A method and system for intelligent review of antitumor drug prescriptions

The intelligent anti-tumor drug prescription review system solves the problem of insufficient multi-dimensional review in existing technologies, realizes efficient and accurate anti-tumor drug prescription review, adapts to the individualized needs of different hospitals, and provides full-process traceability and quality analysis support.

CN123072618APending Publication Date: 2026-11-10BEIJING GAOBO HOSPITAL CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611359376.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-03
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

In the existing technology, the review of anti-tumor drug prescriptions lacks an automated solution, resulting in low efficiency and high error rate of manual review. Furthermore, it cannot simultaneously cover the multi-dimensional review needs such as matching disease diagnosis names, calculating individualized dosages, selecting solvents, and ensuring compliance with drug concentrations. Inconsistent body surface area calculations and a lack of systematic review of solvents and concentrations also contribute to the problem.

Method used

This paper provides an intelligent review method and system for anti-tumor drug prescriptions. It performs four-dimensional linkage review through automated linkage steps, including disease diagnosis matching, individualized dosage stratification calculation, solvent compatibility testing, and compliance verification of dosage concentration. Combined with multi-source data interfaces and modular architecture, it realizes full-process automation and intelligent review.

Benefits of technology

It enables comprehensive and systematic intelligent screening of anti-tumor drug prescriptions, improves review efficiency, reduces the error rate of manual calculation, reduces the risk of solvent mismatch and concentration imbalance, supports localization adaptation for different hospitals, and provides traceability and quality analysis support for the review process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123072618A_ABST
    Figure CN123072618A_ABST
Patent Text Reader

Abstract

This invention relates to the field of medical information technology and intelligent assisted diagnosis and treatment, and discloses an intelligent review method and system for anti-tumor drug prescriptions, including the following automated linkage execution steps: Disease diagnosis – standardized matching and retrieval of chemotherapy regimens: Accessing the hospital's HIS system to obtain the patient's original clinical diagnosis text; constructing a standardized disease diagnosis name database that supports local configuration based on the original clinical diagnosis text for unambiguous matching; upon successful matching, retrieving the corresponding compliant standardized medication parameters for anti-tumor drugs, including standard single dose, daily maximum dose, specific route of administration, upper limit of drug concentration, and chemotherapy regimen. By integrating a four-dimensional linkage automated review chain of disease diagnosis matching, individualized dose calculation, solvent compatibility detection, and drug concentration verification, a comprehensive and systematic intelligent screening of anti-tumor drug prescriptions is achieved, overcoming the shortcomings of existing systems such as fragmented review dimensions and high risk of missed reviews.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of medical information technology and intelligent assisted diagnosis and treatment technology, specifically to an intelligent review method and system for anti-tumor drug prescriptions. Background Technology

[0002] Anti-tumor drugs are among the highest-risk and most difficult-to-review medications in the daily compounding work of intravenous drug preparation centers (PIVAS). Compared with ordinary drugs, the review of prescriptions for anti-tumor drugs faces the following unique challenges:

[0003] First, the influencing factors are multidimensional and complex. The rational use of anti-tumor drugs requires simultaneous consideration of multiple dimensions, including patient disease diagnosis and risk stratification, individualized dosage (based on body surface area or weight), solvent selection suitability, and compliance with dosage concentration regulations. Deviations in any of these aspects can lead to insufficient efficacy, increased toxicity, or even endanger the patient's life. Currently, there is a lack of automated review solutions that can integrate these four review dimensions.

[0004] Secondly, manual review is inefficient and prone to errors. Currently, the vast majority of medical institutions still rely on pharmacists to manually review prescriptions for anti-tumor drugs. Reviewing a single prescription requires a pharmacist to verify the match between diagnosis and medication, calculate the patient's body surface area and dosage, consult the drug's instructions to confirm solvent compatibility, and calculate and verify whether the dosage concentration is within a safe and effective range. This process is time-consuming and prone to errors due to incomplete information gathering or insufficient knowledge. Statistics show that incorrect solvent selection, exceeding the instructions for use, and inappropriate dosage concentrations are among the most common problems in inappropriate anti-tumor drug prescriptions.

[0005] Third, the existing information systems have limited auxiliary capabilities. Although some hospitals have tried using rational drug use monitoring systems to assist in the review of anti-tumor drug prescriptions, the existing systems still lack the ability to review inappropriate drug dosages, incorrect solvent selection, and incorrect routes of administration, and still need to be combined with manual review. There is a lack of an intelligent review method for anti-tumor drug prescriptions that can simultaneously cover four dimensions: disease diagnosis name matching, individualized dosage calculation, solvent compatibility testing, and compliance verification of drug concentration.

