Patient medicine information processing method and system for thoracic surgery department

Abstract

The invention discloses a medicine information processing method and system for a thoracic surgery patient, and relates to the technical field of medical information processing, and the method comprises the steps: collecting prescription medicine information and real-time physiological monitoring data of the patient; the method comprises the following steps: carrying out structured modeling on prescription drugs, and extracting metabolic enzyme categories, pharmacological targets and pharmacokinetic characteristics; sequentially analyzing the comprehensive similarity degree of adjacent drugs in the prescription to identify potential interaction risks; establishing a dynamic risk correction model by combining indexes such as heart rate, blood pressure and oxyhemoglobin saturation of the patient; in the drug risk recursion process, the risk threshold value is adjusted in real time according to the patient state, and the overall risk level of the prescription is output; individualized dose suggestions and blood concentration prediction results are generated according to parameters such as body surface area and liver and kidney functions, the system is composed of a data acquisition module, a drug modeling module, a similarity analysis module, a risk assessment module, a dose adjustment module and a visualization module, and the system is automatically operated when a doctor makes a prescription, prompts drug risks in real time and provides medication optimization suggestions.

Description

Technical Field

[0001] This invention relates to the field of medical information processing technology, and more specifically, to a method and system for processing patient drug information in thoracic surgery. Background Technology

[0002] Thoracic surgery patients often require the combined use of multiple medications before and after surgery, including analgesics, antibiotics, sedatives, anticoagulants, and hemodynamic modulators. These patients frequently experience physiological changes such as weakened respiratory function, circulatory fluctuations, and increased metabolic load, all of which affect the absorption, distribution, metabolism, and clearance of drugs in the body. If there is overlap in metabolic enzyme pathways or pharmacological targets between drugs, drug interactions can easily occur, leading to reduced efficacy or enhanced toxic side effects. Therefore, real-time assessment of the safety of different drug combinations during clinical use remains a key focus and challenge in thoracic surgery medication management.

[0003] Currently, commonly used clinical methods for drug risk control mainly include pharmacist manual review, drug interaction database queries, and rule-based medication decision support systems. While these methods can be effective in general wards, they have significant limitations in multi-drug combination scenarios such as thoracic surgery. Pharmacist manual review relies on experience-based judgment, which is labor-intensive and inefficient when dealing with complex prescription combinations; drug interaction databases are based on static drug control relationships and cannot provide differentiated assessments based on the specific physiological state of each patient; rule-based decision support systems typically operate on single-drug characteristics or simple combinations, lacking a mechanism for calculating the risk aggregation between multiple drug sequences, and thus failing to reflect the true clinical risk in a timely manner.

[0004] On the other hand, the data structure within hospitals is fragmented, with prescription information, physiological monitoring data, and test results distributed across different systems, making direct information linkage impossible. Existing information systems typically only provide single-time risk alerts and lack the ability for continuous tracking and individualized adjustments. As treatment plans for thoracic surgery patients become increasingly diverse, manual or static rules alone are insufficient to support rapid and accurate risk identification and dosage decisions, especially during postoperative monitoring. A medication information processing method that can integrate multi-source data and dynamically adjust judgment criteria based on patient status is needed to provide physicians with real-time and reliable auxiliary references.

[0005] Therefore, there is an urgent need for a patient medication information processing method and system for thoracic surgery to solve these problems. Summary of the Invention

[0006] The purpose of this invention is to solve the technical problems mentioned in the background section and to provide a method and system for processing patient drug information in thoracic surgery.

