Electronic medical record-based healthcare service management system and method
Patent Information
- Application Number
- CN202611279951.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-21
- Publication Date
- 2026-10-13
AI Technical Summary
[0003]但是现有的医疗保健服务管理系统其用药安全规则的权重参数和预警触发阈值一经配置即保持固定，既无法根据各临床科室和不同年资医师的实际使用反馈进行动态校准，也缺乏将医师对预警的拒绝行为作为反馈信号反向优化知识库的闭环机制，导致大量低临床相关性的警报频繁弹出，医师被迫反复点击确认或覆盖，形成严重的警报疲劳，反而削弱了对真正高风险事件的关注，且系统对同一药物禁忌组合在所有临床场景下采用统一的判断标准，既无法区分不同科室、不同适应症下的风险差异，也无法根据医师的历史响应行为进行个性化适配，使得系统的预警行为与临床实际风险分布之间的偏差持续存在且无法自我修正
[0070]与现有技术相比，本发明的有益效果是：通过从电子病历系统中获取目标患者的临床数据，根据医师开具目标药物的名称，从预建的知识库中检索与该目标药物关联的用药安全规则集合，将目标患者的临床数据与用药安全规则集合中的各条规则逐一进行条件匹配，获取被触发的规则列表及其对应的基础风险权重和当前置信度衰减因子，根据各条被触发规则的基础风险权重与当前置信度衰减因子的乘积计算每一条被触发规则的动态风险贡献值，累加所有被触发规则的动态风险贡献值得到综合风险评分，并根据综合风险评分落入的预设区间确定对应的风险等级，获取医师的历史预警响应记录，根据历史预警响应记录计算该医师对各类预警的响应敏感度系数，基于风险等级，利用响应敏感度系数对风险等级进行偏移校正，得到医师适配风险等级，并执行与医师适配风险等级匹配的分级预警操作，记录医师对分级预警操作的响应行为，当医师的响应行为为拒绝预警并提交原医嘱时，将该拒绝行为作为负反馈信号，根据负反馈信号降低用药安全规则集合中被触发规则对应的当前置信度衰减因子，更新后的当前置信度衰减因子存入知识库中以供后续检索使用；通过构建包含基础风险权重和当前置信度衰减因子的动态用药安全规则结构，结合医师历史预警响应行为对预警阈值进行适配校正，并将医师的预警拒绝行为作为负反馈信号反向传播至知识库以更新规则置信度因子，从而形成预警、响应、反馈、优化的闭环学习机制，有效解决现有系统规则权重静态化、预警阈值同质化、缺乏闭环优化能力的技术缺陷。
Smart Images
Figure CN122907385A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of smart healthcare technology, specifically to a healthcare service management system and method based on electronic medical records. Background Technology
[0002] In the existing technology, the healthcare service management system refers to a digital platform system that uses information technology to comprehensively manage the entire process of healthcare services for the daily operation and clinical business of medical institutions. Its goal is to upgrade the traditional manual and paper-based healthcare management process into an information-based and intelligent closed-loop management model.
[0003] However, existing healthcare service management systems maintain fixed weight parameters and warning trigger thresholds for medication safety rules once configured. This prevents dynamic calibration based on actual usage feedback from various clinical departments and physicians of different seniority. Furthermore, they lack a closed-loop mechanism that uses physicians' rejection of warnings as feedback signals to optimize the knowledge base. This results in a large number of low-clinical-relevance alerts popping up frequently, forcing physicians to repeatedly click to confirm or override them, leading to severe alert fatigue. This weakens the focus on truly high-risk events. Moreover, the system uses a uniform judgment standard for the same drug contraindications across all clinical scenarios, failing to differentiate risk differences across departments and indications, and unable to personalize adjustments based on physicians' historical responses. This results in a persistent and uncorrectable discrepancy between the system's warning behavior and the actual clinical risk distribution.
[0004] Therefore, in view of this, the present invention proposes a healthcare service management system and method based on electronic medical records to make up for and improve the deficiencies of the prior art. Summary of the Invention
[0005] To address the aforementioned technical problems, this invention provides a healthcare service management system and method based on electronic medical records, thereby resolving the corresponding technical issues raised in the background section.
