WO2020087970A1 - 一种基于神经网络的病种分值校验方法及计算设备 - Google Patents

一种基于神经网络的病种分值校验方法及计算设备 Download PDF

Info

Publication number
WO2020087970A1
WO2020087970A1 PCT/CN2019/095011 CN2019095011W WO2020087970A1 WO 2020087970 A1 WO2020087970 A1 WO 2020087970A1 CN 2019095011 W CN2019095011 W CN 2019095011W WO 2020087970 A1 WO2020087970 A1 WO 2020087970A1
Authority
WO
WIPO (PCT)
Prior art keywords
case
score
disease
disease type
classification
Prior art date
Application number
PCT/CN2019/095011
Other languages
English (en)
French (fr)
Inventor
刘俊芳
Original Assignee
平安医疗健康管理股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安医疗健康管理股份有限公司 filed Critical 平安医疗健康管理股份有限公司
Publication of WO2020087970A1 publication Critical patent/WO2020087970A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
    • G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms

Definitions

  • the present application relates to the technical field of medical management, and in particular to a method and computing device for verifying the score of a disease based on a neural network.
  • the payment system reform is a reflection of a major change in the system engineering, the concept of medical insurance management, and the role of medical insurance agencies.
  • the implementation of payment by disease type fully embodies the payment system is the key to comprehensive medical reform.
  • the so-called payment by disease refers to scientifically formulating a fixed payment standard for each disease through a unified disease diagnosis classification, and the social security institution pays the hospitalization fee to the designated medical institution according to the standard and the number of hospitalizations, so that the use of medical resources is standardized , That is, the resource consumption of medical institutions is directly proportional to the number of inpatients treated, the complexity of the disease and the intensity of services. In short, it is to clearly stipulate how much a certain disease should cost, so as to avoid the medical unit from abusing medical service items, repeating items and disassembling items, preventing serious illnesses in hospitals, and ensuring the quality of medical services.
  • Standardized medical information is very important for the application of medical information big data to the payment method based on the type of disease.
  • the standardization of medical information is the prerequisite for the application of medical big data.
  • the classification of disease types currently generally adopts the International Classification of Diseases (ICD).
  • ICD-10 divides the disease into 21 chapters, more than 26000 kinds of diseases according to the characteristics of etiology, location, pathology and clinical manifestations, and encodes each disease type.
  • ICD-10 International Classification of Diseases
  • ICD-10 divides the disease into 21 chapters, more than 26000 kinds of diseases according to the characteristics of etiology, location, pathology and clinical manifestations, and encodes each disease type.
  • the number of common diseases in various regions is far less than 26,000, and when medical staff record cases, due to the diversity and complexity of the diseases in the existing technology, medical staff often do not Ranked according to international standards, each region has a localized language description, which brings certain difficulties to the implementation of fee-based disease.
  • the embodiments of the present application provide a method for calculating a disease score of a neural network and a computing device, which can realize the verification of the case score of cases and find abnormal cases in time. .
  • an embodiment of the present application provides a method for verifying a disease score based on a neural network, including:
  • the computing device receives the case data of the first case, and the case data includes multiple case characteristics
  • the computing device inputs the characteristics of the multiple cases into a disease type score learning model to obtain the first disease type score of the first case;
  • the computing device looks up the first basic disease type score corresponding to the disease type classification of the first case in the disease type score dictionary, and calculates the first case's number based on the first basic disease type score Two disease type scores, the disease type score dictionary includes a one-to-one correspondence between disease type classification and basic disease type scores;
  • the computing device determines whether the first case is an abnormal case according to the first disease type score and the second disease type score, and if so, outputs prompt information for prompting the first case abnormality .
  • an embodiment of the present application further provides a computing device, including:
  • a receiving unit configured to receive case data of the first case, and the case data includes multiple case characteristics
  • a first calculation unit configured to input the characteristics of the multiple cases into a disease score learning model to obtain the first disease score of the first case
  • a searching unit used to search for the first basic disease type score corresponding to the disease type classification of the first case in the disease type score dictionary
  • a second calculation unit configured to calculate the second disease type score of the first case according to the first basic disease type score, and the disease type score dictionary includes the classification of the disease type and the basic disease type score One-to-one correspondence;
  • the judging unit is used to judge whether the first case is an abnormal case according to the score of the first disease type and the score of the second disease type,
  • the output unit is configured to, when the judgment result of the judgment unit is yes, output prompt information for prompting the abnormality of the first case.
  • an embodiment of the present application further provides a computing device, the computing device includes a processor, a memory, and a communication module, the processor is coupled to the memory, the communication module, and the processor is used to call The program code stored in the memory executes:
  • the case data including multiple case characteristics
  • the disease category score dictionary includes a one-to-one correspondence between disease category classification and basic disease category scores
  • the first case is an abnormal case according to the score of the first disease type and the score of the second disease type, and if so, output prompt information for prompting the abnormality of the first case.
  • an embodiment of the present application further provides a computer-readable storage medium, which is used for computer software instructions, which when executed by a computer causes the computer to execute as the first
  • a method for checking the score of a disease based on a neural network is used for checking the score of a disease based on a neural network.
  • an embodiment of the present application further provides a computer program, the computer program includes computer software instructions, which when executed by a computer causes the computer to execute any one of the first aspect Method for checking the score of disease based on neural network.
  • the computing device receives the case data of the first case, which includes multiple case characteristics
  • the disease score dictionary includes a one-to-one correspondence between the classification of the disease and the score of the basic disease; further, according to The first disease type score and the second disease type score determine whether the first case is an abnormal case, and if so, output prompt information for prompting the abnormality of the first case.
  • FIG. 1 is a functional architecture diagram of a medical insurance management platform provided by an embodiment of this application.
  • FIG. 2 is a flowchart of a method for checking the score of a disease according to an embodiment of the present application
  • FIG. 3 is a schematic structural diagram of a computing device according to an embodiment of the present application.
  • FIG. 4 is a schematic structural diagram of yet another computing device provided by an embodiment of this application.
  • FIG. 5 is a schematic structural diagram of yet another computing device provided by an embodiment of this application.
  • FIG. 1 is a functional architecture diagram of a medical insurance management platform provided by the embodiments of the present application.
  • the medical insurance management platform can run on computing devices, and provides a series of functions related to cases, medical insurance, and disease scores for the operators of the medical insurance management platform.
  • the medical insurance management platform includes but is not limited to some or all of the following functions achieve:
  • the medical insurance management platform can encode the disease type obtained by the main diagnosis of the case according to the case data of the input case, and the disease type coding method can use ICD-10 code (also called six-digit code code in this application), Other encoding methods may also be used, such as four-bit code encoding (ie, the first 4 bits of the six-bit code), three-bit code encoding (ie, the first three bits of the six-bit code), and so on.
  • a disease classification dictionary applicable to the region can be established through a disease coding method through a case set of a certain region, and the disease classification dictionary includes names of M disease categories and names of the M disease category classifications
  • M is a positive integer.
  • the computing device can identify the classification of the disease corresponding to the case based on the diagnosis name, disease code and other information filled in by medical personnel in the case data of the medical insurance management platform, and then the disease classification code corresponding to the classification of the disease Added to the case data, in order to further calculate the score of the disease type of the case, and then realize the functions of paying by disease type and detecting the authenticity of the case data based on the disease type score.
  • the medical insurance management platform can store a correspondence table of disease types and disease scores or include a program for calculating the scores of disease types, and can determine the insured person in the case by the name of the disease type and the code of the disease type (Ie the patient) the classification of the diseased type, and then the process of determining the disease type score based on the disease type classification according to the corresponding relationship between the disease type classification and the disease type score value or the disease type value calculation program, etc. value.
  • the disease score is the standard score of the region (for example, country, province or city, etc.) that is determined based on the big data of the case and used to calculate medical expenses (such as predicted medical insurance expenses, predicted total expenses, etc.).
  • a disease category score dictionary can be established.
  • the disease category score dictionary includes a one-to-one correspondence between the disease category and the basic disease category score, and then according to the actual situation of the case (such as the age of the insured person, the severity of the disease) Information such as degree, hospital, department, etc.) is adjusted on the basis of the score of the basic disease to obtain the score of the disease suitable for the case.
  • the score of disease type has a positive correlation with medical expenses, that is, the higher the score of disease type, the higher the medical cost of the disease type.
  • the medical insurance management platform can perform statistical analysis on the cases reported by various hospitals in the area according to the evaluation cycle (for example, monthly, quarterly, annual, etc.).
  • the above statistical analysis of cases can support statistical analysis by month, quarter, year, etc., and support the occurrence of the number of cases, total cost, actual cost of one or more combinations of different hospitals, different cost ranges, different diseases, etc.
  • Statistical analysis of medical insurance costs, forecasted medical insurance costs, etc. to adjust the disease scores for each disease type to be used in the next evaluation cycle based on the statistical analysis results. It should be understood that other functions can be implemented based on the statistical analysis results, such as adjusting the hospital level coefficient of the hospital based on statistics on the income and expenditure of each hospital, and this embodiment of the present application is not limited.
  • the medical insurance management platform can detect the authenticity of the case based on the case data in the case.
  • the case is marked and the case is output with false data or abnormal cases Prompt message, so that the operator of the medical insurance management platform can identify the abnormal case in time and analyze the cause of the abnormal case.
  • the medical insurance management platform can visualize the statistical analysis results obtained by the statistical analysis function of case data, and can also visualize the results of statistical analysis of problem cases, so as to facilitate the statistical analysis of the operators of the medical insurance management platform.
  • computing devices may include, but are not limited to, mobile phones, mobile computers, tablet computers, media players, computers, servers, and other devices that include data processing functions.
