A DRG grouping automatic review method and system

Abstract

The present invention relates to the technical field of medical data processing, and in particular to a method and system for automatic review of DRG grouping. The method includes the steps of: S1, extracting all diagnoses and surgical operations from the front page of the patient's inpatient medical record to generate a Cartesian set of the diagnosis set and the surgical operation set; S2, respectively using each combination in the Cartesian set as the primary diagnosis or surgical operation, and the patient's other diagnoses or surgical operations as the secondary diagnoses or surgical operations to form a grouping set, and inputting the grouping set into a DRG grouper to obtain the DRG code of each grouping set; S3, judging whether it holds. If it holds, judge whether the DRG code corresponding to A_COST is consistent with the original DRG code. If it is, the review passes; otherwise, the review fails. The method of the present invention is simple and efficient, and the DRG code is accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a DRG grouping automatic audit method and system. Background Art

[0002] DRG grouping data is sourced from the medical insurance settlement list. To improve work efficiency and reduce labor costs, the medical insurance settlement list data of medical institutions mainly comes from the inpatient medical record front page system. Therefore, DRG grouping has increasingly high requirements for the quality of inpatient medical record front page data filling. Currently, the number of coding personnel in medical institutions is seriously insufficient and the professional level is not high. The disease diagnosis and surgical operation coding system of the medical record front page mainly meets the needs of disease health statistics rather than the needs of medical insurance DRG settlement. Therefore, in order to meet the needs of medical insurance DRG cost settlement, it is also necessary to establish the conversion between the disease diagnosis and surgical operation coding system of the inpatient medical record front page and the medical insurance coding system. The differences between the two are mainly reflected as follows: 1. The selection of the main diagnosis in the health and hygiene system takes the severity of the disease as the primary factor, while the medical insurance system takes the diagnosis with the most resource consumption as the main diagnosis; 2. The scope of the health and hygiene coding system is much larger than that of the medical insurance coding system. For example, some disease incentives, disease states, injuries and poisonings in the health and hygiene coding system are not included in the medical insurance coding system. These differences pose extremely high requirements for the capabilities of medical institution coders, and the distorted data also brings deviations to the accuracy of DRG grouping by the medical insurance department, which in turn causes troubles to medical expense settlement and medical insurance fund budget management. Summary of the Invention

[0003] The first object of the present invention is to solve the problem of deviation in current medical institution DRG coding, and propose a DRG grouping automatic audit method and system to improve the reliability of DRG grouping; the second object of the present invention is to solve the problem of low grouping efficiency of the current DRG grouper, and provide a new DRG grouper.

[0004] To achieve the first object of the present invention, the present invention provides the following technical solutions:

[0005] A DRG grouping automatic audit method includes the following steps:

[0006] S1, extract all diagnoses and surgical operations from the inpatient medical record front page of the patient, generate a diagnosis dataset based on all diagnoses, generate a surgical operation dataset based on all surgeries, and generate the Cartesian product set of the diagnosis dataset and the surgical operation dataset;

[0007] S2, respectively use each combination in the Cartesian product set as the main diagnosis or surgical operation, and use the other diagnoses or surgical operations of the patient as secondary diagnoses or surgical operations to form a grouping set, input it into the DRG grouper for grouping, and obtain the DRG code of each grouping set;

[0008] S3, judge Whether it holds. If it holds, further determine whether the DRG code corresponding to A_COST is the same as the original DRG code. If so, the review passes; if not or it does not hold, the review fails; FEE_TOTAL is the total hospitalization cost of the patient, k is the weight, A_COST is the local average hospitalization cost, |·| represents taking the absolute value, and Y is the set threshold.

[0009] In the above solution, all possible DRG codes are generated by combining all the diagnoses and surgeries in the front page of the hospitalization medical record. Then, the original DRG code generated by the coder is compared with the generated codes to find out whether the original DRG code is correct. On the one hand, since the obtained DRG codes are generated based on the combination of all the diagnoses and surgeries in the front page of the hospitalization medical record, accurate DRG codes will surely not be missed. That is, it can not only review whether the original DRG code is correct, but also give the accurate DRG code when the original DRG code is incorrect (caused by the wrong order of the main diagnosis or surgery and the secondary diagnosis or surgery), and can also find out the situation of missing diagnoses or surgeries in the front page of the hospitalization medical record. That is, almost all possible situations that may lead to DRG code errors can be found through the above review method. On the other hand, the processing steps of the above method are extremely simple, will not generate a large amount of calculation, and will not generate a large number of useless DRG codes. It can very efficiently find the most accurate code from a small number of codes. That is, the steps of this review method are designed based on the real application scenario and more meet the actual application requirements. Generating a small number of more accurate DRG codes has more practical application value than generating a large number of useless codes.

[0010] The above DRG grouping automatic review method further includes step S4. When the review fails, if the DRG code corresponding to A_COST is different from the original DRG code, output the DRG code corresponding to A_COST; if it does not hold, output a prompt message indicating that there may be missing diagnoses or surgeries in the front page of the hospitalization medical record.

[0011] In the above solution, on the basis of realizing the review, when the review fails, the correct DRG code is given, or it is prompted that the original DRG code error may be caused by missing diagnoses or surgeries, for the reference of coders and clinicians, so as to facilitate the timely discovery and correction of errors.

[0012] Step S1 is replaced by step S0. The processing of step S0 includes: extracting all diagnoses and surgical operations from the first page of the patient's inpatient medical record, determining whether the medical insurance code of the diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data. If so, mark the medical insurance code of the diagnosis or surgical operation as not included in the grouping; otherwise, mark it as included in the grouping. Generate a diagnosis set based on all diagnoses marked as included in the grouping, generate a surgical operation set based on all surgical operations marked as included in the grouping, and generate the Cartesian product set of the diagnosis set and the surgical operation set.

[0013] In the above solution, first, exclude the diagnoses or surgical operations that do not belong to the DRG coding (stipulated by the national medical insurance coding rules), and then combine the remaining diagnoses and surgical operations. This can make the amount of combined data smaller, and thus the time consumed to generate DRG codes is also less, further improving the review efficiency.

[0014] To achieve the second object of the present invention, the present invention provides the following technical solutions:

[0015] For the DRG grouper in S2, for each grouping set, perform the following operations:

[0016] S21, determine whether it meets the priority grouping condition according to the patient's attributes. If so, generate the ADRG grouping result based on the priority group grouping rules; otherwise, proceed to step S22;

[0017] S22, generate the ADRG grouping result based on the patient's gender, the medical insurance code of the diagnosis, and the medical insurance code of the surgical operation;

[0018] S23, use the ADRG grouping result as the first three digits of the DRG code, determine the fourth digit of the DRG code based on the medical insurance code of the secondary diagnosis, and adjust the generated four-digit DRG code according to the locally published DRG grouping directory to obtain the final DRG code.

