Unreasonable medical examination item monitoring method and system

Abstract

The application discloses an unreasonable medical examination item monitoring method and system, and the method comprises the following steps: extracting medical examination items, establishing a first observation list, a second observation list and a third observation list, adding the medical examination items of a medical institution to the first observation list, and adding the medical examination items triggering unreasonable charging rules to the second observation list; generalizing and expanding the detail data of the medical examination items associated with the second observation list and training a classifier to obtain a dynamic unreasonable charging decision forest model; using the dynamic unreasonable charging decision forest model to mark and identify the detail data of the examination items associated with the second observation list, when the marking result is higher than a predetermined probability value, determining that the corresponding medical examination item is unreasonable charging, and adding the medical examination item with unreasonable charging to the third observation list for real-time monitoring and early warning.

Description

Technical Field

[0001] This invention relates to the field of medical technology, and in particular to a method and system for monitoring unreasonable medical examination items. Background Technology

[0002] Medical fees have increasingly drawn public attention, with some medical institutions prioritizing economic benefits, leading to a worsening of unreasonable medical examination charges. These include issues such as issuing multiple bills, bundled charges, duplicate charges, and hidden fees. Summary of the Invention

[0003] This application provides a method and system for monitoring unreasonable medical examination items, which solves the problem of relying on manual judgment of reasonableness in the prior art, which is highly subjective and lacks credibility. The monitoring scheme for unreasonable medical examination items in this application effectively solves the current problem of difficulty in supervising unreasonable fee-charging examination items.

[0004] In a first aspect, embodiments of this application provide a method for monitoring unreasonable medical examination items, the method comprising: S1: Establish a primary observation list of unreasonable chargeable examination items, take all current medical institutions as monitoring objects, use statistical analysis and probability distribution methods to extract the medical examination items, and add them to the primary observation list; S2: Pre-configure unreasonable charging rules, establish a secondary observation list of unreasonable charging examination items, and add the medical examination items that trigger the unreasonable charging rules in the primary observation list to the secondary observation list; S3: Based on the medical examination items in the secondary observation list, extract the detailed data of the associated medical examination items. After data anonymization, label the quantitative detailed data with unreasonable charges to obtain a training data set L0. Use a pre-established generalization rule set F to generalize and expand the training data set L0 to obtain a training data set L1. Use a random forest as a classifier RF to train the classifier in the training data set L1 to obtain a dynamic unreasonable charge decision forest model. S4: Establish a three-level observation list of unreasonable-charged examination items, obtain the anonymized examination item details associated with the two-level observation list, and use the dynamic unreasonable-charged decision forest model for labeling and identification. When the labeling result is higher than the predetermined probability value, the corresponding medical examination item is determined to be unreasonable-charged, and the unreasonable-charged medical examination item is added to the three-level observation list and monitored and warned in real time.

[0005] In one embodiment, the unreasonable charging rule in step S2 includes: obtaining the guidance price of the first medical examination item based on the first medical examination item associated with each medical institution in the primary observation list, comparing and analyzing the guidance price with the actual price; when the actual price of any first medical examination item is higher than the limited increase of the guidance price, it is determined to be an unreasonable charge, and the first medical examination item with unreasonable charges is added to the secondary observation list.

[0006] In one embodiment, the unreasonable charging rule in step S2 further includes sorting the unrelated second medical examination items in each medical institution from high to low according to the actual total price of multiple charges and the price of a single charge, and adding the second medical examination items with a predetermined proportion at the top of the sequence to the secondary observation list, wherein the total price of multiple charges is expressed as the product of the number of times any second medical examination item is examined and the price of a single charge.

[0007] In one embodiment, in step S1, the method for extracting the medical examination items using statistical analysis and probability distribution methods is as follows: based on the medical examination item information generated in all monitored medical institutions, the detailed charges and number of examinations for each medical examination item are statistically analyzed; the charges and number of examinations for a single medical examination item are sorted; the charge sequence and number of examinations for each medical examination item in each level of medical institution are obtained; the probability distribution dispersion of each charge sequence and number of examinations sequence is calculated; the coefficient of variation is normalized; and the sequences are sorted according to the coefficient of variation. The medical examination items that rank higher according to a predetermined proportion are added to the first-level observation list.

