A treatment process grouping method and device for insurance intelligent claim settlement

Abstract

The application discloses a treatment process grouping method and device for intelligent insurance claim settlement, which comprises three-stage classification of various bills in a treatment process and exclusion of irrelevant bills; bill identification and extraction of structured information from the remaining bills, wherein the structured information comprises table header information and detailed information; and multi-period grouping of the bills in the treatment process according to the bill categories and the structured information, so as to finally obtain outpatient final grouping and inpatient final grouping. The method and device can gradually arrange disordered medical invoices into ordered structured data meeting the needs of insurance claim settlement, so as to be used for assisting intelligent insurance claim settlement of medical insurance.

Description

Technical Field

[0001] This invention belongs to the field of intelligent insurance claims technology, specifically relating to a treatment process grouping method and apparatus for intelligent insurance claims. Background Technology

[0002] Insurance claims settlement refers to the act of an insurance company fulfilling its liability for compensation or payment according to the contract when an insured event occurs that results in loss of property or personal injury to the insured, or when other insured events stipulated in the policy occur and insurance benefits are due. It directly reflects the insurance function and the fulfillment of insurance responsibilities. Simply put, insurance claims settlement is the act of the insurer processing the claim submitted by the insured after a risk event occurs to the insured object. In insurance operations, insurance claims settlement is the concrete manifestation of the insurance compensation function. With increasing public awareness of insurance and a growing number of insurance accidents and claims, manual insurance claims settlement can no longer meet the needs. Therefore, to improve the efficiency of insurance claims settlement, intelligent insurance claims settlement has emerged.

[0003] Intelligent medical insurance claims processing is a type of intelligent insurance claims processing, primarily used for medical claims. Due to the special nature of medicine and the complexity of diseases, the verification of medical claims is exceptionally complex. Although existing technologies have proposed various claims systems and methods, such as the medical insurance claims management system disclosed in patent document CN106295152A and the medical insurance claims method, device, equipment, and storage medium disclosed in patent document CN111222997A, none of these medical claims processing methods consider the importance of treatment process grouping for claims processing.

[0004] In intelligent medical insurance claims processing, grouping medical cases according to treatment process plays a crucial role. First, user-uploaded receipts may contain duplicates or those irrelevant to the claim. Irrelevant receipts need to be filtered out, while duplicates only need to be calculated once. Second, insurance policies generally have detailed rules governing the insured's treatment process. Different types of insurance have different constraints regarding treatments unrelated to the policy terms, treatments outside the insurance's effective period, the hospital's level, and the medical methods used. This necessitates that intelligent claims processing obtain the treatment process information of the case, i.e., grouping the treatment process. Summary of the Invention

[0005] In view of the above, the purpose of this invention is to provide a treatment process grouping method and apparatus for intelligent insurance claims, which can achieve accurate grouping of treatment processes to assist in intelligent medical insurance claims.

[0006] To achieve the above-mentioned objectives, an embodiment provides a method for grouping treatment processes for intelligent insurance claims, comprising the following steps:

[0007] The various receipts during the treatment process are classified in three stages, and irrelevant receipts are excluded;

[0008] The remaining invoices are identified and their structured information is extracted, including header information and detailed information.

[0009] The invoices during the treatment process are grouped in multiple phases according to the invoice type and structured information, resulting in the final outpatient group and the final inpatient group.

[0010] Preferably, the classification of various documents during the treatment process is carried out in three stages, including:

[0011] First-stage classification: Using image-based and text-based classification methods, various receipts during the treatment process are classified into medical invoices, expense lists, reimbursement settlement statements, medical records, and irrelevant receipts;

[0012] Two-stage classification: Using keyword matching and rule-based methods, medical invoices are classified into outpatient medical invoices and inpatient medical invoices; expense lists are classified into outpatient expense lists and inpatient expense lists; reimbursement settlement forms are classified into outpatient reimbursement settlement forms and inpatient reimbursement settlement forms; and medical records are classified into outpatient medical records and inpatient medical records.

[0013] Three-stage classification: Outpatient medical invoices are classified into outpatient electronic medical invoices and outpatient general medical invoices using keyword matching and rule-based methods, and inpatient medical invoices are classified into inpatient electronic medical invoices and inpatient general medical invoices.

[0014] Preferably, the extraction of structured information from the remaining invoices includes:

[0015] For each invoice, a recognition engine is used to identify it, obtaining its location and text information. Then, based on the location and text information, and by using layout analysis and querying the financial database, the structured information of the invoice is extracted.

[0016] Preferably, the step of extracting structured information from invoices based on location and text information, and by employing layout analysis and querying a financial database, includes the following steps:

[0017] First, layout analysis is used to obtain structured information for all invoices, including: determining the header based on the location information, extracting the text information corresponding to the header position to obtain the header information, determining the detail positions based on the location information, and obtaining the detail information based on the text information corresponding to the detail positions;

[0018] Then, it determines whether the invoice is an outpatient electronic medical invoice or an inpatient electronic medical invoice. If it is not an outpatient electronic medical invoice or an inpatient electronic medical invoice, it directly outputs the structured information obtained from the layout analysis. If it is an outpatient electronic medical invoice or an inpatient electronic medical invoice, it queries the Ministry of Finance database based on the five elements of the header information obtained from the layout analysis, including the invoice code, invoice number, invoice date, verification code, and total amount, to obtain the precise structured information of the outpatient electronic medical invoice and the inpatient electronic medical invoice.

