Departmental operation support systems, methods, equipment and media for DRG disease groups

Abstract

This invention discloses a departmental operation support system, method, equipment, and medium for DRG disease groups, relating to the field of smart healthcare. It includes: a sample group construction module that matches hospital admission numbers with patient record homepages and DRG grouping data, aggregating cases by department, DRG code, and statistical period to form evaluation sample groups, and calculating two types of operational indicators: average length of stay and average cost per hospitalization; a feature construction module that combines current and historical evaluation sample groups to extract three types of model input features: case structure, DRG complexity, and historical operation; a dynamic reasonable range generation module that inputs features into a target benchmark generation model trained from historical samples, outputting the dynamic lower bound, median benchmark, and upper bound of each operational indicator through multiple sub-models; and a status evaluation module that determines the operational performance status based on the degree of deviation of indicators from the dynamic range and preset rules. The entire module avoids the benchmark distortion and inaccurate evaluation problems caused by using fixed thresholds to judge DRG operational indicators.

Description

Technical Field

[0001] This invention relates to the field of smart healthcare, and in particular to a departmental operation support system, method, equipment and medium for DRG disease groups. Background Technology

[0002] With the development of DRG (Diagnosis Related Groups) payment reform and refined hospital operation management, hospitals typically use data from the patient record front page and DRG group data to calculate the average length of stay and average cost per hospitalization in departments, and use this data to evaluate the operational status of departments.

[0003] Existing methods often use fixed target values, historical averages, year-on-year and month-on-month changes, or departmental rankings for judgment. However, the DRG disease groups, case structures, and complexity of patients admitted by the same department in different statistical periods may be significantly different. Fixed or static benchmarks are difficult to reflect the reasonable fluctuation range under the current case composition, and it is easy to misjudge the fluctuation of indicators caused by changes in case structure as operational abnormalities or performance changes.

[0004] In summary, how to avoid distortion of the evaluation benchmark and inaccurate evaluation results due to the use of fixed thresholds to evaluate the operational indicators of DRG disease groups is a problem that needs to be solved in this field. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a departmental operation support system, method, equipment, and medium for DRG disease groups, avoiding the distortion of evaluation criteria and inaccurate assessment results caused by using fixed thresholds to evaluate DRG disease group operation indicators. The specific solution is as follows: In a first aspect, this invention discloses a departmental operation support system for DRG disease groups, comprising: The sample group construction module is used to match the medical record front page data and DRG grouping data with the hospitalization number as the associated field to obtain the discharge case associated data. Within the current statistical period, the discharge cases in the discharge case associated data are merged according to the same discharge department and the same DRG group code to obtain the current evaluation sample groups under the current statistical period, and to determine the operational indicator values ​​of the current evaluation sample groups; wherein, the operational indicators are the average length of stay and the average cost per hospitalization. The feature construction module is used to generate current model input features based on the current evaluation sample group and the historical evaluation sample group under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features, and historical operational features; The dynamic reasonable range generation module is used to input the current model input features into the target benchmark generation model to generate the dynamic reasonable range of each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark and dynamic upper bound of the dynamic reasonable range; The status assessment module is used to generate the operational performance status of the current assessment sample group based on the deviation of the operational indicator value of the current assessment sample group from the dynamic reasonable range and the preset assessment rules; wherein, the deviation includes the operational indicator value being higher than the dynamic upper limit, lower than the dynamic lower limit, or located between the dynamic lower limit and the dynamic upper limit.

[0006] Optionally, the sample group construction module includes: The data acquisition unit is used to acquire medical record homepage data from the medical record management system and DRG grouping data from the DRG grouping platform or medical insurance settlement system; wherein, the medical record homepage data includes hospital number, discharge department, admission date, discharge date, patient age, major surgical procedures, length of stay and total hospitalization cost, and the DRG grouping data includes hospital number, DRG group code, DRG group name, DRG weight and grouping status; The matching unit is used to match the medical record cover page data and the DRG grouping data with the hospitalization number as the associated field to obtain the associated data of each discharged case; The merging unit is used to determine the statistical period based on the discharge date, and to merge the discharged cases in the discharged case association data according to the discharge department, the DRG group code and the discharge date, so as to obtain the current evaluation sample groups corresponding to each discharge department and each DRG disease group in the current statistical period.

[0007] Optionally, the feature construction module includes: The first feature construction unit is used to statistically analyze the case data within the current evaluation sample group to obtain case structure features; wherein, the case structure features include the total number of cases, average age, proportion of elderly patients, and proportion of surgical cases within the current evaluation sample group, the proportion of elderly patients is the ratio of the number of cases with an age not lower than a preset age threshold to the total number of cases, and the proportion of surgical cases is the ratio of the number of cases with a non-empty main surgical operation field to the total number of cases; The second feature construction unit is used to generate DRG complexity features based on the DRG weights; The third feature construction unit is used to construct historical operational features based on the historical average length of stay, the historical average length of stay standard deviation, the historical average cost per hospitalization, and the historical average cost per hospitalization for the historical evaluation sample group under the historical statistical period.

[0008] Optionally, the target benchmark generation model is a quantile regression model. The target benchmark generation model includes a first target benchmark generation model for outputting a first dynamic reasonable range of average length of stay and a second target benchmark generation model for outputting a second dynamic reasonable range of average cost per hospitalization. The first target benchmark generation model includes multiple first sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the first dynamic reasonable range, respectively. The second target benchmark generation model includes multiple second sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the second dynamic reasonable range, respectively. The training label of the first target benchmark generation model is the average length of stay, and the training label of the second target benchmark generation model is the average cost per hospitalization.

[0009] Optionally, the state assessment module includes: The correction deviation calculation unit is used to determine the correction deviation corresponding to each operational indicator of the current evaluation sample group based on the operational indicator value of the current evaluation sample group, the dynamic median benchmark of the corresponding operational indicator, and the width of the dynamic reasonable range. The deviation intensity calculation unit is used to determine the adverse deviation in the adverse direction and the improvement deviation in the improvement direction of each operational indicator according to the correction deviation corresponding to each operational indicator, and to determine the adverse deviation intensity based on the adverse deviation and the effective improvement intensity based on the improvement deviation. The quality risk calculation unit is used to acquire quality risk event data of the current assessment sample group, use Laplace smoothing, determine the current quality risk ratio based on the quality risk event data, and determine the quality risk change value based on the current quality risk ratio and the historical quality risk ratio. The status determination output unit is used to generate the operational performance status determination result of the current evaluation sample group based on the adverse deviation intensity, the effective improvement intensity, and the quality risk change value.

[0010] Optionally, the deviation intensity calculation unit includes: The deviation determination unit is used to determine the portion of the correction deviation that is greater than zero as an unfavorable deviation of the operating indicator, and to determine the absolute value of the portion of the correction deviation that is less than zero as an improvement deviation of the operating indicator. The adverse deviation intensity determination unit is used to perform weighted fusion of the adverse deviations of each operational indicator to obtain the adverse deviation intensity; The improvement deviation intensity determination unit is used to perform weighted fusion of the improvement deviations of each operational indicator, and to perform offsetting processing on the fused improvement deviations based on the adverse deviations, so as to obtain the effective improvement intensity. Accordingly, the quality risk calculation unit includes: The risk ratio determination unit is used to determine the current readmission risk ratio and the current complication risk ratio based on the number of cases readmitted within a preset time after discharge in the current assessment sample group, the number of cases with complication records, and the total number of cases, and to determine the current readmission risk ratio and the current complication risk ratio as the current quality risk ratio. The change value determination unit is used to determine the quality risk change value based on the relative change between the current quality risk ratio and the historical quality risk ratio of the same discharge department and the same DRG group code in the historical statistical period.

