Indication information classification coding method and indication information retrieval recommendation method

Abstract

The invention belongs to the technical field of big data, and particularly relates to an indication information classification coding method and an indication information retrieval recommendation method. The indication information classification coding method comprises the following steps: S0, making normalized codes of attribute characteristics of all dimensions of indication information, and forming a normalized code database; s1, acquiring indication information of each data source; s2, carrying out attribute feature data capture on the obtained indication information to obtain attribute feature information of each dimension of the indication information; s3, normalizing the grabbed attribute feature information of each dimension to obtain specification information conforming to a preset standard; and S4, coding each dimension attribute feature of each indication information by using the standardized coding database constructed in the step S0, and storing a coding result in a preset indication coding database. The method can adapt to medical technology development, meets precise medical requirements, and is uniform in specification and easy to use.

Description

Technical Field

[0001] The present invention belongs to the field of big data technology, and in particular relates to a classification and coding method for indication information and a retrieval and recommendation method for indication information. Background Art

[0002] Indications, as the scope and standard for the appropriate use of medical methods such as drugs and surgeries, are a crucial concept in the medical field. Generally speaking, people usually understand indications as specific diseases or symptoms, but in fact, a complete indication information is far more than that, it also includes a lot of non-disease and symptom information.

[0003] Taking the description of indications in drug instructions as an example, many countries or regions with strong supervision in the world currently have detailed content requirements for the description of drug indications. For example, China's "General Format and Writing Guidelines for Instructions for Chemical Drugs and Therapeutic Biological Products", the United States' "US Federal Regulations" and the European Union's "Summary of Product Characteristics Guidelines" all clearly stipulate that the [Indications] item in the drug instructions should clearly describe the drug's use, applicable population, applicable diseases or symptoms, therapeutic status, and restrictions on use. This means that complete indication information not only covers diseases or symptoms, but also involves characteristic information in multiple dimensions such as use, applicable population, therapeutic status, and restrictions on use.

[0004] However, the sources of indication information are extensive and complex, including public information from domestic and foreign drug regulatory agencies, clinical trial data, medical product instructions, news reports, annual reports, diagnosis and treatment guidelines, professional literature, medical records, diagnostic reports, etc. Due to differences in time, space, culture, language, etc., these indication information has many non-standard and non-uniform problems, which has brought many troubles to medical decision-making, drug research and development and supervision.

[0005] At present, although there are some coding systems in the industry, most of them tend to classify and code "diseases or symptoms", and there are many shortcomings. Take the International Classification of Diseases (ICD) system as an example. Although it is an internationally universal unified classification and coding system developed by the World Health Organization (WHO) to promote the analysis of population health status and the study of differences in causes of death in countries around the world, it has the advantages of clear classification and easy tracking, but its update speed is slow. It often takes several years from proposal to implementation, and it is difficult to keep up with the rapid development of medical technology and pathology research. In addition, the coding form and rules of ICD are relatively complex, and the requirements for professionalism are high, making it difficult to classify and code quickly and accurately, and it is easy to make mistakes or omissions during manual classification. At the same time, the classification granularity of ICD is not detailed enough, and some classification ranges are too large to accurately reflect the condition and meet the needs of the development of precision medicine. More importantly, ICD mainly classifies based on common conditions, ignoring the differences between individuals, and it is difficult to adapt to the development trend of personalized medicine.

[0006] These problems not only affect the accuracy of medical decision-making, but also bring potential risks to drug development, supervision and patient medication safety. Solving these problems requires building a comprehensive, standardized, timely updated and easy-to-use indication information coding system, but this task faces many challenges. For example, how to ensure the comprehensiveness and standardization of the coding system, how to update it in a timely manner to adapt to the development of medical technology, how to reduce the complexity of coding to improve the accuracy and efficiency of classification, and how to refine the classification granularity to meet the needs of precision medicine.

[0007] Therefore, how to construct a classification and coding method for indication information that can adapt to the development of medical technology, meet the needs of precision medicine, and be standardized, unified, and easy to use has become a problem that needs to be solved urgently. Summary of the invention

[0008] In view of the above-mentioned deficiencies in the prior art, the present invention provides a classification and coding method for indication information, which can adapt to the development of medical technology, meet the needs of precision medicine, be standardized and unified, and be easy to use.

