Intelligent evaluation method and system based on test item billing rationality

Abstract

The application provides a kind of based on the intelligent evaluation method and system of test item opening rationality, the method includes: based on the identification of examinee, obtain the diagnosis and treatment data in preset historical period, based on current vital sign monitoring information, generate diagnosis and treatment time sequence track with request time stamp as terminal point;Determine the reference decay function based on the stability of historical observation characteristics, generate dynamic correction coefficient according to key intervention event, execute nonlinear adjustment to reference decay function by dynamic correction coefficient, obtain the time limit confidence of test item;Real-time acquisition and test item associated test equipment dynamic operation data, generate resource matching degree, calculate opening rationalization quantitative score based on spatial logic mapping algorithm, according to the comparison result of rationalization quantitative score and preset risk threshold, dynamically adjust the confirmation control state of business initiation end, the application improves the test item opening response capability of medical system to acute condition or unexpected situation, reduces the error judgment of doctor due to excessive dependence on historical data.

Description

Technical Field

[0001] This application relates to the field of intelligent decision-making technology for medical information, specifically a method and system for intelligent evaluation of the rationality of test item prescriptions. Background Technology

[0002] With the widespread application of smart healthcare, clinical decision support systems have become an important tool for improving compliance in diagnosis and treatment. Existing test ordering schemes typically intercept redundant medical orders based on preset clinical pathways or fixed testing cycles, such as prohibiting the re-ordering of the same type of test within 3 days. The current approach compares the historical records in electronic medical records with the logical relationship of the current test order to output simple warning prompts, aiming to reduce medical costs and optimize the allocation of medical resources.

[0003] The current assessment method only focuses on the time difference of historical tests, ignoring the real-time fluctuations in the examinee's physiological signs. For example, for critically ill patients whose conditions change drastically, even test results from several hours ago may become invalid if their vital signs are significantly abnormal now. Furthermore, allowing doctors to order tests even when equipment is experiencing reagent shortages, hardware maintenance, or extreme congestion leads to excessively long sample dwell times, affecting test accuracy, wasting medical resources, and increasing patient waiting costs.

[0004] Therefore, optimizing the utilization rate of medical resources and improving the compliance of diagnosis and treatment have become key issues that need to be addressed. Summary of the Invention

[0005] To overcome the aforementioned problems in the prior art, this application provides a method and system, which adopts the following technical solution:

[0006] Firstly, this application provides an intelligent evaluation method for the reasonableness of ordering inspection items, including:

[0007] Based on the communication request for order placement for a test item, the system captures the request message of the order placement instruction. By parsing the structured request message, it obtains the examinee's identifier, the proposed test item, and the request timestamp. Based on the examinee's identifier, it acquires treatment data within a preset historical period. Based on current vital sign monitoring information, it generates a treatment time-series trajectory ending at the request timestamp. By parsing the treatment time-series trajectory, it obtains historical observation characteristics. Based on the stability of these historical observation characteristics, it determines a baseline decay function. Dynamic correction coefficients are generated based on key intervention events. These dynamic correction coefficients are used to perform nonlinear adjustment on the baseline decay function to obtain the timeliness confidence of existing test results for the proposed test item. It acquires real-time dynamic operating data of the testing equipment associated with the proposed test item and generates a resource matching degree. The timeliness confidence degree, resource matching degree, and current vital sign monitoring information are mapped to a preset decision coordinate space. A spatial logic mapping algorithm is used to calculate a reasonable quantitative score for order placement. Based on the comparison between the reasonable quantitative score and a preset risk threshold, the system dynamically adjusts the confirmation control status at the business initiation end.

[0008] Furthermore, based on the communication request for issuing an inspection order, the system captures the order issuance instruction request message, and obtains the inspectee identifier, the proposed inspection item, and the request timestamp based on structured parsing of the request message. This includes: real-time detection of the asynchronous communication link between the business initiator and the backend server; non-intrusive data frame capture at the network layer based on a preset traffic filtering algorithm, and identification of the order issuance instruction request message through feature point matching; and parsing the request message through a multi-level protocol engine to obtain the inspectee identifier, the proposed inspection item, and the request timestamp.

[0009] Secondly, this application also provides an intelligent evaluation system for the reasonableness of ordering inspection items, including:

[0010] The data parsing module is used to capture the request message of the order issuance instruction based on the communication request for the inspection item order, and obtain the inspectee's identifier, the item to be inspected, and the request timestamp based on the structured parsing of the request message.

[0011] The diagnosis and treatment timeline generation module is used to obtain diagnosis and treatment data within a preset historical period based on the examinee's identifier, and generate a diagnosis and treatment timeline trajectory with the requested timestamp as the endpoint based on the current vital sign monitoring information.

[0012] The timeliness confidence acquisition module is used to obtain historical observation characteristics by analyzing the diagnosis and treatment timeline trajectory, determine the benchmark decay function based on the stability of the historical observation characteristics, generate dynamic correction coefficients according to key intervention events, perform nonlinear adjustment on the benchmark decay function through the dynamic correction coefficients, and obtain the timeliness confidence of the existing test results for the test item.

[0013] The resource matching degree module is used to obtain the dynamic operating data of the testing equipment associated with the proposed testing project in real time and generate the resource matching degree.

[0014] The order rationality assessment module maps timeliness confidence, resource matching degree, and current vital sign monitoring information to a preset decision coordinate space, calculates the order rationality quantitative score based on the spatial logic mapping algorithm, and dynamically adjusts the confirmation control status at the business initiation end based on the comparison result of the rational quantitative score and the preset risk threshold.

[0015] Thirdly, this application provides an electronic device, comprising:

[0016] One or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the device, cause the device to perform the method as described in the first aspect.

[0017] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described in the first aspect.

[0018] Fifthly, this application provides a computer program that, when executed by a computer, performs the method described in the first aspect.

[0019] In one possible design, the program in the fifth aspect can be stored wholly or partially on a storage medium packaged with the processor, or it can be stored wholly or partially on a memory not packaged with the processor.