[0019] By carefully analyzing the DRG coding rules compiled by the National Healthcare Security Administration, there is a correlation between the four characters that make up the DRG code. In the above solution, first obtain the ADRG grouping result based on the priority grouping condition, and only when it cannot be determined, determine the ADRG grouping result based on the medical insurance code of the diagnosis or surgical operation. Through such a processing sequence design, the amount of data processing can be minimized. For example, there will not be two DRG codes that belong to both the priority group and the main diagnosis coding rules at the same time. Therefore, the efficiency of generating DRG codes can be greatly improved.

[0020] The said step S21 includes the following processing:

[0021] Extract the patient's attributes from the first page of the patient's hospitalization medical record, including neonatal attributes, medical insurance codes for diagnoses, and medical insurance codes for surgical operations;

[0022] First, determine whether the medical insurance codes for diagnoses and surgical operations meet the MDCA grouping conditions. If so, further obtain the ADRG grouping result based on the medical insurance codes for diagnoses and surgical operations;

[0023] If it does not meet the MDCA grouping conditions, then determine whether it belongs to the MDCP grouping conditions based on the neonatal attributes. If so, further obtain the ADRG grouping result based on the neonatal attributes, medical insurance codes for diagnoses, and medical insurance codes for surgical operations;

[0024] If it does not meet the MDCP grouping conditions, then determine whether it meets the MDCY grouping conditions based on the medical insurance codes for diagnoses or surgical operations. If so, obtain the ADRG grouping result based on the medical insurance codes for diagnoses or surgical operations;

[0025] If it does not meet the MDCY grouping conditions, then determine whether it meets the MDCZ grouping conditions based on the medical insurance codes for diagnoses or surgical operations. If so, obtain the ADRG grouping result based on the medical insurance codes for diagnoses or surgical operations, otherwise enter step S22.

[0026] In the above solution, the ADRG grouping result is determined in sequence according to the MDCA, MDCP, MDCY, and MDCZ grouping conditions. Such a design method can not only reduce the data processing volume as much as possible, but also ensure the accuracy of the ADRG grouping result.

[0027] The said step S22 includes the following processing:

[0028] Establish a dataset for the main diagnosis categories and a dataset for ADRG grouping rules. The dataset for the main diagnosis categories contains the attributes DIAG_CODE, DIAG_NAME, and MDC. The dataset for ADRG grouping rules contains the attributes MDC, ADRG, ICD_CODE, and ICD_NAME. DIAG_CODE represents the medical insurance code for diagnoses, DIAG_NAME represents the medical insurance name for diagnoses, ICD_CODE represents the medical insurance code for diagnoses or surgical operations, and ICD_NAME represents the medical insurance name for diagnoses or surgical operations;

[0029] Search the dataset for the main diagnosis categories, and use the value of the MDC attribute item that successfully matches the gender and the medical insurance code for the main diagnosis as the first digit of the ADRG grouping result;

[0030] With the value of the MDC attribute item being the first digit of the determined ADRG grouping result and the value of the ICD_CODE attribute item being the medical insurance code of the diagnosis or surgical operation as conditions, matching data entries are searched in the ADRG grouping rule data set, and the value of the ADRG attribute item in the first-ranked data entry is used as the ADRG grouping result.

[0031] In the above scheme, by establishing a data set, searching in the data set, the processing efficiency of the data set search method is high, and the ADRG grouping rule data set can be composed of MDC subsets of each single character spliced in sequence. The next subset is searched only when the previous subset cannot be found (for example, the MDCC subset is searched when the MDCB subset cannot be found). Therefore, the generation efficiency of DRG code can be further improved.

[0032] The step S23 includes the following processing:

[0033] Establishing a complication or complication catalog data set and a complication or complication exclusion data set, wherein the complication or complication catalog data set includes attributes DIAG_CODE, DIAG_NAME, VALUE, and MCC_CC, and the complication or complication exclusion data set includes attributes DIAG_CODE, DIAG_NAME, and VALUE, wherein VALUE indicates a complication or complication sequence number, and MCC_CC indicates a complication category;

[0034] Match the medical insurance code of the primary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication exclusion data set. If the match is successful, obtain the value of the VALUE attribute item, and use the value of the VALUE attribute item to find the values of all corresponding DIAG_CODE attribute items in the comorbidity or complication catalog data set. Match the medical insurance code of the secondary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication catalog data set, eliminate the medical insurance code of the secondary diagnosis corresponding to the successfully matched data entry, and match the retained medical insurance code of the secondary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication exclusion data set. If the match is successful, use the value of the MCC_CC attribute item as the fourth digit of the DRG code.

[0035] The step S3 comprises the following steps:

[0036] S31. Form a DRG case group result set by combining the DRG codes of all grouped sets. The attribute items included in the DRG case group result set are DRGs_MDF, MAIN_DIAG, and MAIN_SURG. DRGs_MDF represents the DRG code grouped by the DRG grouper. MAIN_DIAG and MAIN_SURG respectively represent the medical insurance codes of the main diagnosis and the main surgical operation for obtaining the DRG code in the DRGs_MDF attribute item.

[0037] S32. Obtain the latest DRG case group measurement directory issued by the local medical insurance bureau. This DRG case group measurement directory includes DRG case group codes and local average hospitalization costs.

[0038] S33. Using the DRG code as the associated field, establish an associated data set between the DRG case group result set and the DRG case group measurement directory, and add the total hospitalization cost of the patient in the patient's hospitalization medical record homepage to this associated data set. The attribute items included in this associated data set are DRGs_MDF, MAIN_DIAG, MAIN_SURG, FEE_TOTAL, DRG case group code, and A_COST.

[0039] S34. Judge Whether it holds. If it holds, then screen out the data entries corresponding to A_COST from the associated data set, extract the DRG codes in the DRGs_BASE attribute and the DRG case group code attribute from this data entry, and judge whether the two are consistent. If they are consistent, the audit passes; if not, output this data entry. If it does not hold, the audit fails.

[0040] In the above solution, since the data entries in the associated data set include the medical insurance codes of the main diagnosis or surgical operation for generating the DRG code, when But the DRG code corresponding to A_COST is inconsistent with the original DRG code, outputting the data entries of the associated data set can not only directly display the correct DRG code, but also directly display the medical insurance code of the main diagnosis and the medical insurance code of the main surgery, which is convenient for the coding personnel to directly correct.

