WO2020087971A1 - 基于预测模型预测住院合理性的方法及相关产品 - Google Patents

基于预测模型预测住院合理性的方法及相关产品 Download PDF

Info

Publication number
WO2020087971A1
WO2020087971A1 PCT/CN2019/095050 CN2019095050W WO2020087971A1 WO 2020087971 A1 WO2020087971 A1 WO 2020087971A1 CN 2019095050 W CN2019095050 W CN 2019095050W WO 2020087971 A1 WO2020087971 A1 WO 2020087971A1
Authority
WO
WIPO (PCT)
Prior art keywords
hospitalization
preoperative
insured
data
item set
Prior art date
Application number
PCT/CN2019/095050
Other languages
English (en)
French (fr)
Inventor
周竹凌
汪丽娟
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 WO2020087971A1 publication Critical patent/WO2020087971A1/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
    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Definitions

  • This application relates to the medical technology field, and specifically relates to a method and related products for predicting the rationality of hospitalization based on a prediction model.
  • the Medical Insurance Co-ordination Foundation reimburses the insured person for the huge medical expenses, but the current medical insurance system is still not perfect, resulting in There are many interests in the reimbursement process.
  • medical examination items are not included in the reimbursement scope of the medical insurance co-ordination fund, and the medical examination expenses of the insured can only be reimbursed personally.
  • the medical insurance pooling fund can be used to reimburse some or even all hospitalization expenses.
  • some insured persons may conspire with doctors to obtain medical insurance co-ordination funds to reimburse the insured for the expenses of the insured, that is, in the case that the insured does not need to be hospitalized, the insured will be admitted to the hospital, and some extra preoperative inspections will be added.
  • the project will increase the cost of pre-operative examinations, increase the hospitalization costs briefly, bring the hospitalization costs to the threshold, and apply for the medical insurance co-ordination fund to pay for the inspection costs.
  • the relevant personnel only check whether the preoperative inspection item is the preoperative inspection item designated by the doctor, and do not consider whether the insured person is qualified and the rationality of the preoperative inspection item.
  • the embodiments of the present application provide a method and related products for predicting the rationality of hospitalization based on a prediction model, with a view to predicting whether the pre-operative examination items of the insured person are abnormal, and determining the rationality of the insured person's hospitalization behavior.
  • an embodiment of the present application provides a method for predicting the rationality of hospitalization based on a prediction model.
  • the method is applied to an electronic device, and the method includes:
  • the actual preoperative inspection item set is compared with the preset preoperative inspection item set to determine the rationality of the insured's hospitalization behavior.
  • an embodiment of the present application provides an electronic device for predicting the rationality of hospitalization based on a prediction model, the electronic device includes:
  • a receiving unit configured to receive the hospitalization data of any insured person entered, the hospitalization data including diagnosis data and an actual preoperative inspection item set;
  • the input unit is used to extract the diagnosis data in the hospitalization data, input the diagnosis data to a pre-trained hospitalization prediction model, and output the hospitalization probability corresponding to the insured person;
  • An obtaining unit configured to obtain a preset set of preoperative examination items corresponding to the diagnostic data when the hospitalization probability is greater than a first threshold
  • the comparison unit is used to compare the actual preoperative inspection item set with the preset preoperative inspection item set to determine the rationality of the insured's hospitalization behavior.
  • an embodiment of the present application provides an electronic device, including one or more processors, one or more memories, one or more transceivers, and one or more programs, where the one or more programs are Stored in the memory and configured to be executed by the one or more processors, the program includes instructions for performing the steps in the method as described in the first aspect.
  • an embodiment of the present application provides a computer-readable storage medium for storing a computer program, where the computer program is executed by a processor to implement the method according to the first aspect.
  • the input of the hospitalization data of the insured person is received, the diagnostic data in the hospitalization data is extracted, and the preset preoperative inspection item set of the insured person is obtained according to the diagnostic data, and the preset The inspection item set is compared with the actual inspection item set in the hospitalization data to determine whether the insured person ’s preoperative inspection item set is abnormal, thereby determining the reasonableness of the insured person ’s hospitalization behavior.
  • the reasonableness of the insured person ’s hospitalization behavior can be detected, so the probability of the medical insurance fund being paid for preoperative examination can be reduced, providing data reference for the reform of the medical system and improving the medical system.
  • 1A is a schematic flowchart of a method for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application;
  • FIG. 1B is a schematic diagram of a binary data matrix determined based on a preoperative inspection item provided by an embodiment of this application;
  • FIG. 2 is a schematic flowchart of another method for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application;
  • FIG. 3 is a schematic flowchart of another method for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application;
  • 3A is a schematic diagram of a process of establishing a frequent pattern tree FP-tree provided by an embodiment of this application;
  • FIG. 4 is a schematic structural diagram of an electronic device for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application;
  • FIG. 5 is a schematic structural diagram of an electronic device for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application.
  • the electronic devices in this application may include smartphones, tablets, PDAs, laptops, mobile Internet devices MID (Mobile Internet Devices, MID) or wearable devices, etc.
  • the above electronic devices are only examples, not exhaustive. Including, but not limited to, the foregoing electronic device.
  • the foregoing electronic device is referred to as user equipment UE (User Equipment, abbreviated as UE).
  • UE User Equipment
  • the above user equipment is not limited to the above-mentioned monetization form, and may also include, for example, intelligent vehicle-mounted terminals, computer equipment, and so on.
  • FIG. 1A is a schematic flowchart of a method for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application.
  • the method is applied to an electronic device.
  • the method includes steps S101 to S104:
  • Step S101 Receive any hospitalization data of any participant.
  • the hospitalization data includes diagnostic data and actual preoperative inspection items.
  • the diagnostic data includes the insured ’s demographic parameters and disease information.
  • the demographic parameters include height, weight, age, disease history, education level, marital status, etc. Etc.
  • the disease information is specifically the name of the disease, the severity of the disease, etc.
  • Step S102 Extract the diagnosis data in the hospitalization data, input the diagnosis data into a pre-trained hospitalization prediction model, and output the hospitalization probability corresponding to the insured person.
  • keyword identification is performed on the hospitalization data to identify the diagnostic data in the hospitalization data, that is, the disease name, disease severity level, and demographic parameters, and the disease name, disease severity level, and demographic parameters are input data Input into the preset trained hospitalization prediction model to get the insured's hospitalization probability.
  • the input data of the disease name, disease severity level, and demographic parameters specifically include: constructing the disease name, disease severity level, and demographic parameters as an initial feature matrix, where the initial feature matrix is non-0-1 Information matrix, normalize the non-0-1 data in the initial feature matrix, convert all the elements in the feature matrix into a data matrix of numbers 0 to 1, and then, the data in the data matrix is 0 Rearrangement, that is, the 0 values in the data matrix are arranged next to each other in the same row or column to form an input data matrix, and the input data matrix is input to the pre-trained hospitalization prediction model to perform multi-layer forward operations, after global pooling Then, the feature vector of the input data matrix is obtained, and the feature vector is input into the softmax classifier to obtain the “category number” corresponding to the feature vector, and the value corresponding to the “category number” is obtained to obtain the hospitalization probability of the insured person.
  • the initial feature matrix is non-0-1 Information matrix
  • Step S103 When the hospitalization probability is greater than the first threshold, obtain a preset set of preoperative examination items corresponding to the diagnostic data.
  • the first threshold is specifically 0.5, 0.6, 0.7, 0.8 or other values.
  • obtaining the preset set of preoperative inspection items corresponding to the diagnostic data specifically includes: obtaining the name of the disease in the diagnostic data, as shown in Table 1, illustrating the name of the disease and the set of preoperative inspection items as an example
  • a pre-operative inspection item set corresponding to the disease name is obtained according to the mapping relationship shown in Table 1, and a preset pre-operative inspection item set for the diagnostic data is obtained.
  • Step S104 Compare the actual preoperative inspection item set with the preset preoperative inspection item set to determine the rationality of the insured's hospitalization behavior.
  • comparing the actual preoperative inspection item set with the preset preoperative inspection item set specifically includes: comparing the type of the actual preoperative inspection item set with the preset preoperative inspection item set Type comparison to determine the difference between the two types, that is, to determine the similarity between the preset set of preoperative inspection items and the actual set of preoperative inspection items, if the similarity is greater than the second threshold, determine the The actual preoperative examination items are normal, and it is determined that the insured ’s hospitalization behavior is reasonable. If the similarity is less than or equal to the second threshold, the actual preoperative examination items are abnormal, and the insured ’s The hospitalization behavior is unreasonable.
  • the second threshold may be 0.5, 0.6, 0.7, 0.8 or other values.
  • the similarity calculation formula is:
  • S is the similarity between the preset set of preoperative examination items and the actual set of preoperative examination items
  • A is the preset set of preoperative examination items
  • B is the set of actual preoperative examination items
  • Card (A ⁇ B) is the number of elements in A ⁇ B
  • Card (A ⁇ B) is the number of elements in A ⁇ B.
  • the input of the hospitalization data of the insured person is received, the diagnostic data in the hospitalization data is extracted, and the preset preoperative inspection item set of the insured person is obtained according to the diagnostic data, and the preset The inspection item set is compared with the actual inspection item set in the hospitalization data to determine the similarity of the two, and whether the actual preoperative inspection item set of the insured person is abnormal according to the similarity degree, and the reasonable inpatient behavior of the insured person is determined Therefore, when the hospitalization request of the insured person is received, the reasonableness of the insured person ’s hospitalization behavior can be detected before considering whether to join the insured person into the hospital, which can reduce the medical insurance fund payment before surgery Check the probability of expenses, increase the way to determine the rationality of hospitalization, provide data reference for the reform of the medical system, improve the persuasion of the medical system reform, and improve the medical system.
  • the hospitalization data of the insured person who received the same disease in the same time period check the similarity between the actual preoperative inspection item set and the preset preoperative inspection item set, which takes a long time and repeats The large number of calculations affects the calculation speed of the server. Based on this, the following algorithm for calculating the similarity between the actual preoperative inspection items of multiple insured persons and the preset preoperative inspection items is provided.
  • the actual preoperative inspection item sets A, B, C, D ... of the N insured persons are obtained, Determine the preset preoperative inspection item set ⁇ corresponding to the N insured persons, obtain all preoperative inspection items of the medical institution, and combine A, B, C, D ..., and ⁇ with all preoperative inspection items
  • the comparison result is obtained, and the binary (ie 0 and 1) data matrix of the actual preoperative inspection item set of the N insured persons and the preset preoperative inspection item set is constructed according to the comparison result, including
  • the preoperative inspection item is used as the reference set j, and compares A, B, C, D ..., and ⁇ with the reference set j to determine that any element of the reference set j is in A, B, C, D ..., And whether there is a corresponding element in ⁇ , if so, mark A, B, C, D ..., and ⁇ is marked as 1 at this position, if
  • J i is the jeckard similarity coefficient of the i-th insured person ’s actual preoperative inspection item set and the preset inspection item set
  • M i 11 is the i-th insured person ’s actual preoperative inspection item set
  • the binary data matrix is 1 and the preset inspection item set ⁇ is also the total number of 1 in the binary data matrix
  • M i 10 is the actual preoperative inspection item set of the i-th insured person in the binary data matrix Is 1 and the preset inspection item set ⁇ is the total number of 0 in the binary data matrix
  • M i 01 is the actual preoperative inspection item set of the i-th insured person is 0 in the binary data matrix
  • the preset inspection The total number of items in the binary data matrix is 1.
  • all the preoperative examination items of this medical institution are ⁇ blood routine, urine routine, electrocardiogram, blood pressure, blood glucose, oral cavity ⁇ , and a, b, c, d, e, f represent blood routine and urine, respectively Routine, electrocardiogram, blood pressure, blood sugar, oral cavity, so the set of all preoperative examination items of the medical institution is ⁇ a, b, c, d, e, f ⁇ .
  • the method further includes:
