A medical information anomaly detection method and system based on AI algorithm
Through the medical information abnormality detection method based on AI algorithm, combined with diagnostic items, diagnosis and treatment items, cost data and patient scores, the total score is calculated to identify abnormal information, which solves the medical process monitoring problem caused by the low level of hospital informationization, and achieves the correct Effective abnormal detection of medical information and monitoring of treatment processes have improved hospital management efficiency and patient satisfaction.
Patent Information
- Application Number
- CN202311259590.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2023-09-11
- Filing Date
- 2023-09-27
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2043-09-27
AI Technical Summary
The prior art is difficult to effectively monitor and detect whether the medical process received by patients in hospitals complies with regulations, especially in hospitals with low level of informationization, which makes it difficult to detect abnormal costs and treatment processes in a timely manner.
Using an abnormal medical information detection method based on AI algorithm, the second and third scores are calculated by obtaining the patient's medical information, including diagnostic items, diagnosis items, cost data and patient scores, and the total score is calculated by combining the three. If the total score is low At the preset value, the corresponding medical information is determined to be abnormal information, thereby realizing abnormal detection of medical information and monitoring of the treatment process.
It realizes the abnormal detection of medical information and comprehensive monitoring of treatment processes, ensuring compliance with diagnosis and treatment processes, costs and patient evaluations, and improving hospital management efficiency and patient satisfaction.
Smart Images
Figure CN117316404B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of Internet medical technology, and in particular to a medical information anomaly detection method and system based on an AI algorithm. Background Art
[0002] With the development of informatization in all walks of life, hospital informatization has become an important way to improve hospital management efficiency.
[0003] Usually, when patients go to the hospital for medical treatment, they cannot effectively know whether the treatment they receive, the fees they pay, etc. are in compliance with regulations, and this is also due to the hospital's imperfect level of informatization and insufficient information analysis.
[0004] Therefore, how to realize abnormal monitoring of the medical process received by patients has become an urgent problem to be solved. Summary of the invention
[0005] The present invention aims to solve one of the technical problems in the related art to at least a certain extent. To this end, the present invention provides a medical information anomaly detection method and system based on an AI algorithm to detect anomalies in medical information and monitor anomalies in the patient's treatment process.
[0006] The present invention provides a medical information anomaly detection method based on an AI algorithm, comprising:
[0007] Acquiring medical information of the patient, the medical information including diagnosis items, various diagnosis and treatment items, cost data, and a first score, the first score being the patient's score for the treatment service;
[0008] Determine a second score according to the diagnostic item and each of the treatment items;
[0009] Determine a third score according to the diagnosis item, the cost data and the preset cost data;
[0010] A total score is determined according to the first score, the second score, and the third score, and when the total score is lower than a preset value, the medical information corresponding to the total score is determined to be abnormal medical information.
[0011] Determining the second score based on the diagnostic items and each of the treatment items includes: determining a corresponding set of standard treatment items in a pre-constructed medical database based on the diagnostic items; determining a similarity value between each of the treatment items and the standard treatment item set; and determining the second score based on the similarity value.
[0012] The method of determining the similarity value between each of the diagnostic and treatment items and the set of standard diagnostic and treatment items includes: preprocessing the name text of each of the diagnostic and treatment items and the name text of each standard diagnostic and treatment item in the set of standard diagnostic and treatment items, wherein the preprocessing includes word segmentation, vocabulary deduplication and vectorization; and determining the similarity value based on the preprocessed diagnostic and treatment items and the preprocessed standard diagnostic and treatment item set.
[0013] The similarity value is obtained by the following formula:
[0014]
[0015] Where C is the similarity value, x is the diagnosis and treatment project, y is the standard diagnosis and treatment project set, n is the number of words after word segmentation and word deduplication between the text of the diagnosis and treatment project and the text of the standard diagnosis and treatment project set, k is the kth word, x is the number of words after word segmentation and word deduplication between the text of the diagnosis and treatment project and the text of the standard diagnosis and treatment project set, k is the frequency of the kth word in the text of the diagnosis and treatment project, y k is the word frequency of the kth word in the text of the standard diagnosis and treatment item set.
[0016] Determining the third score based on the diagnostic item, the cost data and the preset cost data includes: determining the preset cost data based on the diagnostic item; determining the cost difference between the preset cost data and the cost data, and determining the third score based on the cost difference.