[0009] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0010] A method for classifying and encoding indication information comprises the following steps:

[0011] S0. Prepare standardized codes for the attribute characteristics of each dimension of indication information to form a standardized code database; wherein, for a dimension, its standardized code includes each code used in the dimension and the standard information corresponding to each code; a code under a dimension includes a dimension code, a level code, and an additional code field; and specify the association relationship between each code under the same dimension;

[0012] S1. Obtain the indication information from each data source and construct the original indication information dataset;

[0013] S2. Based on the dimensions included in the standardized coding, the attribute characteristic data of the acquired indication information is captured to obtain the attribute characteristic information of each dimension of the indication information;

[0014] S3, normalizing the captured attribute feature information of each dimension to obtain standardized information of the attribute feature information of each dimension that meets the preset standard;

[0015] S4. Use the standardized coding database constructed in S0 to encode the attribute characteristics of each dimension of each indication information, and store the coding results in a preset indication coding database.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] 1. This method constructs a standardized coding database. Through the S0 step, this method creates a standardized coding database containing the attribute characteristics of each dimension. This database not only contains all the codes of each dimension, but also provides the standardized information corresponding to each code and the relationship between codes. This step ensures the standardization and uniformity of the indication information and provides a solid foundation for subsequent classification coding. At the same time, the updatability of the database enables us to keep up with the development of medical technology and maintain the timeliness of the coding system.

[0018] 2. This method can automatically capture and normalize attribute features of each dimension. In steps S1 and S2, indication information is obtained from various data sources, and attribute feature data is captured according to the dimensions of the standardized coding database. Subsequently, in step S3, the captured attribute feature information is normalized to meet the preset standards. This process has a high degree of automation, which greatly improves processing efficiency and reduces errors and omissions in manual classification and coding. Normalization ensures the accuracy and consistency of information and provides reliable data support for subsequent coding.

[0019] 3. This method can achieve efficient coding and storage. In step S4, the constructed standardized coding database is used to encode the attribute characteristics of each dimension of each indication information, and the coding results are stored in the preset indication coding database. In this way, the advantages of the standardized coding database are fully utilized to achieve fast and accurate coding. At the same time, the storage of the coding results facilitates subsequent information retrieval and analysis, and improves work efficiency.

[0020] In summary, this method can build a comprehensive, standardized, timely updated and easy-to-use indication information coding system, which effectively solves the problems mentioned in the background technology, such as the non-standard and non-uniform indication information, as well as the slow update of the existing coding system, complex coding, and insufficient classification granularity. Through automated capture, standardized processing, and efficient coding, this method not only improves processing efficiency and accuracy, but also provides a solid foundation for subsequent information retrieval and analysis, and promotes the intelligent and precise development of medical decision-making, drug research and development, and regulatory work.

[0021] Preferably, in S0, the association relationships between codes in the same dimension include equal to, belongs to and no relationship.

[0022] Such a setting, 1. Makes the information hierarchy clear. Through the "belongs to" relationship, the internal hierarchy of the indication information can be constructed. For example, a disease may belong to a broader disease category. This hierarchical relationship helps to better understand the associations and differences between diseases. Improve coding accuracy. The "equal to" relationship ensures a one-to-one correspondence between codes and avoids classification errors caused by coding confusion. This helps to more accurately describe and classify indication information and improve the accuracy and reliability of coding. "No relationship" allows different information points to be clearly distinguished during coding and avoids unnecessary connections. This helps to more clearly identify key information in the subsequent information integration and analysis process, and improves the accuracy and efficiency of analysis.

[0023] 2. It can improve coding flexibility. Clear association settings make the coding process more flexible. According to actual needs, the coding can be appropriately adjusted and optimized while maintaining coding consistency to adapt to different application scenarios and data requirements. It can also improve information retrieval efficiency. The association between codes helps to build a more efficient index structure, thereby speeding up information retrieval. This enables researchers and decision makers to obtain the required information faster and improve decision-making efficiency.

[0024] Preferably, in S0, the dimensions include disease name, symptoms, causes, biomarkers, pathological characteristics, clinical characteristics, level, use, applicable population, treatment status, and other custom dimensions.

