Medical insurance settlement method based on DRG model
By using a DRG-based medical insurance settlement method, combined with symptom feature vectors and concurrent synergistic effect models, the nonlinear processing and individual differences issues of the DRG payment model are resolved, achieving more accurate and equitable medical insurance payments.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-12
- Publication Date
- 2026-04-10
AI Technical Summary
The existing DRG payment model has problems such as linear handling of the superposition effect of complications, lag and static weight updates, and neglect of individual case differences, resulting in inaccurate and unfair settlement results.
A medical insurance settlement method based on the DRG model is adopted. By combining the symptom feature vector and the concurrent synergy effect model, the concurrent superposition coefficient is calculated. Combined with DRG grouping and regional basic premium rate, the medical insurance payment amount is calculated in a refined manner.
It enables precise quantification of the nonlinear synergistic effects among multiple complications, improves the fairness and accuracy of DRG payments, adapts to advancements in medical technology and cost changes, and safeguards reasonable profits for medical institutions.
Smart Images
Figure CN121836931A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to a medical insurance settlement method based on a DRG model and belongs to the technical field of medical insurance settlement. BACKGROUND
[0002] Under a DRG (disease diagnosis related group) payment system, firstly, a main diagnosis category (MDC) to which a main disease belongs is determined, and a basic DRG is confirmed; then complications are evaluated, and it is analyzed whether a main complication / combination (MCC) exists, whether a complication / combination (CC) exists, or whether a non-complication / combination (Non-CC) exists, so that the DRG grouping is adjusted; finally, a medical insurance payment amount is determined according to a basic rate and in combination with a weight corresponding to the DRG grouping; however, in actual application, the current standard DRG payment model has the following core technical problems:
[0003] 1. Linear processing defects of complication superposition effects, the existing DRG weight only performs stepwise classification based on “whether the MCC / CC exists” (for example, no CC, with CC, and with MCC), and the classification manner cannot finely quantify the synergistic effect and nonlinear cost growth when multiple complications exist at the same time, that is, one complication belonging to the MCC has triggered the highest payment level, and the additional cost of the second and third complications belonging to the MCC or the CC is completely ignored. For example, the resource consumption of a patient with “pneumonia + respiratory failure (MCC)” is exponentially increased compared with a patient with “pneumonia + respiratory failure (MCC) + renal failure (MCC) + sepsis (MCC)”, but the payment standard is the same.
[0004] 2. Hysteresis and staticity of weight updating, the DRG weight is a static value measured based on historical data, and it cannot respond to the cost structure changes caused by new diagnosis and treatment technologies, new drugs on the market and sudden public health events in real time, thereby inhibiting the enthusiasm of technical innovation and efficient service.
[0005] 3. Ignorance of individual case differences, patients in the same DRG group are regarded as “homogeneous”, but there are great differences among actual individuals, such as age, constitution and specific disease differences, such as a 90-year-old weak patient with “cholecystitis with MCC” and a 40-year-old strong patient with “cholecystitis with MCC”, and the treatment complexity and cost are completely different, but the payment standard is the same. SUMMARY
[0006] The technical problem to be solved by the application is to provide a medical insurance settlement method based on a DRG model, which comprehensively considers the main diagnosis disease and each complication to obtain a more objective and more accurate settlement result.
[0007] The present application adopts the following technical solutions to solve the above technical problems: The present application designs a medical insurance settlement method based on a DRG model, and performs the following steps A to step E to determine the medical insurance payment amount corresponding to the target case;
[0008] Step A. According to the basic DRG grouping under the DRG model corresponding to the main diagnosis disease in the target case, the basic weight corresponding to the basic DRG grouping is obtained , and then step B is entered;
[0009] Step B. Determine whether there is a complication in the target case in addition to the main diagnosis disease. If yes, go to step C; otherwise, go to step E;
[0010] Step C. Obtain the comprehensive disease feature vector corresponding to the main diagnosis disease and each complication in the target case, and apply the pre-trained complication synergistic effect model with the comprehensive disease feature vector as the input and the complication superposition coefficient corresponding to the comprehensive disease feature vector as the output to obtain the complication superposition coefficient corresponding to the comprehensive disease feature vector of the target case , and then go to step D; wherein ;
[0011] Step D. According to the regional basic rate under the DRG model, the calculation result of is obtained, which constitutes the medical insurance payment amount corresponding to the target case;
[0012] Step E. According to the regional basic rate under the DRG model, the calculation result of is obtained, which constitutes the medical insurance payment amount corresponding to the target case.