[0017] Determining the preset cost data according to the diagnostic item includes: acquiring corresponding historical cost data according to the diagnostic item, and determining the average cost corresponding to the diagnostic item according to the historical cost data; and determining the preset cost data according to the average cost.
[0018] The present invention also provides a medical information anomaly detection system based on an AI algorithm, comprising: an anomaly detection device as described in claim 7, used to obtain the patient's medical information and determine abnormal medical information based on the medical information, the medical information including diagnostic items, various treatment items, cost data and a first score, the first score being the patient's score for the treatment service; a first user terminal, the first user terminal is connected to the anomaly detection device, the first user terminal is used to obtain the first score and send the first score to the anomaly detection device; a second user terminal, the second user terminal is connected to the anomaly detection device, and the second user terminal is used to receive the abnormal medical information sent by the anomaly detection device.
[0019] The abnormality detection device includes: an acquisition module, used to acquire the patient's medical information, the medical information including diagnostic items, various diagnostic and treatment items, cost data and a first score, the first score being the patient's score for the treatment service; a first determination module, used to determine a second score based on the diagnostic items and each of the diagnostic and treatment items; a second determination module, used to determine a third score based on the diagnostic items, the cost data and preset cost data; and an abnormality detection module, used to determine a total score based on the first score, the second score and the third score, and when the total score is lower than a preset value, determine that the medical information corresponding to the total score is abnormal medical information.
[0020] The first determination module is specifically used to: determine a corresponding set of standard diagnosis and treatment items in a pre-constructed medical database according to the diagnosis items; determine a similarity value between each of the diagnosis and treatment items and the set of standard diagnosis and treatment items; and determine the second score according to the similarity value.
[0021] The first determination module is also specifically used to: preprocess the name text of each of the medical items and the name text of each standard medical item in the standard medical item set, the preprocessing including word segmentation, vocabulary deduplication and vectorization; determine the similarity value based on the preprocessed medical items and the preprocessed standard medical item set.
[0022] In the embodiment of the present application, the main improvements are as follows: (1) The medical information includes a diagnosis item, each diagnosis and treatment item, cost data, and a first score. (2) A second score is determined based on the diagnosis item and each diagnosis and treatment item. (3) A third score is determined based on the diagnosis item, the cost data, and the preset cost data. (4) A total score is determined based on the first score, the second score, and the third score, and when the total score is lower than a preset value, the medical information corresponding to the total score is determined to be abnormal medical information. (5) Similarity value calculation formula.
[0023] The AI algorithm-based medical information anomaly detection method and system provided by the present invention obtain the patient's medical information, the medical information includes diagnostic items, various diagnostic and treatment items, cost data and a first score, the first score is the patient's score for the treatment service; then determine the second score based on the diagnostic items and various diagnostic and treatment items; determine the third score based on the diagnostic items, cost data and preset cost data; finally, determine the total score based on the first score, the second score and the third score, and when the total score is lower than the preset value, determine that the medical information corresponding to the total score is abnormal medical information, thereby realizing abnormal detection of medical information and abnormal monitoring of the patient's treatment process, and comprehensively determine whether the medical information is abnormal in terms of the three aspects of the diagnostic and treatment process, cost and patient evaluation. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0025] Figure 1 A schematic diagram of a scenario of the medical information anomaly detection method based on an AI algorithm provided by the present invention;
[0026] Figure 2 It is one of the flow charts of the medical information anomaly detection method based on AI algorithm in an embodiment of the present invention;
[0027] Figure 3 This is the second flowchart of the medical information anomaly detection method based on AI algorithm according to an embodiment of the present invention;
[0028] Figure 4 This is the third flowchart of the medical information anomaly detection method based on AI algorithm according to an embodiment of the present invention;
[0029] Figure 5 is a block diagram of a medical information anomaly detection system based on an AI algorithm according to an embodiment of the present invention;
[0030] Figure 6 is a block diagram of a medical information anomaly detection device based on an AI algorithm according to an embodiment of the present invention;
[0031] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0032] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0033] For ease of understanding, this specification provides an example scenario of a medical information anomaly detection method based on an AI algorithm. The example scenario is applied in Figure 1 In the application environment shown, there is a medical information anomaly detection device based on an AI algorithm, a first user terminal and a second user terminal. The first user terminal and the second user terminal can establish a communication connection with the anomaly detection device through the Internet.
