DIP home page coding optimization recommendation modeling visualization method and system
Through DIP homepage coding optimization recommendation modeling method, an intuitive electronic medical record visual summary is generated, which solves the problem of large deviations from logic and actuality in the existing technology, and achieves more accurate recommendation and intuitive display effects.
Patent Information
- Application Number
- CN202411126200.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing electronic medical record visual summary generation method cannot effectively reduce recommended medical projects with large deviations from logic and actuality. The content is not in line with reality and is not intuitive in the display of color and illustration size.
Using DIP homepage coding optimization recommended modeling method, through data processing, data conversion, vectorization processing, color processing and correlation graph visualization, Pearson correlation coefficient and Euclidean distance are calculated, and multi-column conversion method and clustering algorithm are used to generate intuitive visualization. Relationship diagram, representing the pros and cons of the recommended item by color and size.
Effectively reduce recommended items with better logic but greater actual deviation, and the content is more accurate and consistent with reality. Through the display of color and size, users can quickly judge more preferred items, and improve the speed and accuracy of recommendation calculations.
Smart Images

Figure CN119065654A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical record visualization, and in particular to a DIP homepage coding optimization recommendation modeling visualization method and system. Background Art
[0002] With the development of medical information technology, electronic medical records have gradually replaced paper medical records as an important document for recording patients' conditions, examination results and treatment processes. During a disease, a large number of electronic medical record forms are often generated. As patients age, their medical record information continues to expand. In order to extract a certain content, doctors often spend a lot of time on querying and searching, which reduces the efficiency of diagnosis and treatment.
[0003] Medical record visualization technology is developing rapidly. The electronic medical record summary is designed to describe a patient's condition as a whole, making it convenient for doctors to quickly and accurately view and even retrieve the overall status of the patient's diagnosis, consultation, and treatment for a certain patient or a certain medical event.
[0004] Existing patents such as a method and device for generating a visual summary of an electronic medical record with application number CN201610800083.1 belong to the field of medical informationization. The method includes: obtaining the electronic medical record file of the patient; parsing the content of the electronic medical record file; dividing at least one physiological period according to the parsed content; and loading a human body model corresponding to at least one physiological period to generate a visual summary of the electronic medical record.
[0005] Although the above patent provides a visual summary generation method and device for electronic medical records, which can help doctors quickly obtain the required electronic medical record information, it cannot effectively reduce the logic when displaying recommendations, and the recommended medical items with large deviations from the actual content are not in line with the actual situation. It is impossible to display the screen color and icon size, ignoring the more intuitive understanding. Summary of the invention
[0006] The present invention aims to solve the above technical problems and provides a DIP homepage coding optimization recommendation modeling visualization method and system.
[0007] In order to solve the above technical problems, the technical solution provided by the present invention is:
[0008] A DIP homepage coding optimization recommendation modeling visualization method includes the following steps:
[0009] S1. Data processing:
[0010] Obtain medical records and pass the information of medical records to the model, use the DIP grouping tool to pre-calculate the DIP score of each document, and obtain the parameter vector and weight required for the recommendation calculation;
[0011] S2. Data conversion:
[0012] The surgical procedures in the surgical record sheet, the main surgical code on the front page of the medical record and the probability of co-occurrence were used as factors to calculate their Pearson correlation coefficients as the basic surgical correlation parameter set P for each combination. The main diagnosis and main surgical code were converted into numerical three-dimensional vector data using a multi-column conversion method.
[0013] S3, vectorized processing:
[0014] The actual total cost of the case, drug costs, consumables costs, examination costs, and DIP scores are quantified using the following formula: The Euclidean distance is calculated by the following formula to compare the difference between the actual case V and the model d,
[0015] The weighted product of each vector that meets the conditions is obtained by the following formula and recorded as set B. B will be used as the parameter used for color mapping.