[0007] This invention is achieved through the following technical solution: One aspect of the present invention provides a method for processing patient drug information in thoracic surgery, comprising the following steps: S1. Prescription and Physiological Data Collection: Obtain prescription drug information for thoracic surgery patients from the hospital information system, and arrange the drugs in the prescription according to the time of prescription or the order of use to form a drug sequence; at the same time, collect real-time physiological monitoring data of patients, including heart rate, respiratory rate, blood oxygen saturation, mean arterial pressure, weight, height, and test data reflecting liver and kidney function, for subsequent individualized analysis. S2. Drug structured modeling: Each drug in the drug sequence is converted into structured feature data. The feature data includes at least the drug's metabolic enzyme category information, pharmacological target information, and pharmacokinetic parameter information. The pharmacokinetic parameter information is used to characterize the distribution and clearance characteristics of the drug in vivo. S3. Adjacent drug similarity analysis: Drugs in the drug sequence that have overlapping or continuous relationships in clinical administration time are identified as adjacent drugs. The metabolic enzyme characteristics, pharmacological target characteristics and pharmacokinetic characteristics of the adjacent drugs are comprehensively compared, and the comprehensive similarity between adjacent drugs is calculated according to preset weights to characterize the potential interaction strength of the two drugs in combination. S4. Physiological state correction: A physiological state correction model is established based on the real-time physiological monitoring data. When at least one of the following conditions is detected in the patient: increased heart rate, decreased mean arterial pressure, or decreased blood oxygen saturation, the sensitivity of subsequent risk assessment is improved to reflect the impact of the patient's current physiological state on medication safety. S5. Risk Progression and Accumulation: The comprehensive similarity of the adjacent drugs is combined with the physiological state correction results. The risk of each adjacent drug is calculated step by step according to the order of the drug sequence. Risk weight and time smoothing parameters are introduced in the calculation process to combine the current risk value with the risk value of the previous stage, thereby obtaining the cumulative risk result within the prescription range. S6. Dynamic Threshold Comparison and Early Warning: The cumulative risk result is compared with the risk threshold that is dynamically updated based on the patient's current physiological state. When the cumulative risk result is greater than the dynamic risk threshold, a high-risk early warning message for thoracic surgery medication is generated and the corresponding drug combination is marked.

[0008] As a preferred technical solution of the present invention, the time smoothing parameter in S5 is adaptively adjusted according to the patient's respiratory rate so as to enhance the influence of historical risk on current risk when the respiratory rate increases, thereby achieving smooth risk progression.

[0009] As a preferred technical solution of the present invention, the cumulative risk result in S6 is an overall risk index obtained by weighted summation of the risk calculation results of all adjacent drugs in the drug sequence, so as to eliminate the influence of different prescription lengths on the risk assessment results.

[0010] As a preferred embodiment of the present invention, it further includes: calculating the body surface area based on the patient's height and weight, and using the body surface area as the basis parameter for subsequent individualized adjustment of drug dosage.

[0011] As a preferred technical solution of the present invention, it further includes: calculating a liver and kidney function correction coefficient based on the patient's serum creatinine index and liver function index; when the liver and kidney function correction coefficient is lower than a preset normal range, adjusting the target drug dose in the drug dosage calculation.

[0012] As a preferred technical solution of the present invention, the body surface area, the liver and kidney function correction coefficient and the output results of the physiological state correction model are combined to obtain the individualized drug dose for the patient. When the physiological state correction model reflects that the patient is in a high-risk physiological state, the individualized drug dose is automatically reduced or the dosing interval is automatically extended.

[0013] As a preferred embodiment of the present invention, the method further includes: predicting the plasma concentration of the drug after administration based on the individualized drug dosage and the corresponding pharmacokinetic parameters, comparing the predicted plasma concentration with the safe therapeutic concentration range of the drug, and generating an overdose warning when the prediction result exceeds the upper limit of the safe therapeutic concentration range, and using the warning to update subsequent dosing recommendations.

[0014] As a preferred technical solution of the present invention, it further includes: storing the cumulative risk results at multiple time points in chronological order to generate a risk time series, performing trend analysis on the risk time series, and outputting trend-based medication risk warning information when the trend continues to rise or exceeds the dynamic risk threshold within a continuous preset time period.

[0015] The present invention also provides a patient medication information system for thoracic surgery, comprising: The data acquisition module is used to acquire prescription drug information, physiological monitoring data, and laboratory test data; The drug modeling module is used to convert the prescription drug information into structured feature data containing metabolic enzymes, pharmacological targets and pharmacokinetic parameters; The similarity analysis module is used to perform comprehensive similarity analysis on adjacent drugs in a prescription sequence; The physiological correction module is used to generate physiological state correction results based on real-time physiological monitoring data; The risk assessment module is used to perform risk extrapolation and cumulative calculations based on adjacent drugs, and compare the calculation results with dynamic risk thresholds to obtain risk determination results; The dosage adjustment and concentration prediction module is used to generate individualized drug dosages based on body surface area, liver and kidney function correction coefficients, and physiological state correction results, and to predict plasma drug concentrations. The risk trend analysis module is used to perform trend analysis on risk outcomes at different points in time and generate trend warning information; The alarm and visualization module is used to display risk levels, drug combinations, and dosage adjustment suggestions to doctors' workstations or mobile terminals.