[0013] As a preferred technical solution of the present application, in step C, the complication superposition coefficient corresponding to the comprehensive disease feature vector of the target case is obtained, and the following is also performed:
[0014] Obtain the excess consumption amount of each preset extraordinary resource item of the target case diagnosis and treatment process exceeding the upper limit of the normal range corresponding to the basic DRG grouping, and then combine the standard unit price corresponding to each excess consumption amount to obtain the excess cost of the target case compared with the normal range of the basic DRG grouping ; then go to step D;
[0015] In step D, according to the regional basic rate under the DRG model, the calculation result of is obtained, which constitutes the medical insurance payment amount corresponding to the target case.
[0016] As a preferred technical solution of the present invention: the preset extraordinary resource items satisfy at least one of the following conditions 1 and 2, and the usage can be quantified;
[0017] Condition 1. The standard unit price of the extraordinary resource item exceeds the preset unit price threshold;
[0018] Condition 2. The average usage of the extraordinary resource item in historical single-case statistics is greater than the preset usage threshold;
[0019] Based on historical statistics of each first historical case enrolled to the target case and corresponding to the basic DRG group, and each of them including the primary diagnosis disease and at least one complication, for each preset abnormal resource item, the upper limit of the data range consisting of the data corresponding to the abnormal resource item of each first historical case is statistically obtained by the percentage of the preset upper limit, and then constitutes the upper limit of the normal range of each abnormal resource item corresponding to the basic DRG group of the target case.
[0020] In step C, the formula is as follows:
[0021]
[0022] Calculate the excess cost of obtaining the target case relative to its baseline DRG grouping standard range. ,in, Indicates the number of abnormal resource items. This indicates the corresponding case in the diagnosis and treatment process of the target case. The actual consumption of each extraordinary resource item This indicates the baseline DRG group corresponding to the target case. The upper limit of the normal range for each extraordinary resource item. Indicates the first The standard unit price of an extraordinary resource item.
[0023] As a preferred embodiment of the present invention, the preset upper limit percentage is 95%.
[0024] As a preferred technical solution of the present invention, in step C, the comprehensive symptom feature vector corresponding to the main diagnosed disease and various complications in the target case is obtained according to the following steps C1-1 to C1-2;
[0025] Step C1-1. Apply the embedding layer in the deep learning model to map and obtain the low-dimensional embedding vectors corresponding to the main diagnosed disease and each complication in the target case, and then proceed to step C1-2;
[0026] Step C1-2. For each low-dimensional embedding vector corresponding to the target case, perform concatenation, re-averaging, or weighted aggregation through attention mechanism, and the result constitutes the comprehensive disease feature vector corresponding to the main diagnosis disease and the whole of each complication in the target case.
[0027] As a preferred technical solution of the present application, after obtaining the comprehensive disease feature vector corresponding to the main diagnosis disease and the whole of each complication in the target case in step C1-2, further obtain the statistical features of the patient's age, the number of complications, and whether there is a major complication / comorbidity in each complication in the target case, and add them to the comprehensive disease feature vector for updating.