[0020] This application has the following beneficial effects:

[0021] 1. This application assesses the timeliness of historical test data under the current physiological state by nonlinearly adjusting the dynamic correction coefficient and the benchmark decay function, avoiding reliance on outdated historical data in acute or severe cases. Especially in the case of critically ill patients, this application can automatically adjust the weight of historical data, reducing the risk of relying on invalid or outdated data. This application improves the accuracy of clinical decision-making through real-time feedback based on physiological fluctuations and the decay method of historical test data, especially in complex cases, and can provide more scientific and real-time diagnosis and treatment suggestions.

[0022] 2. This application generates resource matching scores by acquiring real-time dynamic operating data of testing equipment. When testing equipment resources are scarce, it automatically prompts doctors to adjust the priority of testing items or postpone testing tasks, avoiding misdiagnosis, test failure, or waste of resources due to equipment overload or insufficient reagents. This application also dynamically adjusts prescription suggestions by sensing equipment maintenance cycles and task congestion, reducing waste of medical resources and patient waiting time.

[0023] 3. This application dynamically maps changes in the patient's physiological data into the decision space based on real-time vital sign monitoring information, and adjusts the timeliness of test results in combination with changes in current vital signs; when vital signs fluctuate significantly, this application will automatically adjust the reference weight of relevant test results to adapt to rapid changes in the condition.

[0024] 4. This application uses a spatial logic mapping algorithm to map timeliness confidence, resource matching degree, and vital sign monitoring information to the decision coordinate space, calculates a reasonable quantitative score for order placement, evaluates the rationality of each order placement request, and dynamically adjusts the order confirmation control status by comparing it with a preset risk threshold. This application can not only evaluate the timeliness and feasibility of each request, but also consider multiple factors such as equipment load and resource utilization, providing multi-dimensional decision support.

[0025] 5. This application automatically adjusts the decision-making process according to different situations by comprehensively evaluating timeliness, resource availability and physiological status from multiple dimensions; in emergency or critical cases, patients can avoid unnecessary medical waiting and delays caused by equipment resource limitations, inaccurate historical data or failure to reflect physiological fluctuations in a timely manner. Attached Figure Description

[0026] Figure 1 This is a flowchart of the intelligent evaluation method for the rationality of inspection item invoicing according to an embodiment of this application;

[0027] Figure 2 This is a flowchart of the structured parsing request message according to an embodiment of this application;

[0028] Figure 3 This is a schematic diagram of the process for obtaining timeliness confidence in an embodiment of this application;

[0029] Figure 4 This is a schematic diagram illustrating the specific implementation process of obtaining timeliness confidence in an embodiment of this application;

[0030] Figure 5 This is a schematic diagram of the system flow according to an embodiment of this application. Detailed Implementation

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects and are not used to describe a particular order.

[0032] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily indicate the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0033] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0034] Please refer to Figure 1 The following is a flowchart of an intelligent evaluation method for the rationality of inspection item invoicing, provided as an embodiment of this application. The specific implementation process is as follows:

[0035] Step 110: Based on the communication request for issuing an inspection order, capture the request message of the order issuance instruction, and obtain the inspectee identifier, the inspection item to be inspected, and the request timestamp based on the structured parsing of the request message.

[0036] For details, please refer to Figure 2 The following is a detailed implementation process of the structured parsing request message provided in the embodiments of this application:

[0037] Step 21: Real-time detection of the asynchronous communication link between the service initiator and the backend server; based on a preset traffic filtering algorithm, non-intrusive data frame capture is performed at the network layer.

[0038] Step 22 involves identifying the request message for issuing a test order through feature point matching. The request message is then parsed using a multi-level protocol engine to obtain the examinee's identifier, the proposed test item, and the request timestamp. The parsing process also includes establishing a temporal causal chain between the parsed fields and verifying the message sequence number to ensure a strong logical correlation between the examinee information and the test item obtained in an asynchronous network environment. Feature points are characteristic code fragments in the data message sequence. This application identifies these feature code fragments amidst the massive and chaotic flow of medical networks, enabling rapid identification of test order instructions from doctors within a preset time period. For example, by real-time monitoring of the data stream and scanning the bitstream at fixed offset positions in the message header, when the feature point POST / lis / api is scanned, it can be determined as a test order request. Scanning the feature point action:create at the Payload offset of 128 bits indicates a new instruction, and scanning item_code:0502 indicates a blood routine test item. However, if all the above feature points are met, the corresponding fragmented message will be reassembled into a structured order placement object. This application, through feature point matching, can complete the reasonable pre-review of medical orders simultaneously without parsing all the data and without affecting the response speed of the medical system.

[0039] In one possible implementation, the asynchronous communication link between the service initiator and the backend server is detected in real time. Based on a preset traffic filtering algorithm, non-intrusive data frame capture is performed at the network layer, including: obtaining a copy of the original request message based on the traffic mirror node; using the extended Berkeley packet filter program mounted on the trace point of the kernel protocol stack to perform signature collision on the incoming request message packets; when a specific protocol header (such as the HL7L3 layer message identifier) ​​that matches the characteristics of the test items prescribed by the doctor is detected, cloning is triggered; the request message is intercepted in real time; the intercepted instruction characteristics are asynchronously reported to the application layer via the kernel buffer, thereby achieving accurate capture of high-concurrency request messages without any intrusion into the main business flow.

[0040] In this embodiment, the preset traffic filtering algorithm is based on deep packet inspection with bitmap offset. It employs a filtering operator array pre-configured in the kernel space, containing service protocol feature fingerprints, to perform a one-way transparent scan of the massive packet stream in the mirror path. The preset traffic filtering algorithm identifies the diagnostic and treatment transaction feature components in the packet payload, enabling accurate reassembly of test order request packets while isolating noise traffic unrelated to the target service. The traffic mirror node is deployed on the core interactive machine or aggregation switch of the hospital's LAN. It acts as a bypass point for network traffic, not processing or forwarding parameter data, but only monitoring and copying. This application obtains packets through the traffic mirror node without consuming the CPU of the hospital information system server. When parsing is slow, it captures the lag caused by the doctor's order processing node. When the data packet passes through the switch, it can be copied immediately. A copy of the original request packet is obtained from the mirror node. Since it is the original bitstream, there is no need to wait for a server response; it can directly enter the feature point matching stage to identify the order processing instruction.

