A multidimensional fusion-based dry eye diagnosis and treatment data management system and method

Abstract

This invention discloses a dry eye diagnosis and treatment data management system and method based on multidimensional fusion, belonging to the field of data management technology. This invention defines three types of quality control metadata standards. When a user undergoes dry eye diagnosis and treatment, all three types of quality control metadata are encapsulated into data packets using an API interface. The rule compliance score, statistical stability score, and historical consistency score of each data packet are calculated. Based on these three scores, a multidimensional quality scoring model is constructed to output a comprehensive quality score for each data packet. A quality threshold is preset, and an automatic workflow is configured to display the diagnosis and treatment data in the user data packets to the doctor's workstation according to the quality threshold. A data passport is constructed for each data packet, and a "credibility details" icon is set when displaying user diagnosis and treatment data. Doctors can click the icon to view the data passport of the diagnosis and treatment data. An analysis cycle is set, and the data passports of all data packets are analyzed according to the analysis cycle, outputting a data quality optimization report.

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically to a dry eye diagnosis and treatment data management system and method based on multidimensional fusion. Background Technology

[0002] Dry eye is a common chronic ocular surface disease in clinical ophthalmology. Its diagnosis and treatment involve multiple professional examinations, such as tear secretion testing, tear film breakup time measurement, and ocular surface staining. The resulting diagnostic and treatment data are characterized by multi-dimensionality, high frequency, and strong correlation. Furthermore, the diagnosis, treatment plan adjustment, and disease follow-up of dry eye disease all rely heavily on continuous and reliable diagnostic and treatment data.

[0003] With the development of digital ophthalmology, various dry eye examination devices have achieved electronic data collection, and medical institutions have gradually accumulated a large amount of dry eye diagnosis and treatment data. However, the existing data management system is still at the stage of simple storage and extensive management, and a professional and standardized quality control and management system for dry eye diagnosis and treatment data has not yet been formed. At the same time, the generation of dry eye diagnosis and treatment data is affected by multiple factors such as the condition of the examination equipment, the standardization of the operators, and the parameters of the examination environment. Abnormalities in any of these factors can lead to data distortion. If distorted data flows into the clinical diagnosis and treatment process, it is very easy to cause misdiagnosis and missed diagnosis, affecting the diagnosis and treatment effect of dry eye diseases. When abnormal values ​​appear in the data, it is impossible to quickly trace back whether it is due to equipment failure, operational error or real pathological changes, affecting the reliability of the data. Summary of the Invention

[0004] The purpose of this invention is to provide a dry eye diagnosis and treatment data management system and method based on multidimensional fusion, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for managing dry eye diagnosis and treatment data based on multidimensional fusion, characterized in that the method includes the following steps:

[0007] S100. Define three types of quality control metadata standards. When a user undergoes dry eye treatment, collect all the user's treatment data and store the three types of quality control metadata standards. Use the API interface to encapsulate all three types of quality control metadata of the user's treatment into a data packet.

[0008] Furthermore, the specific steps for encapsulating all three types of quality control metadata for user diagnosis and treatment into a data packet using the API interface are as follows:

[0009] S101. The three types of quality control metadata standards include equipment dimension, operation dimension, and environment dimension. Among them, the equipment dimension standard includes the equipment unique identifier D. ID Final calibration timestamp T cal Equipment self-test status code Sself Sensor reading V senser The device self-test status codes include abnormal 0 and normal 1; the three categories of quality control metadata standards—equipment dimension, operation dimension, and environment dimension—are clearly defined, and the data is standardized from the core impact dimensions of data generation, so that the quality control of dry eye diagnosis and treatment data has a unified and quantifiable basis, solving the problem of traditional diagnosis and treatment data collection lacking unified standards and having a single dimension.

[0010] The operational dimension standard includes operator ID Z. ID Operational compliance completed, marked with F proc The compliance completion mark for the operation is obtained through an expert experience scoring mechanism.

[0011] The environmental standards include the examination room temperature (Temp), relative humidity (RH), and data collection timestamp (T). acq ;

[0012] S102. When a user performs a diagnostic examination, while acquiring the examination result data V, three types of quality control metadata are collected according to the three types of quality control metadata standards; the three types of quality control metadata are encapsulated into a data packet Packet={D} via the API interface. ID T cal S self V senser Z ID F proc Temp, RH, T acq Simultaneous collection of diagnosis and treatment results data and quality control metadata enables integrated collection of data "results + traceability information," providing complete original evidence for subsequent data quality assessment and problem tracing, and avoiding the drawback of focusing only on result data without being able to judge data reliability.

[0013] By encapsulating standardized data packets through API interfaces, standardized data transmission and storage can be achieved, improving the efficiency of data interaction within the system while ensuring data integrity and preventing information loss and format corruption during transmission.

[0014] S200. Calculate the rule compliance score, statistical stability score, and historical consistency score for each data packet. Based on the three scores, construct a multi-dimensional quality scoring model and output the comprehensive quality score for each data packet.

[0015] Furthermore, the specific steps for constructing a multi-dimensional quality scoring model based on the three scores and outputting a comprehensive quality score for each data packet are as follows:

[0016] S201. Predefined hard rules, with a preset rule compliance score of 100. The rule compliance score is calculated using a deduction method. The hard rules include: when S selfWhen the value is 0, it is judged as an equipment detection anomaly, and the rule compliance score R is set to 0. rule =0;

[0017] The preset normal temperature difference threshold is T. nor When abs(Temp-22℃) > T nor When the temperature exceeds the normal temperature difference, the rule compliance score is set to R. rule Deduct 30 points;

[0018] Preset compliance threshold, when F proc When the compliance threshold is reached, let the rule compliance score R be... rule Deduct 20 points; the rule compliance score is calculated based on the hard rules.