[0006] Fourth, there is a lack of uniformity in the rules for calculating body surface area and dosage. The dosage of antitumor drugs is usually calculated based on the patient's body surface area (BSA), but different medical institutions use different formulas for calculating BSA (such as the Xu Wensheng formula). The existing system lacks a standardized method that can uniformly configure, automatically calculate, and verify the rationality of the dosage.

[0007] Fifth, the review of solvents and concentrations lacks a systematic approach. Different types of antitumor drugs have strict specific requirements for the type and volume of solvents (e.g., some drugs are only compatible with 0.9% sodium chloride injection, some drugs are only compatible with 5% glucose injection, and some drugs are only compatible with 100ml of solvent), and most drugs have specified safe dosing concentration ranges.

[0008] Therefore, a solution is proposed. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention provides an intelligent review method and system for anti-tumor drug prescriptions, which solves the problems mentioned in the background section.

[0010] To achieve the above objectives, the present invention provides the following technical solution: an intelligent review method for anti-tumor drug prescriptions, comprising the following automated and linked execution steps:

[0011] S1 Disease Diagnosis - Standardized Matching and Retrieval Steps: Access the hospital's HIS system to obtain the patient's original clinical diagnosis text, construct a standardized disease diagnosis name database that supports localized configuration based on the original clinical diagnosis text, and perform standardized matching. After successful matching, retrieve the corresponding compliant anti-tumor drug's standardized medication parameters. The standardized medication parameters include standard single dose, daily maximum dose, dosing frequency, exclusive route of administration, upper limit of dosing concentration, and medication regimen. Automatically remove medication regimens that are off-label, outdated according to guidelines, or without instruction manual support, and retain only compliant medication regimens based on drug instructions and national oncology clinical guidelines.

[0012] S2 Personalized Dosing Dosing Stratification Calculation and Verification Steps: Based on the drug dosing baseline, two types of calculation logics are automatically switched: S2-1 body surface area dosing calculation logic and S2-2 body weight dosing calculation logic.

[0013] S2-1 Body Surface Area Dosing Calculation Logic: Retrieves the patient's height and weight physiological data, calculates the patient's body surface area (BSA) according to the hospital's preset formula, supports one-click configuration and switching of multiple algorithms such as Xu Wensheng's formula, and the single dose = standard body surface area dose of the drug × BSA; the calculated dose is simultaneously compared with the single safe dose range and single dose concentration limit of the disease matching plan, and the dose is judged to be unreasonable if either threshold is exceeded.

[0014] S2-2 Body Weight Dosing Calculation Logic: Single Dosage = Standard Weight Dosage of Drug × Patient's Actual Weight; Differentiated safe dosage thresholds are set based on the patient's liver and kidney function test indicators, and the upper limit of the dosage is automatically lowered or a dosage adjustment is prompted for patients with liver and kidney function impairment;

[0015] S3 Drug-Specific Solvent Compatibility Testing Steps: For the target anti-tumor drug matched in S1, the system retrieves the built-in whitelist of compatible solvents and blacklist of incompatible solvents from the drug database; it automatically verifies the solvents entered into the doctor's order based on compatibility stability, acid-base compatibility, precipitation risk, and drug degradation rate; only solvents on the whitelist are deemed qualified, while solvents on the blacklist are directly marked with compatibility risks, and the type of solvent conflict risk is output.

[0016] S4 Dosage Concentration Compliance Range Quantitative Verification Steps: Calculate the actual dosage concentration using the individualized single-dose dose verified in S2 as the numerator and the total infusion volume of the prescribed solvent as the denominator; retrieve the upper and lower limits of the legally mandated safe concentration of the drug and compare the actual concentration with the standard range; distinguish between insufficient efficacy due to excessively low concentration and risk due to excessively high concentration and mark them accordingly.

[0017] S5 Tiered Audit Result Output, Differential Early Warning and Closed-Loop Review Steps:

[0018] If all four-dimensional verifications from S1 to S4 are compliant, the first-level review result will be output with the compliance level marked and directly pushed to the PIVAS drug dispensing scheduling port.

[0019] If any of the S1-S4 dimensions fails to pass the verification, a risk warning will be triggered, and the abnormal dimension, the quantified deviation value, and the corresponding clinical safety risk will be marked.

[0020] When the system identifies off-label drug use, it will not release the medication directly, but will prompt "It is recommended that the pharmacist verify whether there is hospital filing or informed consent from the patient". The medication can only be released after the pharmacist confirms it. The authority to block or release the medication is determined by the pharmacist during the manual review process.

[0021] All audit process data, early warning records, and manual review operations are archived and retained in a unified manner.

[0022] Among them, steps S1, S2, S3, and S4 are a serial linkage automated review chain. If the verification of the preceding steps is unqualified, the subsequent calculation will be terminated directly and the corresponding warning will be output, realizing the four-dimensional integrated automated review of anti-tumor drug prescription diagnosis matching, individualized dosage, solvent compatibility, and infusion concentration.