[0016] As a preferred technical solution of the present invention, the system is connected to the hospital's electronic medical record system and prescription review system to display the drug interaction analysis results, the risk level of the current prescription, individualized dosage recommendations, and blood drug concentration prediction curves in real time when doctors prescribe or modify thoracic surgery prescriptions, thereby realizing dynamic monitoring and decision support for the entire process of medication use for thoracic surgery patients.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention achieves unified calculation and expression of drug-metabolizing enzymes, pharmacological targets, and pharmacokinetic parameters by structurally modeling prescription drug information and physiological monitoring data of thoracic surgery patients, thus constructing a quantifiable drug similarity analysis mechanism. Compared with traditional static drug interaction database retrieval methods, this scheme can perform dynamic risk calculations in real time based on the patient's specific drug combinations and physiological state, thereby realizing personalized and real-time medication safety assessment. The system adopts a chain-like risk recursive model, which effectively avoids abrupt changes and misjudgments in risk assessment by calculating the cumulative risk between adjacent drugs and introducing a time smoothing factor, thereby improving the stability and accuracy of the assessment results.

[0018] 2. This invention can automatically monitor prescription risks throughout the entire process of patient care, including preoperative, postoperative, and perioperative periods. It dynamically adjusts risk thresholds and recommended dosages based on changes in the patient's heart rate, blood pressure, blood oxygen saturation, and liver and kidney function. When a decrease in blood oxygen saturation or weakened metabolic function is detected, the system automatically triggers a dose reduction or dosing interval extension recommendation and predicts blood drug concentration curves to verify the adjusted safety. This mechanism avoids the lag and subjectivity of traditional manual interpretation, significantly improving the safety and rationality of medication use in complex multi-drug combination therapy for thoracic surgery patients, reducing the incidence of adverse reactions, and demonstrating clear clinical application value.

[0019] 3. The system provided by this invention has a clear structure and a high degree of modularity, enabling seamless integration with existing hospital electronic medical record systems, prescription review systems, and monitoring systems. This achieves a complete closed loop of automatic data collection, risk analysis, trend warning, and visualization. At the algorithm level, the system supports self-learning and parameter optimization, continuously refining the risk assessment model through historical case data, thus achieving an intelligent evolution from "rule-driven" to "data-driven." This solution is not only applicable to thoracic surgery patients but can also be extended to multidisciplinary medication management scenarios such as cardiac surgery, neurosurgery, and critical care medicine, demonstrating broad application value in improving the overall medication safety management level of hospitals. Attached Figure Description

[0020] Figure 1 This is a flowchart of a patient drug information processing method for thoracic surgery proposed in this invention; Figure 2 This invention presents a system block diagram of a patient drug information processing system for thoracic surgery. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the following description is provided in conjunction with embodiments and appendices. Figure 1-2 The present invention will be further described in detail below. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0022] Example 1: In this example, the hospital's information system provides the system with patient prescription data and real-time physiological monitoring data. The system is deployed on the hospital's local area network server and interfaces with HIS (Hospital Information System), EMR (Electronic Medical Record System), LIS (Laboratory System), and monitors in the thoracic surgery ward.

[0023] The core operating process of the system is as follows: First, extract the drugs in the prescription, model them uniformly, then compare these drugs in pairs (in order, adjacent to each other) to see if there are any drugs that are too similar or have overlapping metabolic pathways, targets, or pharmacokinetics. Then, tie this "relationship between drugs" to the patient's "current physiological state" to create a risk recursion chain. Finally, based on this risk, adjust the dosage, predict the blood drug concentration, and give early warnings.