[0019] By using a hard-rule deduction method to quantify whether data meets basic quality control requirements, low-quality data such as equipment malfunctions, environmental exceedances, and non-compliant operations can be quickly screened out, ensuring data validity from a basic compliance perspective and excluding obviously unreliable data sources for clinical diagnosis and treatment.

[0020] S202. When a user undergoes a diagnostic examination, obtain n examination result data V, calculate the mean and standard deviation of the n examination result data, and calculate the coefficient of variation of the examination result data, using the following formula:

[0021] ;

[0022] In the formula, CV represents the coefficient of variation, and V st V represents the standard deviation of the inspection results data. p This represents the average value of the inspection results data;

[0023] The statistical stability score of data packets is calculated using the coefficient of variation, using the following formula:

[0024] ;

[0025] In the formula, S stat The coefficient of variation represents the statistical stability score, and k represents the amplification factor. The dispersion of the test results data is calculated by the coefficient of variation, which quantifies the stability of the data itself and avoids clinical misjudgment caused by large fluctuations in a single test. This ensures the reliability of the diagnostic data from the perspective of the data's own characteristics.

[0026] S203. Collect historical inspection result data of users and calculate the standard score of the historical inspection result data of users, using the following formula:

[0027] ;

[0028] In the formula, Z represents the standard score, and V history-st V represents the standard deviation of historical inspection results data.history-p This represents the average value of historical inspection results data;

[0029] Calculate the historical consistency score C using standard scores. consist The formula is: C consist =100-10×Z; Combine the user's historical examination data to calculate the standard score, quantify the matching degree between the current data and historical data, fit the chronic and progressive characteristics of dry eye disease, evaluate the rationality of the data from the perspective of individual diagnosis and treatment patterns, and avoid the interpretation of isolated data that is detached from the patient's medical history.

[0030] S204. Weighted fusion of the rule compliance score, statistical stability score, and historical consistency score of each data packet, and construct a multi-dimensional quality scoring model based on the three scores, with the formula: Q=α×R rule +β×S stat +γ×C consist Q represents the overall quality score. A weighted, fusion-based multidimensional quality scoring model is constructed, integrating the scores from the three dimensions into a comprehensive quality score. This enables a holistic, quantitative assessment of data quality, replacing traditional subjective judgments of data reliability and making data quality evaluation more scientific and objective.

[0031] S300: Preset quality threshold, configure automatic workflow based on quality threshold to display diagnosis and treatment data in user data package to doctor workstation;

[0032] Furthermore, the specific steps for configuring an automatic workflow based on a quality threshold to display the diagnostic data in the user data packet to the doctor's workstation are as follows:

[0033] S301, Preset quality threshold Q high and Q low The specific steps to configure the automatic workflow are as follows:

[0034] When Q>Q high If the data packet is deemed to be of high quality, it is marked as "highly reliable" and directly transmitted to the doctor's workstation for priority display.

[0035] When Q low <Q<Q high If the data packet is determined to be of medium quality, it is marked as "Reference Available" and transmitted to the doctor's workstation with the message "Slight deviation, please make a clinical judgment" displayed.

[0036] When Q low When a data packet is determined to be of low quality, it is marked as "pending verification," the data packet is automatically frozen, and an alert is triggered simultaneously.

[0037] Send an alarm to the department's quality control personnel, along with the data packet's unique device identifier D. ID ​The comprehensive quality score is sent out along with the alarm;

[0038] Push a reminder to the operator corresponding to the data packet, and recommend that the operator review the operation process;

[0039] Divide the data into three quality levels of high, medium, and low according to the comprehensive quality score and configure differential push rules to achieve hierarchical and classified management of data, allowing doctors to view high-trust data first, improving the efficiency of diagnosis and treatment work, and avoiding interference from low-quality data in clinical decision-making.

[0040] Automatically freeze low-quality data and trigger multi-terminal warnings to achieve timely warning and closed-loop processing of problem data, quickly locate the person responsible for data quality problems, and prevent low-quality data from flowing into the clinical diagnosis and treatment process.

[0041] S302. For the historical consistency score C consist Separate preset consistency threshold Cin, when C consist < Cin, ignore the comprehensive quality score Q value, automatically create a "clinical review task", and prompt "The test result is significantly different from the patient's historical pattern, please confirm" when transmitting the data packet to the doctor's workstation.

[0042] Separate clinical review tasks are created for data with too low historical consistency scores, specifically prompting doctors to pay attention to the differences between the data and the patient's historical pattern, conforming to the diagnostic and treatment characteristics of dry eye diseases, reducing misdiagnosis and missed diagnosis caused by data abnormalities, and improving the accuracy of clinical diagnosis and treatment.

[0043] S400. Build a data passport for each data packet, set a "credibility details" icon when displaying the user's diagnostic data, and the doctor clicks on the icon to view the data passport of the diagnostic data;

[0044] Furthermore, the specific steps for the doctor to click on the icon to view the data passport of the diagnostic data are as follows:

[0045] S401. The data passport includes: identity summary, quality sub-report, original metadata, handling and warning log;

[0046] The identity summary includes the unique identification code of the data packet device, acquisition timestamp, comprehensive quality score, and data packet mark;

[0047] The quality sub-report includes the rule compliance score, statistical stability score, historical consistency score of the data packet displayed through a dashboard, and the corresponding deduction item details;

[0048] The original metadata includes the examination equipment, operator, operation record, and environmental metadata during the user's diagnosis and treatment;

[0049] The treatment and early warning logs include the specific content of the automated workflow; a data passport is constructed for each data packet, containing an identity summary, quality sub-report, original metadata, and treatment and early warning logs, to achieve full-dimensional visualization of data quality information, enabling doctors to quickly trace the quality control details and deductions of the data, and providing a complete reference for clinical data interpretation.