[0011] Optionally, the state determination output unit includes: The first state output unit is used to generate a normal fluctuation state when both the adverse deviation intensity and the effective improvement intensity are less than a preset deviation threshold; wherein, the normal fluctuation state indicates that the average length of hospital stay and the average cost per hospitalization of the current evaluation sample group are within a reasonable range. The second state output unit is used to generate a true degradation state when the intensity of the adverse deviation is not less than a preset deviation threshold; wherein, the true degradation state indicates that the average length of stay and / or the average cost per hospitalization of the current evaluation sample group is higher than the corresponding dynamic reasonable range and there is an adverse deviation in the average length of stay and / or the average cost per hospitalization; The third state output unit is used to generate a true improvement state when the adverse deviation intensity is less than a preset deviation threshold, the effective improvement intensity is not less than a preset deviation threshold, and the quality risk change value is less than a preset quality risk threshold; wherein, the true improvement state indicates that the average length of stay and / or the average cost per hospitalization of the current evaluation sample group is lower than the corresponding dynamic reasonable range and the average length of stay and / or the average cost per hospitalization has been effectively improved. The fourth state output unit is used to generate an indicator deviation state when the adverse deviation intensity is less than a preset deviation threshold, the effective improvement intensity is not less than a preset deviation threshold, and the quality risk change value is not less than a preset quality risk threshold; wherein, the indicator deviation state indicates that the average length of stay and / or the average cost per hospitalization of the current evaluation sample group is lower than the corresponding dynamic reasonable range and that there is a superficial improvement in the average length of stay and / or the average cost per hospitalization, accompanied by an increase in quality risk.

[0012] Secondly, this invention discloses a departmental operation support method for DRG disease groups, including: Using the hospital admission number as the associated field, the medical record front page data and DRG group data are matched to obtain the discharge case association data. Within the current statistical period, the discharge cases in the discharge case association data are merged according to the same discharge department and the same DRG group code to obtain the current evaluation sample groups under the current statistical period, and the operational indicator values ​​of the current evaluation sample groups are determined. Among them, the operational indicators are the average length of stay and the average cost per hospitalization. The current model input features are generated based on the current assessment sample group and the historical assessment sample group under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features, and historical operational features; The current model input features are input into the target benchmark generation model to generate the dynamic reasonable range of each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark and dynamic upper bound of the dynamic reasonable range; The operational performance status of the current evaluation sample group is generated based on the deviation of the operational indicator value of the current evaluation sample group from the dynamic reasonable range and the preset evaluation rules; wherein, the deviation includes the operational indicator value being higher than the dynamic upper limit, lower than the dynamic lower limit, or located between the dynamic lower limit and the dynamic upper limit.

[0013] Thirdly, the present invention discloses an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the aforementioned disclosed departmental operation support method for DRG disease groups.

[0014] Fourthly, the present invention discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the steps of the aforementioned disclosed departmental operation support method for DRG disease groups.

[0015] The beneficial effects of this invention are as follows: The sample group construction module of this invention is used to match the medical record front page data and DRG grouping data with the hospitalization number as the associated field to obtain the discharge case association data, and merge the discharge cases in the discharge case association data according to the same discharge department and the same DRG group code within the current statistical period to obtain each current evaluation sample group under the current statistical period, and determine the operational indicator value of the current evaluation sample group; wherein, the operational indicator is the average length of stay and the average cost per hospitalization; the feature construction module is used to generate the current model input features based on the current evaluation sample group and the historical evaluation sample groups under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features and historical operational features; the dynamic reasonable range generation module, The system is used to input the current model input features into a target benchmark generation model to generate a dynamic reasonable range for each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the dynamic reasonable range; the state evaluation module is used to generate the operational performance status of the current evaluation sample group based on the deviation of the operational indicator values ​​of the current evaluation sample group from the dynamic reasonable range and preset evaluation rules; wherein, the deviation includes operational indicator values ​​being higher than the dynamic upper bound, lower than the dynamic lower bound, or located between the dynamic lower bound and the dynamic upper bound.Therefore, this invention matches medical record front page data and DRG grouping data using the hospitalization number as the associated field, enabling data such as department information, length of stay, hospitalization costs, and DRG group codes for the same discharged case to be associated with the same data object. This provides complete and consistent discharged case association data for subsequent evaluation. Furthermore, this invention merges discharged cases according to the same discharge department and the same DRG group code within the current statistical period, forming a current evaluation sample group at the departmental DRG group level. This refines the operational performance evaluation object from the traditional overall departmental evaluation to an evaluation object of "department-DRG group-statistical period," avoiding indicator distortion caused by the mixed statistics of cases from different DRG groups. Based on this, this invention generates current model input features based on the current evaluation sample group and historical evaluation sample groups, including case structure features, DRG complexity features, and historical operational features. This ensures that the model input not only reflects the current number of cases, age structure, and surgical ratio, but also... Differences in case structure also reflect the complexity of DRG groups and the historical operational level of the same department and the same DRG group. Therefore, the target benchmark generation model can generate dynamic lower bounds, dynamic median benchmarks, and dynamic upper bounds for each operational indicator for the current assessment sample group. This ensures that the reasonable judgment benchmarks for average length of stay and average cost per hospitalization are no longer fixed thresholds, simple historical averages, or departmental rankings, but rather reasonable ranges that dynamically adjust with changes in the current case structure, DRG complexity, and historical operational level. Furthermore, the status assessment module compares the operational indicator values ​​of the current assessment sample group with the corresponding dynamic reasonable ranges, identifying deviations where operational indicator values ​​are above the dynamic upper bound, below the dynamic lower bound, or within the dynamic reasonable range. This improves the accuracy of identifying abnormal fluctuations in the department's DRG group operational indicators, reduces misjudgments caused by changes in case structure, DRG group complexity, or historical benchmark differences, and provides more objective and interpretable data for subsequent operational performance status assessments. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of a departmental operation support system for DRG disease groups disclosed in this invention; Figure 2 This is a schematic diagram of a specific departmental operation support system for DRG disease groups disclosed in this invention; Figure 3This is a flowchart of a departmental operation support method for DRG disease groups disclosed in this invention; Figure 4 This is a structural diagram of an electronic device disclosed in this invention. Detailed Implementation

[0018] 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.

[0019] With the advancement of DRG payment reform and refined hospital operation management, hospitals are paying increasing attention to operational indicators such as average length of stay and average cost per hospitalization under different DRG disease groups in departmental management. Average length of stay reflects inpatient service efficiency and bed turnover, while average cost per hospitalization reflects the consumption of medical resources and cost control. Hospitals typically conduct statistical analysis of the operation of each clinical department based on data from patient records, DRG grouping data, and performance management systems, and use the results for departmental operation management, DRG management, and performance review.

[0020] In current hospital management practices, average length of stay and average cost per hospitalization are typically assessed using fixed target values, historical averages, year-on-year and month-on-month changes, or departmental rankings. For example, if a department's average length of stay or average cost per hospitalization increases in a given month, the system might directly flag it as a decline in operational efficiency or abnormal cost control. However, the DRG group structure, patient age structure, proportion of surgical cases, and case complexity may vary significantly across different months within the same department. Judging solely based on fixed thresholds or simple fluctuations can easily lead to misinterpreting reasonable fluctuations caused by changes in case structure as departmental operational problems. Furthermore, a decrease in average length of stay or average cost per hospitalization does not necessarily indicate operational improvement. For some DRG groups, if a decrease in length of stay or cost is accompanied by an increase in 30-day readmission rates or complication rates, it may suggest that the apparent improvement in operational indicators is accompanied by an increase in medical quality risks. Existing departmental operational analysis systems often focus on indicator display, ranking, and early warning, paying insufficient attention to whether improvements in operational indicators are accompanied by changes in quality risks, making it difficult to distinguish between genuine improvement and indicator deviations.

[0021] Furthermore, existing hospital performance analysis methods typically use departments as a whole as the statistical object, easily overlooking the differences in cases between different DRG groups within the same department. For departments that treat a large number of complex cases, the lack of a dynamic and reasonable range for specific DRG groups may lead to unfair evaluation of departmental operational indicators; and for disease groups that do have high length of stay or high costs, it is also difficult to promptly locate the specific department, specific DRG group, and specific indicators.