  • each department with clinical operation qualification obtain the hospitalization data of each department within a preset time period, determine the multiple insured persons whose actual preoperative examination items are abnormal in the hospitalization data of each department, and calculate each department at the preset time
  • the hospitalization data of each department within a preset period of time is analyzed to obtain the proportion of the preoperative cost abnormality of the insured persons who are treated in each department, and each department is sorted according to this ratio to obtain a ranking result, according to the ranking
  • the top-ranked departments were adjusted, and the hospitalization system of the departments was adjusted accordingly, providing data reference for the reform of the hospitalization system of medical institutions, which was conducive to improving the overall quality of medical institutions and reducing the adoption of medical insurance funds by each department to pay for preoperative inspections.
  • the probability of expenses is analyzed to obtain the proportion of the preoperative cost abnormality of the insured persons who are treated in each department, and each department is sorted according to this ratio to obtain a ranking result, according to the ranking
  • the method further includes:
  • the occurrence frequency monitors the doctors corresponding to all the doctor information, which specifically includes: the doctors corresponding to the top five doctors information of the occurrence frequency take regular spot checks on the hospitalization data sheets they have taken, such as the preoperative examination in the hospitalization data sheet
  • the preset threshold is related to the ranking of doctors, the higher the ranking The smaller the preset threshold.
  • the hospitalization data of all the insured persons of the preoperative examination expense item is analyzed, the frequency of occurrence of each doctor's information in the hospitalization data is obtained, and the hospitalization data sheet that the doctor takes during the clinical operation is monitored based on the frequency , To reduce the probability that doctors will enroll the insured persons into hospitals with low standards to apply for medical insurance funds to pay for preoperative examination costs.
  • FIG. 2 is a schematic flowchart of another method for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application.
  • the method is applied to an electronic device, and the method includes steps S201-S209:
  • Step S201 Receive any hospitalization data of any insured person input.
  • the hospitalization data includes diagnostic data and actual preoperative inspection itemsets.
  • Step S202 Extract the diagnosis data in the hospitalization data, input the diagnosis data into a pre-trained hospitalization prediction model, and output the hospitalization probability of the insured person.
  • Step S203 When the hospitalization probability is greater than the first threshold, obtain a preset set of preoperative examination items corresponding to the diagnostic data.
  • Step S204 comparing the actual preoperative inspection item set with the preset preoperative inspection item set to determine the similarity between the preset preoperative inspection item set and the actual preoperative inspection item set.
  • Step S205 If the similarity is greater than the second threshold, determine that the actual pre-operative examination items are normal, and determine that the insured's hospitalization behavior is reasonable.
  • Step S206 If the similarity is less than or equal to the second threshold, determine that the actual preoperative examination item is abnormal, and determine that the insured person's hospitalization behavior is unreasonable.
  • Step S207 When the hospitalization probability is less than or equal to the first threshold, send the insured's hospitalization data to the online verification center, prompting the online verification center to verify the actual pre-operation in the hospitalization data Check if the item set is abnormal.
  • an online verification center is established in advance.
  • the hospitalization probability is less than the first threshold, it means that the insured person does not need to be hospitalized to a large extent.
  • the comparison of inspection item sets cannot accurately determine the reasonableness of the hospitalization behavior, so the insured ’s hospitalization data is sent to the online verification center, prompting the online verification center to verify the actual preoperative inspection in the hospitalization data Whether the item set is abnormal, and whether the verification is abnormal specifically includes: if it is determined that the insured person does not need to be hospitalized, but there are still pre-operative examination items, the pre-operative examination item of the insured person is determined to be abnormal, if it is determined that the insured person needs If you are hospitalized, but the pre-operative examination item of the insured person does not correspond to the disease of the insured person, it is determined that the pre-operative examination item of the insured person is abnormal.
  • Step S208 Receive the verification result from the online verification center, and when the verification result confirms that the actual preoperative inspection item set is normal, determine that the insured person ’s hospitalization behavior is reasonable, and determine the actual status of the insured person.
  • the set of preoperative inspection items is transferred to the network equipment related to the preoperative inspection.
  • the actual preoperative inspection item set of the insured person is transferred to a network device related to the preoperative inspection, so as to include the information of the insured person and the preoperative inspection item list of the insured person into the network device, It is determined that the insured person is an authorized user in terms of the actual inspection item, so that the insured person can query the information of the insured person in the network device when doing the preoperative inspection item.
  • Step S209 When the verification result is that the actual preoperative examination item set is abnormal, it is determined that the insured person's hospitalization behavior is unreasonable, and it is forbidden to transfer the insured person's actual preoperative examination item set to the network
  • the device prompts to re-enter the pre-operative inspection items of the insured person.
  • the input of the hospitalization data of the insured person is received, the diagnostic data in the hospitalization data is extracted, and the diagnostic data is input into the hospitalization prediction model to obtain the hospitalization probability of the insured person.
  • obtain the preset preoperative inspection item set of the insured person compare the preset inspection item set with the actual inspection item set in the hospitalization data, and determine the similarity between the two, based on the similarity To determine whether the insured person ’s actual preoperative inspection item set is abnormal and to determine the reasonableness of the insured person ’s hospitalization behavior.
  • the insured person needs to be hospitalized, you can compare the preset preoperative inspection item set, To determine whether the insured person has taken the medical insurance fund to pay for the preoperative inspection cost, based on the model to determine the probability of hospitalization, it can realize the targeted determination of the insured person whose probability of hospitalization meets the conditions, and improve the accuracy of the judgment.
  • the hospitalization probability is less than the first threshold, the hospitalization data is forwarded to the online verification center to verify the rationality of the insured person ’s preoperative inspection items online. Therefore, When the insured person does not need to be hospitalized, it can be combined with the online verification system to assist in verifying the rationality of the insured person's hospitalization.
  • the organic combination of the two can completely solve the behavior of drawing the medical insurance fund to pay for the preoperative inspection. Increased the way to determine the rationality of hospitalization, provide data reference for the medical system reform, improve the persuasion of the medical system reform, and improve the medical system.
  • FIG. 3 is a schematic flowchart of another method for predicting the rationality of hospitalization based on a prediction model provided by an embodiment of the present application.
  • the method is applied to an electronic device.
  • the method includes steps S301 to S312:
  • Step S301 Receive any hospitalization data of any participant.
  • the hospitalization data includes diagnostic data and actual preoperative inspection itemsets.
  • Step S302 Extract the diagnosis data in the hospitalization data, input the diagnosis data into a pre-trained hospitalization prediction model, and output the hospitalization probability of the insured person.
  • Step S303 When the hospitalization probability is greater than the first threshold, obtain a preset set of preoperative examination items corresponding to the diagnostic data.
  • Step S304 Compare the actual preoperative inspection item set with the preset preoperative inspection item set to determine the similarity between the preset preoperative inspection item set and the actual preoperative inspection item set.
  • Step S305 If the similarity is less than or equal to the second threshold, determine that the actual preoperative inspection item is abnormal, and determine that the insured person's hospitalization behavior is unreasonable.
  • Step S306 When the hospitalization probability is less than or equal to the first threshold, send the insured's hospitalization data to the online verification center, prompting the online verification center to verify the actual operation in the hospitalization data Before checking whether the program set is abnormal.
  • Step S307 Receive the verification result from the online verification center, and when the verification result is to confirm that the actual preoperative inspection item set is abnormal, determine that the insured's hospitalization behavior is unreasonable.
  • Step S308 Count the insurers whose medical institutions have unreasonable hospitalization behavior within a preset time period.
  • Step S309 If the number of the insured person is single, obtain the inpatient data of the insured person, extract the actual preoperative inspection item set of the insured person from the inpatient data of the insured person, and The actual preoperative inspection item set is marked as the target preoperative inspection item set of the medical institution.
  • the preset time period is specifically 1 week, 1 month, 6 months, 1 year or other values.
  • Step S310 If the number of the insured persons is multiple, obtain the hospitalization data of the multiple insured persons, and extract each of the multiple insured persons from the hospitalized data of the multiple insured persons An actual preoperative inspection item set of one insured person, the actual preoperative inspection item set of each insured person is marked as a thing set, and a plurality of thing sets of the plurality of insured persons are obtained.
  • Step S311 Determine the frequent item set in the plurality of thing sets based on the frequent pattern growth FP-Growth algorithm, and mark the frequent item set as the target preoperative inspection item set of the medical institution.
  • the preoperative inspection items in the target preoperative inspection item set are preoperative inspection items that are prone to abnormalities in the preoperative inspection of the medical institution.
  • determining the frequent item set in the multiple transaction sets based on the FP-Growth algorithm specifically includes: setting a minimum support P, filtering multiple frequent elements in the multiple transaction sets based on the minimum support P, and using the multiple Frequent elements construct the FP-tree of the multiple thing sets, set the number of elements required in the frequent item set, read multiple frequent item sets matching the number of elements from the FP-tree, and obtain the multiple frequent items Centralized frequent itemsets with the highest degree of support. If the number of frequent itemsets with the highest degree of support is multiple, obtain the intersection of the multiple frequent itemsets, and use the intersection as the target preoperative inspection item set or obtain the multiple Union of frequent itemsets, and use this union as the target preoperative inspection item set.
  • the following is an example to illustrate the specific process of determining frequent item sets and target preoperative inspection item sets.
  • Thing set number Set of things Set of things 001 ⁇ r, z, h, j, p ⁇ ⁇ r, z ⁇ 002 ⁇ z, y, x, w, v, u, t, s ⁇ ⁇ z, y, x, t, s ⁇ 003 ⁇ z ⁇ ⁇ z ⁇ , 004 ⁇ r, x, n, o, s ⁇ ⁇ r, x, s ⁇ 005 ⁇ y, r, x, z, q, t, p ⁇ ⁇ y, r, x, z, t ⁇ 006 ⁇ y, z, x, e, q, s, t, m ⁇ ⁇ y, z, x, s, t ⁇
  • Figure 3A FP-tree
  • the tree node of the FP-tree gives the single element in the set and its total number of occurrences in the new set of things
  • the number of occurrences of the element of the root node of the path shows the corresponding The number of occurrences of the sequence (ie support).
  • all elements on each path constitute a frequent item set
  • the root node on the path shows the support of the frequent item set; set the number of elements required in the frequent item set, from the FP-
  • the tree node of the tree intercepts the root node corresponding to the number of elements, and the elements between the tree node and the root node form a frequent item set.
  • the frequent item set can be obtained ⁇ z, x, y, s ⁇ , ⁇ z, x, y, r ⁇ , so the union of the frequent itemsets ⁇ z, x, y, s ⁇ , ⁇ z, x, y, r ⁇ ⁇ z , x, y, r, s ⁇ is marked as the final frequent item set of the 6 thing sets, that is, the target preoperative check item set; or set the second support degree P2, and conduct the third round based on the second support degree P2 Scan, scan the FP-tree with a support level greater than or equal to P2, and use the set as a frequent item set.
  • the set ⁇ z ⁇ , ⁇ z, x ⁇ and ⁇ z, x, y ⁇ can be obtained
  • the support levels are 5, 3 and 3, respectively, so ⁇ z ⁇ , ⁇ z, x ⁇ and ⁇ z, x, y ⁇ can be obtained, so ⁇ z ⁇ , ⁇ z, x ⁇ and ⁇ z, x, y can be taken
  • the union ⁇ z, x, y ⁇ of ⁇ serves as the final frequent item set of the 6 thing sets, that is, the target pre-operative check item set.
  • Step S312 Send the target preoperative inspection item set to the network device corresponding to the medical institution to adjust the management system of the medical institution in terms of the target preoperative examination.
  • the input of the hospitalization data of the insured person is received, the diagnostic data in the hospitalization data is extracted, and the diagnostic data is input into the hospitalization prediction model to obtain the hospitalization probability of the insured person.
  • obtain the preset preoperative inspection item set of the insured person compare the preset inspection item set with the actual inspection item set in the hospitalization data, and determine the similarity between the two, based on the similarity To determine whether the insured person ’s actual preoperative inspection item set is abnormal and to determine the reasonableness of the insured person ’s hospitalization behavior.