[0024] A method for processing drug information for thoracic surgery patients, in actual operation, the system first performs prescription data collection. Assuming that a thoracic surgery patient's current prescription includes several items such as postoperative analgesics, antibiotics, proton pump inhibitors, and anticoagulants, the prescription drugs read by the system from the HIS are organized into an ordered prescription drug sequence: , in, This represents the complete set of medications currently being used by the patient. arrive This indicates the first to nth medications, ordered by prescription time or medical order, where n is the current patient's medication dosage. This step also retrieves physiological and laboratory data related to medication safety from the monitor and testing system, including heart rate. (unit: breaths / min), respiratory rate (Unit: times / min), blood oxygen saturation (Unit: %) Mean Arterial Pressure (Unit: mmHg) Weight (Unit: kg), Height (unit: cm), serum creatinine (Used to estimate glomerular filtration rate (GFR), in μmol / L) and liver function indicator alanine aminotransferase (ALT) AST (Unit: U / L)

[0025] In order for the system to perform algorithmic calculations on these drugs, the drugs need to be represented in a unified structured way, for each drug in the prescription sequence. (i=1...n), the system is modeled in the following form: , in, This represents the set of metabolic enzyme classes that the i-th drug mainly depends on, such as CYP3A4, CYP2D6, etc., which are stored here in encoded form or one-hot form; The set of major pharmacological targets of the i-th drug, such as bacterial cell wall synthesis, opioid receptors, proton pumps, etc., are also represented by sets or encoding methods; It indicates the apparent volume of distribution of the drug (in L or L / kg), reflecting the extent to which the drug is distributed in the body; The clearance rate of a drug (unit: L / h or mL / min) is used to describe the body's ability to clear the drug. The absorption rate constant of a drug (units) For oral or other dosage forms that require absorption, the significance of this step is to transform the messy text, drug codes, and routes of administration in the original HIS into a set of "computable feature vectors" that contain both category information and pharmacokinetic numerical information, so that they can directly participate in subsequent similarity and risk calculations.

[0026] After modeling is completed, this embodiment does not compare all drugs pairwise using Cartesian products, but rather performs chain comparisons of adjacent drugs according to the prescription order, that is, it compares... and Compare once, put and Compare again, until... and This comparison reduces computational load and better reflects the more likely interactions between drugs used concurrently and prescribed close together in clinical practice. The overall similarity between two adjacent drugs is defined as follows: , in, These are weighted coefficients for the five characteristic components, used to adjust the relative importance of metabolic enzyme similarity, target similarity, and the similarity of the three pharmacokinetic parameters. The sum of the five coefficients is 1. This indicates the size of the intersection of the sets of enzymes that metabolize the two drugs. This component represents the size of the union of the sets of enzymes that metabolize the two drugs; it reflects whether they are metabolized via the same pathway. and Similarly, it only reflects the degree of overlap of pharmacodynamic targets; , , These are the average distribution volume, average clearance rate, and average absorption rate constant of similar drugs pre-statistically collected within the system, used for normalization to allow differences in different dimensions to be included in the same formula; the exponential term... The purpose is to make the greater the difference, the faster the similarity decays. This formula is calculated as follows. Between 0 and 1, the closer to 1, the more similar the two drugs are in metabolism or pharmacokinetics, and the higher the possibility of interaction or superimposed risks. Thoracic surgery patients often experience hypoxia, hypotension, and rapid heart rate after surgery, and these phenomena themselves can narrow the safety margin of some drugs. Therefore, this embodiment introduces a "physiological correction function" after drug similarity to take into account the patient's current state.

[0027] The physiological correction function is defined as: , in, It is an overall correction coefficient based on the patient's current physiological state; This is the currently monitored heart rate. This is a normal reference heart rate (e.g., 70 beats per minute). It refers to the degree of deviation in heart rate; It is the current mean arterial pressure. It is a reference mean arterial pressure (e.g., 90 mmHg). Used to reflect the degree of low pressure, the reason is This is because the lower the blood pressure, the higher the risk. This refers to the current blood oxygen saturation level. 95% is often considered a healthy lower limit, hence the use of 95% here. This indicates the severity of hypoxia; These are the weights of three channels: heart rate, blood pressure, and blood oxygen saturation, used to control which physiological indicator's change has a greater impact on medication risk. The result is that if a patient has just undergone surgery and their blood oxygen saturation has dropped to 90%, It will be greater than 1, thus pushing up all the subsequent risks and causing the system to issue an early warning.

