Medical consumable document management method based on ai model, chip and intelligent device

Abstract

This invention provides a medical consumables document management method, chip, and intelligent device based on an AI model, belonging to the field of data management technology. Specifically, it includes: managing and processing medical consumables document data using a document management strategy to obtain abnormal requisition data; identifying associated medical consumables based on the abnormal requisition data; determining the impact type of abnormal requisition based on the correlation between the requisition data of associated medical consumables in different departments and the overall requisition data of medical consumables; determining the impact type of abnormal requisition based on a risk type of medical consumables; and combining the requisition data of departments for different risk types of medical consumables to determine the document management method for each department, thereby improving the reliability of medical consumables inventory management.

Description

Technical Field

[0001] This invention belongs to the field of data management technology, and in particular relates to a medical consumables document management method, chip and intelligent device based on an AI model. Background Technology

[0002] With the rapid development of HIS systems, medical consumables documents often involve multiple stages of the medical consumables circulation, such as procurement and warehousing, inventory management, clinical use and billing. Therefore, as the quantity of medical consumables and the number of circulation stages increase, it is inevitable that the management of medical consumables documents will become increasingly difficult.

[0003] To address the aforementioned technical issues, the invention patent application CN114021749A, "Intelligent Management System for Clinical Laboratory Reagents and Consumables," creates multi-dimensional and refined user profile tags for each user. Personalized lottery rules are then formulated based on these user profile tags, improving user experience and enhancing the credibility and appeal of the activity. However, the above technical solution has the following drawbacks: In the management of medical consumables, discrepancies in the requisition of medical consumables often occur due to their similar placement or appearance. Therefore, determining a document management strategy for medical consumables based on the types of risks caused by shortages and the differences in confusion risks resulting from the dispersed nature of requisitions is a pressing technical problem. This strategy aims to quickly identify related medical consumables that are prone to requisition errors and to develop targeted document verification management strategies to ensure the reliability of requisition processing.

[0004] Therefore, there is an urgent need for a medical consumables document management method, chip, and intelligent device based on AI models. Summary of the Invention

[0005] To achieve the objectives of this invention, the following technical solution is adopted: Specifically, this application provides a medical consumables document management method based on an AI model, which includes: S1 determines the impact data of different departments when the medical consumables are in short supply based on the requisition data of medical consumables. Based on the impact data and the AI ​​model, it determines the risk type of the medical consumables shortage impact. Based on the risk type of the medical consumables shortage impact and combined with the dispersion of the medical consumables requisition data on different dates, it determines the document management strategy for medical consumables. S2 uses the document management strategy to manage and process the document data of the medical consumables, obtains abnormal requisition data, determines the associated medical consumables with abnormal requisition based on the abnormal requisition data, and determines the impact type of abnormal requisition of the medical consumables according to the degree of correlation between the requisition data of the associated medical consumables in different departments and the requisition data of the medical consumables. S3 determines the document management method for the department based on the abnormal impact type of the requisition of medical consumables of the aforementioned risk type, combined with the department's requisition data for medical consumables of different risk types.

[0006] The beneficial effects of this invention are as follows: Based on the risk type of medical consumable shortages and the dispersion of medical consumable requisition data across different dates, a document management strategy for medical consumables is determined. According to the shortage risk level of medical consumables and the co-occurrence characteristics of their historical requisition data, the most suitable document verification sensitivity is dynamically matched for each consumable. By assessing the overall monitoring reliability of the target consumable's requisition scenario, a decision is made on whether to initiate independent, rigorous verification. Furthermore, considering the general situation of medical consumables when requisition anomalies occur, the verification processing of medical consumables in the documents is determined, thereby achieving optimized allocation of management resources while ensuring effective monitoring.

[0007] Based on the correlation between the requisition data of related medical consumables in different departments and the requisition data of medical consumables, the types of impact of abnormal requisition of medical consumables are determined. Taking departments as units, based on the proportion of Class I impact medical consumables in the whole hospital, the scale of the source of confusion of Class I risk consumables, the department's dependence on consumables, and the number of highly dependent consumables in the department, the document verification strategy of the department is dynamically determined. While ensuring effective monitoring of high-risk consumables, the focus is on accelerating the identification efficiency of related consumables, and accumulating data for subsequent management.

[0008] Furthermore, the medical consumables requisition data includes requisition data for the medical consumables on different dates.

[0009] Furthermore, the impact data on different departments is determined based on the department's requisition data on different dates.

[0010] Furthermore, the method for determining the risk type of the shortage of medical consumables is as follows: S11 Based on the impact data of the shortage of medical consumables on different departments, determine the requisition data of the departments on different dates; S12 uses the date on which the requisition data exists as the requisition date; S13 determines the type of risk impact of the shortage of medical consumables based on the requisition date data in different departments and the AI ​​model.

[0011] Furthermore, the method for determining the document management method for the aforementioned department is as follows: S41 Based on the abnormal impact type of the requisition of medical consumables of the aforementioned risk type, determine the medical consumables of the aforementioned risk type with a certain impact type, and treat them as medical consumables with a certain impact. S42 determines the requisition date of a certain type of risk medical consumable in the department based on the department's usage data for different medical consumables; S43 determines the department's document management method based on one type of medical consumables affecting the first type of risk, and the department's requisition date for the first type of medical consumables.

[0012] Secondly, the present invention provides an intelligent device that employs the aforementioned AI model-based medical consumables document management method, specifically including: The module includes a management strategy determination module, an impact type output module, and a management method output module. The management strategy determination module is responsible for determining the document management strategy for medical consumables. The impact type output module is responsible for determining the impact type of the abnormal requisition of the medical consumables; The management method output module is responsible for determining the document management method for the department.

[0013] Thirdly, this application provides a chip for use in a smart device, comprising: a memory and a processor connected in communication, and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the aforementioned AI model-based medical consumables document management method when running the computer program.

[0014] Other features and advantages will be set forth in the following description, and the objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0015] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0016] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0017] Figure 1This is a flowchart of a medical consumables document management method based on an AI model; Figure 2 This is a flowchart illustrating the method for determining the risk types impacted by shortages of medical consumables; Figure 3 This is a flowchart illustrating the method for determining the document management strategy for medical consumables; Figure 4 This is a flowchart illustrating the method for determining the impact types of abnormal requisition of medical consumables. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0019] Example 1 like Figure 1 As shown, this application provides a medical consumables document management method based on an AI model, specifically including: S1 determines the impact data of different departments when the medical consumables are in short supply based on the requisition data of medical consumables. Based on the impact data and the AI ​​model, it determines the risk type of the medical consumables shortage impact. Based on the risk type of the medical consumables shortage impact and combined with the dispersion of the medical consumables requisition data on different dates, it determines the document management strategy for medical consumables. S2 uses the document management strategy to manage and process the document data of the medical consumables, obtains abnormal requisition data, determines the associated medical consumables with abnormal requisition based on the abnormal requisition data, and determines the impact type of abnormal requisition of the medical consumables according to the degree of correlation between the requisition data of the associated medical consumables in different departments and the requisition data of the medical consumables. S3 determines the document management method for the department based on the abnormal impact type of the requisition of medical consumables of the aforementioned risk type, combined with the department's requisition data for medical consumables of different risk types.

[0020] Furthermore, the medical consumables requisition data includes requisition data for the medical consumables on different dates.

[0021] Furthermore, the impact data on different departments is determined based on the department's requisition data on different dates.

[0022] Specifically, such as Figure 2 As shown, the method for determining the risk type of the shortage of medical consumables is as follows: The core decision-making objective of this plan is to accurately and hierarchically assess the potential impact of a shortage of a medical consumable on the overall operation of the hospital, and classify it into different risk types in order to adopt differentiated response strategies.