[0006] To achieve the above objectives, the technical solution adopted by this invention is: a healthcare service management system based on electronic medical records, comprising a data acquisition unit, a rule retrieval unit, a conflict detection unit, an adaptation and early warning unit, and a feedback learning unit, wherein:
[0007] The data acquisition unit is used to acquire clinical data of the target patient from the electronic medical record system.
[0008] The rule retrieval unit is used to retrieve a set of medication safety rules associated with the target drug from a pre-built knowledge base based on the name of the target drug prescribed by the physician.
[0009] The conflict detection unit is used to match the clinical data of the target patient with each rule in the medication safety rule set one by one, obtain the list of triggered rules and their corresponding basic risk weights and current confidence decay factors, calculate the dynamic risk contribution value of each triggered rule based on the product of the basic risk weight and the current confidence decay factor of each triggered rule, accumulate the dynamic risk contribution values of all triggered rules to obtain a comprehensive risk score, and determine the corresponding risk level based on the preset interval in which the comprehensive risk score falls.
[0010] The adaptive warning unit is used to obtain the physician's historical warning response records, calculate the physician's response sensitivity coefficient to various warnings based on the historical warning response records, and use the response sensitivity coefficient to correct the risk level based on the risk level to obtain the physician's adapted risk level, and execute the graded warning operation that matches the physician's adapted risk level.
[0011] The feedback learning unit is used to record the physician's response to the graded warning operation. When the physician's response is to refuse the warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval.
[0012] Preferably, in the data acquisition unit:
[0013] The clinical data includes the target patient's current medication list, allergy history record, and current physiological status parameters;
[0014] The current medication list records the names, dosages, usage methods, and prescription dates of all medications the patient is currently using;
[0015] The allergy history record includes the name of the allergen and its corresponding severity level;
[0016] The current physiological parameters include age, sex, pregnancy status, and liver and kidney function indicators;
[0017] The liver and kidney function indicators include serum alanine aminotransferase, serum creatinine, and estimated glomerular filtration rate.
[0018] Preferably, in the rule retrieval unit:
[0019] The knowledge base pre-stores medication safety rules covering various types of drugs. Each rule includes a rule identifier, associated contraindications, basic risk weight, and current confidence decay factor.
[0020] The initial value of the current confidence decay factor is 1, which is used to reflect the credibility of the rule in current clinical practice. The lower the value, the worse the applicability of the rule in actual clinical practice.
[0021] The set of medication safety rules specifically includes three categories: drug interaction rules, drug and allergen contraindication rules, and drug and physiological state contraindication rules.
[0022] Each medication safety rule in the knowledge base is also pre-configured with applicable department tags and applicable age group tags. During retrieval, the search results are also filtered according to the clinical department to which the physician belongs and the age of the target patient, returning only medication safety rules whose applicable department tags match the clinical department to which the physician belongs and whose applicable age group tags match the age of the target patient.
[0023] Preferably, in the conflict detection unit, the specific process of matching the target patient's clinical data with each rule in the medication safety rule set is as follows:
[0024] Each drug in the current medication list is matched with the target drug according to drug interaction rules to detect the risk of interaction between the target drug and other drugs being used by the target patient.
[0025] Each allergen in the allergy history record is matched with the target drug according to the drug-allergen incompatibility rule, and the known allergen components of the target patient contained in the target drug are detected.
[0026] The parameters of the current physiological state are matched with the conditions for use of the target drug according to the drug-physiological state contraindication rules, and the indicators of the current physiological state of the target patient that are contraindicated for the use of the target drug are detected.