[0008] In one embodiment, a formula is used. Calculate the coefficient of variation (CV) for all charge sequences and examination frequency sequences, where SD is the standard deviation and MN is the mean.

[0009] In one embodiment, the coefficient of variation can be used to eliminate the difficulty in comparing the degree of dispersion of a sequence due to differences in mean, length, and units, and to measure the degree of dispersion of the charging sequence and the number of inspections sequence.

[0010] In one embodiment, the dynamic unreasonable charging decision forest model in step S3 includes a decision forest and a decision engine. The decision forest is dynamically updated to identify and update regulatory rules in real time. The decision forest automatically identifies and adds decision trees or decision branches. The decision engine automatically performs the decision forest decision identification process on the details of medical examination items and the examination objects, and judges whether the current medical examination item charge is reasonable based on the decision identification result value.

[0011] In one embodiment, the dynamic unreasonable charging decision forest model in step S3 uses the random forest algorithm as the basic ensemble classifier of the decision forest model; the training dataset L1 is divided into regions, and random forests are trained in each region to establish multiple forests to adapt to different regions, policies, prices and medical service market demands.

[0012] In one embodiment, the feature information L1 of the training data set in step S3 includes, but is not limited to, the region, medical institution level, age of the examinee, gender of the examinee, symptoms of the examinee, examination time, name of examination item, examination fee, medical equipment, examination site, consumables used, detailed breakdown of fee items and costs, and regional guidance price of examination items. The feature information is classified using number identifiers, represented by normalized values, and represented by time features using time period inclusion values.

[0013] Secondly, embodiments of this application provide a monitoring system for unreasonable medical examination items, employing the method described in any one of the first aspects, the system comprising: The primary data processing module is configured to establish a primary observation list of unreasonable chargeable examination items, taking all current medical institutions as monitoring objects, extracting the medical examination items using statistical analysis and probability distribution methods, and adding them to the primary observation list; The secondary data processing module is configured to pre-configure unreasonable charging rules, establish a secondary observation list of unreasonable charging examination items, and add the medical examination items that trigger the unreasonable charging rules in the primary observation list to the secondary observation list. The decision model training module is configured to extract the associated medical examination item fee details data based on the medical examination items in the secondary observation list, and after data anonymization, label the quantitative fee details data as unreasonable charges to obtain a training data set L0; generalize and expand the training data set L0 using a pre-established generalization rule set F to obtain a training data set L1; and train the training data set L1 using a random forest as a classifier RF to obtain a dynamic unreasonable charge decision forest model. The third-level data processing module is configured to establish a third-level observation list of unreasonable-charged examination items, obtain the anonymized detailed data of the examination items associated with the second-level observation list, and use the dynamic unreasonable-charged decision forest model for labeling and identification. When the labeling result is higher than a predetermined probability value, the corresponding medical examination item is determined to be unreasonable-charged, and the unreasonable-charged medical examination item is added to the third-level observation list and monitored and warned in real time.

[0014] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor, wherein the memory stores instructions, and the memory and the processor are interconnected via a circuit; the processor invokes the instructions in the memory to implement the method for monitoring unreasonable medical examination items as described in any one of the first aspects.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for monitoring unreasonable medical examination items as described in any of the first aspects.

[0016] The technical solutions provided in this application embodiment have at least the following technical effects: This invention employs statistical analysis and probability distribution methods to extract medical examination items and obtain associated medical examination items. It utilizes primary, secondary, and tertiary observation lists to arrange medical examination items under different scenarios, thereby solving the problems of cost data dispersion and arbitrary pricing in medical examination items. Furthermore, it uses regional guidance prices to address price differences across different regions. A dynamic unreasonable charging decision forest model is used to identify medical examination items with unreasonable charges that ultimately require regulation. It also supports manually adding regulatory rules for dynamic regulation, achieving multi-faceted oversight and effectively solving the current difficulties in regulating unreasonable charging examination items.