[0023] Its core logic is a progressively layered decision tree. First, it examines whether there is a single high-risk department with an extremely high dependence on this consumable. If not, it examines whether the scope of affected departments is broad. If the scope is broad, it further uses an artificial intelligence model to comprehensively assess the dependence of each department to determine the overall risk level. The entire logic delves deeper layer by layer, from the "point" (extreme dependence of a single department) to the "surface" (the number of affected departments) and then to the "volume" (the depth of the overall impact), ensuring the comprehensiveness and accuracy of the risk assessment.

[0024] S11 Based on the impact data of the shortage of medical consumables on different departments, determine the requisition data of the departments on different dates; Shortage status: refers to the immediate situation where the inventory of the medical consumable is below the preset safety stock level and cannot meet normal clinical needs.

[0025] Impact Data: This specifically refers to the records of the impact on clinical departments during the period when the consumable was in short supply, due to their inability to obtain it normally. Examples include departmental substitution requests, emergency restocking requests, and surgery postponement records. In this step, it is concretized as analyzing the actual consumption patterns of departments during past shortage periods.

[0026] Requisition data: refers to the detailed records of each department's actual receipt of the consumable from the warehouse in history, including information such as the date of receipt, the department that received it, and the quantity received.

[0027] Analyzing historical data on the impact of shortages and transforming it into objective requisition data establishes a reliable and quantifiable analytical foundation. Requisition data directly reflects a department's actual need for consumables and serves as the primary basis for all subsequent risk assessments. The significance of this step lies in concretizing the abstract concept of "impact" into extractable and calculable records in the database, ensuring the objectivity and operability of the assessment.

[0028] Specific examples: For example, when analyzing the shortage risk of "disposable intravascular ultrasound catheters", the system retrieved the consumable management system records of three departments—cardiology, emergency department, and interventional operating room—during the past three months when the consumable was in short supply, and obtained the daily usage data of "disposable intravascular ultrasound catheters" in these three departments during this period.

[0029] S12 uses the date on which the requisition data exists as the requisition date; Issuance Date: Refers to the specific calendar date in the historical records on which a particular department actually received the medical consumable.

[0030] This step involves cleaning and feature extraction of the original requisition data. Marking the dates on which requisition data exists separately transforms the department's demand for the consumable from a continuous timeline to discrete points in time. This allows subsequent analysis to focus on "on which days the department needs the consumable," thus more accurately calculating the department's reliance frequency and avoiding fluctuations caused by variations in the quantity requisitioned on a particular day. Its significance lies in constructing a standardized indicator that measures demand frequency on a daily basis.

[0031] Specific examples: In the requisition data obtained from S11, the cardiology department had 20 requisition records in the past three months. These 20 days are then marked as requisition dates for the department. Even if the cardiology department requisitions 10 catheters on a single day, it is only counted as one requisition date.

[0032] S13 determines the type of risk impact of the shortage of medical consumables based on the requisition date data in different departments and the AI ​​model.

[0033] Issuance date data: After processing by S12, this refers to the set of data for each department that is marked with the issuance date. It reflects the frequency and pattern of each department's demand for this consumable.

[0034] AI model: Refers to a pre-trained machine learning model, such as a neural network or support vector machine. This model learns the complex non-linear relationship between the characteristics of requisition dates in different departments and the final comprehensive impact in historical data, thereby being able to predict the "comprehensive impact value" of new consumables.

[0035] Shortage Impact Risk Type: This defines the risk level output by this method. Based on the proposed solution, it can be categorized into three risk types: Type 1 (high risk), Type 2 (medium risk), and Type 3 (low risk).

[0036] This is the core decision-making step of the entire methodology. The introduction of AI models aims to move beyond simple rule-based judgments (such as looking only at the number of departments or the usage rate of a single department), enabling the processing of complex data from multiple departments and dimensions. It seeks to uncover the synergistic effects of demand patterns across different departments, thereby deriving a more accurate and intelligent comprehensive risk assessment result. Its significance lies in achieving a leap from "rule-based assessment" to "data-driven intelligent assessment."

[0037] Specifically, based on the requisition date data from different departments, the types of risks associated with the shortage of the aforementioned medical consumables are determined, including: S131 determines the shortage impact value of the department based on the proportion of the requisition date in the department, and determines whether there is a department whose shortage impact value is greater than the preset impact threshold. If so, the shortage impact risk type of the medical consumables is determined to be a type of risk. If not, proceed to step S132. Percentage of requisition dates: This refers to the ratio of the number of requisition dates for a particular department to the total number of days in a specific statistical period (e.g., one year).

[0038] Shortage Impact Value: A value quantified by the percentage of requisition dates, used to measure the potential impact of a shortage of this consumable on the department's daily operations. The higher the percentage, the more frequently the department needs the consumable and the stronger its dependence on it; therefore, the greater the impact of a shortage on the department.

[0039] Preset impact threshold: A critical value set by hospital administrators based on management objectives. It is used to determine whether a department's dependence on the consumable has reached an "extremely high" level, such that in the event of a shortage, the impact on that department alone would be sufficient to classify the entire consumable's risk level as the highest level (Category 1 risk).

[0040] Risk type 1: This indicates that the shortage of this consumable will have the most severe and direct impact on the hospital, requiring the highest priority response strategy.

[0041] This step, the first layer of the decision tree, is used to capture the risk of "single-point high dependence." The logic is: if a department has an extremely high dependence on a particular consumable (for example, the department needs to use it almost daily), then a shortage of that consumable is itself a significant risk, regardless of the situation in other departments. This ensures that the most urgent risks can be identified and addressed immediately. Its significance lies in establishing a rapid-response "red alert" mechanism for high-risk consumables.

[0042] Specific examples: Assuming a statistical period of one year (365 days) and a preset impact threshold of 80%, if the cardiology department's usage date is 300 days, then the usage date percentage is 300 / 365 ≈ 82.2%, and the calculated shortage impact value is 82.2%. This value is greater than the preset impact threshold of 80%, therefore the system directly classifies the shortage impact risk type of "single-use intravascular ultrasound catheter" as a Class I risk type.

[0043] S132: Obtain the departments with requisition data, designate the departments with requisition data as requisitioning departments, and determine whether the number of requisitioning departments is greater than a preset threshold for the number of requisitioning departments. If yes, proceed to step S133; otherwise, determine that the shortage of medical consumables is a type three risk. Departments that issued the product: These are departments that have issued the product at least once within the statistical period. This is the initial screening of the affected departments.

[0044] Preset threshold for the number of departments using the consumable: A critical quantity value set by hospital administrators. It is used to determine whether the range of departments affected by the consumable is wide. If the quantity exceeds the threshold, it indicates that the consumable is a "general-purpose" or "common-use" consumable used by many departments.

[0045] Three risk types: This indicates that the shortage of this consumable has a limited impact on the hospital and the hospital is not highly dependent on it, which is considered low risk.

[0046] After eliminating the risk of "single-point high dependence," the second step is to assess the breadth of the risk. If a consumable is used by only a very few departments, then its shortage impact is localized, and the risk level is naturally low. The logic of this step is "if the impact is narrow, the risk is low." Its significance lies in quickly filtering out a large number of low-risk "niche" consumables through simple statistical analysis, thereby focusing management resources and subsequent AI analysis on consumables with a wider impact.

[0047] Specific examples: Continuing with the example in S131, if the shortage impact value for all departments does not exceed the threshold, the process proceeds to S132. System statistics show that in the past year, only the emergency department and interventional operating room, besides cardiology, had requisition records. Therefore, the number of departments that requisitioned the consumable is 3. If the hospital's preset threshold for the number of departments requisitioning the consumable is 5, since 3 is no more than 5, the system classifies the shortage impact risk type of this consumable as a Class III risk type. However, if the number of departments requisitioning the consumable reaches 6, the process proceeds to S133 for further analysis.

[0048] S133 uses the shortage impact value of different departments as the input of the AI ​​model, determines the comprehensive impact value based on the output of the AI ​​model, and judges whether the comprehensive impact value is greater than the preset impact value. If so, the shortage impact risk type of the medical consumables is determined to be a type I risk type; otherwise, the shortage impact risk type of the medical consumables is determined to be a type II risk type.