[0050] S402. When displaying the examination results of the user data package's diagnostic data in the doctor's workstation, set a "credibility details" icon. Doctors can click the icon to view the data passport of the diagnostic data. The "credibility details" icon in the doctor's workstation adopts a lightweight interactive design that does not interfere with the doctor's routine viewing of diagnostic data, while allowing them to view detailed quality control information as needed, thus balancing the needs of diagnostic efficiency and data traceability.

[0051] The comparison view allows for side-by-side comparison of data passports from different time points for the same examination items. This multi-time-point comparison view of identical examination items aligns with the chronic management characteristics of dry eye disease, enabling doctors to intuitively compare changes in the quality and outcomes of patient treatment data across different periods. This provides a more comprehensive basis for assessing the disease's progression and adjusting treatment plans.

[0052] S500: Set the analysis cycle, analyze the data passport of all data packets according to the analysis cycle, and output a data quality optimization report.

[0053] Furthermore, the specific steps for the output data quality optimization report are as follows:

[0054] S501, Set the analysis period T f The data passports of all data packets are analyzed according to the analysis period, specifically as follows:

[0055] Using quantitative analysis, with any one of the following factors—operator, data collection timestamp, equipment unique identifier, inspection room temperature, and relative humidity—constantly selected as the analysis item, m comprehensive quality scores were collected and plotted. The average of the m comprehensive quality scores was calculated. When the average comprehensive quality score decreased compared to the historical average, the historical average score Q was used to determine the trend. h The average quality score Q of the same analysis item in the current period c The formula for calculating the decrease P is:

[0056] P=[(Q) h -Q c ) / Q h ×100%;

[0057] For the equipment dimension, an anomaly detection threshold is set as Y1; for the operational dimension, it is Y2; and for the environmental dimension, it is Y3. The requirement is that Y1... <Y2<Y3;

[0058] For different types of analysis items, select an anomaly judgment threshold. When P ≥ the corresponding anomaly judgment threshold, the corresponding analysis item is judged to be abnormal. Equipment dimension anomalies are defined as level 1 anomalies, operation dimension anomalies as level 2 anomalies, and environmental dimension anomalies as level 3 anomalies.

[0059] S502. Experts set the minimum threshold H for samples based on experience. m When the number of data packet samples in the current analysis period is less than the minimum sample threshold H m When the decrease magnitude P is adjusted by the sample size, the formula is as follows:

[0060] P' = P × R;

[0061] R represents the correction factor, which is preset by experts; P' represents the reduction after correction.

[0062] Quantitative analysis is performed on all data passports at fixed intervals to uncover patterns of data quality changes across core dimensions such as operators, equipment, environment, and time. This enables routine and systematic monitoring of data quality and timely detection of potential anomalies.

[0063] S503. Based on the analysis results of the quantitative analysis method, a data quality optimization report is compiled and sent to the equipment maintenance personnel and management personnel. The maintenance personnel and management personnel maintain and process the abnormal analysis items according to the priority of the abnormality level from small to large, with level 1 being the highest priority and level 3 being the lowest priority.

[0064] Output data quality optimization reports and push them to relevant responsible persons to achieve root-cause solutions to data quality problems, form a closed-loop management of "collection-quality control-early warning-optimization", and continuously improve the overall quality of dry eye diagnosis and treatment data.

[0065] A dry eye diagnosis and treatment data management system based on multi-dimensional fusion, comprising a data acquisition module, a data quality scoring module, an automated operation module, a display module, and an anomaly tracing module;

[0066] The data acquisition module is used to define three types of quality control metadata standards. When a user undergoes dry eye treatment, it collects all the user's treatment data and stores the three types of quality control metadata standards. It also uses an API interface to encapsulate all three types of quality control metadata of the user's treatment into a data packet.

[0067] The data quality scoring module is used to calculate the rule compliance score, statistical stability score, and historical consistency score of each data packet, and to construct a multi-dimensional quality scoring model based on the three scores to output the comprehensive quality score of each data packet.

[0068] The automated workflow module is used to preset quality thresholds and configure an automated workflow based on the quality thresholds to display the diagnosis and treatment data in the user data package to the doctor's workstation.

[0069] The display module is used to construct a data passport for each data packet. When displaying user medical data, a "credibility details" icon is set. Doctors can click the icon to view the data passport of the medical data.

[0070] The anomaly tracing module is used to set the analysis cycle, analyze the data passports of all data packets according to the analysis cycle, and output a data quality optimization report.

[0071] The data quality scoring module includes a rule compliance unit, a statistical stability unit, and a historical consistency unit;

[0072] The rule compliance unit is used to predefine hard rules. The maximum rule compliance score is 100, and the rule compliance score is calculated using a deduction method.

[0073] The statistical stability unit is used to acquire n test result data V when a user performs a diagnostic examination, calculate the mean and standard deviation of the n test result data respectively, calculate the coefficient of variation of the test result data, and use the coefficient of variation to calculate the statistical stability score of the data packet.

[0074] The historical consistency unit is used to collect users' historical inspection result data, calculate the standard score of users' historical inspection result data, and use the standard score to calculate the historical consistency score.

[0075] The display module includes a data passport unit and an icon display unit;

[0076] The data passport unit is used to construct the data passport for each data packet. The data passport includes: identity digest, quality sub-report, raw metadata, and processing and early warning logs.

[0077] The icon display unit is used to set a "credibility details" icon when displaying the examination results data of the user data package diagnosis data in the doctor's workstation.

[0078] Compared with the prior art, the beneficial effects of the present invention are:

[0079] 1. This invention establishes quality control standards from three dimensions: equipment, operation, and environment. It combines rule compliance, statistical stability, and historical consistency to construct a multi-dimensional quality scoring model, thereby achieving full-dimensional and quantitative quality control of diagnostic and treatment data. This ensures the effectiveness and reliability of the data from the source and process, and solves the problems of traditional dry eye diagnosis and treatment data quality lacking a unified evaluation standard and being highly subjective.