[0019] Preferably, the header information is data in key-value pair format, including consultation date, hospitalization date, invoice code, invoice number, invoice date, verification code, fund payment, out-of-pocket, self-paid, name, and hospital name.

[0020] Preferably, the detailed information is data in list format, including item name, amount, medical insurance type, ratio, unit price, and quantity.

[0021] Preferably, the step of grouping the treatment process tickets into multiple phases according to ticket type and structured information includes:

[0022] (1) According to the type of invoice, the invoices in medical cases are divided into 8 initial groups, namely: outpatient medical invoice initial group, inpatient medical invoice initial group, outpatient expense list initial group, inpatient expense list initial group, outpatient medical record initial group, inpatient medical record initial group, outpatient settlement statement initial group and inpatient settlement statement initial group;

[0023] (2) For the initial group of non-empty outpatient-related documents, the documents are grouped according to the visit date or invoice number in the header information according to certain rules to obtain the second-stage outpatient group, including the second-stage outpatient case group, the second-stage outpatient settlement statement group, the second-stage outpatient expense list group, and the second-stage outpatient medical invoice group; For the initial group of non-empty inpatient-related documents, the documents are grouped according to the admission date or invoice number in the header information according to certain rules to obtain the second-stage inpatient group, including the second-stage inpatient case group, the second-stage inpatient settlement statement group, the second-stage inpatient expense list group, and the second-stage inpatient medical invoice group;

[0024] (3) For each second-phase group, calculate the similarity sim1 between any two tickets based on the structured information. If the similarity sim1 is greater than the given threshold thresh1, output the similarity sim2 between the text information and relative position information of the two tickets. If the similarity sim2 is greater than the given threshold thresh2, determine that the two tickets are duplicate tickets and add them to the duplicate ticket set. Keep only the structured information of one ticket in the duplicate ticket set and clear the structured information of the other tickets to obtain the second-phase group after clearing the duplicate tickets.

[0025] (4) For the second-phase group after clearing duplicate invoices, the second-phase group of outpatient medical invoices and the second-phase group of outpatient expense list are matched and grouped according to the visit date or invoice number in the header information to obtain the third-phase group of outpatient medical invoices. The second-phase group of inpatient medical invoices and the second-phase group of inpatient expense list are matched and grouped according to the admission date or invoice number in the header information to obtain the third-phase group of inpatient medical invoices.

[0026] (5) For outpatient expense list receipts that are grouped separately due to missing invoice numbers or missing consultation dates in step (4), if the difference between the total amount of a single outpatient medical invoice in the outpatient three-phase group and the detailed amount of at least one outpatient expense list in the group is equal to the detailed amount of the outpatient expense list receipts that are grouped separately, then the outpatient expense list receipts that are grouped separately will be merged into the corresponding outpatient three-phase group; For inpatient expense list receipts that are grouped separately due to missing invoice numbers or missing hospitalization dates in step (4), if the difference between the total amount of a single inpatient medical invoice in the inpatient three-phase group and the detailed amount of at least one inpatient expense list in the group is equal to the detailed amount of the inpatient expense list receipts that are grouped separately, then the inpatient expense list receipts that are grouped separately will be merged into the corresponding inpatient three-phase group;

[0027] (6) For each three-phase group obtained in steps (4) and (5), if the sum of the detailed amounts of all expense lists in the three-phase group is equal to the sum of the total amount of all medical invoices, then clear the detailed information in the medical invoices in the three-phase group and clear the header information in the expense list; if the medical invoices in the three-phase group are electronic medical invoices, and the expense list corresponding to the electronic medical invoices is obtained by querying the Ministry of Finance database, then according to the text information and location information, match and overwrite the detailed information of the expense list returned by the Ministry of Finance database to obtain the four-phase group by analyzing the layout. The three-phase group includes the outpatient three-phase group and the inpatient three-phase group, the expense list includes the outpatient expense list and the inpatient expense list, the medical invoices include the outpatient medical invoices and the inpatient medical invoices, and the four-phase group includes the outpatient four-phase group and the inpatient four-phase group.

[0028] (7) Match and group the outpatient settlement statement group II and group IV according to the date of visit or invoice number to obtain the outpatient group V; match and group the inpatient settlement statement group II and group IV according to the admission date or invoice number to obtain the inpatient group V;

[0029] (8) Match and group the outpatient medical record phase II group and outpatient phase V group according to the date of visit or invoice number to obtain the final outpatient group; match and group the inpatient medical record phase II group and inpatient phase V group according to the admission date or invoice number to obtain the final inpatient group.

[0030] Preferably, in step (2), when the initial group related to outpatient visits and the initial group related to hospitalization are grouped in two phases according to certain rules, tickets with the same visit date are grouped together, tickets with the same hospitalization date are grouped together, and tickets with the same invoice number are grouped together.