[0049] Shortage Impact Value: Same as S131, referring to the percentage of requisition dates for each department. Here, these serve as the input feature vector for the AI ​​model.

[0050] AI Model: Similar to S13, this is a trained machine learning model. Its function is to perform a complex nonlinear fusion of the shortage impact values ​​of multiple departments to simulate the synergistic impact effect of "1+1>2" when these departments face shortages simultaneously.

[0051] Overall Impact Value: The AI ​​model calculates and outputs an overall score based on the input shortage impact values ​​for each department. This score represents the overall risk level posed by the consumable to the hospital's overall operation, taking into account the dependence of all departments using it.

[0052] Impact preset value: Another threshold value set by hospital administrators to define the boundary between "high risk" and "medium risk" on the continuous value of comprehensive impact.

[0053] Category 2 risk: This indicates that the shortage of this consumable will have a moderate impact, requiring attention and the development of routine contingency plans.

[0054] This AI model employs a Multilayer Perceptron (MLP) neural network architecture. The number of nodes in its input layer corresponds to the number of departments using the consumables to be evaluated, with each node receiving the shortage impact value of one department. The model contains at least one hidden layer to learn the nonlinear interaction and synergistic effects between the shortage impact values ​​of different departments. The output layer is a single node that generates a comprehensive impact value between 0 and 100. The model is trained using supervised learning, with a training set constructed based on historical consumable data. The input feature vector of each training sample consists of the shortage impact values ​​of each department using consumables in a past shortage event. The sample label (i.e., the target output) is the true comprehensive impact score, comprehensively evaluated by hospital management experts based on the actual impact consequences of the shortage event (such as the number of canceled surgeries, clinical complaint rates, and emergency procurement costs). During training, the model's internal weight parameters are continuously optimized using a backpropagation algorithm to minimize the mean squared error between the model's predicted comprehensive impact value and the expert-assessed true value, until the model converges. This enables the model to accurately predict overall risk from departmental dependency patterns.

[0055] This is the final and most complex layer of the decision tree. When a consumable is widely used across departments (passing the S132 screening) but lacks a single, extremely dependent department, its overall risk becomes less readily apparent. In this case, introducing an AI model is for "deep assessment." The AI ​​model can capture the significant pressure that multiple departments, while individually having low dependence, may collectively exert on hospital operations (e.g., multiple departments simultaneously requiring alternative solutions). Its significance lies in using artificial intelligence to finely differentiate consumables of medium risk levels, avoiding grouping all widely used consumables into the same category, thereby achieving better allocation of management resources.

[0056] Specific examples: Assuming that more than five departments requested the consumable in step S132, the process proceeds to step S133. The system inputs the shortage impact values ​​of six departments—cardiology (shortage impact value 82.2%), emergency department (65%), interventional operating room (70%), general surgery (40%), vascular surgery (55%), and intensive care unit (60%)—into a pre-trained AI model. The model calculates and outputs a comprehensive impact value of 88 (out of 100). If the hospital's preset impact value is 85, since 88 > 85, the system classifies the consumable's shortage impact risk as a Class I risk type. If the model's output comprehensive impact value is 75, which is less than 85, it is classified as a Class II risk type.

[0057] Specifically, such as Figure 3 As shown, the method for determining the document management strategy for medical consumables is as follows: The core decision-making objective of this solution is to dynamically match the most suitable document verification sensitivity for each medical consumable based on its shortage risk level and the co-occurrence characteristics of its historical requisition data. By assessing the overall monitoring reliability of the requisition scenario in which the target consumable is located, it determines whether an independent, rigorous verification is necessary, thereby optimizing the allocation of management resources while ensuring effective monitoring. Its core logic is based on the type of shortage impact risk. For intermediate states like Category II risk, the monitoring reliability of the "consumable system" is assessed through progressive conditions: if the target consumable always appears simultaneously with a large number of Category I risk consumables, then the scenario has high monitoring reliability due to the presence of high-risk consumables, and a lenient correlation-based sensitivity strategy can be adopted; if the target consumable appears frequently alone or its requisition pattern is highly discrete, then the monitoring reliability of the scenario is low, requiring an upgrade to a strict global sensitivity strategy.

[0058] S21 Based on the medical consumables requisition data for different dates, determine the medical consumables requisitioned on different dates and use them as requisitioned consumables; Consumable requisition data: refers to the daily details of consumables requisitioned by each department from the warehouse in the historical records, including information such as consumable name, specifications, requisitioning department, requisition date, requisition quantity, and the corresponding document number.

[0059] Consumables requisitioned: This specifically refers to the collection of all medical consumables requisitioned by any department on a particular date. This collection reflects the material needs structure of clinical activities on that day.

[0060] Aggregating consumables by requisition date yields a daily "basket" of consumables actually consumed by the hospital. This "basket" forms the basis for subsequent analysis of co-occurrence relationships among different consumables. Only by identifying which consumables are requisitioned simultaneously each day can the degree of association between the target consumable and other consumables be determined. Its significance lies in transforming discrete requisition records into a daily consumable combination view, establishing a foundation for co-occurrence analysis.

[0061] Taking March 1, 2024 as an example, the system retrieved all the requisition records for that day and found that the requisitioned consumables included "disposable intravascular ultrasound catheters", "disposable sterile syringes", "medical sutures", and "medical surgical masks". These consumables constituted the set of consumables requisitioned on March 1, 2024.

[0062] S22 determines the deviation of consumable requisition between different dates based on the dispersion of consumable requisition on different dates; Dispersion: refers to the degree of difference between the sets of consumables requisitioned on different dates.

[0063] Deviation: A quantification of "dispersion" used to describe how different the types of consumables requisitioned are between any two dates.

[0064] This step aims to capture the stability and diversity of consumable usage patterns. Deviations reflect the changing patterns of requisition combinations across different dates, providing a basis for subsequent segmentation into "requisition pattern groups," thereby assessing the impact of varying usage scenarios on system monitoring reliability.

[0065] Specific examples: The system compares the sets of consumables requisitioned on March 1st and March 2nd, 2024. {A, B, C, D} were requisitioned on March 1st, and {A, B, C, E} were requisitioned on March 2nd. By calculating the percentage of discrepancies, the deviation for these two days is determined to be moderate to small.

[0066] S23 determines the document management strategy for the medical consumables based on the risk type of the shortage impact of the medical consumables, the requisition data on different dates, and the deviation of the requisition of consumables between different dates.

[0067] Shortage impact risk type: refers to the risk level of the consumable as determined by the aforementioned methods, whether it is classified as Class I, Class II, or Class III.

[0068] Document Management Strategy: This refers to whether to initiate an inspection and processing of this consumable when any discrepancy is found between the requisition records of other consumables and the actual needs. There are two types: one is the "global sensitivity strategy," which means that this consumable will be inspected whenever any other consumable is found to have a discrepancy; the other is the "related sensitivity strategy," which means that this consumable will only be inspected if discrepancies are found between it and other consumables on the same document.

[0069] This step combines risk level assessment with scenario reliability evaluation to formulate refined monitoring rules. Consumables with different risk levels have different monitoring priorities. For Category 1 risks, regardless of scenario reliability, the most sensitive global strategy must be adopted; for Category 3 risks, an economical correlation strategy can be used; for Category 2 risks, it is necessary to assess the overall monitoring reliability of the scenario in which the consumables are used to determine the leniency of the strategy. Its significance lies in achieving precise allocation of monitoring resources and avoiding over-monitoring of Category 2 consumables in high-reliability scenarios.

[0070] Furthermore, the document management strategy for the aforementioned medical consumables is determined, specifically including: Scenario 1: If the risk type of the shortage of medical consumables is classified as a Class I risk, then the document management strategy for the medical consumables shall be determined as follows: whenever medical consumables that are not in accordance with the demand are found to be requisitioned, the requisition data of the medical consumables shall be checked and processed, that is, whether the requisition record is consistent with the actual outbound data, and whether there is a requisition deviation.