[0016]
[0017] S4. Color processing:
[0018] Sort the elements in set B from large to small, select the one closest to 0 as the benchmark, and select the adjacent positive and negative n ones as candidate suggestions to be displayed
[0019] S5. Visualization of association graph:
[0020] The diagnosis code and surgical code corresponding to each specific element in set D are processed to obtain the diagnosis code and surgical code sub-items and establish associations. Then the image is expanded with the four-digit surgical sub-items as the core, with the specific surgical code as the first-level connection, and the code and name are displayed. For each surgical code, the four-digit diagnosis sub-item code is derived and displayed. The size indicates the DIP score of the group. The larger the value, the higher the score. The color is determined by the data of the corresponding element in set B, indicating its proximity to the optimal target. The thickness of the connecting line indicates the number of times this combination appears in the previous medical records. The larger the number, the thicker the line, and a visual relationship diagram is obtained.
[0021] S5. Visual display:
[0022] The visual relationship diagram can be viewed and operated through the terminal.
[0023] Preferably, the weight is to take the main diagnosis code, main surgery code and code that can be used as MCC / CC in the medical record as a code combination, and take the DRG-RW value and DIP score of the code combination as the comprehensive maximum as the optimization target, and obtain the parameters of each code combination by calculating past data as the weight parameters for the recommendation algorithm ranking.
[0024] The weight function is
[0025] Preferably, wherein x is a data point in the window, m is the current cluster center, h is the window radius, and ||xm|| is the Euclidean distance between point x and cluster center m.
[0026] Preferably, the weighted average of all data points in the neighborhood of the cluster center m is calculated, that is, the new cluster center m′, and the calculation formula is:
[0027] Preferably, where N(m) is the set of data points in a window centered at m, W(x) is the weight of the data point x, the cluster center m is moved to the position of the new weighted average m′, and this step is repeated iteratively until the stopping condition is met.
[0028] Preferably, each element bi in the set B is color mapped through a mapping function, and the deviation within ±0.5 is displayed in orange, above +0.5 is displayed in red, and below -0.5 is displayed in green.
[0029] A DIP homepage coding optimization recommendation modeling visualization system, including
[0030] Data acquisition module: used to obtain medical records and transmit the information of medical records to the model;
[0031] Data conversion module: converts data information into numerical three-dimensional vector data and transmits it to the model; Data analysis module: processes and analyzes the information of medical records and transmits it to the data conversion module;
[0032] Model: the parameters required for calculation;
[0033] Visualization processing module: displays data information in the form of images, text and lines;
[0034] Color processing module: The data of the visualization processing module in set B is used to determine its closeness to the optimal target, making the visualization relationship diagram output by the visualization processing module clearer;
[0035] Terminal: used to view the visual relationship diagram.
[0036] A visualization platform includes at least one memory, a processor and an input-output module. The memory is used to store computer resources, and the processor is used to call computer programs and data stored in a visualization system to execute a DIP homepage coding optimization recommendation modeling visualization method.
[0037] After adopting the above structure, the present invention has the following advantages:
[0038] The present invention optimizes the algorithm and simplifies the content, so that when displaying the recommended code, it can effectively reduce the recommended items with better logic but large actual deviation, and the content is more accurate and in line with the actual situation. At the same time, through color and size, it is convenient for users to quickly judge the better options, and the thickness of the auxiliary line provides visualization of the probability selection.
[0039] The present invention obtains the parameters such as vectors, weights, etc. required for the recommendation calculation through data modeling, and eliminates the parameters that have theoretical significance but low actual correlation, so that the recommendation calculation speed is faster and more accurate. The recommendation program uses an adjustable parameter to make the recommendation result as close to the Pareto optimal curve as possible in the direction expected by the user, and maximizes the utility. The main diagnosis code, main surgery code, and code that can be used as mcc / CC in the medical record are used as a coding combination, and the DRG-RW value and DIP score of the combination are used as the comprehensive maximum optimization target. The parameters of each coding combination are obtained by calculating the past data as the weight parameters for the recommendation algorithm sorting. With the optimization of the algorithm and the simplification of the content, when displaying the recommendation code, the recommended items with better logic but larger actual deviation can be effectively reduced, and the content is more accurate and in line with the actual situation. Through the display of the screen color and the size of the icon, it is more intuitive and understandable.
