A management system and application of DTP pharmacy

Abstract

The present invention relates to the field of medical health. The present invention discloses a management system and application of a DTP pharmacy, which includes a condition input and prescription module, which is used for doctors to judge the patient's condition on site and input it into the management system and match relevant drugs, a prescription selection and prescription module, which is used for doctors to select and search for relevant drugs and prescribe for patients on the management system, and send it to the DTP pharmacy, a drug conflict detection module, which is used for the system to automatically detect whether there is a conflict between the prescribed multiple drugs, assign weights according to the importance of the drugs, and decide whether to continue to use these drugs, and a pharmacy information sharing module, which is used to transmit the electronic prescription issued by the doctor to the DTP pharmacy, and the DTP pharmacy dispenses drugs. The management system of the DTP pharmacy of the present application can enable the system to automatically detect drug conflicts and improve the safety of drug use. Patients can directly pick up drugs at the DTP pharmacy with electronic prescriptions, simplifying the medical treatment process.

Description

Technical Field

[0001] The present invention relates to the field of medical health, and in particular to a management system and application of a DTP pharmacy. Background Art

[0002] DTP (Direct-to-Patient) pharmacy is a pharmacy model that provides special drugs and professional services directly to patients. This model usually establishes cooperative relationships with hospitals, doctors and drug manufacturers to ensure that patients can obtain high-quality drugs and personalized medication guidance. DTP pharmacies have significant advantages in improving patient medication compliance and optimizing drug supply chain management.

[0003] At present, the pharmacy system faces multiple challenges in the field of healthcare. These problems not only affect patients' medication experience, but also bring difficulties to drug sales and supply chain management. Different hospitals and different doctors may give different diagnosis results and prescriptions when diagnosing the same disease. These differences cause patients to feel confused when choosing drugs and find it difficult to determine which drug is most suitable for their condition. Due to the relative independence of information, it is difficult for patients to obtain comprehensive medical advice, which affects the treatment effect. Patients have a vague understanding of drug selection, administration method and efficacy. Patients who lack professional knowledge often find it difficult to understand doctors' prescriptions and medication instructions, and are prone to improper medication. This lack of cognition not only affects the treatment effect, but may also lead to drug waste and medical resources. irrational use; when treating multiple diseases, patients often seek different doctors, and these different doctors may prescribe multiple drugs, which may interact with each other, leading to adverse reactions or reducing drug efficacy; the lack of an effective drug conflict detection mechanism makes patients face risks when using multiple drugs; patients need to spend a lot of time queuing up at the hospital pharmacy window to get medicine, which increases the workload of pharmacy staff; long waiting times not only affect patients' medical experience, but may also lead to low pharmacy work efficiency; inaccurate drug sales data and information island problems make it difficult for sales teams to make accurate market decisions; the lack of effective sales monitoring and analysis tools makes drug inventory management difficult and affects the stability of the supply chain. Summary of the invention

[0004] In order to solve the deficiencies of the prior art, the present invention discloses a DTP pharmacy management system that can avoid strong drug conflicts when doctors prescribe drugs, facilitate doctors to prescribe and guide medication for patients, and improve patients' efficiency in obtaining drugs.

[0005] The invention discloses a management system for a DTP pharmacy, which comprises a condition input and prescription module, which is used for doctors to judge the condition of patients on site and input the condition into the management system, the management system matches the medicines related to the condition of patients, a prescription selection and prescription module, which is used for doctors to select and search for related medicines and prescribe for patients on the management system, and send the prescription information to related DTP pharmacies, a medicine conflict detection module, which is used for the system to automatically detect whether there are conflicts between multiple prescribed medicines, give weights according to the importance of the medicines, and decide whether to continue to use these medicines, and a pharmacy information sharing module, which is used for transmitting the electronic prescriptions issued by doctors to the cooperating DTP pharmacies through a secure channel, the DTP pharmacies receive the electronic prescriptions and prepare the medicines, and the DTP pharmacies dispense the medicines according to the patient's medicine collection voucher and in accordance with the electronic prescriptions.

[0006] Furthermore, the doctor inputs the condition and prescribes the medicine module, including an input submodule and an output submodule. The input submodule is used by the doctor to input the patient's condition and symptoms, and the output submodule is used by the system to provide diagnosis suggestions and medicine selection according to the doctor's input of the patient's condition and symptoms. Intelligent diagnosis and medicine recommendation find the shortest path P shortest , from P to the diagnosis node D: ,P: patient's condition and symptoms,D: diagnosis node,P(P,D): the set of all paths from patient input P to diagnosis node D,P′: a path in the path set,(u,v): an edge in path P′,u: a node in the path, representing a symptom, disease or drug,v: another node in the path, representing another symptom, disease or drug,w(u,v): the weight of edge (u,v), representing the strength of the relationship between node u and node v,P shortest : The shortest path from P to D, used to recommend the most appropriate diagnosis and medicine.

[0007] Furthermore, the prescription selection and issuance module includes a prescription selection submodule and a prescription issuance submodule. The prescription selection submodule is used by doctors to select medicines and can input multiple interchangeable medicines of the same category. The prescription issuance submodule is used by the system to generate electronic prescriptions and send them to the cooperating DTP pharmacies.