[0028] After obtaining the similarity between adjacent drugs and the physiological correction function, chain-like risk recursion can be performed. The recursion formula used in this embodiment is: , in, This represents the cumulative risk value when recursively applying to the i-th drug edge (that is, between the i-th and i+1-th drugs); It is the weight of the i-th drug on the overall risk. For example, the weight of opioids and aminoglycosides currently used after surgery can be set higher. The similarity score calculated earlier indicates how "similar" the two drugs are; the input is the risk smoothing or memory coefficient, which smooths out previously formed risks. By appropriately incorporating these risks into the current step, the risk is not abrupt but rather accumulates gradually with the drug sequence. Generally, the value is between 0.5 and 0.9. This indicates that the risk is 0 before the recursion begins. The intuitive meaning of this formula is: the risk of the current drug pair = the risk inherent in the current drug pair itself + the accumulated risk from all previous medications, multiplied by a factor amplifying the real-time physiological state. After completing one cycle of recursion, an overall measure is needed to determine whether this patient's prescription has exceeded the department's safety threshold. Therefore, the overall risk is defined as: , in, This is the average risk value for the entire prescription sequence. It calculates the risk between all adjacent drugs, sums the results, and then divides by the sequence length minus 1 to make it independent of the number of drugs used. The system will then use this... Compare this to a risk threshold, which is not a fixed constant but is adjusted according to the patient's level of hypoxia. This makes it easier to detect critically ill patients. The formula for calculating the threshold is: , in, It is the dynamic risk threshold at the current moment; These are the basic thresholds for department or system configuration; This is the blood oxygen correction factor, indicating how much the threshold should decrease for every 196 decrease in blood oxygen saturation; other symbols are the same as before. If Normally, the part in parentheses is close to 1, and the threshold remains unchanged; if blood oxygen levels decrease, the threshold decreases accordingly to trigger an earlier alarm. After assessing the risk, individualized dosage adjustments are necessary. This is another layer of logic in this plan: it's not just about telling the doctor "there is a risk," but rather "how much you should administer." To achieve this, the body surface area is first calculated: , in, This is the patient's body surface area, in m2. It is body weight, in kg; It is height, in cm; 0.725 is the standard coefficient of the DuBois formula, and the purpose of body surface area is to make a basic individualization of the dose based on the body surface area.

[0029] Next, liver and kidney function should also be taken into account, because many thoracic surgery patients experience fluctuations in kidney function before or after surgery. The liver and kidney function correction factor is defined as: , in, It is a comprehensive organ function correction coefficient; This is the glomerular filtration rate calculated based on serum creatinine, in mL / min / 1.73 , This roughly indicates the percentage of kidney function; if it is greater than 1, it is counted as 1 and not exaggerated. and These are liver function transaminase levels. Summing them and dividing by 200 normalizes the values ​​to approximately... between; This is a weighting coefficient for liver function, indicating the degree to which abnormal liver function affects the dosage. The smaller K is, the worse the organ function, and the lower the dosage should be subsequently adjusted. Considering body surface area, organ function, and physiological state, the final individualized dosage adjustment is: , in, It is the adjusted dose of the i-th drug; This is the standard dosage listed in the instruction manual or the commonly used hospital routes; This converts the patient's body surface area to a value of 1.73. The standard person as the benchmark; These are the liver and kidney function correction coefficients calculated earlier; It is a dose-downregulation function based on physiological state, defined as follows: , in, This is the dose reduction factor, indicating how much the dose should be reduced to control increased heart rate and decreased blood oxygenation; other symbols have the same meaning as before. Thus, if the patient has a fast heart rate and poor oxygenation, this will lead to… The dosage automatically decreases.

[0030] In order to help doctors see whether the concentration of the drug you are prescribing in the patient will exceed the limit, the system also needs to predict the plasma drug concentration.

[0031] This embodiment uses a common one-compartment model oral / absorption formula: , in, This represents the theoretical plasma concentration of the i-th drug at t hours after administration, expressed in mg / L. This is the adjusted dosage of the drug that was just calculated; It is the absorption rate constant of the drug; It is the distributed volume; It is the elimination rate constant, where It is the clearance rate; and It is an exponential term representing both the elimination and absorption processes. If the calculated... Exceeding the upper limit of the therapeutic window of the drug The system will then provide a message indicating that "the dosage is too high or the dosing interval needs to be extended".