[0028] As a preferred technical solution of the present application, the concurrent synergistic effect model in step C is obtained as follows:
[0029] Step C2-1. Collect a preset number of each second historical case each including a main diagnosis disease and at least one complication, and obtain the comprehensive disease feature vector corresponding to each second historical case and the actual total cost corresponding to each second historical case in the manner of steps C1-1 to C1-2. Then go to step C2-2;
[0030] Step C2-2. For each second historical case, calculate the cost ratio corresponding to the second historical case according to the following formula:
[0031]
[0032] , and then go to step C2-3;
[0033] Step C2-3. Based on each second historical case, take the comprehensive disease feature vector corresponding to the second historical case as input and the cost ratio corresponding to the second historical case as output, train the preset classification model, obtain the trained model, and take the cost ratio as the concurrent superposition coefficient , that is, obtain the concurrent synergistic effect model;
[0034] Wherein, the concurrent superposition coefficient = 1, indicating that the complication has no superposition effect compared with the basic DRG grouping corresponding to the case; the concurrent superposition coefficient > 1, indicating that the complication has a nonlinear superposition effect compared with the basic DRG grouping corresponding to the case.
[0035] As a preferred technical solution of the present application: the preset classification model in step C2-3 is any one of the three classification models of gradient boosting decision tree model GBDT, extreme gradient boosting tree model XGBoost improved based on gradient boosting decision tree model GBDT, and light gradient boosting tree model LightGBM.
[0036] As a preferred technical solution of the present application: based on the blockchain, the data involved in steps A to E are stored on the chain.
[0037] As a preferred technical solution of the present application: the correspondence between the basic DRG grouping and the corresponding basic weight in step A is constructed according to the following steps A1 to A4;
[0038] Step A1. Based on a preset number of third historical cases in a preset historical period, the actual total cost corresponding to each third historical case is counted, and then step A2 is entered;
[0039] Step A2. Under the DRG model, the basic DRG grouping operation is performed on the main diagnostic diseases in each third historical case, and the DRG grouper is used to perform the basic DRG grouping operation on each third historical case to obtain the third historical case corresponding to each basic DRG grouping, and then step A3 is entered;
[0040] Step A3. The first average cost value of the actual total cost of the third historical cases corresponding to each basic DRG grouping is counted, and then the first average cost value corresponding to each basic DRG grouping is obtained, and then step A4 is entered;
[0041] Step A4. The total average cost value of the first average cost values of each basic DRG grouping is counted, and then the ratio of the first average cost value corresponding to each basic DRG grouping to the total average cost value is used to form the basic weight corresponding to the basic DRG grouping.
[0042] The DRG model-based medical insurance settlement method has the following technical effects compared with the prior art:
[0043] The DRG model-based medical insurance settlement method is designed, which first determines the basic weight according to the basic DRG grouping of the main diagnostic diseases in the target case ; then the concurrent synergistic effect model is applied to the comprehensive disease feature vector corresponding to the main diagnostic diseases and each complication to obtain the corresponding complication superposition coefficient ; then the regional basic rate under the DRG model , the medical insurance payment amount corresponding to the target case is calculated; the design scheme introduces a concurrent superposition coefficient, and the cost growth caused by the nonlinear synergistic effect between multiple complications is accurately quantified, the roughness of the traditional DRG model 'jumping' grouping is overcome, the design method significantly improves the fairness and accuracy of the DRG payment, which can not only guarantee the reasonable income of medical institutions for treating complex cases, but also adapt to the progress of medical technology and cost changes, and provides core technical support for the scientific and fine management of the medical insurance payment system. BRIEF DESCRIPTION OF DRAWINGS
[0044] Figure 1 It is the architecture schematic diagram of the medical insurance settlement method based on the DRG model designed by the application. DETAILED DESCRIPTION
[0045] The specific embodiments of the application will be further described in detail below in combination with the drawings of the specification.