[0008] Furthermore, in the drug conflict detection module, the system checks the interaction between each pair of drugs and calculates the conflict score C(M i , M j ), M i and M j : Two drugs were prescribed, w i and w j :Drug M i and M jThe weight indicates the importance of the drug. The main drug has a higher weight and the auxiliary drug has a lower weight. The weight is given according to the doctor's advice and past experience. i , M j ): Drug M i and M j The conflict coefficient between them indicates the degree of adverse reactions when these two drugs are used simultaneously, C(M i , M j ): Drug M i and M j The conflict scores between them are used to calculate the total conflict scores C of all drug pairs. total : , C total : The total conflict score of all drug pairs, conflict threshold: The conflict threshold is T, C total ≤T, the drug combination can continue to be used, C total >T, prompts the doctor to re-select or carefully select the drug, T: preset conflict threshold, indicating the maximum acceptable total conflict score.

[0009] Furthermore, the drug sales monitoring and analysis module is used to input drug information and prices. Among the multiple drugs prescribed by doctors, the system gives priority to recommending replaceable drugs with the same efficacy from the sales company based on the price fluctuation range. The time series prediction module is used to predict the drug sales trend in the future. The sales data analysis module is used to discover sales patterns and potential market opportunities through cluster analysis and association rule learning. The anomaly detection module is used to monitor sales data in real time and detect abnormal sales fluctuations.

[0010] Furthermore, the drug sales monitoring and analysis module includes a sales information input submodule and a price range recommendation submodule. The sales information input submodule is used for the sales company to input drug information and prices, which are stored by the system. The price range recommendation submodule is used for the system to give priority to recommending the sales company's replaceable drugs with the same efficacy according to the price fluctuation range among the multiple drugs recommended by the doctor. , Input: Multiple drugs recommended by doctors and their prices, Output: Prioritize the sales company drugs whose price fluctuation does not exceed a% of the drugs recommended by doctors, Formula: P s : The price of the company's drug, P d : The price of the drug recommended by the doctor, a: The preset price fluctuation percentage, set by the sales company. If the price fluctuation of the sales company's drugs does not exceed a% of the drugs recommended by the doctor, the sales company's drugs will be recommended first.

[0011] Furthermore, the time series prediction module is used to predict the drug sales trend in the future based on the time series prediction algorithm of the long short-term memory network. , , x t : Historical sales data at time t, as input features, y t : The predicted value at time t, indicating the drug sales volume at a certain point in the future, W x : Input weight matrix, used to transform the input features into part of the hidden state, W h : Hidden state weight matrix, used to convert the hidden state of the previous moment into the hidden state of the current moment, W hx : The weight matrix from hidden state to output, used to convert the hidden state to the final predicted value, h t−1 : The hidden state of the previous moment, indicating the memory of the model at the previous moment, b h : bias term, used to adjust the output of the model, σ: activation function, used to introduce nonlinear characteristics so that the model can capture complex patterns.

[0012] Furthermore, the sales data analysis module uses the K-means clustering algorithm to discover sales patterns and potential market opportunities: , C: a set of clusters, representing the number of groups into which sales records are divided, k: the number of clusters, i.e. the number of groups, C i : The i-th cluster represents a group, x j : a data point in the sales record, μ i : The center of the i-th cluster, representing the average value of the group, ∥x j −μ i ∥ 2 : Data point x j With cluster center μ i The square of the Euclidean distance between them is used to measure the proximity of a data point to the cluster center. Association rule learning is used to calculate support and confidence through the Apriori algorithm, discover frequent item sets and association rules, and help the sales team discover potential market opportunities and sales models: , X and Y: item sets, representing a group of drugs or sales records, N: total number of records, representing the total number of sales records, count(X): the number of times item set X appears, representing the number of sales records containing X, support(X): the support of item set X, representing the frequency of X appearing in all sales records, support(X∪Y): the support of item sets X and Y appearing at the same time, representing the number of sales records containing X and Y, confidence(X→Y): the confidence from item set X to item set Y, representing the proportion of sales records containing X that also contain Y.

[0013] Furthermore, the anomaly detection module is used to monitor sales data in real time, detect abnormal sales fluctuations, and convert sales data into dimensionless standardized values ​​through standardization processing: , Si : Sales volume on the i-th day, μ S : The average value of sales volume, σ S : Standard deviation of sales volume, z i : The degree of deviation between the sales volume on the i-th day and the average sales volume, ,θ: preset threshold, indicating the maximum acceptable standardized deviation, |z i ∣>θ, assuming that the current sales volume is x t Deviations from historical averages that exceed preset thresholds trigger alerts, alerting the sales team to potential market changes or supply chain issues.

[0014] The present invention discloses an application of a DTP pharmacy, which comprises the application of any one of the above-mentioned DTP pharmacy management systems in the DTP pharmacy.

[0015] Beneficial effects of the present invention:

[0016] The management system of the DTP pharmacy of this application can make the entire process from seeing a doctor to picking up medicine more efficient, reduce waiting time, and improve patient satisfaction. Through automation and information technology, the allocation of medical resources is optimized, the overall efficiency of medical services is improved, and it is convenient for doctors to prescribe and guide patients on medication. The system automatically detects drug conflicts, discovers potential adverse reactions in advance, and improves the safety of medication. Patients can pick up medicines directly at the DTP pharmacy with electronic prescriptions without carrying paper prescriptions, which simplifies the medical process. The management system of the DTP pharmacy not only improves the efficiency and safety of medical services, but also improves the patient's medical experience and optimizes the management of the pharmacy.

