WO2020078053A1 - 医疗数据异常检测方法、装置、设备及存储介质 - Google Patents

医疗数据异常检测方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2020078053A1
WO2020078053A1 PCT/CN2019/095406 CN2019095406W WO2020078053A1 WO 2020078053 A1 WO2020078053 A1 WO 2020078053A1 CN 2019095406 W CN2019095406 W CN 2019095406W WO 2020078053 A1 WO2020078053 A1 WO 2020078053A1
Authority
WO
WIPO (PCT)
Prior art keywords
medical data
detected
preset
data
historical
Prior art date
Application number
PCT/CN2019/095406
Other languages
English (en)
French (fr)
Inventor
荣絮
吴亚博
郑毅
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 WO2020078053A1 publication Critical patent/WO2020078053A1/zh

Links

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
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients

Definitions

  • the present application relates to the technical field of medical data processing, and in particular to a method, device, equipment, and storage medium for medical data abnormality detection.
  • Social medical insurance is a social insurance system established by the state and society in accordance with certain laws and regulations to provide workers within the scope of protection with basic medical needs protection when they are sick.
  • the social medical insurance card (referred to as the medical insurance card in this application) is a unified plan issued by the Ministry of Labor and Social Security and issued by the labor security departments to the society. It is an integrated circuit card (IC card) used in various fields of labor and social security.
  • the medical insurance card is an important voucher for insured persons to buy daily medical treatment and is a carrier of personal accounts for medical insurance. It is used to record basic information of insured persons, including the basic medical insurance premiums paid by individuals, and paid by the employer in accordance with the prescribed proportion. Basic medical insurance premiums, other funds, and interest. It is a non-financial special card for insured personnel for basic medical consumption.
  • the medical insurance card is a special card for personal account of medical insurance.
  • the personal ID card is used as the identification code to store detailed information such as the personal ID number, name, gender, payment of account money, and consumption.
  • the medical insurance card can only be used by me and can only be used for medical services, but some people rent other people ’s social security cards or the insured people use the social insurance cards to cash out, or the insured people use their own social security cards to help others pay. These abnormal behaviors have broken the rules for the use of medical insurance cards. We need to prevent such unreasonable use of medical insurance cards. Therefore, we urgently need a method for detecting abnormal medical data of medical insurance cards.
  • the main purpose of the present application is to provide an abnormal detection method of medical data, which aims to solve the technical problem that the behavior data of medical insurance card swiping cannot be detected abnormally.
  • the medical data abnormality detection method includes the following steps:
  • the medical data to be detected is marked abnormally.
  • the present application also provides a medical data abnormality detection device
  • the medical data abnormality detection device includes:
  • An obtaining module configured to obtain medical data to be detected corresponding to the abnormal detection request when an abnormal detection request is detected
  • the history query module is used to judge whether there is an effective history record of the target insured person corresponding to the medical data to be detected in the history database;
  • the abnormality determination module is used to determine whether the medical data to be tested exists based on the effective historical record of the target insured person if the valid historical record of the target insured person corresponding to the medical data to be detected exists in the history database abnormal;
  • the abnormal marking module is configured to mark the medical data to be detected abnormally if the medical data to be detected is abnormal.
  • the present application also provides a medical data abnormality detection device
  • the medical data abnormality detection device includes a processor, a memory, and medical data stored on the memory and executable by the processor Anomaly detection readable instruction, wherein when the medical data anomaly detection readable instruction is executed by the processor, the steps of the medical data anomaly detection method as described above are implemented.
  • the present application also provides a storage medium storing medical data abnormality detection readable instructions, wherein when the medical data abnormality detection readable instructions are executed by the processor, the implementation is as described above Steps of the medical data anomaly detection method.
  • FIG. 1 is a schematic structural diagram of a medical data abnormality detection device in a hardware operating environment involved in an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a first embodiment of a medical data abnormality detection method of the application
  • FIG. 3 is a schematic flowchart of a second embodiment of the medical data abnormality detection method of the present application.
  • FIG. 4 is a schematic flowchart of a third embodiment of the medical data abnormality detection method of the present application.
  • FIG. 5 is a schematic diagram of functional modules of the first embodiment of the medical data abnormality detection device of the present application.
  • FIG. 1 is a schematic diagram of the hardware structure of the medical data abnormality detection device provided by the present application.
  • the medical data anomaly detection device may be a PC, or a device with display function such as a smart phone, tablet computer, portable computer, desktop computer, etc.
  • the medical data anomaly detection device may be a server device, where medical data exists An anomaly detection back-end management system through which users manage medical data anomaly detection equipment.
  • the medical data abnormality detection device may include components such as a processor 10 and a memory 20.
  • the processor 10 is connected to the memory 20, the medical data abnormality detection readable instructions are stored on the memory 20, and the processor 10 can call the medical data abnormality stored in the memory 20 Detect readable instructions and implement the steps of each embodiment of the medical data abnormality detection method described below.
  • the memory 20 may be used to store software readable instructions and various data.
  • the memory 20 may mainly include a storage readable instruction area and a storage data area, where the storage readable instruction area may store an operating system, application readable instructions required by at least one function (such as medical data abnormality detection readable instructions), etc .;
  • the data area may include a database, for example, the application needs to query to obtain valid historical records.
  • the memory 20 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.
  • the processor 10 is the control center of the medical data anomaly detection device, uses various interfaces and lines to connect the various parts of the entire medical data anomaly detection device, and runs or executes the software readable instructions and / or modules stored in the memory 20, And call the data stored in the memory 20 to perform various functions and process data of the medical data abnormality detection device, thereby performing overall monitoring of the medical data abnormality detection device.
  • the processor 10 may include one or more processing units ;.
  • the structure of the medical data abnormality detection device shown in FIG. 1 does not constitute a limitation on the medical data abnormality detection device, and may include more or fewer components than the illustration, or a combination of certain components, Or different component arrangements.
  • the detection device in the following is an abbreviation of medical data abnormality detection device.
  • This application provides a medical data abnormality detection method.
  • FIG. 2 is a schematic flowchart of a first embodiment of a medical data abnormality detection method of this application.
  • the medical data abnormality detection method includes the following steps:
  • Step S10 When an abnormality detection request is detected, obtain medical data to be detected corresponding to the abnormality detection request;
  • the anomaly detection request is an instruction that triggers the start of the medical data anomaly detection readable instruction corresponding to the medical data anomaly detection method of the present application.
  • the anomaly detection request can be triggered by a preset external event / page request, such as a click operation for anomaly detection input by a user; You can also use the timed JOB (task) as the starting interface for readable instructions for medical data anomaly detection, which triggers anomaly detection of medical data at intervals, or you can perform anomaly detection immediately after receiving medical data to be tested It is also possible to perform abnormality detection after the data amount of the received medical data to be detected reaches a certain threshold.
  • the anomaly detection of medical data in the present application may be one or more process nodes in the medical data review process, and the circulation of the process nodes may automatically trigger the readable instructions for anomaly detection of medical data.
  • the medical data to be detected is the medical data pointed to by the abnormal detection request and requested to be detected.
  • Medical data includes but is not limited to: type of current consultation, identification of treatment items, time of consultation, diagnosis result, medication situation, amount of treatment, amount of medication, etc.
  • the card swiping device After the insured person swipes the medical insurance card, the card swiping device will upload the relevant medical data to the hospital device, and then the hospital device will send it to the testing device or directly upload it to the testing device. It can be directly stored in the memory of the testing device.
  • the detection device directly obtains the medical data to be detected from the local storage according to the abnormal detection request, and can also be stored in the data storage device / cloud of the medical data for the detection device to detect the abnormal medical data from the data according to the abnormal detection request The storage device / cloud gets medical data to be tested.
  • Step S20 Determine whether there is an effective history record of the target insured person corresponding to the medical data to be detected in the history database;
  • the historical database is used to store the historical medical data of the insured person, and can also store the data obtained by further processing the historical medical data of the insured person, for example, the characteristic data obtained by processing the medical data, such as: Insurer A swipes the card 12 times a year, each time the amount is 150.
  • the effective history record refers to medical data determined to be free of abnormality after abnormality detection.
  • the effective history record of the insured person is used to determine whether the new medical data of the insured person is abnormal.
  • the medical insurance number is the medical insurance account number of the insured person. It is the medical insurance identification number of each insured person. According to the medical insurance number, a unique and specific insured person can be queried. At the same time, each time the insured swipes the card, a piece of medical data is generated. In this embodiment, the medical insurance number is used as the identification of the medical data.
  • the corresponding medical insurance number to be detected is obtained based on the medical data to be detected; the historical database is queried based on the medical insurance number to be detected, and whether there is a valid historical record corresponding to the medical insurance number to be detected in the historical database , That is, the effective history record of the target insured corresponding to the medical data to be tested.
  • the insured person uses the medical insurance card for the first time, there is no history of the insured person ’s card swipe record in the history database, or the insured person ’s historical card swipe record is an abnormal medical insurance card swipe record, and the history database does not exist at this time.
  • the effective history of the insured person If there is no valid historical record of the target insured person in the historical database, the information of the target insured person can be marked out for manual review, or the corresponding medical treatment data can be obtained from the preset diagnosis and treatment specification database according to the medical data to be detected Diagnosis and treatment specifications for the comparison of abnormality judgment by the testing equipment.