[0046] The application designs a medical insurance settlement method based on a DRG model, which is actually applied in the operation and implementation of a medical insurance settlement system, the medical insurance settlement system is connected to the HIS (hospital information system), LIS (laboratory information system) and PACS (picture archiving and communication system) of the hospital, and real-time collection of anonymized resource consumption data (such as ICU hours, special drug consumption, high-value consumables, operation time, nursing working hours, etc.) is performed, and as shown in the figure, the medical insurance settlement system specifically performs the following steps A to E to determine the medical insurance payment amount corresponding to the target case. Figure 1
[0047] Step A. According to the basic DRG grouping of the target case under the DRG model corresponding to the main diagnostic disease, the basic weight corresponding to the basic DRG grouping is obtained Then step B is entered.
[0048] Wherein, the corresponding relationship between the basic DRG grouping and the corresponding basic weight is constructed according to the following steps A1 to A4.
[0049] Step A1. Based on a preset number of third historical cases in a preset historical period, the actual total cost corresponding to each third historical case is counted, and then step A2 is entered.
[0050] Step A2. Under the DRG model, according to the basic DRG grouping operation for the main diagnostic disease in each third historical case, the DRG grouper is used to perform the basic DRG grouping operation for each third historical case, and each third historical case corresponding to each basic DRG grouping is obtained, and then step A3 is entered.
[0051] Step A3. For each base DRG group, respectively, count the first average cost value of the actual total cost of each third historical case corresponding to the base DRG group, and then obtain the first average cost value corresponding to each base DRG group, respectively, and then enter step A4.
[0052] Step A4. Count the total average cost value of the first average cost values of each base DRG group, and then for each base DRG group, respectively, construct the base weight corresponding to the base DRG group with the ratio of the first average cost value corresponding to the base DRG group to the total average cost value.
[0053] Step B. As shown in Figure 1 , determine whether there is a complication in addition to the primary diagnosis disease in the target case. If yes, enter step C; otherwise, enter step E.
[0054] Next, consider the cost increase caused by the nonlinear synergistic effect between multiple complications, that is, enter the execution of step C.
[0055] Step C. As shown in Figure 1 , first obtain the comprehensive disease feature vector corresponding to the primary diagnosis disease and each complication in the target case by following steps C1-1 to C1-2.
[0056] Step C1-1. First, obtain the statistical features of the patient's age, the number of complications, and whether there is a major complication / combination in each complication in the target case, and then apply the embedding layer Embedding in the deep learning model to map the low-dimensional embedding vectors corresponding to the patient's age, the number of complications, whether there is a major complication / combination in each complication, the primary diagnosis disease, and each complication in the target case, and then enter step C1-2.
[0057] Step C1-2. First, for each low-dimensional embedding vector corresponding to the target case, perform concatenation, then average, or perform weighted aggregation through attention mechanism, and the obtained result constitutes the comprehensive disease feature vector corresponding to the primary diagnosis disease and each complication in the target case.
[0058] Then apply the pre-trained concurrent synergistic effect model with the comprehensive disease feature vector as input and the concurrent superposition coefficient corresponding to the comprehensive disease feature vector as output to obtain the concurrent superposition coefficient corresponding to the comprehensive disease feature vector of the target case , wherein , such as the target case including pneumonia and respiratory failure, the obtained concurrent superposition coefficient , or such as the target case including pneumonia, and respiratory failure and renal failure, the obtained concurrent superposition coefficient .
[0059] In practical applications, the following steps C2-1 to C2-3 are performed to obtain the concurrent synergistic effect model.
[0060] Step C2-1. Collect a preset number of second historical cases each including a primary diagnostic disease and at least one complication, and obtain the corresponding comprehensive disease feature vector of each second historical case and the actual total cost corresponding to each second historical case according to the manner of steps C1-1 to C1-2. Then go to step C2-2.
[0061] Step C2-2. For each second historical case, the following formula is used:
[0062]
[0063] The cost ratio corresponding to the second historical case is calculated as , and the cost ratio corresponding to each second historical case is obtained, and then go to step C2-3.
