Medical record information treatment effect display system

Abstract

The present application relates to the field of medical data management and clinical auxiliary technology, and in particular to a medical record information treatment effect display system to solve the problems of data fragmentation, static and single treatment effect display, lack of personalized adaptation, and insufficient evaluation accuracy in existing systems. The system includes five modules: data acquisition, processing, treatment effect modeling, display adaptation, and interactive control. The data acquisition module integrates multi-source data and standardizes it. The data processing module preprocesses the data and filters the treatment effect related indicators, while synchronously encrypting and tracing the data. The treatment effect modeling module dynamically models and generates trend curves, combines individual correction factors to output accurate results and predict treatment effects. The display adaptation module provides customized interfaces according to user roles and clinical scenarios. The interactive control module supports data backtracking, multi-case comparison, and three-level early warning linkage. This system realizes multi-source data fusion and dynamic accurate evaluation, improves the quality and efficiency of diagnosis and treatment, and ensures data security and compliance.

Description

Technical Field

[0001] This invention relates to the field of medical data management and clinical support technology, specifically a medical record information treatment effect display system. Background Technology

[0002] In the healthcare system, medical records are the core carrier of information recording the entire process of patient diagnosis and treatment. They are not only important evidence for medical quality assessment and clinical research, but also crucial support for doctors to judge treatment effectiveness and adjust treatment plans. With the rapid development of medical informatization, electronic medical record systems have been widely used in medical institutions at all levels, gradually replacing traditional paper medical records and realizing the digital storage and initial sharing of medical record information. However, existing electronic medical record systems and related efficacy display technologies still have many shortcomings, making it difficult to meet the actual needs of precise clinical diagnosis and treatment and refined medical management. Specific problems are as follows: First, existing systems suffer from severe data fragmentation and limited dimensions for efficacy evaluation. Most current electronic medical record systems focus solely on storing structured data such as electronic medical records and laboratory reports, failing to effectively integrate multi-source data including real-time physiological monitoring data, unstructured information in medical imaging reports, outpatient follow-up records, and patients' subjective symptom descriptions. Regarding efficacy display, it primarily presents isolated laboratory indicator values, lacking comprehensive analysis of indicator trends, multi-indicator correlations, and individual patient differences. This makes it difficult for doctors to fully assess the actual effectiveness of treatment plans. For example, for patients with chronic diseases, existing systems only display a single blood glucose or blood pressure measurement, failing to consider the patient's diet, exercise records, and long-term trend changes to assess the suitability of the treatment plan, easily leading to biased diagnostic and treatment decisions.

[0003] Secondly, the efficacy display lacks dynamism and personalized adaptability. Existing technologies mostly adopt a static display method, presenting efficacy-related data only at specific points in time. They cannot build dynamic efficacy models based on time-series data, making it difficult to predict subsequent treatment effect trends and providing support for advance adjustments to treatment plans. Furthermore, the system interfaces are mostly generic designs, failing to customize for the needs of different user roles. Clinicians require complete efficacy indicators, related medical record data, and trend analysis to aid decision-making; patients need concise and easy-to-understand efficacy results and health guidance; and healthcare administrators need group efficacy statistics to control medical quality. Existing systems cannot simultaneously meet these diverse needs, resulting in low data utilization efficiency.

[0004] Furthermore, there is a lack of effective clinical interaction and early warning linkage mechanisms. Existing systems mostly display efficacy data in a one-way manner, with limited interactive functions, making it difficult to perform commonly used clinical operations such as comparing efficacy in multiple cases and tracing data back through the treatment phase, thus increasing the workload of physicians. At the same time, when efficacy indicators show abnormal fluctuations, the system often only issues simple prompts, failing to link with related medical record data (such as past medical history and medication records) for auxiliary analysis, and lacking a tiered early warning mechanism, making it difficult to help physicians quickly locate the cause of the abnormality, potentially delaying timely diagnosis and treatment. In addition, some systems have insufficient data security and inadequate data traceability capabilities, failing to meet the requirements for compliant medical data management.

[0005] Finally, existing technologies have shortcomings in the accuracy of efficacy assessment. Most systems use uniform efficacy assessment standards, failing to fully consider the impact of individual factors such as patient age, underlying diseases, physical characteristics, and lifestyle habits on treatment outcomes. This leads to biased efficacy assessment results and cannot provide precise support for personalized treatment. Furthermore, existing systems have weak visualization capabilities, struggling to transform complex, multi-source data into intuitive charts and graphs. This hinders doctors from quickly grasping key efficacy information and patients from understanding their treatment progress, impacting treatment adherence.

[0006] In summary, existing medical record information systems have significant shortcomings in data integration, dynamic modeling, personalized adaptation, clinical interaction, and accurate assessment. They fail to fully leverage the supporting role of medical record information in evaluating treatment effectiveness, thus hindering the improvement of clinical diagnosis and treatment quality and medical management efficiency. Therefore, developing a medical record information treatment effect display system capable of multi-source data fusion, dynamic and accurate assessment, personalized display adaptation, and efficient clinical interaction has become an urgent need in the field of medical informatics. Summary of the Invention

[0007] To address the problems of fragmented multi-source data, static and singular efficacy display, lack of personalized adaptation, weak clinical interaction capabilities, and insufficient assessment accuracy in existing medical record systems, this paper provides a medical record information treatment effect display system. This system achieves deep integration of medical record information and efficacy data, dynamically, accurately, and intuitively presenting treatment effects. At the same time, it adapts to the needs of different user roles, providing comprehensive support for clinical diagnosis and treatment, patient management, and medical quality control, thereby improving the accuracy and efficiency of medical services.