[0034] In this scenario example, the anomaly detection device can be a server that can run a program of a medical information anomaly detection method based on an AI algorithm. The anomaly detection device can establish a connection with the medical information system of each hospital to obtain part of the patient's medical information.
[0035] The first user terminal may be a user terminal used by the patient. The patient may upload his or her rating of the treatment service (i.e., the first rating) through the first user terminal. The patient may also query information related to his or her diagnosis and treatment items, cost-related information, and the total rating obtained when the abnormality detection device performs abnormality detection, and other information through the first user terminal.
[0036] The second user terminal may be a user terminal used by a doctor. The doctor may view the abnormal medical information detected by the abnormality detection device through the second user terminal, and perform certain rectifications on the abnormal medical information to improve the patient's satisfaction with the medical service.
[0037] In this scenario example, the first user terminal and the second user terminal may be mobile terminals such as mobile phones, laptop computers, PADs (tablet computers), etc., and fixed terminals such as digital TVs, desktop computers, etc. The first user terminal and the second user terminal listed here are only examples and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.
[0038] With reference to the scenario example of the medical information anomaly detection method based on the AI algorithm in the aforementioned embodiment, the medical information anomaly detection method based on the AI algorithm in an embodiment of the present invention is introduced in detail below.
[0039] Figure 2 is one of the flowcharts of the medical information anomaly detection method based on AI algorithm in an embodiment of the present invention. The medical information anomaly detection method based on AI algorithm can be applied to the above-mentioned anomaly detection device, such as Figure 2 As shown, the medical information anomaly detection method based on AI algorithm may include the following steps:
[0040] Step 210: Obtain the patient's medical information, which includes diagnostic items, various treatment items, cost data, and a first score, where the first score is the patient's score for the treatment service.
[0041] Step 220: Determine a second score according to the diagnosis item and each treatment item.
[0042] Step 230: Determine a third score according to the diagnosis item, the cost data, and the preset cost data.
[0043] Step 240: Determine a total score based on the first score, the second score, and the third score, and if the total score is lower than a preset value, determine that the medical information corresponding to the total score is abnormal medical information.
[0044] Diagnostic items refer to the types of diseases determined based on clinical diagnosis, and treatment items refer to the treatment methods for the diagnosed diseases, such as surgical items, examination items, medication, etc.
[0045] In this embodiment, the abnormality detection device can establish a connection with the medical information system of each hospital and obtain the patient's diagnosis items, various treatment items and cost data from the medical information system. In addition, the patient can upload the first score through the first user terminal, and the first user terminal then sends the first score to the abnormality detection device.
[0046] After obtaining the diagnostic items and each diagnosis and treatment item, the second score can be determined based on the diagnostic items and each diagnosis and treatment item. It can be understood that the second score is used to evaluate whether the diagnosis and treatment methods received by the patient are abnormal. Specifically, for the diagnostic item, the standard diagnosis and treatment items for the diagnostic item can be compared with the various diagnosis and treatment items received by the patient. If they are the same or similar, it indicates that there is no violation or abnormality in the diagnosis and treatment methods for the patient. Among them, the standard diagnosis and treatment items can be obtained through historical diagnosis and treatment data and expert experience.
[0047] After obtaining the expense data, the third score can be determined based on the diagnostic items, expense data, and preset expense data. It is understandable that the third score is used to evaluate whether there are any abnormalities in the fees paid by the patient. Specifically, a correspondence table between diagnostic items and preset expense data can be prepared in advance, and the correspondence table can be stored in the abnormality detection device. When the third score needs to be calculated, the preset expense data corresponding to the diagnostic item can be found by calling the correspondence table, and then the preset expense data is compared with the expense data actually paid by the patient to obtain the third score.
[0048] After obtaining the first score, the second score, and the third score, the total score can be determined based on the first score, the second score, and the third score. Here, the total score can be determined by adding the first score, the second score, and the third score to obtain the total score; or by first setting a weight value for each of the first score, the second score, and the third score, and then multiplying each score by the corresponding weight value and adding them together to obtain the total score. The calculation method of the total score is not specifically limited here. The weight values of each score can be manually set by the staff according to actual needs, and are not specifically limited here.