[0064] Step C2-3. Based on each second historical case, the comprehensive disease feature vector corresponding to the second historical case is input and the cost ratio corresponding to the second historical case is output, and the preset classification model is trained to obtain the trained model, and the cost ratio is used as the concurrent superposition coefficient , that is, the concurrent synergistic effect model is obtained; wherein the concurrent superposition coefficient =1, indicating that the complication has no superposition effect compared to the basic DRG grouping corresponding to the case; the concurrent superposition coefficient >1, indicating that the complication has a nonlinear superposition effect compared to the basic DRG grouping corresponding to the case; the preset classification model is any one of the three classification models of gradient boosting decision tree model GBDT, extreme gradient boosting tree model XGBoost improved based on gradient boosting decision tree model GBDT, and light gradient boosting tree model LightGBM.
[0065] When the concurrent superposition coefficient is obtained, the excess consumption of each preset extraordinary resource item of the target case diagnosis and treatment process that exceeds the upper limit of the normal range of its basic DRG grouping is obtained, and the excess cost of the target case compared to the normal range of its basic DRG grouping is obtained by combining the standard unit price corresponding to each excess consumption. Then go to step D.
[0066] In practical applications, each preset extraordinary resource item satisfies at least one of the following conditions 1 and 2, and the amount is quantifiable.
[0067] Condition 1. The standard unit price of the extraordinary resource item exceeds the preset unit price threshold;
[0068] Condition 2. The average usage of the extraordinary resource item in historical statistics of single cases is greater than the preset usage threshold.
[0069] Based on historical statistics of each first historical case enrolled to the target case and corresponding to the basic DRG group, and each including the primary diagnosis and at least one complication, for each preset extraordinary resource item, the upper limit of the data range consisting of the data corresponding to each extraordinary resource item of the first historical cases is statistically obtained by a preset upper limit percentage number of first historical cases. This constitutes the upper limit of the normal range of each extraordinary resource item corresponding to the basic DRG group of the target case. In practical applications, for example, if 95% of the data of the first historical cases for a certain extraordinary resource item does not exceed 'a', then 'a' is the upper limit of the normal range of the basic DRG group of the target case for that extraordinary resource item. In specific implementation, for example, 95% of patients with simple cholecystitis stay in the ICU for no more than 12 hours, and 95% of patients with simple cholecystitis do not need a ventilator at all.
[0070] Based on the upper limit of the normal range of the extraordinary resource item corresponding to the basic DRG group of the target case, in specific applications, step C above is performed according to the following formula:
[0071]
[0072] Calculate the excess cost of obtaining the target case relative to its baseline DRG grouping standard range. ,in, Indicates the number of abnormal resource items. This indicates the corresponding case in the diagnosis and treatment process of the target case. The actual consumption of each extraordinary resource item This indicates the baseline DRG group corresponding to the target case. The upper limit of the normal range for each extraordinary resource item. Indicates the first The standard unit price of an extraordinary resource item.
[0073] Step D. As Figure 1 As shown, based on the regional basic rate under the DRG model ,get The calculation results constitute the medical insurance payment amount corresponding to the target case.
[0074] Step E. Based on the regional base rate under the DRG model ,get The calculation results constitute the medical insurance payment amount corresponding to the target case.
[0075] In actual application, in order to prevent coding game, based on the blockchain, the key coding evidence (such as the hash value of the key examination report and the consultation record) of the main diagnosis and the complications is stored on the chain, that is, the data involved in the execution of steps A to E are all stored by being chained, the non-tamperable nature of the blockchain is used to provide the audit tracking function after the event, the coding authenticity is ensured, and the system credibility is increased.