[0008] The technical solution adopted by this invention to solve its technical problem is: a medical record information treatment effect display system, comprising: The data acquisition module is used to acquire patients' electronic medical records, real-time physiological monitoring data, medical imaging reports, laboratory test results, outpatient follow-up records, and diagnosis and treatment operation records. It uses natural language processing semantic parsing technology to extract key diagnosis and treatment information from unstructured data, integrates multi-source data according to the diagnosis and treatment timeline through a time-series alignment algorithm, and outputs a standardized medical record dataset containing structured, unstructured, and real-time data. The built-in data verification unit ensures data validity. The data processing module performs cleaning, deduplication, and format unification preprocessing on standardized medical record datasets. It uses a feature extraction model to screen indicators that are strongly correlated with treatment effects, constructs a efficacy evaluation dataset that integrates objective indicators and subjective symptoms, and simultaneously uses encryption technology to achieve full-process security control of medical record information and establishes a data traceability mechanism to ensure compliance. The efficacy modeling module dynamically models the efficacy assessment dataset based on time series analysis, generates a treatment effect trend curve, constructs individual correction factors by combining patient age, underlying diseases, and physical characteristics, and outputs accurate efficacy assessment results by calibrating the trend curve. The display adaptation module provides a customized interface based on the user roles (clinical doctors, patients, and medical administrators) and preset display rules: the doctor's side displays complete efficacy indicators, trend changes, and abnormal warning signs; the patient's side presents simplified efficacy results and health guidance; and the management side displays group efficacy statistics and treatment quality analysis. The interactive control module supports retrospective analysis of treatment data by stage of diagnosis and treatment, comparison of treatment effects in multiple cases, triggering an early warning signal and linking related medical record data when treatment indicators fluctuate abnormally, and providing data export and printing functions. All modules achieve bidirectional communication through a data bus to ensure real-time data transmission.

[0009] Specifically, the data acquisition module also includes an off-site device interface, which can access real-time data from smart wearable devices and home health monitoring instruments. Data transmission uses Bluetooth and 5G dual-mode adaptation to ensure real-time uploading of off-site data.

[0010] Specifically, the individual correction factor in the efficacy modeling module is also dynamically updated in conjunction with the patient's medication adherence and lifestyle adjustments, with the update cycle synchronized with the treatment follow-up cycle.

[0011] Specifically, the display adaptation module also supports scene switching function. In clinical emergency scenarios, it automatically prioritizes displaying critical values ​​associated with efficacy indicators, while in outpatient follow-up scenarios, it focuses on displaying efficacy change trends and treatment plan adjustment suggestions.

[0012] Specifically, the multi-case comparison function of the interactive control module can group and filter patients according to disease type, treatment plan, and patient physical characteristics, and generate comparative analysis reports.

[0013] Specifically, the feature extraction model of the data processing module uses a gradient boosting algorithm to optimize the accuracy of indicator screening and eliminate redundant data and interfering indicators.

[0014] Specifically, the efficacy modeling module also has an efficacy prediction function, which predicts the trend of subsequent treatment effects based on existing diagnosis and treatment data and similar historical cases, providing a reference for optimizing the diagnosis and treatment plan.

[0015] Specifically, the interface of the display adaptation module supports custom configuration. Users can add commonly used efficacy indicators to the homepage display and adjust the display order and visualization format of the indicators.

[0016] Specifically, the warning signals of the interactive control module are divided into three levels, corresponding to mild, moderate and severe abnormalities in efficacy indicators, respectively. Different levels of warning correspond to different linkage response mechanisms.

[0017] Specifically, the data processing module also has a data traceability function, which marks the source, processing time and processing process of each efficacy indicator data to ensure data traceability.

[0018] The beneficial effects of this invention are: (1) The medical record information treatment effect display system of the present invention integrates structured, unstructured, real-time and out-of-hospital multi-source data, breaks down the data fragmentation barrier, and combines semantic parsing and time sequence alignment technology to realize the integrated integration of medical record information and efficacy data, providing more comprehensive data support for efficacy evaluation and avoiding the one-sided evaluation caused by single data.

[0019] (2) The medical record information treatment effect display system of the present invention constructs a dynamic efficacy model based on time series data, combines individual correction factor calibration evaluation results, fully considers the impact of individual patient differences on treatment effect, and has efficacy prediction function, which can provide accurate basis for early adjustment of treatment plan, which is superior to the existing static and unified evaluation method.

[0020] (3) The medical record information treatment effect display system described in this invention provides customized display interfaces for different user roles and clinical scenarios, meets the differentiated needs of doctors' diagnosis and treatment decisions, patients' health cognition, and managers' quality control, avoids data redundancy and information omission caused by a general interface, and improves the work efficiency of each role.

[0021] (4) The medical record information treatment effect display system described in this invention reduces the workload of doctors through rich interactive functions (multi-case comparison, data backtracking), and the three-level early warning mechanism and related data linkage analysis help doctors quickly locate the cause of abnormal efficacy, avoid delaying the diagnosis and treatment time, and improve the safety of clinical diagnosis and treatment.

[0022] (5) The medical record information treatment effect display system of the present invention adopts encryption technology and data traceability mechanism to realize the secure storage, transmission and operation traceability of medical record data, meet the requirements of medical data compliance management and avoid data security risks. Attached Figure Description

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] Figure 1 This invention provides an architecture diagram of a medical record information treatment effect display system. Detailed Implementation

[0025] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0026] like Figure 1 As shown, the medical record information treatment effect display system of the present invention includes a data acquisition module, a data processing module, an efficacy modeling module, a display adaptation module, and an interactive control module. Each module achieves bidirectional communication through a data bus to ensure the real-time performance and stability of data transmission. The specific structure and functions are as follows: Data Acquisition Module: This module integrates multiple interface adapters, enabling simultaneous access to medical institution electronic medical record systems, LIS laboratory systems, PACS imaging systems, physiological monitoring equipment, outpatient smart wearable devices, and follow-up management platforms, achieving comprehensive acquisition of multi-source data. For structured data such as electronic medical records and test results, standardized interfaces are used to directly read and convert them into a unified format. For unstructured data such as imaging reports and medical records, natural language processing semantic parsing technology is employed to extract key diagnostic and treatment information (such as symptom descriptions, imaging conclusions, and medication adjustment suggestions). For real-time physiological data (such as heart rate, blood pressure, and blood glucose) and outpatient follow-up data, a time-series alignment algorithm is used to integrate the data according to the treatment timeline, eliminating time bias and ultimately outputting a standardized medical record dataset containing structured, unstructured, and real-time data. Simultaneously, the module has a built-in data verification unit to perform preliminary integrity checks on the collected data, automatically removing invalid data to ensure data validity.