[0017] 2. This invention uses decision trees for spatial partitioning and random forests to generate multiple trees for voting decisions, avoiding the bias of single decisions. The generation of multiple forests also fits different regional special conditions and scenarios. In addition, the dynamic unreasonable charging decision forest model also takes into account the application of regulatory policies, that is, it leverages the theoretical algorithm model's ability to automatically identify and process practical problems, while also taking into account the regulatory authorities' supervision of the services provided by medical institutions, further optimizing the medical service market. Attached Figure Description

[0018] Figure 1 This is a flowchart of the monitoring method for unreasonable medical examination items in Embodiment 1 of this application; Figure 2 This is a block diagram of the monitoring system for unreasonable medical examination items in Embodiment 2 of this application; Figure 3 This is a block diagram of the electronic device in Embodiment 3 of this application. Detailed Implementation

[0019] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0020] Example 1 Reference Appendix Figure 1 As shown in the figure, this application provides a method for monitoring unreasonable medical examination items, which includes the following steps.

[0021] Step S1: Establish a primary observation list of unreasonable examination items, taking all current medical institutions as monitoring targets, extracting medical examination items using statistical analysis and probability distribution methods, and adding them to the primary observation list.

[0022] In step S1, the method for extracting medical examination items using statistical analysis and probability distribution methods is as follows: based on the medical examination item information generated in the associated medical institutions, the detailed charges and number of examinations for each medical examination item are statistically analyzed; the charges and number of examinations for a single medical examination item are sorted to obtain the charge sequence and examination frequency sequence for each medical examination item of each level of medical institutions; the coefficient of variation, a normalized measure of the dispersion of the probability distribution of each charge sequence and examination frequency sequence, is calculated; the sequences are sorted according to the coefficient of variation; and the medical examination items that rank higher according to a predetermined proportion are added to the first-level observation list.

[0023] Using formula Calculate the coefficient of variation (CV) for all billing sequences and examination frequency sequences, where SD is the standard deviation and MN is the mean. The coefficient of variation eliminates the difficulty in comparing the degree of dispersion caused by differences in mean, length, and units of measurement of data sets, and measures the dispersion of billing sequences and examination frequency sequences.

[0024] For example, the process involves acquiring medical examination items generated by medical institutions of various levels across different regions. After data anonymization, detailed examination data is obtained. The charges and frequency of each single examination item and examination for each body part are statistically analyzed, and then sorted by frequency to obtain the charging sequences for each medical examination item and examination for each body part at each level of medical institution. The coefficient of variation (COP) is calculated to normalize the dispersion of the probability distribution of each sequence, and the sequences are sorted according to their COP values. The top 20% of the ranked examination item sequences and examination body part sequences are selected as a preliminary observation list of unreasonable charges. Specifically, for the same region, statistical data is collected by medical institution level and by examination item. For example, the fee schedule for a CT scan of the brain at a Grade A tertiary institution in Province S is: [500.0, 500.0, 460.5, 460.5, 460.5, 460.5, 460.5, 460.5, 300.0, 300.0...263.2, 263.2...215.8, 215.8...]; the fee schedule for a CT scan of the brain at a Grade B tertiary institution in Province S is: [480.0, 480.0, 460.5, 460.5, 460.5, 460.5, 460.5, 460.5, 300.0, 300.0...263.2, 263.2...255.8, 255.8...]. Further using the formula... Calculate the coefficient of variation (CV) for all sequences, and obtain: Then, sort the items by CV from largest to smallest, and take the top 20% of the items to place in the primary observation list. An example of the primary observation list is shown in the table below: Therefore, it can be seen that the reason for using the coefficient of variation to measure the dispersion of each sequence is to eliminate the difficulty in comparing the dispersion caused by different means, different sequence lengths, or even different units.