[0080] 2. This invention enables the hierarchical and categorized delivery of diagnostic and treatment data through an automated workflow, allowing doctors to prioritize viewing highly reliable data. It also provides data passports with traceability visualization and multi-time point comparison functions, providing a complete basis for clinical data interpretation and disease progression assessment. This reduces the interference of low-quality data on clinical decision-making and aligns with the diagnostic and treatment characteristics of chronic dry eye management, thereby reducing the probability of misdiagnosis and missed diagnosis.

[0081] 3. The solution fully considers the clinical characteristics of dry eye disease, which is chronic, progressive, and requires long-term follow-up. It focuses on designing functions such as historical consistency scores, data passport comparison views, and separate clinical review tasks, so that the data management system is deeply matched with the clinical diagnosis and treatment needs of dry eye disease, and provides scientific data support for the chronic management of dry eye disease and the adjustment of treatment plans. Attached Figure Description

[0082] Figure 1 This is a schematic diagram illustrating the steps of a dry eye diagnosis and treatment data management method based on multidimensional fusion according to the present invention. Detailed Implementation

[0083] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0084] Example: Figure 1 As shown, the present invention provides a technical solution.

[0085] A method for managing dry eye diagnosis and treatment data based on multidimensional fusion, characterized in that the method includes the following steps:

[0086] S100. Define three types of quality control metadata standards. When a user undergoes dry eye treatment, collect all the user's treatment data and store the three types of quality control metadata standards. Use the API interface to encapsulate all three types of quality control metadata of the user's treatment into a data packet.

[0087] The specific steps for encapsulating all three types of quality control metadata for user diagnosis and treatment into a data packet using the API interface are as follows:

[0088] S101. The three types of quality control metadata standards include equipment dimension, operation dimension, and environment dimension. Among them, the equipment dimension standard includes the equipment unique identifier D. ID Final calibration timestamp T cal Equipment self-test status code S self Sensor reading V senserThe device self-test status codes include abnormal 0 and normal 1; the three categories of quality control metadata standards—equipment dimension, operation dimension, and environment dimension—are clearly defined, and the data is standardized from the core impact dimensions of data generation, so that the quality control of dry eye diagnosis and treatment data has a unified and quantifiable basis, solving the problem of traditional diagnosis and treatment data collection lacking unified standards and having a single dimension.

[0089] The operational dimension standard includes operator ID Z. ID Operational compliance completed, marked with F proc The compliance completion mark for the operation is obtained through an expert experience scoring mechanism.

[0090] The environmental standards include the examination room temperature (Temp), relative humidity (RH), and data collection timestamp (T). acq ;

[0091] S102. When a user performs a diagnostic examination, while acquiring the examination result data V, three types of quality control metadata are collected according to the three types of quality control metadata standards; the three types of quality control metadata are encapsulated into a data packet Packet={D} via the API interface. ID T cal S self V senser Z ID F proc Temp, RH, T acq Simultaneous collection of diagnosis and treatment results data and quality control metadata enables integrated collection of data "results + traceability information," providing complete original evidence for subsequent data quality assessment and problem tracing, and avoiding the drawback of focusing only on result data without being able to judge data reliability.

[0092] By encapsulating standardized data packets through API interfaces, standardized data transmission and storage can be achieved, improving the efficiency of data interaction within the system while ensuring data integrity and preventing information loss and format corruption during transmission.

[0093] S200. Calculate the rule compliance score, statistical stability score, and historical consistency score for each data packet. Based on the three scores, construct a multi-dimensional quality scoring model and output the comprehensive quality score for each data packet.

[0094] The specific steps for constructing a multidimensional quality scoring model based on three scores and outputting a comprehensive quality score for each data packet are as follows:

[0095] S201. Predefined hard rules, with a preset rule compliance score of 100. The rule compliance score is calculated using a deduction method. The hard rules include: when S self When the value is 0, it is judged as an equipment detection anomaly, and the rule compliance score R is set to 0. rule =0;

[0096] The preset normal temperature difference threshold is T. nor When abs(Temp-22℃) > T nor When the temperature exceeds the normal temperature difference, the rule compliance score is set to R. rule Deduct 30 points;

[0097] Preset compliance threshold, when F proc When the compliance threshold is reached, let the rule compliance score R be... rule Deduct 20 points; the rule compliance score is calculated based on the hard rules.

[0098] By using a hard-rule deduction method to quantify whether data meets basic quality control requirements, low-quality data such as equipment malfunctions, environmental exceedances, and non-compliant operations can be quickly screened out, ensuring data validity from a basic compliance perspective and excluding obviously unreliable data sources for clinical diagnosis and treatment.

[0099] S202. When a user undergoes a diagnostic examination, obtain n examination result data V, calculate the mean and standard deviation of the n examination result data, and calculate the coefficient of variation of the examination result data, using the following formula:

[0100] ;

[0101] In the formula, CV represents the coefficient of variation, and V st V represents the standard deviation of the inspection results data. p This represents the average value of the inspection results data;

[0102] The statistical stability score of data packets is calculated using the coefficient of variation, using the following formula:

[0103] ;

[0104] In the formula, S stat The coefficient of variation represents the statistical stability score, and k represents the amplification factor. The dispersion of the test results data is calculated by the coefficient of variation, which quantifies the stability of the data itself and avoids clinical misjudgment caused by large fluctuations in a single test. This ensures the reliability of the diagnostic data from the perspective of the data's own characteristics.

[0105] S203. Collect historical inspection result data of users and calculate the standard score of the historical inspection result data of users, using the following formula:

[0106] ;

[0107] In the formula, Z represents the standard score, and V history-st V represents the standard deviation of historical inspection results data. history-p This represents the average value of historical inspection results data;

[0108] Calculate the historical consistency score C using standard scores. consist The formula is: C consist =100-10×Z; Combine the user's historical examination data to calculate the standard score, quantify the matching degree between the current data and historical data, fit the chronic and progressive characteristics of dry eye disease, evaluate the rationality of the data from the perspective of individual diagnosis and treatment patterns, and avoid the interpretation of isolated data that is detached from the patient's medical history.