[0027] Data Processing Module: This module performs preprocessing on standardized medical record datasets, including cleaning (removing null and outlier values), deduplication (eliminating duplicate records), and format standardization (converting to a system-compatible format) to ensure data quality. It utilizes feature extraction models to select indicators strongly correlated with treatment efficacy (such as disease-specific indicators, physiological function indicators, and symptom improvement indicators) to construct a efficacy evaluation dataset that integrates objective indicators and subjective symptoms. Individual information such as patient age, underlying diseases, physical characteristics, and lifestyle habits are also incorporated as auxiliary variables. The module employs encryption technology for end-to-end security control of medical record data, including storage and transmission encryption. A data traceability log is established, marking the source, processing time, processing procedure, and operator for each data entry to ensure data security and compliance. Furthermore, the module includes a data update unit that can update the efficacy evaluation dataset in real time based on newly collected diagnostic and treatment data.

[0028] The efficacy modeling module, based on the efficacy assessment dataset, employs time-series analysis methods (such as time-series neural networks) to construct a dynamic efficacy assessment model. It continuously tracks and analyzes efficacy indicators, generating trend curves of treatment effects (such as the magnitude and rate of change of indicators). Individual correction factors are constructed by incorporating patient age, underlying diseases, and physical characteristics. Multiple regression analysis is used to calibrate the trend curves, eliminating the influence of individual differences on efficacy assessment and outputting accurate efficacy assessment results, including the current efficacy level, analysis of the causes of indicator changes, and predictions of future efficacy. The module includes a built-in efficacy assessment standard library covering treatment guidelines for different diseases. It can automatically match the corresponding assessment standard based on the patient's disease type, while also supporting manual adjustment of assessment parameters by physicians to adapt to specific case needs.

[0029] Display Adaptation Module: This module pre-sets multi-role display rules, providing customized interfaces for clinicians, patients, and medical administrators. The doctor's interface uses a split-column layout: the left side displays basic patient medical record information; the middle shows dynamic efficacy trend curves, multi-indicator correlation analysis charts, and abnormal warning indicators; the right side provides suggestions for adjusting treatment plans and a quick access point for historical data, supporting custom filtering and display priority settings for efficacy indicators. The patient's interface adopts a simplified design, displaying efficacy results (such as symptom improvement and comparison of normal indicator ranges), medication reminders, and health guidance in a combination of text and graphics, avoiding excessive technical jargon and improving readability. The administrator's interface focuses on group efficacy statistics, displaying comparative efficacy data, medical quality scores, and abnormal case statistics for different departments and treatment plans, supporting filtering and analysis by time, disease type, and other dimensions. Simultaneously, the module supports scene switching, automatically adapting to the display needs of different clinical scenarios such as emergency, outpatient, inpatient, and follow-up.

[0030] Interactive Control Module: This module provides rich clinical interactive functions, supporting the retrospective analysis of efficacy data by treatment stage (e.g., initial, intermediate, and recovery phases). Users can view efficacy changes and corresponding treatment operations at different stages. It supports multi-case efficacy comparisons, allowing grouping and filtering by disease type, treatment plan, patient constitution, and other characteristics, generating comparative analysis reports to assist doctors in optimizing treatment plans. The module establishes a three-level early warning mechanism. When efficacy indicators show mild, moderate, or severe abnormalities, different intensities of warning signals (audio prompts + color indicators) are issued, simultaneously linking related medical record data (past medical history, medication records, and laboratory reports) for auxiliary analysis, displaying possible causes of the abnormalities. Furthermore, the module supports exporting efficacy data (formats include PDF and Excel) and printing, meeting the needs of medical record archiving and research data organization. It also sets access control, allowing different roles to access data only within their corresponding permission range.

[0031] Example 1: Application of the outpatient hypertension patient medical record information and treatment effect display system: This embodiment applies the system of the present invention to the display and management of treatment effects for hypertension patients in community hospital outpatient clinics. Considering the characteristics of outpatient scenarios such as fast diagnosis and treatment processes, high patient mobility, and the need to focus on blood pressure control trends and medication adherence, the system deployment adopts a cloud + terminal architecture. The terminal devices include doctor workstations and patient self-service query machines, while the cloud is used for data storage and model calculation to ensure smooth system operation.

[0032] The specific configuration and operation process of each module of the system are as follows: The data acquisition module integrates the existing electronic medical record system and outpatient laboratory system interface of the community hospital, and also connects to the smart blood pressure monitor worn by patients (supporting Bluetooth transmission) and follow-up management applet to realize real-time acquisition of multi-source data. For structured data, such as patient name, age, hypertension classification, past medication history, outpatient blood pressure test results, blood routine and liver and kidney function test data, it is directly read through a standardized interface and converted into JSON format for storage. For unstructured data, such as doctor's progress notes ("patient reports dizziness symptoms have improved compared to before, medication is taken regularly every day, with occasional missed doses") and echocardiogram reports ("left ventricular hypertrophy has not progressed significantly compared to before"), a BERT-based semantic parsing model is used to extract key information (degree of symptom improvement, medication adherence, changes in cardiac structure) and convert it into structured labels. For real-time data uploaded by smart blood pressure monitors (blood pressure values ​​at 8 am and 8 pm daily) and dietary and exercise information recorded by the follow-up mini-program, a timestamp alignment algorithm is used to integrate the data into the patient's treatment timeline by date, forming a standardized medical record dataset. The data verification unit automatically removes invalid data (such as blood pressure values ​​outside the reasonable range or data without timestamps) to ensure data validity.