[0025] Step S2: Pre-configure unreasonable charging rules, establish a secondary observation list of unreasonable charging examination items, and add medical examination items that trigger unreasonable charging rules from the primary observation list to the secondary observation list.

[0026] The unreasonable charging rules in step S2 include: obtaining the guide price for the first medical examination item related to each medical institution in the primary observation list, and comparing and analyzing the guide price with the actual price; when the actual price of any first medical examination item exceeds the limit increase of the guide price, it is judged as unreasonable charging, and the first medical examination item with unreasonable charging is added to the secondary observation list. The unreasonable charging rules in step S2 also include sorting the non-related second medical examination items in each medical institution from high to low according to the actual total price of multiple charges and the price per charge, and adding the second medical examination items with a predetermined proportion at the top of the sequence to the secondary observation list. The total price of multiple charges is expressed as the product of the number of times any second medical examination item is performed and the price per charge. For example, for some common medical examination items in several medical institutions in the same area, when the actual price of a certain examination item is higher than the guide price by 20%, then that medical examination item will be added to the secondary observation list as a key observation list. For example, for certain specialized medical examination items at medical institutions—that is, self-funded items without guiding prices—the top 20% of these items will be added to a secondary observation list as a key observation list, ranked according to their prices. Furthermore, for unreasonable instances of duplicate examinations, the top 20% of the total examination fees (the product of the unit price and the number of examinations) will be added to a secondary observation list as a key observation list.

[0027] Step S3: Based on the medical examination items in the secondary observation list, extract the detailed data of the associated medical examination items. After data anonymization, label the quantitative detailed data with unreasonable charges to obtain the training dataset L0. Use the pre-established generalization rule set F to generalize and expand the training dataset L0 to obtain the training dataset L1. Use Random Forest as the classifier RF to train the classifier on the training dataset L1 to obtain the dynamic unreasonable charge decision forest model.

[0028] For example, the training data set L0 is generalized and expanded using a pre-established set of generalization rules F. In L0, if a test subject has undergone hepatitis B surface antigen testing more than once within 5 days, it is marked as unreasonable charging. The hepatitis B surface antigen testing is further generalized to: any test for hepatitis B, hepatitis C, hepatitis A, etc., more than once within 5 days is judged as unreasonable charging.

[0029] To further clarify, the detailed data for medical examination items includes, but is not limited to, the identity details of the examinee, the location and level of the medical institution, and the examination details.

[0030] The identity details of the inspected individuals include, but are not limited to, their name, ID, contact information, mailing address, gender, age, and existing medical conditions. The name and ID of the inspected individuals are considered sensitive private data. In this embodiment, when monitoring unreasonable charges, the name and ID of the monitored individuals will be anonymized and can be directly replaced with codes. For example, if the inspected individual's "Name: Zhang San, ID: 321321******1234, Contact Information: 123**123, Mailing Address: Room D, No. C, B Street, S City" can be represented as "A123123". This protects the privacy of the inspectors while still allowing the extraction of the necessary data for the inspection.

[0031] The location and level of a medical institution can be understood as the region to which the institution belongs and its level. The charges for medical examinations vary between different regions or between different levels of medical institutions within the same region. The following is a breakdown of the charges for a CT scan of the brain at a Grade III Class A medical institution in a certain province: [500.0, 500.0, 460.5, 460.5, 460.5, 460.5, 460.5, 460.5, 300.0, 300.0...263.2, 263.2...2] The fee schedule for a CT scan of the brain at a Class B tertiary medical institution is as follows: [480.0, 480.0, 460.5, 460.5, 460.5, 460.5, 460.5, 460.5, 300.0, 300.0...263.2, 263.2...255.8, 255.8...]. Therefore, in this embodiment, when an unreasonable medical examination is performed, it is necessary to make reasonable judgments and limits based on the different regions and levels of medical institutions.