[0071] Medical consumables requisitioned but not in accordance with demand: refers to consumables whose actual requisitioned type, specifications, or quantity does not match clinical needs in any requisition record, regardless of whether the consumable is the target consumable itself.

[0072] Verification and processing of requisition data: This refers to reviewing the requisition records of target consumables.

[0073] Category I high-risk consumables are the most valuable and have the greatest impact from shortages, requiring the most stringent monitoring. Their monitoring strategy does not rely on scenario reliability assessments. Any deviation from other consumables could lead to chaos in warehouse management, which could potentially affect Category I high-risk consumables located nearby or with similar appearances. Therefore, any "alarm" in the system must immediately trigger an inspection of Category I high-risk consumables to ensure their safety.

[0074] Specific examples: "Disposable intravascular ultrasound catheters" are classified as a risk category. One day, the warehouse discovered a discrepancy between the amount of "disposable sterile syringes" issued and the demand. Upon detecting this discrepancy, the system immediately triggered an inspection of the "disposable intravascular ultrasound catheters," requiring the warehouse to check its inventory to confirm whether it was mistakenly taken due to being placed too close to the syringes.

[0075] Scenario 2: If the risk type of the shortage of medical consumables is classified as a Category III risk, then the document management strategy for the medical consumables is determined as follows: whenever medical consumables that are inconsistent with the demand are found to have been requisitioned, and the medical consumables that are inconsistent with the demand are on the same document, the requisition data of the medical consumables will be checked and processed. That is, whether the requisition record is consistent with the actual outbound data, and whether there is a requisition deviation.

[0076] On the same document: This refers to the fact that the consumables with the deviation and the target consumables both appeared in the same historical record of the requisition form or delivery form.

[0077] Category III consumables are low-value and highly substitutable, requiring no comprehensive monitoring. However, if a consumable has appeared in the same order as another, it indicates that both may serve the same clinical scenario, posing a potential for confusion. If the consumable in that same order is misrepresented, the target consumable also risks being mistakenly received. This strategy uses historical order relationships as a proxy indicator to capture potential confusion at low cost.

[0078] Specific examples: "Medical cotton swabs" are classified as a third-risk item. The system detected a discrepancy between the requisition and demand for "medical tape." Historical data revealed that "medical tape" and "medical cotton swabs" had previously appeared on the same document. Therefore, the system triggered an inspection of the "medical cotton swabs."

[0079] Scenario 3: If the shortage of medical consumables is classified as a Class II risk type, obtain Class I risk type medical consumables from the list of medical consumables and determine whether the proportion of Class I risk type medical consumables in the requisition of medical consumables on different dates is greater than a preset consumable proportion threshold. If so, determine that the document management strategy for the medical consumables is as follows: whenever medical consumables with inconsistent requisition and demand are found, and the medical consumables with inconsistent requisition and demand are on the same document as the medical consumables, the requisition data of the medical consumables will be checked to determine whether there is a requisition deviation. If not, proceed to the next step. Category 1 risk-type medical consumables: refers to those that are transported via... Figure 2 Other consumables that are classified as having the highest risk level are themselves under close monitoring by a globally sensitive strategy.

[0080] Preset consumable proportion threshold: A percentage threshold used to determine whether the proportion of risky consumables in the daily requisition set of target Class II consumables is consistently higher than this value.

[0081] "All greater than": This means that the percentage of each day with a usage record exceeds this threshold.

[0082] If each time a target consumable is requisitioned, its consumable combination contains a large number of Class 1 risk consumables (each exceeding a threshold percentage), then the presence of these Class 1 risk consumables indicates that the entire requisition scenario is under intensive monitoring. According to scenario 1, any deviation in consumables will trigger the inspection of Class 1 risk consumables. This intensive monitoring covers the entire scenario, ensuring high system monitoring reliability in the environment where the target consumable is located. In this case, the target consumable does not require independent, rigorous monitoring; a correlation-based sensitivity strategy can effectively cover its risk, thereby reducing management costs.

[0083] Specific examples: "Medical sutures" are classified as a Class II risk type. The system retrieves a list of all medical consumables classified as Class I risk. The system then calculates the percentage of Class I risk consumables in the daily requisition data for sutures across all past requisition dates. If the preset threshold for this percentage is set to 30%, and the percentage is consistently greater than 30% on any given day, it indicates that sutures are consistently associated with a large number of high-risk consumables. The system considers the presence of these high-risk consumables to have high monitoring reliability, and therefore, a "correlation-sensitive strategy" is adopted. If the percentage is less than 30% on any given day, the process proceeds to the next step.

[0084] The date on which the proportion of medical consumables requisitioned is not greater than the preset threshold for the proportion of medical consumables requisitioned is used as the verification risk date. It is determined whether the proportion of the number of verification risk dates is greater than the preset threshold for the proportion of verification risk dates. If so, the document management strategy for medical consumables is determined to be that whenever medical consumables requisitioned are found to be inconsistent with the demand, the requisition data of the medical consumables will be checked to determine whether there is a requisition deviation. If not, proceed to the next step. Risk Verification Dates: These refer to dates on which the proportion of a particular type of high-risk consumable falls below a preset threshold among the dates the target consumable was issued. On these days, the scenario where the target consumable is located lacks a sufficient number of high-risk consumables, leading to a decrease in the overall monitoring reliability of that scenario.

[0085] Preset threshold for the percentage of verification risk dates: A percentage threshold used to determine whether the proportion of "verification risk dates" among all requisition dates is too high.

[0086] If the target consumable is not surrounded by a single type of risky consumable every day, then we need to look at how frequently it is in scenarios with low monitoring reliability. If the proportion of risky verification dates (i.e., dates with low scenario monitoring reliability) is high, it indicates that the consumable spends a significant proportion of its time in an environment lacking high-intensity monitoring. In this case, the overall consumable system's monitoring reliability for that consumable is low, and therefore it needs to be upgraded to a more stringent global sensitivity strategy to compensate for the shortcomings of scenario monitoring.

[0087] Specific examples: Continuing the previous example, the sutures were used over 120 days, with less than 30% of them being Class I risk consumables (i.e., the verification risk dates). The preset threshold for the percentage of verification risk dates is set at 20%. 30 / 120 = 25%, which is greater than 20%, indicating that the sutures were used in scenarios with low monitoring reliability for a considerable proportion of the time, resulting in low overall system monitoring reliability. Therefore, a "global sensitivity strategy" is adopted. If the verification risk dates are only 18 days (15%), which is less than 20%, then proceed to the next step.

[0088] Based on the discrepancies in consumable requisition between different dates, dates with identical requisitions are grouped into the same group. It is then determined whether the number of groups exceeds a preset group size threshold. If so, the document management strategy for medical consumables is determined to be: whenever a medical consumable requisition is found to be inconsistent with the demand, the requisition data for that medical consumable is checked to determine if a requisition discrepancy exists. If not, the document management strategy for medical consumables is determined to be: whenever a medical consumable requisition is found to be inconsistent with the demand, and that inconsistent medical consumable is on the same document as the medical consumable, the requisition data for that medical consumable is checked to determine if the requisition record matches the actual outbound data and whether a requisition discrepancy exists.

[0089] Consumable requisition dates are completely identical: This refers to two or more dates on which the sets of consumables requisitioned are exactly the same.

[0090] Group: Group all dates that are "completely consistent with the requisition of consumables" into one group.

[0091] Preset group number threshold: An integer threshold used to determine the diversity of usage patterns.

[0092] This is the final assessment of the stability of the target consumable's usage scenario. If its usage pattern is highly diverse (many groups), it indicates that it may appear in various different consumable combinations, potentially causing confusion with different consumables. This variable usage scenario makes it difficult for the system to achieve stable and reliable coverage, reducing overall monitoring reliability. Therefore, even if its order frequency is low, the variable usage scenario reduces the system's monitoring reliability, and a global sensitivity strategy should be adopted. If its usage pattern is very simple (few groups), its usage scenario is relatively stable, monitoring reliability is high, and an association-based sensitivity strategy can be maintained.