  • Step S30 if there is an effective history record of the target insured person corresponding to the medical data to be detected in the history database, based on the effective history record of the target insured person, it is judged whether the medical data to be detected has abnormality;
  • the judgment indicators involved include but are not limited to: type (type of diagnosis and treatment item / drug type), time (time of diagnosis and treatment item / time of drug purchase), frequency (frequency of diagnosis and treatment item / frequency of medication), amount ( Amount of medical treatment / medicine amount).
  • the historical data corresponding to each indicator is obtained based on the analysis of the effective history records, for example, through the analysis of the effective history records of the insured person A: the insured person A is diabetes, the type of drug purchase is a, and the time of drug purchase is On the 1st to the 5th of the month, the drug purchase frequency is 8 times a month and the amount is 200 each time.
  • the abnormality of different indicators corresponds to different abnormality types, such as abnormality of type, abnormality of time, abnormality of frequency, abnormality of amount, which can distinguish abnormality detection medical data identifying different types of abnormality.
  • Some chronic patients' medical data such as the type of treatment, identification of treatment items, time of treatment, time of medication, amount of money, etc. have certain rules, and it is not easy to break this rule.
  • This rule is related to the type of disease and the specific situation of the patient. If this rule is broken, it is very likely that the medical insurance card will be abnormally used.
  • a historical database is established in advance, which is used to store all the medical insurance card swiping data of the insured person, and the acquired medical data is standardized in the data and stored in the historical database with a preset data structure.
  • data extraction is performed on each piece of medical data to obtain historical data corresponding to the preset index, and the preset index and its corresponding data are stored in association.
  • the detection device calculates the corresponding average value for each preset index, the structure can be directly The acquired data is quickly acquired and calculated, and the calculated average value is used for abnormal detection of the medical data to be detected by the testing equipment.
  • the step S30 includes:
  • Step S301 If there is an effective history record of the target insured person corresponding to the medical data to be detected in the history database, obtain the number of effective history records of the target insured person;
  • the effective historical record of the target insured person here refers to any piece of effective historical medical data of the target insured person.
  • the number of effective historical records is the number of effective historical records.
  • the acquired effective historical medical data is subjected to data standardization processing and stored in the historical database with a preset data structure.
  • the number of effective historical records is counted based on the preset data structure and output, It can improve the efficiency of obtaining the number of effective historical records.
  • Step S302 if the number of valid historical records is greater than a preset threshold, then based on the valid historical records of the target insured, it is determined whether the medical data to be tested is abnormal;
  • the preset threshold can be temporarily set by the user, or the user can set the system default value in advance. If the number of effective historical records is greater than the preset threshold, the number of effective historical records is sufficient to support the mining of the medical rules of the target insured person, and then directly based on the effective historical records of the target insured person, determine whether the medical data to be tested exists abnormal.
  • Step S303 If the number of valid historical records is less than a preset threshold, query the diagnosis and treatment specification database according to the medical data to be detected to obtain the diagnosis and treatment specification corresponding to the medical data to be detected; based on the diagnosis and treatment criterion, determine the to be detected Is there any abnormality in the medical data?
  • step S40 if the medical data to be detected has abnormality, the medical data to be detected is marked abnormally.
  • the medical data to be detected is marked abnormally, and the abnormality detection result is output to prompt the user to perform offline verification. If there is no abnormality in the medical data to be detected, the historical database is updated, and the medical data to be detected is added to the historical database as a valid historical record.
  • the medical to be detected when the medical to be detected is detected When the data is abnormal, determine whether the corresponding target insured person has the latest diagnosis result, that is: if the medical data to be tested is abnormal, query to determine whether the target insured person's diagnosis result is updated; if the target insured person's diagnosis result If there is an update, the latest diagnosis result of the target insured person will be obtained; query the diagnosis and treatment specification database according to the latest diagnosis result, obtain the normal medical specification corresponding to the latest diagnosis result, compare the medical data to be tested with the normal medical specification, and judge the pending It is detected whether the medical data is abnormal; if the medical data to be detected is abnormal, the medical data to be detected is marked abnormally.
  • the detection device and / or the user selects a corresponding processing method for processing the medical data to be detected
  • the Anomaly type if the medical data to be detected is abnormal, the Anomaly type; according to the anomaly type, the corresponding anomaly identification is obtained and the medical data to be detected is anomaly marked.
  • the detection status of the medical data to be detected is changed to normal, and the medical data to be detected is added to the history database as a valid historical record.
  • the medical data to be detected corresponding to the abnormality detection request is obtained; it is determined whether there is an effective historical record of the target participant corresponding to the medical data to be detected in the historical database; if the historical database Where there is an effective history record of the target insured person corresponding to the medical data to be tested, based on the effective historical record of the target insured person, it is determined whether the medical data to be tested is abnormal; if the medical data to be tested If there is an abnormality, the medical data to be tested is abnormally marked, so that the detection device performs abnormal detection on the medical data of the insured person according to the medical habits of each insured person.
  • the medical data to be detected may be in violation of regulations, and users need to verify it offline, which can effectively detect abnormal card swiping behavior and regulate the use of medical insurance cards.
  • the method includes:
  • Step S50 if there is no valid historical record of the target insured person corresponding to the medical data to be tested in the history database, query the diagnosis and treatment specification database according to the medical data to be detected to obtain the diagnosis and treatment specification corresponding to the medical data to be tested ;
  • diagnosis and treatment specification database in advance, store disease types and corresponding diagnosis and treatment specification samples in the diagnosis and treatment specification database, and regularly collect samples of new diseases and diagnosis and treatment specifications.
  • diagnosis and treatment specification database also stores abnormal diagnosis and treatment specifications.
  • the medical data queries the diagnosis and treatment specification database to determine whether the medical data to be tested conforms to the diagnosis and treatment specification or the abnormal diagnosis and treatment specification. If it conforms to the abnormal diagnosis and treatment specification, it is directly determined that the medical data to be detected is abnormal.
  • Obtaining the diagnosis and treatment specifications corresponding to the medical data to be tested may specifically include: extracting the diagnosis result in the medical data to be tested, the diagnosis result includes a main diagnosis (disease type) and a subdiagnosis (subdivision type, including complications, etc.) ); Query the diagnosis and treatment specification database according to the diagnosis result to obtain the diagnosis and treatment specification corresponding to the diagnosis result.
  • the diagnosis result may be a code or other identification symbol.
  • Step S51 based on the diagnosis and treatment specification, determine whether the medical data to be tested has abnormality
  • diagnosis and treatment specifications corresponding to a diagnosis result including the type of medicine to be used, diagnosis and treatment items, frequency of diagnosis / purchase of medicine, diagnosis and treatment items and amount of medicine, etc.
  • Diagnosis / drug purchase frequency, diagnosis and treatment items and drug amount, etc. are compared with the corresponding types of drugs, diagnosis and treatment items, diagnosis / treatment drug purchase frequency, diagnosis and treatment items and drug amount, etc., if there is a difference, it is determined that the medical data to be tested exists abnormal.
  • step S52 if the medical data to be detected has abnormality, the medical data to be detected is marked abnormally.
  • step S52 is the same as step S40, in FIG. 3, step S40 is also step S52.
  • the medical data to be detected is abnormally marked, and optionally, an abnormality detection result is output to notify the user to perform offline verification.
  • the detection status of the insured person's medical data to be tested is changed to normal, and the insured person's medical data to be tested is added to the historical database as a valid historical record in.
  • the offline personnel are verifying In the actual situation, if there is no abnormal card swiping, the detection result can be corrected.
  • the normal medical data to be tested is taken as an effective historical record.
  • the diagnosis and treatment specification database is queried according to the medical data to be detected, and the diagnosis and treatment specification corresponding to the medical data to be detected is obtained. Based on the diagnosis and treatment criterion, the judgment is determined Whether the medical data to be detected is abnormal, so as to realize the abnormal detection of the cardholder's first card swipe data.
  • step S30 based on the effective history record of the target insured, it is determined whether the medical data to be detected has abnormality.
  • the steps include:
  • Step S31 Obtain a preset first analysis index, and obtain historical data and an error threshold corresponding to each preset first analysis index according to the effective history record of the target insured person;
  • the medical data to be tested is compared with the valid historical records / diagnosis and treatment specifications to preset the first analysis index, so as to realize abnormal judgment of the medical data to be tested.
  • the preset first analysis index may include: type (diagnosis and treatment item type / drug type), time (diagnosis and treatment item time / drug purchase time), frequency (diagnosis and treatment item frequency / medication frequency), amount (diagnosis and treatment item amount / drug amount)
  • the preset first analysis index may also include diagnosis results, etc. This embodiment does not limit the preset first analysis index.
  • the preset first analysis index may be set by the system by default, or may be set by the user before anomaly detection
  • the preset first analysis index can be directly stored in the preset address, and can be obtained directly from the preset address.
  • each preset first analysis indicator for example, the type of medicine is aspirin, the time of drug purchase is from the 1st to the 5th of the month, the frequency of medication is once every two months, and the amount of the drug is 200 once.
  • the drug Type / time of drug purchase / frequency of drug use / medicine amount are preset first analysis indicators, and “Aspirin”, “No. 1 to No. 5”, “Once every two months”, and “200” correspond to preset first analysis indicators Historical data.
  • each card swipe record will form an effective historical record.
  • the historical data corresponding to each preset first analysis index in each effective historical record may be slightly different, so all valid
  • the average value of the historical records, as historical data corresponding to each preset first analysis index in this embodiment, is used to compare with the medical data to be tested for abnormal judgment.
  • the medical data to be detected is subjected to data standardization processing, and stored in the historical database with a preset data structure, and the corresponding average value is calculated for each preset index at the detection device At the time, directly obtain and calculate the structured data directly.