[0032] The inspection details include, but are not limited to, inspection time, project name, medical equipment, inspection fee, inspection site, consumables used, detailed breakdown of fee items and costs, and regional guidance price for inspection items. To simplify the processing flow, this embodiment classifies and represents the feature information other than time and numerical value in the form of coded identifiers, and uses a normalization method to unify the numerical format for numerical feature representation. For time features, the form of time period containing values ​​is used for feature representation.

[0033] The dynamic unreasonable charging decision forest model in step S3 includes a decision forest and a decision engine. The decision forest can be understood as numerous decision trees. The decision forest is dynamically updated to identify and update regulatory rules in real time, automatically identifying and adding decision trees or decision branches. The decision engine automatically executes the decision forest's decision identification process on the details of medical examination items and the examination subjects, judging the reasonableness of the current medical examination charge based on the decision identification result. To further explain, by dynamically updating the decision forest, regulatory rules can be identified and updated in real time. For example, if the cost of various medical examination items already includes the cost of fixed-use consumables, but the medical institution still charges for consumables again, this is considered unreasonable charging. The decision forest automatically identifies and adds decision trees or decision branches. Taking a decision branch as an example, since the internal data storage of the decision tree uses feature values ​​or feature numbers, for example, a decision branch storage rule might be: Province S -> Grade A Tertiary Hospital -> Blood Routine Examination -> Disposable Blood Collection Device: Charge -> 0.9. If a decision branch shows that a blood routine examination is performed on a subject at a Grade A Tertiary Hospital in Province S, and the cost of a disposable blood collection device is charged, an alert is triggered.

[0034] Since many medical examination items are unnecessary, the corresponding charges are unreasonable. Therefore, decision forests can be used to identify and add decision trees or decision branches to determine whether the charges for each examination item are reasonable. Further, using decision tree JX1003 as an example, one decision branch storage rule is: 1: Province S -> Grade III Class B -> 5*24-hour range -> more than 1 hepatitis B surface antigen test -> 0.9; another decision branch storage rule is: 2: Province S -> Grade II Class A -> 5*24-hour range -> more than 1 hepatitis C antibody test -> 0.9; using decision tree JS993 as an example, the decision branch storage rule is: Province M -> Grade II -> Infusion -> Rh blood type identification -> 0.65, etc.

[0035] The dynamic unreasonable charging decision forest model in this embodiment also supports manually adding regulatory rules. For example, an organization in province XX requires that certain consumables in medical examinations within the province should not be charged separately. Manual addition can be configured as follows: (A. Inspection, a. Consumables, charge, 1.0) The storage rules for the decision tree automatically generated internally by the dynamic unreasonable charging decision forest model are as follows: xx1001, xx province -> Level 1 -> A inspection -> a consumable: charge -> 1.0, xx province -> Level 2 -> A inspection -> a consumable: charge -> 1.0, ... xx province -> Level 3 Grade A -> A inspection -> a consumable: charge -> 1.0.

[0036] In step S3, the dynamic unreasonable charging decision forest model uses the random forest algorithm as the basic ensemble classifier for the decision forest model. The training dataset L1 is divided into regions, and random forests are trained in each region to establish multiple forests to adapt to different regions, policies, prices, and medical service market demands.

[0037] The feature information of the training dataset L1 in step S3 includes, but is not limited to, the region, medical institution level, age of the examinee, gender of the examinee, symptoms of the examinee, examination time, examination item name, examination fee, medical equipment, examination site, consumables used, detailed breakdown of fees and costs, and regional guidance price for examination items. Features are classified using identification numbers, represented numerically using normalized values, and represented temporally using time period inclusion values. Further, in the time feature represented by time period inclusion values, if the examination time is divided into a 5-day cycle, the value is the same for examinations within adjacent 5 days. In this embodiment, during feature processing, the feature information corresponding to the medical detail data of each medical examination item is extracted and quantified. Only after the feature information is processed is it used as training data for the model and input into the dynamic unreasonable charging decision forest model to train the random forest model.