[0023] Preferably, it also includes the S6 review log full-link archiving and quality analysis steps: automatically collecting the review time, patient basic information, original medical order parameters, four-dimensional verification sub-items results, warning type, pharmacist review operation, medical order modification record, and final release status for each medical order; supporting multi-dimensional retrieval of logs by drug type, clinical department, abnormal warning type, and time interval, generating PIVAS prescription review quality statistical reports, which include the proportion of various unreasonable medical orders, the ranking of high-frequency risk drugs, and the comparison data of departmental prescription review pass rates, for continuous iteration and optimization of the review rule base.

[0024] Preferably, the solvent compatibility test in step S3 is performed by pre-establishing a drug-solvent incompatibility knowledge base, which stores quantitative data on adverse reactions after different cytotoxic drugs are combined with incompatible solvents. When the system determines that the solvents are incompatible, it simultaneously retrieves the corresponding adverse reaction data and displays it along with the warning information, such as risk labels for intravenous irritation, worsening of bone marrow suppression, tubing blockage, and drug failure, to assist pharmacists in quickly assessing the degree of risk.

[0025] Preferably, the risk warning in step S5 is divided into three levels: Level 1 warning is a minor parameter deviation that can be easily adjusted and corrected by a pharmacist; Level 2 warning is a moderate medication risk that requires contacting a clinician to verify and adjust the prescription before it can be reviewed and released; Level 3 warning is a severe safety risk that requires in-hospital approval before it can be released.

[0026] An intelligent review system for anti-tumor drug prescriptions includes three integrated hardware and software modules: a data terminal, a review engine, and an application terminal that communicate with each other in sequence.

[0027] The data terminal, used for multi-source clinical data storage and interfacing with external system data, includes:

[0028] Disease Diagnosis Name Database: Stores diagnosis names;

[0029] Drug Information Database: Stores standard dosage, compatible solvent white and black lists, safe dosing concentration range, upper limit of dosing concentration, and incompatibility knowledge base for anti-tumor drugs introduced from the PIVAS Center;

[0030] Patient information API interaction interface: bidirectional connection to the hospital HIS system, real-time synchronization of patient diagnosis, height and weight, liver and kidney function test results, age and gender, and electronic records of informed consent forms;

[0031] The audit engine is the core computing and processing unit of the system, with five functional modules that work in parallel: disease diagnosis matching module, dose stratification calculation module, solvent compatibility detection module, concentration verification module, and risk warning module.

[0032] The application terminal is a pharmacist's visual interactive terminal, which includes a module for displaying review results, a module for risk warning pop-ups, a manual review operation interface, and a preferred interface for log querying and exporting quality reports.

[0033] Preferably, the review engine also has a built-in knowledge base update and maintenance module. This module is configured with a scheduled automatic update interface and a manual import channel. The scheduled automatic update interface connects to the hospital's existing drug instruction manual and clinical guideline database or a third-party rational drug use information platform. By acquiring authoritative materials such as drug instruction manuals and oncology clinical guidelines, it enables automatic updates of drug review rules. The manual import channel supports the maintenance and updating of medication plans and in-house self-made medication plans. After the update, the entire system takes effect in real time, ensuring the timeliness and adaptability of the review rules within the hospital.

[0034] Preferably, the dose stratification calculation module has a built-in multi-body surface area calculation formula switching unit, stores standard algorithms such as Xu Wensheng's formula, and also reserves a custom formula editing input port; the hospital administrator can select the default calculation algorithm in the system background, or configure independent BSA calculation formulas for pediatrics and adults respectively, and the module automatically retrieves the corresponding formula according to the patient information to complete the calculation of body surface area and drug dosage.

[0035] Preferably, the manual review interface of the application terminal has an online modification and linkage verification function for medical orders; after the pharmacist completes the parameter correction and submits the warning medical order, the system automatically sends the modified medical order back to the review engine to re-execute the S1-S4 full four-dimensional automated review, and after the secondary verification is completed, the review results and logs are updated.

[0036] This invention provides a method and system for intelligent review of antitumor drug prescriptions. It has the following beneficial effects:

[0037] 1. This invention integrates a four-dimensional automated review process involving disease diagnosis matching, individualized dosage calculation, solvent compatibility testing, and drug concentration verification to achieve comprehensive and systematic intelligent screening of anti-tumor drug prescriptions, overcoming the shortcomings of existing systems such as fragmented review dimensions and high risk of missed reviews.

[0038] 2. This invention automatically captures patients' physiological and diagnostic data by connecting to the hospital's HIS and automatically completes the calculation of body surface area, drug dosage, and infusion concentration by incorporating standardized calculation rules. This replaces a large amount of repetitive manual calculation and data retrieval work by pharmacists, greatly improving the efficiency of PIVAS prescription review and reducing manual calculation errors.