[0033] The data processing module preprocesses the collected standardized dataset: it supplements missing blood pressure data using mean imputation, removes outliers (such as significantly elevated blood pressure values ​​due to equipment malfunction) using clustering algorithms, and deletes duplicate test reports. A feature extraction model is built based on the gradient boosting algorithm to select indicators strongly correlated with hypertension treatment efficacy, including systolic blood pressure, diastolic blood pressure, improvement in dizziness symptoms, medication adherence, liver and kidney function indicators (AST, serum creatinine), and left ventricular hypertrophy. Individual information such as patient age (62 years), underlying diseases (5-year history of diabetes), body mass index (24.5 kg / m²), and exercise habits (30 minutes of daily walking) are also incorporated to construct a efficacy evaluation dataset. The dataset is encrypted using the AES-256 encryption algorithm, and a data traceability log records the collection time and processing personnel (outpatient nurses, doctors) for each data point to ensure data security and compliance. The module automatically updates the efficacy evaluation dataset every 24 hours, incorporating newly collected clinical data for that day.

[0034] The efficacy modeling module matches the hypertension efficacy assessment standard library (referencing the "Chinese Guidelines for the Prevention and Treatment of Hypertension 2023"). It uses a time series neural network to construct a dynamic efficacy model, tracks and analyzes the patient's systolic and diastolic blood pressure data for three consecutive months, generates a blood pressure change trend curve, and calculates the weekly blood pressure control target achievement rate (number of days with systolic blood pressure ≤130 mmHg and diastolic blood pressure ≤80 mmHg / total number of monitoring days per week). Correction factors were constructed based on individual patient information: a history of diabetes led to a 5 mmHg reduction in the target blood pressure; age resulted in a 1.1 correction coefficient for blood pressure fluctuation; and medication adherence (85%) led to an adjustment of the efficacy assessment weight to 0.95. The trend curve was calibrated using multivariate regression analysis to output accurate efficacy assessment results: the patient's current treatment effect is "good," with systolic blood pressure decreasing from 155 mmHg at the beginning of treatment to the current 132 mmHg, a decrease of 14.8%. Dizziness symptoms were significantly relieved, and liver and kidney function indicators were normal. However, due to occasional missed doses, the blood pressure fluctuation was slightly higher than the ideal range. It is predicted that if medication is taken regularly in the next month, systolic blood pressure can be stabilized below 130 mmHg. If the frequency of missed doses increases, it may lead to a rebound in blood pressure.

[0035] The display module switches to the outpatient doctor's interface, which adopts a three-column layout: the left side displays the patient's basic medical record information (name, age, hypertension classification, diabetes history, current medication regimen); the upper half of the middle area displays the dynamic trend curve of blood pressure (the horizontal axis is the date, the vertical axis is the blood pressure value, and the normal range is marked), with green lines representing systolic blood pressure and blue lines representing diastolic blood pressure. Abnormal values ​​(blood pressure rises to 148 / 92 mmHg after a missed medication dose) are marked with red dots. The lower half displays a multi-indicator correlation analysis chart (correlation curve between medication adherence and blood pressure target achievement rate) and warning indicators (mild warning: indicating that insufficient medication adherence may affect efficacy); the right side provides suggestions for adjusting the treatment plan ("It is recommended to strengthen medication reminders, which can be sent daily medication reminders through the follow-up mini-program") and a historical data retrieval portal, allowing doctors to quickly review the efficacy data in the early stages of treatment. The patient self-service inquiry machine displays the patient-side interface, showing the treatment results in large font ("Your blood pressure is well controlled, significantly lower than at the beginning of treatment, and dizziness symptoms have been relieved"), along with a simplified blood pressure trend chart (showing only the weekly average blood pressure value). Below, it displays medication reminders ("Take one amlodipine tablet on an empty stomach every morning, please do not miss a dose") and dietary recommendations ("Reduce the intake of high-salt foods, daily salt intake ≤ 5g"). It avoids technical jargon and improves patient readability.

[0036] The interactive control module allows doctors to compare multiple cases, selecting 5 hypertensive patients of the same age group with diabetes, and generating a efficacy comparison report. This report displays the blood pressure control achievement rate of different medication regimens (amlodipine monotherapy and amlodipine + valsartan combination therapy), assisting doctors in optimizing treatment plans. It also supports retrospective analysis by treatment stage, allowing doctors to view changes in patient efficacy after 1, 2, and 3 months of treatment, comparing the correlation between medication adjustments and blood pressure changes at different stages. When a patient's blood pressure rises to 150 / 95 mmHg (moderately abnormal), the system issues a moderate warning (yellow light and sound alert), simultaneously linking related data to display possible causes ("missed medication dose yesterday, and excessive intake of high-salt foods"), helping doctors quickly pinpoint the abnormal factors. The module allows doctors to export 3 months of patient efficacy data (PDF format) for archiving in outpatient medical records, while also setting access controls so that patients can only view their own efficacy data and cannot access other patients' information.

[0037] In this embodiment, the system operates stably, with a data acquisition delay of ≤30 seconds, a semantic parsing accuracy rate of ≥92%, and a 95% consistency between the efficacy evaluation results and the doctor's clinical judgment. Through this system, doctors can quickly and comprehensively grasp the patient's treatment effect and influencing factors, making adjustments to treatment plans more targeted. Patients can understand their own condition through self-service inquiries, medication adherence has increased from 85% to 93%, and the blood pressure control rate of outpatient hypertension patients has increased by 12% compared to before, significantly improving the quality and efficiency of outpatient diagnosis and treatment.