[0049] After obtaining the total score, it is determined whether the total score is lower than the preset value. If the total score is lower than the preset value, it indicates that the corresponding medical information is abnormal, and the medical information is determined to be abnormal medical information; if the total score is greater than or equal to the preset value, it indicates that the corresponding medical information is not abnormal, and the medical information is determined to be normal medical information. The preset value can be manually set by the staff according to actual needs, and is not specifically limited here.
[0050] As an example, the full marks of the first score, the second score, and the third score can be set to 10 points, and the preset value can be set to 25 points. If the second score can be calculated to be 8 points, the third score to be 7 points, and the first score to be 9 points through diagnostic items, various treatment items, and cost data, then the total score can be obtained by adding up to 24 points. If the total score is lower than the preset value, the corresponding medical information is determined to be abnormal medical information.
[0051] Therefore, by calculating the scores of the three aspects of the medical treatment items received by the patient, the cost data and the patient's evaluation, and finally determining whether there are any abnormalities in the medical information based on the total score, anomaly detection of medical information is achieved, and abnormal monitoring of the patient's treatment process is carried out.
[0052] Figure 3 This is the second flowchart of the medical information anomaly detection method based on AI algorithm according to an embodiment of the present invention. Step 220 may include the following steps:
[0053] Step 310: Determine a corresponding set of standard diagnosis and treatment items in a pre-built medical database according to the diagnosis items.
[0054] Step 320: Determine the similarity value between each diagnosis and treatment item and the standard diagnosis and treatment item set.
[0055] Step 330: Determine a second score according to the similarity value.
[0056] Specifically, a medical database can be pre-built, which stores the names of several diagnostic items and the corresponding standard diagnosis and treatment item sets for each diagnostic item. The standard diagnosis and treatment item sets can be determined based on the clinical history data of the diagnostic items and various disease experts. As an example, if the diagnostic item is chronic obstructive pulmonary emphysema, the standard diagnosis and treatment item set may include anti-infection, bronchodilators, antiasthmatic, expectorant, bronchoscopic treatment, pulmonary function test, CT, and blood gas analysis.
[0057] After obtaining the standard diagnosis and treatment item set corresponding to the diagnostic item, the similarity value between each diagnosis and treatment item and the standard diagnosis and treatment item set can be calculated by using the text of each diagnosis and treatment item accepted by the patient and the text in the standard diagnosis and treatment item set. Finally, the second score is determined according to the similarity value. As an example, if the similarity value is 90% and the full score of the second score is 10 points, the similarity value can be multiplied by the full score of the second score, and the second score can be obtained as 9 points.
[0058] Determining the similarity value between each diagnosis and treatment item and the standard diagnosis and treatment item set may include: preprocessing the name text of each diagnosis and treatment item and the name text of each standard diagnosis and treatment item in the standard diagnosis and treatment item set, the preprocessing including word segmentation, vocabulary deduplication and vectorization; determining the similarity value based on the preprocessed diagnosis and treatment item and the preprocessed standard diagnosis and treatment item set.
[0059] The similarity value between each diagnosis and treatment project and the standard diagnosis and treatment project set can be calculated based on the word frequency of the text of each diagnosis and treatment project and the word frequency of the text in the standard diagnosis and treatment project set. Specifically, the text of each diagnosis and treatment project is combined into one text; then the text of each diagnosis and treatment project and the text of the standard diagnosis and treatment project set are first segmented, and the existing word segmentation method can be used to segment the text of each diagnosis and treatment project and the text of the standard diagnosis and treatment project set. After the word segmentation, take the union of all the words in the text of each diagnosis and treatment project and the text of the standard diagnosis and treatment project set, that is, put all the words that appear in the text of each diagnosis and treatment project and the text of the standard diagnosis and treatment project set into one set to obtain a vocabulary union set. After obtaining the vocabulary union set, it is necessary to perform vocabulary deduplication processing on the vocabulary union set. Vocabulary deduplication processing means that when there are identical words, only one word is retained.