[0025] Such a setting can, 1. Improve the comprehensiveness of information. By covering multiple core dimensions such as disease name, symptoms, and etiology, it can more comprehensively capture all aspects of indication information and ensure the completeness and accuracy of the information. This will help to more accurately reflect the true situation of the indication in the subsequent classification and coding process.

[0026] 2. It can enhance classification accuracy. The refinement of dimension settings, such as biomarkers and pathological characteristics, makes classification coding more refined. This not only improves the accuracy of coding, but also helps meet the needs of precision medicine for detailed classification. The introduction of custom dimensions enables the method to flexibly adapt to the needs of different data sources and specific application scenarios. This enhances the versatility and practicality of the method, enabling it to be widely used in various medical decision-making, drug development and regulatory work.

[0027] 3. Clear dimensional division and construction of standardized coding database make the coding process more efficient. Through automated capture and standardized processing, the classification and coding of indication information can be completed quickly, greatly saving manpower and time costs. The coded indication information is stored in the preset database, which greatly facilitates subsequent information retrieval and analysis. This helps researchers and decision makers quickly obtain the required information and make more accurate and scientific decisions.

[0028] Preferably, in S1, the data sources include public information from medical product regulatory agencies, clinical trials, medical product instructions, news, annual reports, diagnosis and treatment guidelines, literature, medical records, and diagnostic reports.

[0029] With this setup, 1. The comprehensiveness and accuracy of information are improved. By integrating multiple data sources, more comprehensive and accurate information on indications can be obtained. Different data sources may contain information that complements or verifies each other, which helps us understand the full picture of the indications more comprehensively and improves the accuracy and reliability of information. Data diversity is enhanced. Diversified data sources provide a rich variety of data types and information forms, including text, charts, statistical data, etc. This diversity helps to use data more flexibly and dig out deeper information in the subsequent analysis and decision-making process.

[0030] 2. Adapt to different application scenarios. Different data sources may be suitable for different application scenarios. For example, public information from medical product regulatory agencies may be more suitable for regulatory compliance and compliance checks, while clinical trial data is more suitable for drug development and efficacy evaluation. By selecting the right data source, the needs of specific application scenarios can be better met.

[0031] 3. Enhance information credibility. Certain data sources (such as public information from medical product regulatory agencies, clinical trials, diagnosis and treatment guidelines, etc.) are highly authoritative and credible. By integrating these data sources, the quality and credibility of the overall information can be improved, providing strong support for subsequent decision-making and analysis.

[0032] Preferably, in S2, attribute feature data is captured manually or using an AI algorithm based on machine learning.

[0033] This setting, through combining manual and machine learning-based AI algorithms to capture attribute feature data, not only ensures the accuracy and flexibility of the data, but also improves the efficiency of data capture, realizes intelligent identification and classification, has the ability of continuous learning and optimization, and adapts to the needs of different data sources. These effects together provide a high-quality data foundation for subsequent standardized processing and coding work, and promote the intelligent and precise development of medical decision-making, drug research and development, and supervision.

[0034] The method also provides a method for retrieving and recommending indication information, using the indication coding database obtained by the above indication information classification coding method; comprising the following steps:

[0035] The first step is to encode the indication information to be replaced according to the above-mentioned indication information classification and encoding method;

[0036] The second step is to set the dimensions to be compared and the weight coefficients of each dimension according to the needs;

[0037] The third step is to calculate the similarity between the indication information to be replaced and each indication information in the indication coding database according to the set dimensions and the weight coefficients of each dimension;

[0038] The fourth step is to make similar recommendations for indication information based on the similarity calculation results.

[0039] Beneficial effects:

[0040] 1. Accurate matching and recommendation. By encoding the indication information to be replaced and comparing it with the information in the indication coding database, this method can accurately find similar indication information. This precise matching capability helps to quickly find the indication information that best meets the patient's needs in scenarios such as drug replacement and treatment plan selection, thereby improving the scientificity and accuracy of medical decision-making. Thereby improving the quality of medical decisions. The similarity calculation and usage methods provided by this method can provide more comprehensive and accurate information support for medical decisions. By comparing the similarity of different indication information, doctors can more scientifically choose treatment plans, drug replacements, etc., and improve the quality and effectiveness of medical decisions.