[0031] Preferably, in step (7), the outpatient settlement statement group II and outpatient settlement statement group with the same visit date or invoice number are matched to form outpatient settlement statement group V, and the inpatient settlement statement group II and inpatient settlement statement group with the same hospitalization date or invoice number are matched to form inpatient settlement statement group V;

[0032] Preferably, in step (8), the outpatient medical record group of phase II and outpatient medical record group with the same visit date or invoice number are matched to form the outpatient final group, and the inpatient medical record group of phase II and inpatient medical record group with the same hospitalization date or invoice number are matched to form the inpatient final group.

[0033] To achieve the above-mentioned objectives, the embodiments also provide a treatment process grouping device for intelligent insurance claims, including a classification module, a structural information extraction module, and a multi-stage grouping module.

[0034] The classification module is used to classify various documents during the treatment process in three stages and exclude irrelevant documents;

[0035] The structure information extraction module is used to identify the remaining invoices and extract structured information, which includes header information and detail information.

[0036] The multi-stage grouping module is used to group the bills in the treatment process into multiple stages according to the bill type and structured information, and finally obtain the outpatient final grouping and the inpatient final grouping.

[0037] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0038] By employing automation, the system connects invoice recognition and claims rules in the insurance claims process, streamlining the entire intelligent insurance claims chain from invoice recognition to rule matching and cost calculation. The process utilizes invoice classification, invoice recognition, structured information extraction, and treatment process grouping to gradually organize disordered medical invoices into structured data that meets the needs of insurance claims, thereby assisting in intelligent medical insurance claims processing. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart of a treatment process grouping method for intelligent insurance claims provided in the embodiments;

[0041] Figure 2 This is a flowchart of ticket classification provided in the embodiment;

[0042] Figure 3 This is a flowchart of structured information extraction provided in the embodiment;

[0043] Figure 4 This is a flowchart of the ticket grouping provided in the embodiment;

[0044] Figure 5 This is a schematic diagram of the treatment process grouping device for intelligent insurance claims provided in the embodiment. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of this invention.

[0046] This embodiment provides a method for grouping treatment processes in intelligent insurance claims. The method includes classifying medical invoices in medical cases to obtain invoice category information; after classification, a recognition engine is used to obtain the location and text information of the invoices; further, structured information of the invoices is obtained through layout analysis and searching the Ministry of Finance database; finally, based on the invoice category and structured information, the invoices are grouped according to the treatment process. Specifically, as follows... Figure 1 As shown, it includes the following steps:

[0047] Step 1: Classify all receipts during the treatment process into three stages and exclude irrelevant receipts.

[0048] In this embodiment, various receipts are obtained during the insured's treatment process, and these are submitted along with the claim. The grouping system processes different receipts differently; therefore, the receipts for the entire claim case need to be categorized first. This embodiment employs a three-stage classification method based on image and text classification, dividing the receipts into 11 categories: irrelevant receipts, outpatient general medical invoices, outpatient electronic medical invoices, outpatient expense lists, outpatient reimbursement settlement statements, outpatient medical records, inpatient general medical invoices, inpatient electronic medical invoices, inpatient expense lists, inpatient reimbursement settlement statements, and inpatient medical records. Figure 2 As shown, the specific three-stage classification includes:

[0049] First-stage classification: Using image-based and text-based classification methods, various receipts during the treatment process are classified into medical invoices, expense lists, reimbursement settlement statements, medical records, and irrelevant receipts;

[0050] Two-stage classification: Using keyword matching and rule-based methods, medical invoices are classified into outpatient medical invoices and inpatient medical invoices; expense lists are classified into outpatient expense lists and inpatient expense lists; reimbursement settlement forms are classified into outpatient reimbursement settlement forms and inpatient reimbursement settlement forms; and medical records are classified into outpatient medical records and inpatient medical records.

[0051] Three-stage classification: Outpatient medical invoices are classified into outpatient electronic medical invoices and outpatient general medical invoices using keyword matching and rule-based methods, and inpatient medical invoices are classified into inpatient electronic medical invoices and inpatient general medical invoices.

[0052] In this embodiment, irrelevant invoices refer to invoices other than the following 10 types of invoices: outpatient general medical invoices, outpatient electronic medical invoices, outpatient expense lists, outpatient reimbursement settlement forms, outpatient medical records, inpatient general medical invoices, inpatient electronic medical invoices, inpatient expense lists, inpatient reimbursement settlement forms, and inpatient medical records. These irrelevant invoices are then removed.

[0053] In this embodiment, the first-stage classification uses an image classification and text classification method. Specifically, an image classification network based on ResNet18 is used to classify receipts into medical invoices, expense lists, reimbursement settlement forms, medical records, and irrelevant receipts. After the recognition engine performs text recognition on the receipts, a text classification network based on Fasttext is used to perform binary classification of receipts / irrelevant receipts. Receipts whose images are classified as medical invoices, expense lists, reimbursement settlement forms, or medical records, and whose text is classified as irrelevant receipts, are set as irrelevant receipts.

[0054] In the keyword matching and rule-based method used for the two-stage classification, outpatient, emergency, and consultation dates are used as keywords for outpatient invoices, while inpatient, admission, and discharge are used as keywords for inpatient invoices. Based on whether the keywords are matched, medical invoices are divided into outpatient medical invoices and inpatient medical invoices, expense lists are divided into outpatient expense lists and inpatient expense lists, reimbursement settlement forms are divided into outpatient reimbursement settlement forms and inpatient reimbursement settlement forms, and medical records are divided into outpatient medical records and inpatient medical records.