  • the insured person needs to be hospitalized, you can compare the preset preoperative inspection item set, To determine whether the insured person has taken the medical insurance fund to pay for the preoperative inspection fee, based on the model to determine the probability of hospitalization, it can realize the targeted determination of the insured person whose hospitalization probability meets the conditions, improve the accuracy of the judgment, and, if the When the hospitalization probability is less than the first threshold, the hospitalization data is forwarded to the online verification center to verify the rationality of the insured person ’s preoperative inspection items online. Therefore, When the insured person does not need to be hospitalized, the online verification system can be used to assist in checking the rationality of the insured person's hospitalization.
  • the organic combination of the two can completely solve the behavior of drawing the medical insurance fund to pay for the preoperative inspection fee.
  • determine the rationality of hospitalization provide data reference for the reform of the medical system, improve the persuasion of the medical system reform, and improve the medical system.
  • obtain multiple insured persons whose hospitalization behavior is unreasonable for a preset period of time determine the target preoperative examination item set of the medical institution based on the inpatient data of the multiple insured persons, and feed back the target preoperative examination item set to the network
  • the side equipment can realize the targeted reform of the medical system.
  • the target preoperative inspection item set can be determined in time, which can improve the efficiency of medical reform.
  • the electronic device 400 includes a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are different from the one or more application programs, and the one or more programs are Stored in the aforementioned memory and configured to be executed by the aforementioned processor, and the aforementioned program includes instructions for performing the following steps;
  • the actual preoperative inspection item set is compared with the preset preoperative inspection item set to determine the rationality of the insured's hospitalization behavior.
  • the instructions in the above program are also used to perform the following operations:
  • the item set is transferred to the network equipment related to the preoperative inspection; when the verification result is to confirm that the actual preoperative inspection item set is abnormal, it is determined that the insured ’s hospitalization behavior is unreasonable, and the The set of preoperative inspection items is transmitted to the network device, prompting to re-enter the preoperative inspection items of the insured person.
  • the diagnosis data includes the name of the disease, the severity of the disease, and the demographic parameters of the insured.
  • the diagnosis data is input into a pre-trained hospitalization prediction model, and the parameters are output.
  • the insured s hospitalization probability
  • the instructions in the above procedure are specifically used to perform the following operations:
  • the input data matrix composed of the disease name, disease severity level, and demographic parameters is input to the pre-trained hospitalization prediction model to perform a forward operation, and the hospitalization probability of the insured person is output.
  • the instructions in the above program are specifically used to perform the following operations:
  • the preset preoperative inspection item set of the diagnostic data is obtained.
  • the instructions in the above procedure are specific Used to perform the following operations:
  • S is the similarity between the preset set of preoperative examination items and the actual set of preoperative examination items
  • A is the preset set of preoperative examination items
  • B is the set of actual preoperative examination items
  • Card (A ⁇ B) is the number of elements in A ⁇ B
  • Card (A ⁇ B) is the number of elements in A ⁇ B.
  • the instructions in the above program are also used to perform the following operations:
  • the person's actual preoperative inspection item set, the actual preoperative inspection item set of each insured person is marked as a thing set, to obtain the multiple thing sets of the multiple insured persons, based on the frequent pattern growth FP-
  • the Growth algorithm determines the frequent item sets in the plurality of thing sets, and marks the frequent item sets as the target preoperative inspection item set of the medical institution;
  • the preoperative inspection items in the target preoperative inspection item set are preoperative inspection items that are prone to abnormalities in the preoperative inspection of the medical institution.
  • the instructions in the above program are also used to perform the following operations:
  • the instructions in the above program are also used to perform the following operations:
  • the instructions in the above program are also used to perform the following operations:
  • FIG. 5 shows a block diagram of a possible functional unit composition of an electronic device 500 for predicting the rationality of hospitalization based on the hospitalization model involved in the above embodiment.
  • the electronic device 500 includes a receiving unit 510, an input unit 520, and an acquisition unit. Unit 530, comparison unit 540, among them;
  • the receiving unit 510 is configured to receive the hospitalization data of any one of the insured persons, and the hospitalization data includes diagnosis data and an actual preoperative inspection item set;
  • the input unit 520 is configured to extract the diagnosis data in the hospitalization data, input the diagnosis data to a pre-trained hospitalization prediction model, and output the hospitalization probability corresponding to the insured person;
  • the obtaining unit 530 is configured to obtain a preset set of preoperative examination items corresponding to the diagnostic data when the hospitalization probability is greater than the first threshold;
  • the comparison unit 540 is configured to compare the actual preoperative inspection item set with the preset preoperative inspection item set to determine the rationality of the insured's hospitalization behavior.
  • the electronic device 500 further includes a sending unit 550;
  • the sending unit 550 is configured to send the hospitalization data of the insured to the online verification center when the hospitalization probability is less than or equal to the first threshold, prompting the online verification center to verify the hospitalization data Whether the actual preoperative inspection item set is abnormal; and used to receive the verification result from the online verification center; when the verification result is to confirm that the actual preoperative check item set is normal, determine the insured ’s hospitalization Act reasonably, transfer the actual preoperative inspection item set of the insured person to the network equipment related to the preoperative inspection; when the verification result is to confirm that the actual preoperative check item set is abnormal, determine the insured person ’s The hospitalization behavior is unreasonable, and it is forbidden to transfer the actual preoperative inspection item set of the insured person to the network device, prompting to re-enter the preoperative inspection item of the insured person.
  • the diagnosis data includes the disease name, the severity of the disease, and the demographic parameters of the insured person.
  • the insured person After inputting the diagnostic data into a pre-trained hospitalization prediction model, the insured person ’s In the probability of hospitalization, the input unit 520 is specifically used for: extracting the disease name, disease severity level, and demographic parameters of the insured person in the diagnostic data; and for the disease name, disease severity level, and demographic parameters
  • the input data matrix is composed and input to the pre-trained hospitalization prediction model to perform a forward operation, and the hospitalization probability of the insured person is output.
  • the obtaining unit 530 when acquiring the preset preoperative inspection item set corresponding to the diagnostic data, is specifically configured to: based on the mapping relationship between the disease name and the preoperative inspection item set and the diagnosis For the name of the disease in the data, a preset set of preoperative examination items of the diagnostic data is obtained.
  • the comparing unit 540 when comparing the actual preoperative inspection item set with the preset preoperative inspection item set to determine the reasonableness of the insured ’s hospitalization behavior, the comparing unit 540, specifically Used to: determine the similarity between the preset set of preoperative inspection items and the actual set of preoperative inspection items, if the similarity is greater than a second threshold, determine that the actual preoperative inspection items are normal, and determine the parameters The insured ’s hospitalization behavior is reasonable, if the similarity is less than or equal to the second threshold, it is determined that the actual preoperative inspection item is abnormal, and the insured ’s hospitalization behavior is determined to be unreasonable;
  • S is the similarity between the preset set of preoperative examination items and the actual set of preoperative examination items
  • A is the preset set of preoperative examination items
  • B is the set of actual preoperative examination items
  • Card (A ⁇ B) is the number of elements in A ⁇ B
  • Card (A ⁇ B) is the number of elements in A ⁇ B.
  • the obtaining unit 530 is further used to: count the insured persons whose hospitalization behavior of the medical institution is unreasonable within a preset period of time; and when the number of insured persons is a single, Obtain the insured person's hospitalization data, extract the insured person's actual preoperative examination item set from the insured person's inpatient data, and mark the actual preoperative examination item set as the medical institution's The target preoperative inspection item set; and for obtaining the hospitalization data of the plurality of insured persons when the number of the insured persons is multiple, and extracting the hospitalization data from the multiple insured persons
  • the sending unit 550 is further configured to: send the target preoperative inspection item set to the network device corresponding to the medical institution to adjust the management of the medical institution in the target preoperative examination system.
  • the electronic device 500 further includes an adjustment unit 560, which is used to obtain various departments qualified for clinical surgery; obtain the hospitalization data of each department within a preset time period; determine the hospitalization data of each department The ratio of insured persons with abnormal preoperative examination items to the total number of insured persons in the preset period of time; according to the ratio, each department is ranked in terms of preoperative examination items, and the adjustment is ranked in the top five The management strategy of the department in the preoperative inspection project.
  • an adjustment unit 560 which is used to obtain various departments qualified for clinical surgery; obtain the hospitalization data of each department within a preset time period; determine the hospitalization data of each department The ratio of insured persons with abnormal preoperative examination items to the total number of insured persons in the preset period of time; according to the ratio, each department is ranked in terms of preoperative examination items, and the adjustment is ranked in the top five The management strategy of the department in the preoperative inspection project.
  • the electronic device 500 further includes an adjustment unit 560, which is used to acquire all insured persons whose actual preoperative examination items are abnormal in the hospitalization data within a preset time period; acquire all the parameters All doctor information in the insured's hospitalization data; determine the frequency of occurrence of each doctor's information, and supervise the doctor corresponding to each doctor's information according to the frequency of occurrence.
  • an adjustment unit 560 which is used to acquire all insured persons whose actual preoperative examination items are abnormal in the hospitalization data within a preset time period; acquire all the parameters All doctor information in the insured's hospitalization data; determine the frequency of occurrence of each doctor's information, and supervise the doctor corresponding to each doctor's information according to the frequency of occurrence.
  • An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, and the computer program enables the computer to execute any of the methods described in the above method embodiments based on The predictive model predicts some or all steps of the method of hospitalization rationality.
  • An embodiment of the present application further provides a computer program product, the computer program product includes a non-transitory computer-readable storage medium storing a computer program, the computer program is operable to cause the computer to execute as described in the above method embodiments Some or all steps of any method to predict the rationality of hospitalization based on a prediction model.
  • the disclosed device may be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or may Integration into another system, or some features can be ignored, or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical or other forms.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or software program modules.
  • the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it may be stored in a computer-readable memory.
  • the technical solution of the present application essentially or part of the contribution to the existing technology or all or part of the technical solution can be embodied in the form of a software product, the computer software product is stored in a memory, Several instructions are included to enable a computer device (which may be a personal computer, server, network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application.
  • the aforementioned memory includes: U disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.
  • the program may be stored in a computer-readable memory, and the memory may include: a flash disk , Read-Only Memory (English: Read-Only Memory, abbreviation: ROM), Random Access Device (English: Random Access Memory, abbreviation: RAM), magnetic disk or optical disk, etc.
  • ROM Read-Only Memory
  • RAM Random Access Device
  • magnetic disk or optical disk etc.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • General Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Tourism & Hospitality (AREA)
  • Finance (AREA)
  • Game Theory and Decision Science (AREA)
  • Accounting & Taxation (AREA)
  • Educational Administration (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Biomedical Technology (AREA)
  • Primary Health Care (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Technology Law (AREA)
  • Medical Treatment And Welfare Office Work (AREA)