  • each preset first analysis index has a corresponding error threshold. Taking into account this reasonable difference, it can reduce misjudgment and increase the flexibility of abnormal judgment.
  • Step S32 extract the current data corresponding to each preset first analysis index from the medical data to be tested, and compare the current data corresponding to each preset first analysis index with the corresponding historical data to obtain each preset first Analyze the difference corresponding to the index;
  • Step S33 Determine whether there is a preset first analysis indicator whose corresponding difference is greater than the corresponding error threshold;
  • the medical data to be tested belongs to the normal error range, that is, there is no abnormality in the medical data to be tested, and if the difference corresponding to each preset first analysis index is greater than the error threshold, the medical data to be tested does not belong to the normal error range .
  • Step S34 if there is a preset first analysis indicator whose corresponding difference value is greater than the corresponding error threshold, it is determined that the medical data to be detected has abnormality.
  • the preset first analysis index is obtained, and according to the effective history record of the target insured person, the historical data and the error threshold corresponding to each preset first analysis index are obtained; each pre-examination is extracted from the medical data to be tested
  • Set the current data corresponding to the first analysis index compare the current data corresponding to each preset first analysis index with the corresponding historical data, and obtain the difference corresponding to each preset first analysis index; determine whether there is a corresponding difference greater than A preset first analysis index corresponding to the error threshold; if there is a preset first analysis index corresponding to the difference greater than the corresponding error threshold, it is determined that the medical data to be detected is abnormal, because only one preset first analysis index corresponds to If the difference is greater than the corresponding error threshold, it is determined that the medical data to be detected is abnormal, and more suspicious data can be obtained to avoid missing abnormal data.
  • the step of determining whether the medical data to be detected is abnormal based on the effective history record of the target insured in step S30 includes:
  • Step S35 Obtain the preset second analysis index and the weight corresponding to each preset second analysis index, and obtain historical data corresponding to each preset second analysis index based on the effective history record of the target insured person;
  • the preset second analysis index may include: type (diagnosis and treatment item type / drug type), time (diagnosis and treatment item time / drug purchase time), frequency (diagnosis and treatment item frequency / medication frequency), amount (diagnosis and treatment item amount / drug amount,
  • the preset second analysis index in this embodiment may be the same as the preset first analysis index in the third embodiment.
  • the preset second analysis index is the same as the preset first analysis index, the related description and preset The first analysis index is the same as the analysis index, and the second analysis index and the first analysis index may be different.
  • Different preset second analysis indicators correspond to their respective weights.
  • the type of medication is generally fixed, and the treatment effects of the drugs used are similar.
  • the type of drug purchase is quite different, that is, the similarity is low, and it is most likely to be a non-standard card swiping. Therefore, compared with the time and frequency of drug purchase, the type of drug purchase has a greater impact on abnormal judgment.
  • This preset second analysis index corresponds to a larger weight.
  • the historical data corresponding to each preset second analysis index in this embodiment refers to a value obtained after performing integrated calculation on multiple pieces of historical medical data and used for abnormal judgment.
  • the medical data to be detected is subjected to data standardization processing and stored in the historical database with a preset data structure.
  • the detection device calculates the corresponding average value for each preset index, the structured data is quickly obtained directly And operation. For example, for the time of drug purchase, the date of drug purchase in the first piece of historical data is July 3, the date of drug purchase in the second piece is August 4, and the date of drug purchase in the third piece is September 2.
  • the average value of the drug purchase date calculated on the 3rd of each month is the historical data corresponding to the drug purchase time in this embodiment.
  • Step S36 extract the current data corresponding to each preset second analysis index from the medical data to be tested, and compare the current data corresponding to each preset second analysis index with the corresponding historical data to obtain each preset second Analyze the difference corresponding to the index;
  • Analyze the medical data to be tested obtain the current data corresponding to each preset second analysis index, compare the current data corresponding to each preset second analysis index with the corresponding historical data, in all preset second analysis index data .
  • Some are numerical values you can directly compare the values to get a numerical difference, such as: time (diagnosis and treatment items time / drug purchase time), frequency (diagnosis and treatment items frequency / medication frequency), amount (diagnosis and treatment item amount / drug amount).
  • time diagnosis and treatment items time / drug purchase time
  • frequency diagnosis and treatment items frequency / medication frequency
  • amount diagnosis and treatment item amount / drug amount
  • the numerical difference cannot be directly obtained, and the two can be calculated and obtained. Based on the similarity to obtain the corresponding difference.
  • Step S37 Calculate the weighted average error value corresponding to the medical data to be detected based on the difference and weight corresponding to each preset second analysis index;
  • the sum of the product of the difference and the weight corresponding to each preset second analysis index is the weighted average error value corresponding to the medical data to be detected.
  • Step S38 Compare the weighted average error value with a preset reasonable error threshold to determine whether the weighted average error value is greater than a reasonable error threshold;
  • different disease types correspond to different reasonable error thresholds.
  • a reasonable error threshold can be obtained by processing medical data samples using big data technology. Specifically, the average value of the medical data samples is calculated, and the error value of all medical data samples and the average is obtained, and the error threshold corresponding to each medical data sample is calculated based on the error value Calculate the average error threshold of all medical data samples based on the error threshold corresponding to each medical data sample.
  • the average error threshold is the reasonable error threshold.
  • Step S39 If the weighted average error value is greater than a reasonable error threshold, it is determined that the medical data to be detected has an abnormality.
  • weighted average error value is greater than the reasonable error threshold, it indicates that the medical data to be tested is significantly different from the normal data and is likely to be abnormal. Therefore, it is determined that the medical data to be tested is abnormal.
  • the weighted average error value corresponding to the medical data to be detected is calculated based on the difference and weight corresponding to each preset second analysis index; the weighted average error value is compared with a preset reasonable error threshold, Determine whether the weighted average error value is greater than a reasonable error threshold; if the weighted average error value is greater than a reasonable error threshold, it is determined that the medical data to be tested is abnormal, because the weighted average error value used for abnormal judgment is all preset
  • the weighted result of the difference and weight corresponding to the second analysis index only when the weighted value of all preset second analysis indexes is greater than the reasonable error threshold, can the medical data to be tested be determined as abnormal data, which can interpret the abnormal data more accurately. Reduce the amount of data confirmed offline and reduce the workload of offline audits.
  • step S36 the current data corresponding to each preset second analysis index is compared with the corresponding historical data to obtain each preset second
  • the steps to analyze the difference between the indicators include:
  • Step S40 when the preset second analysis indicator is a diagnosis item index or a drug type index, extract the current diagnosis item or current drug type from the current data, and extract the history diagnosis item or historical drug from the historical data kind;
  • time diagnosis and treatment item time / purchase time
  • frequency diagnosis and treatment item frequency / medication frequency
  • amount amount of diagnosis and treatment items / Amount of drugs
  • this preset second analysis indicator cannot be directly obtained after comparison
  • the numerical difference therefore, similarity calculation is performed on this preset second analysis index, and a corresponding difference value is obtained based on the similarity.
  • Step S41 calculating the similarity between the current diagnosis and treatment item and the historical diagnosis and treatment item or the similarity between the current drug type and the historical drug type;
  • the current diagnosis and treatment items (or types of drugs) and historical diagnosis and treatment items (or types of drugs) can be calculated first.
  • the sum of the items, in one embodiment, the first ratio is the similarity in this embodiment; in another embodiment, the current diagnosis item (or drug type) and the historical diagnosis item (or drug type) are obtained the same sum
  • the number of identical and similar diagnosis and treatment items accounts for the second proportion of the total number of diagnosis and treatment items, and the second proportion is the similarity in this embodiment.
  • Step S42 Calculate the difference corresponding to the medical treatment item index or the drug type index based on the similarity.
  • the negative correlation coefficient can be obtained by the user through calculation, or can be obtained by the system based on big data processing and analysis.
  • the preset second analysis index is a diagnosis item index or a drug type index
  • the current diagnosis item or current drug type is extracted from the current data
  • the history diagnosis item or history is extracted from the historical data Drug type
  • the solutions of the embodiments described in steps S41 and S42 can be applied to the third embodiment, wherein the preset second analysis index can be replaced with the preset first analysis index.
  • the present application also provides a medical data abnormality detection device corresponding to each step of the foregoing medical data abnormality detection method.
  • FIG. 5 is a schematic diagram of functional modules of an embodiment of a medical data abnormality detection device of the present application.
  • the medical data abnormality detection device of the present application includes:
  • the obtaining module 10 is configured to obtain medical data to be detected corresponding to the abnormal detection request when an abnormal detection request is detected;
  • the history query module 20 is used to determine whether there is an effective history record of the target insured person corresponding to the medical data to be detected in the history database;
  • the abnormality determination module 30 is configured to determine whether the medical data to be tested is based on the effective history records of the target insured person corresponding to the medical data to be detected in the history database There is an exception;
  • the abnormal marking module 40 is configured to mark the medical data to be detected abnormally if the medical data to be detected is abnormal.
  • the present application also proposes a storage medium on which computer-readable instructions are stored.
  • the storage medium may be the memory 20 in the medical data abnormality detection device of FIG. 1, or may be a ROM (Read-Only Memory, read-only memory) / RAM (Random Access Memory (random access memory) storage medium may be a non-volatile readable storage medium, and may perform the methods described in various embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • Health & Medical Sciences (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Pathology (AREA)
  • Economics (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Databases & Information Systems (AREA)
  • Biomedical Technology (AREA)
  • Development Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