[0017] Through the drug sales monitoring and analysis module, time series prediction module, sales data analysis module and anomaly detection module, the DTP pharmacy management system can recommend lower-priced but equally effective alternative drugs to reduce patients' medical expenses. By predicting future sales trends, inventory management can be optimized to reduce inventory backlogs and out-of-stock phenomena. Through cluster analysis and association rule learning, new market opportunities and associations between drugs can be discovered to develop more effective sales strategies. Through real-time monitoring and early warning mechanisms, abnormal sales fluctuations can be detected and responded to in a timely manner, losses can be reduced, and management efficiency can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of the management system of a DTP pharmacy in an embodiment of the present application.

[0019] Figure 2 Another schematic diagram of the management system of a DTP pharmacy in an embodiment of the present application.

[0020] Figure 3Another schematic diagram of the management system of a DTP pharmacy in an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the specific implementation modes of the present invention will be clearly and completely described below.

[0022] The present application discloses a management system for a DTP pharmacy, which includes a condition input and prescription module, which is used for doctors to judge the patient's condition on site and input it into the management system, and the management system matches the relevant drugs for the patient's condition; the doctor can quickly input the patient's condition information, and the system automatically matches the relevant drugs, which reduces the time for doctors to manually search for drugs and improves the efficiency of diagnosis and treatment. The drug recommendations provided by the system are based on a large amount of clinical data and professional knowledge, which helps to reduce the misdiagnosis caused by doctors' memory errors or insufficient information. The system can provide personalized drug recommendations based on the patient's specific condition and medical history to ensure that the treatment plan is more accurate. The prescription selection and prescription module is used for doctors to select and search for relevant drugs and prescribe for patients on the management system, and send the prescription information to the relevant DTP pharmacy; the doctor can select and search for drugs through the system, reduce errors when writing prescriptions manually, and improve the accuracy and standardization of prescriptions. The use of electronic prescriptions simplifies the doctor's workflow and reduces the management and storage costs of paper prescriptions. Patients can directly pick up drugs at the DTP pharmacy with electronic prescriptions without carrying paper prescriptions, which improves the patient's medical experience. The drug conflict detection module is used for the system to automatically detect whether there is a conflict between the multiple drugs prescribed, assign weights according to the importance of the drugs, and decide whether to continue to use these drugs; the system automatically detects drug conflicts, discovers potential drug interactions in advance, reduces adverse reactions caused by drug conflicts, and improves drug safety. The system provides scientific decision support based on the importance of the drugs and the specific conditions of the patients, helping doctors make more reasonable drug choices. By timely detecting and prompting drug conflicts, the occurrence of medical accidents is reduced and medical risks are reduced. The pharmacy information sharing module is used to transmit the electronic prescriptions issued by doctors to the cooperating DTP pharmacies through a secure channel. The DTP pharmacy receives the electronic prescriptions and prepares the drugs. The DTP pharmacy dispenses the drugs according to the patient's drug collection voucher and the corresponding electronic prescription. The electronic prescription has a fast transmission speed, and the DTP pharmacy can quickly receive the prescription information, speeding up the preparation and delivery of drugs. The electronic prescription is transmitted through a secure channel to ensure the integrity and security of the prescription information and prevent information leakage and tampering. The management system of the DTP pharmacy of this application can make the entire process from patient consultation to drug collection more efficient, reduce waiting time, and improve patient satisfaction. Through automation and information technology, the allocation of medical resources is optimized and the overall efficiency of medical services is improved. The system automatically detects drug conflicts, discovers potential adverse reactions in advance, and improves the safety of medication. Patients can directly pick up medicines at DTP pharmacies with electronic prescriptions, without having to carry paper prescriptions, which simplifies the medical process. The management system of DTP pharmacies not only improves the efficiency and safety of medical services, but also improves the patient's medical experience and optimizes the management of pharmacies. This system provides comprehensive support for DTP pharmacies and helps to improve the overall level of medical services.

[0023] As a specific implementation, the doctor inputs the condition and prescribes medicine module, including an input submodule and an output submodule, the input submodule is used by the doctor to input the patient's condition and symptoms, and the output submodule is used by the system to provide diagnosis suggestions and medicine selection according to the doctor's input of the patient's condition and symptoms;

[0024] Intelligent diagnosis and drug recommendation to find the shortest path P shortest , from P to the diagnosis node D:

[0025] ;

[0026] P: patient's condition and symptoms, for example, the patient inputs "headache" and "fever";

[0027] D: Diagnosis node, for example, outputs "flu", "cold";

[0028] P(P, D): the set of all paths from patient input P to diagnosis node D, for example, all possible paths from "headache" and "fever" to "flu";

[0029] P′: a path in the path set, for example, a possible path is from “headache” to “fever” to “flu”;

[0030] (u, v): an edge in the path P′;

[0031] u: a node in the path, representing a symptom, disease or drug;

[0032] v: another node in the path, representing another symptom, disease or drug;

[0033] w(u, v): the weight of edge (u, v), which indicates the strength of the relationship between node u and node v;

[0034] P shortest : The shortest path from P to D, used to recommend the most appropriate diagnosis and medicine. Example

[0035] Assuming that the symptoms entered by the patient are "headache" and "fever", the system workflow is as follows:

[0036] The doctor enters the patient's symptoms: "headache" and "fever".