[0022] Therefore, there is an urgent need to propose a method and system for judging the operational performance status of departmental DRG disease groups. This system should be able to generate a dynamic and reasonable range for the average length of stay and the average cost per hospitalization based on hospital medical record data, DRG grouping data, and historical operational indicators. It should also be able to determine whether the current monthly evaluation unit of the departmental DRG disease group is in a state of normal fluctuation, real decline, real improvement, or indicator deviation, based on actual indicator deviations and quality risk indicators such as readmission rate and complication rate. This would provide an auxiliary basis for hospital operational performance management and performance review.

[0023] Therefore, the present invention provides a departmental operation support scheme for DRG disease groups, which avoids the distortion of evaluation criteria and inaccurate evaluation results caused by using fixed thresholds to evaluate the operation indicators of DRG disease groups.

[0024] See Figure 1 As shown, this embodiment of the invention discloses a departmental operation support system for DRG disease groups, including: The sample group construction module 11 is used to match the medical record front page data and DRG grouping data with the hospitalization number as the associated field to obtain the discharge case associated data, and to merge the discharge cases in the discharge case associated data according to the same discharge department and the same DRG group code within the current statistical period to obtain each current evaluation sample group under the current statistical period, and to determine the operational indicator value of the current evaluation sample group; wherein, the operational indicators are the average length of stay and the average cost per hospitalization; The feature construction module 12 is used to generate current model input features based on the current evaluation sample group and the historical evaluation sample group under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features and historical operation features; The dynamic reasonable range generation module 13 is used to input the current model input features into the target benchmark generation model to generate the dynamic reasonable range of each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark and dynamic upper bound of the dynamic reasonable range respectively; The status assessment module 14 is used to generate the operational performance status of the current assessment sample group based on the deviation of the operational indicator value of the current assessment sample group from the dynamic reasonable range and the preset assessment rules; wherein, the deviation includes the operational indicator value being higher than the dynamic upper limit, lower than the dynamic lower limit, or located between the dynamic lower limit and the dynamic upper limit.

[0025] It is understood that the sample group construction module 11 includes: a data acquisition unit, used to acquire medical record homepage data from the medical record management system and DRG grouping data from the DRG grouping platform or medical insurance settlement system; wherein, the medical record homepage data includes hospital number, discharge department, admission date, discharge date, patient age, major surgical procedure, length of stay and total hospitalization cost, and the DRG grouping data includes hospital number, DRG group code, DRG group name, DRG weight and grouping status; a matching unit, used to match the medical record homepage data and the DRG grouping data with the hospital number as the association field to obtain the associated data of each discharged case; and a merging unit, used to determine the statistical period according to the discharge date and merge the discharged cases in the associated data of the discharged cases according to the discharge department, the DRG group code and the discharge date to obtain the current evaluation sample groups corresponding to each discharge department and each DRG disease group in the current statistical period.

[0026] The system data acquisition unit first acquires inpatient case data from the hospital's existing business systems to form the basis for subsequent operational performance status assessment. The inpatient case data includes medical record front page data and DRG grouping data.

[0027] The medical record front page data is exported from the hospital's medical record management system and includes the hospital admission number, discharge department, admission date, discharge date, patient age, major surgical procedure, length of stay, and total hospitalization cost. Among them, the hospital admission number is used to identify a hospitalization record, the discharge department is used to determine the clinical department to which the case belongs, the discharge date is used to determine the statistical month, the patient age and major surgical procedure are used to construct the case structure characteristics, and the length of stay and total hospitalization cost are used to calculate the average length of stay and average cost per hospitalization for this evaluation unit.

[0028] DRG grouping data is exported from the hospital's DRG grouping platform or medical insurance settlement system and includes the inpatient number, DRG group code, DRG group name, DRG weight, and grouping status. The DRG group code indicates the DRG group to which the case is assigned; the DRG group name describes the disease diagnosis or treatment type corresponding to the DRG group; and the grouping status determines whether the case has been successfully grouped into a DRG group. The DRG weight indicates the resource consumption level of the DRG group relative to ordinary inpatient cases; a higher weight generally indicates higher treatment complexity and resource consumption. If the DRG grouping data already contains DRG weights, the system directly reads this field. If the DRG grouping data only contains the DRG group code, the system matches the corresponding weight in the hospital's DRG weight parameter table based on the DRG group code. The DRG weight parameter table is maintained by the DRG grouping platform or medical insurance settlement system.

[0029] The system matching unit uses the hospital admission number as the associated field to match the medical record cover page data with the DRG grouping data, obtaining the discharge department, discharge month, DRG group code, DRG group name, DRG weight, patient age, major surgical procedure, length of stay, and total hospitalization cost for each discharged case. After completing the case matching, the system merging unit merges the cases according to the discharge department, DRG group code, and discharge month, constructing a monthly evaluation unit for the same department's DRG group, i.e., the current evaluation sample group, for discharged cases within the same discharge department, the same DRG group, and the same month. Each current evaluation sample group includes the department name, DRG group code, DRG group name, statistical month, DRG weight, number of cases, average length of stay, and average cost per hospitalization. The number of cases is obtained by counting the number of discharged cases within the current evaluation sample group; the average length of stay is obtained by averaging the length of stay for each case within the current evaluation sample group; and the average cost per hospitalization is obtained by averaging the total hospitalization cost for each case within the current evaluation sample group. For cases whose grouping status shows that they were not successfully enrolled, or which lack any of the key fields such as the discharge department, DRG group code, length of stay, or total hospitalization cost, the system will not include them in the current assessment sample group construction and will record them separately as cases pending review.

[0030] For example, the system merges orthopedic cases discharged in March 2026 that belong to the same hip replacement-related DRG group, forming a current assessment sample group called "Orthopedics - Hip Replacement-Related DRG Group - March 2026". The system then calculates the number of cases, average length of stay, and average cost per hospitalization for this current assessment sample group. This current assessment sample group serves as the basic data object for subsequent feature construction, dynamic benchmark generation, and operational performance status assessment.

[0031] In this embodiment, the feature construction module 12 includes: a first feature construction unit, used to statistically analyze the case data within the current evaluation sample group to obtain case structure features; wherein, the case structure features include the total number of cases, average age, proportion of elderly patients, and proportion of surgical cases within the current evaluation sample group, the proportion of elderly patients is the ratio of the number of cases with an age not lower than a preset age threshold to the total number of cases, and the proportion of surgical cases is the ratio of the number of cases with a non-empty main surgical operation field to the total number of cases; a second feature construction unit, used to generate DRG complexity features based on the DRG weights; and a third feature construction unit, used to construct historical operational features based on the historical average length of stay, historical average length of stay standard deviation, historical average cost per hospitalization, and historical average cost per hospitalization standard deviation of the historical evaluation sample group under historical statistical periods.

[0032] After obtaining the current assessment sample group, the system statistically summarizes the case data within the same current assessment sample group and constructs input features for the dynamic benchmark model. The input features include case structure features, DRG complexity features, and historical operational features.

[0033] The first feature construction unit statistically analyzes the case data within the current assessment sample group to obtain case structure features. These features are derived from the data on the medical record front page and include the number of cases, average age, proportion of elderly patients, and proportion of surgical cases. Specifically, the number of cases is the total number of discharged cases within the current assessment sample group; the average age is the average of the ages of all cases; the proportion of elderly patients is the percentage of cases with an age not lower than a preset age threshold (e.g., 65 years old) out of the total number of cases; and the proportion of surgical cases is the percentage of cases where the main surgical procedure field is not empty out of the total number of cases.

[0034] The second feature construction unit generates DRG complexity features based on DRG weights. Specifically, DRG complexity features are represented using DRG weights. DRG weights are provided by the hospital's DRG grouping platform or medical insurance settlement system; if the case details do not directly contain DRG weights, the system obtains them from the hospital's DRG weight parameter table based on the DRG group code. DRG weights are used to represent the resource consumption level of the DRG group relative to ordinary inpatient cases.