Abstract

本申请公开了一种基于住院模型预测住院合理性的方法及相关产品，该方法应用于电子设备，该方法包括：接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率；在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。本申请实施例有利于增加检测住院合理行的方式，为医疗体制改革提供数据参考。

Description

基于预测模型预测住院合理性的方法及相关产品
本申请要求于2018年10月30日提交中国专利局、申请号为2018112769268、申请名称为“基于预测模型预测住院合理性的方法及相关产品”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
本申请涉及医疗技术领域，具体涉及一种基于预测模型预测住院合理性的方法及相关产品。
背景技术
随着国家基本医保制度不断加强，越来越多的人加入医保行列，参保人就诊时，医保统筹基金会为参保人报销绝大的医疗费用，但是目前的医保制度仍然不够完善，导致报销过程中存在很多利益问题。在很多城市，对体检项目未纳入医保统筹基金的报销范畴，参保人的体检费用只能个人报销。但是对于参保人来说，住院费用只要符合起付线等条件，可使用医保统筹基金报销部分甚至全部住院费用。因此部分参保人可能与医生合谋，套取医保统筹基金为参保人报销检查费用，即在参保人无需住院的情况下，将参保人收入院，再额外增加一些无关的术前检查项目，从而增加术前检查费用，简接增加住院费用，将住院费用达到起付线，套取医保统筹基金支付检查费用。目前，参保人在做术前检查项目时，相关人员仅核对该术前检查项目是否为医生指定的术前检查项目，未考虑参保人是否具有资格以及该术前检查项目的合理性。
现有的方法确认参保人是否具有住院资格的方式单一、准确度低，存在套取统筹基金支付检查费用的行为。因此，亟需提供一种从术前检查项目角度判断参保人住院是否具有住院资格的方法。
发明内容
本申请实施例提供了一种基于预测模型预测住院合理性的方法及相关产品，以期预测参保人的术前检查项目是否异常，确定该参保人的住院行为的合理性。
第一方面，本申请实施例提供一种基于预测模型预测住院合理性的方法，所述方法应用于电子设备，所述方法包括：
接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；
提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人对应的住院概率；
在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；
将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
第二方面，本申请实施例提供一种基于预测模型预测住院合理性的电子设备，所述电子设备包括：
接收单元，用于接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；
输入单元，用于提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好 的住院预测模型，输出所述参保人对应的住院概率；
获取单元，用于在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；
比对单元，用于将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
第三方面，本申请实施例提供一种电子设备，包括一个或多个处理器、一个或多个存储器、一个或多个收发器，以及一个或多个程序，所述一个或多个程序被存储在所述存储器中，并且被配置由所述一个或多个处理器执行，所述程序包括用于执行如第一方面所述的方法中的步骤的指令。
第四方面，本申请实施例提供一种计算机可读存储介质，其用于存储计算机程序，其中，所述计算机程序被处理器执行，以实现如第一方面所述的方法。
实施本申请实施例，具有如下有益效果：
可以看出，在本申请实施例中，接收输入的参保人的住院数据，提取住院数据中的诊断数据，依据诊断数据获取该参保人的预设术前检查项目集，将该预设检查项目集与该住院数据中的实际检查项目集比对，判断该参保人的术实际前检查项目集是否异常，从而确定该参保人住院行为的合理性，因此，在接收到参保人的住院请求时，可检测出该参保人住院行为的合理性，故可减少套取医保基金支付术前检查费用的概率，为医疗体制改革提供数据参考，完善医疗制度。
附图说明
图1A为本申请实施例提供的一种基于预测模型预测住院合理性的方法的流程示意图；
图1B为本申请实施例提供的一种基于术前检查项目确定出的二元数据矩阵的示意图；
图2为本申请实施例提供的另一种基于预测模型预测住院合理性的方法的流程示意图；
图3为本申请实施例提供的另一种基于预测模型预测住院合理性的方法的流程示意图；
图3A为本申请实施例提供的一种建立频繁模式树FP-tree过程的示意图；
图4为本申请实施例提供的一种基于预测模型预测住院合理性的电子设备的结构示意图；
图5为本申请实施例提供的一种基于预测模型预测住院合理性的电子设备的结构示意图。
具体实施方式
本申请的说明书和权利要求书及所述附图中的术语“第一”、“第二”、“第三”和“第四”等是用于区别不同对象，而不是用于描述特定顺序。此外，术语“包括”和“具有”以及它们任何变形，意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、系统、产品或设备没有限定于已列出的步骤或单元，而是可选地还包括没有列出的步骤或单元，或可选地还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。
本申请中的电子设备可以包括智能手机、平板电脑、掌上电脑、笔记本电脑、移动互联网设备MID(Mobile Internet Devices，简称：MID)或穿戴式设备等，上述电子设备仅是举例，而非穷举，包含但不限于上述电子设备，为了描述的方便，下面实施例中将上述电子设备称为用户设备UE(User equipment，简称：UE)。当然在实际应用中，上述用户 设备也不限于上述变现形式，例如还可以包括：智能车载终端、计算机设备等等。
参阅图1A，图1A为本申请实施例提供的一种基于预测模型预测住院合理性的方法的流程示意图，该方法应用于电子设备，该方法包括步骤S101～S104中所示的内容：
步骤S101、接收输入的任意一个参保人的住院数据。
其中，该住院数据包括诊断数据和实际术前检查项目集，诊断数据包括参保人的人口学参数和疾病信息，人口学参数包括身高、体重、年龄、疾病史、文化程度、婚姻状况，等等；疾病信息具体为疾病名称、疾病严重等级，等等。
步骤S102、提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人对应的住院概率。
可选的，对该住院数据进行关键字识别，识别出该住院数据中的诊断数据，即疾病名称、疾病严重等级、人口学参数，将该疾病名称、疾病严重等级、人口学参数组成输入数据输入到预设训练好的住院预测模型，得到该参保人的住院概率。其中，将该疾病名称、疾病严重等级、人口学参数组成输入数据具体包括：将该疾病名称、疾病严重等级、人口学参数构建为初始特征矩阵，其中，该初始特征矩阵为非0-1的信息矩阵，对于该初始特征矩阵中的非0-1数据进行归一化处理，将该特征矩阵中的所有元素转化为数字0～1的数据矩阵，然后，对该数据矩阵中为0的数据重排列，即将该数据矩阵中的0值紧邻排列在同一行或者同一列组成输入数据矩阵，将该输入数据矩阵输入到该预先训练好的住院预测模型执行多层正向运算，经过全局池化后得到该输入数据矩阵的特征向量，将该特征向量输入到softmax分类器，得到该特征向量对应的“类别号”，获取该“类别号”对应的数值，得到该参保人的住院概率。
步骤S103、在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集。
其中，该第一阈值具体为0.5、0.6、0.7、0.8或者其他值。
可选的，获取所述诊断数据对应的预设术前检查项目集具体包括：获取所述诊断数据中的疾病名称，如表1所示，举例示出了疾病名称和术前检查项目集的映射关系，依据表1所示的映射关系获取与所述疾病名称对应的术前检查项目集，得到所述诊断数据的预设术前检查项目集。
表1
Figure PCTCN2019095050-appb-000001
步骤S104、将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
可选的，将所述实际术前检查项目集和所述预设术前检查项目集比对具体包括：将所 述实际术前检查项目集中的种类与所述预设术前检查项目集的种类比对，确定两者在种类上的差异性，即确定所述预设术前检查项目集与所述实际术前检查项目集的相似度，如所述相似度大于第二阈值，确定所述实际术前检查项目正常，确定所述参保人的住院行为合理，如所述相似度小于或者等于所述第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理。
其中，第二阈值可以为0.5、0.6、0.7、0.8或者其他值。
所述相似度计算公式为：
Figure PCTCN2019095050-appb-000002
其中，S为所述预设术前检查项目集与所述实际术前检查项目集的相似度，A为所述预设术前检查项目集，B为所述实际术前检查项目集，Card(A∩B)为A∩B中的元素个数，Card(A∪B)为A∪B中的元素个数。