[0039] 3. This invention establishes a blacklist and whitelist of drug solvents and verifies solvents from multiple dimensions such as compatibility stability, acid-base conflict, precipitation and degradation. At the same time, it quantitatively compares the dosage concentration with the legally safe range, thereby achieving a dual rigid interception of two clinically common drug use risks: solvent mismatch and concentration imbalance, thus reducing the probability of serious adverse drug reactions from the source.

[0040] 4. This invention incorporates multiple body surface area calculation formulas and supports flexible configuration and switching by ward and patient population, enabling adaptation to different hospitals, adult and pediatric clinical calculation standards, thereby improving the system's localization and cross-institutional applicability.

[0041] 5. This invention achieves full traceability of the review process by completely recording the review data of each medical order throughout the entire process and supporting multi-dimensional retrieval and automatic generation of prescription review quality analysis reports. At the same time, it provides quantitative data support for the iteration of medication rules and specialized training for pharmacists, forming a closed loop for continuous improvement of prescription review quality.

[0042] 6. This invention achieves automatic transfer of compliant medication orders to the dispensing stage by natively connecting the system to the PIVAS platform and integrating data interfaces for review, dispensing, label printing, and drug delivery, thus constructing an integrated intelligent management and control business process for prescription review and dispensing of cytotoxic drugs. Attached Figure Description

[0043] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0044] The following will refer to the appendix to the specification of this invention. Figure 1 The technical solutions in the embodiments of the present invention are clearly and completely described herein. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0045] I. Example 1: Implementation Process of Intelligent Review Standard for Antitumor Drug Prescriptions

[0046] The S1-S5 fully automated review steps described in this embodiment use a 56-year-old non-small cell lung cancer patient as the review sample to intuitively demonstrate the entire four-dimensional linkage review process.

[0047] S1 Disease Diagnosis Name Matching and Search Implementation Process

[0048] The system reads the original patient diagnosis text pushed by the HIS system in real time via the patient information interface API: "Non-small cell lung cancer". The system calls the disease diagnosis name database to construct a standardized disease diagnosis name library that supports localized configuration from the original clinical diagnosis text for standardized matching:

[0049] Extract the core disease keyword "non-small cell lung cancer" to complete basic diagnostic matching;

[0050] Based on the doctor's prescription, the system automatically searches for compliant medication regimens approved by the instructions and CSCO guidelines under this stratum. Based on the complete set of parameters such as drug standard body surface area dose, route of administration, upper limit of administration concentration, suitable solvent, and safe concentration range, the system automatically filters out old medication regimens that lack guideline support or are clinically obsolete.

[0051] S2 Individualized Dosing Stratification Validation Implementation Process

[0052] The system synchronously retrieves the patient's physiological data stored in the HIS: Male, 56 years old, height 172cm, weight 68kg, clinical diagnosis "locally advanced non-small cell lung cancer", liver and kidney function indicators normal. The hospital's backend uses adult body surface area calculation, with the following formula logic (Xu Wensheng's formula):

[0053] BSA(m 2 BSA = 0.0061 × height (cm) + 0.0128 × weight (kg) - 0.1529; Substituting the data, the system will automatically calculate: BSA = 0.0061 × 172 + 0.0128 × 68 − 0.1529 = 1.0492 + 0.8704 − 0.1529 ≈ 1.767m 2 .

[0054] Target drug gemcitabine hydrochloride for injection, matched to the standard dose of non-small cell lung cancer regimen, 1000 mg / m². 2 The system automatically calculates the single dose as 1000 × 1.767 = 1767 mg.

[0055] The verification process simultaneously compares two thresholds: ① The single-dose safe range is calculated based on body surface area, 1000 mg / m². 2 ① This is the standard adult dose, 1767mg, which is a compliant calculation based on the patient's BSA; ② The maximum cumulative dose of gemcitabine per treatment course is based on the periodic dosing regimen. The single dose is calculated based on body surface area, and the cumulative dose is determined by the number of doses. The current single dose has not triggered the cumulative upper limit warning, and the dose review item is marked "Review passed".

[0056] If the patient has liver or kidney dysfunction, the system will automatically retrieve the differential dosage threshold and prompt a reduction in the upper limit of the dosage. If the drug being reviewed is one that requires weight calculation, the system will automatically switch to the weight-based dosing calculation logic: single dose = standard mg / kg dose × patient weight.

[0057] Among them, renal function: 1. HIS system data extraction: age (years), weight (kg), gender, serum creatinine (μmol / L), serum cystatin C (mg / L).

[0058] 2. Calculation formulas: Glomerular filtration rate (eGFR) was calculated using the CKD-EPI2021 formula; Creatinine clearance (CrCl) was calculated using the Cockcroft-Gault formula.

[0059] 3. The pharmacist adjusts the drug dosage based on the eGFR or CrCl values ​​calculated by the system.