[0093] Specific examples: Assuming the risk percentage of suture requisition dates is 15% (not exceeding the threshold), proceed to this step. The system compares all requisitioned consumable sets for all requisition dates and finds four completely identical combination patterns: Pattern A (appearing on the same day as ultrasound catheters and syringes), Pattern B (appearing on the same day as cardiac stents and angiography catheters), Pattern C (appearing on the same day as gauze and cotton swabs), and Pattern D (appearing alone). Therefore, the number of groups is 4. The preset threshold for the number of groups is set to 3. Since 4 > 3, it indicates that the suture requisition pattern is relatively discrete, the scenarios are varied, and the system monitoring reliability is low. Therefore, a "global sensitive strategy" is adopted. If the number of groups is 2 (≤ 3), it indicates that the requisition pattern is stable, the scenario monitoring reliability is high, and an "association sensitive strategy" is adopted.

[0094] It should be noted that during the verification process, the requisition records are matched with the actual outbound data to determine if any anomalies exist.

[0095] Furthermore, the abnormal requisition data includes the number of times the requisition data and demand data for the medical consumables are inconsistent.

[0096] Furthermore, the associated medical consumables with abnormal requisition are other medical consumables whose abnormal requisition is caused by the abnormal requisition of the aforementioned medical consumables. This abnormal requisition situation may occur due to proximity or label misidentification.

[0097] Specifically, such as Figure 4 As shown, the method for determining the impact type of abnormal requisition of medical consumables is as follows: The core decision-making objective of this solution is to assess the potential actual impact of requisition anomalies (such as mis-delivery or incorrect issuance) by analyzing the degree of overlap between the target medical consumable and its associated medical consumables in departmental usage. This allows for the classification of target consumables into different types of requisition anomaly impacts, providing a basis for subsequent differentiated management. The core logic is: the more associated consumables there are, the greater the probability of confusion and the greater the impact of shortages; the more discrete consumables (without fixed departments), the easier it is for the target consumable to be mistakenly received by non-needing departments, resulting in a greater impact on departments that actually need it when in short supply; conversely, if the departments using associated and target consumables highly overlap, even if mis-delivery occurs, the actual negative impact is relatively small because the mistakenly received consumable is also commonly used in that department. Through a progressively layered assessment, a refined classification of the impact of requisition anomalies can be achieved.

[0098] S31 uses the usage data of the associated medical consumables in different departments to determine the usage ratio of the associated medical consumables in the respective departments, and determines the associated departments of the associated medical consumables based on the usage ratio. It should be noted that the associated departments are those whose usage of the associated medical consumables accounts for a proportion greater than a preset usage proportion threshold.

[0099] Related medical consumables: refers to consumables that may be confused with the target medical consumables, such as those that are easily confused by the system due to similar appearance, similar names, or close storage locations.

[0100] Usage percentage: This refers to the proportion of a specific department's usage of a particular medical consumable relative to the total usage of that consumable.

[0101] Related departments: These refer to departments whose usage of a particular related medical consumable exceeds a preset threshold. These departments are the primary users of that related consumable.

[0102] Preset usage percentage threshold: A percentage threshold used to filter out departments with stable and significant demand for a certain consumable.

[0103] Each consumable has its primary users, and these departments are highly dependent on it. Identifying related departments by the percentage of usage quantifies the consumable's "home turf." When analyzing overlapping departments later, only these primary users are representative; departments with sporadic usage are excluded to avoid interference. The significance lies in focusing on the core user group, making the overlap analysis more practical.

[0104] Specific examples: For the associated medical consumable "disposable intravascular ultrasound catheter," the system statistically analyzes its usage data over the past year and calculates the usage percentage for each department. Cardiology accounts for 82%, interventional operating room accounts for 15%, and other departments account for 3% in total. If the preset usage percentage threshold is set to 10%, then cardiology and interventional operating room are identified as the associated departments for this consumable.

[0105] S32 Based on the degree of association between the associated department of the associated medical consumable and the department of the medical consumable, determine the department that simultaneously belongs to both the associated department of the associated medical consumable and the department of the medical consumable, and treat it as an overlapping associated department. Specifically, the overlapping and related departments are those departments that use the medical consumables and are associated with the related medical consumables.

[0106] Department of medical consumables: refers to the main department that uses the target medical consumables, which can be determined using the same method as S31 (the proportion of consumption exceeds the threshold).

[0107] Degree of correlation: refers to the size of the intersection between two department sets.

[0108] Overlapping related departments: These refer to departments that belong to both the related department of a particular medical consumable and the department of the target medical consumable. These departments are the primary users of both types of consumables.

[0109] Overlapping departments serve as a bridge connecting the target consumable with its related consumables. If a related consumable and the target consumable have a large number of overlapping departments, it indicates a strong symbiotic or substitutive relationship in clinical use. When an error occurs (such as mistakenly taking B instead of A), if the receiving department is also the primary user of B, the error may be masked or have a smaller impact. Its significance lies in quantifying this degree of overlap in usage.

[0110] Specific examples: The target medical consumable, "medical sutures," is distributed among general surgery, orthopedics, and cardiology. For the related consumable, "vascular sutures," the related departments are cardiology and vascular surgery. The intersection of these two is cardiology; therefore, cardiology is the overlapping related department.

[0111] S33 determines the type of abnormal requisition impact of the medical consumables based on the overlapping related departments in different related medical consumables and the related medical consumables of the medical consumables.

[0112] Impact Type of Abnormal Requisition: This refers to the classification of the potential severity of consequences when an error occurs in requisitioning the target consumables. This plan divides it into three impact types: Type 1 (High Risk), Type 2 (Medium Risk), and Type 3 (Low Risk).

[0113] This step is the comprehensive decision-making stage. Through progressive judgment in subsequent sub-steps, information such as the quantity, dispersion, and overlap of related consumables are integrated to arrive at a risk level reflecting the overall impact. Its significance lies in mapping multi-dimensional data to a single risk label, facilitating the formulation of subsequent management strategies.

[0114] The above steps include the following: S331 Obtain the quantity of associated medical consumables of the medical consumables, and determine whether the quantity of associated medical consumables of the medical consumables is greater than a preset threshold for the quantity of associated consumables. If yes, determine that the abnormal requisition impact type of the medical consumables is a type of impact. If no, proceed to step S332. Number of associated medical consumables: refers to the total number of consumable types that the system identifies as potentially confusing with the target consumable.

[0115] Preset threshold for associated consumable quantity: An integer threshold used to determine if there are too many sources of confusion.

[0116] Type 1 Impact: This indicates that abnormal requisition carries the highest risk and requires the strictest control.

[0117] If there are many types of consumables that could be confused with the target consumable, the probability of requisition errors is naturally high. In a shortage situation, once confusion occurs, the impact on meeting clinical needs is enormous, because the incorrect consumable may not be able to replace the function of the target consumable, or it may require complex traceability and exchange procedures. Therefore, when the number of associated consumables exceeds a threshold, it is directly classified as a Class 1 impact. This is the first rapid screening step used to identify high-risk consumables that are the "target of everyone's attention."

[0118] Specific examples: The target consumable, "disposable sterile syringe," has 15 associated consumables (such as syringes of different specifications, infusion sets, needles, etc.). The preset threshold for the number of associated consumables is 10. Since 15 > 10, its abnormal requisition impact type is directly determined to be a Class I impact type.

[0119] S332 Based on the associated departments of different associated medical consumables, determine the associated medical consumables that do not have associated departments and treat them as discrete medical consumables. Determine whether the quantity of the discrete medical consumables is greater than the preset discrete consumables quantity threshold. If yes, determine that the abnormal requisition impact type of the medical consumables is a type of impact. If no, proceed to step S333. Related medical consumables without associated departments: These refer to related consumables whose usage in all departments does not exceed a preset threshold. The use of these consumables is highly dispersed, with no clearly defined primary department of use.