[0027] Preferably, in the conflict detection unit, the specific process for determining the corresponding risk level based on the preset range into which the comprehensive risk score falls is as follows:
[0028] Obtain the list of triggered rules and their corresponding base risk weights and current confidence decay factors. Calculate the dynamic risk contribution value of each triggered rule based on the product of its base risk weight and current confidence decay factor. The formula is as follows:
[0029]
[0030] in, The dynamic risk contribution value for the i-th triggered rule;
[0031] The basic risk weight for the i-th rule;
[0032] Let be the current confidence decay factor for the i-th rule, and its value range is . ;
[0033] The overall risk score is obtained by summing the dynamic risk contribution values of all triggered rules. Its formula is:
[0034]
[0035] Where n is the total number of medication safety rules that are triggered;
[0036] The risk level is determined based on the preset range into which the comprehensive risk score falls, as follows:
[0037] when If the value is greater than or equal to the first preset threshold, it is determined to be in the first level;
[0038] when If the value is greater than or equal to the second preset threshold and less than the first preset threshold, it is determined to be the second level;
[0039] when If the value is less than the second preset threshold, it is judged as the third level.
[0040] Preferably, in the adaptive early warning unit:
[0041] Extract all warning response records of the physician within a preset historical time period from the audit log of the electronic medical record system, count the number of times the physician accepted and rejected each type of warning, and calculate the response sensitivity coefficient. Its formula is:
[0042]
[0043] in, This represents the number of times the physician rejected a certain type of warning;
[0044] This represents the total number of times the physician responded to this type of warning;
[0045] The closer the value is to 1, the lower the physician's acceptance of this type of warning, and the more likely they are to believe that this type of warning is not clinically relevant.
[0046] The closer the value is to 0, the higher the doctor's acceptance of this type of warning;
[0047] The risk level output by the conflict detection unit is obtained, and the risk level is offset by the response sensitivity coefficient to obtain the physician-appropriate risk level as follows:
[0048] When the risk level is Level 1, it remains Level 1.
[0049] When the risk level is level two and If the risk level exceeds a preset threshold, the risk level will be downgraded to level three.
[0050] When the risk level is level three and If the risk level falls below a preset threshold, the risk level will be upgraded to the second level.
[0051] At the same time, the following graded early warning procedures will be implemented, matching the physician's risk level:
[0052] If the physician's risk level is Level 1, it means that the target drug is absolutely contraindicated for the target patient, generating a blocking warning and preventing the submission of the target drug prescription;
[0053] If the physician's risk level is Level 2, it means that the target drug poses a significant clinical risk to the target patient. An alert is generated and the physician is shown alternative medication options.
[0054] If the physician's risk level is level three, it means that there are general precautions for the target drug for the target patient. A reminder message will be generated and displayed on the target drug's prescription submission interface.
[0055] Preferably, in the feedback learning unit:
[0056] Record the physician's response behavior to the graded early warning operation, and the response behavior is as follows:
[0057] When a physician chooses to accept the recommendation for a graded alert, the acceptance is recorded and the physician is allowed to modify the prescription for the target drug accordingly before submission.
[0058] When a physician chooses to refuse the recommendation for a graded warning procedure, the refusal should be recorded and the physician should be required to include the reason for refusal when submitting the original prescription for the target drug.
[0059] When a physician's response is to refuse the warning and submit the original medical order, this refusal is treated as a negative feedback signal. Based on this negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced, using the following formula:
[0060]
[0061] in, The current confidence decay factor before the triggered rule update;
[0062] The preset attenuation coefficient;
[0063] This is the updated current confidence decay factor.
[0064] The approach to managing healthcare services based on electronic medical records includes the following steps:
[0065] Step 1: Obtain the target patient's clinical data from the electronic medical record system;
[0066] Step 2: Based on the name of the target drug prescribed by the physician, retrieve the set of medication safety rules associated with the target drug from the pre-built knowledge base;
[0067] Step 3: Match the clinical data of the target patient with each rule in the medication safety rule set to obtain the list of triggered rules and their corresponding basic risk weights and current confidence decay factors. Calculate the dynamic risk contribution value of each triggered rule based on the product of its basic risk weight and current confidence decay factor. Accumulate the dynamic risk contribution values of all triggered rules to obtain a comprehensive risk score. Determine the corresponding risk level based on the preset interval into which the comprehensive risk score falls.
[0068] Step 4: Obtain the physician's historical early warning response records, calculate the physician's response sensitivity coefficient for various early warnings based on the historical early warning response records, and use the response sensitivity coefficient to correct the risk level based on the risk level to obtain the physician's appropriate risk level, and execute the graded early warning operation that matches the physician's appropriate risk level.