[0037] Construct a path set P(P, D). Based on the input condition and symptoms, the system constructs a set of all paths from the patient input P to the diagnosis node D. The system evaluates the cost of each path and calculates the sum of the weights of all edges in the path.

[0038] From “headache” to “fever” to “flu”;

[0039] From "headache" to "fever" to "cold";

[0040] From "headache" to "fever" to "heat stroke";

[0041] Evaluate the path cost:

[0042] The cost of going from "headache" to "fever" to "flu" is 0.1;

[0043] The cost of going from "headache" to "fever" to "cold" is 0.6;

[0044] The cost from "headache" to "fever" to "heat stroke" is 0.7;

[0045] Path 1: From "headache" to "fever" to "flu";

[0046] Edge 1: (headache, fever), weight w(headache, fever)=0.8w;

[0047] Edge 2: (fever, flu), weight w(fever, flu)=0.7w;

[0048] Path cost: 0.8+0.1=0.9;

[0049] Path 2: From "headache" to "fever" to "cold"

[0050] Edge 1: (headache, fever), weight w(headache, fever)=0.8w;

[0051] Edge 2: (fever, cold), weight w(fever, cold) = 0.6w;

[0052] Path cost: 0.8+0.6=1.4;

[0053] Path 3: From "headache" to "fever" to "heat stroke";

[0054] Edge 1: (headache, fever), weight w(headache, fever)=0.8w;

[0055] Edge 2: (fever, heat stroke), weight w(fever, heat stroke) = 0.1w;

[0056] Path cost: 0.8+0.7=1.5;

[0057] Find the shortest path P shortest : Path 1: From "headache" to "fever" to "flu";

[0058] From “headache” to “fever” to “flu”;

[0059] Diagnostic suggestions: the system provides diagnostic suggestions: "influenza";

[0060] Drug recommendation, system recommended drug: Oseltamivir.

[0061] As a specific implementation method, the prescription selection and prescribing module includes a prescription selection submodule and a prescription prescribing submodule. The prescription selection submodule is used for doctors to select drugs and can input multiple interchangeable drugs of the same category. Doctors can select multiple different interchangeable drugs from the same category of prescription drugs for patients to choose according to their economic conditions. The prescription prescribing submodule is used for the system to generate electronic prescriptions and send them to the cooperating DTP pharmacies. Provide a variety of interchangeable drugs, and patients can choose the most suitable drugs according to their economic conditions and preferences. Reduce errors in manual prescription writing and improve the accuracy and standardization of prescriptions. Provide convenient ways to pick up drugs and transparent drug selection to improve patients' medical experience. Improve drug distribution efficiency, optimize inventory management, and reduce drug backlogs and out-of-stock phenomena.

[0062] As a specific implementation, the drug conflict detection module system checks the interaction between each pair of drugs and calculates the conflict score C(M i , M j );

[0063]

[0064] M i and M j : Two medications prescribed;

[0065] w i and w j :Drug M i and M j The weight indicates the importance of the drug, with primary drugs having a higher weight and auxiliary drugs having a lower weight, which are assigned based on the doctor's advice and past experience;

[0066] c(M i , M j ): Drug M i and M j The conflict coefficient between them indicates the degree of adverse reactions when these two drugs are used simultaneously;

[0067] C(M i , M j ): Drug M i and M j The conflict score between

[0068] Calculate the total conflict score C of all drug pairs total :

[0069] ;

[0070] Ctotal : The total conflict score of all drug pairs;

[0071] Conflict threshold: The conflict threshold is T, C total ≤T, the drug combination can continue to be used, C total >T, prompting doctors to re-select or carefully choose drugs;

[0072] T: The preset conflict threshold, which indicates the maximum acceptable total conflict score.

[0073] In practical applications, the setting of the threshold T is usually the result of comprehensive consideration of the above-mentioned factors. For example, referring to clinical guidelines and standards, first refer to the guidelines of national or international drug regulatory agencies to obtain the recommended conflict threshold. Combined with clinical experience, according to the clinical experience of doctors and pharmacists, adjust the recommended threshold to make it more in line with the actual situation. Consider the specific situation of the patient, and adjust the threshold appropriately according to the health status and medical history of the specific patient. Drug property analysis, according to the half-life, metabolic pathway and dosage of the drug, further optimize the threshold. The above-mentioned data related to the conflict threshold T can be directly input into the management system of the DTP pharmacy, and the system calculates the appropriate conflict threshold T based on the relevant data when in use. Example

[0074] Suppose a doctor prescribes three drugs for a patient:

[0075] Drug A (primary drug);

[0076] Drug B (adjuvant drug);

[0077] Drug C (adjuvant drug);

[0078] Weights and conflict coefficients:

[0079] Weight:

[0080] w A =0.8 (primary drug);

[0081] w B =0.3 (adjuvant medication);