[0120] Discrete medical consumables: The naming of the above-mentioned consumables.

[0121] Preset discrete consumable quantity threshold: An integer threshold used to determine if there are too many discrete obfuscation sources.

[0122] If a related medical consumable does not have a designated primary department of use, it indicates that it may be a generic consumable that any department could use. If this consumable is confused with the target consumable, it means that the target consumable may be mistakenly requisitioned by numerous departments that do not require it. In a shortage situation, the department that truly needs the consumable may be unable to obtain it due to other departments mistakenly requisitioning it, resulting in significant impact. Therefore, an excessive number of discrete medical consumables is also considered a type of impact.

[0123] Specific examples: The target consumable "medical gauze pads" has 5 associated consumables, of which 2 (such as cotton balls and bandages) have no associated departments (they are used evenly across departments) and are marked as discrete medical consumables. The preset threshold for the quantity of discrete consumables is 1. Since 2 > 1, it is determined to be a type of influence.

[0124] S333 determines the correlation coefficient of the associated medical consumables based on the proportion of overlapping associated departments in the associated departments of the associated medical consumables, and determines whether there are associated medical consumables with a correlation coefficient greater than a preset correlation coefficient threshold. If yes, proceed to step S334; otherwise, determine that the abnormal requisition impact type of the medical consumables is a type of impact. Correlation coefficient: For a given associated medical consumable, the correlation coefficient = number of overlapping associated departments of the consumable / number of associated departments of the consumable. It reflects the degree of overlap between the associated consumable and the target consumable in the main departments where they are used.

[0125] Preset correlation coefficient threshold: A percentage threshold used to determine whether the correlation is strong enough.

[0126] When the number of related consumables is small and their dispersion is also limited, it's necessary to further examine the departmental overlap between each related consumable and the target consumable. If the correlation coefficients of all related consumables are low, it means that even if a mistake occurs, the wrongly taken consumable is likely not commonly used by the receiving department, which would directly render the department unable to use it, having a significant impact in a shortage situation. Therefore, if no related consumables are strongly correlated, it is classified as a single type of impact. This step identifies the risk of confusion due to "completely unrelated" consumables.

[0127] Specific examples: The target consumable "medical sutures" has three associated consumables: A, B, and C. The correlation coefficients for A are 0.8, B is 0.6, and C is 0.2. The preset threshold for the correlation coefficient is 0.7. Since the correlation coefficient for A is 0.8 > 0.7, the process proceeds to step S334. If the correlation coefficients for all associated consumables are ≤ 0.7, the product is directly classified as a single type of influence.

[0128] S334 identifies medical consumables with a correlation coefficient greater than a preset correlation coefficient threshold as strongly correlated consumables, and determines whether the proportion of strongly correlated consumables in the total number of related medical consumables is greater than a preset strongly correlated proportion threshold. If so, the abnormal requisition impact type of the medical consumables is determined to be a third-class impact type. If not, proceed to step S335. Strongly related consumables: refers to related consumables with a correlation coefficient greater than a threshold, that is, consumables that highly overlap with the department of the target consumable.

[0129] Preset strong correlation percentage threshold: A percentage threshold used to determine whether strongly correlated consumables are dominant.

[0130] Three types of impact: This indicates that the impact risk of abnormal requisition is the lowest, and more lenient management measures can be taken.

[0131] If there are strongly related consumables, and they constitute the vast majority of all related consumables, it means that most of the potentially confused consumables share a common primary department of use with the target consumable. In this case, even if the wrong consumable is taken, the department can still use it because the wrongly taken consumable is also commonly used in that department, and the negative impact of the shortage is relatively small. Therefore, it is classified as a Category III impact. This step reflects the core idea that "high departmental overlap results in low impact."

[0132] Specific examples: The target consumable has 5 associated consumables, of which 4 have an association coefficient greater than 0.7 (strongly associated consumables). The preset threshold for the proportion of strong associations is 70%. 4 / 5 = 80% > 70%, therefore it is judged as a type 3 influence. If there are only 3 strongly associated consumables (60% ≤ 70%), then proceed to S335.

[0133] S335 determines the impact value based on the correlation coefficient of different related medical consumables and the proportion of discrete medical consumables in the related medical consumables. It then determines whether the impact value is greater than a preset impact threshold. If so, the abnormal requisition impact type of the medical consumables is determined to be a type II impact type. If not, the abnormal requisition impact type of the medical consumables is determined to be a type III impact type.

[0134] Impact Value: A comprehensive score calculated by combining the correlation coefficients (usually the average) of the related consumables and the proportion of discrete medical consumables. The higher the proportion of discrete consumables, the greater the impact value.

[0135] Preset impact threshold: A numerical threshold used to classify risks into Class II and Class III based on continuous impact values.

[0136] Category II impact type: This indicates that the impact risk of abnormal requisition is moderate.

[0137] When strongly correlated consumables exist but do not constitute an absolute majority, a comprehensive assessment of the overall situation is necessary. The average correlation coefficient reflects the overall overlap; the lower the average correlation coefficient, the more sources of confusion there are that do not match the target department, and the greater the impact. The discrete proportion reflects the impact of sources of confusion without a fixed department; the higher the discrete proportion, the more dispersed and difficult to trace the sources of confusion, and the higher the risk of mis-receiving by non-demanding departments, resulting in a greater impact. Combining these two factors yields a continuous impact value, which is then used to distinguish between medium and low risk. This step achieves refined quantification of intermediate states, reflecting the principle that "the more discrete consumables, the higher the risk."

[0138] Specific examples: The target consumable has 5 related consumables, including 3 strongly related ones, 2 weakly related ones (correlation coefficients of 0.3 and 0.4 respectively), and 1 discrete medical consumable (correlation coefficient counted as 0). The average correlation coefficient is calculated as (0.8 + 0.75 + 0.9 + 0.3 + 0.4) / 5 = 0.63, and the discrete percentage is 1 / 5 = 0.2. The impact value is calculated using the formula: Impact Value = (1 - Average Correlation Coefficient) × (1 + Discrete Percentage) = (1 - 0.63) × (1 + 0.2) = 0.37 × 1.2 = 0.444. The preset impact threshold is 0.5. Since 0.444 < 0.5, it is classified as a third-type impact. If the average correlation coefficient is lower (e.g., 0.4) and the discrete percentage is the same, then the impact value = 0.6 × 1.2 = 0.72 > 0.5, and it is classified as a second-type impact.

[0139] Furthermore, the method for determining the document management method for the aforementioned department is as follows: The core decision-making objective of this plan is to dynamically determine the document verification strategy for each department based on the proportion of Class I medical consumables affecting overall medical use in the hospital, the scale of sources of confusion for Class I risk consumables, the department's dependence on consumables, and the number of highly dependent consumables within the department. This ensures effective monitoring of high-risk consumables while prioritizing the acceleration of the identification efficiency of related consumables, thus accumulating data for subsequent management. The core logic is as follows: First, assess the overall verification reliability across the hospital. If the reliability is high, all departments adopt a uniformly lenient strategy. If the reliability is low, determine whether to accelerate identification across the hospital based on the scale of sources of confusion for Class I risk consumables. Then, based on the department's dependence on specific consumables, maintain a lenient strategy for departments with low dependence, while accelerating the identification of related consumables in departments with high dependence through comprehensive document verification.

[0140] S41 Based on the abnormal impact type of the requisition of medical consumables of the aforementioned risk type, determine the medical consumables of the aforementioned risk type with a certain impact type, and treat them as medical consumables with a certain impact. Category 1 risk-related medical consumables: refers to those produced using the aforementioned methods ( Figure 2 Medical consumables that were identified as having a shortage impact risk type 1.

[0141] Type of impact of abnormal requisition: refers to the impact caused by the aforementioned methods ( Figure 3 The types of abnormal impacts of the requisition of medical consumables are determined.

[0142] Medical consumables with a Class I impact type: refers to medical consumables whose abnormal requisition impact type is determined to be Class I.