[0055] In the keyword matching and rule-based method used for the three-stage classification, "electronic" is used as the keyword for electronic invoices. Based on whether the keyword is matched, outpatient medical invoices are divided into outpatient electronic medical invoices and outpatient regular medical invoices, and inpatient medical invoices are divided into inpatient electronic medical invoices and inpatient regular medical invoices.

[0056] Step 2: Perform ticket identification and extract structured information from the remaining tickets, including header information and detailed information.

[0057] In this embodiment, the grouping process relies on structured information on the invoices, such as the date of visit, the date of hospitalization, the invoice number, and the date of issuance. Therefore, after obtaining the category of the invoices, it is necessary to identify invoices of other categories, excluding irrelevant invoices, and obtain the structured information of each invoice. This embodiment divides the structured information of the invoices into header information and detail information. The header information is data in key-value pair format, including data such as the date of visit, the date of hospitalization, the invoice code, the invoice number, the date of issuance, and the check code. The detail information is data in list format, containing attributes such as the item name, the amount, and the medical insurance type.

[0058] In this embodiment, a recognition engine is used to identify each invoice, obtaining its location and text information. Then, based on the location and text information, and by employing layout analysis and querying a financial database, structured information is extracted from the invoice. For example... Figure 3 As shown, the process of obtaining structured information includes:

[0059] First, layout analysis is used to obtain structured information for all invoices, including: determining the header based on the location information, obtaining the header information by extracting the text information corresponding to the header location, determining the detail location based on the location information, and obtaining the detail information based on the text information corresponding to the detail location.

[0060] In this embodiment, location information refers to the coordinates of the selected area on the ticket, while text information refers to the text within the selected area. When structured information is obtained through layout analysis, the header and detail positions are located based on the location information. Then, the text within the corresponding areas is aligned, and the header and detail information are extracted based on the semantic meaning of the text.

[0061] Then, it determines whether the invoice is an outpatient electronic medical invoice or an inpatient electronic medical invoice. If it is not an outpatient electronic medical invoice or an inpatient electronic medical invoice, it directly outputs the structured information obtained from the layout analysis. If it is an outpatient electronic medical invoice or an inpatient electronic medical invoice, it queries the Ministry of Finance database based on the five elements of the header information obtained from the layout analysis, including the invoice code, invoice number, invoice date, verification code, and total amount, to obtain the precise structured information of the outpatient electronic medical invoice and the inpatient electronic medical invoice.

[0062] Since electronic medical invoices have a corresponding Ministry of Finance database, the structured information stored in the Ministry of Finance database is more accurate. Therefore, after obtaining the header information of electronic medical invoices through layout analysis, the Ministry of Finance database is queried based on the header information to obtain more accurate structured information.

[0063] Step 3: Group the invoices in the treatment process into multiple groups according to the invoice type and structured information, and finally obtain the outpatient final group and the inpatient final group.

[0064] In the embodiments, such as Figure 4 As shown, the receipts during the treatment process are grouped into multiple phases according to receipt type and structured information, including:

[0065] (1) According to the type of invoice, the invoices in medical cases are divided into 8 initial groups, namely: outpatient medical invoice initial group, inpatient medical invoice initial group, outpatient expense list initial group, inpatient expense list initial group, outpatient medical record initial group, inpatient medical record initial group, outpatient settlement statement initial group and inpatient settlement statement initial group.

[0066] (2) For the initial group of non-empty outpatient related documents, the documents are grouped according to the date of visit or the invoice number in the header information according to certain rules to obtain the second-stage outpatient group, including the second-stage outpatient case group, the second-stage outpatient settlement statement group, the second-stage outpatient expense list group, and the second-stage outpatient medical invoice group; for the initial group of non-empty inpatient related documents, the documents are grouped according to the admission date or the invoice number in the header information according to certain rules to obtain the second-stage inpatient group, including the second-stage inpatient case group, the second-stage inpatient settlement statement group, the second-stage inpatient expense list group, and the second-stage inpatient medical invoice group.

[0067] Specifically, when performing secondary grouping of outpatient-related and inpatient-related initial groups according to certain rules, the grouping is specifically within the invoice group. That is, for each initial group, invoices with the same visit date are grouped together, invoices with the same hospitalization date are grouped together, and invoices with the same invoice number are grouped together. It should be noted that each type of invoice is grouped within the group based on its own header information. For example, if the header information of an outpatient medical record includes the visit date, then the grouping is based on the visit date, that is, outpatient medical records with the same visit date are grouped together. Similarly, if the header information of an inpatient medical invoice includes the invoice number, then the grouping is based on the invoice number, that is, inpatient medical invoices with the same invoice number are grouped together.

[0068] (3) For each second-phase group, calculate the similarity sim1 between any two tickets based on the structured information. If the similarity sim1 is greater than the given threshold thresh1, output the similarity sim2 of the text information of the two tickets. If the similarity sim2 is greater than the given threshold thresh2, determine that the two tickets are duplicate tickets and add them to the duplicate ticket set. Keep only the structured information of one ticket in the duplicate ticket set and clear the structured information of the other tickets to obtain the second-phase group after clearing the duplicate tickets.