[0082] w C =0.2 (adjuvant medication);

[0083] Conflict coefficient (hypothetical conflict coefficient value):

[0084] c(A, B)=0.5;

[0085] c(A, C)=0.3;

[0086] c(B, C)=0.4;

[0087] Conflict score calculation:

[0088] Calculate the conflict score C(M) between each pair of drugs i , M j ):

[0089] C(A,B)=w A ·w B ·c(A,B)=0.8·0.3·0.5=0.12;

[0090] C(A,C)=w A ·w C ·c(A,C)=0.8·0.2·0.3=0.048;

[0091] C(B,C)=w B ·w C ·c(B,C)=0.3·0.2·0.4=0.02;

[0092] Calculate the total conflict score C of all drug pairs total :

[0093] C total =C(A,B)+C(A,C)+C(B,C)=0.12+0.048+0.024=0.192;

[0094] Set the conflict threshold T:

[0095] Assume that the preset conflict threshold T=0.2;

[0096] Determine whether the drug combination can continue to be used:

[0097] If C total ≤T, the drug combination can continue to be used;

[0098] If C total >T, it prompts the doctor to re-select or carefully select the drug.

[0099] Example results:

[0100] Total conflict score C total =0.192;

[0101] The preset conflict threshold T=0.2;

[0102] Due to C total =0.192≤0.2, so the drug combination can continue to be used.

[0103] As a specific implementation method, the management system of the DTP pharmacy also includes: a drug sales monitoring and analysis module, which is used to input drug information and prices, and among the multiple drugs prescribed by the doctor, the system gives priority to recommending replaceable drugs with the same efficacy from the sales company according to the price fluctuation range. When the doctor prescribes the drug, the system can input drug information and prices. The system can reduce the patient's medical expenses and improve the patient's economic affordability by recommending replaceable drugs with lower prices but equivalent efficacy according to the price fluctuation range. The sales company can increase the sales opportunities of drugs and increase market share by recommending its own replaceable drugs. Patients can choose drugs with higher cost performance and improve their satisfaction with treatment. A time series prediction module is used to predict the sales trend of drugs in the future. By predicting future sales trends, the DTP pharmacy can optimize inventory management, reduce inventory backlogs and out-of-stock phenomena, and improve operational efficiency. The pharmacy can prepare stocks in advance according to the prediction results to ensure sufficient supply of drugs to meet patient needs. A sales data analysis module is used to discover sales patterns and potential market opportunities through cluster analysis and association rule learning. Through cluster analysis, pharmacies can discover new market opportunities and develop new drug combinations to meet the needs of different patients. Through association rule learning, pharmacies can discover the associations between drugs, develop more effective joint sales strategies, and increase sales. Anomaly detection module is used to monitor sales data in real time and detect abnormal sales fluctuations. Through real-time monitoring and early warning mechanisms, pharmacies can promptly discover and respond to abnormal sales fluctuations to reduce losses. By promptly discovering and handling abnormal sales fluctuations, pharmacies can prevent potential risks and ensure the stable operation of the business. Through the drug sales monitoring and analysis module, time series prediction module, sales data analysis module, and anomaly detection module, the DTP pharmacy management system can recommend alternative drugs with lower prices but equivalent efficacy to reduce patients' medical expenses. By predicting future sales trends, inventory management is optimized to reduce inventory backlogs and out-of-stock phenomena. Through cluster analysis and association rule learning, new market opportunities and associations between drugs are discovered, and more effective sales strategies are developed. Through real-time monitoring and early warning mechanisms, abnormal sales fluctuations are promptly discovered and responded to, losses are reduced, and management efficiency is improved.

[0104] As a specific implementation method, the drug sales monitoring and analysis module includes a sales information input submodule and a price range recommendation submodule. The sales information input submodule is used for the sales company to input drug information and prices, and the system stores them. The price range recommendation submodule is used for the system to give priority to recommending the sales company's replaceable drugs with the same efficacy according to the price fluctuation range among multiple drugs recommended by doctors.

[0105] ;

[0106] Input: Multiple drugs recommended by doctors and their prices;

[0107] Output: Prioritize the recommendation of drugs from sales companies whose price fluctuation does not exceed a% of the drugs recommended by doctors;

[0108] formula:

[0109] P s : The price of the company's drugs;

[0110] P d : Price of medicines recommended by doctors;

[0111] a: The preset price fluctuation percentage, set by the sales company;

[0112] If the price fluctuation of the sales company's drugs does not exceed a% of the doctor's recommended drugs, the sales company's drugs will be recommended first.

[0113] The present invention allows patients to choose drugs with lower prices but equivalent efficacy, thereby reducing medical expenses and alleviating economic burdens by recommending drugs from sales companies whose price fluctuations do not exceed a% of the drugs recommended by doctors. For patients with limited economic conditions, more economical alternative drugs can be selected to improve the accessibility of drugs. Sales companies can increase sales opportunities and market share of drugs by recommending their own replaceable drugs. By providing reasonably priced replaceable drugs, sales companies can strengthen their cooperative relationship with medical institutions and establish good cooperative relationships. Patients can choose the most suitable drugs according to their economic situation and preferences, thereby improving their satisfaction with treatment. By providing reasonably priced replaceable drugs with equivalent efficacy, patients' trust in medical services will be enhanced, and their satisfaction with hospitals and pharmacies will be improved. By recommending reasonably priced replaceable drugs, pharmacies can optimize inventory management, reduce inventory backlogs and out-of-stock phenomena, and improve operational efficiency. Example

[0114] A doctor recommended three antihypertensive drugs to a patient with hypertension. Among these recommended drugs, the sales company hopes to give priority to the sales company's drugs whose price fluctuation does not exceed 10% of the drugs recommended by the doctor.