[0069] Step 5: Record the physician's response to the graded warning operation. When the physician's response is to refuse the warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval.
[0070] Compared with existing technologies, the beneficial effects of this invention are as follows: Clinical data of the target patient is obtained from an electronic medical record system; based on the name of the target drug prescribed by the physician, a set of medication safety rules associated with the target drug is retrieved from a pre-built knowledge base; the clinical data of the target patient is matched with each rule in the medication safety rule set one by one to obtain a list of triggered rules and their corresponding basic risk weights and current confidence decay factors; the dynamic risk contribution value of each triggered rule is calculated based on the product of its basic risk weight and current confidence decay factor; the dynamic risk contribution values of all triggered rules are summed to obtain a comprehensive risk score; and the corresponding risk level is determined based on the preset interval into which the comprehensive risk score falls; the physician's historical warning response records are obtained; the physician's response sensitivity coefficient to various warnings is calculated based on the historical warning response records; and the risk level is determined using the response sensitivity coefficient based on the risk level. Offset correction is performed to obtain the physician's risk level, and a graded early warning operation matching the physician's risk level is executed. The physician's response to the graded early warning operation is recorded. When the physician's response is to refuse the early warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval. By constructing a dynamic medication safety rule structure that includes basic risk weights and current confidence decay factors, and combining the physician's historical early warning response behavior to adjust the early warning threshold, the physician's early warning refusal behavior is used as a negative feedback signal to propagate back to the knowledge base to update the rule confidence factor. This forms a closed-loop learning mechanism of early warning, response, feedback, and optimization, effectively solving the technical defects of existing systems such as static rule weights, homogeneous early warning thresholds, and lack of closed-loop optimization capabilities. Attached Figure Description
[0071] Figure 1 This is a schematic diagram of the overall structure of a preferred embodiment of the present invention;
[0072] Figure 2 This is a flowchart of the healthcare service management method of the present invention. Detailed Implementation
[0073] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0074] Embodiments of the present invention:
[0075] Please refer to Figures 1 to 2 As shown, the healthcare service management system based on electronic medical records includes a data acquisition unit, a rule retrieval unit, a conflict detection unit, an adaptation and early warning unit, and a feedback learning unit, wherein:
[0076] The data acquisition unit is used to acquire clinical data of the target patient from the electronic medical record system.
[0077] The rule retrieval unit is used to retrieve a set of medication safety rules associated with the target drug from a pre-built knowledge base based on the name of the target drug prescribed by the physician.
[0078] The conflict detection unit is used to match the clinical data of the target patient with each rule in the medication safety rule set one by one, obtain the list of triggered rules and their corresponding basic risk weights and current confidence decay factors, calculate the dynamic risk contribution value of each triggered rule based on the product of the basic risk weight and the current confidence decay factor of each triggered rule, accumulate the dynamic risk contribution values of all triggered rules to obtain a comprehensive risk score, and determine the corresponding risk level based on the preset interval in which the comprehensive risk score falls.
[0079] The adaptive warning unit is used to obtain the physician's historical warning response records, calculate the physician's response sensitivity coefficient to various warnings based on the historical warning response records, and use the response sensitivity coefficient to correct the risk level based on the risk level to obtain the physician's adapted risk level, and execute the graded warning operation that matches the physician's adapted risk level.
[0080] The feedback learning unit is used to record the physician's response to the graded warning operation. When the physician's response is to refuse the warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval.
[0081] In the data acquisition unit:
[0082] Clinical data includes the target patient's current medication list, allergy history, and current physiological parameters;
[0083] The current medication list records the names, dosages, methods of administration, and dates of administration of all medications the patient is currently using.
[0084] Allergy history records include the names of allergens and their corresponding severity levels;
[0085] Current physiological parameters include age, sex, pregnancy status, and liver and kidney function indicators;
[0086] Among them, liver and kidney function indicators include serum alanine aminotransferase, serum creatinine, and estimated glomerular filtration rate.