[0076] The above technical solution designs a medical insurance settlement method based on the DRG model. First, the basic weight is determined according to the basic DRG grouping of the main diagnosis disease in the target case ; then, the corresponding superposition coefficient of the complications is obtained by applying the concurrent synergistic effect model to the comprehensive disease characteristic vector corresponding to the main diagnosis disease and each complication ; and then, the medical insurance payment amount corresponding to the target case is calculated according to the regional basic rate under the DRG model . The design scheme realizes accurate quantification of the cost growth caused by the nonlinear synergistic effect among multiple complications by introducing the superposition coefficient, overcomes the roughness of the traditional DRG model "jumping" grouping, significantly improves the fairness and accuracy of the DRG payment, can guarantee the reasonable income of the medical institutions for treating complex cases, and can adapt to the progress of medical technology and cost changes, thereby providing core technical support for the scientific and fine management of the medical insurance payment system.
[0077] The embodiments of the application are described in detail above in combination with the drawings, but the application is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the purpose of the application.
Claims
1. A medical insurance settlement method based on a DRG model, characterized by: The following steps A to E are performed to determine the medical insurance payment amount corresponding to the target case; Step A. According to the basic DRG grouping under the DRG model corresponding to the main diagnosis disease in the target case, the basic weight corresponding to the basic DRG grouping is obtained Then go to step B; Step B. Determine whether there is a complication in the target case in addition to the main diagnosis disease. If yes, go to step C; otherwise, go to step E; Step C. Obtain the comprehensive disease syndrome feature vector corresponding to the main diagnosis disease and each complication in the target case as a whole, and apply the pre-trained complication synergistic effect model with the comprehensive disease syndrome feature vector as the input and the complication superposition coefficient corresponding to the comprehensive disease syndrome feature vector as the output to obtain the complication superposition coefficient corresponding to the comprehensive disease syndrome feature vector of the target case Then go to step D; wherein ; Step D. Obtain the calculation result of the DRG model under the regional basic rate , the calculation result of the DRG model under the regional basic rate , the calculation result of the DRG model under the regional basic rate Step E. Obtain the calculation result of the DRG model under the regional basic rate , the calculation result of the DRG model under the regional basic rate , the calculation result of the DRG model under the regional basic rate 2. The medical insurance settlement method based on the DRG model according to claim 1, wherein, The concurrent superposition coefficients corresponding to the comprehensive symptom feature vector of the target case obtained in step C Also included is performing the following: Obtaining the excess consumption amount of each preset extraordinary resource item corresponding to the target case diagnosis and treatment process exceeding the upper limit of the normal range of the basic DRG grouping, and combining the standard unit price corresponding to each excess consumption amount, the excess cost of the target case compared with the normal range of the basic DRG grouping is obtained ; then step D is entered; In Step D, the calculation result of the DRG model-based regional basic rate is obtained to constitute the medical insurance payment amount corresponding to the target case. 3.The medical insurance settlement method based on the DRG model of claim 2, wherein: The preset each extraordinary resource item meets at least one of the following conditions 1 and condition 2, and the usage is quantifiable; Condition 1. The standard unit price of the extraordinary resource item exceeds the preset unit price threshold; Condition 2. The average usage of the extraordinary resource item in the historical statistics of each first historical case in the target case corresponding to the basic DRG grouping is greater than the preset usage threshold; Based on the historical statistics, each first historical case in the target case corresponding to the basic DRG grouping and each first historical case respectively including the main diagnosis disease and at least one complication, for each preset extraordinary resource item, the upper limit value of the data range formed by the data of the preset upper limit percentage number of first historical cases corresponding to the extraordinary resource item is obtained, and then the upper limit of the normal range corresponding to each extraordinary resource item of the basic DRG grouping corresponding to the target case is formed; In step C, the following formula is used: ； The excess cost of the target case compared to the regular range of the base DRG grouping is calculated , wherein represents the number of extraordinary resource items, represents the actual consumption of the target case in the course of treatment corresponding to the th extraordinary resource item, represents the upper limit of the regular range of the target case corresponding to the base DRG grouping corresponding to the th extraordinary resource item, represents the standard unit price of the th extraordinary resource item.