[0040] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present invention will be readily apparent by reference to the full text and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 It is a clustering effect diagram of the present invention;
[0042] Figure 2 It is a list-like display diagram of set B of the present invention;
[0043] Figure 3 It is a 3-row version list display diagram of set B of the present invention;
[0044] Figure 4 is a correlation diagram of the present invention;
[0045] Figure 5Each of the four sub-items of the present invention expands multiple relationship diagrams;
[0046] Figure 6 It is a system diagram of the present invention. DETAILED DESCRIPTION
[0047] Specific embodiments of the present invention will now be mentioned in detail. Although the present invention is described in conjunction with these specific embodiments, it should be appreciated that it is not intended to limit the present invention to these specific embodiments. On the contrary, these embodiments are intended to cover substitutions, changes or equivalent embodiments that may be included in the spirit and scope of the invention defined by the claims. In the following description, a large number of specific details are set forth in order to provide a comprehensive understanding of the present invention. The present invention can be implemented without some or all of these specific details. In other cases, in order not to make the present invention unnecessarily obscure, well-known process operations are not described in detail.
[0048] When used in conjunction with "including," "methods comprising," or similar language in this specification and the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0049] The present invention is further described in detail below in conjunction with the full text.
[0050] Combined with Figure 1 , a DIP homepage coding optimization recommendation modeling visualization method, comprising the following steps:
[0051] S1. Data processing:
[0052] Obtain medical records and pass the information of medical records to the model, use the DIP grouping tool to pre-calculate the DIP score of each document, and obtain the parameter vector and weight required for the recommendation calculation;
[0053] S2. Data conversion:
[0054] The surgical procedures in the surgical record sheet, the main surgical code on the front page of the medical record and the probability of co-occurrence were used as factors to calculate their Pearson correlation coefficients as the basic surgical correlation parameter set P for each combination. The main diagnosis and main surgical code were converted into numerical three-dimensional vector data using a multi-column conversion method.
[0055] S3, vectorized processing:
[0056] The actual total cost of the case, drug costs, consumables costs, examination costs, and DIP scores are quantified using the following formula: The Euclidean distance is calculated by the following formula to compare the difference between the actual case V and the model d,
[0057] The weighted product of each vector that meets the conditions is obtained by the following formula and recorded as set B. B will be used as the parameter used for color mapping.
[0058]
[0059] S4. Color processing:
[0060] Sort the elements in set B from large to small, select the one closest to 0 as the benchmark, and select the adjacent positive and negative n ones as candidate suggestions to be displayed
[0061] S5. Visualization of association graph:
[0062] The diagnosis code and surgical code corresponding to each specific element in set D are processed to obtain the diagnosis code and surgical code sub-items and establish associations. Then the image is expanded with the four-digit surgical sub-items as the core, with the specific surgical code as the first-level connection, and the code and name are displayed. For each surgical code, the four-digit diagnosis sub-item code is derived and displayed. The size indicates the DIP score of the group. The larger the value, the higher the score. The color is determined by the data of the corresponding element in set B, indicating its proximity to the optimal target. The thickness of the connecting line indicates the number of times this combination appears in the previous medical records. The larger the number, the thicker the line, and a visual relationship diagram is obtained.
[0063] S5. Visual display:
[0064] The visual relationship diagram can be viewed and operated through the terminal.
[0065] The weight is to take the main diagnosis code, main surgery code and code that can be used as MCC / CC in the medical record as a code combination, and take the DRG-RW value and DIP score of the code combination as the comprehensive maximum as the optimization goal. The parameters of each code combination are obtained by calculating past data as the weight parameters for the recommendation algorithm ranking.
[0066] The weight function is
[0067] Where x is the data point in the window, m is the current cluster center, h is the window radius, and ||xm|| is the Euclidean distance between point x and cluster center m.
[0068] According to the cluster center m, the weighted average of all data points in its neighborhood is calculated, that is, the new cluster center m′. The calculation formula is:
[0069] Among them, N(m) is the set of data points in the window centered on m, W(x) is the weight of the data point x, and the cluster center m is moved to the position of the new weighted average m′. This step is repeated iteratively until the stopping condition is met.
[0070] Each element bi in the set B is color mapped through a mapping function, and the deviation within ±0.5 is displayed in orange, above +0.5 is displayed in red, and below -0.5 is displayed in green.