[0041] In this embodiment, deep packet probing with bitmap offsets locks the key feature fields of service packets using a preset offset vector table, and performs non-blocking scanning of the target fields using bitmap masks. This application transforms complex semantic recognition into bitwise operation matching through deep packet probing with bitmap offsets, enabling instantaneous recognition of heterogeneous medical protocol packets at the kernel-mode protocol stack. This ensures the integrity of capturing massive requests while reducing additional system latency.

[0042] In this embodiment, the offset vector table is an addressing index structure for fast data location, and is a data organization and mapping structure. For example, when processing high-concurrency hospital network packets, it is not necessary to read the entire data packet from beginning to end. Instead, the offset vector table is directly consulted. When the packet enters the network card, the patient ID is obtained in the 42nd byte by querying the offset vector table. The probe jumps to the 42nd byte to extract 16 bits of data, and then jumps to the 128th byte to extract the test item code. Different medical protocols correspond to different offset vector tables. By changing the offset vector table, compatibility with HIS systems from different manufacturers can be achieved without changing the parsing engine.

[0043] In this embodiment, by deploying traffic mirroring nodes on the physical switching side, a bypass listening path parallel to the main business link is established to achieve unidirectional isolation and replication of business traffic; by extending the Berkeley packet filter to build a non-intrusive probe operator, tracking points are mounted in real time in the operating system kernel mode to clone and intercept asynchronous communication packets generated by the business initiator in situ; asynchronous distribution of packets is achieved based on the kernel buffer, ensuring that data does not interfere with the main business logic during the capture process and the execution process, improving the transparency and efficiency of data collection, and effectively avoiding the single point of failure risk of traditional interception methods.

[0044] It should be noted that this application, by extending the Berkeley packet filter, can directly filter request messages in kernel space, unpacking only packets that meet the characteristics of the verification project application and uploading them to user space, thus reducing the frequent data copying overhead between user space and kernel space. User space directly uses operating system resources, resulting in high execution efficiency; kernel space requires system calls and interaction with the kernel during execution, leading to lower efficiency.

[0045] It's important to note that a traffic mirroring node is a physical port or function of a network device that copies all raw data packets from a specified port and sends them to a pre-defined analysis node. The extended Berkeley packet filter can execute directly in kernel space without interrupting or interfering with the main business process. Mounted trace points can intercept and clone request packets as they pass through. Data packets captured by the extended Berkeley packet filter within kernel space are stored in a kernel buffer, awaiting retrieval by user-space business programs.

[0046] It should be noted that the business initiation end includes clinical terminal systems, electronic medical record systems, and vital sign monitoring systems. The clinical terminal system receives test requests from doctors and nurses, including the patient's basic information, the proposed test items, and the request timestamp. The electronic medical record system includes the patient's historical medical record data, such as historical test results, diagnostic information, and medication records. The vital sign monitoring system includes real-time collected patient vital sign data, such as blood pressure, heart rate, body temperature, and blood oxygen saturation.

[0047] For example, when processing real-time requests from outpatient doctor workstations, communication requests involving Order_Entry actions are captured from massive amounts of network data packets. These request messages record the doctor's instructions for specific patients' tests, with the aim of making reasonable pre-emptive interventions before the data enters the database. This application ensures that the evaluation system intercepts or alerts the test requests before they officially take effect through real-time capture, thereby avoiding unnecessary over-testing or inefficient testing.

[0048] For example, suppose an emergency room doctor issues a test order for troponin I and cardiac enzyme profile to a patient suspected of having an acute myocardial infarction at 10:15:30 on 2025-11-20. When the communication message is captured, its core payload contains the subject identifier ID_789, the test item set {cTnI, CK-MB}, and the corresponding request time.

[0049] Step 120: Based on the examinee's identifier, obtain the diagnosis and treatment data within a preset historical period, and based on the current vital sign monitoring information, generate a diagnosis and treatment timeline trajectory with the requested timestamp as the endpoint.

[0050] The extraction of medical data within the preset historical period includes: extracting historical test results and medical information from the hospital information management system and electronic medical record database based on the examinee's identification. This data includes, but is not limited to, test items, test times, diagnostic conclusions, and prescription records. Current vital sign monitoring information includes current vital sign monitoring data collected through medical devices. Data collection and processing in this application comply with hospital management regulations and patient privacy protection regulations, and data is collected according to the hospital's specific specifications.

[0051] Specifically, after receiving a test order request, the system retrieves the patient's medical data from the hospital information system within a preset historical period using the patient's identifier. This data is then integrated with the currently collected vital sign monitoring information to construct a complete timeline of the patient's medical treatment. The timeline uses the timestamp of the current test request as the endpoint and connects historical medical events with the current status information in chronological order to form structured trajectory data representing the changes in the patient's medical treatment within a preset time period.

[0052] For example, when a doctor submits a test order for a patient, the doctor accesses the electronic medical record system, the testing system, and the vital signs monitoring system based on the patient's identification, such as patient ID or medical card number, to obtain the patient's medical records from the past period. The medical records may include previous test results, diagnostic information, medication records, surgical records, and vital signs change data, providing a basic data source for assessing the rationality of test order orders.

[0053] Specifically, retrieving medical data within a preset historical period involves extracting historical records associated with the examinee's identifier from the medical database based on a pre-defined time range, such as the past 7 days, 30 days, or a single hospitalization cycle. This medical data may include fields such as test item name, ordering department, test result value, diagnosis code, medication records, and treatment operation time. Through a data interface or data query module, the medical data information is extracted and organized into a unified data object structure, such as a JSON-formatted collection of medical records. Each record contains the event time, event category, and corresponding medical information. This application uses a unified data object structure to integrate medical data scattered across different systems, avoiding information loss caused by data silos.

[0054] For example, a subject may have undergone blood routine tests, liver function tests, and blood glucose tests in the past week, and may also have been diagnosed with hypertension in an outpatient clinic. These historical records are stored in the laboratory system and the electronic medical record system, respectively. After being linked by the subject's identifier, a list of medical records sorted by time is formed, such as: [(October 10, 2025, blood routine test), (October 12, 2025, liver function test), (October 14, 2025, hypertension diagnosis record)].