[0069] In this embodiment, duplicate receipts are filtered out for all outpatient phase II groups and inpatient phase II groups obtained in step (2). Specifically, for each phase II group, similarity calculation is performed between receipts within the group. The similarity sim1 is calculated based on the header information and detail information. When the similarity sim1 is greater than a given threshold thresh1, the similarity sim2 is calculated based on the text information and relative position information. When the similarity sim2 is greater than the given threshold thresh2, the two receipts are considered to be duplicates.

[0070] (4) For the second-phase group after clearing duplicate invoices, the second-phase group of outpatient medical invoices and the second-phase group of outpatient expense list are matched and grouped according to the visit date or invoice number in the header information to obtain the third-phase group of outpatient medical invoices. The second-phase group of inpatient medical invoices and the second-phase group of inpatient expense list are matched and grouped according to the admission date or invoice number in the header information to obtain the third-phase group of inpatient medical invoices.

[0071] In this embodiment, some medical invoices contain amount information for broad categories of items. The specific details and corresponding amounts for each category are presented in the expense list, which also contains invoice numbers. Therefore, matching and grouping of medical invoices and expense lists is necessary. Specifically, for the second-phase group of outpatient medical invoices and the second-phase group of outpatient expense lists, groups of second-phase outpatient medical invoices and the second-phase outpatient expense lists belonging to the same visit date or invoice number are matched to form a third-phase outpatient group. Similarly, for the second-phase group of inpatient medical invoices and the second-phase inpatient expense lists, groups of second-phase inpatient medical invoices and the second-phase inpatient expense lists belonging to the same admission date or invoice number are matched to form a third-phase inpatient group.

[0072] (5) For outpatient expense list receipts that are grouped separately due to missing invoice numbers or missing consultation dates in step (4), if the difference between the total amount of a single outpatient medical invoice in the outpatient three-phase group and the detailed amount of the outpatient expense list is equal to the detailed amount of the separately grouped outpatient expense list receipts, then the separately grouped outpatient expense list receipts will be merged into the corresponding outpatient three-phase group; For inpatient expense list receipts that are grouped separately due to missing invoice numbers or missing hospitalization dates in step (4), if the difference between the total amount of a single inpatient medical invoice in the inpatient three-phase group and the detailed amount of the inpatient expense list is equal to the detailed amount of the separately grouped inpatient expense list receipts, then the separately grouped inpatient expense list receipts will be merged into the corresponding inpatient three-phase group.

[0073] In the embodiment, after the inter-group matching in step (4), some expense lists (outpatient expense lists and inpatient expense lists) will be grouped separately due to missing invoice numbers or missing consultation dates. They are not matched with medical invoices to form a three-phase group. Therefore, for these separately grouped expense lists, it is necessary to match them into the three-phase group. Specifically, the matching is performed based on the fact that the total amount of the invoice is equal to the total amount of the details as a tuple. The specific implementation process is as follows: For a separately grouped outpatient expense list invoice P, if the difference between the total amount of a single outpatient medical invoice in outpatient phase three group A and the detailed amount of at least one outpatient expense list in outpatient phase three group A is equal to the detailed amount of the separately grouped outpatient expense list invoice P, then the separately grouped outpatient expense list invoice P will be merged into the corresponding outpatient phase three group A; For a separately grouped inpatient expense list invoice Q, if the difference between the total amount of a single inpatient medical invoice in inpatient phase three group B and the detailed amount of at least one inpatient expense list in inpatient phase three group B is equal to the detailed amount of the separately grouped inpatient expense list invoice Q, then the separately grouped inpatient expense list invoice Q will be merged into the corresponding inpatient phase three group B.

[0074] (6) For each three-phase group obtained in steps (4) and (5), if the sum of the detailed amounts of all expense lists in the three-phase group is equal to the sum of the total amount of all medical invoices, then clear the detailed information in the medical invoices in the three-phase group and clear the header information in the expense list; if the medical invoices in the three-phase group are electronic medical invoices, and the expense list corresponding to the electronic medical invoices is obtained by querying the Ministry of Finance database, then according to the text information and location information, match and overwrite the detailed information obtained by layout analysis to obtain the four-phase group.

[0075] In this embodiment, the three-phase group includes an outpatient three-phase group and an inpatient three-phase group; the expense list includes an outpatient expense list and an inpatient expense list; and the medical invoice includes an outpatient medical invoice and an inpatient medical invoice. The four-phase group includes an outpatient four-phase group and an inpatient four-phase group. After obtaining each three-phase group, the four-phase group is constructed by removing semantically redundant information based on verifying that the detailed amounts of the expense list and the total amount of the medical invoices within the group are equal.

[0076] Taking the outpatient phase three group as an example, if the sum of the detailed amounts of all outpatient expense lists in the outpatient phase three group equals the sum of the total amounts of all outpatient medical invoices, then the detailed information in the outpatient medical invoices in the outpatient phase three group is cleared, and the header information in the outpatient expense list is also cleared. Based on this, to obtain more accurate detailed information, if the outpatient medical invoices in the outpatient phase three group are electronic outpatient medical invoices, and the corresponding outpatient expense list is obtained by querying the Ministry of Finance's database, then based on text and location information, the detailed information of the outpatient expense list returned by the Ministry of Finance's database is matched and overwritten. The detailed information obtained through layout analysis is then used to obtain the outpatient phase four group.