[0035] The third feature construction unit constructs historical operational characteristics based on the historical average length of stay, historical average length of stay standard deviation, historical average cost per hospitalization, and historical average cost per hospitalization for the historical assessment sample group under the historical statistical period. These historical operational characteristics are calculated from the assessment sample group records of the same department and the same DRG disease group for several consecutive months prior to the current month. The historical window can be set to the past 3 months or 6 months. The system calculates the average length of stay, average length of stay standard deviation, average cost per hospitalization, and average cost per hospitalization standard deviation within the historical window, respectively.

[0036] The model input features include case structure features, DRG complexity features, and historical operational features. Specifically, the model input features are: ; Where d represents the department, g represents the DRG disease group, and t represents the month of statistics; This represents the model input features generated by the d-th department in month t for the g-th DRG group. This indicates the structural characteristics of the cases, including the number of cases, average age, proportion of elderly patients, and proportion of surgical cases. Indicates DRG weights, This indicates historical operating characteristics, including the historical average length of stay, the historical average length of stay standard deviation, the historical average cost per hospitalization, and the historical average cost per hospitalization standard deviation.

[0037] In this embodiment, the target benchmark generation model is a quantile regression model. The target benchmark generation model includes a first target benchmark generation model for outputting a first dynamic reasonable range of average length of stay and a second target benchmark generation model for outputting a second dynamic reasonable range of average cost per hospitalization. The first target benchmark generation model includes multiple first sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the first dynamic reasonable range. The second target benchmark generation model includes multiple second sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the second dynamic reasonable range. The training label of the first target benchmark generation model is the average length of stay, and the training label of the second target benchmark generation model is the average cost per hospitalization.

[0038] Understandably, following the current process for generating assessment sample groups, historical assessment sample groups for historical statistical periods are generated. The average length of stay and average cost per hospitalization for these historical assessment sample groups are used as training labels, and the model input features of these historical assessment sample groups are used as model inputs. A target benchmark generation model is constructed using the quantile regression model LightGBM. Specifically, a first target benchmark generation model is constructed to output the first dynamic reasonable range of the average length of stay, and a second target benchmark generation model is constructed to output the second dynamic reasonable range of the average cost per hospitalization. The first target benchmark generation model includes multiple first sub-models used to output the dynamic lower bound, dynamic median benchmark, and dynamic upper bound within the first dynamic reasonable range. The training label for the first target benchmark generation model is the average length of stay. The second target benchmark generation model includes multiple second sub-models used to output the dynamic lower bound, dynamic median benchmark, and dynamic upper bound within the second dynamic reasonable range. The training label for the second target benchmark generation model is the average cost per hospitalization.

[0039] In other words, the system establishes six quantile regression models, including the lower quantile model of average length of stay, the median model of average length of stay, the upper quantile model of average length of stay, the lower quantile model of average cost per hospitalization, the median model of average cost per hospitalization, and the upper quantile model of average cost per hospitalization. Among them, the lower quantile model is used to output the reasonable lower bound of the operating indicators, the median model is used to output the dynamic median benchmark of the operating indicators, and the upper quantile model is used to output the reasonable upper bound of the operating indicators. In this way, the system obtains a dynamic reasonable range, rather than a single predicted value.

[0040] Each LightGBM quantile regression model consists of multiple regression trees, with the model input being the model input features. This includes case structure characteristics, DRG weights, and historical operational characteristics. During model training, historical assessment sample groups are used as training samples. For the average length of stay model, the training label is the actual average length of stay of the historical assessment sample groups; for the average cost per hospitalization model, the training label is the actual average cost per hospitalization of the historical assessment sample groups.

[0041] For any operating metric ,in The quantile regression model predicts the average length of stay or average cost per hospitalization, as follows: ; in, This represents a quantile, with values ​​of 0.1, 0.5, or 0.9. Indicates the number of regression trees. Indicates the first The regression tree in the 1st The first operational metric, the Output results at each quantile This represents the predicted value for the corresponding quantile.

[0042] After the model training is complete, the system will input the current evaluation sample group. By inputting the lower quantile model, median model, and upper quantile model respectively, operational indicators are obtained. The dynamic reasonable range: ; in, Indicates the first The department in the first Monthly targeting the first The first DRG disease group, the first The dynamic reasonable range of each operational indicator To provide a dynamically reasonable lower bound, As a dynamic median reference, It is a dynamically reasonable upper bound.

[0043] For example, for the evaluation unit "Orthopedics-Hip Replacement Related DRG Group - March 2026", the system inputs the number of cases, average age, proportion of elderly patients, proportion of surgical cases, DRG weight, and historical operational characteristics of the past 6 months into the corresponding model to obtain the dynamic reasonable range of average length of stay and the dynamic reasonable range of average cost per hospitalization.

[0044] In this embodiment, the status assessment module 14 includes: a correction deviation calculation unit 141, used to determine the correction deviation corresponding to each operational indicator of the current assessment sample group based on the operational indicator value of the current assessment sample group, the dynamic median benchmark of the corresponding operational indicator, and the width of the dynamic reasonable range; a deviation intensity calculation unit 142, used to determine the adverse deviation in the adverse direction and the improvement deviation in the improvement direction of each operational indicator based on the correction deviation corresponding to each operational indicator, and to determine the adverse deviation intensity based on the adverse deviation and the effective improvement intensity based on the improvement deviation; a quality risk calculation unit 143, used to acquire the quality risk event data of the current assessment sample group, use Laplace smoothing, and determine the current quality risk ratio based on the quality risk event data, and determine the quality risk change value based on the current quality risk ratio and the historical quality risk ratio; and a status judgment output unit 144, used to generate the operational performance status judgment result of the current assessment sample group based on the adverse deviation intensity, the effective improvement intensity, and the quality risk change value.

[0045] The status assessment module 14 includes a correction deviation calculation unit 141, a deviation intensity calculation unit 142, a quality risk calculation unit 143, and a status determination output unit 144.

[0046] The correction deviation calculation unit 141 determines the correction deviation for each operational indicator in the current evaluation sample group based on the operational indicator values ​​of the current evaluation sample group, the dynamic median benchmark of the corresponding operational indicator, and the width of the dynamic reasonable range. For any operational indicator... The system records the actual index values ​​of the current evaluation sample group as follows: And using the dynamic median benchmark and the dynamic reasonable range width output by the model as references, the correction deviation value is calculated: ; in, Indicates the first The department in the first Monthly targeting the first The first DRG disease group, the first Corrected deviation values ​​for each operational indicator This indicates the actual operational indicator values ​​of the evaluation sample group. To prevent extremely small constants with a denominator of zero.

[0047] The correction deviation value is used to indicate the degree of deviation of actual operating indicators from the dynamic baseline. For average length of stay and average cost per hospitalization, excessively high values ​​usually indicate longer hospital stays or increased costs. A value significantly greater than 0 indicates that the actual indicator is higher than the dynamic median benchmark, indicating an unfavorable deviation; when When it is close to 0, it indicates that the actual indicator is close to the dynamic benchmark; when A value less than 0 indicates that the actual indicator is lower than the dynamic median benchmark.

[0048] The system further combines dynamic reasonable range to determine whether the actual indicator exceeds the reasonable range. When the actual indicator is within... to When the actual indicator is between [a certain value] and [another value], the system marks it as an indicator within the baseline range; when the actual indicator is higher than [a certain value], the system marks it as an indicator within the baseline range. When the actual index is below the upper limit of the dynamic benchmark, the system marks it as being above the upper limit; when the actual index is below the upper limit... When this happens, the system marks it as below the lower bound of the dynamic baseline.