一种基于大数据的医疗数据异常检测方法、装置、设备及存储介质，该方法包括：在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据（S10）；判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录（S20）；若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常（S30）；若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记（S40）。基于关系网络分析、数据标准化、数据更新等技术，实现对异常刷卡行为，可帮助规范医保卡的使用。

Description

医疗数据异常检测方法、装置、设备及存储介质
本申请要求于2018年10月19日提交中国专利局、申请号为201811226395.1、发明名称为“医疗数据异常检测方法、装置、设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及医疗数据处理技术领域，尤其涉及一种医疗数据异常检测方法、装置、设备及存储介质。
背景技术
社会医疗保险是国家和社会根据一定的法律法规，为向保障范围内的劳动者提供患病时基本医疗需求保障而建立的社会保险制度。社会医疗保险卡（本申请简称医保卡）是由劳动和社会保障部统一规划，由各地劳动保障部门面向社会发行，应用于劳动和社会保障各项业务领域的集成电路卡(IC卡)。
医保卡是参保人员日常看病购买的重要凭证，是医疗保险个人账户的载体，用于记录参保人员基本信息，包括个人缴纳的基本医疗保险费、按照规定比例划入的由用人单位缴纳的基本医疗保险费、其他资金、利息。它是参保人员进行基本医疗消费的非金融专用卡。医保卡是医疗保险个人帐户专用卡，以个人身份证为识别码，储存记载着个人身份证号码，姓名，性别以及帐户金的拨付，消费情况等详细资料信息。
随着社会医疗服务以及医保卡的普及，越来越多的人使用医保卡享受一系列相关的医疗服务。根据相关规定，医保卡只能限定本人使用以及只能用于医疗服务上，但是有些人租用他人社保卡或参保人自己使用社保卡套现，或参保人用自己的社保卡帮别人付费，这些异常行为都破坏了医保卡的使用规范，我们有必要杜绝这种不合理的医保卡使用，因此，我们急需一种医保卡异常医疗数据的检测方法。
发明内容
本申请的主要目的在于提供一种医疗数据异常检测方法，旨在解决无法检测异常刷医保卡行为数据的技术问题。
为实现上述目的，本申请提供一种医疗数据异常检测方法，所述医疗数据异常检测方法包括以下步骤：
在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
此外，为实现上述目的，本申请还提供一种医疗数据异常检测装置，所述医疗数据异常检测装置包括：
获取模块，用于在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
历史查询模块，用于判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
异常判断模块，用于若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
异常标记模块，用于若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
此外，为实现上述目的，本申请还提供一种医疗数据异常检测设备，所述医疗数据异常检测设备包括处理器、存储器、以及存储在所述存储器上并可被所述处理器执行的医疗数据异常检测可读指令，其中所述医疗数据异常检测可读指令被所述处理器执行时，实现如上述的医疗数据异常检测方法的步骤。
此外，为实现上述目的，本申请还提供一种存储介质，所述存储介质上存储有医疗数据异常检测可读指令，其中所述医疗数据异常检测可读指令被处理器执行时，实现如上述的医疗数据异常检测方法的步骤。
附图说明
图1是本申请实施例方案涉及的硬件运行环境的医疗数据异常检测设备结构示意图；
图2为本申请医疗数据异常检测方法第一实施例的流程示意图；
图3为本申请医疗数据异常检测方法第二实施例的流程示意图；
图4为本申请医疗数据异常检测方法第三实施例的流程示意图；
图5为本申请医疗数据异常检测装置第一实施例的功能模块示意图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
请参见图1，图1为本申请所提供的医疗数据异常检测设备的硬件结构示意图。医疗数据异常检测设备可以是PC，也可以是智能手机、平板电脑、便携计算机、台式计算机等具有显示功能的设备设备，可选地，所述医疗数据异常检测设备可以是服务器设备，存在医疗数据异常检测的后端管理系统，用户通过所述后端管理系统对医疗数据异常检测设备进行管理。
所述医疗数据异常检测设备可以包括：处理器10以及存储器20等部件。在所述医疗数据异常检测设备中，所述处理器10与所述存储器20连接，所述存储器20上存储有医疗数据异常检测可读指令，处理器10可以调用存储器20中存储的医疗数据异常检测可读指令，并实现如下述医疗数据异常检测方法各实施例的步骤。
所述存储器20，可用于存储软件可读指令以及各种数据。存储器20可主要包括存储可读指令区和存储数据区，其中，存储可读指令区可存储操作系统、至少一个功能所需的应用可读指令（比如医疗数据异常检测可读指令）等；存储数据区可包括数据库，例如本申请需查询获取有效历史记录等。此外，存储器20可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他易失性固态存储器件。
处理器10，是医疗数据异常检测设备的控制中心，利用各种接口和线路连接整个医疗数据异常检测设备的各个部分，通过运行或执行存储在存储器20内的软件可读指令和/或模块，以及调用存储在存储器20内的数据，执行医疗数据异常检测设备的各种功能和处理数据，从而对医疗数据异常检测设备进行整体监控。处理器10可包括一个或多个处理单元；。
本领域技术人员可以理解，图1中示出的医疗数据异常检测设备结构并不构成对医疗数据异常检测设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
基于上述硬件结构，提出本申请方法各个实施例，在下文中的检测设备为医疗数据异常检测设备的简称。
本申请提供一种医疗数据异常检测方法。
参照图2，图2为本申请医疗数据异常检测方法第一实施例的流程示意图。
本实施例中，所述医疗数据异常检测方法包括以下步骤：
步骤S10，在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
异常检测请求是触发本申请医疗数据异常检测方法对应医疗数据异常检测可读指令启动的指令，可以通过一个预设的外部事件/页面请求触发异常检测请求，如用户输入的异常检测的点击操作；也可以通过使用定时JOB（任务）的方式作为医疗数据异常检测可读指令的启动接口，每隔一段时间就触发进行医疗数据的异常检测，也可以在接收到待检测医疗数据后立即进行异常检测，也可以在接收的待检测医疗数据的数据量达到一定阈值后进行异常检测。可选地，本申请对医疗数据的异常检测，可以是医疗数据审核流程中的一个或多个流程节点，可由流程节点的流转自动触发医疗数据的异常检测可读指令。
待检测医疗数据即异常检测请求指向的、请求检测的医疗数据。医疗数据包括但不限于：当次就诊类型、诊疗项目标识、就诊时间、诊断结果、用药情况、诊疗金额、用药金额等。
参保人刷医保卡后，刷卡设备会将相关医疗数据上传到医院设备，再由医院设备发送到检测设备或直接上传到检测设备，可以直接在检测设备的存储器中存储，在进行异常检测时，检测设备直接根据异常检测请求从本地存储器中获取待检测医疗数据，也可以存储在医疗数据的数据存储设备/云端，以供检测设备在进行异常医疗数据的检测时，根据异常检测请求从数据存储设备/云端获取待检测医疗数据。
步骤S20，判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
历史数据库用于存储参保人的历史医疗数据，还可存储将参保人的历史医疗数据进行进一步处理后得到的数据，例如，通过对医疗数据进行数据处理后获得的特征数据，如：参保人A一年刷卡12次，每次金额150。
有效历史记录指经过异常检测后，确定无异常的医疗数据，本实施例中，参保人的有效历史记录用于判断该参保人新的医疗数据是否存在异常。
每个参保人都有专属于自己的医保号，医保号是参保人的医疗保险账号，是每一个参保人的医保标识号，根据医保号可以查询到唯一、特定的参保人，同时，参保人每刷卡一次，就产生一条医疗数据，本实施例中，以医保号作为医疗数据的标识。具体地，在获取到待检测医疗数据后，基于该待检测医疗数据获得对应的待检测医保号；基于待检测医保号查询历史数据库，判断历史数据库中是否存在待检测医保号对应的有效历史记录，即待检测医疗数据对应的目标参保人的有效历史记录。
在参保人首次使用医保卡时，历史数据库中还没有该参保人的历史刷卡记录，或者参保人的历史刷卡记录都为异常的医保刷卡记录，此时的历史数据库中也不存在该参保人的有效历史记录。若是历史数据库中不存在目标参保人的有效历史记录，则可将目标参保人的信息标出，以供人工审核，也可以根据待检测医疗数据从预置的诊疗规范数据库中获取对应的诊疗规范，以供检测设备自行对比进行异常判断。
步骤S30，若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
待检测医疗数据的异常判断，涉及的判断指标包括但不限于：种类（诊疗项目种类/药物种类）、时间（诊疗项目时间/购药时间）、频次（诊疗项目频次/用药频次）、金额（诊疗项目金额/药物金额）。