[0060] After deduplication of the vocabulary union set, the vocabulary set of the text of each diagnosis and treatment project and the text of the standard diagnosis and treatment item set are obtained. The number of all words in the vocabulary set can be recorded as n, where n is an integer; then according to the vocabulary set, the word frequency of all words in the text of each diagnosis and treatment project and the text of the standard diagnosis and treatment item set are counted respectively, and the word frequency vector of the text of each diagnosis and treatment project and the word frequency vector of the text of the standard diagnosis and treatment item set can be obtained respectively.
[0061] As an example, when it is necessary to determine the similarity between each diagnosis and treatment item x and the standard diagnosis and treatment item set y, the word frequency vector corresponding to each diagnosis and treatment item x can be recorded as x={x1,x2…x k …x n}, the word frequency vector corresponding to the standard diagnosis and treatment item set y can be recorded as y={y1,y2…y k …y n}, where x k is the frequency of the kth word in the text of the diagnosis and treatment project, y kis the frequency of the kth word in the text of the standard diagnosis and treatment item set.
[0062] After obtaining the word frequency of each diagnosis and treatment item x and the standard diagnosis and treatment item set y, the similarity value is determined according to each diagnosis and treatment item x and the standard diagnosis and treatment item set y. The similarity value can be determined by the following formula:
[0063]
[0064] Among them, C is the similarity value, x is the diagnosis and treatment project, y is the standard diagnosis and treatment project set, n is the number of words after word segmentation and word deduplication between the text of the diagnosis and treatment project and the text of the standard diagnosis and treatment project set, k is the kth word, x is the number of words after word segmentation and word deduplication, k is the frequency of the kth word in the text of the diagnosis and treatment project, y k is the frequency of the kth word in the text of the standard diagnosis and treatment item set.
[0065] Figure 4 This is the third flowchart of the medical information anomaly detection method based on AI algorithm according to an embodiment of the present invention. Step 230 may include the following steps:
[0066] Step 410: Determine preset cost data according to the diagnosis item.
[0067] Step 420: Determine the cost difference between the preset cost data and the cost data, and determine a third score according to the cost difference.
[0068] Specifically, the corresponding relationship table can be retrieved to find the preset cost data corresponding to the diagnosis item; then the cost difference between the preset cost data and the cost data can be calculated; and finally, the third score can be determined according to the cost difference.
[0069] The preset fee data corresponding to each diagnostic item can be determined by the following steps: obtaining corresponding historical fee data according to the diagnostic item, and determining the average fee corresponding to the diagnostic item according to the historical fee data; and determining the preset fee data according to the average fee.
[0070] Specifically, the extreme values in the historical cost data may be removed first, and then the average value of the remaining historical cost data may be calculated to obtain the average cost, and finally the average cost may be used as the preset cost data.
[0071] As an example, if the diagnosis item is chronic obstructive pulmonary emphysema, its historical cost data is: 6000, 5000, 5500, 1500 and 9000 yuan. It can be obviously seen that 1500 yuan and 9000 yuan in this set of data are the maximum or minimum values. Therefore, these two data are removed, and the average value of the remaining three data is calculated to be 5500 yuan. The preset cost data corresponding to the diagnosis item chronic obstructive pulmonary emphysema can be set to 5500 yuan.
[0072] Figure 5 is a block diagram of a medical information anomaly detection system based on an AI algorithm according to an embodiment of the present invention. Figure 5 As shown, the medical information anomaly detection system 500 based on the AI algorithm may include:
[0073] The abnormality detection device 510 is used to obtain the patient's medical information and determine abnormal medical information based on the medical information. The medical information includes diagnostic items, various treatment items, cost data and a first score. The first score is the patient's score for the treatment service.
[0074] The first user terminal 520 is connected to the anomaly detection device, and is used to obtain a first score and send the first score to the anomaly detection device.
[0075] The second user terminal 530 is connected to the abnormality detection device, and is used to receive abnormal medical information sent by the abnormality detection device.
[0076] Figure 6 It is a block diagram of a medical information anomaly detection device based on an AI algorithm according to an embodiment of the present invention.
[0077] like Figure 6 As shown, the abnormality detection device 600 may include: an acquisition module 610, which is used to acquire the patient's medical information, the medical information includes diagnostic items, various diagnostic and treatment items, cost data and a first score, and the first score is the patient's score for the treatment service. A first determination module 620 is used to determine a second score based on the diagnostic items and various diagnostic and treatment items. A second determination module 630 is used to determine a third score based on the diagnostic items, cost data and preset cost data. An abnormality detection module 640 is used to determine a total score based on the first score, the second score and the third score, and when the total score is lower than a preset value, determine that the medical information corresponding to the total score is abnormal medical information.