可以看出，在本申请实施例中，接收输入的参保人的住院数据，提取住院数据中的诊断数据，依据诊断数据获取该参保人的预设术前检查项目集，将该预设检查项目集与该住院数据中的实际检查项目集比对，确定两者的相似度，依据相似度判断该参保人的实际术前检查项目集是否异常，确定该参保人住院行为的合理性，因此，在接收到参保人的住院请求时，可在检测出该参保人住院行为的合理性后，再考虑是否将该参保人接纳入院，可减少套取医保基金支付术前检查费用的概率，增加确定住院合理性的方式，为医疗体制改革提供数据参考，提高医疗体制改革的说服力，完善医疗制度。
在一可能的示例中，如在同一时间段接收到相同疾病的参保人的住院数据，一一核对实际术前检查项目集与预设术前检查项目集的相似度，耗时久，重复计算的次数多，影响服务器的运算速度，基于此提供下面用于计算多个参保人的实际术前检查项目与预设术前检查项的相似度的算法。
假定在同一时间段接收到N(N为大于等于2的整数)个参保人的住院数据，获取该N个参保人中的实际术前检查项目集A、B、C、D……，确定该N个参保人对应的预设术前检查项目集Ω，获取该医疗机构的所有术前检查项目，将该A、B、C、D……，以及Ω与该所有术前检查项目比对得到比对结果，根据比对结果构建该N个参保人的实际术前检查项目集与该预设术前检查项目集的二元(即0和1)数据矩阵，包括将该所有术前检查项目作为参考集合j，将该A、B、C、D……，以及Ω与该参考集合j比较，确定该参考集合j的任意一个元素在A、B、C、D……，以及Ω中是否存在对应元素，如是，将A、B、C、D……，以及Ω在该位置标记为1，如不包含，在该位置标记为0，从而将该A、B、C、D……，以及Ω转换为与参考集合j相同维度的二元集合，将转换后的A、B、C、D……，以及Ω组成该二元数据矩阵，所以二元数据矩阵的行数为所有术前检查项目的数量，列数为(N+1)；确定二元数据矩阵确定每个实际术前检查项目与预设术前检查项目集的杰卡德相似系数Jaccard，依据该杰卡德相似系数Jaccard确定该N个参保人的实际术前检查项目的合理性，判断该参保人住院行为的合理性；
即：
Figure PCTCN2019095050-appb-000003
其中，J i为第i个参保人的实际术前检查项目集与该预设检查项目集的杰卡德相似系数，M i11为第i个参保人的实际术前检查项目集在二元数据矩阵中为1且预设检查项目集 Ω在二元数据矩阵中也为1的总数量，M i10为第i个参保人的实际术前检查项目集在二元数据矩阵中为1且预设检查项目集Ω在二元数据矩阵中为0的总数量，M i01为第i个参保人的实际术前检查项目集在二元数据矩阵中为0且预设检查项目集Ω在二元数据矩阵中为1的总数量。
下面以一个实际的例子具体说计算杰卡德相似系数Jaccard的过程。
举例来说，假定该医疗机构的所有术前检查项目集{血常规、尿常规、心电图、血压、血糖、口腔}，且以a,b,c,d,e,f分别表示血常规、尿常规、心电图、血压、血糖、口腔，故得到所述医疗机构的所有术前检查项目集为{a,b,c,d,e,f}。在N＝4，如4个参保人的术前检查项目集具体为A＝{b,c,d}，B＝{a,b,c,d,e}，C＝{b,d}，D＝{d,f}时，且该4个人对应的预设术前检查项目集为Ω＝{a,d}，得到如图1B所示的二元数据矩阵，确定A与Ω的杰卡德相似系数J 1＝(1/(2+1+1))＝1/4，B与Ω的杰卡德相似系数J 2＝(2/(3+0+2))＝2/5，C与Ω的杰卡德相似系数J 3＝(1/(1+1+1))＝1/3，D与Ω的杰卡德相似系数J 4＝(1/(1+1+1))＝1/3，如第二阈值为0.6，则可确定集合C和D对应的术前检查项目正常，确定其对应的参保人的住院行为合理，集合A和B对应的术前检查项目异常，确定其对应的参保人的住院行为不合理。
可以看出，在本示例中，如预设时间段内接收到多个病因相同的参保人的住院数据时，先构建二元数据矩阵，依据二元数据矩阵确定各个参保人的住院行为的合理性，可批次处理多个参保人的住院数据，提高服务器的运算速度。
可选的，在一可能的示例中，所述方法还包括：
获取具有临床手术资格的各个科室，获取各个科室在预设时间段内的住院数据，确定各个科室的住院数据中实际术前检查项目异常的多个参保人，计算各个科室在该预设时间段内术前检查项目异常的参保人的总数量与各个科室在该预设时间段内住院的参保人的总数量的比例，依据该比例对各个科室在术前检查项目方面进行排序，调整排序在前五的科室在术前检查项目方面的管理策略。
可以看出，本示例中分析各个科室在预设时间段内的住院数据，得到在各个科室就诊的参保人术前费用异常的比例，根据该比例对各个科室进行排序得到排序结果，依据排序结果调整排名靠前的科室，从而针对性调整科室的住院制度，为医疗机构改革住院制度提供数据参考，有利于提高医疗机构的综合质量，减少各个科室采用住院模式套取医保基金支付术前检查费用的概率。
可选的，在一可能的示例中，所述方法还包括：
获取在预设时间段内住院数据中的实际术前检查项目异常的所有参保人，获取所述所有参保人的住院数据中所有医生信息，确定每个医生信息的出现频率，根据所述出现频率监管所述所有医生信息对应的医生，其具体包括：对出现频率排名前五的医生信息对应的医生采取定期抽查其所开取的住院数据单，如该住院数据单中的术前检查项目异常次数大于各自的预设阈值时，将该医生信息上传至住院管理系统，以监管该医生开取的住院数据单的合理性，其中，该预设阈值与医生排名相关，排名越靠前预设阈值越小。
可以看出，本示例中分析术前检查费用项目的所有参保人的住院数据，获取该住院数据中各个医生信息的出现频率，依据该出现频率监管医生在临床手术时开取的住院数据单，减少医生通过低标准将参保人收入住院来套取医保基金支付术前检查费用的概率。
参阅图2，图2为本申请实施例提供的另一种基于预测模型预测住院合理性的方法的流程示意图，该方法应用于电子设备，该方法包括步骤S201～S209中所示的内容：
步骤S201、接收输入的任意一个参保人的住院数据。
其中，所述住院数据中包括诊断数据和实际术前检查项目集。
步骤S202、提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率。
步骤S203、在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集。
步骤S204、将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述预设术前检查项目集与所述实际术前检查项目集的相似度。
步骤S205、如所述相似度大于第二阈值，确定所述实际术前检查项目正常，确定所述参保人的住院行为合理。
步骤S206、如所述相似度小于或者等于所述第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理。
步骤S207、在所述住院概率小于或者等于所述第一阈值时，向线上核对中心发送所述参保人的住院数据，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常。
其中，预先建立线上核对中心，在住院概率小于第一阈值时，表示该参保人很大程度上无需住院治疗，由于无需住院治疗，故再将实际术前检查项集与预设术前检查项目集比对，无法准确判断该住院行为的合理性，故将该参保人的住院数据发送至线上核对中心，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常，验证是否异常具体包括：如确定所述参保人无需住院，但仍存在术前检查项目时，确定该参保人的术前检查项目异常，如确定所述参保人需要住院，但该参保人的术前检查项目与该参保人所患疾病不对应时，确定该参保人的术前检查项目异常。
步骤S208、接收来自所述线上核对中心的验证结果，在验证结果为确认所述实际术前检查项目集正常时，确定所述参保人的住院行为合理，将所述参保人的实际术前检查项目集传输至与术前检查相关的网络设备。
可选的，将该参保人的实际术前检查项目集传输至与术前检查相关的网络设备，以便将该参保人的信息以及该参保人的术前检查项目录入该网络设备，确定该参保人在该实际检查项目方面为授权用户，以便该参保人在做术前检查项目时，可以在该网络设备中查询到该参保人的信息。
步骤S209、在验证结果为确认所述实际术前检查项目集异常时，确定所述参保人的住院行为不合理，禁止将所述参保人的实际术前检查项目集传输至所述网络设备，提示重新输入所述参保人的术前检查项目。
可以看出，在本申请实施例中，接收输入的参保人的住院数据，提取住院数据中的诊断数据，将该诊断数据输入住院预测模型得到该参保人的住院概率，在该住院概率大于第一阈值时，获取该参保人的预设术前检查项目集，将该预设检查项目集与该住院数据中的实际检查项目集比对，确定两者的相似度，依据相似度判断该参保人的实际术前检查项目集是否异常，确定该参保人住院行为的合理性，因此，在参保人需要住院的情况下，可通 过比较预设的术前检查项目集，判断该参保人是否存在套取医保基金支付术前检查费用的行为，基于模型判断住院概率，可实现针对性的判断住院概率满足条件的参保人，提高判断的准确度，而且，如该住院概率小于第一阈值时，将该住院数据转发至线上核对中心，以在线核对该参保人术前检查项目的合理性，因此，在该参保人无需住院的时候，可结合线上核对系统辅助核对该参保人住院的合理性，因此，两者的有机集合，可以完全解决套取医保基金支付术前检查费用的行为，增加了确定住院合理性的方式，为医疗体制改革提供数据参考，提高医疗体制改革的说服力，完善医疗制度。
参阅图3，图3为本申请实施例提供的另一种基于预测模型预测住院合理性的方法的流程示意图，该方法应用于电子设备，该方法包括步骤S301～S312中所示的内容：
步骤S301、接收输入的任意一个参保人的住院数据。
其中，所述住院数据中包括诊断数据和实际术前检查项目集。
步骤S302、提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率。
步骤S303、在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集。