一实施例中，基于有效历史记录分析获得各指标对应的历史数据，例如，通过对参保人A的有效历史记录分析获得：参保人A为糖尿病，购药种类为a，购药时间在月初的1-5号，购药频次为8次每月，金额为每次200。将待检测医疗数据中的种类、时间、频次、金额等指标对应的数据与正常的历史数据进行比较，判断是否存在某种或多种异常。其中，不同指标的异常对应不同异常类型，如种类异常、时间异常、频次异常、金额异常，可区分标识不同异常类型的异常检测医疗数据。
有些慢性病人的就诊类型、诊疗项目标识、就诊时间、用药时间、金额等医疗数据具有一定规律，且轻易不会打破这种规律，这种规律与病种相关，也与病人的具体情况相关，若这种规律被打破，极有可能出现了医保卡异常使用的情况。
为获取参保人医疗数据的规律，预先建立历史数据库，用于存储参保人所有的医保刷卡数据，将获取的医疗数据进行数据标准化处理后，以预设的数据结构存储在历史数据库中，可选地，对每条医疗数据进行数据提取获得预设指标对应的历史数据，将预设指标与其对应的数据关联存储，在检测设备对各预设指标计算对应的均值时，可直接对结构化的数据进行快速获取及运算，计算获得的均值用于供检测设备对待检测医疗数据进行异常检测。
可选地，为提升异常判断的准确性，一实施例中，只有在有效历史记录的数量达到一定阈值时，才将其作为异常判断的标准，即提高样本数量，以提高规律挖掘的准确性，进而提升异常判断的准确性。具体地，所述步骤S30包括：
步骤S301，若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，获取所述目标参保人的有效历史记录数量；
这里的目标参保人的有效历史记录，指目标参保人的任意一条有效历史医疗数据，有效历史记录数量即有效历史记录的条数，在将有效的历史医疗数据存储到历史数据库之前，对获取的有效历史医疗数据进行数据标准化处理，以预设的数据结构存储在历史数据库中，在检测到获取有效历史记录数量的请求时，基于预设的数据结构统计有效历史记录数量，并输出，可提升获取有效历史记录数量的效率。
步骤S302，若所述有效历史记录数量大于预设阈值，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
预设阈值可以由用户临时自定义设置，也可以由用户预先设置系统默认值。若有效历史记录数量大于预设阈值，则有效历史记录的数量足以支撑起目标参保人的医疗规律的挖掘，则直接基于目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常。
步骤S303，若所述有效历史记录数量小于预设阈值，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；基于该诊疗规范，判断所述待检测医疗数据是否存在异常。
若有效历史记录数量小于预设阈值，说明有效历史记录样本数量还不够，由此得出的医疗规律可能存在一定的偶然性，为避免这种情况，获得更为准确的医疗规律，基于诊疗规范数据库中的诊疗规范判断所述待检测医疗数据是否存在异常。
步骤S40，若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
若判定待检测医疗数据存在异常，则将待检测医疗数据进行异常标记，并输出异常检测结果，以提示用户进行线下核实。若待检测医疗数据不存在异常，则对历史数据库进行数据更新，将待检测医疗数据作为有效历史记录添加到历史数据库。
因为参保人病情发生变化会导致其医疗数据发生变化，为了避免检测设备将这种情况下的正常变化标记为异常，减少线下审核的工作量，一实施例中，在检测到待检测医疗数据存在异常时，判断对应的目标参保人是否有最新的诊断结果，即：若待检测医疗数据存在异常，查询判断目标参保人的诊断结果是否有更新；若目标参保人的诊断结果有更新，则获取目标参保人的最新诊断结果；根据该最新诊断结果查询诊疗规范数据库，获取该最新诊断结果对应的正常医疗规范，将待检测医疗数据与该正常医疗规范进行对比，判断待检测医疗数据是否存在异常；若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
进一步地，为更好地进行异常标记，以便检测设备和/或用户对待检测医疗数据选择对应的处理方式进行处理，另一实施例中，若待检测医疗数据存在异常，获取待检测医疗数据的异常类型；根据该异常类型获取对应的异常标识对待检测医疗数据进行异常标记。
进一步地，在检测到用户线下核实后输入的修正指令时，将待检测医疗数据的检测状态修改为正常，并将待检测医疗数据作为有效历史记录添加到历史数据库中。
本实施例通过在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记，使得检测设备根据每个参保人的医疗习惯对参保人待检测医疗数据进行异常检测，当待检测医疗数据与参保人的医疗习惯发生偏离时，待检测医疗数据可能存在违规情况，需要用户进一步线下核实，可有效检测出异常刷卡行为，规范医保卡的使用。
进一步地，参照图3，在本申请医疗数据异常检测方法第二实施例中，所述步骤S20之后包括：
步骤S50，若历史数据库中不存在所述待检测医疗数据对应的目标参保人的有效历史记录，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；
在历史数据库中不存在目标参保人的有效历史记录时，基于诊疗规范对待检测医疗数据进行异常判断。
预先建立诊疗规范数据库，将病种及对应诊疗规范样本存入诊疗规范数据库，并定时收集新的病种及诊疗规范的样本，可选地，诊疗规范数据库还存储有异常诊疗规范，根据待检测医疗数据查询诊疗规范数据库，判断该待检测医疗数据是否符合诊疗规范或异常诊疗规范，若是符合异常诊疗规范，则直接判定待检测医疗数据存在异常。
获取所述待检测医疗数据对应的诊疗规范，具体可包括：提取待检测医疗数据中的诊断结果，所述诊断结果包括主诊断（病种）和副诊断（细分病种，包括并发症等）；根据诊断结果查询所述诊疗规范数据库，获取该诊断结果对应的诊疗规范。
可选地，所述诊断结果可以为一个编码或其他标识符号。
步骤S51，基于该诊疗规范，判断所述待检测医疗数据是否存在异常；
一个诊断结果（或称病种）对应的诊疗规范，包括要用到的药物种类、诊疗项目、诊疗/购药频次、诊疗项目与药物金额等，将待检测医疗数据中药物种类、诊疗项目、诊疗/购药频次、诊疗项目与药物金额等与诊疗规范中对应的药物种类、诊疗项目、诊疗/购药频次、诊疗项目与药物金额等进行比较，若有区别，则判定待检测医疗数据存在异常。
步骤S52，若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
应注意的是，因步骤S52与步骤S40相同，在图3中，步骤S40也为步骤S52。
如果待检测医疗数据存在异常，则将待检测医疗数据进行异常标记，可选地，输出异常检测结果，以通知用户进行线下核实。在检测到用户线下核实后输入的修正指令时，将该参保人的待检测医疗数据的检测状态改为正常，并将该参保人的待检测医疗数据作为有效历史记录添加到历史数据库中。可能存在个别病人确实有特殊情况，其需要违反通用正常的诊疗规范进行治疗，或者因其他特殊原因导致医保刷卡异常，或者因未收集的诊断结果导致的诊疗方案的改变，线下人员在核实了实际情况，认为不存在异常刷卡时，可修正检测结果。
若所述待检测医疗数据不存在异常，则将正常的待检测医疗数据作为有效历史记录。
本实施例通过在历史数据库中不存在目标参保人的有效历史记录时，根据待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范，基于该诊疗规范，判断所述待检测医疗数据是否存在异常，以实现对参保人首次刷卡数据的异常检测。
进一步地，如图4，在本申请医疗数据异常检测方法第三实施例中，所述步骤S30中基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常这一步骤包括：
步骤S31，获取预设第一分析指标，根据所述目标参保人的有效历史记录，获得各预设第一分析指标对应的历史数据和误差阈值；
将待检测医疗数据与有效历史记录/诊疗规范进行预设第一分析指标的对比，以实现对待检测医疗数据的异常判断。预设第一分析指标，可包括：种类（诊疗项目种类/药物种类）、时间（诊疗项目时间/购药时间）、频次（诊疗项目频次/用药频次）、金额（诊疗项目金额/药物金额），预设第一分析指标还可包括诊断结果等，本实施例不对预设第一分析指标进行限制，预设第一分析指标可以由系统默认设置，也可以在异常检测前由用户自定义设置，预设第一分析指标可直接存储于预设地址，直接从该预设地址获取即可。
各预设第一分析指标对应的历史数据，举例而言，药物种类为阿司匹林，购药时间为月份的一号到五号，用药频次为每俩月一次，药物金额为200一次，其中，药物种类/购药时间/用药频次/药物金额为预设第一分析指标，“阿司匹林”、“一号到五号”、“每俩月一次”、“200”为各预设第一分析指标对应的历史数据。
因为一个参保人可能有多次历史刷卡记录，每一次刷卡记录都会形成一条有效历史记录，每条有效历史记录中各预设第一分析指标对应的历史数据可能有些区别，则可获取所有有效历史记录的平均值，作为本实施例中各预设第一分析指标对应的历史数据，用于与待检测医疗数据对比，以进行异常判断。可选地，在判定待检测医疗数据不存在异常时，将待检测医疗数据进行数据标准化处理后，以预设的数据结构存储在历史数据库中，在检测设备对各预设指标计算对应的均值时，直接对结构化的数据进行快速获取及运算。
因患同一病种的不同病人实际情况不同，所需的诊疗项目、就诊时间、用药金额与频率等也有细微不同，例如，病种A，吃的B种类药物的频次为10±3次/月，即患A病的人，吃B种类药物的频次在10±3次/月，都是正常的。因此，各预设第一分析指标具有对应的误差阈值，因考虑到这种合理的差异性，可减少误判，提高异常判断的灵活性。
步骤S32，从所述待检测医疗数据中提取各预设第一分析指标对应的当前数据，将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第一分析指标对应的差值；