  • Computing devices running various functions of the medical insurance management platform can receive cases reported from institutions and individuals such as hospitals.
  • the medical insurance management platform provided by the present application may also include the implementation of other functions, for example, optimization of disease scores, etc., and this embodiment of the present application is not limited.
  • FIG. 2 is a method for verifying the score of a disease based on a neural network provided by this application.
  • the execution subject of the calculation of the disease score is described as an example of a computing device (a device that runs various functions of the case management platform). It can be understood that the calculation of the score of the disease can also be performed by other terminals or servers For devices with data processing functions, this embodiment of the present application is not limited.
  • the method may include, but is not limited to, some or all of the following steps:
  • S2 Receive case data of the first case, and the case data includes multiple case characteristics.
  • the case is the patient diagnosis and treatment process recorded by the hospital for the patient.
  • the case data may include but is not limited to one or a combination of one or more of the patient's personal information, diagnosis information, treatment information, cost information, actual medical insurance costs, and so on.
  • the diagnosis information may include a diagnosis identifier for identifying the classification of the insured person's disease type.
  • the diagnosis identifier may be a diagnosis name, such as a main diagnosis name; it may also be a diagnosis code, such as an ICD diagnosis code, etc .; or it may be a surgery ID, which may be a surgery name, a surgery code, etc.
  • personal information may include but is not limited to information such as the age, gender, medical history, etc.
  • the treatment information is the process information of the insured person's treatment recorded in the case.
  • the fee information includes, but is not limited to, one or more combinations of the surgical fee, hospitalization fee, testing fee, registration fee, drug fee, and total cost incurred by the insured during the treatment of the disease.
  • the computing device may identify and extract the case characteristics of the case data according to the case data, and the case characteristics may include, but are not limited to, diagnostic identification, drug identification, drug dosage, drug cost, test item identification, test item cost, and surgical identification , One or more combinations of surgery cost, hospitalization days, hospitalization cost, complication identification, secondary symptom identification, participant's age, participant's gender, etc.
  • S4 Input the characteristics of the multiple cases into a disease score learning model to obtain the first disease score of the first case.
  • the disease type score learning model in this application is a neural network model, which is used to identify the disease type score corresponding to the case data according to multiple case characteristics of the case.
  • the data input to the disease score learning model may include, but is not limited to, one or more combinations of case characteristics listed above.
  • the disease score learning model needs to train the disease score learning model through sample data before predicting the disease score, in order to learn to identify the disease score corresponding to the case based on multiple case characteristics of the case .
  • the sample data may be data obtained before calculating the medical insurance cost without implementing the score of the disease type to ensure the accuracy of the sample data. It is understandable that the case data obtained before the calculation of the medical insurance cost without implementing the score of the disease type does not suspect that the medical staff makes false information such as the replacement of the main diagnosis and the subdiagnosis to increase the score of the disease type, and increases the hospitalization items. The data has good reliability.
  • the actual disease type score of the case in the sample data can be calculated by the actual medical insurance cost of the case.
  • the second case is any case in the sample data. The second case in this application is used as an example to explain the calculation of the true disease score corresponding to each case in the sample data.
  • the true disease score of the second case may be the ratio of the medical insurance cost of the second case to a preset constant, where the preset constant is configurable and set by a benchmarkable institution, and the preset constant may be 50, 100 or other numerical values are not limited in the embodiments of the present application. You can supervise and predict the disease score and real disease score to train the disease score learning model.
  • the disease species score dictionary includes a one-to-one correspondence between disease species classification and basic disease species scores.
  • the case data of the first case includes a diagnostic mark for identifying the classification of the disease.
  • the first case has recognized the disease classification of the first case through the disease classification recognition method and added the recognized disease classification (eg, the identification of the disease classification) to the first case In the case data.
  • the diagnosis identification is the identification of the disease classification (the classification name or classification code). It can be understood that the identification of the disease classification is the name or classification code of the disease classification included in the disease classification dictionary, where the disease classification dictionary includes a one-to-one correspondence between the name of the disease classification and the disease classification code, M is a positive integer.
  • the disease classification code is an ICD code
  • the disease classification dictionary is an ICD dictionary
  • the disease type classification code is the first N digits of the ICD code, and N is a positive integer less than 6.
  • the disease classification code in the disease classification dictionary is the first N digits of the ICD code, choose the first four digits, the first three digits or the first two digits of the ICD code of the disease, depending on the case classification of the cases in the concentrated case
  • the number of the first four digits of the ICD code such as greater than 10 cases, choose "four-digit code” as the disease classification code; if less than 10 cases, choose "three-digit code” as the disease classification code.
  • the diagnosis identifier in the first case is a diagnosis name or a diagnosis code
  • the disease category to which the first case belongs cannot be directly obtained through the diagnosis name or the diagnosis code.
  • the computing device may determine the disease type classification of the first case according to the diagnostic identifier for identifying the disease type classification of the insured person in the first case.
  • the computing device may determine that the classification of the disease type to which the first case belongs based on the diagnosis identifier of the first case may be: the diagnosis ID of the first disease type may include the diagnosis name, and the calculation device may pre-store the disease name comparison table
  • the disease name comparison table includes M disease type classifications and one or more diagnostic names corresponding to each of the M disease type classifications. Furthermore, the computing device may determine the disease category corresponding to the diagnosis name of the first disease according to the disease category name comparison table, and further determine the disease category code corresponding to the disease category of the second case according to the disease category dictionary.
  • the computing device may determine that the classification of the disease category to which the first case belongs according to the diagnosis identifier of the first case may be: the diagnosis identifier of the first disease type may include a diagnosis code, and the diagnosis code may be ICD-10 Code, ICD-9-CM3 surgical code, tumor morphology code (also called M code) or Chinese medicine disease code, etc.
  • the diagnosis code may be ICD-10 Code, ICD-9-CM3 surgical code, tumor morphology code (also called M code) or Chinese medicine disease code, etc.
  • the computing device can directly look up the disease classification code in the disease classification dictionary that matches the diagnostic code in the first case.
  • the M code can be converted into an ICD code or a four-digit ICD code.
  • the MCD “M8140 / 6" corresponds to the ICD code "C78.7”
  • the M code "M8140 / 3" corresponds to the ICD code "C34.9”.
  • the computing device can convert the M code in the first case to an ICD code according to the M code conversion table (the ICD code can be a disease classification code in a disease classification dictionary, or an ICD code with a six-digit code, etc.), Then look up the disease classification code in the disease classification dictionary that matches the converted ICD code.
  • an implementation manner in which the computing device calculates the second disease type score of the first case according to the first basic disease type score may be:
  • Y is the score of the second disease of the first case
  • a 1 is the score of the first basic disease
  • C 1 is the hospital level coefficient of the hospital where the first case is located
  • E i is the additional disease Species score
  • i is the index of the additional disease species score
  • i is a positive integer.
  • the second disease score needs to be appended with the first
  • the additional score of the disease type corresponding to the item satisfied by the case It can be understood that for the first case, the additional scores of the disease types corresponding to different items may be different. For the same item, the additional score of the disease type corresponding to cases of different disease types may be different.
  • S8 Determine whether the first case is an abnormal case according to the score of the first disease type and the score of the second disease type, and if so, output prompt information for prompting the abnormality of the first case.
  • the disease type score corresponding to the disease type classification can be found from the disease type dictionary by identifying the disease type classification of the case.
  • the case data of the case there may be false data in the case data of the case, so that the case score found by the case from the disease score dictionary is far from the actual situation of the case.
  • the diagnosis information, treatment information, cost information, and medical insurance costs in the case do not meet the diagnosis and treatment methods of the disease category.
  • the treatment methods used include surgery and medicines include Drugs used to treat tumors.
  • the treatment method used for the case of "diabetes" in the case is only anti-inflammatory drugs, then the case data may be false.
  • an implementation manner in which the computing device judges whether the first case is an abnormal case may be: A threshold, if yes, the first case is an abnormal case.
  • the first threshold may be 10, 20, 35 or other numerical values.
  • another implementation manner of the computing device judging whether the first case is an abnormal case may be: the computing device judging whether the ratio of the second disease type score to the first disease type score is greater than the first
  • the second threshold is less than the third threshold, and if it is, the first case is an abnormal case.
  • the value range of the third threshold may be 1.1-3, for example, 1.5, 2, 2.4 or other values, which are not limited in the embodiments of the present application.
  • the second threshold may be 0.1-0.9, such as 0.4, 0.5, 0.7 or other values, which are not limited in the embodiments of the present application.
  • a prompt message for indicating that the disease type score of the first case is too high is output; when the second disease type When the ratio of the score to the score of the first disease is less than the first threshold and greater than the second threshold, a prompt message is displayed to indicate that the score of the first case is within the normal range; when the score of the second disease is equal to When the proportion of the first disease type score is less than the second threshold, a prompt message for indicating that the disease type score of the first case is too low is output.
  • the method further includes:
  • the case set includes multiple cases, and the first case may be any one of the multiple cases;
  • the first image is displayed according to the case data of the abnormal case in the case set, and the first image includes at least one of the following: the correspondence between the hospital where the abnormal case in the case set is located and the number of cases, and the abnormal case in the case set The corresponding relationship between the number of attending physicians and the number of cases, the classification of the disease types of the abnormal cases in the case set and the number of cases.
  • the abnormal cases are analyzed by visualizing the case data of the abnormal cases, so that the management personnel of the medical insurance management platform can quickly find the problem of the abnormal case according to the visualized first image.