步骤S304、将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述预设术前检查项目集与所述实际术前检查项目集的相似度。
步骤S305、如所述相似度小于或者等于第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理。
步骤S306、在所述住院概率小于或者等于所述第一阈值时，向线上核对中心发送所述参保人的住院数据，，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常。
步骤S307、接收来自所述线上核对中心的验证结果，在验证结果为确认所述实际术前检查项目集异常时，确定所述参保人的住院行为不合理。
步骤S308、统计所述医疗机构在预设时间段内住院行为不合理的参保人。
步骤S309、如所述参保人的数量为单个时，获取所述参保人的住院数据，从所述参保人的住院数据中提取所述参保人的实际术前检查项目集，将所述实际术前检查项目集标记为所述医疗机构的目标术前检查项目集。
其中，所述预设时间段具体为1星期、1个月、6个月、1年或者其他值。
步骤S310、如所述参保人的数量为多个时，获取所述多个参保人的住院数据，从所述多个参保人的住院数据中提取所述多个参保人中每一个参保人的实际术前检查项目集，将所述每一个参保人的实际术前检查项目集标记为一个事物集，得到所述多个参保人的多个事物集。
步骤S311、基于频繁模式增长FP-Growth算法确定所述多个事物集中的频繁项集，将所述频繁项集标记为所述医疗机构的目标术前检查项目集。
其中，所述目标术前检查项目集中的术前检查项目为所述医疗机构在术前检查方面易发生异常的术前检查项目。
可选的，基于FP-Growth算法确定所述多个事物集中的频繁项集具体包括：设置最小支持度P，基于该最小支持度P筛选该多个事物集中的多个频繁元素，利用该多个频繁元 素构建所述多个事物集的FP-tree，设置频繁项集中所需的元素数量，从该FP-tree中读取符合该元素数量的多个频繁项集，获取该多个频繁项集中支持度最大的频繁项集，如该支持度最大的频繁项集的数量为多个时，获取该多个频繁项集的交集，将该交集作为该目标术前检查项目集或者获取该多个频繁项集的并集，将该并集作为该目标术前检查项目集。
下面举例说明确定频繁项集和目标术前检查项目集的具体过程。
假定在预设时间段内获取到6个参保人的住院行为不合理，获得6个实际术前检查项目集，将该6个实际术前检查项目集分别进行编号为001、002、003、004、005、006，将该6个实际术前检查项目集中的元素(即术前检查项目)分别以字母表示，例如，001中的术前检查项目集分别为血常规、尿常规、心电图、血压、血糖、口腔，然后，将该血常规、尿常规、心电图、血压、血糖、口腔分别以字母r,z,h,j,p表示，故可得001＝{r,z,h,j,p}，基于得到001的方式，可对002、003、004、005和006的术前检查项目字母化，得到如表2所示的事物集。
表2
事物集编号 事物集 剔除后的事物集
001 {r,z,h,j,p} {r,z}
002 {z,y,x,w,v,u,t,s} {z,y,x,t,s}
003 {z} {z}，
004 {r,x,n,o,s} {r,x,s}
005 {y,r,x,z,q,t,p} {y,r,x,z,t}
006 {y,z,x,e,q,s,t,m} {y,z,x,s,t}
然后，开启第一轮扫描，设置第一最小支持度P1＝3，将001、002、003、004、005、006中元素出现次数小于支持度P1剔除,即剔除q、n、o、h、j、p、w、v、u和e，如表1所示，得到新的事物集{r,z}，{z,y,x,t,s}，{z}，{r,x,s}，{y,r,x,z,t}，{y,z,x,s,t}；开启第二轮扫描依次扫描新的事物集中的元素，如图3A所示，以空集null为根节点开始创建FP-tree，在扫描每一个新的事物集时依次往FP-tree中添加元素，如扫描{r,z}时，可添加元素r、z，扫描{z,y,x,t,s}时，可在第一次添加元素r、z得到的FP-tree中添加元素z,y,x,t,s，全部扫描完后，可得到如图3A最右方示出的FP-tree，其中，该FP-tree的树节点上给出集合中的单个元素及其在新的事物集中出现的总次数，路径的根节点的元素的出现次数示出该路径对应的序列的出现次数(即支持度)。如图3A所示，每个路径上的所有元素构成一个频繁项集，且该路径上的根节点示出该频繁项集的支持度；设置频繁项集中所需的元素数量，从该FP-tree的树节点截取与该元素数量对应的根节点，将该树节点与该根节点之间的元素组成频繁项集，例如，设置频繁项集中的元素数量为4时，可获得频繁项集{z,x,y,s}，{z,x,y,r}，故可将该频繁项集{z,x,y,s}，{z,x,y,r}的并集{z,x,y,r,s}标记为该6个事物集的最终频繁项集，即目标术前检查项目集；或者设置第二支持度P2，以第二支持度P2为基础进行第三轮扫描，扫描FP-tree中支持度大于或者等于P2的集合，将该集合作为频繁项集，例如P2＝3时，可得集合{z}，{z,x}和{z,x,y}的支持度分别为5、3和3，故可得{z}，{z,x}和{z,x,y}， 故可取{z}，{z,x}和{z,x,y}的并集{z,x,y}作为该6个事物集的最终频繁项集，即目标术前检查项目集。
步骤S312、向所述医疗机构对应的网络设备发送所述目标术前检查项目集，以调整所述医疗机构在所述目标术前检查方面的管理体制。
可以看出，在本申请实施例中，接收输入的参保人的住院数据，提取住院数据中的诊断数据，将该诊断数据输入住院预测模型得到该参保人的住院概率，在该住院概率大于第一阈值时，获取该参保人的预设术前检查项目集，将该预设检查项目集与该住院数据中的实际检查项目集比对，确定两者的相似度，依据相似度判断该参保人的实际术前检查项目集是否异常，确定该参保人住院行为的合理性，因此，在参保人需要住院的情况下，可通过比较预设的术前检查项目集，判断该参保人是否存在套取医保基金支付术前检查费用的行为，基于模型判断住院概率，可实现针对性的判断住院概率满足条件的参保人，提高判断的准确度，而且，如该住院概率小于第一阈值时，将该住院数据转发至线上核对中心，以在线核对该参保人术前检查项目的合理性，因此，在该参保人无需住院的时候，可结合线上核对系统辅助核对该参保人住院的合理性，因此，两者的有机集合，可以完全解决套取医保基金支付术前检查费用的行为，增加了确定住院合理性的方式，为医疗体制改革提供数据参考，提高医疗体制改革的说服力，完善医疗制度。而且，获取预设时间段住院行为不合理的多个参保人，依据该多个参保人的住院数据确定医疗机构的目标术前检查项目集，将该目标术前检查项目集反馈至网络侧设备，可实现针对性改革医疗体制，基于FP-Growth算法及时确定出该目标术前检查项目集，可提高医疗改革的效率。
与上述图1A、图2、图3所示的实施例一致的，请参阅图4，图4为本申请实施例提供的一种基于住院模型预测住院合理性的电子设备400的结构示意图，如图4所示，该电子设备400包括处理器、存储器、通信接口以及一个或多个程序，其中，上述一个或多个程序不同于上述一个或多个应用程序，且上述一个或多个程序被存储在上述存储器中，并且被配置由上述处理器执行，上述程序包括用于执行以下步骤的指令；
接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；
提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率；
在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；
将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
在一可能的示例中，上述程序中的指令还用于执行以下操作：
在所述住院概率小于或者等于所述第一阈值时，向线上核对中心发送所述参保人的住院数据，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常；
接收来自所述线上核对中心的验证结果；在验证结果为确认所述实际术前检查项目集正常时，确定所述参保人的住院行为合理，将所述参保人的实际术前检查项目集传输至与术前检查相关的网络设备；在验证结果为确认所述实际术前检查项目集异常时，确定所述参保人的住院行为不合理，禁止将所述参保人的实际术前检查项目集传输至所述网络设备，提示重新输入所述参保人的术前检查项目。
在一可能的示例中，在所述诊断数据中包括疾病名称、疾病严重等级、参保人的人口学参数，所述将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率方面，上述程序中的指令具体用于执行以下操作：
提取所述诊断数据中的疾病名称、疾病严重等级、参保人的人口学参数；
将所述疾病名称、疾病严重等级、人口学参数组成输入数据矩阵输入到所述预先训练好的住院预测模型执行正向运算，输出所述参保人的住院概率。
在一可能的示例中，在获取所述诊断数据对应的预设术前检查项目集方面，上述程序中的指令具体用于执行以下操作：
获取所述诊断数据中的疾病名称；
依据所述疾病名称和术前检查项目集的映射关系以及所述诊断数据中的疾病名称，获取所述诊断数据的预设术前检查项目集。
在一可能的示例中，在将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性方面，上述程序中的指令具体用于执行以下操作：
确定所述预设术前检查项目集与所述实际术前检查项目集的相似度，如所述相似度大于第二阈值，确定所述实际术前检查项目正常，确定所述参保人的住院行为合理，如所述相似度小于或者等于所述第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理；