[0109] S204. Weighted fusion of the rule compliance score, statistical stability score, and historical consistency score of each data packet, and construct a multi-dimensional quality scoring model based on the three scores, with the formula: Q=α×R rule +β×S stat +γ×C consist Q represents the overall quality score. A weighted, fusion-based multidimensional quality scoring model is constructed, integrating the scores from the three dimensions into a comprehensive quality score. This enables a holistic, quantitative assessment of data quality, replacing traditional subjective judgments of data reliability and making data quality evaluation more scientific and objective.

[0110] S300: Preset quality threshold, configure automatic workflow based on quality threshold to display diagnosis and treatment data in user data package to doctor workstation;

[0111] The specific steps for configuring an automated workflow based on a quality threshold to display the diagnostic data from the user data packet to the doctor's workstation are as follows:

[0112] S301, Preset quality threshold Q high and Q low The specific steps to configure the automatic workflow are as follows:

[0113] When Q>Q high If the data packet is deemed to be of high quality, it is marked as "highly reliable" and directly transmitted to the doctor's workstation for priority display.

[0114] When Q low <Q<Q high If the data packet is determined to be of medium quality, it is marked as "Reference Available" and transmitted to the doctor's workstation with the message "Slight deviation, please make a clinical judgment" displayed.

[0115] When Q low When a data packet is determined to be of low quality, it is marked as "pending verification," the data packet is automatically frozen, and an alert is triggered simultaneously.

[0116] Send an alarm to the department's quality control personnel, along with the data packet's unique device identifier D. ID The overall quality score was issued along with the alarm.

[0117] ​Push a reminder to the operator corresponding to the data packet, and recommend that the operator review the operation process;

[0118] Divide the data into three quality levels of high, medium, and low according to the comprehensive quality score, and configure differentiated push rules to achieve hierarchical and classified management of data, enabling doctors to view highly credible data first, improving the efficiency of diagnosis and treatment work, and avoiding interference from low-quality data in clinical decision-making.

[0119] Automatically freeze low-quality data and trigger multi-terminal warnings to achieve timely warning and closed-loop processing of problem data, quickly locate the responsible person for data quality problems, and prevent low-quality data from flowing into the clinical diagnosis and treatment process.

[0120] S302. For the historical consistency score C consist Separate the preset consistency threshold Cin. When C consist < Cin, regardless of the comprehensive quality score Q value, automatically create a "clinical review task", and prompt "The test result is significantly different from the patient's historical pattern. Please confirm" when transmitting the data packet to the doctor's workstation.

[0121] Separate clinical review tasks are created for data with too low historical consistency scores, specifically prompting doctors to pay attention to the differences between the data and the patient's historical pattern, conforming to the characteristics of the diagnosis and treatment of dry eye diseases, reducing misdiagnosis and missed diagnosis caused by data anomalies, and improving the accuracy of clinical diagnosis and treatment.

[0122] S400. Construct a data passport for each data packet, set a "credibility details" icon when displaying the user's diagnosis and treatment data, and doctors click on the icon to view the data passport of the diagnosis and treatment data;

[0123] The specific steps for the doctor to click on the icon to view the data passport of the diagnosis and treatment data are as follows:

[0124] S401. The data passport includes: identity summary, quality sub-report, original metadata, disposal and warning log;

[0125] The identity summary includes the unique device identification code of the data packet, collection timestamp, comprehensive quality score, and data packet mark;

[0126] The quality sub-report includes the rule compliance score, statistical stability score, historical consistency score of the data packet displayed through a dashboard, and the corresponding details of deduction items;

[0127] The original metadata includes the examination equipment, operator, operation record, and environmental metadata during the user's diagnosis and treatment; <000038'4>

[0128] The treatment and early warning logs include the specific content of the automated workflow; a data passport is constructed for each data packet, containing an identity summary, quality sub-report, original metadata, and treatment and early warning logs, to achieve full-dimensional visualization of data quality information, enabling doctors to quickly trace the quality control details and deductions of the data, and providing a complete reference for clinical data interpretation.

[0129] S402. When displaying the examination results of the user data package's diagnostic data in the doctor's workstation, set a "credibility details" icon. Doctors can click the icon to view the data passport of the diagnostic data. The "credibility details" icon in the doctor's workstation adopts a lightweight interactive design that does not interfere with the doctor's routine viewing of diagnostic data, while allowing them to view detailed quality control information as needed, thus balancing the needs of diagnostic efficiency and data traceability.

[0130] The comparison view allows for side-by-side comparison of data passports from different time points for the same examination items. This multi-time-point comparison view of identical examination items aligns with the chronic management characteristics of dry eye disease, enabling doctors to intuitively compare changes in the quality and outcomes of patient treatment data across different periods. This provides a more comprehensive basis for assessing the disease's progression and adjusting treatment plans.

[0131] S500: Set the analysis cycle, analyze the data passport of all data packets according to the analysis cycle, and output a data quality optimization report.

[0132] The specific steps for generating the output data quality optimization report are as follows:

[0133] S501, Set the analysis period T f The data passports of all data packets are analyzed according to the analysis period, specifically as follows:

[0134] Using quantitative analysis, with any one of the following factors—operator, data collection timestamp, equipment unique identifier, inspection room temperature, and relative humidity—constantly selected as the analysis item, m comprehensive quality scores were collected and plotted. The average of the m comprehensive quality scores was calculated. When the average comprehensive quality score decreased compared to the historical average, the historical average score Q was used to determine the trend. h The average quality score Q of the same analysis item in the current period c The formula for calculating the decrease P is:

[0135] P=[(Q) h -Q c ) / Q h ×100%;

[0136] For the equipment dimension, an anomaly detection threshold is set as Y1; for the operational dimension, it is Y2; and for the environmental dimension, it is Y3. The requirement is that Y1... <Y2<Y3;

[0137] For different types of analysis items, select an anomaly judgment threshold. When P ≥ the corresponding anomaly judgment threshold, the corresponding analysis item is judged to be abnormal. Equipment dimension anomalies are defined as level 1 anomalies, operation dimension anomalies as level 2 anomalies, and environmental dimension anomalies as level 3 anomalies.

