outliers: For each data point, if the data point is less than the lower limit or greater than the upper limit, it is marked as an outlier. The authenticity of the outlier is further verified by checking with the patient. For confirmed outliers, the mean of the adjacent data points is used for correction. Unconfirmed outliers are removed and the median of the adjacent data points is interpolated for correction.

4. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 3, characterized in that: The dynamic data cleaning unit obtains the pain score data of cancer patients at home, obtains the home pain score personality characteristic value through the preset pain score personality criterion number formula, obtains the home pain score characteristic category of cancer patients based on the home pain score personality characteristic value according to the preset threshold range, and obtains the sensitivity parameters of the cancer patients' home pain score according to the rules set for each category, and then detects and processes the patients' pain scores for abnormal values ​​through the cancer patients' home pain score interquartile range method.

5. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 4, characterized in that: In the dynamic data cleaning unit, the preset pain score individual criterion number formula nonlinearly combines the volatility, discreteness and data trend of the pain score data in the patient's active report for the past month to obtain the pain score individual criterion number of home-based cancer patients. The preset pain score individual criterion number formula is: Where: C is the number of personality criteria for pain scores of cancer patients at home, k MAD is the median absolute deviation of the pain score data in the past month, μ recent is the average pain score in the past month, σ recent is the standard deviation of pain scores in the past month, k trend The data trend of pain scores in the past month was calculated by the moving average method.

6. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 5, characterized in that: In the dynamic data cleaning unit, based on the set classification threshold, the pain scores of cancer patients at home are divided into low-fluctuation pain, medium-fluctuation pain and high-fluctuation pain according to the values ​​of the pain score personality criteria. The classification criteria for low-fluctuation pain, medium-fluctuation pain and high-fluctuation pain are C≤C L , C L <C≤C M 、C>C M , where C L and C M They are the set low category threshold and medium category threshold respectively.

7. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 6, characterized in that: In the dynamic data cleaning unit, the pain score of cancer patients at home is low-fluctuation pain, and the rules for obtaining sensitive parameters of the pain score are: Where: ξ0 is the initial pain score sensitivity parameter, which is 1.5; α1 is the index adjustment parameter, which is set according to demand.

8. A cancer patient home pain management decision system based on patient-reported outcomes according to claim 7, characterized in that: In the dynamic data cleaning unit, the pain score of cancer patients at home is classified as moderately volatile pain. The rules for obtaining sensitive parameters of the pain score are as follows: Where: β1 and β2 are Item and The impact factor of the item is set according to the requirements.

9. The cancer patient home pain management decision system based on patient-reported outcomes according to claim 7, characterized in that: In the dynamic data cleaning unit, the pain score of cancer patients at home is classified as high-volatility pain. The rules for obtaining sensitive parameters of the pain score are as follows: Where: γ1, γ2, γ3 are item, item, The impact factor of the item is set according to the requirements.

Images