[0032] Considering that the condition of thoracic surgery patients can change and cannot remain at a fixed risk level indefinitely, the system also generates a time series of the calculated total risk to show whether the risk is increasing or decreasing. The time series is defined as follows: , in, It is the overall risk trend value at the current moment; It is a time smoothing coefficient used to control the ratio of new risks to old risks; This reflects the risk trend of the previous moment; This is the overall prescription risk calculated in this study; It is a weighted rate of change of physiological parameters, used to indicate whether a patient's vital signs are getting worse or better. For example, when the heart rate increases, blood pressure decreases, or blood oxygen decreases, this item is positive, indicating that the risk should be raised. It is the rate of change weight, used to control the intensity of the impact on risk when physiological deterioration occurs rapidly.

[0033] In terms of system implementation, the data acquisition module is responsible for interacting with the HIS and monitors, periodically retrieving prescriptions and vital signs data; the drug modeling module converts all prescription drugs into... A unified format; the similarity calculation module calculates all similarities sequentially. The risk assessment module executes a chain recursive formula to generate the risk for each segment and the overall risk; the dosage adjustment module executes... The formula is then passed to the concentration prediction module; finally, the visualization output module outputs the results. Each drug The curve and the text results such as "suggestion to reduce dosage / suggestion to change medication / suggestion to extend dosing interval" are pushed to the doctor's workstation or ward terminal.

[0034] Example 2: In a preferred embodiment, the thoracic surgery patient drug information processing system provided by the present invention is deployed in the hospital information platform server. The system adopts a distributed architecture design and includes functional units such as a data acquisition module, a drug modeling module, a similarity analysis module, a physiological correction module, a risk assessment module, a dose adjustment and concentration prediction module, a risk trend analysis module, and an alarm and visualization module. The modules interact with each other through an internal communication bus or API interface.

[0035] The system acquires patient information through interfaces with the Hospital Information System (HIS), Electronic Medical Record System (EMR), and patient monitors. The data acquisition module continuously monitors patient prescriptions and monitoring data. When a new prescription is entered or physiological parameters fluctuate significantly, a risk analysis process is automatically triggered. The collected data includes basic patient identification information, ward information, diagnosis results, prescription drug information (drug code, dosage, route of administration, frequency, and prescription time), physiological monitoring information (heart rate, respiratory rate, mean arterial pressure, and blood oxygen saturation), and laboratory test data (serum creatinine, alanine aminotransferase, aspartate aminotransferase, etc.).

[0036] After receiving prescription drug information, the drug modeling module calls upon the built-in drug knowledge base to extract features such as the metabolic enzyme category, main targets, volume of distribution, clearance rate, and absorption rate for each drug. The knowledge base is built upon a pharmacopoeia database and hospital medication big data, and can output structured drug feature data in a unified format. This module converts drug names in the original medical order into standardized drug codes and generates feature vectors for subsequent algorithmic analysis.

[0037] After receiving the modeling results, the similarity analysis module compares adjacent drugs according to the order in which they are prescribed. Adjacent drugs refer to two drugs whose administration times overlap or are consecutive, such as postoperative analgesics and antibiotics used simultaneously. The module comprehensively calculates similarity data between drugs based on three dimensions: similarity of metabolic enzyme categories, similarity of targets, and differences in pharmacokinetic parameters, and generates comprehensive similarity data according to weighted settings. The module outputs the potential interaction strength value for each pair of adjacent drugs, which serves as input for subsequent risk recursion.

[0038] The physiological correction module is used to capture the patient's physiological signals in real time and output correction coefficients. This module has a built-in multi-parameter monitoring interface that can receive real-time data from bedside monitors or surgical anesthesia systems. Based on the deviations in heart rate, blood pressure, and blood oxygen saturation, the system dynamically corrects the overall risk model, automatically increasing the risk assessment weight when blood oxygen or blood pressure decreases, thus reflecting the differences in the individual patient's condition.

[0039] The risk assessment module integrates similarity analysis results with physiological correction coefficients and executes a chain-recursive logic. The module calculates risk values ​​pairwise according to the order of the prescribed drugs, and introduces a time smoothing parameter in each iteration to balance current and historical risks. The system ultimately outputs an overall risk index for the prescription, which is updated in real time and recorded in the background database. If the risk value exceeds a dynamic threshold, the system immediately issues a warning.