[0049] In this embodiment, operational metrics This includes the average length of stay and the average cost per hospitalization. To facilitate subsequent status determination, the system expresses the correction deviations for the average length of stay and the average cost per hospitalization as follows: ; ; For example, if the actual average length of stay for a certain evaluation unit is 9.0 days, and the dynamic reasonable range for the average length of stay generated by S3 is 7.6 to 8.5 days with a median baseline of 8.0 days, then the system determines that the actual average length of stay for this evaluation unit is higher than the upper limit of the dynamic baseline, and calculates the corresponding corrected deviation value for the average length of stay. This corrected deviation value serves as the basis for subsequent calculations of the adverse deviation intensity and the effective improvement intensity.

[0050] The deviation intensity calculation unit 142 relies on the correction deviation corresponding to each operational indicator to extract the adverse deviation component and the improvement deviation component of the indicator respectively, and then quantifies and calculates two discrimination parameters, adverse deviation intensity and effective improvement intensity, through the two types of deviation components respectively.

[0051] The quality risk calculation unit 143 retrieves the quality risk event data of the current assessment sample group, uses the Laplace smoothing algorithm to calculate the current quality risk ratio, and then combines it with the historical quality risk ratio in the same dimension to calculate the quality risk change value that can reflect the magnitude of risk rise and fall.

[0052] The status determination output unit 144 integrates the three quantitative parameters obtained from the aforementioned calculations: adverse deviation intensity, effective improvement intensity, and quality risk change value, to comprehensively determine the operational performance status of the current evaluation sample group and output the corresponding determination result.

[0053] The beneficial effects of this invention are as follows: The sample group construction module of this invention is used to match the medical record front page data and DRG grouping data with the hospitalization number as the associated field to obtain the discharge case association data, and merge the discharge cases in the discharge case association data according to the same discharge department and the same DRG group code within the current statistical period to obtain each current evaluation sample group under the current statistical period, and determine the operational indicator value of the current evaluation sample group; wherein, the operational indicator is the average length of stay and the average cost per hospitalization; the feature construction module is used to generate the current model input features based on the current evaluation sample group and the historical evaluation sample groups under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features and historical operational features; the dynamic reasonable range generation module, The system is used to input the current model input features into a target benchmark generation model to generate a dynamic reasonable range for each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the dynamic reasonable range; the state evaluation module is used to generate the operational performance status of the current evaluation sample group based on the deviation of the operational indicator values ​​of the current evaluation sample group from the dynamic reasonable range and preset evaluation rules; wherein, the deviation includes operational indicator values ​​being higher than the dynamic upper bound, lower than the dynamic lower bound, or located between the dynamic lower bound and the dynamic upper bound.Therefore, this invention matches medical record front page data and DRG grouping data using the hospitalization number as the associated field, enabling data such as department information, length of stay, hospitalization costs, and DRG group codes for the same discharged case to be associated with the same data object. This provides complete and consistent discharged case association data for subsequent evaluation. Furthermore, this invention merges discharged cases according to the same discharge department and the same DRG group code within the current statistical period, forming a current evaluation sample group at the departmental DRG group level. This refines the operational performance evaluation object from the traditional overall departmental evaluation to an evaluation object of "department-DRG group-statistical period," avoiding indicator distortion caused by the mixed statistics of cases from different DRG groups. Based on this, this invention generates current model input features based on the current evaluation sample group and historical evaluation sample groups, including case structure features, DRG complexity features, and historical operational features. This ensures that the model input not only reflects the current number of cases, age structure, and surgical ratio, but also... Differences in case structure also reflect the complexity of DRG groups and the historical operational level of the same department and the same DRG group. Therefore, the target benchmark generation model can generate dynamic lower bounds, dynamic median benchmarks, and dynamic upper bounds for each operational indicator for the current assessment sample group. This ensures that the reasonable judgment benchmarks for average length of stay and average cost per hospitalization are no longer fixed thresholds, simple historical averages, or departmental rankings, but rather reasonable ranges that dynamically adjust with changes in the current case structure, DRG complexity, and historical operational level. Furthermore, the status assessment module compares the operational indicator values ​​of the current assessment sample group with the corresponding dynamic reasonable ranges, identifying deviations where operational indicator values ​​are above the dynamic upper bound, below the dynamic lower bound, or within the dynamic reasonable range. This improves the accuracy of identifying abnormal fluctuations in the department's DRG group operational indicators, reduces misjudgments caused by changes in case structure, DRG group complexity, or historical benchmark differences, and provides more objective and interpretable data for subsequent operational performance status assessments.

[0054] Reference Figure 2 As shown, this embodiment of the invention discloses a specific departmental operation support system module framework for DRG disease groups. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution. Specifically: The deviation intensity calculation unit 142 includes: a deviation determination unit 1421, used to determine the portion of the correction deviation that is greater than zero as the adverse deviation of the operating indicator, and to determine the absolute value of the portion of the correction deviation that is less than zero as the improvement deviation of the operating indicator; an adverse deviation intensity determination unit 1422, used to perform weighted fusion of the adverse deviations of each operating indicator to obtain the adverse deviation intensity; and an improvement deviation intensity determination unit 1423, used to perform weighted fusion of the improvement deviations of each operating indicator, and to perform offsetting processing on the fused improvement deviations based on the adverse deviations to obtain the effective improvement intensity.

[0055] The quality risk calculation unit 143 includes: a risk ratio determination unit 1431, used to determine the current readmission risk ratio and the current complication risk ratio based on the number of cases readmitted within a preset time after discharge in the current assessment sample group, the number of cases with complication records, and the total number of cases, and to determine the current readmission risk ratio and the current complication risk ratio as the current quality risk ratio; and a change value determination unit 1432, used to determine the quality risk change value based on the relative change between the current quality risk ratio and the historical quality risk ratio of the same discharge department and the same DRG group code in the historical statistical period.

[0056] The state determination output unit 144 includes: a first state output unit 1441, used to generate a normal fluctuation state when both the adverse deviation intensity and the effective improvement intensity are less than a preset deviation threshold; wherein the normal fluctuation state indicates that the average length of stay and average cost per hospitalization of the current assessment sample group are within a reasonable fluctuation range; a second state output unit 1442, used to generate a true degradation state when the adverse deviation intensity is not less than a preset deviation threshold; wherein the true degradation state indicates that the average length of stay and / or average cost per hospitalization of the current assessment sample group are higher than the corresponding dynamic reasonable range and the average length of stay and / or average cost per hospitalization have an adverse deviation; and a third state output unit 1443, used to generate a true degradation state when the adverse deviation intensity is less than a preset deviation threshold and the effective improvement intensity is not less than a preset deviation threshold. When the quality risk change value is less than the preset quality risk threshold, a true improvement state is generated; wherein, the true improvement state indicates that the average length of stay and / or average cost per hospitalization of the current assessment sample group is lower than the corresponding dynamic reasonable range and the average length of stay and / or average cost per hospitalization has effectively improved; the fourth state output unit 1444 is used to generate an indicator deviation state when the adverse deviation intensity is less than the preset deviation threshold, the effective improvement intensity is not less than the preset deviation threshold, and the quality risk change value is not less than the preset quality risk threshold; wherein, the indicator deviation state indicates that the average length of stay and / or average cost per hospitalization of the current assessment sample group is lower than the corresponding dynamic reasonable range and the average length of stay and / or average cost per hospitalization has superficially improved but is accompanied by an increase in quality risk.

[0057] The deviation intensity calculation unit 142 includes a deviation determination unit 1421, an adverse deviation intensity determination unit 1422, and an improved deviation intensity determination unit 1423.