[0038] Example 2: Application of the treatment effect display system for hospitalized patients with acute pancreatitis: This embodiment applies the system of the present invention to the monitoring and management of treatment effects for hospitalized patients with acute pancreatitis in the gastroenterology department of a tertiary hospital. In view of the characteristics of hospitalized patients with rapid changes in their condition, high demand for multidisciplinary collaboration, and the need for real-time monitoring of key efficacy indicators and organ function status, the system is deployed with a local server + distributed terminal architecture. The terminals cover doctor workstations, nurse stations, ICU monitoring equipment, and multidisciplinary consultation room displays. The local server ensures data processing speed and security and supports real-time data synchronization across multiple terminals.

[0039] The specific configuration and operation flow of each module of the system are as follows: The data acquisition module integrates the hospital's electronic medical record system, inpatient laboratory system, PACS imaging system, ICU physiological monitoring instruments (heart rate, respiration, blood oxygen saturation, urine output), bedside ultrasound equipment, and gastrointestinal function monitoring instrument interface. It also connects to the nursing station's nursing record system to achieve comprehensive collection of multi-dimensional data. Structured data includes patient gender (male), age (45 years old), acute pancreatitis type (biliary-related moderate to severe), amylase level upon admission, complete blood count, electrolytes, coagulation function test results, medication regimen (sophostane infusion, cefoperazone / sulbactam for infection), and nursing records (daily gastrointestinal decompression, abdominal pain score), which are directly read and standardized through the interface. Unstructured data includes abdominal CT reports ("Pancreatic edema has decreased, and peripheral effusion has decreased"), physician's rounds records ("Patient's abdominal pain score has decreased from 8 to 4, flatus has returned, and small amounts of liquid diet are possible"), and multidisciplinary consultation opinions ("It is recommended to continue the current anti-infection regimen"). The system monitors liver and kidney function and employs an improved BERT semantic parsing model (optimized for medical terminology recognition) to extract key information (changes in pancreatic morphology, changes in abdominal pain scores, recovery of gastrointestinal function, and consultation recommendations). Real-time physiological data (hourly heart rate, blood oxygen saturation, and urine output every 4 hours) and bedside ultrasound images (with pancreatic morphology annotations) are uploaded in real-time via a 5G interface. A millisecond-level timestamp alignment algorithm is used to integrate the data into the treatment timeline by hour, forming a standardized medical record dataset. The data verification unit verifies the physiological monitoring data in real-time, and immediately issues a device fault warning when the data exceeds the device's measurement range to ensure data accuracy.

[0040] The data processing module preprocesses the standardized dataset: interpolation is used to supplement a small amount of missing urine volume data; an outlier detection algorithm (based on the 3σ principle) is used to remove abnormal heart rate data caused by equipment interference; and duplicate CT reports (uploaded multiple times at the same time point) are removed. A feature extraction model is constructed based on the gradient boosting algorithm to screen out core indicators strongly correlated with the treatment effect of acute pancreatitis, including serum amylase, lipase, abdominal pain score (VAS score), gastrointestinal decompression, urine volume, pancreatic morphology (CT / ultrasound findings), inflammatory markers (white blood cell count, C-reactive protein), and liver and kidney function (alanine aminotransferase, serum creatinine). At the same time, individual information such as patient age, weight (78 kg), past medical history (3-year history of gallstones), and smoking history (20 years, now quit) are incorporated to construct a treatment efficacy evaluation dataset. The system uses AES-256 encryption algorithm combined with blockchain technology for encrypted data storage. The data traceability log records in detail the collection equipment, processing personnel (nurses, laboratory technicians, doctors), and processing time for each piece of data, supporting full-process traceability. The module automatically updates the efficacy evaluation dataset every hour to ensure real-time reflection of changes in the patient's condition.

[0041] The efficacy modeling module matches the acute pancreatitis efficacy assessment standard library (referencing the "Chinese Guidelines for the Diagnosis and Treatment of Acute Pancreatitis 2021"). A dynamic efficacy model is constructed using a time-series neural network combined with an LSTM algorithm. This model continuously tracks and analyzes core efficacy indicators for 7 days after patient admission, generating trend curves for amylase and lipase changes (the horizontal axis represents the number of days in hospital, and the vertical axis represents the indicator value), abdominal pain score changes, and urine output changes. Correction factors are constructed based on individual patient information: a history of gallstones leading to a faster inflammation resolution is corrected with a coefficient of 0.9, and weight-related efficacy weights for medication dosage are adjusted to 1.05. The trend curves are calibrated using multivariate regression analysis to output accurate efficacy assessment results. On the third day of hospitalization, the assessment result was "improvement": serum amylase decreased from 1200 U / L at admission to 450 U / L, lipase decreased from 800 U / L to 320 U / L, abdominal pain score decreased from 8 to 4, gastrointestinal decompression decreased from 800 ml per day to 300 ml, urine output returned to normal (about 1500 ml per day), pancreatic edema was less than at admission, and inflammatory markers decreased, but C-reactive protein was still high (25 mg / L). It was predicted that if the inflammation continued to be controlled in the next two days, somatostatin infusion could be discontinued and the patient could switch to a liquid diet; if C-reactive protein increased, it might indicate a worsening infection, and the anti-infection regimen would need to be adjusted.

[0042] The display adaptation module automatically switches to a multi-role interface based on the hospitalization scenario: The doctor's interface uses a four-column layout. The left side displays the patient's basic medical record information and admission diagnosis. The middle left side displays the dynamic trend curves of core efficacy indicators, marking the normal range of each indicator and treatment nodes (such as medication adjustments and consultation times). The middle right side displays a comparison of pancreatic CT / ultrasound images (the images at admission and the current image are displayed in a split screen, marking areas of edema and exudate changes). The right side displays abnormal warning information and treatment suggestions. The nurse's station interface focuses on displaying real-time physiological data and nursing-related efficacy indicators (gastrointestinal decompression, urine output, abdominal pain score), and supports the linked input of nursing records and efficacy data. The multidisciplinary consultation room display screen uses a full-screen layout to display the patient's complete efficacy data, trend curves, and imaging data, facilitating joint analysis of the condition by doctors from gastroenterology, hepatobiliary surgery, and ICU. At the same time, the system supports emergency scenario switching. When a patient presents with shortness of breath (blood oxygen saturation 88%) in the early stages of admission, it automatically prioritizes displaying respiratory function-related indicators and abnormal warnings to assist doctors in rapid rescue.