[0041] A DRG grouping automatic audit system includes:

[0042] A combination generation unit for extracting all diagnoses and surgical operations from the patient's hospitalization medical record homepage, generating a diagnosis data set based on all diagnoses, generating a surgical operation data set based on all surgical operations, and generating the Cartesian product set of the diagnosis data set and the surgical operation data set.

[0043] The coding generation unit is used to take each combination in the Cartesian set as the main diagnosis or surgical operation respectively, and the other diagnoses or surgical operations of the patient as the secondary diagnosis or surgical operation, to form a grouping set, and input it into the DRG grouper for grouping to obtain the DRG code of each grouping set;

[0044] The coding review unit is used to judge whether it holds. If it holds, further judge whether the DRG code corresponding to A_COST is the same as the original DRG code. If so, the review passes; if they are inconsistent or it does not hold, then the review fails; FEE_TOTAL is the total hospitalization cost of the patient, k is the weight, A_COST is the local average hospitalization cost, |·| represents taking the absolute value, and Y is the set threshold.

[0045] The DRG grouper includes:

[0046] The data exclusion module is used to judge whether the medical insurance code of the diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data and output the judgment result;

[0047] The preferred group decision module is used to, for each grouping set whose judgment result is no, judge whether it meets the preferred grouping conditions according to the attributes of the patient. If so, generate the ADRG grouping result based on the preferred group grouping rules;

[0048] The main diagnosis decision module is used to generate the ADRG grouping result based on the gender of the patient, the medical insurance code of the diagnosis and the medical insurance code of the surgical operation when the attribute judgment of the patient does not meet the preferred grouping conditions;

[0049] The DRG code determination module is used to take the ADRG grouping result as the first three digits of the DRG code, determine the fourth digit of the DRG code based on the medical insurance code of the secondary diagnosis or surgical operation, and adjust the generated four-digit DRG code according to the locally published DRG grouping directory to obtain the final DRG code.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1) By reviewing the original DRG code, it is possible to avoid incorrect medical insurance settlement costs caused by incorrect original DRG codes, and it can also provide more accurate data references for the medical insurance fund budget management.

[0052] 2) Through the review method of the present invention, when an incorrect original DRG code is found, an accurate DRG code can be directly provided for timely correction.

[0053] 3) Through the audit method of the present invention, not only can the DRG coding errors caused by the incorrect order of the main diagnosis or surgery and the secondary diagnosis or surgery be discovered, but also the DRG coding errors caused by the omission of the diagnosis or surgery in the front page of the inpatient medical record can be discovered.

[0054] 4) On the basis of realizing the audit, the audit process of the present invention is extremely simple and will not generate a large amount of computing, thus improving the audit efficiency.

[0055] 5) The audit process of the present invention is designed based on the real application scenario, and the generated DRG coding has high accuracy and will not generate a large number of useless DRG codings. Therefore, when the audit fails, the most accurate coding can be quickly found from a small number of codings.

[0056] 6) When but the DRG coding corresponding to A_COST is inconsistent with the original DRG coding, not only the correct DRG coding is directly displayed, but also the medical insurance coding of the main diagnosis and the medical insurance coding of the main surgery are directly displayed, which is convenient for the coding personnel to directly and quickly correct and improve the efficiency.

[0057] 7) The newly designed DRG coding grouper has high grouping efficiency and can quickly group out DRG codings, improving the grouping efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0059] Figure 1 It is a flowchart of the DRG grouping automatic audit method provided in the embodiment.

[0060] Figure 2 It is a refined flowchart of step S3 in the DRG grouping automatic audit method.

[0061] Figure 3 It is a flowchart of the grouping process implemented by the DRG grouper.

[0062] Figure 4 It is a block diagram of the composition of the DRG grouping automatic audit system provided in the embodiment.

[0063] Figure 5 It is a block diagram of the composition of the DRG grouper in the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to embodiments and drawings. The illustrative embodiments and descriptions thereof of the present invention are only used to explain the present invention and shall not be construed as limiting the present invention.

[0065] Please refer to Figure 1 , a DRG grouping automatic review method is provided in this embodiment, including the following steps:

[0066] S1, extract all diagnoses and surgical operations from the front page of the patient's inpatient medical record, generate a diagnosis dataset based on all diagnoses, generate a surgical operation dataset based on all surgical operations, and generate the Cartesian product set of the diagnosis dataset and the surgical operation dataset.

[0067] In specific implementation, obtain the front page of the patient's inpatient medical record, and the inpatient number, description of the diagnosis or surgical operation can be extracted from the front page of the inpatient medical record. The description includes the type identifier of the diagnosis or surgical operation, the medical insurance code of the diagnosis or surgical operation, the medical insurance name of the diagnosis or surgical operation, and the identifier indicating whether the diagnosis or surgical operation is the main diagnosis or surgical operation. Write these data items such as the inpatient number, description of the diagnosis or surgical operation into the corresponding attributes in the patient dataset to generate the patient dataset. It should be noted that there may be multiple diagnoses or surgical operations in the front page of the inpatient medical record. When generating the patient dataset, the descriptions of all diagnoses or surgical operations need to be recorded in the patient dataset.

[0068] As an example, the form of the patient dataset can be as shown in Table 1 below:

[0069] Table 1: Patient Dataset

[0070]

[0071] The INPATIENT_NO attribute represents the inpatient number. The purpose of extracting the inpatient number is to distinguish and identify which patient the patient dataset belongs to. The ICD_LX attribute represents the type identifier of the diagnosis or surgical procedure, which can facilitate the screening of the medical insurance codes for diagnoses or surgeries from the patient dataset. The ICD_CODE attribute represents the medical insurance code of the diagnosis or surgical procedure, and the ICD_NAME attribute represents the medical insurance name of the diagnosis or surgical procedure. The medical insurance code and the medical insurance name are corresponding and consistent. For DRG coding, either the medical insurance code or the medical insurance code can be selected. Since the medical insurance code is composed of numbers and / or letters, it is more conducive to direct computer recognition. Therefore, it is preferred to generate DRG codes based on the medical insurance code. The MAINICD_FLAG attribute represents the identifier indicating whether the diagnosis or surgical procedure is the main diagnosis or surgical procedure (for example, the value of "yes" is 1, and the value of "no" is 0). The DRG code is closely related to the order of the main diagnosis or surgery, secondary diagnosis or surgery. Providing this attribute item facilitates the identification of the main and secondary diagnoses or surgeries corresponding to the DRG code. The IS_MAIN attribute represents the identifier indicating whether the diagnosis or surgical procedure is the main diagnosis or main surgical procedure on the first page of the inpatient medical record (for example, the value of "yes" is 1, and the value of "no" is 0), that is, to identify whether it is the original DRG code.