[0138] S502. Experts set the minimum threshold H for samples based on experience. m When the number of data packet samples in the current analysis period is less than the minimum sample threshold H m When the decrease magnitude P is adjusted by the sample size, the formula is as follows:

[0139] P' = P × R;

[0140] R represents the correction factor, which is preset by experts; P' represents the reduction after correction.

[0141] Quantitative analysis is performed on all data passports at fixed intervals to uncover patterns of data quality changes across core dimensions such as operators, equipment, environment, and time. This enables routine and systematic monitoring of data quality and timely detection of potential anomalies.

[0142] S503. Based on the analysis results of the quantitative analysis method, a data quality optimization report is compiled and sent to the equipment maintenance personnel and management personnel. The maintenance personnel and management personnel maintain and process the abnormal analysis items according to the priority of the abnormality level from small to large, with level 1 being the highest priority and level 3 being the lowest priority.

[0143] Output data quality optimization reports and push them to relevant responsible persons to achieve root-cause solutions to data quality problems, form a closed-loop management of "collection-quality control-early warning-optimization", and continuously improve the overall quality of dry eye diagnosis and treatment data.

[0144] A dry eye diagnosis and treatment data management system based on multi-dimensional fusion, comprising a data acquisition module, a data quality scoring module, an automated operation module, a display module, and an anomaly tracing module;

[0145] The data acquisition module is used to define three types of quality control metadata standards. When a user undergoes dry eye treatment, it collects all the user's treatment data and stores the three types of quality control metadata standards. It also uses an API interface to encapsulate all three types of quality control metadata of the user's treatment into a data packet.

[0146] The data quality scoring module is used to calculate the rule compliance score, statistical stability score, and historical consistency score of each data packet, and to construct a multi-dimensional quality scoring model based on the three scores to output the comprehensive quality score of each data packet.

[0147] The automated workflow module is used to preset quality thresholds and configure an automated workflow based on the quality thresholds to display the diagnosis and treatment data in the user data package to the doctor's workstation.

[0148] The display module is used to construct a data passport for each data packet. When displaying user medical data, a "credibility details" icon is set. Doctors can click the icon to view the data passport of the medical data.

[0149] The anomaly tracing module is used to set the analysis cycle, analyze the data passports of all data packets according to the analysis cycle, and output a data quality optimization report.

[0150] The data quality scoring module includes a rule compliance unit, a statistical stability unit, and a historical consistency unit;

[0151] The rule compliance unit is used to predefine hard rules. The maximum rule compliance score is 100, and the rule compliance score is calculated using a deduction method.

[0152] The statistical stability unit is used to acquire n test result data V when a user performs a diagnostic examination, calculate the mean and standard deviation of the n test result data respectively, calculate the coefficient of variation of the test result data, and use the coefficient of variation to calculate the statistical stability score of the data packet.

[0153] The historical consistency unit is used to collect users' historical inspection result data, calculate the standard score of users' historical inspection result data, and use the standard score to calculate the historical consistency score.

[0154] The display module includes a data passport unit and an icon display unit;

[0155] The data passport unit is used to construct the data passport for each data packet. The data passport includes: identity digest, quality sub-report, raw metadata, and processing and early warning logs.

[0156] The icon display unit is used to set a "credibility details" icon when displaying the examination results data of the user data package diagnosis data in the doctor's workstation.

[0157] Example: Taking the dry eye treatment center of a tertiary hospital as an application scenario, the center is equipped with professional equipment such as a dry eye comprehensive analyzer and a tear secretion tester. It completes about 80 dry eye examinations per day. The dry eye treatment data management system and method of the present invention are used to carry out data management work and manage the dry tear secretion test data of patient Zhang throughout the whole process.

[0158] Based on industry standards and equipment operation requirements for dry eye treatment, three categories of quality control metadata standards have been established: equipment maintenance, operational maintenance, and environmental maintenance. Specific parameter thresholds have been set as follows:

[0159] Equipment maintenance: The unique identification code of the equipment is the fixed code of the equipment at the factory; the last calibration timestamp should be ≤90 days; the equipment self-test status code is 1 for normal and 0 for abnormal; the sensor readings must be within the calibrated range of the equipment.

[0160] Operation dimension: Operator ID is the hospital employee's work number; the operation compliance completion mark adopts an expert experience scoring mechanism, with a full score of 10 points and a compliance threshold set at 8 points.

[0161] Environmental parameters: Standard temperature of the inspection room is 22℃, normal temperature difference threshold is ±2℃; relative humidity is 40%-60%; data collection timestamp is the time of data collection accurate to the second.

[0162] Patient Zhang underwent tear secretion testing at this center at 10:00 AM on February 11, 2026, due to dry eye symptoms. During the treatment, the system simultaneously collected test result data and three types of quality control metadata. The specific information collected is as follows:

[0163] Inspection result data (V): A total of 5 detection values ​​were collected, namely 5mm / 5min, 6mm / 5min, 4mm / 5min, 5mm / 5min, and 6mm / 5min;

[0164] Equipment Maintenance: D ID =LY-FY2023001,T cal =January 5, 2026, 09:30 (37 days from the data collection date, ≤90 days), S self =1 (normal), V senser =0.85 (within the calibrated range of 0.5-1.0);

[0165] Operational dimension: Z ID =YK2022015,F proc =9 points (≥8 points compliance threshold);

[0166] Environmental dimension: T emp =23℃, RH=55%, T acq =2026-02-11 10:00:00.