[0087] In the rule retrieval unit:
[0088] The knowledge base pre-stores medication safety rules covering various types of drugs. Each rule includes a rule identifier, associated contraindications, basic risk weights, and current confidence decay factors.
[0089] The current confidence decay factor has an initial value of 1, which reflects the credibility of the rule in current clinical practice. The lower the value, the worse the applicability of the rule in actual clinical practice.
[0090] The set of medication safety rules specifically includes three categories: rules on drug interactions, rules on drug-allergen contraindications, and rules on drug-physiological contraindications.
[0091] Each medication safety rule in the knowledge base is also pre-configured with applicable department tags and applicable age group tags. During the search, the search results are also filtered according to the clinical department to which the physician belongs and the age of the target patient, returning only medication safety rules whose applicable department tags match the clinical department to which the physician belongs and whose applicable age group tags match the age of the target patient.
[0092] In the conflict detection unit, the specific process of matching the target patient's clinical data with each rule in the medication safety rule set is as follows:
[0093] Each drug in the current medication list is matched with the target drug according to drug interaction rules to detect the risk of interaction between the target drug and other drugs being used by the target patient.
[0094] Each allergen in the allergy history record is matched with the target drug according to the drug-allergen incompatibility rule, and the known allergen components of the target patient contained in the target drug are detected.
[0095] The parameters of the current physiological state are matched with the conditions for use of the target drug according to the drug-physiological state contraindication rules, and the indicators of the current physiological state of the target patient that are contraindicated for the use of the target drug are detected.
[0096] In the conflict detection unit, the specific process for determining the corresponding risk level based on the preset range into which the comprehensive risk score falls is as follows:
[0097] Obtain the list of triggered rules and their corresponding base risk weights and current confidence decay factors. Calculate the dynamic risk contribution value of each triggered rule based on the product of its base risk weight and current confidence decay factor. The formula is as follows:
[0098]
[0099] in, The dynamic risk contribution value for the i-th triggered rule;
[0100] The basic risk weight for the i-th rule;
[0101] Let be the current confidence decay factor for the i-th rule, and its value range is . ;
[0102] The overall risk score is obtained by summing the dynamic risk contribution values of all triggered rules. Its formula is:
[0103]
[0104] Where n is the total number of medication safety rules that are triggered;
[0105] The risk level is determined based on the preset range into which the comprehensive risk score falls, as follows:
[0106] when If the value is greater than or equal to the first preset threshold, it is determined to be in the first level;
[0107] when If the value is greater than or equal to the second preset threshold and less than the first preset threshold, it is determined to be the second level;
[0108] when If the value is less than the second preset threshold, it is judged as the third level.
[0109] In the adaptive early warning unit:
[0110] Extract all warning response records of the physician within a preset historical time period from the audit log of the electronic medical record system, count the number of times the physician accepted and rejected each type of warning, and calculate the response sensitivity coefficient. Its formula is:
[0111]
[0112] in, This represents the number of times the physician rejected a certain type of warning;
[0113] This represents the total number of times the physician responded to this type of warning;
[0114] The closer the value is to 1, the lower the physician's acceptance of this type of warning, and the more likely they are to believe that this type of warning is not clinically relevant.
[0115] The closer the value is to 0, the higher the doctor's acceptance of this type of warning;
[0116] The risk level output by the conflict detection unit is obtained, and the risk level is offset by the response sensitivity coefficient to obtain the physician-appropriate risk level as follows:
[0117] When the risk level is Level 1, it remains Level 1.
[0118] When the risk level is level two and If the risk level exceeds a preset threshold, the risk level will be downgraded to level three.
[0119] When the risk level is level three and If the risk level falls below a preset threshold, the risk level will be upgraded to the second level.
[0120] At the same time, the following graded early warning procedures will be implemented, matching the physician's risk level:
[0121] If the physician's risk level is Level 1, it means that the target drug is absolutely contraindicated for the target patient, generating a blocking warning and preventing the submission of the target drug prescription;
[0122] If the physician's risk level is Level 2, it means that the target drug poses a significant clinical risk to the target patient. An alert is generated and the physician is shown alternative medication options.