  • the computing device in the embodiment of the present application receives the case data of the first case, and the case data includes multiple case characteristics; the multiple case characteristics are input to the disease score learning model to obtain the first case of the first case Disease score; and, in the disease score dictionary, look up the first basic disease score corresponding to the disease classification of the first case, and calculate the second disease type of the first case based on the first basic disease score
  • the score determines whether the first case is an abnormal case based on the score of the first disease type and the score of the second disease type, and if so, outputs prompt information for prompting the abnormality of the first case.
  • a receiving unit 31 a first computing unit 32, a searching unit 33, a second computing unit 34, a judging unit 35, an output unit 36, and the like. among them,
  • the receiving unit 31 is configured to receive case data of the first case, and the case data includes multiple case characteristics;
  • the first calculation unit 32 is used to input the characteristics of the multiple cases into the disease type score learning model to obtain the first disease type score of the first case;
  • the searching unit 33 is configured to search for the first basic disease category score corresponding to the disease category to which the first case belongs in the disease category score dictionary;
  • the second calculation unit 34 is configured to calculate the second disease type score of the first case according to the first basic disease type score, and the disease type score dictionary includes a disease type classification and a basic disease type score One-to-one correspondence;
  • the judging unit 35 is configured to judge whether the first case is an abnormal case according to the score of the first disease type and the score of the second disease type,
  • the output unit 36 is configured to, when the judgment result of the judgment unit is yes, output prompt information for prompting the abnormality of the first case.
  • the individual case characteristics include diagnostic identification, drug identification, drug dosage, drug cost, test item identification, test item cost, surgical identification, surgical cost, hospitalization days, hospitalization cost, complication identification, secondary disease identification, A combination of one or more of the age of the insured person and the gender of the insured person.
  • the computing device 40 further includes:
  • the classification and identification unit 37 is configured to determine the classification of the disease type of the first case according to the diagnostic identifier used to identify the disease type classification of the insured in the first case, wherein the classification of the disease type is the classification of the disease type
  • the disease classification dictionary includes the names of the M disease classifications and the M disease classification codes corresponding to the names of the M disease classifications one by one
  • the disease classification codes are ICD codes Or the first N digits of the ICD code, where N is a positive integer less than 6, and M is a positive integer.
  • the second calculating unit 34 specifically calculates the second disease type score of the first case by the following formula:
  • Y is the score of the second disease of the first case
  • a 1 is the score of the first basic disease
  • C 1 is the hospital level coefficient of the hospital where the first case is located
  • E i is the additional disease Species score
  • i is the index of the additional disease species score
  • i is a positive integer.
  • the judgment unit 35 is specifically configured to:
  • the judging unit 35 is specifically configured to judge whether the ratio of the second disease type score to the first disease type score is greater than the second threshold or less than the third threshold;
  • the output unit 36 is specifically used for:
  • a prompt message for prompting that the score of the disease type of the first case is too high is output;
  • a prompt message for prompting that the score of the disease type of the first case is too low is output.
  • the receiving unit 31 is further configured to receive a case set, the case set includes multiple cases, and the first case is any one of the multiple cases;
  • the output unit 36 is further configured to display a first image according to the case data of the abnormal case in the case set, the first image including at least one of the following: the correspondence of the hospital where the abnormal case in the case set is located and the number of cases The relationship, the correspondence relationship between the attending physicians of the abnormal cases in the case set and the number of cases, and the correspondence relationship between the classification of the disease types of the abnormal cases in the case set and the number of cases.
  • the computing device 500 may include: a baseband chip 510, a memory 515 (one or more computer-readable storage media), a communication module 516 (eg, a radio frequency (RF) module 5161, and / or communication Interface 5162), peripheral system 517. These components can communicate on one or more communication buses 514.
  • a baseband chip 510 one or more computer-readable storage media
  • a communication module 516 eg, a radio frequency (RF) module 5161, and / or communication Interface 5162
  • peripheral system 517 can communicate on one or more communication buses 514.
  • the peripheral system 517 is mainly used to realize the interactive function between the computing device 510 and the user / external environment, and mainly includes input / output devices of the computing device 500.
  • the peripheral system 517 may include: a touch screen controller 518, a camera controller 519, an audio controller 520, and a sensor management module 521. Wherein, each controller may be coupled with their corresponding peripheral devices (such as touch screen 523, camera 524, audio circuit 525, and sensor 526). It should be noted that the peripheral system 517 may also include other I / O peripherals.
  • the baseband chip 510 may include one or more processors 511, a clock module 522, and a power management module 513.
  • the clock module 522 integrated in the baseband chip 510 is mainly used to generate a clock required for data transmission and timing control for the processor 511.
  • the power management module 513 integrated in the baseband chip 510 is mainly used to provide a stable, high-precision voltage for the processor 511, the radio frequency module 5161, and peripheral systems.
  • the radio frequency (RF) module 5161 is used to receive and transmit radio frequency signals, and mainly integrates the receiver and transmitter of the computing device 500.
  • the radio frequency (RF) module 5161 communicates with the communication network and other communication devices through radio frequency signals.
  • the radio frequency (RF) module 5161 may include, but is not limited to: an antenna system, an RF transceiver, one or more amplifiers, a tuner, one or more oscillators, a digital signal processor, a CODEC chip, a SIM card, and Storage media, etc.
  • the radio frequency (RF) module 5161 may be implemented on a separate chip.
  • the communication module 516 is used for data exchange between the computing device 500 and other devices.
  • the memory 515 is coupled to the processor 511 and is used to store various software programs and / or multiple sets of instructions.
  • the memory 515 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more disk storage devices, flash memory devices, or other non-volatile solid-state storage devices.
  • the memory 515 may store an operating system (hereinafter referred to as a system), such as an embedded operating system such as ANDROID, IOS, WINDOWS, or LINUX.
  • the memory 515 may also store a network communication program, which may be used to communicate with one or more additional devices, one or more computing device devices, or one or more network devices.
  • the memory 515 can also store a user interface program, which can display the content of the application program vividly through a graphical operation interface, and receive user control operations on the application program through input controls such as menus, dialog boxes, and keys. .
  • the memory 515 may also store one or more application programs. As shown in FIG. 5, these applications may include: social applications (such as Facebook), image management applications (such as albums), map applications (such as Google Maps), browsers (such as Safari, Google Chrome), etc. .
  • social applications such as Facebook
  • image management applications such as albums
  • map applications such as Google Maps
  • browsers such as Safari, Google Chrome
  • the processor 511 may be used to read and execute computer-readable instructions. Specifically, the processor 511 may be used to call a program stored in the memory 515, for example, an implementation program of a method for verifying a disease score based on a neural network provided in this application, and execute instructions contained in the program.
  • the processor 511 may be used to call a program stored in the memory 515, such as an implementation program of a method for verifying a disease score based on a neural network provided in this application, and execute the following process:
  • the case data including multiple case characteristics
  • the disease category score dictionary includes a one-to-one correspondence between disease category classification and basic disease category scores
  • the first case is an abnormal case according to the score of the first disease type and the score of the second disease type, and if so, output prompt information for prompting the abnormality of the first case.
  • the individual case characteristics include diagnostic identification, drug identification, drug dosage, drug cost, test item identification, test item cost, surgical identification, surgical cost, hospitalization days, hospitalization cost, complication identification, secondary disease identification, A combination of one or more of the age of the insured person and the gender of the insured person.
  • the processor before the processor executes the search for the first basic disease score corresponding to the disease classification of the first case in the disease score dictionary, the processor is further configured to execute:
  • the disease classification of the first case is determined according to the diagnostic identifier used to identify the disease classification of the insured in the first case, wherein the disease classification is an item in the disease classification dictionary, the The disease classification dictionary includes names of M disease classifications and M disease classification codes corresponding to the names of the M disease classifications one by one, and the disease classification codes are ICD codes or the first N digits of ICD codes Code, the N is a positive integer less than 6, and M is a positive integer.
  • the processor executing the calculation of the second disease type score of the first case according to the first basic disease type score includes:
  • Y is the score of the second disease of the first case
  • a 1 is the score of the first basic disease
  • C 1 is the hospital level coefficient of the hospital where the first case is located
  • E i is the additional disease Species score
  • i is the index of the additional disease species score
  • i is a positive integer.
  • the processor executes the judging whether the first case is an abnormal case according to the first disease type score and the second disease type score, specifically including executing:
  • the processor executes the judgment of whether the first case is an abnormal case according to the score of the first disease type and the score of the second disease type, and if yes, an output is used to prompt the
  • the prompt information of the first case abnormality includes:
  • a prompt message for prompting that the score of the disease type of the first case is too high is output;
  • a prompt message for prompting that the score of the disease type of the first case is too low is output.
  • the processor is further configured to: receive a case set through the communication module 516, the case set includes multiple cases, and the first case is any one of the multiple cases; according to the case concentration
  • the case data of the abnormal case in the first screen is displayed on the touch screen 523, and the first image includes at least one of the following: the correspondence between the hospital where the abnormal case in the case set is located and the number of cases, and the abnormal case in the case set The corresponding relationship between the number of attending physicians and the number of cases, the classification of the disease types of the abnormal cases in the case set and the number of cases.
  • computing device 500 is only an example provided by the embodiments of the present application, and the computing device 500 may have more or less components than those shown, may combine two or more components, or may have Different configurations of components are implemented.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM), etc.
  • the modules in the device of the embodiment of the present application may be combined, divided, and deleted according to actual needs.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • General Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Epidemiology (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Development Economics (AREA)
  • Biomedical Technology (AREA)
  • Physics & Mathematics (AREA)
  • Technology Law (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Medical Treatment And Welfare Office Work (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