[0060] Liver function: 1. HIS system data extraction: age (years), gender, aspartate aminotransferase (AST) (U / L), alanine aminotransferase (ALT) (U / L), alkaline phosphatase (ALP) (U / L), total bilirubin (TBil) (umol / L).

[0061] 1. If the measured value is greater than the following, the system will prompt the pharmacist reviewing the prescription: ALT ≥ 3 times ULN; ALP ≥ 2 times ULN; TBil ≥ 2 times ULN; AST ≥ 1.5 times ULN.

[0062] Based on the results, the pharmacist adjusting the dosage is recommended.

[0063] S3 Drug Delivery Solvent Compatibility Comprehensive Testing Implementation Process

[0064] The system retrieves the gemcitabine-specific solvent knowledge base from the drug information database and generates a whitelist of compatible solvents: 0.9% sodium chloride injection.

[0065] The system automatically verifies the solvent used in medical orders from five dimensions: compatibility stability, acid-base matching, risk of precipitation, drug degradation rate, and formation of toxic products.

[0066] In this case, the prescription solvent was set to 0.9% sodium chloride injection, which matched the whitelist, indicating no acid-base conflict or risk of degradation and precipitation. The solvent test result was marked as "approved".

[0067] S4 Dosage Concentration Compliance Verification Implementation Process

[0068] After the doctor's order was entered and the solution was reconstituted, the volume of the solvent was further diluted to 100 mL. The system automatically calculated the drug concentration = single dose ÷ solvent volume = 1767 ÷ 100 = 17.67 mg / mL, which meets both the requirements of ≤40 mg / mL and ≥0.1 mg / mL. The concentration item was approved.

[0069] If the amount of solvent used during reconstitution is insufficient (e.g., only 2 mL of solvent is added for a 200 mg dose), the calculated concentration will be >40 mg / mL, exceeding the upper limit of concentration. The system will classify the risk as too high concentration and indicate that it will cause incomplete drug dissolution and increased infusion risk. If the prescription is further diluted to a volume of 2000 mL, the calculated concentration will be <0.1 mg / mL, and the risk will be marked as too low drug concentration, which may affect the stability of efficacy.

[0070] S5 Tiered Audit Result Output and Closed-Loop Early Warning Implementation Process

[0071] In this case, all four sub-items from S1 to S4 passed the verification, the system output a Level 1 compliance result, automatically pushed the audit data to the ePIVAS system's drug waiting module, and simultaneously generated instructions for printing dispensing labels.

[0072] All audit parameters, sub-item results, warning information, and review operations are written to the audit log storage module in real time and permanently retained for querying.

[0073] II. Example 2: Complete Implementation of an Intelligent Review System for Antitumor Drug Prescriptions

[0074] This embodiment fully constructs a hardware and software integrated system consisting of three core modules: data terminal, audit engine, and application terminal. Each module is deployed independently and has bidirectional data communication, with a knowledge base update module and an audit log module to assist in operation.

[0075] Data deployment and implementation

[0076] (1) Disease diagnosis name database: Locally deployed, doctors select disease diagnoses by standardized names in the system.

[0077] (2) Drug information database: Stores all parameters of commonly used cytotoxic drugs in clinical practice, including dosage, solvent blacklist / whitelist, safe concentration, and quantitative data on adverse drug reactions;

[0078] (3) Patient information API interface: The standard HL7 protocol is used to connect to the hospital HIS system to obtain diagnosis, height and weight, liver and kidney function and informed consent documents in real time and one-way, and to transmit the review results and medical order modification records in two directions.

[0079] Audit engine deployment and implementation (system core computing unit)

[0080] The engine has built-in functional modules that execute audit logic sequentially; if the preceding module fails to verify, subsequent operations are terminated directly.

[0081] Disease diagnosis matching module: performs stratified diagnosis matching and screening of compliant solutions;

[0082] Dosage stratification calculation module: Switches between BSA / body weight dual calculation logic, supports configuration of multi-body surface area formula;

[0083] Solvent testing module: Automatically verifies five compatibility dimensions and outputs the risk level;

[0084] Concentration calculation module: Quantitatively calculates concentration and compares it with the safe range;

[0085] Risk warning module: Provides warnings and alerts for inappropriate medical orders.

[0086] It also includes two additional sub-modules: a knowledge base update and maintenance module, and an audit log recording module.

[0087] Clinical interaction implementation on the application side (pharmacist operating terminal)

[0088] Deployed on the PIVAS prescription review pharmacist workstation, the visual interactive interface includes four main functional areas:

[0089] The review results are displayed in sections: the review status of the four sub-items, namely diagnosis, dosage, solvent, and concentration, is displayed in separate columns.