[0041] 2. Flexible setting of comparison dimensions and weights. This method allows users to flexibly set the dimensions to be compared and the weight coefficients of each dimension according to actual needs. This flexibility enables the method to be applicable to different application scenarios and decision-making needs, ensuring the pertinence and practicality of the similarity calculation results.

[0042] 3. It can realize the integration and utilization of indication information. By calculating the similarity, this method can integrate and classify the indication information to form a clearer and more orderly information structure. This information integration capability helps researchers and decision makers better understand and utilize indication information and tap into its potential value.

[0043] In summary, this technical solution provides more scientific, accurate and efficient information support for medical decision-making, drug research and development, and supervision through precise matching and recommendation, flexible setting of comparison dimensions and weights, efficient calculation and rapid response, information integration and utilization, and improvement of the quality of medical decision-making, thereby promoting the intelligent and precise development of the medical industry.

[0044] Preferably, in the third step, the similarity between the indication information to be replaced and the indication information in the indication coding database is calculated according to the following formula:

[0045]

[0046] In the formula, F represents the similarity between indication information; n is the number of dimensions to be compared; a i is the weight coefficient of the i-th dimension to be compared; Z i is the number of codes of the i-th dimension that need to be compared for the indication information to be replaced; X i The number of codes for which the correlation is equal to or belongs to in the i-th dimension for which the two indications need to be compared.

[0047] Preferably, X i The calculation formula is:

[0048]

[0049] Where m is the number of codes of the i-th dimension of the indication information to be compared; b k is 0 or 1, b k =1 indicates that the kth code of the i-th dimension of the indication information to be compared is equal to or belongs to one of the codes of the i-th dimension to be compared with the indication information to be replaced; b k =0 indicates that the kth code of the i-th dimension of the indication information to be compared is irrelevant to all the codes of the i-th dimension to be compared with the indication information to be replaced.

[0050] Such a setting, 1. The similarity can be accurately quantified. The above formula comprehensively considers multiple dimensions and their weight coefficients, as well as the number of codes whose association relationship is equal to or belongs to each dimension, so as to accurately quantify the similarity between the indication information to be replaced and other information in the database. This precise quantification capability helps to more accurately evaluate the correlation and differences between different indication information in scenarios such as medical decision-making and drug development. The n dimensions involved in the formula make the similarity calculation more comprehensive and detailed. This multi-dimensional comparative analysis helps to discover the potential connections and common features between different indication information, and provide more comprehensive information support for medical decision-making. Based on the precise quantified similarity results, this technical solution can support intelligent recommendation and decision-making functions. For example, in scenarios such as drug replacement and treatment plan selection, the indication information or treatment plan that best meets the patient's needs can be automatically recommended based on the similarity calculation results, thereby improving the scientificity and accuracy of medical decision-making.

[0051] 2. The weight coefficient can be flexibly adjusted. The weight coefficient ai in the formula can be flexibly adjusted according to actual needs, so as to highlight the importance of certain key dimensions in similarity calculation. This flexibility enables the method to be applicable to different application scenarios and decision-making needs, and improves the pertinence and practicality of the similarity calculation results.

[0052] Preferably, in the fourth step, the results of similarity calculation are sorted from large to small, and the first P indication information in the sorting are used as recommended entities.

[0053] With such a setting, by sorting and selecting the top P indication information, the technical solution can provide users with accurate recommendation results. These recommended entities have a high degree of similarity with the indication information to be replaced. Therefore, in scenarios such as drug replacement and treatment plan selection, they can provide users with valuable reference information and support more scientific medical decisions. When faced with a large amount of indication information, it is often difficult to quickly find the information that best meets the needs. Through the sorting and recommendation functions of this method, users can quickly lock in the top P most relevant indication information, thereby greatly improving decision-making efficiency and reducing unnecessary search and comparison time. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0055] Figure 1 is a flow chart of the method for classifying and encoding indication information in an embodiment;

[0056] Figure 2 This is a flowchart of a recommended method for retrieving indication information in an embodiment. DETAILED DESCRIPTION