[0043] The interactive control module supports multi-case comparison, selecting 8 patients with moderate to severe acute pancreatitis of biliary origin admitted during the same period, comparing the efficacy of different anti-infection regimens (cefotaxime vs. piperacillin-tazobactam), and displaying data such as the rate of decline of inflammatory markers and length of hospital stay, providing a basis for decision-making in multidisciplinary consultations; it also supports retrospective analysis by treatment stage (hospitalization and rescue period, stable condition period, recovery period), allowing doctors to view changes in efficacy at different stages and compare the correlation between medication adjustments, surgical interventions (such as gallbladder puncture and drainage), and efficacy indicators. A three-tiered early warning mechanism was established: On the second day of admission, when the patient's C-reactive protein level rose to 35 mg / L (moderately abnormal), the system issued a moderate warning (yellow alert sound + pop-up window), linking related data ("Cholecystoscopic drainage fluid culture shows positive E. coli"), indicating worsening infection, and the doctor promptly adjusted the anti-infection regimen (adding metronidazole); On the fourth day of admission, when the patient's urine output dropped to 300 ml / 24h (severely abnormal), the system issued a severe warning (red alert sound + flashing light), linking renal function indicators and medication records, indicating possible kidney damage, and immediately initiating an ICU consultation. The module supports exporting efficacy data to Excel format for clinical research data organization, while setting strict access control, authorizing only doctors and nurses to access patient data, and consulting doctors can only view efficacy indicators related to the consultation, ensuring data security.

[0044] In this embodiment, the system data processing latency is ≤10 seconds, the semantic parsing accuracy is ≥95%, and the consistency between the efficacy evaluation results and the multidisciplinary consultation conclusions reaches 96%. Through this system, doctors can monitor changes in the patient's condition and treatment effects in real time, respond quickly to abnormal situations, improve the efficiency of multidisciplinary collaboration by 30%, shorten the average length of hospital stay by 2 days, and reduce the incidence of complications (such as kidney damage and infection spread) from 18% to 8%, significantly improving the safety and efficacy of diagnosis and treatment for hospitalized patients with acute pancreatitis.

[0045] Example 3: Application of the Diabetes Chronic Disease Management Medical Record Information Treatment Effect Display System: This embodiment applies the system of the present invention to the management of diabetes chronic diseases in urban community health service centers. In view of the characteristics of chronic disease patients, such as long management cycle, importance of out-of-hospital data, need for long-term monitoring of blood glucose control and risk of complications, and high demand for patient self-management, the system deployment adopts a cloud + multi-terminal (doctor workstation, patient smart terminal, community health management center display screen) architecture, which supports seamless connection of in-hospital and out-of-hospital data, and the cloud provides long-term data storage and model calculation services.

[0046] The specific configuration and operation flow of each module of the system are as follows: The data acquisition module integrates the electronic medical record system of the community health service center, the outpatient laboratory system, and the interface of the glycated hemoglobin testing equipment. It also connects to the patient's out-of-hospital smart devices (blood glucose meter, blood glucose meter, smartwatch) and chronic disease management APP to achieve integrated data collection from both inside and outside the hospital. Structured data includes the patient's gender (female), age (58 years), type of diabetes (type 2 diabetes), disease duration (8 years), glycated hemoglobin (HbA1c) value, fasting blood glucose, 2-hour postprandial blood glucose, blood lipids, renal function indicators, and medication regimen (metformin + gliclazide), which is read and standardized through a standardized interface. Unstructured data includes the doctor's chronic disease follow-up records ("The patient reported no significant worsening of foot numbness symptoms, regular daily blood glucose monitoring, and good dietary control"), fundus examination reports ("No new retinal microaneurysms compared to before"), and the patient's dietary records from the APP. The log ("Daily staple food intake is about 200g, exercise 4 times a week") uses a lightweight semantic parsing model (adapted to mobile data processing) to extract key information (complication progression, blood glucose monitoring adherence, diet and exercise status); real-time data uploaded from out-of-hospital smart devices (daily fasting and 2-hour post-meal blood glucose values, daily steps, and heart rate) are integrated into the patient's chronic disease management timeline by date using a timestamp alignment algorithm to form a standardized medical record dataset. The data verification unit automatically removes invalid data (such as abnormal blood glucose values ​​caused by uncalibrated blood glucose meters) and reminds patients to re-monitor.

[0047] The data processing module preprocesses the standardized dataset: it uses median imputation to supplement a small number of missing postprandial blood glucose data, removes outliers (such as a significant increase in blood glucose due to improper diet) using a clustering algorithm, and removes duplicate glycated hemoglobin test reports. A feature extraction model is built based on the gradient boosting algorithm to select indicators strongly correlated with the treatment effect of type 2 diabetes, including fasting blood glucose, 2-hour postprandial blood glucose, glycated hemoglobin, severity of foot numbness, progression of fundus lesions, lipid indicators (total cholesterol, triglycerides), and blood glucose monitoring compliance. Individual information such as patient age, body mass index (23.2 kg / m²), family history (mother has type 2 diabetes), dietary habits (low-fat diet), and exercise intensity are also incorporated to construct a efficacy evaluation dataset. The dataset is encrypted and stored using the AES-256 encryption algorithm. A data traceability log records the collection source (in-hospital electrocardiogram, out-of-hospital blood glucose meter) and processing time for each data point, supporting full-process data traceability. The module automatically updates the efficacy evaluation dataset every 7 days, incorporating the week's in-hospital and out-of-hospital diagnosis and monitoring data.