[0167] The system encapsulates all the above data into a unified data packet through the API interface. The data packet content is: Packet={LY-FY2023001, 2026-01-05 09:30, 1, 0.85, YK2022015, 9, 23℃, 55%, 2026-02-11 10:00:00, [5,6,4,5,6]};

[0168] The weighting coefficients of the multidimensional quality scoring model are set as follows: α=0.4 (weight of rule compliance score), β=0.3 (weight of statistical stability score), γ=0.3 (weight of historical consistency score), and all scores are out of 100.

[0169] The score for compliance with the rules is calculated using a deduction method, with a maximum score of 100 points. The hard rules are determined as follows:

[0170] The device self-test status code Sself=1, no points deducted;

[0171] The inspection room temperature Temp = 23℃, abs (23-22) = 1℃ ≤ 2℃ (Tnor), no points deducted;

[0172] Operational compliance completion is marked with Fproc=9 points (≥8 points, compliance threshold), no points deducted.

[0173] Calculate the statistical stability score, and calculate the mean and standard deviation of the 5 test results: V p =5.2 mm / 5min; V st ≈0.748 mm / 5min;

[0174] Calculate the coefficient of variation (CV): CV = (0.748 / 5.2) × 100% ≈ 14.38%;

[0175] Calculate the statistical stability score, with the amplification factor k set to 2, using the formula S. stat =100-2×14.38=71.24 points, statistical stability score S stat =71.24 points.

[0176] To calculate the historical consistency score, retrieve the patient Zhang's three most recent dry tear secretion test data (October 8, 2025: 5mm / 5min; December 12, 2025: 6mm / 5min; January 15, 2026: 5mm / 5min). Calculate the mean and standard deviation of the historical data:

[0177] Historical average (V) history-p ) = ≈5.33 mm / 5min;

[0178] Historical standard deviation (V) history-st )≈0.577 mm / 5min;

[0179] The average value of this test, V p =5.2 mm / 5min, calculate the standard score Z = |5.2 - 5.33| / 0.577≈0.225;

[0180] Historical consistency score formula C consist =100-10×Z, calculated to: C consist =100 - 10 × 0.225 = 97.75 points, historical consistency score C consist =97.75 points.

[0181] Substituting the data into the multidimensional quality scoring model formula, we calculate: Q = 0.4 × 100 + 0.3 × 71.24 + 0.3 × 97.75 = 40 + 21.372 + 29.325 = 90.697 points, and the overall quality score Q ≈ 90.70 points.

[0182] The preset quality threshold at the center is: Q high =85 points, Q low =60 points, historical consistency score consistency threshold Cin=70 points, judge and push this data packet:

[0183] The overall quality score Q≈90.70 points>85 points, which is considered a high-quality data packet and marked as "highly reliable". The system directly transmits the data packet to the ophthalmologist's workstation and displays it prominently on the diagnosis and treatment interface without any additional prompts.

[0184] This historical consistency score is C. consist =97.75 points > 70 points, no need to create a clinical review task.

[0185] The system constructs a complete data passport for patient Zhang's tear secretion test data package and sets a "Reliability Details" icon on the test results interface of the doctor's workstation. Doctors can click to view the full contents of the data passport. At the same time, doctors can use the system's comparison view function to compare this data passport with Zhang's tear secretion test data passports from October 8, 2025, December 12, 2025, and January 15, 2026, to intuitively view the changes in quality scores, result data, and quality control metadata of each test.

[0186] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for managing dry eye diagnosis and treatment data based on multidimensional fusion, characterized in that: The method includes the following steps: S100. Define three types of quality control metadata standards. When a user undergoes dry eye treatment, use the API interface to encapsulate all three types of quality control metadata into a data packet. S200. Calculate the rule compliance score, statistical stability score, and historical consistency score for each data packet. Based on the three scores, construct a multi-dimensional quality scoring model and output the comprehensive quality score for each data packet. S300: Preset quality threshold, configure automatic workflow based on quality threshold to display diagnosis and treatment data in user data package to doctor workstation; S400: Construct a data passport for each data packet, and set a "credibility details" icon when displaying user medical data. Doctors can click the icon to view the data passport of the medical data. S500: Set the analysis cycle, analyze the data passport of all data packets according to the analysis cycle, and output a data quality optimization report.

2. The method for managing dry eye diagnosis and treatment data based on multidimensional fusion according to claim 1, characterized in that: The rule compliance score in S200 is specifically as follows: Predefined hard rules are used, with a maximum rule compliance score of 100. A deduction method is employed to calculate the rule compliance score. The hard rules include: when S... self When the value is 0, it is judged as an equipment detection anomaly, and the rule compliance score R is set to 0. rule =0; The preset normal temperature difference threshold is T. nor When abs(Temp-22℃) > T nor When the temperature exceeds the normal temperature difference, the rule compliance score is set to R. rule Deduct 30 points; Preset compliance threshold, when F proc When the compliance threshold is reached, let the rule compliance score R be... rule Deduct 20 points; the rule compliance score is calculated through hard rules.

3. The method for managing dry eye diagnosis and treatment data based on multidimensional fusion according to claim 1, characterized in that: The statistical stability score in S200 is specifically as follows: When a user undergoes a diagnostic examination, n examination result data points V are obtained. The mean and standard deviation of each of the n examination result data points are calculated, and the coefficient of variation of the examination result data is calculated using the following formula: ； In the formula, CV represents the coefficient of variation, and V st V represents the standard deviation of the inspection results data. p This represents the average value of the inspection results data; The statistical stability score of data packets is calculated using the coefficient of variation, using the following formula: ； In the formula, S stat This represents the statistical stability score, and k represents the amplification factor.