分析待检测医疗数据，获得各预设第一分析指标对应的当前数据，将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，在所有预设第一分析指标的数据中，有些是数值，可以直接进行数值比较得到一个数值差，如：时间（诊疗项目时间/购药时间）、频次（诊疗项目频次/用药频次）、金额（诊疗项目金额/药物金额）。在所有预设第一分析指标的数据中，还有些不是数值，如种类（诊疗项目种类/药物种类），这种预设第一分析指标对比后无法直接得到数值差，则可计算获得二者的相似度，基于相似度获得对应的差值。
步骤S33，判断是否存在对应差值大于对应误差阈值的预设第一分析指标；
将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第一分析指标对应的差值，若是各预设第一分析指标对应的差值小于所述误差阈值，则待检测医疗数据属于正常的误差范围，即待检测医疗数据不存在异常，若是各预设第一分析指标对应的差值大于所述误差阈值，则待检测医疗数据不属于正常的误差范围。
步骤S34，若存在对应差值大于对应误差阈值的预设第一分析指标，则判定所述待检测医疗数据存在异常。
待检测医疗数据中各预设第一分析指标对应的当前数据中，只要一个预设第一分析指标对应的差值大于对应误差阈值，就说明待检测医疗数据存在异常。
本实施例获取预设第一分析指标，根据所述目标参保人的有效历史记录，获得各预设第一分析指标对应的历史数据和误差阈值；从所述待检测医疗数据中提取各预设第一分析指标对应的当前数据，将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第一分析指标对应的差值；判断是否存在对应差值大于对应误差阈值的预设第一分析指标；若存在对应差值大于对应误差阈值的预设第一分析指标，则判定所述待检测医疗数据存在异常，因只要一个预设第一分析指标对应的差值大于对应误差阈值，就判定待检测医疗数据存在异常，可获取较多的可疑数据，避免漏掉异常数据。
进一步地，在本申请医疗数据异常检测方法第四实施例中，所述步骤S30中基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常这一步骤包括：
步骤S35，获取预设第二分析指标以及各预设第二分析指标对应的权重，并基于所述目标参保人的有效历史记录，得到各预设第二分析指标对应的历史数据；
预设第二分析指标，可包括：种类（诊疗项目种类/药物种类）、时间（诊疗项目时间/购药时间）、频次（诊疗项目频次/用药频次）、金额（诊疗项目金额/药物金额，本实施例中的预设第二分析指标与第三实施例中的预设第一分析指标可以相同，在预设第二分析指标与预设第一分析指标相同时，相关描述与预设第一分析指标相同，不赘述，预设第二分析指标与预设第一分析指标也可以不相同。
不同预设第二分析指标，对应各自的权重。待检测医疗数据中有多个预设第二分析指标的数据，对于异常判断具有不同程度的影响，例如，对于特定病种，其用药类型一般较为固定，所用药物的治疗效果具有相似性，若是购药类型差异较大，即相似性较低，则极有可能是不规范地刷卡，因此，相较于购药时间、购药频次，购药类型对于异常判断的影响较大，购药类型这一预设第二分析指标对应的权重较大。
本实施例中的各预设第二分析指标对应的历史数据，指对多条历史医疗数据进行整合计算后获得的、用于进行异常判断的值，可选地，在判定待检测医疗数据不存在异常时，将待检测医疗数据进行数据标准化处理后，以预设的数据结构存储在历史数据库中，在检测设备对各预设指标计算对应的均值时，直接对结构化的数据进行快速获取及运算。例如，对于购药时间，第一条历史数据中的购药日期为7月3日，第二条的购药日期为8月4日，第三条的购药日期为9月2日，则计算获得购药日期的均值为每月的3号，这个均值即本实施例中的购药时间对应的历史数据。
步骤S36，从所述待检测医疗数据中提取各预设第二分析指标对应的当前数据，将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第二分析指标对应的差值；
分析待检测医疗数据，获得各预设第二分析指标对应的当前数据，将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，在所有预设第二分析指标的数据中，有些是数值，可以直接进行数值比较得到一个数值差，如：时间（诊疗项目时间/购药时间）、频次（诊疗项目频次/用药频次）、金额（诊疗项目金额/药物金额）。在所有预设第二分析指标的数据中，还有些不是数值，如种类（诊疗项目种类/药物种类），这种预设第二分析指标对比后无法直接得到数值差，则可计算获得二者的相似度，基于相似度获得对应的差值。
步骤S37，基于各预设第二分析指标对应的差值和权重，计算所述待检测医疗数据对应的加权平均误差值；
各预设第二分析指标对应的差值和权重的乘积之和，即为待检测医疗数据对应的加权平均误差值。
步骤S38，将所述加权平均误差值与预置的合理误差阈值进行对比，判断所述加权平均误差值是否大于合理误差阈值；
一实施例中，不同病种对应不同的合理误差阈值。合理误差阈值，可以使用大数据技术处理医疗数据样本获得，具体地，计算医疗数据样本的均值，并获取所有医疗数据样本与均值的误差值，基于该误差值计算各医疗数据样本对应的误差阈值，基于各医疗数据样本对应的误差阈值计算所有医疗数据样本的平均误差阈值，该平均误差阈值即为合理误差阈值。
步骤S39，若所述加权平均误差值大于合理误差阈值，则判定所述待检测医疗数据存在异常。
若加权平均误差值大于合理误差阈值，说明待检测医疗数据与通常数据存在较大区别，很可能存在异常，因此判定待检测医疗数据存在异常。
本实施例通过基于各预设第二分析指标对应的差值和权重，计算所述待检测医疗数据对应的加权平均误差值；将所述加权平均误差值与预置的合理误差阈值进行对比，判断所述加权平均误差值是否大于合理误差阈值；若所述加权平均误差值大于合理误差阈值，则判定所述待检测医疗数据存在异常，因为用于异常判断的加权平均误差值是所有预设第二分析指标对应的差值和权重的加权结果，只有所有预设第二分析指标的加权值大于合理误差阈值，才确定待检测医疗数据为异常数据，可更为精准地判读异常数据，可减少线下确认的数据数量，降低线下审核工作量。
进一步地，在本申请医疗数据异常检测方法第五实施例中，所述步骤S36中所述将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第二分析指标对应的差值的步骤包括：
步骤S40，在所述预设第二分析指标为诊疗项目指标或药物种类指标时，从所述当前数据中提取当前诊疗项目或当前药物种类，从所述历史数据中提取历史诊疗项目或历史药物种类；
在所有预设第二分析指标的数据中，有些是数值，可以直接进行数值比较得到一个数值差，如：时间（诊疗项目时间/购药时间）、频次（诊疗项目频次/用药频次）、金额（诊疗项目金额/药物金额）；在所有预设第二分析指标的数据中，还有些不是数值，如种类（诊疗项目种类/药物种类），这种预设第二分析指标对比后无法直接得到数值差，因此，对这种预设第二分析指标进行相似度计算，基于相似度获得对应的差值。
步骤S41，计算所述当前诊疗项目与历史诊疗项目的相似度或当前药物种类与历史药物种类的相似度；
可先计算当前诊疗项目（或药物种类）与历史诊疗项目（或药物种类）中，相同诊疗项目的数量占诊疗项目总数量的第一比例，这里的诊疗项目总数量指当前诊疗项目与历史诊疗项目之和，在一实施例中，第一比例即该实施例中的相似度；在另一实施例中，获取当前诊疗项目（或药物种类）与历史诊疗项目（或药物种类）中相同和相似的诊疗项目，计算相同和相似的诊疗项目的数量占诊疗项目总数量的第二比例，第二比例即该实施例中的相似度。
步骤S42，基于所述相似度计算所述诊疗项目指标或药物种类指标对应的差值。
相似度越高，差值越小，即：相似度与差值呈负相关关系，负相关系数可以由用户通过计算获得，也可以由系统基于大数据处理分析获得。
本实施例通过在所述预设第二分析指标为诊疗项目指标或药物种类指标时，从所述当前数据中提取当前诊疗项目或当前药物种类，从所述历史数据中提取历史诊疗项目或历史药物种类；计算所述当前诊疗项目与历史诊疗项目的相似度或当前药物种类与历史药物种类的相似度；基于所述相似度计算所述诊疗项目指标或药物种类指标对应的差值，可对无法直接进行数值计算获得数值差的预设第二分析指标进行相似度计算，进而保证后续顺利计算获得加权平均误差值或诊疗项目指标或药物种类指标对应的差值，以便进行异常判断。
可选地，步骤S41、S42所述实施例方案可适用于第三实施例，其中的预设第二分析指标可替换为预设第一分析指标。
此外，本申请还提供一种与上述医疗数据异常检测方法各步骤对应的医疗数据异常检测装置。
参照图5，图5为本申请医疗数据异常检测装置一实施例的功能模块示意图。在本实施例中，本申请医疗数据异常检测装置包括：
获取模块10，用于在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
历史查询模块20，用于判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
异常判断模块30，用于若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
异常标记模块40，用于若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
本申请还提出一种存储介质，其上存储有计算机可读指令。所述存储介质可以是图1的医疗数据异常检测设备中的存储器20，也可以是如ROM（Read-Only Memory，只读存储器）/RAM（Random Access Memory，随机存取存储器）存储介质可以为非易失性可读存储介质，可以执行本申请各个实施例所述的方法。