[0123] If the physician's risk level is level three, it means that there are general precautions for the target drug for the target patient. A reminder message will be generated and displayed on the target drug's prescription submission interface.
[0124] In the feedback learning unit:
[0125] Record the physicians' response behavior to the graded early warning procedure. The response behavior is as follows:
[0126] When a physician chooses to accept the recommendation for a graded alert, the acceptance is recorded and the physician is allowed to modify the prescription for the target drug accordingly before submission.
[0127] When a physician chooses to refuse the recommendation for a graded warning procedure, the refusal should be recorded and the physician should be required to include the reason for refusal when submitting the original prescription for the target drug.
[0128] When a physician's response is to refuse the warning and submit the original medical order, this refusal is treated as a negative feedback signal. Based on this negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced, using the following formula:
[0129]
[0130] in, The current confidence decay factor before the triggered rule update;
[0131] The preset attenuation coefficient;
[0132] This is the updated current confidence decay factor;
[0133] When the same rule is rejected multiple times, its confidence decay factor continues to decay cumulatively, but it will not be lower than the preset lower limit value to prevent the rule weight from being completely decayed to zero.
[0134] The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval. When the same rule is not rejected within a period of time, the feedback learning unit also gradually increases the confidence decay factor of the rule according to the preset recovery coefficient until it is restored to the initial value of 1.
[0135] It also includes a dynamic optimization unit, which is used to count the rejection rate of various warning operations in the same clinical department within a preset historical time period. When the rejection rate of a certain type of warning operation exceeds a preset threshold, the warning level corresponding to that type of warning operation will be automatically downgraded by one level.
[0136] By setting a basic risk weight and a current confidence decay factor for each medication safety rule in the knowledge base, and propagating physicians' warning rejection behavior as a negative feedback signal back to the knowledge base, the dynamic updating of the rule confidence factor is achieved. This solves the technical defect that the rule weights in the existing system are fixed once configured, enabling the system to continuously optimize the knowledge base parameters using real clinical feedback data and reduce the deviation between the rule weights and the actual clinical risk distribution.
[0137] By acquiring physicians' historical warning response records and calculating their response sensitivity coefficient, and using this coefficient to correct the risk level offset, a physician-adapted risk level that matches the physician's warning preferences and clinical judgment habits is obtained. This solves the problem of the one-size-fits-all approach caused by the existing system using a uniform warning threshold for all physicians, effectively reduces the interference of invalid alarms for senior or specific specialist physicians, and avoids excessive alarm overload for junior physicians.
[0138] The approach to managing healthcare services based on electronic medical records includes the following steps:
[0139] Step 1: Obtain the target patient's clinical data from the electronic medical record system;
[0140] Step 2: Based on the name of the target drug prescribed by the physician, retrieve the set of medication safety rules associated with the target drug from the pre-built knowledge base;
[0141] Step 3: Match the clinical data of the target patient with each rule in the medication safety rule set to obtain the list of triggered rules and their corresponding basic risk weights and current confidence decay factors. Calculate the dynamic risk contribution value of each triggered rule based on the product of its basic risk weight and current confidence decay factor. Accumulate the dynamic risk contribution values of all triggered rules to obtain a comprehensive risk score. Determine the corresponding risk level based on the preset interval into which the comprehensive risk score falls.
[0142] Step 4: Obtain the physician's historical early warning response records, calculate the physician's response sensitivity coefficient for various early warnings based on the historical early warning response records, and use the response sensitivity coefficient to correct the risk level based on the risk level to obtain the physician's appropriate risk level, and execute the graded early warning operation that matches the physician's appropriate risk level.
[0143] Step 5: Record the physician's response to the graded warning operation. When the physician's response is to refuse the warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval.
[0144] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.
[0145] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.
[0146] In the two embodiments provided in this application, it should be understood that the disclosed apparatus and system can be implemented in other ways; for example, the apparatus embodiments described above are merely illustrative, and the division of modules is merely a logical functional division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed; furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of the apparatus or module may be electrical, mechanical or other forms.