[0058] The system obtains the mean length of stay correction deviation for the current evaluation unit. Correction deviation of average cost per inpatient stay .in, This indicates the normalized deviation of the actual average length of stay in the current evaluation unit from the dynamic baseline of average length of stay. This indicates the normalized deviation of the actual average cost per hospitalization in the current evaluation unit from the dynamic baseline for average cost per hospitalization. Since both the average length of stay and the average cost per hospitalization are indicators that are detrimental to departmental operational performance when their values ​​are too high, a positive correction deviation indicates an unfavorable deviation, while a negative correction deviation indicates an improvement relative to the dynamic baseline. That is, the deviation determination unit 1421 determines the portion of the correction deviation that is greater than zero as the unfavorable deviation of the operational indicator, and determines the absolute value of the portion of the correction deviation that is less than zero as the improvement deviation of the operational indicator.

[0059] The adverse deviation intensity determination unit 1422 performs weighted fusion of adverse deviations of various operational indicators to obtain the adverse deviation intensity. The system first extracts the deviation information of average length of stay and average cost per hospitalization in the adverse direction, and calculates the adverse deviation intensity according to the preset indicator weights: ; in, This indicates the intensity of the adverse bias in the current evaluation unit. This indicates an unfavorable bias that occurs when the average length of hospital stay exceeds the dynamic baseline. This indicates an unfavorable bias that occurs when the average cost per hospitalization is higher than the dynamic baseline. and These represent the weights of average length of stay and average cost per hospitalization in the assessment of operational performance status, respectively, and satisfy the following conditions: In this embodiment, the average length of stay and the average cost per hospitalization are both used as core operational indicators in the status determination, and the weight of each is set to 0.5.

[0060] The improvement deviation intensity determination unit 1423 performs weighted fusion of the improvement deviations of each operational indicator, and performs offsetting processing on the fused improvement deviations based on unfavorable deviations to obtain the effective improvement intensity; the system further extracts the deviation information of average length of stay and average cost per hospitalization in the direction of improvement, and calculates the favorable improvement intensity according to the same weight: ; in, This indicates that the strength has been effectively improved. This indicates an improvement bias resulting from the average length of hospital stay being lower than the dynamic baseline. This indicates an improvement bias resulting from the average cost per hospitalization being lower than the dynamic baseline. This represents the offsetting factor between adverse deviations and the strength of improvement, set to 0.6. It is used to reduce adverse deviations in the opposite direction while preserving the improvement signal. Through this method, the system extracts deviations from two different operational indicators—average length of stay and average cost per hospitalization—and merges them according to hospital management weights, avoiding a simple direct comparison of the two different indicators.

[0061] The quality risk calculation unit 143 includes a risk ratio determination unit 1431 and a change value determination unit 1432.

[0062] The risk ratio determination unit 143 determines the current readmission risk ratio and the current complication risk ratio based on the number of cases readmitted within a preset time after discharge, the number of cases with recorded complications, and the total number of cases in the current assessment sample group. These current readmission risk ratio and current complication risk ratio are then defined as the current quality risk ratio. The system further acquires quality risk indicators and calculates the quality risk change value. Quality risk indicators include the 30-day readmission rate and the complication rate. The 30-day readmission rate is obtained by dividing the number of cases readmitted within 30 days of discharge into the total number of discharged cases in the assessment unit, and the complication rate is obtained by dividing the number of cases with recorded complications into the total number of discharged cases in the assessment unit. To reduce the impact of individual cases in a small sample assessment group on the quality risk ratio, the system uses Laplace smoothing to calculate the current quality risk ratio. ; ; in, This indicates the total number of discharged cases within the current assessment sample group. This indicates the number of cases that were readmitted within 30 days of discharge. This indicates the number of cases with recorded complications.

[0063] The change value determination unit 1432 determines the quality risk change value based on the relative change between the current quality risk ratio and the historical quality risk ratio of the same discharge department and the same DRG group code within the historical statistical period. The system further uses the average 30-day readmission rate of the same department and the same DRG group within the historical window. and the mean complication rate Calculate the change in quality risk: ; in, Indicates the change value of quality risk. and These represent the mean 30-day readmission rate and the mean complication rate for the same department and the same DRG group within the historical window, respectively. Both are obtained by averaging the corresponding indicators for each evaluation period within the historical window. An index of 0.6 is used to improve the impact of the 30-day readmission rate on the change value of quality risk, while an index of 0.4 is used to preserve the constraint effect of the complication rate on the safety risk of the hospitalization process.

[0064] In this embodiment, the system sets a quality risk threshold. This is used to determine whether the current quality risk exceeds historical normal levels. When this indicates that the current quality risk is not significantly higher than historical levels; when When this occurs, it indicates that the current quality risk has increased beyond the permissible range compared to historical levels.

[0065] The state determination output unit 144 includes a first state output unit 1441, a second state output unit 1442, a third state output unit 1443, and a fourth state output unit 1444. This is based on the intensity of the adverse deviation. Effectively improves strength and quality risk change value The system performs a final status determination on the current department's DRG monthly evaluation unit (i.e., the current assessment sample group) and generates readable output results for hospital administrators. The system has a preset deviation threshold. and quality risk threshold ,in, Used to determine whether operational metrics have deviated significantly from dynamic benchmarks. Used to determine whether the quality risk exceeds the allowable range. In this embodiment, Set it to 0.8. Set it to 1.10.

[0066] The first state output unit 1441 outputs when both the adverse deviation intensity and the effective improvement intensity are less than a preset deviation threshold. When this occurs, a normal fluctuation state is generated; where normal fluctuation state indicates that the average length of stay and average cost per hospitalization in the current assessment sample group are within reasonable fluctuations. In other words, when and When the system determines the current evaluation unit to be in a normal fluctuation state, it outputs "The current average length of stay and average cost per hospitalization have not deviated significantly from the dynamic reasonable range".

[0067] The second state output unit 1442 generates a true degradation state when the intensity of the adverse deviation is not less than a preset deviation threshold; wherein, the true degradation state indicates that the average length of stay and / or the average cost per hospitalization in the current assessment sample group is higher than the corresponding dynamic reasonable range and that there is an adverse deviation in the average length of stay and / or the average cost per hospitalization. In other words, when When this happens, the system will determine that the current evaluation unit is in a true state of degradation and output "The current operating indicators are higher than the dynamic reasonable level, indicating that there is a decline in hospitalization efficiency or insufficient cost control".

[0068] The third state output unit 1443 generates a true improvement state when the adverse deviation intensity is less than a preset deviation threshold, the effective improvement intensity is not less than a preset deviation threshold, and the quality risk change value is less than a preset quality risk threshold. The true improvement state indicates that the average length of stay and / or average cost per hospitalization in the current assessment sample group is lower than the corresponding dynamic reasonable range, and that there is effective improvement in the average length of stay and / or average cost per hospitalization. In other words, when... and When the system determines that the current evaluation unit is in a state of genuine improvement, it outputs "The current operating indicators are below the dynamic reasonable level, and the quality risk has not increased significantly."

[0069] The fourth state output unit 1444 is used to generate an indicator deviation state when the adverse deviation intensity is less than a preset deviation threshold, the effective improvement intensity is not less than a preset deviation threshold, and the quality risk change value is not less than a preset quality risk threshold. The indicator deviation state indicates that the average length of stay and / or average cost per hospitalization in the current assessment sample group is lower than the corresponding dynamic reasonable range, and that there is superficial improvement in the average length of stay and / or average cost per hospitalization accompanied by an increase in quality risk. In other words, when... and When this happens, the system will determine that the current evaluation unit is in a state of indicator deviation and output "Although the current operating indicators have improved, the quality risk has increased simultaneously, and it is not appropriate to directly identify it as a performance improvement".

[0070] The system further generates key problem manifestations based on the operational performance status assessment results. If the current assessment sample group is determined to be in a true deterioration state, the system compares the mean length of stay with the corrected deviation. Correction deviation of average cost per inpatient stay If the deviation in the average length of stay is large, the main problem generated will be "the average length of stay is too high"; if the deviation in the average cost per hospitalization is large, the main problem generated will be "the average cost per hospitalization is too high"; if both reach the deviation threshold, the main problem generated will be "both the average length of stay and the average cost per hospitalization are too high".