[0057] The following is a further detailed description through specific implementation methods:

[0058] Embodiment 1

[0059] like Figure 1 As shown, this embodiment discloses a classification and coding method for indication information, comprising the following steps:

[0060] S0. Prepare standardized codes for the attribute characteristics of each dimension of indication information to form a standardized code database; wherein, for a dimension, its standardized code includes each code used in the dimension and the standard information corresponding to each code; a code under a dimension includes a dimension code, a level code, and an additional code field; and specify the association relationship between each code under the same dimension;

[0061] Among them, the association relationships between codes in the same dimension include equal, belongs to and no relationship.

[0062] In this way, the information hierarchy is clear. Through the "belongs to" relationship, the internal hierarchy of the indication information can be constructed. For example, a disease may belong to a broader disease category, and this hierarchical relationship helps to better understand the associations and differences between diseases. Improve coding accuracy. The "equal to" relationship ensures a one-to-one correspondence between codes and avoids classification errors caused by coding confusion. This helps to more accurately describe and classify indication information and improve the accuracy and reliability of coding. "No relationship" allows different information points to be clearly distinguished during coding and avoids unnecessary connections. This helps to more clearly identify key information in the subsequent information integration and analysis process, improve the accuracy and efficiency of analysis. It can also improve coding flexibility. Clear association relationship settings make the coding process more flexible. According to actual needs, while maintaining coding consistency, the coding can be appropriately adjusted and optimized to adapt to different application scenarios and data requirements. It can also improve information retrieval efficiency. The association relationship between codes helps to build a more efficient index structure, thereby speeding up information retrieval. This enables researchers and decision makers to obtain the required information faster and improve decision-making efficiency.

[0063] In specific implementation, the dimensions of indication information include disease name, symptoms, causes, biomarkers, pathological characteristics, clinical characteristics, level, use, applicable population, treatment status, and other custom dimensions.

[0064] In this way, the comprehensiveness of information can be improved. By covering multiple core dimensions such as disease name, symptoms, and etiology, it is possible to capture all aspects of indication information more comprehensively and ensure the completeness and accuracy of the information. This helps to more accurately reflect the true situation of the indication in the subsequent classification and coding process. It can also enhance the accuracy of classification. The refinement of dimension settings, such as biomarkers and pathological characteristics, makes the classification coding more refined. This not only improves the accuracy of coding, but also helps to meet the needs of precision medicine for detailed classification. The introduction of custom dimensions enables the method to flexibly adapt to the needs of different data sources and specific application scenarios. This enhances the versatility and practicality of the method, enabling it to be widely used in various medical decision-making, drug research and development, and regulatory work. In addition, clear dimensional division and the construction of a standardized coding database make the coding process more efficient. Through automated capture and standardized processing, the classification and coding of indication information can be completed quickly, greatly saving manpower and time costs. The coded indication information is stored in a preset database, which provides great convenience for subsequent information retrieval and analysis. This helps researchers and decision makers quickly obtain the required information and make more accurate and scientific decisions.

[0065] For better understanding, Table 1 is used to illustrate the normalized encoding.

[0066] Table 1. Examples of normalized coding

[0067]

[0068]

[0069]

[0070] S1. Obtain the indication information from each data source and construct the original indication information dataset.

[0071] During specific implementation, data sources include public information from medical product regulatory agencies, clinical trials, medical product instructions, news, annual reports, diagnosis and treatment guidelines, literature, medical records, and diagnostic reports.

[0072] In this way, the comprehensiveness and accuracy of information can be improved. By integrating multiple data sources, more comprehensive and accurate indication information can be obtained. Different data sources may contain information that complements or verifies each other, which helps us understand the full picture of the indication more comprehensively and improves the accuracy and reliability of information. Data diversity is enhanced. Diversified data sources provide a rich variety of data types and information forms, including text, charts, statistics, etc. This diversity helps to use data more flexibly and dig out deeper information in the subsequent analysis and decision-making process. It can also adapt to different application scenarios. Different data sources may be suitable for different application scenarios. For example, public information from medical product regulatory agencies may be more suitable for regulatory compliance and compliance inspections, while clinical trial data is more suitable for drug development and efficacy evaluation. By selecting the right data source, the needs of specific application scenarios can be better met. In addition, the credibility of information can be enhanced. Some data sources (such as public information from medical product regulatory agencies, clinical trials, diagnosis and treatment guidelines, etc.) are highly authoritative and credible. By integrating these data sources, the quality and credibility of the overall information can be improved, providing strong support for subsequent decision-making and analysis.