[0071] A DIP homepage coding optimization recommendation modeling visualization system, including
[0072] Data acquisition module: used to obtain medical records and transmit the information of medical records to the model;
[0073] Data conversion module: converts data information into numerical three-dimensional vector data and transmits it to the model; Data analysis module: processes and analyzes the information of medical records and transmits it to the data conversion module;
[0074] Model: the parameters required for calculation;
[0075] Visualization processing module: displays data information in the form of images, text and lines;
[0076] Color processing module: The data of the visualization processing module in set B is used to determine its closeness to the optimal target, making the visualization relationship diagram output by the visualization processing module clearer;
[0077] Terminal: used to view the visual relationship diagram.
[0078] Modeling requires a part
[0079] Medical records, and use the DIP grouping tool to pre-calculate the DIP score for each record. The medical records select the target user's data for 1 to 3 years, and need to cover as many DIP grouping situations as possible.
[0080] The surgical procedure in the surgical record sheet, the main surgical code on the front page of the medical record, and the probability of co-occurrence were taken as factors, and their Pearson correlation coefficient was calculated as the basic surgical correlation parameter set P for each combination.
[0081] The main diagnosis and main surgery codes are converted into numerical three-dimensional vector data using a multi-column conversion method.
[0082] The specific steps are as follows:
[0083] 1. Extract the categories, subcategories, detailed codes and complete codes respectively.
[0084] 2. Each level is converted into numerical data, and the original numerical data can be used directly.
[0085] 3. Each level can be treated as a separate vector column.
[0086] Through the differences in dip scores of different surgical procedures for the same main diagnosis, the influencing factor of each diagnosis is calculated, and the data is normalized to obtain the parameter set C.
[0087] The average cost, standard deviation of average cost, drug cost, consumable cost, examination cost, and DIP score of each DIP group are associated with the parameter sets P and C according to the diagnosis and surgery codes.
[0088] The corresponding sets P, C, average cost, average cost standard deviation, drug cost, consumables cost, inspection cost, and DIP score are clustered using mean shift
[0089] 3. The mean shift clustering method is as follows
[0090] 3.1 Initialization phase:
[0091] Select the starting point:
[0092] The selection of the initial point can be done by random sampling method, that is, randomly selecting a data point from the composite data set as the initial cluster center.
[0093] Density-based methods can also be used, such as selecting the point with the highest density in the data set as the initial cluster center. Density can be calculated using kernel density estimation, and the data point with the highest density estimate is selected.
[0094] Set the window radius (Bandwidth):
[0095] 1. The window radius can be determined by the following methods:
[0096] 2. Use Silverman's rule to estimate the bandwidth based on the standard deviation of the data set, that is, bandwidth h = 1.06*standard deviation*N^(-1 / 5), where N is the number of data points.
[0097] 3. Select the optimal bandwidth through cross-validation, such as minimizing the average silhouette coefficient of the clustering results.
[0098] 4. Specific values: Assume that the window radius h=0.5 is determined through cross-validation.
[0099] Mean calculation stage:
[0100] Define the weight function:
[0101] The weight function uses a Gaussian kernel function, which is defined as W(x)=exp(-(||xm||^2) / (2*h^2)), where x is the data point in the window, m is the cluster center, and h is the window radius.
[0102] Update cluster center stage:
[0103] Stop condition:
[0104] Convergence threshold: Set the convergence threshold to ε = 0.001, that is, when the moving distance of the cluster center is less than 0.001, stop the iteration.
[0105] Maximum number of iterations: Set the maximum number of iterations to T = 1000. Even if the cluster center is still moving, the iteration will stop after 1000 iterations.
[0106] Cluster assignment phase:
[0107] Data point assignment: After all cluster centers are stable, the distance from each data point to the cluster center is calculated using Euclidean distance, and the data point is assigned to the nearest cluster center.
[0108] Post-processing stage:
[0109] Noise processing: For clusters in the clustering results that contain data points less than a certain threshold (for example, less than 3% of the total number of data points), they are regarded as noise and removed from the results.