Claims (20)

  1. 一种医疗数据异常检测方法，其特征在于，所述医疗数据异常检测方法包括以下步骤：
    在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
    判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
    若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
    若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
  2. 如权利要求1所述的医疗数据异常检测方法，其特征在于，所述判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录的步骤之后包括：
    若历史数据库中不存在所述待检测医疗数据对应的目标参保人的有效历史记录，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；
    基于该诊疗规范，判断所述待检测医疗数据是否存在异常；
    若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
  3. 如权利要求1所述的医疗数据异常检测方法，其特征在于，所述判断所述待检测医疗数据是否存在异常的步骤之后包括：
    若所述待检测医疗数据不存在异常，则将所述待检测医疗数据作为有效历史记录添加到历史数据库。
  4. 如权利要求1所述的医疗数据异常检测方法，其特征在于，所述基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常的步骤包括：
    获取预设第一分析指标，根据所述目标参保人的有效历史记录，获得各预设第一分析指标对应的历史数据和误差阈值；
    从所述待检测医疗数据中提取各预设第一分析指标对应的当前数据，将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第一分析指标对应的差值；
    判断是否存在对应差值大于对应误差阈值的预设第一分析指标；
    若存在对应差值大于对应误差阈值的预设第一分析指标，则判定所述待检测医疗数据存在异常。
  5. 如权利要求1所述的医疗数据异常检测方法，其特征在于，所述基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常的步骤包括：
    获取预设第二分析指标以及各预设第二分析指标对应的权重，并基于所述目标参保人的有效历史记录，得到各预设第二分析指标对应的历史数据；
    从所述待检测医疗数据中提取各预设第二分析指标对应的当前数据，将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第二分析指标对应的差值；
    基于各预设第二分析指标对应的差值和权重，计算所述待检测医疗数据对应的加权平均误差值；
    将所述加权平均误差值与预置的合理误差阈值进行对比，判断所述加权平均误差值是否大于合理误差阈值；
    若所述加权平均误差值大于合理误差阈值，则判定所述待检测医疗数据存在异常。
  6. 如权利要求5所述的医疗数据异常检测方法，其特征在于，所述将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第二分析指标对应的差值的步骤包括：
    在所述预设第二分析指标为诊疗项目指标或药物种类指标时，从所述当前数据中提取当前诊疗项目或当前药物种类，从所述历史数据中提取历史诊疗项目或历史药物种类；
    计算所述当前诊疗项目与历史诊疗项目的相似度或当前药物种类与历史药物种类的相似度；
    基于所述相似度计算所述诊疗项目指标或药物种类指标对应的差值。
  7. 如权利要求1所述的医疗数据异常检测方法，其特征在于，所述若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常的步骤包括：
    若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，获取所述目标参保人的有效历史记录数量；
    若所述有效历史记录数量大于预设阈值，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
    若所述有效历史记录数量小于预设阈值，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；基于该诊疗规范，判断所述待检测医疗数据是否存在异常。
  8. 一种医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置包括：
    获取模块，用于在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
    历史查询模块，用于判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
    异常判断模块，用于若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
    异常标记模块，用于若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
  9. 如权利要求8所述的医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置还包括：
    规范获取模块，用于若历史数据库中不存在所述待检测医疗数据对应的目标参保人的有效历史记录，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；
    第一异常判断模块，用于基于该诊疗规范，判断所述待检测医疗数据是否存在异常；
    所述异常标记模块，还用于若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
  10. 如权利要求8所述的医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置还包括：
    数据更新模块，用于若所述待检测医疗数据不存在异常，则将所述待检测医疗数据作为有效历史记录添加到历史数据库。
  11. 如权利要求8所述的医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置还包括：
    第一获取模块，用于获取预设第一分析指标，根据所述目标参保人的有效历史记录，获得各预设第一分析指标对应的历史数据和误差阈值；
    第二获取模块，用于从所述待检测医疗数据中提取各预设第一分析指标对应的当前数据，将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第一分析指标对应的差值；
    第一判断模块，用于判断是否存在对应差值大于对应误差阈值的预设第一分析指标；
    第二异常判断模块，用于若存在对应差值大于对应误差阈值的预设第一分析指标，则判定所述待检测医疗数据存在异常。
  12. 如权利要求8所述的医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置还包括：
    第三获取模块，用于获取预设第二分析指标以及各预设第二分析指标对应的权重，并基于所述目标参保人的有效历史记录，得到各预设第二分析指标对应的历史数据；
    第四获取模块，用于从所述待检测医疗数据中提取各预设第二分析指标对应的当前数据，将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第二分析指标对应的差值；
    第一计算模块，用于基于各预设第二分析指标对应的差值和权重，计算所述待检测医疗数据对应的加权平均误差值；
    第二判断模块，用于将所述加权平均误差值与预置的合理误差阈值进行对比，判断所述加权平均误差值是否大于合理误差阈值；
    第三异常判断模块，用于若所述加权平均误差值大于合理误差阈值，则判定所述待检测医疗数据存在异常。
  13. 如权利要求12所述的医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置还包括：
    提取模块，用于在所述预设第二分析指标为诊疗项目指标或药物种类指标时，从所述当前数据中提取当前诊疗项目或当前药物种类，从所述历史数据中提取历史诊疗项目或历史药物种类；
    第二计算模块，用于计算所述当前诊疗项目与历史诊疗项目的相似度或当前药物种类与历史药物种类的相似度；基于所述相似度计算所述诊疗项目指标或药物种类指标对应的差值。
  14. 如权利要求8所述的医疗数据异常检测装置，其特征在于，所述医疗数据异常检测装置还包括：
    第三判断模块，用于若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，获取所述目标参保人的有效历史记录数量；
    第四异常判断模块，用于若所述有效历史记录数量大于预设阈值，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
    第五异常判断模块，若所述有效历史记录数量小于预设阈值，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；基于该诊疗规范，判断所述待检测医疗数据是否存在异常
  15. 一种医疗数据异常检测设备，其特征在于，所述医疗数据异常检测设备包括处理器、存储器、以及存储在所述存储器上并可被所述处理器执行的医疗数据异常检测可读指令，其中所述医疗数据异常检测可读指令被所述处理器执行时，执行以下步骤：
    在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
    判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
    若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
    若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
  16. 如权利要求15所述的医疗数据异常检测设备，其特征在于，所述医疗数据异常检测可读指令被所述处理器执行时，实现以下步骤：
    若历史数据库中不存在所述待检测医疗数据对应的目标参保人的有效历史记录，则根据所述待检测医疗数据查询诊疗规范数据库，获取所述待检测医疗数据对应的诊疗规范；
    基于该诊疗规范，判断所述待检测医疗数据是否存在异常；
    若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
  17. 如权利要求15所述的医疗数据异常检测设备，其特征在于，所述医疗数据异常检测可读指令被所述处理器执行时，实现以下步骤：
    若所述待检测医疗数据不存在异常，则将所述待检测医疗数据作为有效历史记录添加到历史数据库。
  18. 如权利要求15所述的医疗数据异常检测设备，其特征在于，所述医疗数据异常检测可读指令被所述处理器执行时，所述基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常的步骤包括：
    获取预设第一分析指标，根据所述目标参保人的有效历史记录，获得各预设第一分析指标对应的历史数据和误差阈值；
    从所述待检测医疗数据中提取各预设第一分析指标对应的当前数据，将各预设第一分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第一分析指标对应的差值；
    判断是否存在对应差值大于对应误差阈值的预设第一分析指标；
    若存在对应差值大于对应误差阈值的预设第一分析指标，则判定所述待检测医疗数据存在异常。
  19. 如权利要求15所述的医疗数据异常检测设备，其特征在于，所述医疗数据异常检测可读指令被所述处理器执行时，所述基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常的步骤包括：
    获取预设第二分析指标以及各预设第二分析指标对应的权重，并基于所述目标参保人的有效历史记录，得到各预设第二分析指标对应的历史数据；
    从所述待检测医疗数据中提取各预设第二分析指标对应的当前数据，将各预设第二分析指标对应的当前数据与对应的历史数据进行对比，获得各预设第二分析指标对应的差值；
    基于各预设第二分析指标对应的差值和权重，计算所述待检测医疗数据对应的加权平均误差值；
    将所述加权平均误差值与预置的合理误差阈值进行对比，判断所述加权平均误差值是否大于合理误差阈值；
    若所述加权平均误差值大于合理误差阈值，则判定所述待检测医疗数据存在异常。
  20. 一种存储介质，其特征在于，所述存储介质上存储有医疗数据异常检测可读指令，其中所述医疗数据异常检测可读指令被处理器执行时，实现以下的步骤：
    在检测到异常检测请求时，获取该异常检测请求对应的待检测医疗数据；
    判断历史数据库中是否存在所述待检测医疗数据对应的目标参保人的有效历史记录；
    若历史数据库中存在所述待检测医疗数据对应的目标参保人的有效历史记录，则基于所述目标参保人的有效历史记录，判断所述待检测医疗数据是否存在异常；
    若所述待检测医疗数据存在异常，则将所述待检测医疗数据进行异常标记。
PCT/CN2019/095406 2018-10-19 2019-07-10 医疗数据异常检测方法、装置、设备及存储介质 WO2020078053A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811226395.1 2018-10-19
CN201811226395.1A CN109636623A (zh) 2018-10-19 2018-10-19 医疗数据异常检测方法、装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2020078053A1 true WO2020078053A1 (zh) 2020-04-23