本申请实施例公开了一种基于神经网络的病种分值校验方法及计算设备，该方法包括计算设备通过接收第一病例的病例数据，该病例数据包括多个病例特征；将该多个病例特征输入到病种分值学习模型，得到第一病例的第一病种分值；以及，在病种分值字典中查找第一病例所属病种分类对应的第一基础病种分值，并根据第一基础病种分值计算第一病例的第二病种分值，进而，根据第一病种分值与第二病种分值判断第一病例是否为异常病例，如果是，则输出用于提示第一病例异常的提示信息。通过执行上述方法，可以实现病例的病例分值的校验，以及时发现异常病例。

Description

一种基于神经网络的病种分值校验方法及计算设备
本申请要求于2018年10月30日提交中国专利局、申请号为201811282996.4、申请名称为“一种基于神经网络的病种分值校验方法及计算设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及医疗管理技术领域，具体涉及一种基于神经网络的病种分值校验方法及计算设备。
背景技术
随着国家公共医疗的布局和医疗改革的不断深入，支付制度改革是系统工程是医保管理理念和医保经办机构角色发生重大转变的体现。推行按照病种付费，充分体现了支付制度才是全面医改的关键所在。
所谓按病种付费，就是指通过统一的疾病诊断分类，科学地制定出每一种疾病的定额偿付标准，社保机构按照该标准与住院人次，向定点医疗机构支付住院费用，使得医疗资源利用标准化，即医疗机构资源消耗与所治疗的住院病人的数量、疾病复杂程度和服务强度成正比。简而言之，就是明确规定某一种疾病该花多少钱，从而既避免了医疗单位滥用医疗服务项目、重复项目和分解项目，防止医院小病大治，又保证了医疗服务质量。
标准化的医疗信息对于医疗信息大数据应用于按病种收费的支付方式非常重要，医疗信息的标准化是实现医疗大数据进行应用的前提。病种的分类目前通常采用国际疾病分类(international Classification of diseases,ICD)。ICD-10根据病因、部位、病理及临床表现等特征将疾病划分为21章节、26000多种病种，并对各个病种进行编码。然而，对于中国的医疗的整体环境来说，各个地区常见的病种远远少于26000，且医疗人员在记录病例时，由于现有技术中病种的多样性、复杂性，医务人员往往不按照国际的标准来等级，各个地区有地区化的语言描述，给按病种付费的实施带来一定困难。
按照规定，各地确定按病种付费支付标准时，应充分考虑医疗服务成本、既往实际发生费用、医保基金承受能力和参保人负担水平等因素，结合病种主要操作和治疗方式，通过与医疗机构协商谈判合理确定，如何根据各地的医疗情况确定支付标准，如何管理按病 种付费的医疗费用，检测分析病例、医疗付费标准等都是目前急需解决的技术问题。
发明内容
本申请实施例提供了一种神经网络的病种分值校验方法及计算设备，可以实现病例的病例分值的校验，以及时发现异常病例。。
第一方面，本申请实施例提供一种基于神经网络的病种分值校验方法，包括：
计算设备接收第一病例的病例数据，所述病例数据包括多个病例特征；
所述计算设备将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
所述计算设备在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值，并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
所述计算设备根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息。
第二方面，本申请实施例还提供了一种计算设备，包括：
接收单元，用于接收第一病例的病例数据，所述病例数据包括多个病例特征；
第一计算单元，用于将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
查找单元，用于在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值；
第二计算单元，用于并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
判断单元，用于根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，
输出单元，用于，当所述判断单元的判断结果为是时，输出用于提示所述第一病例异常的提示信息。
第三方面，本申请实施例还提供了一种计算设备，该计算设备包括处理器、存储器以及通信模块，所述处理器耦合到所述存储器、所述通信模块，所述处理器用于调用所述存储器存储的程序代码执行：
通过所述通信模块接收第一病例的病例数据，所述病例数据包括多个病例特征；
将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值，并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息。
第四方面，本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质用于计算机软件指令，所述计算机软件指令当被计算机执行时使所述计算机执行如第一方面所述任意一种基于神经网络的病种分值校验方法。
第五方面，本申请实施例还提供了一种计算机程序，所述计算机程序包括计算机软件指令，所述计算机软件指令当被计算机执行时使所述计算机执行如第一方面所述的任意一种基于神经网络的病种分值校验方法。
综上，计算设备通过接收第一病例的病例数据，该病例数据包括多个病例特征；
将该多个病例特征输入到病种分值学习模型，得到第一病例的第一病种分值；以及，在病种分值字典中查找第一病例所属病种分类对应的第一基础病种分值，并根据第一基础病种分值计算第一病例的第二病种分值，该病种分值字典包括病种分类与基础病种分值的一一对应关系；进而，根据第一病种分值与第二病种分值判断第一病例是否为异常病例，如果是，则输出用于提示第一病例异常的提示信息。通过执行上述方法，可以实现病例的病例分值的校验，以及时发现异常病例。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍。
图1为本申请实施例提供的一种医保管理平台的功能架构图；
图2为本申请实施例提供的一种病种分值校验方法的流程图；
图3为本申请实施例提供的一种计算设备的结构示意图；
图4为本申请实施例提供的又一种计算设备的结构示意图；
图5为本申请实施例提供的又一种计算设备的结构示意图。
具体实施方式
需要说明的是，在本申请实施例中使用的术语是仅仅出于描述特定实施例的目的，而非旨在限制本申请。在本申请实施例和所附权利要求书中所使用的单数形式的“一种”、“所述”和“该”也旨在包括多数形式，除非上下文清楚地表示其他含义。还应当理解，本文中使用的术语“和/或”是指并包含一个或多个相关联的列出项目的任何或所有可能组合。
为了更好理解本申请实施例，下面先对本申请实施例适用的医保管理平台的各个功能进行描述，请参阅图1，图1为本申请实施例提供的一种医保管理平台的功能架构图，该医保管理平台可以运行在计算设备中，为医保管理平台的运行商提供的一系列和病例、医保、病种分值等相关的功能，该医保管理平台包括但不限于如下部分或全部功能的实现：
病种编码，医保管理平台可以根据输入病例的病例数据对该病例中主诊断得到的病种进行编码，该病种编码方法可以采用ICD-10编码(本申请中也称六位码编码)、也可以采用其他编码方法，例如四位码编码(即六位码的前4位)、三位码编码(即六位码的前3位)等。可以理解，可以通过某一地区发生病例集通过病种编码方法建立适用该地区的病种分类字典，该病种分类字典包括M个病种分类的名称以及与所述M个病种分类的名称一一对应的M个病种分类码，M为正整数。可选地，计算设备可以基于医保管理平台识别病例数据中有医务人员填写的诊断名称、病种编码等信息识别到该病例对应的病种分类，进而将该病种分类对应的病种分类码补充到病例数据中，以便于进一步地计算该病例的病种分值，进而实现按病种付费、基于病种分值进行病例数据真实性的检测等功能。
病种分值计算，医保管理平台可以存储病种与病种分值的对应关系表或者包含病种分值计算程序，可以通过病例中病种名称、病种编码等确定该病例中参保人(即病人)所患病种分类，进而根据病种分类与病种分值的对应关系或者病种分值计算程序等基于病种分类确定病种分值的实现过程确定该病例的病种分值。其中，病种分值为地区(比如，国家、省或市等)基于病例大数据确定的用于计算医疗费用(比如预测医保费用、预测总费用等)的标准分值。具体的，可以建立病种分值字典，该病种分值字典包括病种分类与基础病种分值的一一对应关系，再根据病例的实际情况(比如参保人年龄、患病的严重程度、所在医院、所属科室等信息)在基础病种分值的基础上进行调整，以得到适合该病例的病种分值。病种分值与医疗费用呈正相关关系，即病种分值越高，该病种的医疗费用越高。
病例数据的统计分析，医保管理平台可以按照评估周期(比如，月度、季度、年度等) 对该地区内各个医院上报的病例进行统计分析。上述对病例的统计分析可以支持按月度、季度、年度等进行统计分析，支持对不同医院、不同费用区间、不同病种等中的一种或多种的组合进行发生例数、总费用、实际医保费用、预测医保费用等的统计分析，以基于统计分析结果对下个评估周期所采用的各个病种的病种分值进行调整。应理解，基于统计分析结果还可以实现对其他功能，比如基于统计得到的各个医院的收入和支出对医院的医院级别系数进行调整等，对此，本申请实施例不作限定。
病例的真实性检测，医保管理平台可以基于病例中的病例数据对该病例的真实性进行检测，当检测到该病例包含虚假数据时，对该病例进行标记、输出该病例包含虚假数据或异常病例的提示消息等，以便于医保管理平台的运行商及时识别到异常病例，并分析异常病例原因。
数据可视化，医保管理平台可以对病例数据的统计分析功能得到的统计分析结果进行可视化，也可以对问题病例进行统计分析的结果进行可视化，以便于医保管理平台的运行商统计分析结果。