[0090] Risk warning prompts are displayed in sections: abnormal medical orders are indicated by pop-up windows with risk type and quantified deviation value;

[0091] Manual review interface: Supports online modification of medical order parameters, and automatically sends the results back to the review engine for secondary review after submission;

[0092] Log query and report export section: Supports multi-condition log retrieval and one-click export of prescription review quality statistics reports.

[0093] III. Example 3: Configuration and Implementation of Multi-body Surface Area Calculation Formula

[0094] This embodiment demonstrates the implementation process of switching between multiple algorithms in the dosage calculation module, adapting to differentiated clinical standards for adults and children:

[0095] The system's backend configuration interface incorporates a native algorithm.

[0096] Xu Wensheng's formula (the default algorithm commonly used in hospitals): BSA = 0.0061 × height + 0.0128 × weight - 0.1529;

[0097] Children / newborns weighing ≥10kg: BSA(m 2 ) = (weight (kg) × 4 + 7) / (weight (kg) + 90); For children / newborns weighing < 10 kg: BSA (m 2 = 0.024265 × height (cm)^0.3964 × weight (kg)^0.5378;

[0098] Administrator partition configuration implementation: The backend is divided by age tag, with Xu Wensheng's formula used by default for adult tumors, and a special formula used for children;

[0099] Audit operation logic: The system reads the relevant information tags of the medical order, automatically retrieves the corresponding calculation formula to complete the BSA and dosage calculation, without the need for pharmacists to manually switch, and realizes the individualized calculation standard of the whole hospital for patients.

[0100] IV. Example 4: Closed-loop implementation of manual review of abnormal early warning medical orders

[0101] Using medical orders with contraindications to solvent compatibility as implementation samples, this study fully demonstrates the closed-loop process of early warning, review, and re-examination.

[0102] Basic information: The patient was diagnosed with non-small cell lung cancer and was prescribed a cytotoxic drug. The only compatible solvent for this drug is 0.9% sodium chloride injection. The prescribed solvent is 5% glucose injection.

[0103] System review process: S1 diagnostic matching and S2 dosage calculation were qualified. When performing S3 solvent testing, the system matched the blacklist of incompatible solvents and triggered a risk warning. The application pop-up window displayed the abnormal description and compatibility risk: the drug is prone to degradation and precipitation when mixed with glucose injection, which may block the infusion line and aggravate venous inflammation.

[0104] Pharmacist manual intervention: Pharmacists can view the warning details on the review interface, contact clinicians through the hospital's communication system, confirm that the solvent has been changed to 0.9% sodium chloride injection, and submit the modified medical order online;

[0105] Automatic second review: After the modification is completed, the system will automatically restart the complete review process of S1 to S4. Once all four sub-items are compliant, the system will push the process to the dispensing stage.

[0106] Log archiving: Completely record the original medical orders, risk warnings, modifications, review pharmacist accounts, and review approval times, and store them in the audit log for future reference.

[0107] V. Example 5: Implementation of Automatic Update and Maintenance of Knowledge Base

[0108] This example describes a dual-channel update implementation method for the knowledge base update module to ensure the timeliness of review rules:

[0109] Automatic timed update channel: The system connects to the hospital's existing drug instructions and clinical guidelines database or a third-party rational drug use information platform to automatically add / revise drug dosage, solvent, and concentration standards. After the update, the system's review rules take effect in real time.

[0110] Manual import channel: For medication information that has been updated but not synchronized with the system, as well as hospital-specific medication plans and hospital-registered off-label medication plans, administrators can maintain medication rules through the maintenance interface;

[0111] Update verification mechanism: Before all updated data is entered into the database, a verification node is set to remove conflicting and erroneous parameters, so as to avoid the audit logic disorder after the update.

[0112] VI. Example 6: Implementation of Audit Log Archiving and Recipient Review Quality Analysis

[0113] This example demonstrates the entire application process of the log module for data storage, retrieval, and statistics:

[0114] Data storage rules: Each medical order review generates an independent log entry, and the stored fields include review time, patient ID, diagnosis name, drug name, original medical order parameters, review result, risk warning, reviewing pharmacist, medical order modification record, and final release status;

[0115] Multi-dimensional search function: Administrators can filter by time range, clinical department, drug category, and alarm anomaly type in the backend to accurately retrieve target logs;

[0116] Automatically generate quality analysis reports: The system statistically outputs multi-dimensional data, including the proportion of various unreasonable medical orders such as solvent mismatch, abnormal dosage, and excessive concentration, the ranking of high-frequency and high-risk drugs, and the approval rate of medical orders in each department;

[0117] Clinical application implementation: PIVAS managers use reports to locate high-frequency prescription review errors in departments and conduct targeted training for physicians and pharmacists on medication review; at the same time, based on long-term statistical data, the system's review thresholds and matching rules are iteratively optimized to form a continuous improvement closed loop of "intelligent review - log recording - data analysis - rule optimization".