[0143] Category 1 Impact on Medical Consumables: Medical consumables that simultaneously meet the criteria of "Category 1 Risk Type" and "Category 1 Impact Type" are the consumables with the highest risk and the greatest impact.

[0144] The highest-risk and most impactful consumables are identified and used as the basis for subsequent assessments of the reliability of hospital-wide verification. This is significant because it focuses on core risk points and uses the proportion of these consumables to evaluate the effectiveness of overall monitoring.

[0145] Specific examples: A hospital has five types of medical consumables classified as Class I risk: A, B, C, D, and E. Based on the classification of impact types caused by abnormal requisition, A and B are classified as Class I impact types, C and D as Class II, and E as Class III. Therefore, A and B are marked as Class I impact medical consumables.

[0146] S42 determines the requisition date of a certain type of risk medical consumable in the department based on the department's usage data for different medical consumables; Usage data: Detailed records of all types of medical consumables historically used by each department.

[0147] Issuance Date: For a specific department and a particular type of risky consumable, this refers to the exact calendar date on which the department actually issued the consumable. This provides the basic data for subsequently calculating the correlation coefficient between the department and the consumable issuance. The number of issuance dates directly reflects the frequency of the department's reliance on consumables.

[0148] Specific examples: For cardiology, the usage period for consumable A is 50 days, for consumable B it is 30 days, and for consumable C it is 20 days.

[0149] S43 determines the department's document management method based on one type of medical consumables affecting the first type of risk, and the department's requisition date for the first type of medical consumables.

[0150] Document Management Methods: Verification rules for departmental requisition documents include a "lenient strategy" (verification based solely on the consumables' own policy) and a "proactive strategy" (comprehensive verification of the entire document to expedite identification), combining hospital-wide risk characteristics with departmental usage characteristics to formulate differentiated departmental management rules.

[0151] Furthermore, if the proportion of a certain type of medical consumables affecting a certain risk level exceeds a preset threshold, then all related medical consumables affecting a certain risk level will be checked when they are requisitioned. This means determining whether the related medical consumables have caused an abnormal requisition of the medical consumables affecting a certain risk level. In this case, the overall verification reliability of the medical consumables affecting a certain risk level is relatively high. Therefore, the department's document management method is that unless the document contains related medical consumables affecting a certain risk level, verification is only required if the document management strategy for medical consumables is met.

[0152] The proportion of Class I impacted medical consumables: The proportion of Class I impacted medical consumables to the total number of Class I risk consumables across the entire hospital.

[0153] Preset threshold for the percentage of consumables affected: A threshold at the hospital level used to determine the overall reliability of verification.

[0154] Verification reliability: This refers to the extent to which existing verification mechanisms can effectively detect abnormalities in consumable requisition. A high proportion of consumables affecting a particular category indicates that most high-risk consumables are under strict monitoring, resulting in high overall reliability.

[0155] This is a hospital-wide assessment. If the proportion of Class I consumables is high across the hospital, it indicates strong overall verification reliability, and a uniform, lenient baseline strategy can be adopted for all departments: in addition to verifying related consumables of Class I consumables, other consumables are only verified if they meet their own document management policies. This avoids over-verification across the hospital.

[0156] Specific examples: There are 5 types of Class I risk consumables in the entire hospital, of which 4 are Class I impact consumables, accounting for 80%. The preset threshold for the percentage of impact consumables is 70%. Since 80% > 70%, the overall verification reliability is high. Therefore, a lenient strategy is adopted for all departments: only related consumables of Class I impact consumables are checked, while other consumables are handled according to their own strategies.

[0157] Furthermore, if the proportion of one type of medical consumables affecting a certain risk level is not greater than a preset threshold for the proportion of such consumables, then all related medical consumables affected by this risk level will be subject to verification upon requisition. In this case, the overall verification reliability of the medical consumables of this risk level is not high, and this also includes the following: S431 determines whether the average number of associated medical consumables of different risk types is less than the preset threshold for the number of associated consumables. If so, the number of associated medical consumables of a certain risk type is relatively small. Therefore, for all departments, as long as there is a document for a certain risk type of medical consumables, the outbound status of all medical consumables in the document is checked and processed to determine whether the outbound of medical consumables in this document has caused confusion in the outbound of medical consumables of a certain risk type. If not, proceed to step S432. Number of associated medical consumables: For each type of risk consumable, the number of associated consumables.

[0158] Average: The arithmetic mean of the number of associated consumables for all consumables of a certain type of risk.

[0159] Preset threshold for the number of associated consumables: Used to determine whether the size of the source of confusion is too small.

[0160] If a category of risky consumables has few varieties, the number of related consumables identified will inevitably be small. This means our understanding of the sources of confusion is still limited, and we need to accelerate the identification process. Therefore, we adopt a proactive strategy of comprehensive review of all orders across all departments. By expanding the scope of the review, we can quickly identify and record potential sources of confusion, accumulating data for subsequent correlation analysis. This reflects the principle of "strengthening detection when understanding is insufficient."

[0161] Specific examples: There are three types of risky consumables: A, B, and C. The quantities of associated consumables are 2, 3, and 2 respectively, with an average value of approximately 2.33. The preset threshold for the number of associated consumables is 3. Since 2.33 < 3, the scale of the source of confusion is considered small, and identification needs to be accelerated. Therefore, for all departments, if any document contains any of A, B, or C, all consumables on that document will be fully checked.

[0162] S432 determines the requisition correlation coefficient between the department and the medical consumables of the first risk type based on the proportion of the requisition dates of the department. It then determines whether there are any medical consumables of the first risk type with a requisition correlation coefficient greater than a preset requisition correlation coefficient threshold. If yes, proceed to step S433. If no, since there are no medical consumables of the first risk type with a requisition correlation coefficient greater than the preset requisition correlation coefficient threshold, the impact of a shortage of medical consumables of the first risk type is not significant. Therefore, in the department, all medical consumables of the first risk type are checked and processed when related medical consumables are requisitioned. It should be noted that if the requisition correlation coefficient is small, the impact of a shortage will be small. Therefore, in the aforementioned departments, all medical consumables affected by this correlation are subject to verification and processing when they are requisitioned. Requisition correlation coefficient: The proportion of the number of days a department requisitions a certain consumable to the total number of days the consumable is requisitioned in the entire hospital, reflecting the department's dependence on the consumable.

[0163] Preset requisition correlation coefficient threshold: Used to determine whether a department's dependence on a certain consumable is significant.

[0164] If a department's correlation coefficient for the requisition of a certain consumable is not high, it indicates that the consumable is evenly distributed throughout the hospital, and the department is not its primary user. Therefore, even if the consumable is in short supply or its requisition is abnormal, the impact on the department will be relatively small. In this case, the department does not need to adopt additional strict strategies; maintaining basic checks on consumables affecting this category and their related consumables is sufficient. This reflects the principle of "less impact, more lenient strategy."

[0165] Specific examples: The correlation coefficients for consumables A, B, and C in the cardiology department are 50%, 37.5%, and 16.7%, respectively. The preset threshold is 35%. If A and B exceed the threshold, proceed to step S433. If the correlation coefficients for all consumables are ≤35%, it is determined that there are no strongly correlated consumables, and the cardiology department only needs to maintain the basic strategy: check the related consumables that affect the consumables.

[0166] S433 identifies medical consumables of a risk type with a requisition correlation coefficient greater than a preset requisition correlation coefficient threshold as strongly correlated consumables. It then determines whether the quantity of strongly correlated consumables in the department exceeds a preset threshold. If so, a shortage of strongly correlated consumables would significantly impact the department's overall health. Therefore, in the department, for any document containing a risk type of medical consumables, the outbound status of all medical consumables in the document is checked to determine if the outbound issuance of medical consumables in this document has led to the confused outbound issuance of other risk type medical consumables. This allows for the rapid identification of associated medical consumables of different risk types. If not, in the department, all related medical consumables of a risk type are checked during requisition. Furthermore, all medical consumables in documents containing strongly correlated consumables are checked to determine if the outbound issuance of medical consumables in this document has led to the confused outbound issuance of other strongly correlated consumables, thus rapidly identifying the associated medical consumables of the strongly correlated consumables.