[0048] The efficacy modeling module matches the type 2 diabetes efficacy assessment standard library (referencing the "China Type 2 Diabetes Prevention and Treatment Guidelines 2022"). It uses a time series neural network combined with the ARIMA algorithm to construct a dynamic efficacy model, which tracks and analyzes the patient's blood glucose data and glycated hemoglobin values ​​for 6 consecutive months, generating blood glucose trend curves and glycated hemoglobin fluctuation curves. At the same time, it calculates the monthly blood glucose control target achievement rate (number of days with fasting blood glucose of 4.4-7.0 mmol / L and 2-hour postprandial blood glucose <10.0 mmol / L / total number of monitoring days per month). Correction factors were constructed based on individual patient information: the correction coefficient for the difficulty of blood glucose control due to family history was 1.08; the weight of the risk of complication progression due to age was adjusted to 1.1; and the weight of efficacy assessment due to blood glucose monitoring compliance (90%) was 0.98. The trend curve was calibrated through multivariate regression analysis to output accurate efficacy assessment results: the patient's current treatment effect is "controllable". Fasting blood glucose has decreased from 8.2 mmol / L at the beginning of management to 6.8 mmol / L currently, 2-hour postprandial blood glucose has decreased from 11.5 mmol / L to 9.2 mmol / L, and glycated hemoglobin has decreased from 8.5% to 7.3%. The symptoms of numbness in the feet have not worsened, and the fundus lesions have not progressed. However, postprandial blood glucose control has not yet reached the target. It is predicted that if the exercise intensity is increased and the gliclazide dose is adjusted in the next 3 months, the postprandial blood glucose can be reduced to below 8.0 mmol / L, and the glycated hemoglobin is expected to be reduced to below 7.0%, thus reducing the risk of complications.

[0049] The display adaptation module provides a multi-role interface: The doctor's interface adopts a column layout, with the left side displaying the patient's basic medical record information and chronic disease management plan, the middle displaying the dynamic trend curve of blood glucose, the change curve of glycated hemoglobin and the progress of complications, and the right side providing treatment plan adjustment suggestions ("It is recommended to adjust the gliclazide dose from 80mg twice a day to 100mg twice a day, and increase the number of exercise sessions to 5 times a week, 40 minutes each time") and efficacy reference data of similar patients; The patient's smart terminal (mobile APP) displays a simplified interface, showing the weekly average blood glucose value and glycated hemoglobin results in graphic form, along with a blood glucose control target achievement score (82 points), and personalized health guidance ("Brisk walking can be done 1 hour after meals, which helps with blood glucose control"), medication reminders and the next follow-up time; The community health management center display screen shows the efficacy statistics of the diabetic patient group in the jurisdiction, including the blood glucose control target achievement rate, the incidence of complications, and the efficacy comparison of different age groups, providing support for managers to optimize chronic disease management plans.

[0050] The interactive control module allows doctors to compare multiple cases, selecting 10 type 2 diabetes patients with the same disease course and medication regimen. It compares the impact of different exercise intensities on blood glucose control, generating comparative analysis reports to assist doctors in developing personalized chronic disease management plans. The module also supports retrospective analysis by management stage (months 1-2, 3-4, and 5-6), allowing doctors to view changes in patient efficacy at different stages and compare the correlation between dietary and exercise adjustments and blood glucose control. When a patient's blood glucose level rises to 12.3 mmol / L (moderately abnormal) 2 hours after a meal, the system issues a moderate warning (mobile app push notification + doctor's workstation pop-up), linking relevant data ("Daily staple food intake exceeded the standard, no post-meal exercise") to help patients and doctors pinpoint the cause. The module allows doctors to export 6 months of patient efficacy data (PDF format) for chronic disease follow-up archiving, and also allows patients to export their own blood glucose data to share with family members or referring doctors. Regarding access control, community administrators can only view group statistics and cannot access individual patient privacy information.

[0051] In this embodiment, the system operates stably, with out-of-hospital data transmission delay ≤1 minute, semantic parsing accuracy ≥90%, and efficacy evaluation results consistent with doctors' chronic disease management judgments reaching 94%. Through this system, patients' self-management awareness has significantly improved, blood glucose monitoring compliance has increased from 90% to 96%, postprandial blood glucose control target achievement rate has increased by 15%, glycated hemoglobin target achievement rate (<7.0%) has increased by 10%, and the progression of complications such as foot numbness and fundus lesions has been effectively controlled, significantly improving the quality of diabetes chronic disease management.

[0052] To verify the superiority of the system of the present invention, the following 6 sets of comparative examples were set up, all based on the application scenarios of the corresponding embodiments, to compare the therapeutic effects and application value of the system of the present invention with those of the prior art or schemes lacking some modules of the present invention: Comparative Example 1: An existing traditional electronic medical record system (which only displays structured data and lacks multi-source data fusion capabilities) was applied to the outpatient hypertension scenario in Example 1. The results showed that data collection only covered electronic medical records and laboratory data, failing to acquire outpatient blood pressure and dietary / exercise information. The efficacy display only showed isolated blood pressure values ​​without trend analysis or individual correction, making it difficult for doctors to comprehensively assess efficacy. The patient blood pressure control achievement rate improved by only 3%, significantly lower than the 12% of the system of this invention.

[0053] Comparative Example 2: A simplified system lacking a efficacy modeling module (only integrating multi-source data, without dynamic modeling and individual calibration) was applied to the outpatient hypertension scenario in Example 1. Results showed that the system could display multi-source data, but only statically, and could not generate efficacy trend curves or prediction results. The efficacy assessment did not consider individual factors such as the patient's history of diabetes and age, resulting in an assessment bias rate of 18%, and consistency with the doctor's clinical judgment was only 75%.