[0077] (7) Match and group the outpatient settlement statement group II and outpatient settlement statement group IV according to the date of visit or invoice number to obtain the outpatient settlement statement group V; match and group the inpatient settlement statement group II and inpatient settlement statement group IV according to the date of admission or invoice number to obtain the inpatient settlement statement group V.

[0078] In this embodiment, it is necessary to perform inter-group matching between the outpatient settlement statement Phase II group and the outpatient settlement statement Phase IV group. Specifically, the outpatient settlement statement Phase II group and the outpatient settlement statement Phase IV group with the same visit date or invoice number are matched to form the outpatient settlement statement Phase V group, and the inpatient settlement statement Phase II group and the inpatient settlement statement Phase IV group with the same hospitalization date or invoice number are matched to form the inpatient settlement statement Phase V group.

[0079] (8) Match and group the outpatient medical record phase II group and outpatient phase V group according to the date of visit or invoice number to obtain the final outpatient group; match and group the inpatient medical record phase II group and inpatient phase V group according to the admission date or invoice number to obtain the final inpatient group.

[0080] In this embodiment, it is necessary to perform inter-group matching between the outpatient medical record phase II group and the outpatient phase V group. The outpatient medical record phase II group and the outpatient phase V group with the same visit date or the same invoice number are matched to form the final outpatient group. The inpatient medical record phase II group and the inpatient phase V group with the same hospitalization date or the same invoice number are matched to form the final inpatient group.

[0081] Based on the same inventive concept, the embodiment also provides a treatment process grouping device for intelligent insurance claims, including a classification module, a structural information extraction module, and a multi-stage grouping module.

[0082] The classification module is used to classify various receipts in the treatment process in three stages and exclude irrelevant receipts; the structure information extraction module is used to identify the remaining receipts and extract structured information, which includes header information and detailed information; the multi-stage grouping module is used to group the receipts in the treatment process into multiple stages according to receipt category and structured information, and finally obtain the outpatient final grouping and the inpatient final grouping.

[0083] It should be noted that the treatment process grouping device provided in the above embodiments should be illustrated by the division of the above functional modules when grouping treatment processes. The functions can be assigned to different functional modules as needed, that is, the internal structure of the terminal or server can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the treatment process grouping device and the treatment process grouping method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the treatment process grouping method embodiments, which will not be repeated here.

[0084] The specific embodiments described above illustrate the technical solution and beneficial effects of the present invention in detail. It should be understood that the above description is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for grouping treatment processes for intelligent insurance claims, characterized in that, Includes the following steps: The various receipts during the treatment process are classified in three stages, excluding irrelevant receipts. The three stages are as follows: Stage 1: Using image and text classification methods, various receipts during the treatment process are divided into medical invoices, expense lists, reimbursement settlement forms, medical records, and irrelevant receipts. Stage 2: Using keyword matching and rule-based methods, medical invoices are divided into outpatient medical invoices and inpatient medical invoices; expense lists are divided into outpatient expense lists and inpatient expense lists; reimbursement settlement forms are divided into outpatient reimbursement settlement forms and inpatient reimbursement settlement forms; and medical records are divided into outpatient medical records and inpatient medical records. Stage 3: Using keyword matching and rule-based methods, outpatient medical invoices are divided into outpatient electronic medical invoices and outpatient regular medical invoices; and inpatient medical invoices are divided into inpatient electronic medical invoices and inpatient regular medical invoices. The remaining invoices are identified and their structured information is extracted, including header information and detailed information. The invoices during the treatment process are grouped in multiple phases according to invoice type and structured information, ultimately resulting in outpatient final grouping and inpatient final grouping, including: (1) According to the type of invoice, the invoices in medical cases are divided into 8 initial groups, namely: outpatient medical invoice initial group, inpatient medical invoice initial group, outpatient expense list initial group, inpatient expense list initial group, outpatient medical record initial group, inpatient medical record initial group, outpatient settlement statement initial group and inpatient settlement statement initial group; (2) For the initial group of non-empty outpatient related documents, the documents are grouped according to the date of visit or the invoice number in the header information according to certain rules to obtain the second outpatient group, including the second outpatient case group, the second outpatient settlement statement group, the second outpatient expense list group, and the second outpatient medical invoice group; For the initial group of non-empty inpatient related documents, the documents are grouped according to the admission date or the invoice number in the header information according to certain rules to obtain the second inpatient group, including the second inpatient case group, the second inpatient settlement statement group, the second inpatient expense list group, and the second inpatient medical invoice group. (3) For each second-phase group, calculate the similarity sim1 between any two tickets based on the structured information. If the similarity sim1 is greater than the given threshold thresh1, output the similarity sim2 of the text information and relative position information of the two tickets. If the similarity sim2 is greater than the given threshold thresh2, determine that the two tickets are duplicate tickets and add them to the duplicate ticket set. Keep only the structured information of one ticket in the duplicate ticket set and clear the structured information of the other tickets to obtain the second-phase group after clearing the duplicate tickets. (4) For the second group after clearing duplicate invoices, the second group of outpatient medical invoices and the second group of outpatient expense list are matched and grouped according to the date of visit or invoice number in the header information to obtain the third group of outpatient medical invoices. The second group of inpatient medical invoices and the second group of inpatient expense list are matched and grouped according to the date of admission or invoice number in the header information to obtain the third group of inpatient medical invoices. (5) For outpatient expense list receipts that are grouped separately due to missing invoice numbers or missing consultation dates in step (4), if the difference between the total amount of a single outpatient medical invoice in the outpatient three-phase group and the detailed amount of at least one outpatient expense list in the group is equal to the detailed amount of the outpatient expense list receipts that are grouped separately, then the outpatient expense list receipts that are grouped separately will be merged into the corresponding outpatient three-phase group; For inpatient expense list receipts that are grouped separately due to missing invoice numbers or missing hospitalization dates in step (4), if the difference between the total amount of a single inpatient medical invoice in the inpatient three-phase group and the detailed amount of at least one inpatient expense list in the group is equal to the detailed amount of the inpatient expense list receipts that are grouped separately, then the inpatient expense list receipts that are grouped separately will be merged into the corresponding inpatient three-phase group; (6) For each three-phase group obtained in steps (4) and (5), if the sum of the detailed amounts of all expense lists in the three-phase group is equal to the sum of the total amount of all medical invoices, then clear the detailed information in the medical invoices in the three-phase group and clear the header information in the expense list; if the medical invoices in the three-phase group are electronic medical invoices, and the expense list corresponding to the electronic medical invoices is obtained by querying the Ministry of Finance database, then according to the text information and location information, match and overwrite the detailed information obtained by layout analysis to obtain the four-phase group, where the three-phase group includes the outpatient three-phase group and the inpatient three-phase group, the expense list includes the outpatient expense list and the inpatient expense list, the medical invoices include the outpatient medical invoices and the inpatient medical invoices, and the four-phase group includes the outpatient four-phase group and the inpatient four-phase group; (7) Match and group the outpatient settlement statement second-phase group and outpatient fourth-phase group according to the date of visit or invoice number to obtain the outpatient fifth-phase group; match and group the inpatient settlement statement second-phase group and inpatient fourth-phase group according to the admission date or invoice number to obtain the inpatient fifth-phase group; (8) Match and group the outpatient medical record phase II group and outpatient phase V group according to the date of visit or invoice number to obtain the final outpatient group; match and group the inpatient medical record phase II group and inpatient phase V group according to the admission date or invoice number to obtain the final inpatient group.