[0110] Centroid update algorithm formula:
[0111] Weight function (Gaussian Kernel): The weight function W(x) is used to calculate the weight of each data point x in the window to the cluster center m. This function usually uses the Gaussian kernel function, which is defined as:
[0112]
[0113] Where: x is the data point in the window, m is the current cluster center, h is the window radius (Bandwidth). ||xm|| is the Euclidean distance between point x and cluster center m.
[0114] Weighted average calculation: For each cluster center m, calculate the weighted average of all data points in its neighborhood, that is, the new cluster center m′. The calculation formula is:
[0115]
[0116] Where: N(m) is the set of data points in the window centered at m. W(x) is the data point The weight of . x is the data point in the window.
[0117] Centroid update: Move the cluster center m to the new weighted average m'. This step is repeated iteratively until the stopping condition is met.
[0118] Stop condition:
[0119] Convergence threshold ε: When the distance between the cluster center m and the new cluster center m is less than the preset convergence threshold ε, the iteration stops.
[0120] Maximum number of iterations T: Even if the cluster center is still moving, the iteration will stop when the preset maximum number of iterations T (default 1000) is reached.
[0121] In this way, the mean shift clustering algorithm can gradually update the cluster centers until dense areas of data points are found, which are the cluster centroids.
[0122] Through the above method, the centroids of all clusters are found and a set M = {m1', m2', m3', ... mn'} is formed.
[0123] The clustering effect is as follows Figure 1 As shown: (virtual, the actual dimension is higher, it is not necessary, and it is decided whether to keep it according to the situation.)
[0124] The Euclidean distance from each point to the centroid is calculated as shown in the following formula to obtain the vector d from each point to the centroid: i,c , the group information is recorded as i, and the total cost mean of the corresponding point is AC i , drug costs are DC i , the cost of consumables is CC i , inspection fee is EC i 、The average cost of DIP score is DIPAC i , the average cost standard deviation is ACSD i ,
[0125] The total cost mean of cluster c is AˉC i The drug cost of cluster c is DˉC i The consumables cost of cluster c is CˉC i , the inspection cost of cluster c is EˉC i The average cost of DIP score of cluster c is DIPAˉC i , the average cost standard deviation of cluster c is ACˉSDAˉ i .
[0126]
[0127] All vectors form a set: as shown below This set is the standard control parameters for the model.
[0128] The actual total cost of the case, drug costs, consumables costs, examination costs, and DIP scores are quantified using the following formula:
[0129]
[0130] The Euclidean distance is calculated by the following formula to compare the difference between the actual case V and the model d,
[0131]
[0132] Define two variables:
[0133] The total cost of actual cases was compared with the average total cost of model cases, and B was set as the benchmark parameter, indicating the total cost deviation.
[0134]
[0135] The DIP value of the actual case is compared with the DIP value of the model case and is set to R, which represents the range radius of the deviation.
[0136] Then, the vector range of the review condition to be searched is determined according to the following formula, and all vectors meeting the condition are retrieved to form a set D'.
[0137]
[0138] The weighted product of each qualified vector is obtained by the following formula and recorded as set B, which will be used as the parameter used for color mapping.
[0139]
[0140] During visualization, the elements in set B are sorted from largest to smallest, the one closest to 0 is selected as the benchmark, and the adjacent positive and negative n elements are selected as candidate suggestions to be displayed. n is defined according to business needs.
[0141] Each element bi in set B is color mapped through a mapping function. Deviations within ±0.5 are displayed in orange, those above +0.5 are displayed in red, and those below -0.5 are displayed in green.
[0142] If displayed in list form, such as Figure 2 As shown, the first line is the recommended coding combination, and the second line is the DIP reference value. The codes and data shown are only for reference and do not represent specific meanings.
[0143] When the mouse is moved over a specific box, more detailed information is displayed, including the specific name of the diagnosis and the name of the surgical procedure. If the display resolution allows, a 3-line version containing all the information can also be displayed directly in the above case, such as Figure 3 shown.
[0144] Through the above visualization, users can easily choose a more appropriate coding combination to optimize the filling of the homepage diagnosis surgery. Moreover, this method takes into account multiple factors such as total cost, drug cost, equipment cost, consumables cost, etc., and its recommendation accuracy is higher.