[0118] VII. Example 7: Integrated Implementation of the System and PIVAS Platform

[0119] This system reserves standardized data interaction interfaces, which can be embedded and integrated into the hospital's existing ePIVAS management system to connect the entire business chain from prescription review to drug dispensing:

[0120] Positive data flow: Compliant medical orders that pass all four-dimensional reviews are automatically and synchronously pushed to the PIVAS medication dispensing workstation, automatically generating a medication dispensing list and infusion label printing instructions, eliminating the need for pharmacists to manually enter medical order information a second time;

[0121] Reverse data feedback: PIVAS allocation completion and ward drug receipt information can be transmitted back to the system's log module, realizing data interoperability throughout the entire process of medical order review, allocation, and delivery;

[0122] Deployment advantages: No need to replace the hospital's existing PIVAS information system; intelligent prescription review function upgrade can be achieved simply through interface integration, significantly reducing the investment and implementation cycle of hospital information system transformation.

[0123] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art can make various equivalent modifications and improvements to the system interface, calculation formula parameters, early warning grading standards, and database storage format without departing from the core four-dimensional intelligent review concept and multi-module system architecture of the present invention. All such modifications and improvements should be included within the scope of protection defined by the claims of the present invention. The embodiments in the specification can be used individually or in any combination, all of which can achieve the core objectives of the present invention: improving the comprehensiveness of anti-tumor drug prescription review, reducing medication errors, and adapting to PIVAS clinical work.

Claims

1. A method for intelligent review of antitumor drug prescriptions, characterized in that, The following automated execution steps are included: S1 Disease Diagnosis - Standardized Matching and Retrieval Steps: Access the hospital's HIS system to obtain the patient's original clinical diagnosis text, construct a standardized disease diagnosis name database that supports localized configuration based on the original clinical diagnosis text, and perform standardized matching. After successful matching, retrieve the corresponding compliant anti-tumor drug's standardized medication parameters. The standardized medication parameters include standard single dose, daily maximum dose, dosing frequency, exclusive route of administration, upper limit of dosing concentration, and medication regimen. Automatically remove medication regimens that are off-label, outdated according to guidelines, or without instruction manual support, and retain only compliant medication regimens based on drug instructions and national oncology clinical guidelines. S2 Personalized Dosing Dosing Stratification Calculation and Verification Steps: Based on the drug dosing baseline, two types of calculation logics are automatically switched: S2-1 body surface area dosing calculation logic and S2-2 body weight dosing calculation logic. S2-1 Body Surface Area Dosing Calculation Logic: Retrieves the patient's height and weight physiological data, calculates the patient's body surface area (BSA) according to the hospital's preset formula, supports one-click configuration and switching of multiple algorithms such as Xu Wensheng's formula, and the single dose = standard body surface area dose of the drug × BSA; the calculated dose is simultaneously compared with the single safe dose range and single dose concentration limit of the disease matching plan, and the dose is judged to be unreasonable if either threshold is exceeded. S2-2 Body Weight Dosing Calculation Logic: Single Dosage = Standard Weight Dosage of Drug × Patient's Actual Weight; Differentiated safe dosage thresholds are set based on the patient's liver and kidney function test indicators, and the upper limit of the dosage is automatically lowered or a dosage adjustment is prompted for patients with liver and kidney function impairment; S3 Drug-Specific Solvent Compatibility Testing Steps: For the target anti-tumor drug matched in S1, the system retrieves the built-in whitelist of compatible solvents and blacklist of incompatible solvents from the drug database; it automatically verifies the solvents entered into the doctor's order based on compatibility stability, acid-base compatibility, precipitation risk, and drug degradation rate; only solvents on the whitelist are deemed qualified, while solvents on the blacklist are directly marked with compatibility risks, and the type of solvent conflict risk is output. S4 Dosage Concentration Compliance Range Quantitative Verification Steps: Calculate the actual dosing concentration using the individualized single-dose dose that passed S2 verification as the numerator and the total infusion volume of the prescribed solvent as the denominator. Retrieve the upper and lower limits of the legally mandated safe concentration of drugs, compare the actual concentration with the standard range, and distinguish between insufficient efficacy due to excessively low concentration and risk due to excessively high concentration; S5 Tiered Audit Result Output, Differential Early Warning and Closed-Loop Review Steps: If all four-dimensional verifications from S1 to S4 are compliant, the first-level review result will be output with the compliance level marked and directly pushed to the PIVAS drug dispensing scheduling port. If any of the S1-S4 dimensions fails to pass the verification, a risk warning will be triggered, and the abnormal dimension, the quantified deviation value, and the corresponding clinical safety risk will be marked. When the system identifies off-label drug use, it will not release the medication directly, but will prompt "It is recommended that the pharmacist verify whether there is hospital filing or informed consent from the patient". The medication can only be released after the pharmacist confirms it. The authority to block or release the medication is determined by the pharmacist during the manual review process. All audit process data, early warning records, and manual review operations are archived and retained in a unified manner. Among them, steps S1, S2, S3, and S4 are a serial linkage automated review chain. If the verification of the preceding steps is unqualified, the subsequent calculation will be terminated directly and the corresponding warning will be output, realizing the four-dimensional integrated automated review of anti-tumor drug prescription diagnosis matching, individualized dosage, solvent compatibility, and infusion concentration.