[0147] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A healthcare service management system based on electronic medical records, characterized in that, It includes a data acquisition unit, a rule retrieval unit, a conflict detection unit, an adaptation warning unit, and a feedback learning unit, among which: The data acquisition unit is used to acquire clinical data of the target patient from the electronic medical record system. The rule retrieval unit is used to retrieve a set of medication safety rules associated with the target drug from a pre-built knowledge base based on the name of the target drug prescribed by the physician. The conflict detection unit is used to match the clinical data of the target patient with each rule in the medication safety rule set one by one, obtain the list of triggered rules and their corresponding basic risk weights and current confidence decay factors, calculate the dynamic risk contribution value of each triggered rule based on the product of the basic risk weight and the current confidence decay factor of each triggered rule, accumulate the dynamic risk contribution values of all triggered rules to obtain a comprehensive risk score, and determine the corresponding risk level based on the preset interval in which the comprehensive risk score falls. The adaptive warning unit is used to obtain the physician's historical warning response records, calculate the physician's response sensitivity coefficient to various warnings based on the historical warning response records, and use the response sensitivity coefficient to correct the risk level based on the risk level to obtain the physician's adapted risk level, and execute the graded warning operation that matches the physician's adapted risk level. The feedback learning unit is used to record the physician's response to the graded warning operation. When the physician's response is to refuse the warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval.
2. The healthcare service management system and method based on electronic medical records according to claim 1, characterized in that, In the data acquisition unit: The clinical data includes the target patient's current medication list, allergy history record, and current physiological status parameters; The current medication list records the names, dosages, usage methods, and prescription dates of all medications the patient is currently using; The allergy history record includes the name of the allergen and its corresponding severity level; The current physiological parameters include age, sex, pregnancy status, and liver and kidney function indicators; The liver and kidney function indicators include serum alanine aminotransferase, serum creatinine, and estimated glomerular filtration rate.
3. The healthcare service management system and method based on electronic medical records according to claim 1, characterized in that, In the rule retrieval unit: The knowledge base pre-stores medication safety rules covering various types of drugs. Each rule includes a rule identifier, associated contraindications, basic risk weight, and current confidence decay factor. The initial value of the current confidence decay factor is 1, which is used to reflect the credibility of the rule in current clinical practice. The lower the value, the worse the applicability of the rule in actual clinical practice. The set of medication safety rules specifically includes three categories: drug interaction rules, drug and allergen contraindication rules, and drug and physiological state contraindication rules. Each medication safety rule in the knowledge base is also pre-configured with applicable department tags and applicable age group tags. During retrieval, the search results are also filtered according to the clinical department to which the physician belongs and the age of the target patient, returning only medication safety rules whose applicable department tags match the clinical department to which the physician belongs and whose applicable age group tags match the age of the target patient.
4. The healthcare service management system and method based on electronic medical records according to claim 1, characterized in that, In the conflict detection unit, the specific process of matching the target patient's clinical data with each rule in the medication safety rule set is as follows: Each drug in the current medication list is matched with the target drug according to drug interaction rules to detect the risk of interaction between the target drug and other drugs being used by the target patient. Each allergen in the allergy history record is matched with the target drug according to the drug-allergen incompatibility rule, and the known allergen components of the target patient contained in the target drug are detected. The parameters of the current physiological state are matched with the conditions for use of the target drug according to the drug-physiological state contraindication rules, and the indicators of the current physiological state of the target patient that are contraindicated for the use of the target drug are detected.
5. The healthcare service management system and method based on electronic medical records according to claim 1, characterized in that, In the conflict detection unit, the specific process of determining the corresponding risk level based on the preset range into which the comprehensive risk score falls is as follows: Obtain the list of triggered rules and their corresponding base risk weights and current confidence decay factors. Calculate the dynamic risk contribution value of each triggered rule based on the product of its base risk weight and current confidence decay factor. The formula is as follows: in, The dynamic risk contribution value for the i-th triggered rule; The basic risk weight for the i-th rule; Let be the current confidence decay factor for the i-th rule, and its value range is . ; The overall risk score is obtained by summing the dynamic risk contribution values of all triggered rules. Its formula is: Where n is the total number of medication safety rules that are triggered; The risk level is determined based on the preset range into which the comprehensive risk score falls, as follows: when If the value is greater than or equal to the first preset threshold, it is determined to be in the first level; when If the value is greater than or equal to the second preset threshold and less than the first preset threshold, it is determined to be the second level; when If the value is less than the second preset threshold, it is judged as the third level.