[0072] The patient dataset can also include the FEE_TOTAL attribute, and FEE_TOTAL represents the total inpatient cost of the patient, which is convenient for use in the review process.

[0073] The patient dataset can also include gender and neonatal attributes. The neonatal attributes include birth age, birth weight, etc., which are convenient for direct invocation during subsequent DRG grouping by the DRG grouper.

[0074] In this step, a diagnosis dataset is generated based on all the diagnoses in the patient dataset, a surgery dataset is generated based on all the surgeries, and a Cartesian set of the diagnosis dataset and the surgery dataset is generated. Assuming there are 7 diagnoses and 7 surgeries in the patient dataset, then there are 49 combinations in the generated Cartesian set.

[0075] However, according to the coding rules formulated by the National Healthcare Security Administration, some diagnoses or surgeries do not participate in DRG coding. Therefore, the GROUP_YES attribute can also be added to the patient dataset. The GROUP_YES attribute represents the identifier indicating whether the diagnosis or surgical procedure is included in the grouping (for example, the value of "yes" is 1, and the value of "no" is 0). When initially establishing the dataset, the value of this attribute is defaulted to 1, and during the subsequent grouping process, the value of the GROUP_YES attribute is modified to 0 according to the exclusion rules (not participating in DRG grouping).

[0076] Therefore, in order to reduce the amount of data processing, the above step S1 can also be replaced by step S0, that is, extract all diagnoses and surgical operations from the first page of the patient's inpatient medical record, and determine whether the medical insurance code of the diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data. If so, mark the medical insurance code of this diagnosis or surgical operation as not included in the grouping, otherwise mark it as included in the grouping; generate a diagnosis dataset based on all diagnoses marked as included in the grouping, generate a surgical operation dataset based on all surgical operations marked as included in the grouping, and generate the Cartesian set of the diagnosis dataset and the surgical operation dataset.

[0077] For example, determine whether the medical insurance code of the diagnosis or surgical operation in the patient dataset belongs to the medical insurance diagnosis exclusion data. If so, modify the corresponding GROUP_YES attribute value in the patient dataset to 0, that is, discard the medical insurance code of this diagnosis or surgical operation so that it does not participate in the DRG grouping. That is to say, if the medical insurance code of a certain diagnosis belongs to the medical insurance diagnosis exclusion data, then modify its corresponding GROUP_YES attribute value to 0; if the medical insurance code of a certain surgical operation belongs to the medical insurance diagnosis exclusion data, then modify its corresponding GROUP_YES attribute value to 0.

[0078] For the medical insurance code of the diagnosis or surgical operation that does not belong to the medical insurance diagnosis exclusion data, keep its corresponding GROUP_YES attribute value as 1.

[0079] Since the medical insurance code of the diagnosis or surgical operation is in one-to-one correspondence with the medical insurance name of the diagnosis or surgical operation, the effect of judging the medical insurance code of the diagnosis or surgical operation is equivalent to that of judging the medical insurance name of the diagnosis or surgical operation.

[0080] Medical insurance diagnosis exclusion data refers to the data that does not participate in the DRG grouping and is formulated by the National Healthcare Security Administration. In specific implementation, an exclusion directory dataset can be established based on the medical insurance diagnosis exclusion data, and search in this exclusion directory dataset whether there is a data item with the same medical insurance code as the diagnosis or surgical operation in the patient dataset. If so, it is judged as medical insurance diagnosis exclusion data, otherwise it is judged as non-medical insurance diagnosis exclusion data.

[0081] As an example, the exclusion directory dataset can include two attribute items, ICD_CODE and ICD_NAME. For example, the form can be as shown in Table 2 below:

[0082] Table 2: Exclusion Directory Dataset

[0083]

[0084] In this step, directly assign the GROUP_YES attribute value of the diagnosis or surgical procedure belonging to the medical insurance diagnosis exclusion data to 0, that is, it does not participate in the subsequent grouping process. The generated diagnosis dataset and surgical dataset only need to contain all diagnoses and surgeries with the GROUP_YES attribute value of 1. That is, generate a diagnosis dataset based on all diagnoses with the GROUP_YES attribute value of 1 in the patient dataset, and generate a surgical dataset based on all surgical procedures with the GROUP_YES attribute value of 1 in the patient dataset. Subsequently, the amount of data processing can be reduced and time can be saved.

[0085] As in the above example, assume that the medical insurance code of a diagnosis belongs to the medical insurance diagnosis exclusion data. Then, the diagnosis dataset contains the medical insurance codes of 6 diagnoses, and the surgical dataset has the medical insurance codes of 7 surgeries. Then, the generated Cartesian set has 42 combinations. Compared with the original 49 combinations, there are 7 fewer combinations. In the subsequent step S2, 7 fewer DRG codes will be generated, which reduces the number of generated DRG codes, is more conducive to searching for the correct code, saves grouping time, and improves grouping efficiency.

[0086] S2. Respectively take each combination in the Cartesian set as the main diagnosis or surgical procedure (MAINDIAG_FLAG attribute is 1), and other diagnoses or surgical procedures of the patient (GROUP_YES attribute value is 1 in the patient dataset) as the secondary diagnosis or surgical procedure (MAINDIAG_FLAG attribute is 0), form a grouping set, and input it into the DRG grouper for grouping to obtain the DRG code of each grouping set.

[0087] S3. Judge Whether it holds. If it holds, further judge whether the DRG code corresponding to A_COST is the same as the original DRG code. If it is, the audit passes. If it is inconsistent or does not hold, the audit fails; FEE_TOTAL is the total hospitalization cost of the patient, k is the weight, A_COST is the local average hospitalization cost, |·| represents taking the absolute value, and Y is the set threshold.

[0088] Specifically, it can be referred to Figure 2 The above step S3 can include the following steps:

[0089] S31. Combine the DRG codes of all grouping sets to form a DRG group result set. The attribute items included in the DRG group result set are DRGs_MDF, MAIN_DIAG, and MAIN_SURG.

[0090] As an example, the form of the DRG group result set can be as shown in Table 3 below:

[0091] Table 3: DRG group result set

[0092]

[0093] DRGs_MDF represents the DRG code obtained by grouping with the DRG grouper; MAIN_DIAG and MAIN_SURG respectively represent the medical insurance codes of the main diagnosis and the main surgical operation for obtaining the DRG code in the DRGs_MDF attribute item.

[0094] The DRG case group result set may also include attributes DRGs_BASE, ADRG, DRGs_ORG, and ID. DRGs_BASE represents the grouping type, with a value of 1 indicating the original DRG grouping and a value of 0 indicating the newly generated DRG grouping, and its main function is to facilitate the identification of the original DRG code; ADRG represents the major diagnosis grouping, and ADRG must be within the grouping rules issued by the state. DRGs_ORG represents the DRG code generated by the grouping rules issued by the National Healthcare Security Administration. ID represents the unique identifier of all patient DRG grouping records (referred to as the primary key in the database). When adjusting the DRGs_ORG grouping to the DRGs_MDF grouping according to the DRG grouping directory issued by the local healthcare security bureau, the data can be updated through ID.