[0167] Strongly correlated consumables: Consumables with a correlation coefficient greater than the threshold are used, i.e. consumables that the department is highly dependent on.

[0168] Preset threshold for the quantity of highly dependent consumables: Used to determine whether there are too many highly dependent consumables in a department.

[0169] The department has a high dependence on strongly related consumables, and abnormal use of these consumables can have a significant impact on the department. Therefore, it is necessary to improve the efficiency of identifying related consumables to detect potential confusion as early as possible. If there are many strongly related consumables, it indicates that the department is highly dependent on multiple consumables. In this case, the most stringent comprehensive audit should be conducted, thoroughly examining all documents containing one type of high-risk consumable to identify all possible confusion relationships with maximum efficiency. If there are not many strongly related consumables, a compromise strategy should be adopted: in addition to basic audits, a comprehensive audit should be conducted on documents containing strongly related consumables to expedite the identification of related consumables for these highly dependent consumables. This reflects the principle of "accelerating identification when dependence is high."

[0170] Specific examples: The cardiology department has two strongly related consumables, A and B. The preset threshold for the number of strongly related consumables is 3. Since 2 < 3, a compromise strategy is adopted: a basic check is performed on the related consumables affecting one type (A and B), while a comprehensive check is performed on any document containing A or B to quickly identify the related consumables of A and B. If there are 4 strongly related consumables (> 3), the strictest strategy is adopted: a comprehensive check is performed on any document containing any type of risky consumable.

[0171] The hospital-wide and departmental two-level linkage enables precise policy implementation: This invention first assesses the overall verification reliability at the hospital level, and then formulates management methods based on the specific usage characteristics of each department, thus achieving an organic combination of macro and micro perspectives, making management strategies both consistent and targeted.

[0172] In this embodiment, the verification intensity is dynamically adjusted to accelerate data accumulation: when the number of related consumables identified is small, the proactive strategy of comprehensive verification of the entire order can quickly accumulate data on confusing relationships, providing rich samples for subsequent correlation analysis, which reflects the scientific management concept of "strengthening detection when there is insufficient understanding".

[0173] For consumables that departments rely heavily on, we will expedite the identification of related consumables through comprehensive order verification to ensure that the risk of confusion with these key consumables can be detected and recorded as early as possible, providing a basis for preventive measures. For departments and consumables with low reliance, we will maintain a lenient basic verification strategy to avoid conducting indiscriminate comprehensive verification of all scenarios, and concentrate limited management resources on high-value targets that truly need to be identified.

[0174] The decision-making logic is clear and configurable: all the thresholds in the method (preset thresholds for the proportion of consumable quantity, preset thresholds for related consumable quantity, preset thresholds for requisition correlation coefficient, and preset thresholds for strongly related consumable quantity) can be flexibly adjusted by the hospital according to its own size, management goals and data accumulation, making the method highly adaptable and scalable.

[0175] Example 2 Secondly, the present invention provides an intelligent device that employs the aforementioned AI model-based medical consumables document management method, specifically including: The module includes a management strategy determination module, an impact type output module, and a management method output module. The management strategy determination module is responsible for determining the document management strategy for medical consumables. The impact type output module is responsible for determining the impact type of the abnormal requisition of the medical consumables; The management method output module is responsible for determining the document management method for the department.

[0176] Example 3 Thirdly, this application provides a chip for use in a smart device, comprising: a memory and a processor connected in communication, and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the aforementioned AI model-based medical consumables document management method when running the computer program.

[0177] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0178] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0179] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A method for managing medical consumables invoices based on an AI model, characterized in that, Specifically, it includes: Based on the requisition data of medical consumables, the impact data of the shortage of medical consumables on different departments is determined. Based on the impact data and the AI ​​model, the shortage impact risk type of medical consumables is determined. Based on the shortage impact risk type of medical consumables and combined with the dispersion of the requisition data of medical consumables on different dates, the document management strategy of medical consumables is determined. The document management strategy is used to manage and process the document data of the medical consumables to obtain abnormal requisition data. Based on the abnormal requisition data, the associated medical consumables with abnormal requisition are identified. According to the degree of correlation between the requisition data of the associated medical consumables in different departments and the requisition data of the medical consumables, the impact type of abnormal requisition of the medical consumables is determined. Based on the impact of abnormal requisition of medical consumables of the aforementioned risk type, and in conjunction with the requisition data of the department for medical consumables of different risk types, the document management method for the department is determined.

2. The medical consumables document management method based on an AI model as described in claim 1, characterized in that, The medical consumables requisition data includes the requisition data of the medical consumables on different dates.

3. The medical consumables document management method based on an AI model as described in claim 1, characterized in that, The impact data for different departments is determined based on the department's requisition data on different dates.

4. The medical consumables document management method based on an AI model as described in claim 1, characterized in that, The method for determining the risk type of the shortage of medical consumables is as follows: Based on the impact data of the shortage of medical consumables on different departments, the requisition data of the departments on different dates is determined; Use the date on which the requisition data exists as the requisition date; Based on the requisition date data from different departments and the AI ​​model, the risk type of the shortage impact of the medical consumables is determined.

5. The medical consumables document management method based on an AI model as described in claim 4, characterized in that, Based on the requisition date data from different departments, the types of risks associated with the shortage of the aforementioned medical consumables are determined, specifically including: Based on the proportion of requisition dates in the department, the shortage impact value of the department is determined. If it is determined that there is a department whose shortage impact value is greater than a preset impact threshold, then the shortage impact risk type of the medical consumables is determined to be a type of risk.

6. The medical consumables document management method based on an AI model as described in claim 1, characterized in that, The method for determining the document management strategy for the aforementioned medical consumables is as follows: Based on the medical consumables requisition data for different dates, identify the medical consumables requisitioned on different dates and use them as the requisitioned consumables; Based on the dispersion of consumable requisition on different dates, determine the deviation of consumable requisition between different dates; Based on the risk type of shortage impact of the medical consumables, the requisition data on different dates, and the deviation of requisition data between different dates, the document management strategy for the medical consumables is determined.

7. The medical consumables document management method based on an AI model as described in claim 6, characterized in that, The document management strategy for the aforementioned medical consumables includes: If the risk type of the shortage of medical consumables is classified as a Class I risk, then the document management strategy for the medical consumables is determined to be that whenever medical consumables that are not in accordance with the demand are found to be requisitioned, the requisition data of the medical consumables will be checked and processed, that is, whether the requisition record is consistent with the actual outbound data, and whether there is a requisition deviation.

8. The medical consumables document management method based on an AI model as described in claim 1, characterized in that, The method for determining the document management method for the aforementioned department is as follows: Based on the abnormal requisition impact type of the medical consumables of the aforementioned risk type, determine the medical consumables of the aforementioned risk type with a type of impact, and classify them as medical consumables with a type of impact. Based on the usage data of different medical consumables in the department, determine the requisition date of a certain type of risk medical consumable in the department; Based on one type of medical consumables with a risk level, and the requisition date of the medical consumables with a risk level in that department, the document management method of that department is determined.

9. A smart device employing the AI-based medical consumables document management method according to any one of claims 1-8, characterized in that, Specifically, it includes: The module includes a management strategy determination module, an impact type output module, and a management method output module. The management strategy determination module is responsible for determining the document management strategy for medical consumables. The impact type output module is responsible for determining the impact type of the abnormal requisition of the medical consumables; The management method output module is responsible for determining the document management method for the department.

10. A chip, used in a smart device according to claim 9, comprising: A memory and processor connected by communication, and a computer program stored on the memory and capable of running on the processor, characterized in that, when the processor runs the computer program, it executes a medical consumables document management method based on an AI model as described in any one of claims 1-8.

Images