Family

ID=66066515

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/095406 WO2020078053A1 (zh) 2018-10-19 2019-07-10 医疗数据异常检测方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN109636623A (zh)
WO (1) WO2020078053A1 (zh)

Families Citing this family (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109636623A (zh) * 2018-10-19 2019-04-16 平安医疗健康管理股份有限公司 医疗数据异常检测方法、装置、设备及存储介质
CN110459276B (zh) * 2019-08-15 2022-05-24 北京嘉和海森健康科技有限公司 一种数据处理方法及相关设备
CN110688406A (zh) * 2019-09-06 2020-01-14 平安医疗健康管理股份有限公司 数据处理方法、装置、计算机设备和存储介质
CN110781222A (zh) * 2019-10-14 2020-02-11 平安医疗健康管理股份有限公司 异常医保申请检测方法、装置、计算机设备和存储介质
CN111028089B (zh) * 2019-11-25 2023-05-02 泰康保险集团股份有限公司 异常操作识别方法及装置、计算机存储介质、电子设备
CN111538311B (zh) * 2020-04-22 2021-08-20 北京航空航天大学 一种基于数据挖掘的机械设备柔性多状态自适应预警方法及装置
CN111612647B (zh) * 2020-05-14 2023-05-23 深圳旦倍科技有限公司 计量表异常数据检测方法、装置、计量表及可读存储介质
CN112036749A (zh) * 2020-08-31 2020-12-04 平安医疗健康管理股份有限公司 基于医疗数据识别风险用户的方法、装置和计算机设备
CN112289403A (zh) * 2020-10-19 2021-01-29 浙江大学山东工业技术研究院 一种基于两阶段筛查流程的医保异常购药的检测方法及装置
CN112365987B (zh) * 2020-10-27 2023-06-06 平安科技（深圳）有限公司 诊断数据异常检测方法、装置、计算机设备及存储介质
CN112445846A (zh) * 2020-12-01 2021-03-05 平安医疗健康管理股份有限公司 医疗项目识别方法、装置、设备及计算机可读存储介质
CN113761340A (zh) * 2021-01-15 2021-12-07 北京京东拓先科技有限公司 信息推荐方法、装置、电子设备和计算机可读介质
CN113049935A (zh) * 2021-03-04 2021-06-29 长鑫存储技术有限公司 半导体智能检测系统、智能检测方法及存储介质
CN113130053A (zh) * 2021-03-24 2021-07-16 大光惠世医疗科技有限公司 一种基于用户活动数据的自助挂号方法及装置
CN113506577A (zh) * 2021-06-25 2021-10-15 贵州电网有限责任公司 一种基于增量采集电话录音完善声纹库的方法
CN114612887B (zh) * 2021-09-01 2023-01-10 腾讯科技（深圳）有限公司 单据异常检测方法、装置、设备及计算机可读存储介质
CN113780792A (zh) * 2021-09-03 2021-12-10 广州云从鼎望科技有限公司 医保违规行为监控方法、装置及计算机可读存储介质
CN114757792B (zh) * 2022-06-15 2022-09-30 南京云联数科科技有限公司 一种基于多领域数据的医保风控管理方法及设备
CN115329905B (zh) * 2022-10-13 2022-12-09 湖南鼎誉检验检测股份有限公司 一种食品检测方法、系统及存储介质
CN117951169A (zh) * 2022-10-18 2024-04-30 华为云计算技术有限公司 查询数据的方法和相关装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107133437A (zh) * 2017-03-03 2017-09-05 平安医疗健康管理股份有限公司 监控药品使用的方法及装置
CN107833039A (zh) * 2017-02-13 2018-03-23 平安医疗健康管理股份有限公司 医疗费用支付方法和系统
CN107871285A (zh) * 2017-12-06 2018-04-03 和金在线（北京）科技有限公司 一种健康保险赔欺诈及滥用的侦测方法和系统
CN108492196A (zh) * 2018-03-08 2018-09-04 平安医疗健康管理股份有限公司 通过数据分析推断医疗保险违规行为的风控方法
CN109636623A (zh) * 2018-10-19 2019-04-16 平安医疗健康管理股份有限公司 医疗数据异常检测方法、装置、设备及存储介质

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102013084A (zh) * 2010-12-14 2011-04-13 江苏大学 用于检测医疗保险门诊欺诈性交易的系统和方法
CN104182824B (zh) * 2014-08-08 2018-02-06 平安养老保险股份有限公司 一种识别医保报销违规行为的规则校验系统及校验方法
WO2018010105A1 (zh) * 2016-07-12 2018-01-18 曹庆恒 处方用药信息合理性数据特征分析系统
CN106447500A (zh) * 2016-11-09 2017-02-22 北京好运到信息科技有限公司 一种根据疾病和药物的匹配度进行打分的方法及其系统
CN107133438B (zh) * 2017-03-03 2018-07-13 平安医疗健康管理股份有限公司 医疗行为监控方法及装置
CN107451401A (zh) * 2017-07-11 2017-12-08 武汉金豆医疗数据科技有限公司 一种医保智能审核方法和系统
CN107451400A (zh) * 2017-07-11 2017-12-08 武汉金豆医疗数据科技有限公司 一种医疗行为监控方法和系统
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
CN107833039A (zh) * 2017-02-13 2018-03-23 平安医疗健康管理股份有限公司 医疗费用支付方法和系统
CN107133437A (zh) * 2017-03-03 2017-09-05 平安医疗健康管理股份有限公司 监控药品使用的方法及装置
CN107871285A (zh) * 2017-12-06 2018-04-03 和金在线（北京）科技有限公司 一种健康保险赔欺诈及滥用的侦测方法和系统
CN108492196A (zh) * 2018-03-08 2018-09-04 平安医疗健康管理股份有限公司 通过数据分析推断医疗保险违规行为的风控方法
CN109636623A (zh) * 2018-10-19 2019-04-16 平安医疗健康管理股份有限公司 医疗数据异常检测方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN109636623A (zh) 2019-04-16

Similar Documents

Publication Publication Date Title
WO2020078053A1 (zh) 医疗数据异常检测方法、装置、设备及存储介质
WO2020078058A1 (zh) 医疗数据异常识别方法、装置、终端及存储介质
WO2020114221A1 (zh) 基于生物特征识别的定点医药机构的监管方法及相关设备
WO2020119384A1 (zh) 基于大数据分析的医保异常检测方法、装置、设备和介质
WO2016165177A1 (zh) 基于面部识别的网络医院自助缴费方法和系统
WO2020119176A1 (zh) 报销数据的排查方法、识别服务端及存储介质
WO2020119403A1 (zh) 住院数据异常检测方法、装置、设备及可读存储介质
WO2020119402A1 (zh) 无关用药的识别方法、装置、终端及计算机可读存储介质
WO2020119131A1 (zh) 用药方案异常的识别方法、装置、终端及可读存储介质
WO2020226456A1 (ko) 처방 정보를 통한 의료 정보 제공 방법 및 장치
WO2016165201A1 (zh) 基于网络医院的自助缴费方法和系统
US11693990B1 (en) Medical data governance
WO2020108111A1 (zh) 医保欺诈行为的识别方法、装置、设备及可读存储介质
WO2020087981A1 (zh) 风控审核模型生成方法、装置、设备及可读存储介质
WO2020082801A1 (zh) 基于生物识别的查房考核方法、装置、服务器及介质
WO2020119118A1 (zh) 异常数据的处理方法、装置、设备及计算机可读存储介质
WO2020107899A1 (zh) 医疗费用预测方法、装置、设备及计算机可读存储介质
WO2020077837A1 (zh) 业务数据核算方法、装置、设备及计算机可读存储介质
WO2016165181A1 (zh) 基于指纹识别的网络医院自助缴费方法和系统
WO2018088585A1 (ko) 복약 관리 방법 및 그 장치
WO2016165180A1 (zh) 基于网络医院挂号号码的自助缴费方法和系统
WO2022149643A1 (ko) 키오스크를 활용한 근로 계약 체결 및 관리 방법, 장치 및 컴퓨터-판독가능 기록매체
WO2020087985A1 (zh) 赔付金额的调整方法、装置、设备及可读存储介质
CN111742372A (zh) 用于检测医疗保健数据中的指示符的学习过滤器
US20090089094A1 (en) System and method for detection of abuse of patient data

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: 19874313

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 11/06/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19874313

Country of ref document: EP

Kind code of ref document: A1