[0115] Various drugs recommended by doctors and their prices:

[0116] Drug A: Amlodipine, price: 100 yuan;

[0117] Drug B: Benazepril, price 80 yuan;

[0118] Drug C: Hydrochlorothiazide, price 50 yuan;

[0119] Drugs sold by the company and their prices:

[0120] Drug D: Amlodipine, price: 90 yuan;

[0121] Drug E: Benazepril, price: 70 yuan;

[0122] Drug F: Hydrochlorothiazide, price: 45 yuan;

[0123] The preset price fluctuation percentage a=10%;

[0124] Preferred drugs:

[0125] Drug D: Price is 90 yuan, price fluctuation is 10%, no more than 10%, priority recommendation.

[0126] Drug E: Price is 70 yuan, and the price fluctuation is 12.5%. If it exceeds 10%, it is not recommended.

[0127] Drug F: Price is 45 yuan, price fluctuation is 10%, no more than 10%, priority recommendation.

[0128] As an implementation method, the time series prediction module is used to predict the drug sales trend in the future period of time based on the time series prediction algorithm of the long short-term memory network;

[0129] ;

[0130] ;

[0131] x t : Historical sales data at time t as input features;

[0132] y t : The predicted value at time t, indicating the drug sales volume at a certain point in the future;

[0133] W x : Input weight matrix, used to transform input features into part of hidden state;

[0134] W h : Hidden state weight matrix, used to convert the hidden state of the previous moment into the hidden state of the current moment;

[0135] W hx : The weight matrix from hidden state to output, used to convert the hidden state into the final predicted value;

[0136] h t−1 : The hidden state of the previous moment, indicating the memory of the model at the previous moment;

[0137] b h : Bias term, used to adjust the output of the model;

[0138] σ: Activation function, used to introduce nonlinear characteristics so that the model can capture complex patterns. Example

[0139] The time series forecasting module is used to predict the sales trend of drugs in the future, helping the sales team make more accurate market decisions. The drug sales dataset records the daily sales volume in the past few months, and the goal is to predict the sales volume in the next week.

[0140] Dataset: Sales data for the past few days: 100, 105, 110, 115, 120, 125, 130, 135, 140, 145, 150, 155, 160, 165, 170.

[0141] ;

[0142] ;

[0143] y t : The predicted value at time t, indicating the drug sales volume at a certain point in the future.

[0144] For example, forecast sales volume for the next 5 days to be 100 units.

[0145] x t : Historical sales data at time t as input features.

[0146] For example, the sales data for the past few days may be [100, 105, 110, 115, 120].

[0147] Weight matrix W x :The input feature x t Transformed into part of the hidden state.

[0148] Sales data for the past 5 days [100, 105, 110, 115, 120], weight matrix W x This data will be converted into an intermediate representation, which is easier for the model to understand and process.

[0149] W x ·x t The sales data for the past few days will be converted into a vector that contains the important features of the data.

[0150] W h : Hidden state weight matrix, used to transform the hidden state h of the previous moment t−1 Convert to the hidden state at the current moment.

[0151] h t−1 : The hidden state of the previous moment, indicating the memory of the model at the previous moment.

[0152] Assume that the hidden state h at the previous moment t−1 =[0.1, 0.2, 0.3], this hidden state records the model’s understanding and state of the sales data at the previous moment.

[0153] W hx : The hidden state to output weight matrix used to transform the hidden state into the final predicted value.

[0154] Assume that the hidden state h at the previous moment t−1 =[0.1, 0.2, 0.3], hidden state weight matrix W h The memory of the current moment will be updated based on the memory of the previous moment.

[0155] Assume that the hidden state at the current moment is [0.4, 0.5, 0.6], the weight matrix from hidden state to output is W hx This hidden state will be converted into the final sales prediction.

[0156] b h : Bias term, used to adjust the output of the model;

[0157] Assume that the bias term b h =0.5, this bias term will add a small correction value to the final prediction value to ensure a more accurate prediction.

[0158] σ: Activation function, used to introduce nonlinear characteristics so that the model can capture complex patterns.

[0159] The input value is transformed into a nonlinear form.

[0160] Enter historical sales data:

[0161] Assume that the past sales data is x t =[100, 105, 110, 115, 120].

[0162] Initialize the weight matrix and bias terms:

[0163] Input weight matrix W x is a 5×3 matrix.

[0164] Hidden state weight matrix W h is a 3×3 matrix.

[0165] The weight matrix W from hidden state to output hx is a 3×1 matrix.

[0166] Bias term b h is a scalar value, b h is 0.5.

[0167] Compute the hidden state:

[0168] Assume that the hidden state h at the previous moment t−1 =[0.1, 0.2, 0.3].