2. The intelligent review method for antitumor drug prescriptions according to claim 1, characterized in that, It also includes the S6 review log full-link archiving and quality analysis steps: automatically collecting the review time, patient basic information, original medical order parameters, four-dimensional verification sub-items results, warning type, pharmacist review operation, medical order modification record, and final release status for each medical order; supporting multi-dimensional retrieval of logs by drug type, clinical department, abnormal warning type, and time interval, generating PIVAS prescription review quality statistical reports. The reports include the proportion of various unreasonable medical orders, the ranking of high-frequency risk drugs, and the comparison data of departmental prescription review pass rates, which are used for continuous iteration and optimization of the review rule base.

3. The intelligent review method for antitumor drug prescriptions according to claim 1, characterized in that, The solvent compatibility test in step S3 pre-establishes a drug-solvent incompatibility knowledge base, which stores quantitative data on adverse reactions after different cytotoxic drugs are combined with incompatible solvents. When the system determines that the solvents are incompatible, it simultaneously retrieves the corresponding adverse reaction data and displays it along with the warning information, such as risk labels for intravenous irritation, worsening of bone marrow suppression, tubing blockage, and drug failure, to help pharmacists quickly assess the degree of risk.

4. The intelligent review method for antitumor drug prescriptions according to claim 1, characterized in that, The risk warnings in step S5 are divided into three levels: Level 1 warnings are for minor parameter deviations that can be easily adjusted and corrected by a pharmacist; Level 2 warnings are for moderate medication risks that require contacting a clinician to verify and adjust the prescription before it can be reviewed and released; Level 3 warnings are for severe safety risks that require hospital approval before release.

5. An intelligent review system for antitumor drug prescriptions, used to execute the intelligent review method according to any one of claims 1-4, characterized in that, It includes three major hardware and software integrated modules: a data terminal, an audit engine, and an application terminal, which are interconnected in sequence. The data terminal, used for multi-source clinical data storage and interfacing with external system data, includes: Disease Diagnosis Name Database: Stores diagnosis names; Drug Information Database: Stores standard dosage, compatible solvent white and black lists, safe dosing concentration range, upper limit of dosing concentration, and incompatibility knowledge base for anti-tumor drugs introduced from the PIVAS Center; Patient information API interaction interface: bidirectional connection to the hospital HIS system, real-time synchronization of patient diagnosis, height and weight, liver and kidney function test results, age and gender, and electronic records of informed consent forms; The audit engine is the core computing and processing unit of the system, with five functional modules that work in parallel: disease diagnosis matching module, dose stratification calculation module, solvent compatibility detection module, concentration verification module, and risk warning module. The application terminal is a pharmacist's visual interactive terminal, which includes a module for displaying review results, a module for risk warning pop-ups, a manual review operation interface, and a log query and quality report export interface.

6. The intelligent review system for antitumor drug prescriptions according to claim 5, characterized in that, The review engine also has a built-in knowledge base update and maintenance module. This module is configured with a scheduled automatic update interface and a manual import channel. The scheduled automatic update interface connects to the hospital's existing drug instruction manual and clinical guideline database or a third-party rational drug use information platform. By acquiring authoritative materials such as drug instruction manuals and oncology clinical guidelines, it enables automatic updates of drug review rules. The manual import channel supports the maintenance and updating of medication plans and in-house self-made medication plans. After the update, the entire system takes effect in real time, ensuring the timeliness and adaptability of the review rules within the hospital.

7. The intelligent review system for antitumor drug prescriptions according to claim 5, characterized in that, The dose stratification calculation module has a built-in multi-body surface area calculation formula switching unit, stores standard algorithms such as Xu Wensheng's formula, and also reserves a custom formula editing input port; Hospital administrators can select the default calculation algorithm in the system backend, or configure independent BSA calculation formulas for pediatrics and adults respectively. The module will automatically retrieve the corresponding formula based on the patient information to complete the calculation of body surface area and drug dosage.

8. The intelligent review system for antitumor drug prescriptions according to claim 5, characterized in that, The manual review interface of the application has the function of online modification and verification of medical orders; after the pharmacist completes the parameter correction and submits the warning medical order, the system automatically sends the modified medical order back to the review engine to re-execute the S1-S4 full four-dimensional automated review, and after the second verification is completed, the review results and logs are updated.