[0078] The first determination module 620 is specifically used to: determine a corresponding set of standard diagnosis and treatment items in a pre-constructed medical database according to the diagnosis items; determine a similarity value between each of the diagnosis and treatment items and the set of standard diagnosis and treatment items; and determine the second score according to the similarity value.
[0079] The first determination module 620 is also specifically used to: preprocess the name text of each of the medical items and the name text of each standard medical item in the standard medical item set, the preprocessing including word segmentation, vocabulary deduplication and vectorization; determine the similarity value based on the preprocessed medical items and the preprocessed standard medical item set.
[0080] Thus, the patient's medical information is acquired through the acquisition module 610, and the medical information includes diagnostic items, various diagnostic and treatment items, cost data and a first score, and the first score is the patient's score for the treatment service; the first determination module 620 then determines the second score based on the diagnostic items and various diagnostic and treatment items; the second determination module 630 determines the third score based on the diagnostic items, cost data and preset cost data; finally, the anomaly detection module 640 determines the total score based on the first score, the second score and the third score, and when the total score is lower than the preset value, determines that the medical information corresponding to the total score is abnormal medical information, thereby realizing anomaly detection of medical information and abnormal monitoring of the patient's treatment process, and comprehensively determining whether the medical information is abnormal in terms of the three aspects of the diagnostic and treatment process, cost and patient evaluation.
[0081] For details not disclosed in the medical information anomaly detection system based on the AI algorithm in this embodiment, please refer to the details disclosed in the embodiment of the medical information anomaly detection method based on the AI algorithm in the embodiments of this specification, which will not be repeated here.
[0082] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the medical information anomaly detection method based on the AI algorithm, the method comprising: obtaining the patient's medical information, the medical information including the diagnosis items, each diagnosis and treatment item, the cost data and the first score, the first score being the patient's score for the treatment service; determining the second score according to the diagnosis items and each diagnosis and treatment item; determining the third score according to the diagnosis items, the cost data and the preset cost data; determining the total score according to the first score, the second score and the third score, and determining the medical information corresponding to the total score as abnormal medical information when the total score is lower than the preset value.
[0083] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the medical information anomaly detection method based on the AI algorithm provided by the above-mentioned methods.
Claims
1. A medical information anomaly detection method based on AI algorithm, It is characterized in that include: Acquiring medical information of the patient, the medical information including diagnosis items, various diagnosis and treatment items, cost data, and a first score, the first score being the patient's score for the treatment service; Determine a second score according to the diagnostic item and each of the treatment items; determining a third score according to the diagnosis item, the cost data and the preset cost data; determining a total score according to the first score, the second score, and the third score, and determining that the medical information corresponding to the total score is abnormal medical information when the total score is lower than a preset value; The determining of the second score according to the diagnostic item and each of the treatment items comprises: Determining a corresponding set of standard diagnosis and treatment items in a pre-constructed medical database according to the diagnosis items; Determining the similarity value between each of the diagnosis and treatment items and the set of standard diagnosis and treatment items; Determining the second score according to the similarity value; The determining of the similarity value between each of the diagnosis and treatment items and the set of standard diagnosis and treatment items includes: Preprocessing the name text of each of the diagnosis and treatment items and the name text of each of the standard diagnosis and treatment items in the standard diagnosis and treatment item set, wherein the preprocessing includes word segmentation, word deduplication and vectorization; Determine the similarity value according to the preprocessed diagnosis and treatment items and the preprocessed standard diagnosis and treatment item set; specifically, combine the texts of each diagnosis and treatment item into one text; then first perform word segmentation on the texts of each diagnosis and treatment item and the texts of the standard diagnosis and treatment item set, and after word segmentation, take the union of all words in the texts of each diagnosis and treatment item and the texts of the standard diagnosis and treatment item set, that is, put all words appearing in the texts of each diagnosis and treatment item and the texts of the standard diagnosis and treatment item set into one set to obtain a vocabulary union set; after obtaining the vocabulary union set, it is necessary to perform word deduplication on the vocabulary union set; word deduplication means that when there are identical words, only one