4. The medical insurance settlement method based on the DRG model according to claim 3, characterized in that: The preset upper limit percentage is 95%. 5.The medical insurance settlement method based on the DRG model according to any one of claims 1 to 4, characterized in that, In step C, the comprehensive disease feature vector corresponding to the main diagnosis disease and each complication in the target case is obtained by the following steps C1-1 to C1-2; Step C1-1. Apply the embedding layer Embedding in the deep learning model to map the low-dimensional embedding vector corresponding to the main diagnosis disease and each complication in the target case, and then go to step C1-2; Step C1-2. For each low-dimensional embedding vector corresponding to the target case, perform concatenation, then average, or perform weighted aggregation through attention mechanism, and the obtained result constitutes the comprehensive disease feature vector corresponding to the main diagnosis disease and each complication in the target case. 6.The medical insurance settlement method based on the DRG model according to claim 5, wherein, After obtaining the comprehensive disease feature vector corresponding to the main diagnosis disease and each complication in the target case in step C1-2, further obtain the statistical features of the patient's age, the number of complications, and whether there is a major complication / combination in each complication, and add them to the comprehensive disease feature vector for updating.
7. The medical insurance settlement method based on the DRG model according to claim 6, wherein, The concurrent synergistic effect model in step C is obtained by the following steps C2-1 to C2-3: Step C2-1. Collect a preset number of second historical cases each including the main diagnostic disease and at least one complication, and obtain a comprehensive disease feature vector corresponding to each second historical case respectively in the manner of Step C1-1 to Step C1-2, and obtain an actual total cost corresponding to each second historical case respectively and then go to Step C2-2; Step C2-2. For each second historical case, the following formula is used: ； The cost ratio corresponding to the second historical case is calculated , and then the cost ratio corresponding to each second historical case is obtained, and then step C2-3 is entered; Step C2-3. Based on each second historical case, taking the comprehensive disease symptom feature vector corresponding to the second historical case as input, taking the cost ratio corresponding to the second historical case as output, training the preset classification model to obtain a trained model, and taking the cost ratio as the concurrent superposition coefficient , that is, obtaining the concurrent synergistic effect model; wherein the concurrent overlay coefficient = 1 indicates that the comorbidity has no overlay effect compared to the underlying DRG grouping to which the case corresponds; and the concurrent overlay coefficient > 1 indicates that the comorbidity has a non-linear overlay effect compared to the underlying DRG grouping to which the case corresponds. 8.The DRG model-based medical insurance settlement method of claim 7, wherein: In step C2-3, the preset classification model is any one of the following three classification models: gradient boosting decision tree model GBDT, extreme gradient boosting tree model XGBoost based on gradient boosting decision tree model GBDT, and light gradient boosting tree model LightGBM.
9. The medical insurance settlement method based on the DRG model according to claim 8, characterized in that: Based on the blockchain, the data involved in steps A to E is stored on the chain. 10.The DRG model-based medical insurance settlement method of claim 1, wherein: The correspondence between the basic DRG grouping and the corresponding basic weight in step A is constructed by the following steps A1 to A4; Step A1. Based on the statistical preset historical period, a preset number of third historical cases are obtained, and the actual total cost corresponding to each third historical case is obtained, and then step A2 is entered; Step A2. Under the DRG model, the basic DRG grouping operation is performed on each third historical case according to the basic DRG grouping of the main diagnosis disease in each third historical case by using a DRG grouper to obtain each third historical case corresponding to each basic DRG grouping respectively, and then the step A3 is entered; Step A3. The first average cost value of the actual total cost of each third historical case corresponding to each basic DRG grouping is counted respectively for each basic DRG grouping, and then the first average cost value corresponding to each basic DRG grouping is obtained, and then the step A4 is entered; Step A4. The total average cost value of the first average cost value of each basic DRG grouping is counted, and then the ratio of the first average cost value corresponding to each basic DRG grouping to the total average cost value is used to form the basic weight corresponding to the basic DRG grouping for each basic DRG grouping.