本申请中，计算设备可以包括但不限于移动电话、移动电脑、平板电脑、媒体播放器、计算机、服务器等包含数据处理功能的设备。运行医保管理平台各个功能的计算设备可以接收到来自医院等机构或个体上报的病例。
不限于图1所示，本申请提供的医保管理平台还可以包括其他功能的实现，例如，病种分值的优化等，对此，本申请实施例不作限定。
请参见图2，图2是本申请提供的一种基于神经网络的病种分值校验方法。在图2实施例中，以病种分值计算的执行主体为计算设备(运行病例管理平台各个功能的设备)为例来描述，可以理解，该病种分值计算还可以由其他终端或服务器等具备数据处理功能的设备，对此，本申请实施例不作限定。如图2所示，该方法可以包括但不限于如下部分或全部步骤：
S2：接收第一病例的病例数据，所述病例数据包括多个病例特征。
其中，病例为由医院针对病人记录的病人诊断治疗过程。病例数据可以包括但不限于病人的个人信息、诊断信息、治疗信息、费用信息、实际医保费用等中的一种或多种的组合。其中，诊断信息可以包括用于识别参保人发生病种分类的诊断标识。其中，诊断标识可以是诊断名称，比如主诊断名称；还可以是诊断编码，如ICD诊断编码等；还可以是手术标识可以是手术名称、手术编码等。应理解，个人信息可以包括但不限于参保人的年龄、 性别、病史等信息。治疗信息为病例中记载参保人治疗的过程信息。费用信息包括但不限于参保人在本次疾病治疗过程中产生的手术费、住院费、检测费、挂号费、药品费、总费用等中的一种或多种的组合。
具体的，计算设备可以根据病例数据识别并提取该病例数据的病例特征，该病例特征可以包括但不限于诊断标识、药品标识、药品剂量、药品费用、检测项的标识、检测项费用、手术标识、手术费用、住院天数、住院费用、并发症标识、继发症标识、参保人年龄、参保人性别等中的一种或多种的组合。
S4：将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值。
本申请中病种分值学习模型为神经网络模型，用于根据病例的多个病例特征识别该病例数据对应的病种分值。输入到病种分值学习模型的数据可以包括但不限于上述列举的一项或多项病例特征的组合。
病种分值学习模型在进行病种分值预测前，需要通过样本数据对该病种分值学习模型进行训练，以学习到根据病例的多个病例特征识别到该病例对应的病种分值。可选地，样本数据可以是未实施病种分值计算医保费用之前获取到的数据，以保证样本数据的准确性。可以理解，未实施病种分值计算医保费用之前获取到的病例数据不存在医务人员为提高病种分值而进行主诊断与副诊断调换、增加住院项等虚假信息的嫌疑，通过其训练的数据具有较好的可靠性。样本数据中病例的真实病种分值可以通过该病例的实际医保费用来计算。第二病例为样本数据中任意一个病例，本申请第二病例为例来说明样本数据中各个病例对应的真实病种分值的计算。
第二病例的真实病种分值可以是第二病例的医保费用与预设常数的比值，其中，预设常数是可配置的，由可定基准的机构设置，该预设常数可以是50、100等或其他数值，本申请实施例不作限定。可以通过监督预测病种分值和真实病种分值，训练病种分值学习模型。
S6：在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值，并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系。
第一病例的病例数据包括用于识别病种分类的诊断标识。在本申请的一种实现中，第一病例已通过病种分类识别方法识别到第一病例的病种分类并将识别到的病种分类(例如， 病种分类的标识)添加到第一病例的病例数据中。该诊断标识即为病种分类的标识(病种分类名称或病种分类码)。可以理解，病种分类的标识为病种分类字典中包括的病种分类的名称或者病种分类码，其中，病种分类字典包括病种分类的名称与病种分类码的一一对应关系，M为正整数。可选地，病种分类码为ICD编码，该病种分类字典即为ICD字典；或，该病种分类码为ICD编码的前N位码，N为小于6的正整数。
对于病种分类字典中病种分类码为ICD编码的前N位码来说，选择病种的ICD编码的前四位、前三位还是前二位，取决于病例集中病例的病种分类为ICD编码的前四位的例数，比如大于10例，选“四位码”作为病种分类码；若小于10例，选“三位码”作为病种分类码。
在本申请的另一种实现中，第一病例中的诊断标识为诊断名称或诊断编码，通过该诊断名称或诊断编码不能直接得到第一病例所属的病种分类。此时，S6之前，计算设备可以根据所述第一病例中用于识别参保人的发生病种分类的诊断标识确定所述第一病例的病种分类。
计算设备可以确定根据第一病例的诊断标识确定该第一病例所属的病种分类的一种实现方式可以是：第一病种的诊断标识可以包括诊断名称，计算设备可以预存病种名称对照表，该病种名称对照表包括M个病种分类以及所述M个病种分类中每一个病种分类对应的一个或多个诊断名称。进而，计算设备可以根据病种名称对照表确定第一病种的诊断名称对应的病种分类，进一步地根据病种分类字典，确定第二病例的病种分类对应的病种分类码。
计算设备可以确定根据第一病例的诊断标识确定该第一病例所属的病种分类的另一种实现方式可以是：第一病种的诊断标识可以包括诊断编码，该诊断编码可以是ICD-10编码、ICD-9-CM3手术编码、肿瘤形态学编码(也称M码)或中医疾病编码等。
对于ICD-10编码或ICD-9-CM3手术编码来说，计算设备可以直接在病种分类字典中查找与第一病例中诊断编码相匹配的病种分类码。
对于M码来说，可以根据M码转换表，将M码转换为ICD编码或四位ICD码。例如，M码“M8140/6”对应的ICD编码“C78.7”，M码“M8140/3”对应的ICD编码“C34.9”。计算设备可以根据M码转换表将第一病例中的M码转换为ICD编码(该ICD编码可以是病种分类字典中的病种分类码，也可以是具有六位编码的ICD编码等)，进而在病种分类字典中查找与转换得到的ICD编码相匹配的病种分类码。
本申请一实施例中，计算设备根据所述第一基础病种分值计算所述第一病例的第二病种分值的一种实现方式可以是：
Figure PCTCN2019095011-appb-000001
其中，Y为所述第一病例的第二病种分值，A 1为所述第一基础病种分值，C 1为所述第一病例所在医院的医院级别系数，E i为附加病种分值，i为所述附加病种分值的索引，i为正整数。
其中，在第一病例满足包含预设手术、预设并发症、预设继发症、预设住院信息、儿科病例等中的一项或多项，第二病种分值需要附加该第一病例所满足的项对应的病种附加分值。可以理解，对于第一病例，不同的项对应的病种附加分值可以不同。对于同一项，不同病种类型的病例对应的病种附加分值可以不同。
可以理解，第二病种分值还可以包括其他计算方式，例如，Y＝∑ iA 1*C 1，本申请实施例不作限定。
S8：根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息。
可以理解，按病种付费的医疗支付方式，通过识别病例的病种分类，进而从病种分值字典中查找到该病种分类对应的病种分值。然而，病例的病例数据可能存在虚假数据，使得病例从病种分值字典中查找到的病种分值远远不符合该病例的实际情况。例如，病例中的诊断信息、治疗信息、费用信息、医保费用不符合该病种分类的诊治方法，例如，病例中针对病种“高血压”，采用的治疗手段包括手术、药品中包括用于治疗肿瘤的药品。或者病例中针对病种“糖尿病”，采用的治疗方法仅仅是消炎药类的药品，则该病例数据可能存在虚假现象。此时，可以将病例的病例数据输入到病种分值学习模型，对该病种分值从治疗过程方面进行病种分值的评估，得到该病例的第一病种分值，进而将其与通过病种分值字典中查找到该病例所属病种分类对应的第二病种分值进行比对分析，当两者差异较大时，认为该病例是否为异常病例。
可选地，计算设备判断所述第一病例是否为异常病例的一种实现方式可以是：计算设备判断所述第一病种分值与所述第二病种分值的差值是否大于第一阈值，如果是，则所述第一病例为异常病例。其中，第一阈值可以是10、20、35或其他数值等。可选地，第一阈值Q可以根据第一病种分值Y’ 1或第二病种分值Y 1设定，例如，Q＝Y’ 1*μ，或Q＝Y 1*μ，其 中，0<μ<1。
可选地，计算设备判断所述第一病例是否为异常病例的另一种实现方式可以是：计算设备判断所述第二病种分值与所述第一病种分值的比值是否大于第二阈值或小于第三阈值，如果是，则所述第一病例为异常病例。其中，第三阈值可以是的取值范围可以是1.1-3，例如，1.5、2、2.4或其他数值，本申请实施例不作限定。第二阈值可以是0.1-0.9，例如0.4、0.5、0.7或其他数值，本申请实施例不作限定。
可选地，当第二病种分值与第一病种分值的比例大于第二阈值时，输出用于提示该第一病例的病种分值过高的提示信息；当第二病种分值与第一病种分值的比例小于第一阈值且大于第二阈值时，输出用于提示该第一病例的病种分值在正常范围的提示信息；当第二病种分值与第一病种分值的比例小于第二阈值时，输出用于提示该第一病例的病种分值过低的提示信息。
本申请一实施例中，该方法还包括：
接收病例集，所述病例集包括多个病例，上述第一病例可以是该多个病例中任意一个；
根据病例集中的异常病例的病例数据，显示第一图像，所述第一图像包括以下至少一项：所述病例集中的异常病例的所在医院和例数的对应关系、所述病例集中的异常病例的主治医师和例数的对应关系、所述病例集中的异常病例的病种分类与例数的对应关系。
可见，通过可视化异常病例的病例数据对异常病例进行分析，以助于医保管理平台的管理人员根据可视化的第一图像迅速找到异常病例的问题所在。
综上，本申请实施例中计算设备通过接收第一病例的病例数据，该病例数据包括多个病例特征；将该多个病例特征输入到病种分值学习模型，得到第一病例的第一病种分值；以及，在病种分值字典中查找第一病例所属病种分类对应的第一基础病种分值，并根据第一基础病种分值计算第一病例的第二病种分值，进而，根据第一病种分值与第二病种分值判断第一病例是否为异常病例，如果是，则输出用于提示第一病例异常的提示信息。通过执行上述方法，可以实现病例的病例分值的校验，以及时发现异常病例。
下面介绍发明实施例涉及的装置。