[0145] The present invention optimizes the algorithm and simplifies the content, so that when displaying the recommended code, it can effectively reduce the recommended items with better logic but large actual deviation, and the content is more accurate and in line with the actual situation. At the same time, through color and size, it is convenient for users to quickly judge the better options, and the thickness of the auxiliary line provides visualization of the probability selection.
[0146] Embodiment 1:
[0147] like Figure 2 As shown, the left side is the home page actually used by the hospital, and the right side is the display diagram displayed by the system of the present invention. The first line is the recommended code combination, and the second line is the DIP reference value. The codes and data shown are only for illustration and do not represent specific meanings. Figure 3 As shown, when the mouse is moved over a specific box, more detailed content is displayed, including the specific name of the diagnosis and the name of the surgical procedure. If the display resolution allows, a 3-line version containing all the content can also be displayed directly in the above case.
[0148] Embodiment 2:
[0149] like Figure 4 and Figure 5 As shown, the diagnostic code and surgical code corresponding to each specific element in set D are processed to obtain the diagnostic code and surgical code sub-items, and establish associations. Then the image is expanded with the four-digit surgical sub-items as the core and the specific surgical code as the first-level connection, and the code and name are displayed in a square box (or rounded box). For each surgical code, the four-digit diagnostic sub-item code is derived and displayed, represented by a circle, where the size of the circle represents the DIP score of the group, and the larger the circle, the higher the score. The color of the circle (background color or border color) is determined by the data of the corresponding element in set B, indicating its closeness to the optimal target. The thickness of the connecting line indicates the number of times this combination appears in previous medical records. The larger the number, the thicker the line. As shown Figure 4 When there are multiple surgical sub-items in set D, multiple relationship graphs are developed for each four-digit sub-item, such as Figure 5 shown.
[0150] A visualization platform includes at least one memory, a processor and an input-output module. The memory is used to store computer resources, and the processor is used to call computer programs and data stored in a visualization system to execute a DIP homepage coding optimization recommendation modeling visualization method.
[0151] The processor may include one or more processing cores. The processor uses various interfaces and lines to connect various parts of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory, and calling data stored in the memory. Optionally, the processor can use digital signal processing
[0152] (Digital Signal Processing, DSP), Field Programmable Gate Array
[0153] The processor can be integrated with one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display; and the modem is used to handle wireless communications. It is understandable that the above-mentioned modem may not be integrated into the processor, but may be implemented separately through a chip.
[0154] Among them, the memory may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory includes a non-transitory computer-readable storage medium. The memory can be used to store instructions, programs, codes, code sets or instruction sets. The memory may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory may also be optionally at least one storage device located away from the aforementioned processor.
[0155] In some implementations, after establishing a resource relationship model based on the corresponding relationship between the application program interfaces of the different hosts, the processor further executes: storing the host model and the resource relationship model in a database.
[0156] The embodiment of the present application also provides a computer-readable storage medium, which stores instructions, and when the instructions are executed on a computer or a processor, the computer or the processor executes one or more steps in the above-mentioned embodiment. If the components of the resource integration device are implemented in the form of software functional units and sold or used as independent products, they can be stored in the computer-readable storage medium.
[0157] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted by the computer-readable storage medium. The computer instructions can be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, a data center, etc. that contains one or more available media integrated. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a digital versatile disc (DVD)), or a semiconductor medium (eg, a solid state disk (SSD)).
[0158] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The aforementioned storage medium includes: a system memory (Read Only Memory, ROM), a random access memory (Random Access Memory, RAM), a disk or an optical disk, and other media that can store program codes. In the absence of conflict, the technical features of this embodiment and the implementation scheme can be arbitrarily combined.
[0159] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.