[0095] S32. Obtain the latest DRG case group measurement directory issued by the local healthcare security bureau, which includes DRG case group codes, local average hospitalization costs, and may also include local average hospitalization days.

[0096] The DRG case group measurement directory issued by the healthcare security bureau generally contains many data items, such as MDC codes, MDC names, ADRG codes, DRG codes, local average hospitalization costs, etc. In this method, only the DRG code and the local average hospitalization cost are concerned, and the local average hospitalization days can also be used as a reference. Therefore, after downloading the latest DRG case group measurement directory issued by the local healthcare security bureau, the directory can be directly used, or these 3 data items can be extracted from it to create a new directory.

[0097] S33. Establish an associated data set between the DRG case group result set and the DRG case group measurement directory with the DRG code as the associated field, and add the total hospitalization cost of the patient in the patient's hospitalization medical record front page to this associated data set. The attribute items included in this associated data set are DRGs_MDF, MAIN_DIAG, MAIN_SURG, FEE_TOTAL, DRG case group code, and A_COST.

[0098] As an example, the form of the associated data set can be as shown in Table 4 below:

[0099] Table 4: Associated Data Set

[0100]

[0101] Among them, INPATIENT_NO, DRGs_BASE, and DRGs_MDF come from the DRG disease group result set. DRG code, A_COST, and A_DAYS come from the data items corresponding to the same DRG code in the DRG disease group measurement directory. The DRG disease group code is the DRG code in the DRG disease group measurement directory. A_COST represents the local average hospitalization cost, and A_DAYS represents the local average length of hospitalization. The values of IN_DAYS and FEE_TOTAL are obtained from the patient's hospitalization medical record homepage or the patient dataset. IN_DAYS represents the length of hospitalization, and FEE_TOTAL represents the total hospitalization cost.

[0102] S34, judgment whether it holds. If it holds, then filter out from the associated dataset the data entry corresponding to A_COST, extract the DRG code in the DRGs_BASE attribute and the DRG disease group code attribute from this data entry, and judge whether the two are consistent. If they are consistent, the audit passes; if not, output this data entry; if it does not hold, the audit fails.

[0103] k is the weight. Generally speaking, the higher the hospital level, the larger the k value; A_COST is the local average hospitalization cost, and |·| represents taking the absolute value. Y is the set threshold, which can be dynamically adjusted according to the calculation result of k*A_COST.

[0104] The closer k*A_COST is to FEE_TOTAL, the smaller the deviation of the DRG code obtained by the DRG grouper grouping. Therefore, the filtered DRG code is the one that most conforms to the actual situation. In the specific implementation, if the filtered DRG code is the same as the original DRG code, the audit passes; if the filtered DRG code is different from the original DRG code, the audit fails, and the filtered DRG code can be pushed to the coder by means of text message or email for further verification.

[0105] If , that is, if the total hospitalization cost is not much different from the local average hospitalization cost, it indicates that the DRG code corresponding to A_COST is relatively accurate and reliable. If the selected DRG code is the same as the original DRG code, it means that the original DRG code grouping is accurate; conversely, if they are inconsistent, it means that the original DRG code may be inaccurate, possibly due to the wrong order of the main diagnosis or operation, that is, recording the secondary diagnosis or operation as the main diagnosis or operation. Therefore, through the review, it is possible to timely discover whether the code is incorrect and the cause of the error. Moreover, in step S2, DRG codes for all combinations of diagnoses and operations are generated. As long as there is no omission in the diagnoses or operations recorded in the front page of the inpatient medical record, the most correct DRG code will not be omitted, that is, through the review, a more accurate DRG code will surely be discovered. Moreover, in this method, the largest computational effort is to calculate the DRG codes for all combinations of diagnoses and operations. Relatively speaking, the overall number of combinations is small, that is, the number of generated DRG codes is small, which is more reference-worthy, that is, the DRG codes participating in the comparison are more accurate; and the computational effort is also small, thus improving the review efficiency.

[0106] If , that is, if the total hospitalization cost is much higher than the local average hospitalization cost, it is possible that there is an omission of diagnosis or surgical operation in the front page of the inpatient medical record, that is, some of the actually occurred diagnoses or surgical operations are not recorded in the front page of the inpatient medical record. Therefore, through the review, it is also possible to discover the situation where the omission of diagnosis or operation leads to incorrect DRG code grouping.

[0107] Therefore, in a more preferred solution, step S4 can also be included. When the review fails, if the DRG code corresponding to A_COST is inconsistent with the original DRG code, the DRG code or data entry corresponding to A_COST is output; if is not established, a prompt message indicating that there may be an omission of diagnosis or operation in the front page of the inpatient medical record is output.

[0108] Since the data entries in the associated dataset contain the medical insurance codes of the main diagnoses for generating DRG codes and the medical insurance codes of the main operations, when but the DRG code corresponding to A_COST is inconsistent with the original DRG code, outputting the data entries of the associated dataset can not only directly display the correct DRG code, but also directly display the medical insurance codes of the main diagnoses and the medical insurance codes of the main operations, which is convenient for the coding personnel to directly correct.

[0109] The DRG grouper involved in the above step S2 can use commercially available products. However, in this embodiment, in order to improve the grouping efficiency of the DRG grouper, a newly designed DRG grouper is provided. For reference, see Figure 3 , the DRG grouper realizes grouping through the following steps:

[0110] S20. Determine whether the medical insurance code of the diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data. If so, discard it; otherwise, proceed to step S21.

[0111] Medical insurance diagnosis exclusion data refers to the data that does not participate in DRG grouping. Therefore, if the medical insurance code of the diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data, it should be directly discarded without grouping.

[0112] It should be noted here that if this step has been performed in the above review method, that is, the diagnosis or surgical operation belonging to the medical insurance diagnosis exclusion data has been excluded, then this step can be skipped and directly proceed to step S21.

[0113] For each grouping set, perform the following same operations in steps S21 - S23:

[0114] S21. Determine whether it meets the priority grouping conditions based on the patient's attributes. If so, generate the ADRG grouping result according to the priority group grouping rules; otherwise, proceed to step S22.

[0115] Specifically, extract the patient's attributes from the patient's inpatient medical record homepage (or directly call from the patient dataset), including neonatal attributes, medical insurance codes of diagnoses, and medical insurance codes of surgical operations.