[0055] Specifically, based on current vital sign monitoring information, the currently collected vital sign data is integrated with historical medical data. Vital sign monitoring information may include information such as body temperature, blood pressure, heart rate, respiratory rate, and blood oxygen saturation. Vital sign monitoring information data usually comes from ward monitoring equipment or outpatient collection terminals. Real-time monitoring data is obtained through data acquisition interfaces and added to the medical data set as current status nodes to form a comprehensive data structure that includes historical information and current status.

[0056] For example, when a doctor requests a blood glucose test for a patient, the system can simultaneously obtain currently collected vital sign data, such as blood pressure of 150 / 95 mmHg, heart rate of 88 beats / minute, and body temperature of 36.7℃. This currently collected vital sign data, together with historical test records, constitutes comprehensive diagnostic information, reflecting the patient's current physical condition and recent trends in their medical history.

[0057] Specifically, constructing the time-series trajectory of diagnosis and treatment involves sorting historical diagnosis and treatment records and current vital sign data according to the time of event occurrence, using the timestamp of the current test order request as the endpoint node to form a continuous time series structure; through the time sorting module, various diagnosis and treatment events are arranged in ascending order according to the time field, and each event node is attached with an event category identifier, such as test record, diagnosis record, vital sign record, etc., forming a time-series trajectory structure with a unified format.

[0058] For example, the treatment timeline of a patient can be represented as: [(October 10, 2025, 09:20, routine blood test result), (October 12, 2025, 10:15, liver function test result), (October 14, 2025, 14:00, hypertension diagnosis), (October 15, 2025, 08:30, blood pressure monitoring data), (October 15, 2025, 09:10, test order request)]; the last node is the timestamp of the test order request, which serves as the endpoint of the entire timeline and is used to identify the evaluation time.

[0059] In this application embodiment, by constructing a treatment timeline trajectory with the request timestamp as the endpoint, it is possible to understand the changing trends of the examinee during recent treatment processes, such as changes in test results, fluctuations in vital signs, and updates to diagnostic information. This application provides complete contextual information for the rationality assessment of test order placement through structured test items, enabling the assessment results to combine historical data with the current status to determine whether the test request involves duplicate, unnecessary, or abnormal order placement. For example, when it is found that an examinee has completed the same test item within a short period of time and the results are stable, and the examinee requests the same test again, the duplicate behavior can be quickly identified based on the treatment timeline trajectory, triggering a rationality check in the subsequent assessment stage.

[0060] Step 130: Obtain historical observation features by analyzing the diagnosis and treatment timeline trajectory, determine the benchmark decay function based on the stability of the historical observation features, generate dynamic correction coefficients according to key intervention events, perform nonlinear adjustment on the benchmark decay function through the dynamic correction coefficients, and obtain the timeliness confidence of the existing test results for the item to be tested.

[0061] For details, please refer to Figure 3 The process of obtaining timeliness confidence in this application embodiment is as follows:

[0062] Step 31 involves analyzing the test value sequence, physiological fluctuation slope, and key intervention event markers in the treatment timeline to obtain structured observation data. The test value sequence represents the trend of the target physiological indicator within the observation period, the physiological fluctuation slope represents the rate of change of vital signs, and the intervention event markers are used to record clinical intervention behaviors that may affect the changes in the indicators.

[0063] Specifically, when parsing historical observation records, observation indicators such as blood glucose, observation time such as 2025-05-01, and observation results such as 6.2 mmol / L are extracted. The observation indicators, observation time, and observation results are separated from the original medical logs and transformed into structured data objects, such as JSON format data structures. Each observation record contains an indicator name, timestamp, and numerical result, transforming the original medical records into computable time series data.

[0064] Step 32: Calculate the discrete characteristics between adjacent test values ​​of the target observation item, and determine the stable state of the indicator by combining the physiological fluctuation slope. When the discrete characteristics and the physiological fluctuation slope meet the preset steady-state threshold, map the physiological half-life of the corresponding physiological indicator to a benchmark decay function. The benchmark decay function = ,in This represents the time difference from the historical verification point to the current timestamp. This represents the half-life of the corresponding physiological indicator. Where the slope of the physiological fluctuation = ,in They represent the first time. With the The test value at time. respectively in the With the A timestamp at a given moment.

[0065] It should be noted that the discrete features in this application include, but are not limited to, coefficient of variation, standard deviation, and the difference between the maximum and minimum values.

[0066] For example, when obtaining a patient's medical records, the patient's test results, such as complete blood count, liver function tests, or blood glucose test results, are retrieved from the electronic medical record. These test result records typically include the test time, the test index, and the test value. The test results are used to form a time-ordered observation sequence. Suppose a patient underwent blood glucose tests on May 1, 2025, May 10, 2025, and May 20, 2025, with corresponding results of 6.2 mmol / L, 6.3 mmol / L, and 6.4 mmol / L. These records constitute a historical observation sequence of the blood glucose test index. By calculating the variation range between adjacent test results and the test time interval, it can be determined that the blood glucose test index is in a relatively stable state over a period of time, thus establishing a relatively gradual baseline decay function.

[0067] Step 33: Key intervention events are transformed into dynamic correction coefficients through feature association. The baseline attenuation function is nonlinearly adjusted using these dynamic correction coefficients to obtain the enhanced attenuation function. The time difference is then substituted into the enhanced attenuation function to calculate the remaining effective component of the existing test results relative to the current requested timestamp, thus obtaining the timeliness confidence of the existing test results for the item to be tested. The timeliness confidence represents the reference validity of historical test results in the current diagnostic and treatment decision.

[0068] For details, please refer to Figure 4 The specific process for obtaining the timeliness confidence level is as follows:

[0069] Step 331 involves performing feature analysis on key intervention events to obtain the event attributes of each key intervention event. Based on the correlation between the event attributes and the project to be tested, feature association analysis is performed to determine the feature association value of each key intervention event on the timeliness of the results of the project to be tested. The event attributes of each key intervention event include, but are not limited to, event category, occurrence time, target group, impact intensity, and duration.