[0169] Compute the linear transformation of the input features: W x ·x t .

[0170] Calculate the linear transformation of the hidden state at the previous moment: W hx ·h t−1 .

[0171] Add the two together and apply the activation function .

[0172] Calculate the final prediction value:

[0173] Assume W x for:

[0174]

[0175] W x ·x t for:

[0176]

[0177]

[0178]

[0179]

[0180] Calculate the linear transformation W of the hidden state at the previous moment h ·h t−1 :

[0181] Assume W h for:

[0182] ;

[0183] Calculate W h ·h t−1 :

[0184]

[0185] Calculate hidden state :

[0186] Assume that the activation function σ is the ReLU function, that is, σ(x)=max(0,x).

[0187] calculate :

[0188] ;

[0189] Apply the ReLU activation function:

[0190] ;

[0191] Calculate the final prediction value :

[0192] ;

[0193] Final prediction value .

[0194] As a specific implementation method, the sales data analysis module uses the K-means clustering algorithm to discover sales patterns and potential market opportunities:

[0195] ;

[0196] C: A set of clusters, which represents the groups into which sales records are divided;

[0197] k: the number of clusters, that is, the number of groups;

[0198] C i : The i-th cluster, representing a group;

[0199] x j : A data point in a sales record;

[0200] μ i : The center of the i-th cluster, indicating the average value of the group;

[0201] ∥x j -μ i ∥ 2 : Data point x j With cluster center μ i The squared Euclidean distance between them is used to measure the proximity of a data point to the cluster center;

[0202] Association rule learning is used to calculate support and confidence through the Apriori algorithm, discover frequent item sets and association rules, and help sales teams discover potential market opportunities and sales models:

[0203] ;

[0204] X and Y: item sets, representing a set of drugs or sales records;

[0205] N: total number of records, indicating the total number of sales records;

[0206] count(X): the number of times item set X appears, indicating the number of sales records containing X;

[0207] support(X): the support of item set X, which indicates the frequency of X appearing in all sales records;

[0208] support(X∪Y): the support of item sets X and Y appearing at the same time, indicating the number of sales records containing X and Y;

[0209] confidence(X→Y): The confidence from item set X to item set Y, which indicates the proportion of sales records containing X that also contain Y.

[0210] As a specific implementation method, the anomaly detection module is used to monitor sales data in real time, detect abnormal sales fluctuations, and convert sales data into dimensionless standardized values ​​through standardization processing:

[0211] ;

[0212] S i : sales volume on the i-th day;

[0213] μ S : The average value of sales volume;

[0214] σ S : Standard deviation of sales volume;

[0215] z i : The degree of deviation between the sales volume on the i-th day and the average sales volume;

[0216] ;

[0217] θ: preset threshold, indicating the maximum acceptable standardized deviation;

[0218] ∣z i ∣>θ, it is considered that the deviation of the current sales volume xt from the historical average exceeds the preset threshold, triggering an alarm to prompt the sales team to pay attention to potential market changes or supply chain problems.

[0219] The present invention also discloses an application of a DTP pharmacy, including the application of any one of the above-mentioned DTP pharmacy management systems in the DTP pharmacy.

[0220] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the protection scope of the appended claims of the present invention.

Claims

1. A DTP pharmacy management system, characterized in that: include: The condition input and prescription module is used by doctors to judge the patient's condition on site and input it into the management system, which then matches the relevant medicines to the patient's condition. Prescription selection and issuance module, which is used by doctors to select and search for relevant drugs and issue prescriptions for patients on the management system, and send prescription information to relevant DTP pharmacies; The drug conflict detection module is used by the system to automatically detect whether there are conflicts between the multiple drugs prescribed, assign weights based on the importance of the drugs, and decide whether to continue using these drugs; The pharmacy information sharing module is used to transmit the electronic prescription issued by the doctor to the cooperating DTP pharmacy through a secure channel. The DTP pharmacy receives the electronic prescription and prepares the medicine. The DTP pharmacy dispenses the medicine according to the patient's medication voucher and the corresponding electronic prescription; In the drug conflict detection module, the system checks the interaction between each pair of drugs and calculates the conflict score C(M i , M j ); ； M i and M j : Two medications prescribed; w i and w j :Drug M i and M j The weight indicates the importance of the drug. The main drug has a higher weight and the auxiliary drug has a lower weight. The weight is given according to the doctor's advice and previous experience; c(M i , M j ): Drug M i and M j The conflict coefficient between them indicates the degree of adverse reactions when these two drugs are used simultaneously; C(M i , M j ): Drug M i and M j The conflict score between Calculate the total conflict score C of all drug pairs total : ； C total : The total conflict score of all drug pairs; Conflict threshold: The conflict threshold is T, C total ≤T, the drug combination can continue to be used, C total >T, prompting doctors to re-select or carefully choose drugs; T: The preset conflict threshold, which indicates the maximum acceptable total conflict score.