[0038] The random forest used in this embodiment is an ensemble learning method in the field of machine learning. It's a machine learning algorithm published in 2001 by Leo Breiman, combining his "Bootstrap aggregating" idea proposed in 1996 and Ho's "random subspace method." It uses decision trees as basic classifiers, with each decision tree acting as a classifier. For an input sample, N trees will produce N classification results. Therefore, random forest integrates all classification voting results, designating the class with the most votes as the final output, overcoming the overfitting problem caused by most classifiers.

[0039] The data details of medical examination items are extensive, the medical service process is complex, patient conditions are varied, and medical service price levels differ across regions. Therefore, a linear classification is used to determine whether the charges for medical examination items are reasonable or unreasonable. To further explain, this embodiment uses a decision tree for spatial partitioning and a random forest to generate multiple trees for voting decisions, avoiding bias from single decisions. The generated multiple forests also fit different regional specificities and scenarios. Furthermore, the dynamic unreasonable charging decision forest model also considers the practical application of regulatory policies. This approach leverages the theoretical algorithm model's ability to automatically identify and process real-world problems while also considering the regulatory oversight of medical institution services, further optimizing the medical service market.

[0040] Step S4: Establish a three-level observation list of unreasonable-charged examination items, obtain the anonymized examination item details associated with the two-level observation list, and use the dynamic unreasonable-charged decision forest model for labeling and identification. When the labeling result is higher than the predetermined probability value, the corresponding medical examination item is judged to be unreasonable-charged, and the unreasonable-charged medical examination item is added to the three-level observation list and monitored and warned in real time.

[0041] In this embodiment, the predetermined probability value is 0.55. This means that based on the data details associated with the secondary observation list, the dynamic unreasonable charging decision forest model is used for identification. If the decision identification result is greater than 0.55, it indicates that the medical examination item is overcharged unreasonably, and an alert is issued for this charging situation. The medical examination item is then added to the tertiary observation list, which is the final list of medical examination items with unreasonable charges. In actual application, once each examination subject (patient)'s medical examination item matches the secondary observation list, the patient's examination item charging information and the patient's anonymized data are automatically pushed into the dynamic unreasonable charging decision forest model. Once the decision identifies unreasonable charging, a dynamic monitoring alert is immediately triggered.

[0042] Example 2 Reference Appendix Figure 2 As shown, this application provides a monitoring system for unreasonable medical examination items, which adopts any one of the unreasonable medical examination item monitoring methods in Embodiment 1. The system includes: a primary data processing module 100, a secondary data processing module 200, a decision model training module 300, and a tertiary data processing module 400.

[0043] The primary data processing module 100 is configured to establish a primary observation list of unreasonable chargeable examination items, taking all current medical institutions as monitoring objects, extracting medical examination items using statistical analysis and probability distribution methods, and adding them to the primary observation list; The secondary data processing module 200 is configured to pre-configure unreasonable charging rules, establish a secondary observation list of unreasonable charging examination items, and add medical examination items that trigger unreasonable charging rules from the primary observation list to the secondary observation list. The decision model training module 300 is configured to extract the detailed billing data of related medical examination items from the secondary observation list, and after data anonymization, label the quantitative detailed billing data with unreasonable charges to obtain the training dataset L0; the training dataset L0 is generalized and expanded using a pre-established set of generalization rules F to obtain the training dataset L1; and the training dataset L1 is trained using a random forest as a classifier RF to obtain the dynamic unreasonable charge decision forest model. The third-level data processing module 400 is configured to establish a third-level observation list of unreasonable-charged examination items, obtain the anonymized detailed data of examination items associated with the second-level observation list, and use a dynamic unreasonable-charged decision forest model for labeling and identification. When the labeling result is higher than the predetermined probability value, the corresponding medical examination item is determined to be unreasonable-charged, and the unreasonable-charged medical examination item is added to the third-level observation list for real-time monitoring and early warning.