6. The healthcare service management system and method based on electronic medical records according to claim 1, characterized in that, In the adaptation and early warning unit: Extract all warning response records of the physician within a preset historical time period from the audit log of the electronic medical record system, count the number of times the physician accepted and rejected each type of warning, and calculate the response sensitivity coefficient. Its formula is: in, This represents the number of times the physician rejected a certain type of warning; This represents the total number of times the physician responded to this type of warning; The closer the value is to 1, the lower the physician's acceptance of this type of warning, and the more likely they are to believe that this type of warning is not clinically relevant. The closer the value is to 0, the higher the doctor's acceptance of this type of warning; The risk level output by the conflict detection unit is obtained, and the risk level is offset by the response sensitivity coefficient to obtain the physician-appropriate risk level as follows: When the risk level is Level 1, it remains Level 1. When the risk level is level two and If the risk level exceeds a preset threshold, the risk level will be downgraded to level three. When the risk level is level three and If the risk level falls below a preset threshold, the risk level will be upgraded to the second level. At the same time, the following graded early warning procedures will be implemented, matching the physician's risk level: If the physician's risk level is Level 1, it means that the target drug is absolutely contraindicated for the target patient, generating a blocking warning and preventing the submission of the target drug prescription; If the physician's risk level is Level 2, it means that the target drug poses a significant clinical risk to the target patient. An alert is generated and the physician is shown alternative medication options. If the physician's risk level is level three, it means that there are general precautions for the target drug for the target patient. A reminder message will be generated and displayed on the target drug's prescription submission interface.
7. The healthcare service management system and method based on electronic medical records according to claim 1, characterized in that, In the feedback learning unit: Record the physician's response behavior to the graded early warning operation, and the response behavior is as follows: When a physician chooses to accept the recommendation for a graded alert, the acceptance is recorded and the physician is allowed to modify the prescription for the target drug accordingly before submission. When a physician chooses to refuse the recommendation for a graded warning procedure, the refusal should be recorded and the physician should be required to include the reason for refusal when submitting the original prescription for the target drug. When a physician's response is to refuse the warning and submit the original medical order, this refusal is treated as a negative feedback signal. Based on this negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced, using the following formula: in, The current confidence decay factor before the triggered rule update; The preset attenuation coefficient; This is the updated current confidence decay factor.
8. A healthcare service management method based on electronic medical records, applied to the healthcare service management system based on electronic medical records as described in any one of claims 1-7, characterized in that, Includes the following steps: Step 1: Obtain the target patient's clinical data from the electronic medical record system; Step 2: Based on the name of the target drug prescribed by the physician, retrieve the set of medication safety rules associated with the target drug from the pre-built knowledge base; Step 3: Match the clinical data of the target patient with each rule in the medication safety rule set to obtain the list of triggered rules and their corresponding basic risk weights and current confidence decay factors. Calculate the dynamic risk contribution value of each triggered rule based on the product of its basic risk weight and current confidence decay factor. Accumulate the dynamic risk contribution values of all triggered rules to obtain a comprehensive risk score. Determine the corresponding risk level based on the preset interval into which the comprehensive risk score falls. Step 4: Obtain the physician's historical early warning response records, calculate the physician's response sensitivity coefficient for various early warnings based on the historical early warning response records, and use the response sensitivity coefficient to correct the risk level based on the risk level to obtain the physician's appropriate risk level, and execute the graded early warning operation that matches the physician's appropriate risk level. Step 5: Record the physician's response to the graded warning operation. When the physician's response is to refuse the warning and submit the original medical order, the refusal behavior is used as a negative feedback signal. Based on the negative feedback signal, the current confidence decay factor corresponding to the triggered rule in the medication safety rule set is reduced. The updated current confidence decay factor is stored in the knowledge base for subsequent retrieval.