[0116] First, determine whether it meets the MDCA grouping conditions based on the medical insurance codes of diagnoses and surgical operations. If so, further obtain the ADRG grouping result according to the medical insurance codes of diagnoses and surgical operations.

[0117] If it does not meet the MDCA grouping conditions, determine whether it belongs to the MDCP grouping conditions based on the neonatal attributes. If so, further obtain the ADRG grouping result according to the neonatal attributes, medical insurance codes of diagnoses, and medical insurance codes of surgical operations.

[0118] If it does not meet the MDCP grouping conditions, determine whether it meets the MDCY grouping conditions based on the medical insurance codes of diagnoses or surgical operations. If so, obtain the ADRG grouping result according to the medical insurance codes of diagnoses or surgical operations.

[0119] If it does not meet the MDCY grouping conditions, determine whether it meets the MDCZ grouping conditions based on the medical insurance codes of diagnoses or surgical operations. If so, obtain the ADRG grouping result according to the medical insurance codes of diagnoses or surgical operations; otherwise, proceed to step S22.

[0120] S22. Generate the ADRG grouping result based on the patient's gender, medical insurance codes of diagnoses, and medical insurance codes of surgical operations.

[0121] Specifically, a main diagnosis category dataset and an ADRG grouping rule dataset can be established. The main diagnosis category dataset contains attributes DIAG_CODE, DIAG_NAME, and MDC. The ADRG grouping rule dataset contains attributes MDC, ADRG, ICD_CODE, and ICD_NAME. DIAG_CODE represents the medical insurance code of the diagnosis, DIAG_NAME represents the medical insurance name of the diagnosis, ICD_CODE represents the medical insurance code of the diagnosis or surgical operation, ICD_NAME represents the medical insurance name of the diagnosis or surgical operation, and MDC represents the first digit value of the ADRG grouping result.

[0122] The form of the main diagnosis category dataset can be as shown in Table 5 below:

[0123] Table 5: Main Diagnosis Category Dataset

[0124]

[0125] CATALOG represents the directory, DIAG_CODE represents the medical insurance code of the diagnosis, and DIAG_NAME represents the medical insurance name of the diagnosis.

[0126] The ADRG grouping rule dataset may also include attributes CLASS (classification, with values of diagnosis or surgery), DIAG_SURG, GROUP_ID, and MAINDIAG_FLAG. As an example, the form of the ADRG grouping rule dataset is as shown in Table 6 below:

[0127] Table 6: ADRG Grouping Rule Dataset

[0128]

[0129] First, search for the medical insurance codes of the same diagnosis in the main diagnosis category dataset, and determine the value of MDC corresponding to the medical insurance codes of the same diagnosis found as the first digit value of the ADRG grouping result. Then, with the value of the MDC attribute item as the first digit of the determined DRG code and the value of the ICD_CODE attribute item as the medical insurance code of the diagnosis or surgical operation as the condition, search for matching data entries in the ADRG grouping rule dataset, and use the value of the ADRG attribute item in the first sorted data entry as the ADRG grouping result.

[0130] More specifically, in the first step, the value of the MDC attribute item is taken as the first digit of the determined ADRG grouping result; in the second step, in the ADRG grouping rule dataset, all ADRG groupings are composed of the attributes of MDC, ADRG, DIAG_SURG, GROUP_ID, and MAINDIAG_FLAG; in the third step, a comparison is established between the valid medical insurance codes of diagnoses or surgical operations in the Cartesian set and the ICD_CODE and MAINDIAG_FLAG in the ADRG grouping rule dataset to obtain the datasets of MDC, ADRG, CLASS, DIAG_SURG, GROUP_ID, and INPATIENT_NO (the unique inpatient identifier of the patient); in the fourth step, all the ADRG groupings obtained in the second step are compared with the datasets obtained in the third step through the attributes of MDC, ADRG, CLASS, DIAG_SURG, and GROUP_ID, and the number of valid records is calculated respectively. When there are multiple ADRG attribute data volumes with the same amount, they are sorted in ascending order according to the ADRG code, and the value of the ADRG attribute item in the first data entry after sorting is taken as the ADRG grouping result.

[0131] For example, for the main diagnosis Z51.003, by querying the MDC as R in the main diagnosis major category dataset, among the results corresponding to the main surgical operation 92.2400x003, the only MDC belonging to R is RL1. Therefore, the ADRG grouping result is determined to be RL1.

[0132] Another example is that for the main diagnosis Z51.801, which belongs to both RU1 and RN2. According to the DRG grouping coding rule, RN2 is preferentially selected because the RN2 coding is sorted before RU1. RN2 belongs to the surgical DRG coding, and RU1 belongs to the medical DRG coding. For the same main diagnosis, the resource consumption of surgical operations is greater than that of medical operations.

[0133] After going through steps S21 and S22, an ADRG grouping that conforms to the grouping rules of the "National Medical Insurance Fund's Payment by Disease Group (DRG) Grouping Scheme (Version 2.0)" is generated.

[0134] S23: Take the ADRG grouping result as the first three digits of the DRG code, determine the fourth digit of the DRG code based on the medical insurance code of the secondary diagnosis, and adjust the generated four-digit DRG code based on the locally released DRG grouping directory to obtain the final DRG code.

[0135] In specific implementation, a complication or complication catalog dataset and a complication or complication exclusion dataset can be established based on the complication or complication (CC) and serious complication or complication (MCC) lists formulated by the National Healthcare Security Administration. The complication or complication catalog dataset contains the attributes DIAG_CODE, DIAG_NAME, VALUE, and MCC_CC. The complication or complication exclusion dataset contains the attributes DIAG_CODE, DIAG_NAME, and VALUE. VALUE represents the complication or complication number, and MCC_CC represents the complication category (according to the DRG coding rules, MCC takes the value 1, CC takes the value 3, and if there is no corresponding value, it takes 5).

[0136] As an example, the form of the comorbidity or complication list data set may be as shown in Table 7 below:

[0137] Table 7: Comorbidity or complication catalog dataset

[0138]

[0139] The format of the comorbidity or complication exclusion data set can be shown in Table 8 below:

[0140] Table 8: Datasets Excluded from Comorbidities or Complications

[0141]

[0142] Match the medical insurance code of the primary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication exclusion data set. If the match is successful, obtain the value of the VALUE attribute item, and use the value of the VALUE attribute item to find the values of all corresponding DIAG_CODE attribute items in the comorbidity or complication catalog data set. Match the medical insurance code of the secondary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication catalog data set, eliminate the medical insurance code of the secondary diagnosis corresponding to the successfully matched data entry, and match the retained medical insurance code of the secondary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication exclusion data set. If the match is successful, use the value of the MCC_CC attribute item as the fourth digit of the DRG code.