[0071] If the current evaluation sample group is determined to be in a state of true improvement, the system will determine the effective improvement intensity. The main problem manifestations are generated based on the sources of deviations in the average length of stay and average cost per hospitalization. If the average length of stay is significantly lower than the dynamic benchmark, the main problem manifestation is "The average length of stay is lower than the dynamic reasonable range"; if the average cost per hospitalization is significantly lower than the dynamic benchmark, the main problem manifestation is "The average cost per hospitalization is lower than the dynamic reasonable range"; if both are lower than the dynamic benchmark and reach the deviation threshold, the main problem manifestation is "Both the average length of stay and the average cost per hospitalization are lower than the dynamic reasonable range".

[0072] If the current evaluation sample group is determined to be in a state of indicator deviation, the system will adjust the quality risk change value accordingly. The corresponding quality risk indicators generate the main problem manifestations. If the 30-day readmission rate is more significantly higher than the historical level, the main problem manifestation is "increased 30-day readmission rate"; if the complication rate is more significantly higher than the historical level, the main problem manifestation is "increased complication rate"; if both the 30-day readmission rate and the complication rate are higher than the historical level, the main problem manifestation is "both 30-day readmission rate and complication rate are increased".

[0073] If the current evaluation sample group is determined to be in a normal fluctuation state, the main problem generated by the system is that "the average length of stay and the average cost per hospitalization have not deviated significantly from the dynamic reasonable range".

[0074] The system ultimately outputs readable results for the current assessment sample group. These results include the department name, DRG group name, statistical month, operational performance status type, and main problems. For example, the system might output: "Orthopedics - Hip Replacement Related DRG Group - March 2026, Operational performance status is true deterioration, mainly manifested as a high average length of stay"; or: "Orthopedics - Hip Replacement Related DRG Group - March 2026, Operational performance status is indicator deviation, mainly manifested as an increased 30-day readmission rate."

[0075] This invention targets the average length of stay and average cost per hospitalization for specific departments and corresponding DRG disease groups. Based on case structure characteristics, DRG weights, and historical operational characteristics, it generates a dynamically reasonable range for operational indicators, encompassing a dynamic reasonable lower bound, a dynamic median benchmark, and a dynamic reasonable upper bound. This constructs a reference range for indicators for the corresponding departments and DRG disease groups. The invention matches the actual average length of stay and actual average cost per hospitalization for the current assessment sample group to the corresponding dynamically reasonable range. It calculates the correction deviation of these two types of indicators based on the dynamic median benchmark and the width of the dynamically reasonable range, quantifying the normalized deviation of actual operational indicators from the dynamically reasonable range. The invention solves for the intensity of adverse deviation and the intensity of effective improvement through the correction deviation of these two types of indicators. Simultaneously, it calculates the quality risk change value by combining the 30-day readmission rate and complication rate, thereby distinguishing the current assessment sample group into four states: normal fluctuation, true decline, true improvement, and indicator deviation. It outputs the department name, DRG disease group, statistical month, state type, and core issue, providing auxiliary judgment criteria for hospital operation management and performance review.

[0076] The system automatically matches heterogeneous data, namely medical record homepage and DRG grouping, using a sample group construction module. It aggregates cases according to department, DRG, and statistical period to generate evaluation sample groups, eliminating the need for manual case screening and summarization, effectively reducing manpower input in medical record data processing. Simultaneously, the feature construction module automatically extracts multi-dimensional features of case structure, DRG complexity, and historical operations. Through a trained multi-sub-model, it outputs dynamic and reasonable ranges corresponding to operational indicators, eliminating the need for manually set fixed evaluation thresholds and reducing evaluation bias caused by fluctuations in case structure. All modules work together to complete all calculations, with all data calculations and status evaluations automatically run by the server, forming a standardized computer-aided judgment process. This is significantly different from manually subjective performance evaluation rules. The system can automatically output deviations of operational indicators for different DRG disease groups, providing hospital administrators with objective and quantifiable data for reference, assisting hospitals in achieving intelligent auxiliary supervision of DRG department operations.

[0077] See Figure 3 As shown, this embodiment of the invention discloses a departmental operation support method for DRG disease groups, including: Step S11: Match the medical record cover page data and DRG group data with the hospitalization number as the associated field to obtain the discharge case associated data. Within the current statistical period, merge the discharge cases in the discharge case associated data according to the same discharge department and the same DRG group code to obtain each current evaluation sample group under the current statistical period, and determine the operational indicator value of the current evaluation sample group; wherein, the operational indicator is the average length of stay and the average cost per hospitalization.

[0078] Step S12: Generate current model input features based on the current evaluation sample group and the historical evaluation sample group under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features and historical operation features.

[0079] Step S13: Input the current model input features into the target benchmark generation model to generate the dynamic reasonable range of each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark and dynamic upper bound of the dynamic reasonable range.

[0080] Step S14: Generate the operational performance status of the current evaluation sample group based on the deviation of the operational indicator value of the current evaluation sample group from the dynamic reasonable range and the preset evaluation rules; wherein, the deviation includes the operational indicator value being higher than the dynamic upper limit, lower than the dynamic lower limit, or located between the dynamic lower limit and the dynamic upper limit.

[0081] Furthermore, embodiments of the present invention also provide an electronic device. Figure 4 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of the invention.

[0082] Figure 4 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Specifically, it may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the departmental operation support method for DRG disease groups disclosed in any of the foregoing embodiments.

[0083] In this embodiment, the power supply 23 is used to provide operating voltage for various hardware devices on the electronic device; the communication interface 24 can create a data transmission channel between the electronic device and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this invention, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0084] The processor 21 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 21 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0085] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored on it include operating system 221, computer program 222 and data 223, etc., and the storage method can be temporary storage or permanent storage.

[0086] The operating system 221 manages and controls the various hardware devices and computer programs 222 on the electronic device to enable the processor 21 to perform calculations and processing on the massive amounts of data 223 in the memory 22. The operating system can be Windows, Unix, Linux, etc. The computer program 222, in addition to including computer programs capable of performing the departmental operation support method for DRG disease groups executed by the electronic device as disclosed in any of the foregoing embodiments, may further include computer programs capable of performing other specific tasks. The data 223 may include data received by the electronic device from external devices, as well as data collected by its own input / output interface 25.

[0087] Furthermore, the present invention also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned disclosed departmental operation support method for DRG disease groups. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0088] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0089] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the invention. The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly in hardware, software modules executed by a processor, or a combination of both. The software module may be located in random access memory (RAM), memory, read-only memory (ROM), electrically programmable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), register, hard disk, removable disk, CD-ROM (Compact Disc Read-Only Memory), or any other form of storage medium known in the art.

[0090] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0091] The above provides a detailed description of a departmental operation support method, device, equipment, and medium for DRG disease groups provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only intended to help understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A departmental operation support system for DRG disease groups, characterized in that, include: The sample group construction module is used to match the medical record front page data and DRG grouping data with the hospitalization number as the associated field to obtain the discharge case associated data. Within the current statistical period, the discharge cases in the discharge case associated data are merged according to the same discharge department and the same DRG group code to obtain the current evaluation sample groups under the current statistical period, and to determine the operational indicator values ​​of the current evaluation sample groups; wherein, the operational indicators are the average length of stay and the average cost per hospitalization. The feature construction module is used to generate current model input features based on the current evaluation sample group and the historical evaluation sample group under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features, and historical operational features; The dynamic reasonable range generation module is used to input the current model input features into the target benchmark generation model to generate the dynamic reasonable range of each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark and dynamic upper bound of the dynamic reasonable range; The status assessment module is used to generate the operational performance status of the current assessment sample group based on the deviation of the operational indicator value of the current assessment sample group from the dynamic reasonable range and the preset assessment rules; wherein, the deviation includes the operational indicator value being higher than the dynamic upper limit, lower than the dynamic lower limit, or located between the dynamic lower limit and the dynamic upper limit.