[0070] The feature parsing process can be implemented using an event terminology dictionary, a medical intervention coding table, a natural language processing model, and rule templates. Specifically, intervention events are standardized and classified based on keyword matching or standard mapping. Attribute extraction is used to extract fields such as event intensity, frequency, duration, and execution time from medical orders, nursing records, or surgical records. The extracted results are then encapsulated into an event attribute vector. This event attribute vector can be represented as {event category, direction of action, affected objects, intensity level, duration, start time, end time}, etc.

[0071] Feature association analysis based on the relationship between event attributes and the proposed test items refers to using a pre-defined knowledge base of associations between intervention events and test items to calculate the degree of influence of key intervention events on the timeliness of the results of the current proposed test items. The association knowledge base can be pre-constructed as a two-dimensional mapping table, knowledge graph, or rule network to record the influence relationship, direction of influence, and level of influence between different event attributes and different test items. For example, the pre-defined rules may include: when the event category is blood transfusion and the proposed test item is complete blood count, the association level is set to high; when the event category is antipyretic drug use and the proposed test item is inflammatory marker detection, the association level is set to medium; when the event category is fluid resuscitation and the proposed test item is electrolyte detection, the association level is set to high; when there is no direct influence between the event category and the proposed test item, the association level is set to low or zero.

[0072] Feature association values ​​are generated by weighting event attribute matching scores, medical rule matching scores, and historical statistical influence scores. The calculation method is as follows: ,in Indicates the first Feature correlation values ​​corresponding to key intervention events This represents the score indicating the match between the event category and the rule for the item to be tested. The semantic similarity score between the mechanism of action and the test index is used to represent the similarity score. This represents the impact strength score obtained based on historical case statistics. The weighting coefficients are preset. The event attribute matching score represents the degree of matching between the physical characteristics of the intervention event and the target observation. The medical rule matching score represents a strong correlation in medical logic, obtained by matching a knowledge graph formed by digitizing authoritative clinical guidelines and expert consensus. For example, when assessing the rationality of blood glucose testing, if it is detected that a patient has just used dexamethasone, the medical rule will immediately output a very high correction strength score because the drug has a clear blood glucose-raising side effect, which falls within the scope of common medical knowledge. The historical statistical influence score is obtained through empirical evidence from big data mining, used to supplement implicit associations or individual differences not yet covered by the medical rule. For example, if historical data shows that 85% of the subjects experienced a numerical shift of more than 10% after intervention, the historical statistical influence score is set to 0.85. This acquisition method corrects the rigidity of pure rules through empirical probability, making the feature association values ​​more robust to noise.

[0073] For example, a patient underwent a serum potassium test on August 1, 2025, with a result of 4.3 mmol / L. During subsequent treatment, key intervention events were recorded, including intravenous potassium supplementation and diuretic adjustments. These key intervention events may directly affect the stability and timeliness of serum potassium levels, thus requiring inclusion in dynamic correction analysis. During feature parsing, the event category field (medication adjustment), event time field (2025-08-02 10:00:00), affected object field (electrolyte balance), and impact intensity field (high intensity) of the key intervention events were extracted and structured. After structuring, each event is represented as a data object in a unified format, such as an event feature vector including {event category, occurrence time, affected object, impact level}, providing a data foundation for subsequent feature association.

[0074] Step 332 involves subtracting the event occurrence time from the current request timestamp to obtain the time difference. Based on this time difference and the event's impact intensity, a time-sensitive weight is determined. A dynamic correction coefficient is then generated using the feature association value and the time-sensitive weight. Step 332 obtains the dynamic correction coefficient for a single event.

[0075] In this embodiment, the time-sensitive weight is determined based on the time difference and the intensity of the event's impact, which can be expressed as follows: ,in Time-sensitive weights For the first The time difference between a key intervention event and the request timestamp The preset time decay constant, As for the intensity of the event's impact, This represents the balance coefficient between time and intensity factors. The intensity of an event's impact can be derived from the intensity level, dosage parameter, and duration parameter in the aforementioned event attributes, or it can be given by a preset event impact coefficient in the rule base. The smaller the time difference and the higher the event's impact intensity, the greater the time-sensitive weight; conversely, the smaller the time difference and the greater the impact intensity. For example, for a glucocorticoid injection event, with an impact intensity set at 0.95, if the event occurred 1 hour ago, the time-sensitive weight is close to 1. For a routine oral vitamin event, although it occurs at the same time, the time-sensitive weight is lower because the impact intensity is only 0.2. The time difference can be calculated in minutes, hours, or days, depending on the type of item to be tested.

[0076] The dynamic correction coefficient, generated by combining feature correlation values ​​and time-sensitive weights, integrates the theoretical correlation between an event and the item to be tested with the event's immediate impact at the current moment, forming a dynamic parameter that reflects the actual corrective effect of the intervention event. This dynamic correction coefficient describes the degree to which a key intervention event weakens the timeliness of existing test results at the current point in time. The dynamic correction coefficient can be expressed as: For example, if the feature correlation value of a key intervention event to the item to be tested is 0.86 and the time-sensitive weight is 0.80, then the dynamic correction coefficient is 0.688.

[0077] For example, the target population of key intervention events in intravenous potassium supplementation therapy is highly correlated with serum potassium test results, resulting in a high characteristic association value between the intravenous potassium supplementation event and the test result. In contrast, routine nursing rounds events have a lower correlation with serum potassium levels, thus generating a relatively lower characteristic association value. This characteristic association method can differentiate the actual impact of different intervention events on the timeliness of different test results.

[0078] Specifically, generating the dynamic correction coefficient also includes: if there are multiple key intervention events, accumulating, weighting, or taking the maximum value of the dynamic correction coefficients for multiple single events to form a comprehensive dynamic correction coefficient. The comprehensive dynamic correction coefficient reflects the combined impact of multiple medical intervention events on the timeliness of test results under superimposed conditions. When a comprehensive dynamic correction coefficient exists, it is incorporated into step 333 to nonlinearly adjust the benchmark attenuation function.