[0054] Comparative Example 3: An existing inpatient monitoring system (which only displays real-time physiological data and lacks multi-source data integration and display adaptation) was applied to the acute pancreatitis scenario in Example 2. Results showed that the system could not integrate imaging reports, laboratory data, and consultation opinions; the efficacy display was limited; different roles needed to switch between multiple systems to view data; multidisciplinary collaboration efficiency was low; the complication rate remained at 17%; and the length of hospital stay was not significantly shortened, failing to meet clinical needs.

[0055] Comparative Example 4: A system lacking an interactive control module (only displaying efficacy data, without early warning linkage or multi-case comparison) was applied to the acute pancreatitis scenario in Example 2. Results showed that the system could present dynamic efficacy trends, but when efficacy indicators were abnormal, there were no tiered early warnings or linked data, making it difficult for doctors to quickly locate the cause. The average delay in abnormal handling increased by 20 minutes, reducing the safety of diagnosis and treatment.

[0056] Comparative Example 5: An existing chronic disease management app (which only records blood glucose data and lacks efficacy modeling and personalized display) was applied to the diabetes chronic disease management scenario in Example 3. Results showed that it could only record blood glucose values, could not generate trend curves or assess complication risks, patients could not obtain personalized health guidance, blood glucose monitoring adherence improved by only 5%, and the glycated hemoglobin target achievement rate did not significantly improve.

[0057] Comparative Example 6: A system lacking a data processing module (no feature extraction and data encryption) was applied to the diabetes chronic disease management scenario in Example 3. Results showed severe data redundancy, with invalid data accounting for 15%, making it impossible to screen core efficacy indicators. Furthermore, there was a risk of data leakage, failing to meet medical data compliance requirements and rendering the system unsuitable for practical application.

[0058] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A system for displaying treatment effects based on medical record information, characterized in that, include: The data acquisition module is used to acquire patients' electronic medical records, real-time physiological monitoring data, medical imaging reports, laboratory test results, outpatient follow-up records, and diagnosis and treatment operation records. It uses natural language processing semantic parsing technology to extract key diagnosis and treatment information from unstructured data, integrates multi-source data according to the diagnosis and treatment timeline through a time-series alignment algorithm, and outputs a standardized medical record dataset containing structured, unstructured, and real-time data. The built-in data verification unit ensures data validity. The data processing module performs cleaning, deduplication, and format unification preprocessing on standardized medical record datasets. It uses a feature extraction model to screen indicators that are strongly correlated with treatment effects, constructs a efficacy evaluation dataset that integrates objective indicators and subjective symptoms, and simultaneously uses encryption technology to achieve full-process security control of medical record information and establishes a data traceability mechanism to ensure compliance. The efficacy modeling module dynamically models the efficacy assessment dataset based on time series analysis, generates a treatment effect trend curve, constructs individual correction factors by combining patient age, underlying diseases, and physical characteristics, and outputs accurate efficacy assessment results by calibrating the trend curve. The display adaptation module provides a customized interface based on the user role's preset display rules: the doctor's side displays complete efficacy indicators, trend changes, and abnormal warning signs; the patient's side presents a simplified version of efficacy results and health guidance; and the management side displays group efficacy statistics and treatment quality analysis. The interactive control module supports retrospective analysis of treatment data by stage of diagnosis and treatment, comparison of treatment effects in multiple cases, triggering an early warning signal and linking related medical record data when treatment indicators fluctuate abnormally, and providing data export and printing functions. All modules achieve bidirectional communication through a data bus to ensure real-time data transmission.

2. The medical record information treatment effect display system according to claim 1, characterized in that: The data acquisition module also includes an external device interface, which can access real-time data from smart wearable devices and home health monitoring instruments. Data transmission uses Bluetooth and 5G dual-mode adaptation to ensure real-time uploading of external data.

3. The medical record information treatment effect display system according to claim 1, characterized in that: The individual correction factor in the efficacy modeling module is also dynamically updated in conjunction with the patient's medication adherence and lifestyle adjustments, with the update cycle synchronized with the treatment follow-up cycle.

4. The medical record information treatment effect display system according to claim 1, characterized in that: The display adaptation module also supports scene switching. In clinical emergency scenarios, it automatically prioritizes displaying critical values ​​associated with efficacy indicators, while in outpatient follow-up scenarios, it focuses on displaying the trend of efficacy changes and suggestions for adjusting treatment plans.

5. The medical record information treatment effect display system according to claim 1, characterized in that: The multi-case comparison function of the interactive control module can group and filter patients according to disease type, treatment plan, and patient physical characteristics, and generate comparative analysis reports.

6. The medical record information treatment effect display system according to claim 1, characterized in that: The feature extraction model of the data processing module uses a gradient boosting algorithm to optimize the accuracy of indicator screening and eliminate redundant data and interfering indicators.

7. A medical record information treatment effect display system according to claim 1, characterized in that: The efficacy modeling module also has an efficacy prediction function, which predicts the trend of subsequent treatment effects based on existing diagnosis and treatment data and similar historical cases, providing a reference for optimizing the diagnosis and treatment plan.

8. A medical record information treatment effect display system according to claim 1, characterized in that: The interface of the display adaptation module supports custom configuration. Users can add commonly used efficacy indicators to the homepage display and adjust the display order and visualization format of the indicators.

9. A medical record information treatment effect display system according to claim 1, characterized in that: The warning signals of the interactive control module are divided into three levels, corresponding to mild, moderate and severe abnormalities in efficacy indicators, respectively. Different levels of warning correspond to different linkage response mechanisms.

10. A medical record information treatment effect display system according to claim 1, characterized in that: The data processing module also has a data traceability function, which marks the source, processing time and processing process of each efficacy indicator data to ensure data traceability.

Images