2. The method for grouping treatment procedures for insurance intelligent claims settlement according to claim 1, characterized in that, The extraction of structured information from the remaining invoices includes: For each invoice, a recognition engine is used to identify the invoice, obtain the invoice's location information and text information, and then extract the invoice's structured information based on the location information and text information, using layout analysis and querying the financial database.

3. The method for grouping treatment procedures for insurance intelligent claims settlement according to claim 2, characterized in that, The extraction of structured information from invoices based on location and text information, using layout analysis and querying of financial databases, includes the following steps: First, layout analysis is used to obtain structured information for all invoices, including: determining the header based on the location information, obtaining the header information by extracting the text information corresponding to the header location, determining the detail location based on the location information, and obtaining the detail information based on the text information corresponding to the detail location. Then, it determines whether the invoice is an outpatient electronic medical invoice or an inpatient electronic medical invoice. If it is not an outpatient electronic medical invoice or an inpatient electronic medical invoice, it directly outputs the structured information obtained from the layout analysis. If it is an outpatient electronic medical invoice or an inpatient electronic medical invoice, it queries the Ministry of Finance database based on the five elements of the header information obtained from the layout analysis, including the invoice code, invoice number, invoice date, verification code, and total amount, to obtain the precise structured information of the outpatient electronic medical invoice and the inpatient electronic medical invoice.

4. The treatment process grouping method for intelligent insurance claims according to claim 1 or 3, characterized in that, The header information is data in key-value pair format, including visit date, hospitalization date, invoice code, invoice number, invoice date, verification code, fund payment, out-of-pocket, self-paid, name, and hospital name.

5. The method for grouping treatment procedures for insurance intelligent claims settlement according to claim 1, characterized in that, The detailed information is in list format, including item name, amount, medical insurance type, ratio, unit price, and quantity.

6. The method for grouping treatment procedures for insurance intelligent claims settlement according to claim 4, characterized in that, The detailed information is in list format, including item name, amount, medical insurance type, ratio, unit price, and quantity.

7. The method for grouping treatment procedures for insurance intelligent claims settlement according to claim 1, characterized in that, In step (2), when the initial group related to outpatient visits and the initial group related to hospitalization are grouped in two phases according to certain rules, tickets with the same visit date are grouped together, tickets with the same hospitalization date are grouped together, and tickets with the same invoice number are grouped together.

8. The method for grouping treatment procedures for insurance intelligent claims settlement according to claim 1, characterized in that, In step (7), the outpatient settlement statement group II and outpatient settlement statement group IV with the same visit date or invoice number are matched to form outpatient settlement statement group V, and the inpatient settlement statement group II and inpatient settlement statement group IV with the same hospitalization date or invoice number are matched to form inpatient settlement statement group V. In step (8), the outpatient medical record group of phase II and outpatient medical record group with the same visit date or invoice number are matched to form the outpatient final group, and the inpatient medical record group of phase II and inpatient medical record group with the same hospitalization date or invoice number are matched to form the inpatient final group.