[0079] For example, a patient had a blood glucose test on August 1, 2025, started intensive insulin therapy on August 3, 2025, and adjusted their dietary intervention plan on August 4, 2025. When the current request is initiated on August 5, 2025, a large dynamic correction coefficient is generated because the intensive insulin therapy and blood glucose test are highly correlated and the events occurred relatively close to the current request time. Although the dietary intervention plan adjustment also has some impact, its weight is slightly lower. Ultimately, multiple correction factors work together on the baseline decay function, causing the enhanced decay function to exhibit a faster effectiveness decay trend near the current time point.

[0080] Step 333: The dynamic correction coefficient is embedded as a nonlinear adjustment factor into the baseline decay function to obtain the enhanced decay function. For example, the baseline decay function can be represented as an exponential function that decreases monotonically with time. The dynamic correction coefficient is introduced through an exponential amplification term, a power function adjustment term, or a sigmoid compression term, so that the decay curve shows a steeper decline segment after the occurrence of a key intervention event, or maintains a relatively gentle trend of change in low-impact event scenarios. Through nonlinear adjustment, this application can not only retain the basic characterization ability of the baseline decay function on the passage of time, but also reflect the sudden impact of the diagnosis and treatment intervention event.

[0081] In this embodiment, the baseline decay function represents the reference validity of existing test results that naturally decreases over time, and usually only reflects the time factor; the enhanced decay function introduces the additional failure impact caused by key intervention events on the basis of the baseline decay function, making the timeliness assessment closer to the real clinical scenario.

[0082] Step 334: Input the time difference into the enhancement decay function to obtain the remaining effective component of the existing test results at the current time. Based on the remaining effective component and the pre-set confidence mapping rule, generate the timeliness confidence of the existing test results for the item to be tested. The enhancement decay function can be expressed as follows: ,in Indicates time difference, This is a dynamic correction factor; when no high-impact key intervention events exist, the comprehensive correction factor is applied. Approaching 1, the enhanced decay function shows little difference from the baseline decay function; when there are highly correlated and intense key intervention events, the comprehensive correction factor... The significant increase in efficiency causes the enhancement decay function to output a lower validity value at the same time difference, reflecting the rapid invalidation of the test results in this diagnostic and treatment scenario.

[0083] It should be noted that the remaining effective component represents the degree to which the test result of the stock can still be referenced at the current moment; the larger the remaining effective component, the closer the test result of the stock is to the current true state; the smaller the effective component, the more representative the test result has become over time and due to the impact of intervention events.

[0084] A segmented interval mapping approach can be used: when the remaining valid component is ≥0.80, the timeliness confidence level is high, indicating that previous results still have sufficient reference value; when 0.50 ≤ remaining valid component <0.80, the timeliness confidence level is medium, indicating that previous results still have some reference value; when the remaining valid component <0.50, the timeliness confidence level is low, indicating that previous results are no longer suitable as a substitute for current prescriptions. For example, if the remaining valid component of a certain stock test result is 0.87, then a high timeliness confidence level is obtained, indicating that the necessity of current duplicate prescriptions is low; if the remaining valid component is 0.42, then a low timeliness confidence level is obtained, indicating that the original result is insufficient to support the current diagnosis and treatment decision, and duplicate prescriptions may be reasonable. For example, for a hospitalized patient, it is identified that an electrolyte test was completed 4 hours ago, but a large amount of fluid resuscitation and diuretic treatment occurred during that time. Although the time difference is short, the remaining effective component calculated by the enhanced attenuation function is only 0.38, which corresponds to a low time-dependent confidence level after mapping. Therefore, it can be judged that the previous test results have a weak ability to replace the current test item, and there is a reasonable basis for re-ordering the electrolyte test.

[0085] The corresponding order assessment prompts are output based on the timeliness confidence level. If the timeliness confidence level is high, it indicates that the existing test results are still within the valid reference period. If the timeliness confidence level is medium, it suggests that the order should be carefully judged in conjunction with the current symptoms. If the timeliness confidence level is low, it indicates that the timeliness of the previous test results is insufficient and the current order is highly reasonable.

[0086] This application uses timeliness confidence to assess timeliness, which not only reflects the natural impact of the passage of time on the validity of test results, but also reflects the dynamic disturbance effect of diagnostic and treatment interventions on the timeliness of results, thereby improving the accuracy and clinical suitability of the judgment on the reuse of existing test results.

[0087] Step 140: Obtain the dynamic operating data of the testing equipment associated with the proposed testing item in real time, and generate the resource matching degree.

[0088] Specifically, based on the order request for the proposed testing item, the system identifies the target set of testing equipment capable of performing the proposed testing item by pre-setting a mapping relationship between testing items and equipment types. It then acquires real-time dynamic operating data of the testing equipment, assesses the current equipment's support for the proposed testing item, and generates a resource matching degree. Real-time acquisition of dynamic operating data is achieved through communication between the equipment interface module and the testing equipment information. This dynamic operating data can include information such as the equipment's current operating status, task queue length, load rate, running time, fault status, and remaining available testing capacity. The system also collects real-time operating data from equipment control or laboratory information via equipment communication interfaces, testing equipment management interfaces, or IoT data acquisition interfaces.

[0089] For example, if a biochemical analyzer is currently running with 15 samples in its task queue, a device load rate of 75%, and an estimated time of 20 minutes to complete the current task, and another similar device is idle with 0 tasks in its task queue, then the device can be identified as having high availability.

[0090] Specifically, after acquiring dynamic operational data, the operational status of each testing device is analyzed, and the device resource matching degree is calculated. The resource matching degree quantifies the extent to which current device resources meet the execution requirements of the proposed testing project. The calculation process can comprehensively consider multiple factors such as device operational status, device processing capacity, task queue length, and device load rate. ,in This indicates that the device is operational and available. Indicates the percentage of tasks in the queue. Indicates the equipment load rate. This is the equipment processing capacity coefficient. These are preset weighting coefficients.

[0091] For example, when a testing device is idle, has a small number of tasks in the queue, and has a low load rate, the calculated resource matching degree is high, indicating that the device can quickly respond to the current testing project requirements; when the device is under high load or has a long queue, its resource matching degree is low.