[0044] Example 3 This application provides an electronic device, characterized in that it includes a memory and a processor, the memory storing instructions, and the memory and processor being interconnected via a circuit; the processor calls the instructions in the memory to implement the method for monitoring unreasonable medical examination items as described in any of the embodiments in Example 1.

[0045] The electronic device 500 can vary considerably depending on its configuration or performance, and may include one or more central processing units (CPUs) 510 (e.g., one or more processors) and memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) for storing application programs 533 or data 532. The memory 520 and storage media 530 may be temporary or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the figure), each module may include a series of instructions for operating on the electronic device 500.

[0046] Furthermore, the processor 510 can be configured to communicate with the storage medium 530 and execute a series of instructions stored in the storage medium 530 on the device 500 for establishing an observation list of unreasonable medical examination items.

[0047] Electronic device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Vista, etc.

[0048] Those skilled in the art will understand that Figure 3 The illustrated electronic device structure does not constitute a limitation on the electronic device and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0049] Example 4 This application provides a computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements a method for monitoring unreasonable medical examination items as described in any of the embodiments in Example 1.

[0050] The computer-readable storage medium can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the irrational medical examination monitoring method in Embodiment 1.

[0051] If the monitoring method for unreasonable medical examination items is implemented in the form of program instructions and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in software. This computer software is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0053] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from the scope of the invention. The spirit and scope of the invention are as follows: Thus, if these modifications and variations of the invention fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.

Claims

1. A method for monitoring unreasonable medical examination items, characterized in that, The method includes: S1: Establish a primary observation list of unreasonable-charged examination items, taking all current medical institutions as monitoring objects, extracting the medical examination items using statistical analysis and probability distribution methods, and adding them to the primary observation list; S2: Pre-configure unreasonable charging rules, establish a secondary observation list of unreasonable charging examination items, and add the medical examination items that trigger the unreasonable charging rules in the primary observation list to the secondary observation list; S3: Based on the medical examination items in the secondary observation list, extract the detailed data of the associated medical examination items. After data anonymization, label the quantitative detailed data with unreasonable charges to obtain a training data set L0. Use a pre-established generalization rule set F to generalize and expand the training data set L0 to obtain a training data set L1. Use a random forest as a classifier RF to train the classifier in the training data set L1 to obtain a dynamic unreasonable charge decision forest model. S4: Establish a three-level observation list of unreasonable-charged examination items, obtain the anonymized examination item details associated with the two-level observation list, and use the dynamic unreasonable-charged decision forest model for labeling and identification. When the labeling result is higher than the predetermined probability value, the corresponding medical examination item is determined to be unreasonable-charged, and the unreasonable-charged medical examination item is added to the three-level observation list and monitored and warned in real time.

2. The method for monitoring unreasonable medical examination items as described in claim 1, characterized in that, The unreasonable charging rule in step S2 includes: obtaining the guidance price of the first medical examination item associated with each medical institution in the first-level observation list, comparing and analyzing the guidance price with the actual price; when the actual price of any first medical examination item is higher than the limit increase of the guidance price, it is determined to be an unreasonable charge, and the first medical examination item with unreasonable charges is added to the second-level observation list.

3. The method for monitoring unreasonable medical examination items as described in claim 1, characterized in that, The unreasonable charging rules in step S2 also include sorting the unrelated second medical examination items in each medical institution from high to low according to the actual total price of multiple charges and the price of a single charge, and adding the second medical examination items with a predetermined proportion at the top of the sequence to the secondary observation list. The total price of multiple charges is expressed as the product of the number of times any second medical examination item is examined and the price of a single charge.