[0143] Since the DRG grouping catalogs issued by local medical security agencies vary according to the actual diagnosis and treatment conditions in their respective regions, for example, the DRG grouping catalog corresponding to the ADRG code RL1, radiotherapy for malignant and proliferative diseases (external irradiation) in Sichuan Province, only includes RL11 and RL15. Therefore, the generated DRG codes need to be adjusted according to the DRG grouping catalog issued by the local medical security bureau. The adjustment method can be the method of rounding up to the nearest mantissa. For example, if the generated DRG code is RL13, since there is no corresponding DRG grouping issued in Sichuan Province, RL13 will be adjusted to RL15. Another example is that adjustments and alignments also need to be made according to the exception grouping rules issued by local medical security agencies. For example, in Sichuan Province, for some ADRG grouped patients, whether they "are accompanied by intensive care" is taken as the digit 8, and separate alignment processing needs to be carried out on the basis of the standard grouping results.

[0144] It should be noted that for the above DRG grouper, the relevant regulations of the National Healthcare Security Administration on DRG grouping rules must be strictly implemented. The present invention does not change the grouping rules, and according to the relevant regulations, the grouping rules cannot be changed either. That is to say, the grouping results of the DRG grouper of the present invention are the same as those of other DRG groupers, but the implementation methods of the grouping process are different. The present invention can improve the grouping efficiency by reasonably designing the implementation methods of those grouping rules, that is, the composition steps of the grouping method and the execution order of the steps.

[0145] See Figure 4 , based on the same inventive concept, the embodiment of the present invention also provides a DRG grouping automatic audit system, including:

[0146] A combination generation unit, configured to extract all diagnoses and surgical operations from the front page of the patient's inpatient medical record, generate a diagnosis dataset based on all diagnoses, generate a surgical operation dataset based on all surgical operations, and generate the Cartesian product set of the diagnosis dataset and the surgical operation dataset;

[0147] A code generation unit, configured to use each combination in the Cartesian product set as the primary diagnosis or surgical operation respectively, and the patient's other diagnoses or surgical operations as secondary diagnoses or surgical operations, to form a grouping set, input it into the DRG grouper for grouping, and obtain the DRG code of each grouping set;

[0148] A code audit unit, configured to judge whether it holds. If it holds, further judge whether the DRG code corresponding to A_COST is consistent with the original DRG code. If so, the audit passes; if not or does not hold, the audit fails; FEE_TOTAL is the total inpatient cost of the patient, k is the weight, A_COST is the local average inpatient cost, |·| represents taking the absolute value, and Y is the set threshold.

[0149] Refer to Figure 5 , the above DRG grouper includes:

[0150] A data exclusion module, configured to determine whether the medical insurance code of a diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data, and output a judgment result;

[0151] An optimal group decision module, configured to, for each group set whose judgment result is negative, determine whether it meets the priority grouping condition according to the attributes of the patient, and if so, generate an ADRG grouping result based on the priority group grouping rule;

[0152] A primary diagnosis decision module, configured to generate an ADRG grouping result based on the gender of the patient, the medical insurance code of the diagnosis, and the medical insurance code of the surgical operation when the attribute judgment of the patient does not meet the priority grouping condition;

[0153] A DRG code determination module, configured to use the ADRG grouping result as the first three digits of the DRG code, determine the fourth digit of the DRG code based on the medical insurance code of the secondary diagnosis or surgical operation, and adjust the generated four-digit DRG code according to the locally published DRG grouping directory to obtain the final DRG code.

[0154] The implementation manners of each component unit of the audit system and each component module of the DRG grouper can be referred to the relevant descriptions in the foregoing method steps. For the sake of brevity, they are not described herein again.

[0155] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for automatic review of DRG grouping, characterized in that, It includes the following steps: S1. Extract all diagnoses and surgical operations from the front page of the patient's inpatient medical record. Generate a diagnosis dataset based on all diagnoses, generate a surgical operation dataset based on all surgical operations, and generate the Cartesian product set of the diagnosis dataset and the surgical operation dataset; S2. Respectively take each combination in the Cartesian product set as the primary diagnosis or surgical operation, and take the patient's other diagnoses or surgical operations as secondary diagnoses or surgical operations to form a grouping set, and input it into the DRG grouper for grouping to obtain the DRG code of each grouping set; S3, Determine whether it holds. If it holds, further determine whether the DRG code corresponding to A_COST is the same as the original DRG code. If so, the audit passes; if they are inconsistent or it does not hold, the audit fails; FEE_TOTAL is the total hospitalization cost of the patient, k is the weight, A_COST is the local average hospitalization cost, |·| represents taking the absolute value, and Y is the set threshold.

2. The DRG grouping automatic review method according to claim 1, wherein, It also includes step S4. When the review fails, if the DRG code corresponding to A_COST is inconsistent with the original DRG code, the DRG code corresponding to A_COST is output; if it does not hold, a prompt message indicating possible missing diagnoses or surgeries in the front page of the inpatient medical record is output.

3. The DRG grouping automatic review method according to claim 1, characterized in that The step S1 is replaced by step S0, and the processing of step S0 includes: Extract all diagnoses and surgical operations from the front page of the patient's inpatient medical record, and determine whether the medical insurance code of the diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data. If so, mark the medical insurance code of this diagnosis or surgical operation as not included in the grouping, otherwise mark it as included in the grouping; Generate a diagnosis dataset based on all diagnoses marked as included in the grouping, generate a surgical operation dataset based on all surgical operations marked as included in the grouping, and generate the Cartesian product set of the diagnosis dataset and the surgical operation dataset.

4. The automatic review method for DRG grouping according to claim 2 or 3, characterized in that, For each grouping set, the DRG grouper in S2 performs the following operations: S21. Judge whether it meets the priority grouping condition according to the patient's attributes. If so, generate the ADRG grouping result based on the priority group grouping rule, otherwise go to step S22; S22. Generate the ADRG grouping result based on the patient's gender, the medical insurance code of the diagnosis, and the medical insurance code of the surgical operation; S23. Take the ADRG grouping result as the first three digits of the DRG code, determine the fourth digit of the DRG code based on the medical insurance code of the secondary diagnosis, and adjust the generated four-digit DRG code according to the locally released DRG grouping directory to obtain the final DRG code.