[0092] When there are multiple testing devices capable of performing the proposed testing items, the resource matching degree of each device can be calculated separately, and the optimal device resource can be determined through a sorting or selection mechanism. For example, the device with the highest resource matching degree can be selected as the priority execution device, or the devices can be scheduled and allocated according to the resource matching degree.

[0093] For example, if a laboratory has three devices capable of performing the same test, and the resource matching degree of the three devices is calculated to be 0.82, 0.65 and 0.91 respectively, then the current test task can be preferentially assigned to the device with a resource matching degree of 0.91, thereby improving the overall test processing efficiency.

[0094] Step 150: Map the timeliness confidence, resource matching degree, and current vital sign monitoring information to the preset decision coordinate space, and calculate the reasonable quantitative score for order opening based on the spatial logic mapping algorithm; dynamically adjust the confirmation control status at the business initiation end according to the comparison result between the reasonable quantitative score and the preset risk threshold.

[0095] Mapping the timeliness confidence level, resource matching degree, and current vital sign monitoring information to a preset decision coordinate space can be represented as: ,in Indicates the confidence level of timeliness. Indicates resource matching degree. This represents an indicator of changes in vital signs. For example, if a subject's existing test results have a timeliness confidence level of 0.85, an equipment resource matching degree of 0.70, and a vital sign abnormality degree of 0.40, then the corresponding decision space feature vector is... .

[0096] In this embodiment, a spatial logic mapping algorithm is used to calculate the reasonableness score of an order. The spatial logic mapping algorithm performs a weighted calculation based on the degree of influence of different dimensions on the reasonableness of the order, and the reasonableness score can be expressed as follows: , , , These are the preset weight parameters. This indicates the degree of invalidity of previous test results. The closer the previous results are to invalidity, the more reasonable the new test order is.

[0097] Specifically, the reasonable quantitative score for order placement is compared with a preset risk threshold to determine whether there are potential unreasonable risks in the current order placement behavior. The status of the confirmation control at the business initiation end is dynamically adjusted based on the comparison results. The business initiation end can be the doctor's order placement terminal, clinical workstation, or order placement interface in medical information. The confirmation control can include an order placement confirmation button, risk warning control, or secondary confirmation control, etc.

[0098] In this embodiment of the application, multiple risk threshold ranges are preset to correspond to different order risk levels. For example, the first threshold range is a reasonable range, the second threshold range is a warning range, and the third threshold range is a restriction range. When the reasonable quantitative score of the order is in different ranges, different control state adjustment strategies will be triggered.

[0099] For example, when the reasonable quantitative score S ≥ T1 for order placement, the order placement is considered reasonable, and the confirmation control on the business initiator remains clickable. When T2 ≤ S < T1, the order placement is considered to have some risk. In this case, a prompt message pops up on the business interface, and the confirmation control is changed to a prompt confirmation mode, requiring the user to confirm again. When S < T2, the order placement is considered to have a high risk, and the confirmation control is changed to a restricted mode, such as requiring additional reasons or approval from a senior physician before submission. For example, when the calculated reasonable quantitative score for order placement is 0.25, and the preset risk threshold T2 is 0.40, the current test order is considered to have a high risk of duplicate order placement. In this case, a prompt message automatically pops up on the doctor's order placement interface, such as the test item has recently been completed and the current result is still valid, and the confirmation button is changed to require a reason before submission.

[0100] In this embodiment, the risk threshold is dynamically optimized based on historical operational data. For example, the relationship between historical prescription behavior and clinical outcomes is analyzed through machine learning models to continuously optimize the mapping relationship between reasonable quantitative scores and risk thresholds, thereby improving the accuracy of judgment.

[0101] This application uses real-time intelligent assessment of the rationality of test order placement and dynamically adjusts the status of the control at the business initiation end to guide doctors to place orders reasonably and reduce duplicate tests.

[0102] Please refer to Figure 5 The diagram illustrates the structure of an intelligent evaluation system for the rationality of test item prescriptions, as provided in this application embodiment. It includes: a data parsing module 510, a treatment timeline trajectory generation module 520, a timeliness confidence acquisition module 530, a resource matching degree generation module 540, and a prescription rationality evaluation module 550, wherein:

[0103] The data parsing module 510 is used to capture the request message of the ordering instruction based on the communication request for ordering the inspection item, and obtain the inspectee's identifier, the item to be inspected, and the request timestamp based on the structured parsing of the request message.

[0104] The diagnosis and treatment timeline generation module 520 is used to obtain diagnosis and treatment data within a preset historical period based on the examinee's identifier, and generate a diagnosis and treatment timeline trajectory with the requested timestamp as the endpoint based on the current vital sign monitoring information.

[0105] The timeliness confidence acquisition module 530 is used to obtain historical observation characteristics by analyzing the diagnosis and treatment timeline trajectory, determine the benchmark decay function based on the stability of the historical observation characteristics, generate dynamic correction coefficients according to key intervention events, perform nonlinear adjustment on the benchmark decay function through the dynamic correction coefficients, and obtain the timeliness confidence of the existing test results for the test item.

[0106] The resource matching degree module generates 540, which is used to obtain the dynamic operation data of the testing equipment associated with the proposed testing project in real time and generate the resource matching degree.

[0107] The order rationality assessment module 550 is used to map timeliness confidence, resource matching degree and current vital sign monitoring information to a preset decision coordinate space, and calculate the order rationality quantitative score based on the spatial logic mapping algorithm; and dynamically adjust the confirmation control status at the business initiation end according to the comparison result of the rational quantitative score and the preset risk threshold.