4. The method for monitoring unreasonable medical examination items as described in claim 1, characterized in that, In step S1, the method for extracting the medical examination items using statistical analysis and probability distribution methods is as follows: based on the medical examination item information already generated in the associated medical institutions, the detailed charges and number of examinations for each medical examination item are statistically analyzed; the charges and number of examinations for a single medical examination item are sorted to obtain the charge sequence and examination frequency sequence for each medical examination item of each level of medical institutions; the normalized measure of the dispersion of the probability distribution of each charge sequence and examination frequency sequence is calculated as the coefficient of variation; the sequences are sorted according to the coefficient of variation; and the medical examination items that rank higher according to a predetermined proportion are added to the first-level observation list.

5. The method for constructing a form for unreasonable medical examination items as described in claim 4, characterized in that, Using formula Calculate the coefficient of variation (CV) for all charge sequences and examination frequency sequences, where SD is the standard deviation and MN is the mean.

6. The method for monitoring unreasonable medical examination items as described in claim 4, characterized in that, The coefficient of variation can be used to eliminate the difficulty in comparing the degree of dispersion of a sequence due to differences in mean, length, and units, and to measure the degree of dispersion of the charging sequence and the number of inspections sequence.

7. The method for monitoring unreasonable medical examination items as described in claim 1, characterized in that, The dynamic unreasonable charging decision forest model in step S3 includes a decision forest and a decision engine. The decision forest is dynamically updated to identify and update regulatory rules in real time. The decision forest automatically identifies and adds decision trees or decision branches. The decision engine automatically performs the decision forest decision identification process on the details of medical examination items and the examination objects, and judges whether the current medical examination item charges are reasonable based on the decision identification result value.

8. The method for monitoring unreasonable medical examination items as described in claim 7, characterized in that, The dynamic unreasonable charging decision forest model in step S3 uses the random forest algorithm as the basic ensemble classifier of the decision forest model; the training dataset L1 is divided into regions, and random forests are trained in each region to establish multiple forests to adapt to different regions, policies, prices and medical service market demands.

9. The method for monitoring unreasonable medical examination items as described in claim 1, characterized in that, The feature information of the training data set L1 in step S3 includes, but is not limited to, the region, medical institution level, age of the examinee, gender of the examinee, symptoms of the examinee, examination time, name of examination item, examination fee, medical equipment, examination site, consumables used, detailed breakdown of fee items and costs, and regional guidance price of examination items. The feature information is classified using number identifiers, represented by normalized values, and represented by time features using time period inclusion values.

10. A monitoring system for unreasonable medical examination items, employing the method described in any one of claims 1-9, characterized in that, The system includes: The primary data processing module is configured to establish a primary observation list of unreasonable chargeable examination items, taking all current medical institutions as monitoring objects, extracting the medical examination items using statistical analysis and probability distribution methods, and adding them to the primary observation list; The secondary data processing module is configured to pre-configure unreasonable charging rules, establish a secondary observation list of unreasonable charging examination items, and add the medical examination items that trigger the unreasonable charging rules in the primary observation list to the secondary observation list. The decision model training module is configured to extract the associated medical examination item fee details data based on the medical examination items in the secondary observation list, and after data anonymization, label the quantitative fee details data as unreasonable charges to obtain a training data set L0; generalize and expand the training data set L0 using a pre-established generalization rule set F to obtain a training data set L1; and train the training data set L1 using a random forest as a classifier RF to obtain a dynamic unreasonable charge decision forest model. The third-level data processing module is configured to establish a third-level observation list of unreasonable-charged examination items, obtain the anonymized detailed data of the examination items associated with the second-level observation list, and use the dynamic unreasonable-charged decision forest model for labeling and identification. When the labeling result is higher than a predetermined probability value, the corresponding medical examination item is determined to be unreasonable-charged, and the unreasonable-charged medical examination item is added to the third-level observation list and monitored and warned in real time.

11. An electronic device, characterized in that, The device includes a memory and a processor, wherein the memory stores instructions and the memory and the processor are interconnected via a circuit; the processor invokes the instructions in the memory to implement the method for monitoring unreasonable medical examination items as described in any one of claims 1-9.

12. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the method for monitoring unreasonable medical examination items as described in any one of claims 1-9.

Images