[0040] The dosage adjustment and concentration prediction module operates based on risk assessment. It calculates body surface area by combining the patient's weight and height, and generates liver and kidney function correction coefficients based on serum creatinine and liver function indicators. The system then adjusts the dosage of the current medication regimen individually, outputting a recommended dose or suggesting an extension of the dosing interval. When the physician confirms the adjustment, the module further predicts the changes in plasma concentration of each drug after administration and compares them with the drug's safe therapeutic range. If the predicted value exceeds the safe upper limit, the module displays a warning of excessive dosage to the physician's workstation; if the predicted value is too low, it provides a warning of insufficient efficacy. The prediction results are simultaneously fed back to the risk assessment module for the next round of risk correction, creating a closed-loop system.

[0041] The risk trend analysis module is responsible for tracking and statistically analyzing changes in risk over time. This module continuously stores risk assessment results, forming continuous risk time series data. The system uses a sliding time window to calculate the rate and direction of risk change. When it detects a sustained increase in risk, continuous exceedance of thresholds, or a deteriorating trend in the patient's physiological indicators, the module generates a trend risk warning. Doctors can visually observe risk trends through time series graphs, allowing for early intervention.

[0042] The alarm and visualization module provides a user-friendly interface. Connecting to doctor workstations, nurse station terminals, and mobile ward round equipment, this module can display drug similarity, risk levels, dosage adjustment recommendations, and predicted blood drug concentrations in the form of lists, heatmaps, or graphs. When a high-risk drug combination exists, the system highlights it in red on the interface and displays text such as "High-risk drug pair" and "Recommended dosage adjustment" in a pop-up window. The module also supports automatically pushing risk reports to the pharmacist review system, enabling collaborative decision-making between physicians and pharmacists.

[0043] In practical applications, when a thoracic surgery patient enters the intensive care unit post-surgery, the system automatically loads the patient's medical orders and initiates risk monitoring. If analgesics and anticoagulants are added to the medical orders, the system immediately compares their characteristics. If it detects identical metabolic enzyme pathways and a decrease in the patient's blood oxygen saturation, the system issues a "high-risk interaction" warning within seconds and displays "It is recommended to adjust the anticoagulant dosage or change the analgesic" on the interface. Doctors can refer to the adjustment suggestions and concentration prediction charts generated by the system to optimize medication use without increasing the risk.

[0044] The system described in this embodiment is compatible with existing hospital information systems and requires no additional hardware. The system adopts a modular deployment approach, running on a single server or in a cloud container, and is scalable and maintainable. Through this system, thoracic surgeons can monitor patients' medication safety status in real time throughout the entire medication process, enabling automatic identification of prescription risks, intelligent dosage adjustment, and dynamic tracking of risk trends, thereby significantly reducing the incidence of adverse medication events.

[0045] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A patient drug information processing method for thoracic surgery, characterized by, Comprise the following steps: S1, prescription and physiological data collection: obtain the prescription drug information of the thoracic surgery patient from the hospital information system, and the drugs in the prescription are arranged in the order of time or use sequence to form a drug sequence; at the same time, the real-time physiological monitoring data of the patient is collected, including heart rate, respiratory rate, oxygen saturation, mean arterial pressure, body weight, height and test data reflecting liver and kidney function, which are used for subsequent individual analysis; S2, drug structured modeling: converting each drug in the drug sequence into structured feature data, the feature data at least including the metabolic enzyme category information, pharmacological target information and pharmacokinetic parameter information of the drug, and the pharmacokinetic parameter information is used to represent the distribution and clearance characteristics of the drug in the body; S3, similarity analysis of adjacent drugs: the drugs in the drug sequence which have overlapping or continuous relationship in clinical administration time are determined as adjacent drugs, the metabolic enzyme characteristics, pharmacological target characteristics and pharmacokinetic characteristics of the adjacent drugs are compared comprehensively, and the comprehensive similarity between the adjacent drugs is calculated according to the preset weight, which is used to represent the potential interaction strength of the two drugs; S4, physiological state correction: a physiological state correction model is established according to the real-time physiological monitoring data, when at least one of the conditions of increased heart rate, decreased mean arterial pressure or decreased oxygen saturation is detected, the sensitivity of subsequent risk assessment is improved to reflect the influence of the current physiological state of the patient on the safety of drug use; S5, risk recursion and accumulation: the comprehensive similarity of the adjacent drugs and the physiological state correction result are combined, the risk of each adjacent drug is calculated in sequence according to the drug sequence, and the risk weight and time smoothing parameter are introduced in the calculation process to combine the current risk value and the previous stage risk value, so as to obtain the cumulative risk result in the range of the prescription; S6, dynamic threshold comparison and early warning: the cumulative risk result is compared with the risk threshold dynamically updated according to the current physiological state of the patient, when the cumulative risk result is greater than the dynamic risk threshold, the thoracic surgery drug high risk early warning information is generated and the corresponding drug combination is marked.