2. The departmental operation support system for DRG disease groups according to claim 1, characterized in that, The sample group construction module includes: The data acquisition unit is used to acquire medical record homepage data from the medical record management system and DRG grouping data from the DRG grouping platform or medical insurance settlement system; wherein, the medical record homepage data includes hospital number, discharge department, admission date, discharge date, patient age, major surgical procedures, length of stay and total hospitalization cost, and the DRG grouping data includes hospital number, DRG group code, DRG group name, DRG weight and grouping status; The matching unit is used to match the medical record cover page data and the DRG grouping data with the hospitalization number as the associated field to obtain the associated data of each discharged case; The merging unit is used to determine the statistical period based on the discharge date, and to merge the discharged cases in the discharged case association data according to the discharge department, the DRG group code and the discharge date, so as to obtain the current evaluation sample groups corresponding to each discharge department and each DRG disease group in the current statistical period.

3. The departmental operation support system for DRG disease groups according to claim 2, characterized in that, The feature construction module includes: The first feature construction unit is used to statistically analyze the case data within the current evaluation sample group to obtain case structure features; wherein, the case structure features include the total number of cases, average age, proportion of elderly patients, and proportion of surgical cases within the current evaluation sample group, the proportion of elderly patients is the ratio of the number of cases with an age not lower than a preset age threshold to the total number of cases, and the proportion of surgical cases is the ratio of the number of cases with a non-empty main surgical operation field to the total number of cases; The second feature construction unit is used to generate DRG complexity features based on the DRG weights; The third feature construction unit is used to construct historical operational features based on the historical average length of stay, the historical average length of stay standard deviation, the historical average cost per hospitalization, and the historical average cost per hospitalization for the historical evaluation sample group under the historical statistical period.

4. The departmental operation support system for DRG disease groups according to any one of claims 1 to 3, characterized in that, The target benchmark generation model is a quantile regression model. The target benchmark generation model includes a first target benchmark generation model for outputting a first dynamic reasonable range of average length of stay and a second target benchmark generation model for outputting a second dynamic reasonable range of average cost per hospitalization. The first target benchmark generation model includes multiple first sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the first dynamic reasonable range. The second target benchmark generation model includes multiple second sub-models for outputting the dynamic lower bound, dynamic median benchmark, and dynamic upper bound of the second dynamic reasonable range. The training label of the first target benchmark generation model is the average length of stay, and the training label of the second target benchmark generation model is the average cost per hospitalization.

5. The departmental operation support system for DRG disease groups according to claim 4, characterized in that, The status assessment module includes: The correction deviation calculation unit is used to determine the correction deviation corresponding to each operational indicator of the current evaluation sample group based on the operational indicator value of the current evaluation sample group, the dynamic median benchmark of the corresponding operational indicator, and the width of the dynamic reasonable range. The deviation intensity calculation unit is used to determine the adverse deviation in the adverse direction and the improvement deviation in the improvement direction of each operational indicator according to the correction deviation corresponding to each operational indicator, and to determine the adverse deviation intensity based on the adverse deviation and the effective improvement intensity based on the improvement deviation. The quality risk calculation unit is used to acquire quality risk event data of the current assessment sample group, use Laplace smoothing, determine the current quality risk ratio based on the quality risk event data, and determine the quality risk change value based on the current quality risk ratio and the historical quality risk ratio. The status determination output unit is used to generate the operational performance status determination result of the current evaluation sample group based on the adverse deviation intensity, the effective improvement intensity, and the quality risk change value.

6. The departmental operation support system for DRG disease groups according to claim 5, characterized in that, The deviation intensity calculation unit includes: The deviation determination unit is used to determine the portion of the correction deviation that is greater than zero as an unfavorable deviation of the operating indicator, and to determine the absolute value of the portion of the correction deviation that is less than zero as an improvement deviation of the operating indicator. The adverse deviation intensity determination unit is used to perform weighted fusion of the adverse deviations of each operational indicator to obtain the adverse deviation intensity; The improvement deviation intensity determination unit is used to perform weighted fusion of the improvement deviations of each operational indicator, and to perform offsetting processing on the fused improvement deviations based on the adverse deviations, so as to obtain the effective improvement intensity. Accordingly, the quality risk calculation unit includes: The risk ratio determination unit is used to determine the current readmission risk ratio and the current complication risk ratio based on the number of cases readmitted within a preset time after discharge in the current assessment sample group, the number of cases with complication records, and the total number of cases, and to determine the current readmission risk ratio and the current complication risk ratio as the current quality risk ratio. The change value determination unit is used to determine the quality risk change value based on the relative change between the current quality risk ratio and the historical quality risk ratio of the same discharge department and the same DRG group code in the historical statistical period.

7. The departmental operation support system for DRG disease groups according to claim 5, characterized in that, The state determination output unit includes: The first state output unit is used to generate a normal fluctuation state when both the adverse deviation intensity and the effective improvement intensity are less than a preset deviation threshold; wherein, the normal fluctuation state indicates that the average length of hospital stay and the average cost per hospitalization of the current evaluation sample group are within a reasonable range. The second state output unit is used to generate a true degradation state when the intensity of the adverse deviation is not less than a preset deviation threshold; wherein, the true degradation state indicates that the average length of stay and / or the average cost per hospitalization of the current evaluation sample group is higher than the corresponding dynamic reasonable range and there is an adverse deviation in the average length of stay and / or the average cost per hospitalization; The third state output unit is used to generate a true improvement state when the adverse deviation intensity is less than a preset deviation threshold, the effective improvement intensity is not less than a preset deviation threshold, and the quality risk change value is less than a preset quality risk threshold; wherein, the true improvement state indicates that the average length of stay and / or the average cost per hospitalization of the current evaluation sample group is lower than the corresponding dynamic reasonable range and the average length of stay and / or the average cost per hospitalization has been effectively improved. The fourth state output unit is used to generate an indicator deviation state when the adverse deviation intensity is less than a preset deviation threshold, the effective improvement intensity is not less than a preset deviation threshold, and the quality risk change value is not less than a preset quality risk threshold; wherein, the indicator deviation state indicates that the average length of stay and / or the average cost per hospitalization of the current evaluation sample group is lower than the corresponding dynamic reasonable range and that there is a superficial improvement in the average length of stay and / or the average cost per hospitalization, accompanied by an increase in quality risk.

8. A departmental operation support method for DRG disease groups, characterized in that, include: Using the hospital admission number as the associated field, the medical record front page data and DRG group data are matched to obtain the discharge case association data. Within the current statistical period, the discharge cases in the discharge case association data are merged according to the same discharge department and the same DRG group code to obtain the current evaluation sample groups under the current statistical period, and the operational indicator values ​​of the current evaluation sample groups are determined. Among them, the operational indicators are the average length of stay and the average cost per hospitalization. The current model input features are generated based on the current assessment sample group and the historical assessment sample group under the historical statistical period; wherein, the model input features include case structure features, DRG complexity features, and historical operational features; The current model input features are input into the target benchmark generation model to generate the dynamic reasonable range of each operational indicator of the current evaluation sample group; wherein, the training samples of the target benchmark generation model include the model input features and operational indicator values ​​of the historical evaluation sample group, and the target benchmark generation model includes multiple sub-models for outputting the dynamic lower bound, dynamic median benchmark and dynamic upper bound of the dynamic reasonable range; The operational performance status of the current evaluation sample group is generated based on the deviation of the operational indicator value of the current evaluation sample group from the dynamic reasonable range and the preset evaluation rules; wherein, the deviation includes the operational indicator value being higher than the dynamic upper limit, lower than the dynamic lower limit, or located between the dynamic lower limit and the dynamic upper limit.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the departmental operation support method for DRG disease groups as described in claim 8.

10. A computer-readable storage medium, characterized in that, Used to store computer programs; wherein, when the computer programs are executed by a processor, they implement the steps of the departmental operation support method for DRG disease groups as described in claim 8.

Images