请参阅图3计算设备30，包括但不限于：接收单元31、第一计算单元32查找单元33、第二计算单元34、判断单元35和输出单元36等。其中，
接收单元31，用于接收第一病例的病例数据，所述病例数据包括多个病例特征；
第一计算单元32，用于将所述多个病例特征输入到病种分值学习模型，得到所述第一 病例的第一病种分值；
查找单元33，用于在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值；
第二计算单元34，用于并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
判断单元35，用于根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，
输出单元36，用于，当所述判断单元的判断结果为是时，输出用于提示所述第一病例异常的提示信息。
可选地，单个病例特征包括诊断标识、药品标识、药品剂量、药品费用、检测项的标识、检测项费用、手术标识、手术费用、住院天数、住院费用、并发症标识、继发症标识、参保人年龄、参保人性别中的一种或多种的组合。
如图4所示的计算设备，所述计算设备40还包括：
分类识别单元37，用于根据所述第一病例中用于识别参保人的发生病种分类的诊断标识确定所述第一病例的病种分类，其中，所述病种分类为病种分类字典中的项，所述病种分类字典包括M个病种分类的名称以及与所述M个病种分类的名称一一对应的M个病种分类码，所述病种分类码为ICD编码或ICD编码的前N位码，所述N为小于6的正整数，M为正整数。
可选地，所述第二计算单元34具体通过下述公式计算所述第一病例的第二病种分值包括：
Figure PCTCN2019095011-appb-000002
其中，Y为所述第一病例的第二病种分值，A 1为所述第一基础病种分值，C 1为所述第一病例所在医院的医院级别系数，E i为附加病种分值，i为所述附加病种分值的索引，i为正整数。
可选地，所述判断单元35具体用于：
判断所述第一病种分值与所述第二病种分值的差值是否大于第一阈值，如果是，则所述第一病例为异常病例。
可选地，所述判断单元35具体用于：判断所述第二病种分值与所述第一病种分值的比 值是否大于第二阈值或小于第三阈值；
所述输出单元36具体用于：
当所述第二病种分值与所述第一病种分值的比例大于所述第二阈值时，输出用于提示所述第一病例的病种分值过高的提示信息；
当所述第二病种分值与所述第一病种分值的比例小于所述第一阈值且大于第二阈值时，输出用于提示所述第一病例的病种分值在正常范围的提示信息；
当所述第二病种分值与所述第一病种分值的比例小于所述第二阈值时，输出用于提示所述第一病例的病种分值过低的提示信息。
可选地，所述接收单元31还用于接收病例集，所述病例集包括多个病例，所述第一病例为所述多个病例中任意一个；
所述输出单元36还用于：根据病例集中的异常病例的病例数据，显示第一图像，所述第一图像包括以下至少一项：所述病例集中的异常病例的所在医院和例数的对应关系、所述病例集中的异常病例的主治医师和例数的对应关系、所述病例集中的异常病例的病种分类与例数的对应关系。
需要说明的是，上述计算设备的各个单元的具体实现可以参见上述方法实施例中相关描述，本申请不再赘述。
如图5所示的计算设备，该计算设备500可包括：基带芯片510、存储器515(一个或多个计算机可读存储介质)、通信模块516(例如，射频(RF)模块5161和/或通信接口5162)、外围系统517。这些部件可在一个或多个通信总线514上通信。
外围系统517主要用于实现计算设备510和用户/外部环境之间的交互功能，主要包括计算设备500的输入/输出装置。具体实现中，外围系统517可包括：触摸屏控制器518、摄像头控制器519、音频控制器520以及传感器管理模块521。其中，各个控制器可与各自对应的外围设备(如触摸屏523、摄像头524、音频电路525以及传感器526)耦合。需要说明的，外围系统517还可以包括其他I/O外设。
基带芯片510可集成包括：一个或多个处理器511、时钟模块522以及电源管理模块513。集成于基带芯片510中的时钟模块522主要用于为处理器511产生数据传输和时序控制所需要的时钟。集成于基带芯片510中的电源管理模块513主要用于为处理器511、射频模块5161以及外围系统提供稳定的、高精确度的电压。
射频(RF)模块5161用于接收和发送射频信号，主要集成了计算设备500的接收器 和发射器。射频(RF)模块5161通过射频信号与通信网络和其他通信设备通信。具体实现中，射频(RF)模块5161可包括但不限于：天线系统、RF收发器、一个或多个放大器、调谐器、一个或多个振荡器、数字信号处理器、CODEC芯片、SIM卡和存储介质等。在一些实施例中，可在单独的芯片上实现射频(RF)模块5161。
通信模块516用于计算设备500与其他设备之间的数据交换。
存储器515与处理器511耦合，用于存储各种软件程序和/或多组指令。具体实现中，存储器515可包括高速随机存取的存储器，并且也可包括非易失性存储器，例如一个或多个磁盘存储设备、闪存设备或其他非易失性固态存储设备。存储器515可以存储操作系统(下述简称系统)，例如ANDROID，IOS，WINDOWS，或者LINUX等嵌入式操作系统。存储器515还可以存储网络通信程序，该网络通信程序可用于与一个或多个附加设备，一个或多个计算设备设备，一个或多个网络设备进行通信。存储器515还可以存储用户接口程序，该用户接口程序可以通过图形化的操作界面将应用程序的内容形象逼真的显示出来，并通过菜单、对话框以及按键等输入控件接收用户对应用程序的控制操作。
存储器515还可以存储一个或多个应用程序。如图5所示，这些应用程序可包括：社交应用程序(例如Facebook)，图像管理应用程序(例如相册)，地图类应用程序(例如谷歌地图)，浏览器(例如Safari，Google Chrome)等等。
本申请中，处理器511可用于读取和执行计算机可读指令。具体的，处理器511可用于调用存储于存储器515中的程序，例如本申请提供的基于神经网络的病种分值校验方法的实现程序，并执行该程序包含的指令。
具体的，处理器511可用于调用存储于存储器515中的程序，如本申请提供的基于神经网络的病种分值校验方法的实现程序，并执行下述流程：
通过所述通信模块516接收第一病例的病例数据，所述病例数据包括多个病例特征；
将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值，并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息。
可选地，单个病例特征包括诊断标识、药品标识、药品剂量、药品费用、检测项的标 识、检测项费用、手术标识、手术费用、住院天数、住院费用、并发症标识、继发症标识、参保人年龄、参保人性别中的一种或多种的组合。
可选地，所述处理器执行所述在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值之前，所述处理器还用于执行：
根据所述第一病例中用于识别参保人的发生病种分类的诊断标识确定所述第一病例的病种分类，其中，所述病种分类为病种分类字典中的项，所述病种分类字典包括M个病种分类的名称以及与所述M个病种分类的名称一一对应的M个病种分类码，所述病种分类码为ICD编码或ICD编码的前N位码，所述N为小于6的正整数，M为正整数。
可选地，所述处理器执行所述根据所述第一基础病种分值计算所述第一病例的第二病种分值包括：
Figure PCTCN2019095011-appb-000003
其中，Y为所述第一病例的第二病种分值，A 1为所述第一基础病种分值，C 1为所述第一病例所在医院的医院级别系数，E i为附加病种分值，i为所述附加病种分值的索引，i为正整数。
可选地，所述处理器执行所述根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，具体包括执行：
判断所述第一病种分值与所述第二病种分值的差值是否大于第一阈值，如果是，则所述第一病例为异常病例。
可选地，所述处理器执行所述根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息，具体包括：
判断所述第二病种分值与所述第一病种分值的比值是否大于第二阈值或小于第三阈值；
当所述第二病种分值与所述第一病种分值的比例大于所述第二阈值时，输出用于提示所述第一病例的病种分值过高的提示信息；
当所述第二病种分值与所述第一病种分值的比例小于所述第一阈值且大于第二阈值时，输出用于提示所述第一病例的病种分值在正常范围的提示信息；
当所述第二病种分值与所述第一病种分值的比例小于所述第二阈值时，输出用于提示所述第一病例的病种分值过低的提示信息。
可选地，所述处理器还用于执行：通过所述通信模块516接收病例集，所述病例集包括多个病例，所述第一病例为所述多个病例中任意一个；根据病例集中的异常病例的病例数据，通过触摸屏523显示第一图像，所述第一图像包括以下至少一项：所述病例集中的异常病例的所在医院和例数的对应关系、所述病例集中的异常病例的主治医师和例数的对应关系、所述病例集中的异常病例的病种分类与例数的对应关系。
可以理解，上述各个流程和各个功能单元的具体实现可以参照上述方法实施例中相关描述，本申请实施例不再赘述。
应当理解，计算设备500仅为本申请实施例提供的一个例子，并且，计算设备500可具有比示出的部件更多或更少的部件，可以组合两个或更多个部件，或者可具有部件的不同配置实现。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见其他实施例的相关描述。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)或随机存储记忆体(Random Access Memory，RAM)等。
本申请实施例方法中的步骤可以根据实际需要进行顺序调整、合并和删减。
本申请实施例装置中的模块可以根据实际需要进行合并、划分和删减。