2. The patient drug information processing method for thoracic surgery according to claim 1, wherein The time smoothing parameter in S5 is adaptively adjusted according to the respiratory rate of the patient, so as to enhance the influence of historical risk on current risk when the respiratory rate is increased, thereby realizing the smooth recursion of risk.

3. The patient drug information processing method for thoracic surgery according to claim 1, wherein The cumulative risk result in S6 is the overall risk index obtained by weighting and summarizing the risk calculation results of all adjacent drugs in the drug sequence, so as to eliminate the influence of different prescription lengths on the risk assessment result.

4. The patient drug information processing method for thoracic surgery according to claim 1, characterized by Further comprising: calculating the body surface area according to the height and weight of the patient, and taking the body surface area as the basic parameter for subsequent individual adjustment of drug dose.

5. The patient drug information processing method for thoracic surgery according to claim 4, wherein Further comprising: calculating the liver and kidney function correction coefficient according to the blood creatinine index and liver function index of the patient, and reducing the target drug dose in the drug dose calculation when the liver and kidney function correction coefficient is lower than the preset normal range.

6. The patient drug information processing method for thoracic surgery according to claim 4, wherein The body surface area, the liver and kidney function correction coefficient, and the output result of the physiological state correction model are combined to obtain an individualized drug dose of the patient, and when the physiological state correction model reflects that the patient is in a high-risk physiological state, the individualized drug dose is automatically reduced or the drug administration interval is automatically extended.

7. The patient drug information processing method for thoracic surgery according to claim 6, wherein Also comprising: Based on the individualized drug dose and the pharmacokinetic parameters corresponding to the drug, the plasma concentration of the drug after administration is predicted, and the predicted plasma concentration is compared with the safe therapeutic concentration range of the drug. When the prediction result exceeds the upper limit of the safe therapeutic concentration range, an overdose prompt is generated and the prompt is used to update the subsequent drug administration recommendation.

8. The patient drug information processing method for thoracic surgery according to claim 3, wherein Also comprising: The cumulative risk results at multiple time points are stored in chronological order to generate a risk time series, and trend analysis is performed on the risk time series. When the trend continues to rise or exceeds the dynamic risk threshold in a continuous preset time period, trend-based drug risk warning information is output.

9. A thoracic surgery patient medication information processing system for implementing the method of any one of claims 1 to 8, characterized by Comprising: A data acquisition module for acquiring prescription drug information, physiological monitoring data, and laboratory test data; A drug modeling module for converting the prescription drug information into structured feature data containing metabolic enzymes, pharmacological targets, and pharmacokinetic parameters; A similarity analysis module for comprehensive similarity analysis of adjacent drugs in a prescription sequence; A physiological correction module for generating a physiological state correction result based on real-time physiological monitoring data; A risk assessment module for performing risk recursion and accumulation calculation based on adjacent drugs, and comparing the calculation result with a dynamic risk threshold to obtain a risk determination result; A dose adjustment and concentration prediction module for generating an individualized drug dose based on body surface area, liver and kidney function correction coefficient, and physiological state correction result, and predicting plasma drug concentration; A risk trend analysis module for trend analysis of risk results at different time points and generation of trend-based warning information; An alarm and visualization module for displaying risk level, drug combination, and dose adjustment suggestion to a doctor workstation or mobile terminal.

10. The system of claim 9, wherein, The system is data-connected with a hospital electronic medical record system and a prescription review system, and is used to display drug interaction analysis results, risk level of the current prescription, individualized dose suggestion, and blood drug concentration prediction curve in real time when a doctor establishes or modifies a thoracic surgery prescription, so as to realize dynamic monitoring and decision support of the whole process of thoracic surgery patient medication.

Images