Claims
1. A DIP homepage coding optimization recommendation modeling visualization method, characterized in that: The following steps are involved: S1. Data processing: Obtain medical records and pass the information of medical records to the model, use the DIP grouping tool to pre-calculate the DIP score of each document, and obtain the parameter vector and weight required for the recommendation calculation; S2. Data conversion: The surgical procedures in the surgical record sheet, the main surgical code on the front page of the medical record and the probability of co-occurrence were used as factors to calculate their Pearson correlation coefficients as the basic surgical correlation parameter set P for each combination. The main diagnosis and main surgical code were converted into numerical three-dimensional vector data using a multi-column conversion method. S3, vectorized processing: The actual total cost of the case, drug costs, consumables costs, examination costs, and DIP scores are quantified using the following formula: The Euclidean distance is calculated by the following formula to compare the difference between the actual case V and the model d, The weighted product of each vector that meets the conditions is obtained by the following formula and recorded as set B. B will be used as the parameter used for color mapping. S4. Color processing: Sort the elements in set B from large to small, select the one closest to 0 as the benchmark, and select the adjacent positive and negative n ones as candidate suggestions to be displayed S5. Visualization of association graph: The diagnosis code and surgical code corresponding to each specific element in set D are processed to obtain the diagnosis code and surgical code sub-items and establish associations. Then the image is expanded with the four-digit surgical sub-items as the core, with the specific surgical code as the first-level connection, and the code and name are displayed. For each surgical code, the four-digit diagnosis sub-item code is derived and displayed. The size indicates the DIP score of the group. The larger the value, the higher the score. The color is determined by the data of the corresponding element in set B, indicating its proximity to the optimal target. The thickness of the connecting line indicates the number of times this combination appears in the previous medical records. The larger the number, the thicker the line, and a visual relationship diagram is obtained. S5. Visual display: The visual relationship diagram can be viewed and operated through the terminal.
2. A DIP homepage coding optimization recommendation modeling visualization method according to claim 1, characterized in that: The weight is to take the main diagnosis code, main surgery code and code that can be used as MCC / CC in the medical record as a code combination, and take the DRG-RW value and DIP score of the code combination as the comprehensive maximum as the optimization goal. The parameters of each code combination are obtained by calculating past data as the weight parameters for the recommendation algorithm ranking.
3. A DIP homepage coding optimization recommendation modeling visualization method according to claim 1, characterized in that: The weight function is 4. A DIP homepage coding optimization recommendation modeling visualization method according to claim 3, characterized in that: x is the data point in the window, m is the current cluster center, h is the window radius, and ||xm|| is the Euclidean distance between point x and cluster center m.
5. A DIP homepage coding optimization recommendation modeling visualization method according to claim 4, characterized in that: According to the cluster center m, the weighted average of all data points in its neighborhood is calculated, that is, the new cluster center m′. The calculation formula is:
6. A DIP homepage coding optimization recommendation modeling visualization method according to claim 5, characterized in that: in, N(m) is the set of data points in the window centered on m, W(x) is the weight of the data point x, and the cluster center m is moved to the position of the new weighted average m′. This step is repeated iteratively until the stopping condition is met.
7. A DIP homepage coding optimization recommendation modeling visualization method according to claim 1, characterized in that: Each element bi in the set B is color mapped through a mapping function, and the deviation within ±0.5 is displayed in orange, above +0.5 is displayed in red, and below -0.5 is displayed in green.
8. A DIP homepage coding optimization recommendation modeling visualization system, characterized by: include Data acquisition module: used to obtain medical records and transmit the information of medical records to the model; Data conversion module: converts data information into numerical three-dimensional vector data and transmits it to the model; Data analysis module: processes and analyzes the information of medical records and transmits it to the data conversion module; Model: the parameters required for calculation; Visualization processing module: displays data information in the form of images, text and lines; Color processing module: The data of the visualization processing module in set B is used to determine its closeness to the optimal target, making the visualization relationship diagram output by the visualization processing module clearer; Terminal: used to view the visual relationship diagram.
9. A visualization platform, characterized in that: It includes at least one memory, a processor and an input-output module, wherein the memory is used to store computer resources, and the processor is used to call computer programs and data stored in a visualization system to execute a DIP homepage coding optimization recommendation modeling visualization method as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Automatic-coding medical insurance disease category score paying system and construction method therefor
CN108831522A
DIP grouping recommendation method and system, terminal and storage medium
CN116487024A
Code selection method and device for reducing blank cases, equipment and medium
CN116881563A
Identifying Medical Codes Applicable to a Patient Based on Patient History and Probability Determination
US20170235884A1