Claims (20)

  1. 一种基于神经网络的病种分值校验方法，其特征在于，包括：
    计算设备接收第一病例的病例数据，所述病例数据包括多个病例特征；
    所述计算设备将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
    所述计算设备在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值，并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
    所述计算设备根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息。
  2. 如权利要求1所述的方法，其特征在于，单个病例特征包括诊断标识、药品标识、药品剂量、药品费用、检测项的标识、检测项费用、手术标识、手术费用、住院天数、住院费用、并发症标识、继发症标识、参保人年龄、参保人性别中的一种。
  3. 如权利要求1所述的方法，其特征在于，所述计算设备在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值之前，所述方法还包括：
    所述计算设备根据所述第一病例中用于识别参保人的发生病种分类的诊断标识确定所述第一病例的病种分类，其中，所述病种分类为病种分类字典中的项，所述病种分类字典包括病种分类的名称与病种分类码的对应关系，所述病种分类码为ICD编码或ICD编码的前N位码，所述N为小于6的正整数，M为正整数。
  4. 如权利要求1所述的方法，其特征在于，所述根据所述第一基础病种分值计算所述第一病例的第二病种分值包括：
    Figure PCTCN2019095011-appb-100001
    其中，Y为所述第一病例的第二病种分值，A 1为所述第一基础病种分值，C 1为所述第一病例所在医院的医院级别系数，E i为附加病种分值，i为所述附加病种分值的索引，i为正整数。
  5. 如权利要求1所述的方法，其特征在于，所述计算设备根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例包括：
    所述计算设备判断所述第一病种分值与所述第二病种分值的差值是否大于第一阈值，如果是，则所述第一病例为异常病例。
  6. 如权利要求1所述的方法，其特征在于，所述计算设备根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息包括：
    所述计算设备判断所述第二病种分值与所述第一病种分值的比值是否大于第二阈值或小于第三阈值；
    当所述第二病种分值与所述第一病种分值的比例大于所述第二阈值时，输出用于提示所述第一病例的病种分值过高的提示信息；
    当所述第二病种分值与所述第一病种分值的比例小于所述第一阈值且大于第二阈值时，输出用于提示所述第一病例的病种分值在正常范围的提示信息；
    当所述第二病种分值与所述第一病种分值的比例小于所述第二阈值时，输出用于提示所述第一病例的病种分值过低的提示信息。
  7. 如权利要求1-6任一项所述的方法，其特征在于，所述方法还包括：
    接收病例集，所述病例集包括多个病例，所述第一病例为所述多个病例中任意一个；
    根据病例集中的异常病例的病例数据，显示第一图像，所述第一图像包括以下至少一项：所述病例集中的异常病例的所在医院和例数的对应关系、所述病例集中的异常病例的主治医师和例数的对应关系、所述病例集中的异常病例的病种分类与例数的对应关系。
  8. 一种计算设备，其特征在于，包括处理器、存储器以及通信模块，所述处理器耦合所述存储器、所述通信模块，所述处理器用于调用所述存储器存储的程序代码执行：
    通过所述通信模块接收第一病例的病例数据，所述病例数据包括多个病例特征；
    将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
    在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值，并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
    根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息。
  9. 如权利要求8所述的计算设备，其特征在于，所述处理器执行所述在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值之前，所述处理器还执行：
    根据所述第一病例中用于识别参保人的发生病种分类的诊断标识确定所述第一病例的病种分类，其中，所述病种分类为病种分类字典中的项，所述病种分类字典包括病种分类的名称与病种分类码的对应关系，所述病种分类码为ICD编码或ICD编码的前N位码，所述N为小于6的正整数，M为正整数。
  10. 如权利要求8所述的计算设备，其特征在于，所述处理器执行所述根据所述第一基础病种分值计算所述第一病例的第二病种分值包括：
    Figure PCTCN2019095011-appb-100002
    其中，Y为所述第一病例的第二病种分值，A 1为所述第一基础病种分值，C 1为所述第一病例所在医院的医院级别系数，E i为附加病种分值，i为所述附加病种分值的索引，i为正整数。
  11. 如权利要求8所述的计算设备，其特征在于，所述处理器执行所述根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，包括执行：
    判断所述第一病种分值与所述第二病种分值的差值是否大于第一阈值，如果是，则所述第一病例为异常病例。
  12. 如权利要求8所述的计算设备，其特征在于，所述处理器执行所述根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，如果是，则输出用于提示所述第一病例异常的提示信息，具体包括执行：
    判断所述第二病种分值与所述第一病种分值的比值是否大于第二阈值或小于第三阈值；
    当所述第二病种分值与所述第一病种分值的比例大于所述第二阈值时，输出用于提示所述第一病例的病种分值过高的提示信息；
    当所述第二病种分值与所述第一病种分值的比例小于所述第一阈值且大于第二阈值时，输出用于提示所述第一病例的病种分值在正常范围的提示信息；
    当所述第二病种分值与所述第一病种分值的比例小于所述第二阈值时，输出用于提示所述第一病例的病种分值过低的提示信息。
  13. 如权利要求8-12任一项所述的计算设备，其特征在于，所述处理器还用于执行：
    接收病例集，所述病例集包括多个病例，所述第一病例为所述多个病例中任意一个；
    根据病例集中的异常病例的病例数据，显示第一图像，所述第一图像包括以下至少一项：所述病例集中的异常病例的所在医院和例数的对应关系、所述病例集中的异常病例的 主治医师和例数的对应关系、所述病例集中的异常病例的病种分类与例数的对应关系。
  14. 一种计算设备，其特征在于，包括
    接收单元，用于接收第一病例的病例数据，所述病例数据包括多个病例特征；
    第一计算单元，用于将所述多个病例特征输入到病种分值学习模型，得到所述第一病例的第一病种分值；
    查找单元，用于在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值；
    第二计算单元，用于并根据所述第一基础病种分值计算所述第一病例的第二病种分值，所述病种分值字典包括病种分类与基础病种分值的一一对应关系；
    判断单元，用于根据所述第一病种分值与所述第二病种分值判断所述第一病例是否为异常病例，
    输出单元，用于，当所述判断单元的判断结果为是时，输出用于提示所述第一病例异常的提示信息。
  15. 如权利要求14所述的计算设备，其特征在于，所述计算设备还包括：
    分类识别单元，用于在所述查找单元执行所述在病种分值字典中查找所述第一病例所属病种分类对应的第一基础病种分值之前，根据所述第一病例中用于识别参保人的发生病种分类的诊断标识确定所述第一病例的病种分类，其中，所述病种分类为病种分类字典中的项，所述病种分类字典包括病种分类的名称与病种分类码的对应关系所述病种分类码为ICD编码或ICD编码的前N位码，所述N为小于6的正整数。
  16. 如权利要求14所述的计算设备，其特征在于，所述第二计算单元具体通过下述公式计算所述第一病例的第二病种分值包括：
    Figure PCTCN2019095011-appb-100003
    其中，Y为所述第一病例的第二病种分值，A 1为所述第一基础病种分值，C 1为所述第一病例所在医院的医院级别系数，E i为附加病种分值，i为所述附加病种分值的索引，i为正整数。
  17. 如权利要求14所述的计算设备，其特征在于，所述判断单元具体用于：
    判断所述第一病种分值与所述第二病种分值的差值是否大于第一阈值，如果是，则所述第一病例为异常病例。
  18. 如权利要求14所述的计算设备，其特征在于，所述判断单元具体用于：判断所述第二病种分值与所述第一病种分值的比值是否大于第二阈值或小于第三阈值；
    所述输出单元具体用于：
    当所述第二病种分值与所述第一病种分值的比例大于所述第二阈值时，输出用于提示所述第一病例的病种分值过高的提示信息；
    当所述第二病种分值与所述第一病种分值的比例小于所述第一阈值且大于第二阈值时，输出用于提示所述第一病例的病种分值在正常范围的提示信息；
    当所述第二病种分值与所述第一病种分值的比例小于所述第二阈值时，输出用于提示所述第一病例的病种分值过低的提示信息。
  19. 如权利要求14-18任一项所述的计算设备，其特征在于，所述接收单元还用于接收病例集，所述病例集包括多个病例，所述第一病例为所述多个病例中任意一个；
    所述输出单元还用于：根据病例集中的异常病例的病例数据，显示第一图像，所述第一图像包括以下至少一项：所述病例集中的异常病例的所在医院和例数的对应关系、所述病例集中的异常病例的主治医师和例数的对应关系、所述病例集中的异常病例的病种分类与例数的对应关系。
  20. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质用于计算机软件指令，所述计算机软件指令当被计算机执行时使所述计算机执行如权利要求1-7中任一项权利要求所述基于神经网络的病种分值校验方法。
PCT/CN2019/095011 2018-10-30 2019-07-08 一种基于神经网络的病种分值校验方法及计算设备 WO2020087970A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811282996.4A CN109493243A (zh) 2018-10-30 2018-10-30 一种基于神经网络的病种分值校验方法及计算设备
CN201811282996.4 2018-10-30