4. The method for managing dry eye diagnosis and treatment data based on multidimensional fusion according to claim 1, characterized in that: The historical consistency score in S200 is specifically as follows: Collect users' historical inspection result data, calculate the standard score Z of the users' historical inspection result data, and use the standard score to calculate the historical consistency score C. consist The formula is: C consist =100-10×Z.

5. The method for managing dry eye diagnosis and treatment data based on multidimensional fusion according to claim 1, characterized in that: The automatic workflow configuration in S300 is specifically as follows: Preset quality threshold Q high and Q low The specific steps to configure the automatic workflow are as follows: When Q>Q high If the data packet is deemed to be of high quality, it is marked as "highly reliable" and directly transmitted to the doctor's workstation for priority display. When Q low <Q<Q high If the data packet is determined to be of medium quality, it is marked as "Reference Available". The data packet is then transmitted to the doctor's workstation with a message displayed: "Slight deviation, please make a clinical judgment." When Q low When a data packet is deemed low-quality, it is marked as "pending verification," the packet is automatically frozen, and an alert is triggered.​ Send an alarm to the department's quality control personnel, along with the data packet's unique device identifier D. ID The overall quality score was issued along with the alarm. A reminder is sent to the operator corresponding to the data packet, suggesting that the operator review the operation process.

6. The method for managing dry eye diagnosis and treatment data based on multidimensional fusion according to claim 1, characterized in that: The data passport in the S400 includes an identity digest, a quality breakdown report, raw metadata, and processing and early warning logs; The identity digest includes the data packet device unique identifier, collection timestamp, comprehensive quality score, and data packet tag; The quality sub-report includes the rule compliance score, statistical stability score, historical consistency score, and corresponding deduction details of the data package displayed on the dashboard; The original metadata includes the examination equipment, operators and operation records, and environmental metadata during user diagnosis and treatment. The handling and early warning log includes the specific content of the automated workflow.

7. The method for managing dry eye diagnosis and treatment data based on multidimensional fusion according to claim 1, characterized in that: The data quality optimization report in S500 is specifically as follows: Set the analysis period T f The data passports of all data packets are analyzed according to the analysis period, specifically as follows: Using quantitative analysis, with any one of the following factors remaining constant—operator, data collection timestamp, equipment unique identifier, inspection room temperature, and relative humidity—we collected m comprehensive quality scores and plotted a curve. Calculate the average of m comprehensive quality scores. When the average comprehensive quality score decreases compared to the historical average comprehensive quality score, use the historical average comprehensive quality score Q. h The average quality score Q of the same analysis item in the current period c The formula for calculating the decrease P is: P=[(Q h -Q c ) / Q h ]×100%; For the equipment dimension, an anomaly detection threshold is set as Y1; for the operational dimension, it is Y2; and for the environmental dimension, it is Y3. The requirement is that Y1... <Y2<Y3; For different types of analysis items, select an anomaly judgment threshold. When P ≥ the corresponding anomaly judgment threshold, the corresponding analysis item is judged to be abnormal. Equipment dimension anomalies are defined as level 1 anomalies, operation dimension anomalies as level 2 anomalies, and environmental dimension anomalies as level 3 anomalies. Experts set the minimum sample threshold H based on experience. m When the number of data packet samples in the current analysis period is less than the minimum sample threshold H m When the decrease magnitude P is adjusted by the sample size, the formula is: P'=P×R; R represents the correction factor, which is preset by experts; P' represents the corrected decrease; The data quality optimization report is compiled based on the results of quantitative analysis and sent to equipment maintenance and management personnel. The maintenance and management personnel maintain and process the anomaly analysis items according to the priority of the anomaly level from lowest to highest, with level 1 being the highest priority and level 3 being the lowest priority.

8. A dry eye diagnosis and treatment data management system based on multidimensional fusion, characterized in that: The dry eye diagnosis and treatment data management system includes a data acquisition module, a data quality scoring module, an automation module, a display module, and an anomaly tracking module; The data acquisition module is used to define three types of quality control metadata standards. When a user undergoes dry eye treatment, it collects all the user's treatment data and stores the three types of quality control metadata standards. It also uses an API interface to encapsulate all three types of quality control metadata of the user's treatment into a data packet. The data quality scoring module is used to calculate the rule compliance score, statistical stability score and historical consistency score of each data packet, and to build a multi-dimensional quality scoring model based on the three scores to output the comprehensive quality score of each data packet. The automated workflow module is used to preset quality thresholds and configure an automated workflow based on the quality thresholds to display the diagnosis and treatment data in the user data package to the doctor's workstation. The display module is used to construct a data passport for each data packet. When displaying user medical data, a "credibility details" icon is set. Doctors can click the icon to view the data passport of the medical data. The anomaly tracing module is used to set the analysis cycle, analyze the data passports of all data packets according to the analysis cycle, and output a data quality optimization report.

9. A dry eye diagnosis and treatment data management system based on multidimensional fusion according to claim 8, characterized in that: The data quality scoring module includes a rule compliance unit, a statistical stability unit, and a historical consistency unit. The rule compliance unit is used to predefine hard rules. The maximum rule compliance score is 100, and the rule compliance score is calculated using a deduction method. The statistical stability unit is used to acquire n test result data V when a user performs a diagnostic examination, calculate the mean and standard deviation of the n test result data respectively, calculate the coefficient of variation of the test result data, and use the coefficient of variation to calculate the statistical stability score of the data packet. The historical consistency unit is used to collect users' historical inspection result data, calculate the standard score of users' historical inspection result data, and use the standard score to calculate the historical consistency score.

10. A dry eye diagnosis and treatment data management system based on multidimensional fusion according to claim 8, characterized in that: The display module includes a data passport unit and an icon display unit; The data passport unit is used to construct the data passport for each data packet. The data passport includes: identity digest, quality sub-report, raw metadata, and processing and early warning logs. The icon display unit is used to set a "credibility details" icon when displaying the examination results data of the user data package diagnosis data in the doctor's workstation.

Images