2. A DTP pharmacy management system according to claim 1, characterized in that: The doctor inputs the condition and prescribes medicine module, including an input submodule and an output submodule. The input submodule is used by the doctor to input the patient's condition and symptoms, and the output submodule is used by the system to provide diagnosis suggestions and medicine selection according to the doctor's input of the patient's condition and symptoms; Intelligent diagnosis and drug recommendation to find the shortest path P shortest , from P to the diagnosis node D: ； P: patient's condition and symptoms; D: Diagnosis node; P(P, D): the set of all paths from the patient input P to the diagnosis node D; P′: a path in the path set; (u, v): an edge in the path P′; u: a node in the path, representing a symptom, disease or drug; v: another node in the path, representing another symptom, disease or drug; w(u, v): the weight of edge (u, v), which indicates the strength of the relationship between node u and node v; P shortest : The shortest path from P to D, used to recommend the most appropriate diagnosis and medicine.

3. According to the DTP pharmacy management system described in claim 1, it is characterized by: The prescription selection and issuance module includes a prescription selection submodule and a prescription issuance submodule. The prescription selection submodule is used by doctors to select drugs and can input multiple interchangeable drugs of the same category. The prescription issuance submodule is used by the system to generate electronic prescriptions and send them to the cooperating DTP pharmacy.

4. According to the management system of a DTP pharmacy in claim 1, it is characterized in that The DTP pharmacy management system also includes: The drug sales monitoring and analysis module is used to input drug information and prices. Among the multiple drugs prescribed by doctors, the system gives priority to recommending alternative drugs with the same efficacy from the sales company based on the price fluctuation range; Time series forecasting module, used to predict drug sales trends in the future; Sales data analysis module, used to discover sales patterns and potential market opportunities through cluster analysis and association rule learning; Anomaly detection module, used to monitor sales data in real time and detect abnormal sales fluctuations.

5. A DTP pharmacy management system according to claim 4, characterized in that: The drug sales monitoring and analysis module includes a sales information input submodule and a price range recommendation submodule. The sales information input submodule is used for the sales company to input drug information and prices, which are then stored by the system. The price range recommendation submodule is used for the system to give priority to recommending the sales company's interchangeable drugs with the same efficacy according to the price fluctuation range among the multiple drugs recommended by doctors. ； Input: Multiple drugs recommended by doctors and their prices; Output: Prioritize the recommendation of drugs from sales companies whose price fluctuation does not exceed a% of the drugs recommended by doctors; formula: P s : The price of the company's drugs; P d : Price of drugs recommended by doctors; a: The preset price fluctuation percentage, set by the sales company; If the price fluctuation of the sales company's drugs does not exceed a% of the doctor's recommended drugs, the sales company's drugs will be recommended first.

6. A DTP pharmacy management system according to claim 4, characterized in that: The time series prediction module is used to predict the drug sales trend in the future based on the time series prediction algorithm of the long short-term memory network; ； ； x t : Historical sales data at time t as input features; y t : The predicted value at time t, indicating the drug sales volume at a certain point in the future; W x : Input weight matrix, used to transform input features into part of hidden state; W h : Hidden state weight matrix, used to convert the hidden state of the previous moment into the hidden state of the current moment; W hx : The weight matrix from hidden state to output, used to convert the hidden state into the final predicted value; h t-1 : The hidden state of the previous moment, indicating the memory of the model at the previous moment; b h : Bias term, used to adjust the output of the model; σ: Activation function, used to introduce nonlinear characteristics so that the model can capture complex patterns.

7. A DTP pharmacy management system according to claim 4, characterized in that: The sales data analysis module uses the K-means clustering algorithm to discover sales patterns and potential market opportunities: ； C: A set of clusters, which represents the groups into which sales records are divided; k: the number of clusters, that is, the number of groups; C i : The i-th cluster, representing a group; x j : a data point in a sales record; μ i : The center of the i-th cluster, indicating the average value of the group; ∥x j -μ i ∥ 2 : data point x j With cluster center μ i The squared Euclidean distance between them is used to measure the proximity of a data point to the cluster center; Association rule learning is used to calculate support and confidence through the Apriori algorithm, discover frequent item sets and association rules, and help sales teams discover potential market opportunities and sales models: ； X and Y: item sets, representing a set of drugs or sales records; N: total number of records, indicating the total number of sales records; count(X): the number of times item set X appears, indicating the number of sales records containing X; support(X): the support of item set X, which indicates the frequency of X appearing in all sales records; support(X∪Y): the support of item sets X and Y appearing at the same time, indicating the number of sales records containing X and Y; confidence(X→Y): The confidence from item set X to item set Y, which indicates the proportion of sales records containing X that also contain Y.

8. A DTP pharmacy management system according to claim 4, characterized in that: The anomaly detection module is used to monitor sales data in real time, detect abnormal sales fluctuations, and convert sales data into dimensionless standardized values ​​through standardization processing: ； S i : sales volume on the i-th day; μ S : The average value of sales volume; σ S : Standard deviation of sales volume; z i : The degree of deviation between the sales volume on the i-th day and the average sales volume; ； θ: preset threshold, indicating the maximum acceptable standardized deviation; ∣z i ∣>θ, assuming that the current sales volume is x t Deviations from historical averages that exceed preset thresholds trigger alerts, alerting the sales team to potential market changes or supply chain issues.

9. An application of a DTP pharmacy, characterized in that: The invention comprises the application of a DTP pharmacy management system as claimed in any one of claims 1 to 8 in a DTP pharmacy.

Images