word is retained; after performing word deduplication on the vocabulary union set, a vocabulary set of the texts of each diagnosis and treatment item and the texts of the standard diagnosis and treatment item set is obtained, and the number of all words in the vocabulary set is recorded as n, where n is an integer; then, according to the vocabulary set, the word frequencies of all words in the texts of each diagnosis and treatment item and the texts of the standard diagnosis and treatment item set are respectively counted, and the word frequency vectors of the texts of each diagnosis and treatment item and the word frequency vectors of the texts of the standard diagnosis and treatment item set are respectively obtained; When it is necessary to determine the similarity between each diagnosis and treatment item x and the standard diagnosis and treatment item set y, the word frequency vector corresponding to each diagnosis and treatment item x is recorded as x={x1,x2…xk…xn}, and the word frequency vector corresponding to the standard diagnosis and treatment item set y is recorded as y={y1,y2…yk…yn}, where xk is the word frequency of the kth word in the text of the diagnosis and treatment item, and yk is the word frequency of the kth word in the text of the standard diagnosis and treatment item set; After obtaining the word frequency of each diagnosis and treatment item x and the standard diagnosis and treatment item set y, the similarity value is determined according to each diagnosis and treatment item x and the standard diagnosis and treatment item set y; the similarity value is determined by the following formula: Where C is the similarity value, x is the diagnosis and treatment project, y is the standard diagnosis and treatment project set, n is the number of words in the text of the diagnosis and treatment project and the text of the standard diagnosis and treatment project set after word segmentation and word deduplication, k is the kth word, xk is the word frequency of the kth word in the text of the diagnosis and treatment project, and yk is the word frequency of the kth word in the text of the standard diagnosis and treatment project set; The determining of the third score according to the diagnosis item, the cost data and the preset cost data includes: Determine the preset cost data according to the diagnostic item; determine the cost difference between the preset cost data and the cost data, and determine the third score according to the cost difference; determine the preset cost data according to the diagnostic item, including: obtain corresponding historical cost data according to the diagnostic item, and determine the average cost corresponding to the diagnostic item according to the historical cost data; determine the preset cost data according to the average cost.
2. A medical information anomaly detection system based on an AI algorithm, the system implementing the method as claimed in claim 1, It is characterized in that include: an abnormality detection device, used to obtain medical information of a patient and determine abnormal medical information based on the medical information, wherein the medical information includes diagnosis items, various diagnosis and treatment items, cost data, and a first score, wherein the first score is a score of the patient for the treatment service; a first user terminal, the first user terminal being connected to the anomaly detection device, and the first user terminal being used to obtain the first score and send the first score to the anomaly detection device; A second user terminal, the second user terminal is connected to the abnormality detection device, and the second user terminal is used to receive the abnormal medical information sent by the abnormality detection device.
3. The medical information anomaly detection system based on AI algorithm according to claim 2, It is characterized in that The abnormality detection device comprises: An acquisition module, used for acquiring medical information of a patient, wherein the medical information includes diagnosis items, various diagnosis and treatment items, cost data, and a first score, wherein the first score is a score of the patient on the treatment service; A first determination module, used to determine a second score according to the diagnosis item and each of the diagnosis and treatment items; A second determination module, configured to determine a third score according to the diagnosis item, the cost data and preset cost data; An abnormality detection module is used to determine a total score based on the first score, the second score and the third score, and when the total score is lower than a preset value, determine that the medical information corresponding to the total score is abnormal medical information.
4. The medical information anomaly detection system based on AI algorithm according to claim 3, It is characterized in that The first determining module is specifically used for: Determining a corresponding set of standard diagnosis and treatment items in a pre-constructed medical database according to the diagnosis items; Determining the similarity value between each of the diagnosis and treatment items and the set of standard diagnosis and treatment items; The second score is determined according to the similarity value.
5. The medical information anomaly detection system based on AI algorithm according to claim 4, It is characterized in that The first determining module is further specifically configured to: Preprocessing the name text of each of the diagnosis and treatment items and the name text of each of the standard diagnosis and treatment items in the standard diagnosis and treatment item set, wherein the preprocessing includes word segmentation, word deduplication and vectorization; The similarity value is determined according to the preprocessed diagnosis and treatment items and the preprocessed standard diagnosis and treatment item set.