所述相似度的计算公式如下：
Figure PCTCN2019095050-appb-000004
其中，S为所述预设术前检查项目集与所述实际术前检查项目集的相似度，A为所述预设术前检查项目集，B为所述实际术前检查项目集，Card(A∩B)为A∩B中的元素个数，Card(A∪B)为A∪B中的元素个数。
在一可能的示例中，上述程序中的指令还用于执行以下操作：
统计所述医疗机构在预设时间段内住院行为不合理的参保人；
如所述参保人的数量为单个时，获取所述参保人的住院数据，从所述参保人的住院数据中提取所述参保人的实际术前检查项目集，将所述实际术前检查项目集标记为所述医疗机构的目标术前检查项目集；
如所述参保人的数量为多个时，获取所述多个参保人的住院数据，从所述多个参保人的住院数据中提取所述多个参保人中每一个参保人的实际术前检查项目集，将所述每一个参保人的实际术前检查项目集标记为一个事物集，得到所述多个参保人的多个事物集，基于频繁模式增长FP-Growth算法确定所述多个事物集中的频繁项集，将所述频繁项集标记为所述医疗机构的目标术前检查项目集；
其中，所述目标术前检查项目集中的术前检查项目为所述医疗机构在术前检查方面易发生异常的术前检查项目。
在一可能的示例中，上述程序中的指令还用于执行以下操作：
向所述医疗机构对应的网络设备发送所述目标术前检查项目集，以调整所述医疗机构在所述目标术前检查方面的管理体制。
在一可能的示例中，上述程序中的指令还用于执行以下操作：
获取具有临床手术资格的各个科室；获取各个科室在预设时间段内的住院数据；确定 各个科室的住院数据中实际术前检查项目异常的参保人相对于在所述预设时间段内容参保人总数量的比例；依据所述比例对各个科室在术前检查项目方面进行排序，调整排序在前五的科室在术前检查项目方面的管理策略。
在一可能的示例中，上述程序中的指令还用于执行以下操作：
获取在预设时间段内住院数据中的实际术前检查项目异常的所有参保人；获取所述所有参保人的住院数据中所有医生信息；确定每个医生信息的出现频率，根据所述出现频率监管每个医生信息对应的医生。
参阅图5，图5示出了上述实施例中所涉及的基于住院模型预测住院合理性的电子设备500的一种可能的功能单元组成框图，电子设备500包括接收单元510、输入单元520、获取单元530、比对单元540、其中；
接收单元510，用于接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；
输入单元520，用于提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人对应的住院概率；
获取单元530，用于在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；
比对单元540，用于将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
在一可能的示例中，电子设备500还包括发送单元550；
其中，发送单元550，用于在所述住院概率小于或者等于所述第一阈值时，向线上核对中心发送所述参保人的住院数据，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常；以及用于接收来自所述线上核对中心的验证结果；在验证结果为确认所述实际术前检查项目集正常时，确定所述参保人的住院行为合理，将所述参保人的实际术前检查项目集传输至与术前检查相关的网络设备；在验证结果为确认所述实际术前检查项目集异常时，确定所述参保人的住院行为不合理，禁止将所述参保人的实际术前检查项目集传输至所述网络设备，提示重新输入所述参保人的术前检查项目。
在一可能的示例中，在诊断数据中包括疾病名称、疾病严重等级、参保人的人口学参数，在将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率时，输入单元520，具体用于：提取所述诊断数据中的疾病名称、疾病严重等级、参保人的人口学参数；以及用于将所述疾病名称、疾病严重等级、人口学参数组成输入数据矩阵输入到所述预先训练好的住院预测模型执行正向运算，输出所述参保人的住院概率。
在一可能的示例中，在获取所述诊断数据对应的预设术前检查项目集时，获取单元530，具体用于：依据所述疾病名称和术前检查项目集的映射关系以及所述诊断数据中的疾病名称，获取所述诊断数据的预设术前检查项目集。
在一可能的示例中，在将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性时，比对单元540，具体用于：确定所述预设术前检查项目集与所述实际术前检查项目集的相似度，如所述相似度大于第二阈值，确定所述实际术前检查项目正常，确定所述参保人的住院行为合理，如所述相似度小于或者等于所述第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理；
所述相似度的计算公式如下：
Figure PCTCN2019095050-appb-000005
其中，S为所述预设术前检查项目集与所述实际术前检查项目集的相似度，A为所述预设术前检查项目集，B为所述实际术前检查项目集，Card(A∩B)为A∩B中的元素个数，Card(A∪B)为A∪B中的元素个数。
在一可能的示例中，获取单元530，还用于：统计所述医疗机构在预设时间段内住院行为不合理的参保人；以及用于如所述参保人的数量为单个时，获取所述参保人的住院数据，从所述参保人的住院数据中提取所述参保人的实际术前检查项目集，将所述实际术前检查项目集标记为所述医疗机构的目标术前检查项目集；以及用于如所述参保人的数量为多个时，获取所述多个参保人的住院数据，从所述多个参保人的住院数据中提取所述多个参保人中每一个参保人的实际术前检查项目集，将所述每一个参保人的实际术前检查项目集标记为一个事物集，得到所述多个参保人的多个事物集，基于频繁模式增长FP-Growth算法确定所述多个事物集中的频繁项集，将所述频繁项集标记为所述医疗机构的目标术前检查项目集；其中，所述目标术前检查项目集中的术前检查项目为所述医疗机构在术前检查方面易发生异常的术前检查项目。
在一可能的示例中，发送单元550，还用于：向所述医疗机构对应的网络设备发送所述目标术前检查项目集，以调整所述医疗机构在所述目标术前检查方面的管理体制。
在一可能的示例中，电子设备500还包括调整单元560，调整单元560，用于获取具有临床手术资格的各个科室；获取各个科室在预设时间段内的住院数据；确定各个科室的住院数据中实际术前检查项目异常的参保人相对于在所述预设时间段内容参保人总数量的比例；依据所述比例对各个科室在术前检查项目方面进行排序，调整排序在前五的科室在术前检查项目方面的管理策略。
在一可能的示例中，电子设备500还包括调整单元560，调整单元560，用于获取在预设时间段内住院数据中的实际术前检查项目异常的所有参保人；获取所述所有参保人的住院数据中所有医生信息；确定每个医生信息的出现频率，根据所述出现频率监管每个医生信息对应的医生。
本申请实施例还提供一种计算机可读存储介质，其中，该计算机可读存储介质存储用于电子数据交换的计算机程序，该计算机程序使得计算机执行如上述方法实施例中记载的任何一种基于预测模型预测住院合理性的方法的部分或全部步骤。
本申请实施例还提供一种计算机程序产品，所述计算机程序产品包括存储了计算机程序的非瞬时性计算机可读存储介质，所述计算机程序可操作来使计算机执行如上述方法实施例中记载的任何一种基于预测模型预测住院合理性的方法的部分或全部步骤。
需要说明的是，对于前述的各方法实施例，为了简单描述，故将其都表述为一系列的动作组合，但是本领域技术人员应该知悉，本申请并不受所描述的动作顺序的限制，因为依据本申请，某些步骤可以采用其他顺序或者同时进行。其次，本领域技术人员也应该知悉，说明书中所描述的实施例均属于可选实施例，所涉及的动作和模块并不一定是本申请所必须的。
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。
在本申请所提供的几个实施例中，应该理解到，所揭露的装置，可通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件程序模块的形式实现。
所述集成的单元如果以软件程序模块的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储器中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储器中，包括若干指令用以使得一台计算机设备(可为个人计算机、服务器或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储器包括：U盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、移动硬盘、磁碟或者光盘等各种可以存储程序代码的介质。
本领域普通技术人员可以理解上述实施例的各种方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成，该程序可以存储于一计算机可读存储器中，存储器可以包括：闪存盘、只读存储器(英文：Read-Only Memory，简称：ROM)、随机存取器(英文：Random Access Memory，简称：RAM)、磁盘或光盘等。
以上对本申请实施例进行了详细介绍，本文中应用了具体个例对本申请的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本申请的方法及其核心思想；同时，对于本领域的一般技术人员，依据本申请的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本申请的限制。