Publications (1)

Publication Number Publication Date
WO2020087970A1 true WO2020087970A1 (zh) 2020-05-07

Family

ID=65692030

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/095011 WO2020087970A1 (zh) 2018-10-30 2019-07-08 一种基于神经网络的病种分值校验方法及计算设备

Country Status (2)

Country Link
CN (1) CN109493243A (zh)
WO (1) WO2020087970A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109493243A (zh) * 2018-10-30 2019-03-19 平安医疗健康管理股份有限公司 一种基于神经网络的病种分值校验方法及计算设备
CN111584021B (zh) * 2020-05-08 2023-08-18 青岛国新健康产业科技有限公司 病案信息校验方法、装置、电子设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105069030A (zh) * 2015-07-20 2015-11-18 赵蒙海 单病种的住院费用估计判定方法
CN106327395A (zh) * 2016-08-31 2017-01-11 杭州逸曜信息技术有限公司 医疗费用信息的处理方法
CN106408141A (zh) * 2015-07-28 2017-02-15 平安科技（深圳）有限公司 异常费用自动提取系统及方法
CN107123072A (zh) * 2017-05-11 2017-09-01 杭州逸曜信息技术有限公司 医疗费用限制规则生成方法
CN109493243A (zh) * 2018-10-30 2019-03-19 平安医疗健康管理股份有限公司 一种基于神经网络的病种分值校验方法及计算设备

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7444291B1 (en) * 2000-08-10 2008-10-28 Ingenix, Inc. System and method for modeling of healthcare utilization
US8510240B2 (en) * 2011-03-31 2013-08-13 Infosys Limited System and method for automatically generating a medical code
US20140149130A1 (en) * 2012-11-29 2014-05-29 Verizon Patent And Licensing Inc. Healthcare fraud detection based on statistics, learning, and parameters
US20140149128A1 (en) * 2012-11-29 2014-05-29 Verizon Patent And Licensing Inc. Healthcare fraud detection with machine learning
CN106202477A (zh) * 2016-07-18 2016-12-07 北京千安哲信息技术有限公司 医疗费用挖掘方法及装置
CN106407686B (zh) * 2016-09-23 2019-04-12 电子科技大学 一种评估慢性病费用的建模方法
CN107609980A (zh) * 2017-09-07 2018-01-19 平安医疗健康管理股份有限公司 医疗数据处理方法、装置、计算机设备及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105069030A (zh) * 2015-07-20 2015-11-18 赵蒙海 单病种的住院费用估计判定方法
CN106408141A (zh) * 2015-07-28 2017-02-15 平安科技（深圳）有限公司 异常费用自动提取系统及方法
CN106327395A (zh) * 2016-08-31 2017-01-11 杭州逸曜信息技术有限公司 医疗费用信息的处理方法
CN107123072A (zh) * 2017-05-11 2017-09-01 杭州逸曜信息技术有限公司 医疗费用限制规则生成方法
CN109493243A (zh) * 2018-10-30 2019-03-19 平安医疗健康管理股份有限公司 一种基于神经网络的病种分值校验方法及计算设备

Also Published As

Publication number Publication date
CN109493243A (zh) 2019-03-19

Similar Documents

Publication Publication Date Title
WO2020087969A1 (zh) 一种基于数据分析的异常病例识别方法及计算设备
WO2021068601A1 (zh) 病历检测方法、装置、设备及存储介质
TWI744542B (zh) 保險服務優化系統
Kishore et al. Clinical risk scores for predicting stroke-associated pneumonia: a systematic review
US9529968B2 (en) System and method of integrating mobile medical data into a database centric analytical process, and clinical workflow
US20200185069A1 (en) Medical coding quality control
CN110504035A (zh) 医疗资料库及系统
CN117271804B (zh) 一种共病特征知识库生成方法、装置、设备及介质
WO2020087970A1 (zh) 一种基于神经网络的病种分值校验方法及计算设备
CN111816319A (zh) 逐步筛选的泌尿系统重疾指标确定方法及风险预测系统
CN117409913A (zh) 一种基于云技术的医疗服务方法和平台
WO2020082804A1 (zh) 一种医疗数据分类存储方法和装置
US20220044328A1 (en) Machine learning systems and methods to evaluate a claim submission
CN113903423A (zh) 用药方案推荐方法、装置、设备及介质
CN109545297A (zh) 一种基于大数据的病种编码方法及计算设备
CN115831298B (zh) 基于医院管理信息系统的临床试验患者招募方法及装置
CN108447533A (zh) 一种多功能智慧医疗系统
US20170308649A1 (en) Integrating trauma documentation into an electronic medical record
CN109523406A (zh) 一种基于数据分析的病种分值计算方法及计算设备
CN109544374B (zh) 一种基于大数据的病种分值调整方法及计算设备
CN110265127A (zh) 病种费用测算方法、装置及终端设备
CN115762704A (zh) 一种处方审核方法、装置、设备和存储介质
CN109544375A (zh) 一种基于大数据的医院盈亏分析方法及计算设备
Velummailum et al. Data challenges for externally controlled trials
US8756234B1 (en) Information theory entropy reduction program

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19878236

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 22/06/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19878236

Country of ref document: EP

Kind code of ref document: A1