5. The DRG grouping automatic review method according to claim 4, wherein The step S21 includes the following processing: Extract the patient's attributes from the front page of the patient's inpatient medical record, including neonatal attributes, the medical insurance code of the diagnosis, and the medical insurance code of the surgical operation; First, judge whether it meets the MDCA grouping condition according to the medical insurance code of the diagnosis and the medical insurance code of the surgical operation. If so, further obtain the ADRG grouping result according to the medical insurance code of the diagnosis and the medical insurance code of the surgical operation; If it does not meet the MDCA grouping condition, then judge whether it belongs to the MDCP grouping condition according to the neonatal attributes. If so, further obtain the ADRG grouping result according to the neonatal attributes, the medical insurance code of the diagnosis, and the medical insurance code of the surgical operation; If it does not meet the MDCP grouping condition, then judge whether it meets the MDCY grouping condition according to the medical insurance code of the diagnosis or surgical operation. If so, obtain the ADRG grouping result according to the medical insurance code of the diagnosis or surgical operation; If it does not meet the MDCY grouping condition, then judge whether it meets the MDCZ grouping condition according to the medical insurance code of the diagnosis or surgical operation. If so, obtain the ADRG grouping result according to the medical insurance code of the diagnosis or surgical operation, otherwise go to step S22.

6. The DRG grouping automatic review method according to claim 4, wherein The step S22 includes the following processing: Create a master diagnosis category dataset and an ADRG grouping rule dataset. The master diagnosis category dataset contains attributes DIAG_CODE, DIAG_NAME, and MDC. The ADRG grouping rule dataset contains attributes MDC, ADRG, ICD_CODE, and ICD_NAME. DIAG_CODE represents the medical insurance code of the diagnosis, DIAG_NAME represents the medical insurance name of the diagnosis, ICD_CODE represents the medical insurance code of the diagnosis or surgical operation, ICD_NAME represents the medical insurance name of the diagnosis or surgical operation, and MDC represents the first digit value of the ADRG grouping result; Search the master diagnosis category dataset, and use the value of the MDC attribute item where the gender and the medical insurance code of the main diagnosis match successfully as the first digit of the ADRG grouping result; With the value of the MDC attribute item as the first digit of the determined ADRG grouping result and the value of the ICD_CODE attribute item as the medical insurance code of the diagnosis or surgical operation as the condition, search for matching data entries in the ADRG grouping rule dataset, and use the value of the ADRG attribute item in the first sorted data entry as the ADRG grouping result.

7. The DRG grouping automatic review method according to claim 4, wherein The step S23 includes the following processing: Create a comorbidity or complication catalog dataset and a comorbidity or complication exclusion dataset. The comorbidity or complication catalog dataset contains attributes DIAG_CODE, DIAG_NAME, VALUE, and MCC_CC. The comorbidity or complication exclusion dataset contains attributes DIAG_CODE, DIAG_NAME, and VALUE. VALUE represents the comorbidity or complication serial number, and MCC_CC represents the complication category; Match the medical insurance code of the main diagnosis with the DIAG_CODE attribute item in the comorbidity or complication exclusion dataset. If the match is successful, obtain the value of the VALUE attribute item, and find all the values of the corresponding DIAG_CODE attribute items in the comorbidity or complication catalog dataset through the value of the VALUE attribute item. Match the medical insurance code of the secondary diagnosis with the DIAG_CODE attribute item in the comorbidity or complication catalog dataset, and eliminate the medical insurance codes of the secondary diagnoses corresponding to the successfully matched data entries. Match the remaining medical insurance codes of the secondary diagnoses with the DIAG_CODE attribute item in the comorbidity or complication exclusion dataset. If the match is successful, use the value of the MCC_CC attribute item as the fourth digit of the DRG code.

8. The DRG grouping automatic review method according to claim 1, wherein The step S3 includes the following steps: S31, Compose the DRG codes of all grouping sets into a DRG case mix result set. The attributes included in the DRG case mix result set are DRGs_MDF, MAIN_DIAG, and MAIN_SURG. DRGs_MDF represents the DRG code grouped by the DRG grouper, and MAIN_DIAG and MAIN_SURG respectively represent the medical insurance codes of the main diagnosis and the main surgical operation for the DRG code in the DRGs_MDF attribute item; S32. Obtain the latest DRG case mix measurement catalog released by the local medical insurance bureau. The DRG case mix measurement catalog includes DRG case mix codes and local average hospitalization costs. S33. Using the DRG code as the associated field, establish an associated dataset between the DRG case mix result set and the DRG case mix measurement catalog, and add the total hospitalization cost of the patient in the patient's hospitalization medical record front page to this associated dataset. The attribute items included in this associated dataset are DRGs_MDF, MAIN_DIAG, MAIN_SURG, FEE_TOTAL, DRG case mix code, and A_COST. S34, Determine whether it holds. If it holds, filter out from the associated dataset the corresponding data entries, extract the DRG codes from the DRGs_BASE attribute and the DRG group coding attribute in these data entries, and determine whether the two are consistent. If they are consistent, the audit passes; if not, output these data entries. If it does not hold, the audit fails.

9. A DRG grouping automatic audit system, characterized in that, It includes: A combination generation unit for extracting all diagnoses and surgical operations from the patient's hospitalization medical record front page, generating a diagnosis dataset based on all diagnoses, generating a surgical operation dataset based on all surgical operations, and generating a Cartesian product set of the diagnosis dataset and the surgical operation dataset. A coding generation unit for using each combination in the Cartesian product set as the main diagnosis or surgical operation respectively, and using the patient's other diagnoses or surgical operations as secondary diagnoses or surgical operations to form a grouping set, and inputting it into the DRG grouper for grouping to obtain the DRG code of each grouping set. The coding review unit is used to judge whether it holds. If it holds, it further judges whether the DRG code corresponding to A_COST is the same as the original DRG code. If so, the review passes; if they are inconsistent or it does not hold, the review fails; FEE_TOTAL is the total hospitalization cost of the patient, k is the weight, A_COST is the local average hospitalization cost, |·| represents taking the absolute value, and Y is the set threshold value.

10. The DRG grouping automatic review system according to claim 9, wherein The DRG grouper includes: A data exclusion module for determining whether the medical insurance code of a diagnosis or surgical operation belongs to the medical insurance diagnosis exclusion data and outputting a judgment result. An optimal group decision module for, for each grouping set with a judgment result of no, judging whether it meets the priority grouping condition according to the patient's attributes, and if so, generating an ADRG grouping result based on the priority group grouping rule. A main diagnosis decision module for generating an ADRG grouping result based on the patient's gender, the medical insurance code of the diagnosis, and the medical insurance code of the surgical operation when the patient's attributes do not meet the priority grouping condition. A DRG code determination module for using the ADRG grouping result as the first three digits of the DRG code, determining the fourth digit of the DRG code based on the medical insurance code of the secondary diagnosis or surgical operation, and adjusting the generated four-digit DRG code based on the locally released DRG grouping directory to obtain the final DRG code.

Images