Claims (20)

  1. 一种基于住院模型预测住院合理性的方法，其特征在于，所述方法应用于电子设备，所述方法包括：
    接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；
    提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率；
    在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；
    将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
  2. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    在所述住院概率小于或者等于所述第一阈值时，向线上核对中心发送所述参保人的住院数据，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常；
    接收来自所述线上核对中心的验证结果；
    在验证结果为确认所述实际术前检查项目集正常时，确定所述参保人的住院行为合理，将所述参保人的实际术前检查项目集传输至与术前检查相关的网络设备；在验证结果为确认所述实际术前检查项目集异常时，确定所述参保人的住院行为不合理，禁止将所述参保人的实际术前检查项目集传输至所述网络设备，提示重新输入所述参保人的术前检查项目。
  3. 根据权利要求1所述的方法，其特征在于，所述诊断数据中包括疾病名称、疾病严重等级、参保人的人口学参数，所述将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率具体包括：
    提取所述诊断数据中的疾病名称、疾病严重等级、参保人的人口学参数；
    将所述疾病名称、疾病严重等级、人口学参数组成输入数据矩阵输入到所述预先训练好的住院预测模型执行正向运算，输出所述参保人的住院概率。
  4. 根据权利要求2所述的方法，其特征在于，所述获取所述诊断数据对应的预设术前检查项目集具体包括：
    依据所述疾病名称和术前检查项目集的映射关系以及所述诊断数据中的疾病名称，获取所述诊断数据的预设术前检查项目集。
  5. 根据权利要求3所述的方法，其特征在于，所述将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性具体包括：
    确定所述预设术前检查项目集与所述实际术前检查项目集的相似度，如所述相似度大于第二阈值，确定所述实际术前检查项目正常，确定所述参保人的住院行为合理，如所述相似度小于或者等于所述第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理；
    所述相似度的计算公式如下：
    Figure PCTCN2019095050-appb-100001
    其中，S为所述预设术前检查项目集与所述实际术前检查项目集的相似度，A为所述预设术前检查项目集，B为所述实际术前检查项目集，Card(A∩B)为A∩B中的元素个数，Card(A∪B)为A∪B中的元素个数。
  6. 根据权利要求4所述的方法，其特征在于，所述方法还包括：
    统计所述医疗机构在预设时间段内住院行为不合理的参保人；
    如所述参保人的数量为单个时，获取所述参保人的住院数据，从所述参保人的住院数据中提取所述参保人的实际术前检查项目集，将所述实际术前检查项目集标记为所述医疗机构的目标术前检查项目集；
    如所述参保人的数量为多个时，获取所述多个参保人的住院数据，从所述多个参保人的住院数据中提取所述多个参保人中每一个参保人的实际术前检查项目集，将所述每一个参保人的实际术前检查项目集标记为一个事物集，得到所述多个参保人的多个事物集，基于频繁模式增长FP-Growth算法确定所述多个事物集中的频繁项集，将所述频繁项集标记为所述医疗机构的目标术前检查项目集；
    其中，所述目标术前检查项目集中的术前检查项目为所述医疗机构在术前检查方面易发生异常的术前检查项目。
  7. 根据权利要求6所述的方法，其特征在于，所述方法还包括：
    向所述医疗机构对应的网络设备发送所述目标术前检查项目集，以调整所述医疗机构在所述目标术前检查方面的管理体制。
  8. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    获取具有临床手术资格的各个科室；
    获取各个科室在预设时间段内的住院数据；
    确定各个科室的住院数据中实际术前检查项目异常的参保人相对于在所述预设时间段内容参保人总数量的比例；
    依据所述比例对各个科室在术前检查项目方面进行排序，调整排序在前五的科室在术前检查项目方面的管理策略。
  9. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    获取在预设时间段内住院数据中的实际术前检查项目异常的所有参保人；
    获取所述所有参保人的住院数据中所有医生信息；
    确定每个医生信息的出现频率，根据所述出现频率监管每个医生信息对应的医生。
  10. 一种基于住院模型预测住院合理性的电子设备，其特征在于，所述电子设备包括：
    接收单元，用于接收输入的任意一个参保人的住院数据，所述住院数据中包括诊断数据和实际术前检查项目集；
    输入单元，用于提取所述住院数据中的诊断数据，将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人对应的住院概率；
    获取单元，用于在所述住院概率大于第一阈值时，获取所述诊断数据对应的预设术前检查项目集；
    比对单元，用于将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性。
  11. 根据权利要求10所述的电子设备，其特征在于，所述电子设备还包括发送单元，所述发送单元，用于在所述住院概率小于或者等于所述第一阈值时，向线上核对中心发送所述参保人的住院数据，提示所述线上核对中心验证所述住院数据中的实际术前检查项目集是否异常；
    接收来自所述线上核对中心的验证结果；
    在验证结果为确认所述实际术前检查项目集正常时，确定所述参保人的住院行为合理，将所述参保人的实际术前检查项目集传输至与术前检查相关的网络设备；在验证结果为确认所述实际术前检查项目集异常时，确定所述参保人的住院行为不合理，禁止将所述参保人的实际术前检查项目集传输至所述网络设备，提示重新输入所述参保人的术前检查项目。
  12. 根据权利要求10所述的电子设备，其特征在于，所述诊断数据中包括疾病名称、疾病严重等级、参保人的人口学参数，
    在将所述诊断数据输入到预先训练好的住院预测模型，输出所述参保人的住院概率方面，所述输入单元，具体用于：
    提取所述诊断数据中的疾病名称、疾病严重等级、参保人的人口学参数；
    将所述疾病名称、疾病严重等级、人口学参数组成输入数据矩阵输入到所述预先训练好的住院预测模型执行正向运算，输出所述参保人的住院概率。
  13. 根据权利要求11所述的电子设备，其特征在于，
    在获取所述诊断数据对应的预设术前检查项目集方面，所述获取单元，具体用于：依据所述疾病名称和术前检查项目集的映射关系以及所述诊断数据中的疾病名称，获取所述诊断数据的预设术前检查项目集。
  14. 根据权利要求12所述的电子设备，其特征在于，
    在将所述实际术前检查项目集和所述预设术前检查项目集比对，确定所述参保人的住院行为的合理性方面，所述比对单元，具体用于：
    确定所述预设术前检查项目集与所述实际术前检查项目集的相似度，如所述相似度大于第二阈值，确定所述实际术前检查项目正常，确定所述参保人的住院行为合理，如所述相似度小于或者等于所述第二阈值，确定所述实际术前检查项目异常，确定所述参保人的住院行为不合理；
    所述相似度的计算公式如下：
    Figure PCTCN2019095050-appb-100002
    其中，S为所述预设术前检查项目集与所述实际术前检查项目集的相似度，A为所述预设术前检查项目集，B为所述实际术前检查项目集，Card(A∩B)为A∩B中的元素个数，Card(A∪B)为A∪B中的元素个数。
  15. 根据权利要求13所述的方法，其特征在于，
    所述获取单元，还用于统计所述医疗机构在预设时间段内住院行为不合理的参保人；
    如所述参保人的数量为单个时，获取所述参保人的住院数据，从所述参保人的住院数据中提取所述参保人的实际术前检查项目集，将所述实际术前检查项目集标记为所述医疗机构的目标术前检查项目集；
    如所述参保人的数量为多个时，获取所述多个参保人的住院数据，从所述多个参保人的住院数据中提取所述多个参保人中每一个参保人的实际术前检查项目集，将所述每一个参保人的实际术前检查项目集标记为一个事物集，得到所述多个参保人的多个事物集，基于频繁模式增长FP-Growth算法确定所述多个事物集中的频繁项集，将所述频繁项集标记为所述医疗机构的目标术前检查项目集；
    其中，所述目标术前检查项目集中的术前检查项目为所述医疗机构在术前检查方面易 发生异常的术前检查项目。
  16. 根据权利要求15所述的电子设备，其特征在于，
    所述发送单元，还用于向所述医疗机构对应的网络设备发送所述目标术前检查项目集，以调整所述医疗机构在所述目标术前检查方面的管理体制。
  17. 根据权利要求10所述的电子设备，其特征在于，所述电子设备还包括调整单元，所述调整单元，用于获取具有临床手术资格的各个科室；获取各个科室在预设时间段内的住院数据；确定各个科室的住院数据中实际术前检查项目异常的参保人相对于在所述预设时间段内容参保人总数量的比例；依据所述比例对各个科室在术前检查项目方面进行排序，调整排序在前五的科室在术前检查项目方面的管理策略。
  18. 根据权利要求10所述的电子设备，其特征在于，所述电子设备还包括调整单元，所述调整单元，用于获取在预设时间段内住院数据中的实际术前检查项目异常的所有参保人；获取所述所有参保人的住院数据中所有医生信息；确定每个医生信息的出现频率，根据所述出现频率监管每个医生信息对应的医生。
  19. 一种电子设备，其特征在于，包括处理器、存储器、通信接口以及一个或多个程序，其中，所述一个或多个程序被存储在所述存储器中，并且被配置由所述处理器执行，所述程序包括用于执行权利要求1-9任一项方法中的步骤的指令。
  20. 一种计算机可读存储介质，其特征在于，其用于存储计算机程序，其中，所述计算机程序被处理器执行，以实现如权利要求1-9任一项所述的方法。
PCT/CN2019/095050 2018-10-30 2019-07-08 基于预测模型预测住院合理性的方法及相关产品 WO2020087971A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811276926.8A CN109584086B (zh) 2018-10-30 2018-10-30 基于预测模型预测住院合理性的方法及相关产品
CN201811276926.8 2018-10-30

Publications (1)

Publication Number Publication Date
WO2020087971A1 true WO2020087971A1 (zh) 2020-05-07

Family

ID=65920849

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/095050 WO2020087971A1 (zh) 2018-10-30 2019-07-08 基于预测模型预测住院合理性的方法及相关产品

Country Status (2)

Country Link
CN (1) CN109584086B (zh)
WO (1) WO2020087971A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109584086B (zh) * 2018-10-30 2024-06-28 平安医疗健康管理股份有限公司 基于预测模型预测住院合理性的方法及相关产品
CN111785372A (zh) * 2020-05-14 2020-10-16 浙江知盛科技集团有限公司 基于关联规则的协同过滤疾病预测系统及其电子设备
CN113643815A (zh) * 2021-08-31 2021-11-12 平安医疗健康管理股份有限公司 疾病并发症的预测方法、装置、计算机设备和存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105005709A (zh) * 2015-08-19 2015-10-28 赵蒙海 基于单病种治疗过程的流程挖掘方法
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
CN109584086A (zh) * 2018-10-30 2019-04-05 平安医疗健康管理股份有限公司 基于预测模型预测住院合理性的方法及相关产品

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140108042A1 (en) * 2012-10-11 2014-04-17 Jayaram Reddy Methods and systems for determining health risk score, risk of hospitalization and forecasting healthcare cost
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
CN104408547B (zh) * 2014-10-30 2017-09-15 浙江网新恒天软件有限公司 一种基于数据挖掘的医保欺诈行为的检测方法
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
CN108573752A (zh) * 2018-02-09 2018-09-25 上海米因医疗器械科技有限公司 一种基于健康大数据的健康信息处理的方法及系统
CN108492196B (zh) * 2018-03-08 2020-11-10 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

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105005709A (zh) * 2015-08-19 2015-10-28 赵蒙海 基于单病种治疗过程的流程挖掘方法
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
CN109584086A (zh) * 2018-10-30 2019-04-05 平安医疗健康管理股份有限公司 基于预测模型预测住院合理性的方法及相关产品

Also Published As

Publication number Publication date
CN109584086B (zh) 2024-06-28
CN109584086A (zh) 2019-04-05

Similar Documents

Publication Publication Date Title
CN109545317B (zh) 基于住院预测模型判定住院行为的方法及相关产品
Muhlestein et al. Predicting inpatient length of stay after brain tumor surgery: developing machine learning ensembles to improve predictive performance
Hull Estimating hospital quality with quasi-experimental data
CN110136788B (zh) 一种基于自动检测的病历质检方法、装置、设备及存储介质
US20200160942A1 (en) Automatic patient recruitment system
Allaudeen et al. Inability of providers to predict unplanned readmissions
US20190043606A1 (en) Patient-provider healthcare recommender system
JP2022523621A (ja) コミュニティデータの集約、完成、修正、および使用
Delen et al. An analytic approach to better understanding and management of coronary surgeries
WO2020087971A1 (zh) 基于预测模型预测住院合理性的方法及相关产品
TW202004636A (zh) 保險服務優化方法、系統及電腦程式產品
Zimmerman et al. A history of outcome prediction in the ICU
US20110282695A1 (en) Methods and systems for fraud detection
US10546654B2 (en) Method and system for intelligent completion of medical record based on big data analytics
Pope et al. Financial and quality impacts of the Medicare physician group practice demonstration
US20210257067A1 (en) State transition prediction device, and device, method, and program for learning predictive model
US11537888B2 (en) Systems and methods for predicting pain level
Chattopadhyay et al. VIRDOCD: A VIRtual DOCtor to predict dengue fatality
Krause Director of nursing current job tenure and past experience and quality of care in nursing homes
Tchetchik et al. From research to practice: Which research strategy contributes more to clinical excellence? Comparing high-volume versus high-quality biomedical research
Cole et al. Assessing the robustness of mixture models to measurement noninvariance
CN116228440B (zh) 一种医保智能管理方法、系统、装置和介质
US20150339602A1 (en) System and method for modeling health care costs
Syed et al. Toolkit to compute time-based Elixhauser comorbidity indices and extension to common data models
US20170186120A1 (en) Health Care Spend Analysis

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: 19880292

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 1205A DATED 31.08.2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19880292

Country of ref document: EP

Kind code of ref document: A1