[0108] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A method for intelligently evaluating the rationality of issuing inspection invoices, characterized in that, include: Based on the communication request for issuing an inspection order, capture the request message of the order issuance instruction, and obtain the inspectee's identifier, the item to be inspected, and the request timestamp based on the structured parsing of the request message; Based on the examinee's identification, obtain the diagnosis and treatment data within a preset historical period, and based on the current vital sign monitoring information, generate a diagnosis and treatment timeline trajectory with the requested timestamp as the endpoint; By analyzing the timeline of diagnosis and treatment to obtain historical observation characteristics, a baseline decay function is determined based on the stability of the historical observation characteristics. Dynamic correction coefficients are generated based on key intervention events. The baseline decay function is nonlinearly adjusted using the dynamic correction coefficients to obtain the timeliness confidence of the stock test results for the inverse test items. Real-time acquisition of dynamic operating data of testing equipment associated with the proposed testing items, and generation of resource matching degree; The system maps timeliness confidence, resource matching degree, and current vital sign monitoring information to a preset decision coordinate space. Based on the spatial logic mapping algorithm, it calculates a reasonable quantitative score for order placement. Based on the comparison between the reasonable quantitative score and the preset risk threshold, it dynamically adjusts the confirmation control status at the business initiation end.

2. The intelligent evaluation method for the rationality of inspection item invoicing as described in claim 1, characterized in that, Based on the communication request for issuing an inspection order, the system captures the order issuance instruction request message. Based on the structured parsing of the request message, it obtains the examinee's identifier, the items to be inspected, and the request timestamp, including: Real-time detection of asynchronous communication links between the service initiator and the backend server; Based on a preset traffic filtering algorithm, non-intrusive data frame capture is performed at the network layer, and the request message of the order opening instruction is identified by feature point matching. The request message is parsed through a multi-level protocol engine to obtain the subject identifier, the item to be inspected, and the request timestamp.

3. The intelligent evaluation method for the rationality of inspection item invoicing as described in claim 1, characterized in that, Based on the communication request for issuing an inspection project order, the request message of the order issuance instruction is captured, including: obtaining a copy of the original request message based on the traffic mirror node, using the extended Berkeley packet filter program mounted on the trace point of the kernel protocol stack to intercept the request message in real time; the intercepted instruction characteristics are asynchronously reported to the application layer via the kernel buffer to realize the capture of the request message.

4. The intelligent evaluation method for the rationality of inspection item invoicing according to claim 1, characterized in that, The confidence level for obtaining timeliness includes: Analyze the sequence of test values, the slope of physiological fluctuations, and key intervention events in the timeline of diagnosis and treatment; Calculate the discrete characteristics between adjacent test values ​​of the target observation item. When the discrete characteristics and the slope of physiological fluctuations meet the preset steady-state threshold, map the physiological half-life of the corresponding physiological index to the benchmark decay function. By using feature association, key intervention events are transformed into dynamic correction coefficients. The benchmark attenuation function is nonlinearly adjusted using the dynamic correction coefficients to obtain the enhanced attenuation function. The time difference is substituted into the enhanced attenuation function to calculate the remaining effective component of the stock inspection results relative to the current request timestamp, and the timeliness confidence of the stock inspection results for the project to be inspected is obtained.

5. The intelligent evaluation method for the rationality of inspection item invoicing according to claim 4, characterized in that, Key intervention events are transformed into dynamic correction coefficients through feature association, including: Feature analysis is performed on key intervention events to obtain the event attributes of each key intervention event. Feature association analysis is then performed based on the relationship between the event attributes and the project to be tested to determine the feature association value of each key intervention event on the timeliness of the results of the project to be tested. The time difference is obtained by subtracting the event occurrence time from the current request timestamp. The time-sensitive weight is determined based on the time difference and the intensity of the event's impact. A dynamic correction coefficient is generated by combining the feature association value and the time-sensitive weight.

6. The intelligent evaluation method for the rationality of inspection item invoicing according to claim 4, characterized in that, An enhanced attenuation function is obtained by nonlinearly adjusting the benchmark attenuation function using a dynamic correction coefficient, including embedding the dynamic correction coefficient as a nonlinear adjustment factor into the benchmark attenuation function to obtain the enhanced attenuation function.

7. The intelligent evaluation method for the rationality of inspection item invoicing according to claim 4, characterized in that, Obtain the timeliness confidence level of the existing inspection results for the item to be inspected, including: Based on the remaining valid components and the pre-set confidence mapping rules, the mapping interval of the remaining valid components is obtained through segmented mapping, and the timeliness confidence of the stock test results on the item to be tested is generated based on the mapping interval.

8. The intelligent evaluation method for the rationality of inspection item invoicing according to claim 1, characterized in that, Real-time acquisition of dynamic operating data of testing equipment associated with the proposed testing item, generating resource matching degree, including: Based on the order request for the proposed inspection item, the system identifies the set of target inspection equipment that can perform the proposed inspection item by pre-setting the mapping relationship between inspection items and equipment types, obtains the dynamic operation data of the inspection equipment in real time, evaluates the current support of the inspection equipment for the proposed inspection item based on the dynamic operation data, and generates a resource matching degree.

9. An intelligent evaluation system for the rationality of inspection item invoicing, used to implement the method of claims 1-8, characterized in that, include: The data parsing module is used to capture the request message of the ordering instruction based on the communication request for the ordering of the test items, and obtain the testee's identifier, the test items to be tested, and the request timestamp based on the structured parsing of the request message; The diagnosis and treatment timeline generation module is used to obtain diagnosis and treatment data within a preset historical period based on the examinee's identifier, and generate a diagnosis and treatment timeline trajectory with the requested timestamp as the endpoint based on the current vital sign monitoring information. The timeliness confidence acquisition module is used to obtain historical observation characteristics by analyzing the diagnosis and treatment timeline trajectory, determine the benchmark decay function based on the stability of the historical observation characteristics, generate dynamic correction coefficients according to key intervention events, perform nonlinear adjustment on the benchmark decay function through the dynamic correction coefficients, and obtain the timeliness confidence of the existing test results for the test item. The resource matching degree module is used to obtain the dynamic operating data of the testing equipment associated with the proposed testing project in real time and generate the resource matching degree. The order rationality assessment module maps timeliness confidence, resource matching degree, and current vital sign monitoring information to a preset decision coordinate space, calculates the order rationality quantitative score based on the spatial logic mapping algorithm, and dynamically adjusts the confirmation control status at the business initiation end based on the comparison result of the rational quantitative score and the preset risk threshold.

10. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

Images