9. A therapeutic procedure grouping apparatus for intelligent claims settlement of insurance, characterized by, It includes a classification module, a structural information extraction module, and a multi-period grouping module. The classification module is used to classify various receipts during the treatment process in three stages and exclude irrelevant receipts. The three-stage classification includes: Stage 1: Using image-based and text-based classification methods, various receipts during the treatment process are divided into medical invoices, expense lists, reimbursement settlement forms, medical records, and irrelevant receipts; Stage 2: Using keyword matching and rule-based methods, medical invoices are divided into outpatient medical invoices and inpatient medical invoices; expense lists are divided into outpatient expense lists and inpatient expense lists; reimbursement settlement forms are divided into outpatient reimbursement settlement forms and inpatient reimbursement settlement forms; and medical records are divided into outpatient medical records and inpatient medical records; Stage 3: Using keyword matching and rule-based methods, outpatient medical invoices are divided into outpatient electronic medical invoices and outpatient regular medical invoices; and inpatient medical invoices are divided into inpatient electronic medical invoices and inpatient regular medical invoices. The structure information extraction module is used to identify the remaining invoices and extract structured information, which includes header information and detail information. The multi-phase grouping module is used to group invoices during the treatment process into multiple phases according to invoice type and structured information, ultimately obtaining outpatient final grouping and inpatient final grouping, including: (1) According to the type of invoice, the invoices in medical cases are divided into 8 initial groups, namely: outpatient medical invoice initial group, inpatient medical invoice initial group, outpatient expense list initial group, inpatient expense list initial group, outpatient medical record initial group, inpatient medical record initial group, outpatient settlement statement initial group and inpatient settlement statement initial group; (2) For the initial group of non-empty outpatient related documents, the documents are grouped according to the date of visit or the invoice number in the header information according to certain rules to obtain the second outpatient group, including the second outpatient case group, the second outpatient settlement statement group, the second outpatient expense list group, and the second outpatient medical invoice group; For the initial group of non-empty inpatient related documents, the documents are grouped according to the admission date or the invoice number in the header information according to certain rules to obtain the second inpatient group, including the second inpatient case group, the second inpatient settlement statement group, the second inpatient expense list group, and the second inpatient medical invoice group. (3) For each second-phase group, calculate the similarity sim1 between any two tickets based on the structured information. If the similarity sim1 is greater than the given threshold thresh1, output the similarity sim2 of the text information and relative position information of the two tickets. If the similarity sim2 is greater than the given threshold thresh2, determine that the two tickets are duplicate tickets and add them to the duplicate ticket set. Keep only the structured information of one ticket in the duplicate ticket set and clear the structured information of the other tickets to obtain the second-phase group after clearing the duplicate tickets. (4) For the second group after clearing duplicate invoices, the second group of outpatient medical invoices and the second group of outpatient expense list are matched and grouped according to the date of visit or invoice number in the header information to obtain the third group of outpatient medical invoices. The second group of inpatient medical invoices and the second group of inpatient expense list are matched and grouped according to the date of admission or invoice number in the header information to obtain the third group of inpatient medical invoices. (5) For outpatient expense list receipts that are grouped separately due to missing invoice numbers or missing consultation dates in step (4), if the difference between the total amount of a single outpatient medical invoice in the outpatient three-phase group and the detailed amount of at least one outpatient expense list in the group is equal to the detailed amount of the outpatient expense list receipts that are grouped separately, then the outpatient expense list receipts that are grouped separately will be merged into the corresponding outpatient three-phase group; For inpatient expense list receipts that are grouped separately due to missing invoice numbers or missing hospitalization dates in step (4), if the difference between the total amount of a single inpatient medical invoice in the inpatient three-phase group and the detailed amount of at least one inpatient expense list in the group is equal to the detailed amount of the inpatient expense list receipts that are grouped separately, then the inpatient expense list receipts that are grouped separately will be merged into the corresponding inpatient three-phase group; (6) For each three-phase group obtained in steps (4) and (5), if the sum of the detailed amounts of all expense lists in the three-phase group is equal to the sum of the total amount of all medical invoices, then clear the detailed information in the medical invoices in the three-phase group and clear the header information in the expense list; if the medical invoices in the three-phase group are electronic medical invoices, and the expense list corresponding to the electronic medical invoices is obtained by querying the Ministry of Finance database, then according to the text information and location information, match and overwrite the detailed information obtained by layout analysis to obtain the four-phase group, where the three-phase group includes the outpatient three-phase group and the inpatient three-phase group, the expense list includes the outpatient expense list and the inpatient expense list, the medical invoices include the outpatient medical invoices and the inpatient medical invoices, and the four-phase group includes the outpatient four-phase group and the inpatient four-phase group; (7) Match and group the outpatient settlement statement second-phase group and outpatient fourth-phase group according to the date of visit or invoice number to obtain the outpatient fifth-phase group; match and group the inpatient settlement statement second-phase group and inpatient fourth-phase group according to the admission date or invoice number to obtain the inpatient fifth-phase group; (8) Match and group the outpatient medical record phase II group and outpatient phase V group according to the date of visit or invoice number to obtain the final outpatient group; match and group the inpatient medical record phase II group and inpatient phase V group according to the admission date or invoice number to obtain the final inpatient group.

Images