[0073] S2. Based on the dimensions included in the standardized coding, the attribute feature data of the acquired indication information is captured to obtain the attribute feature information of each dimension of the indication information.

[0074] In specific implementation, attribute feature data is captured manually or using AI algorithms based on machine learning. In this way, by combining manual and machine learning-based AI algorithms to capture attribute feature data, this technical solution not only ensures the accuracy and flexibility of the data, but also improves the efficiency of data capture, realizes intelligent identification and classification, has the ability of continuous learning and optimization, and adapts to the needs of different data sources. These effects together provide a high-quality data foundation for subsequent standardized processing and coding work, and promote the intelligent and precise development of medical decision-making, drug research and development, and supervision.

[0075] S3. Normalize the captured attribute feature information of each dimension to obtain standardized information of the attribute feature information of each dimension that meets the preset standard.

[0076] S4. Use the standardized coding database constructed in S0 to encode the attribute characteristics of each dimension of each indication information, and store the coding results in a preset indication coding database.

[0077] Through the S0 step, the method produces a standardized coding database containing attribute features of each dimension. This database not only contains all the codes of each dimension, but also provides the standardized information corresponding to each code and the association between the codes. This step ensures the standardization and uniformity of the indication information, and provides a solid foundation for subsequent classification coding. At the same time, the updatability of the database enables us to keep up with the development of medical technology in a timely manner and maintain the timeliness of the coding system. In the S1 and S2 steps, the indication information is obtained from each data source, and the attribute feature data is captured according to the dimensions of the standardized coding database. Subsequently, in the S3 step, the captured attribute feature information is standardized to meet the preset standards. This process has a high degree of automation, which greatly improves the processing efficiency and reduces the errors and omissions of manual classification coding. The standardized processing ensures the accuracy and consistency of the information and provides reliable data support for subsequent coding. In the S4 step, the constructed standardized coding database is used to encode the attribute features of each dimension of each indication information, and the coding results are stored in the preset indication coding database. In this way, the advantages of the standardized coding database are fully utilized to achieve fast and accurate coding. At the same time, the storage of coding results facilitates subsequent information retrieval and analysis and improves work efficiency.

[0078] In summary, this method can build a comprehensive, standardized, timely updated and easy-to-use indication information coding system, which effectively solves the problems mentioned in the background technology, such as the non-standard and non-uniform indication information, as well as the slow update of the existing coding system, complex coding, and insufficient classification granularity. Through automated capture, standardized processing, and efficient coding, this method not only improves processing efficiency and accuracy, but also provides a solid foundation for subsequent information retrieval and analysis, and promotes the intelligent and precise development of medical decision-making, drug research and development, and regulatory work.

[0079] Embodiment 2

[0080] like Figure 2 As shown, the method also provides a method for retrieving and recommending indication information, using the indication coding database obtained by the above indication information classification coding method; comprising the following steps:

[0081] The first step is to encode the indication information to be replaced according to the above-mentioned indication information classification and encoding method;

[0082] The second step is to set the dimensions to be compared and the weight coefficients of each dimension according to the needs;

[0083] The third step is to calculate the similarity between the indication information to be replaced and each indication information in the indication coding database according to the set dimensions and the weight coefficients of each dimension.

[0084] In specific implementation, the similarity between the indication information to be replaced and the indication information in the indication coding database is calculated according to the following formula:

[0085]

[0086] In the formula, F represents the similarity between indication information; n is the number of dimensions to be compared; a i is the weight coefficient of the i-th dimension to be compared; Z i is the number of codes of the i-th dimension that need to be compared for the indication information to be replaced; X i The number of codes for which the correlation is equal to or belongs to in the i-th dimension for which the two indications need to be compared.

[0087] X i The calculation formula is:

[0088]

[0089] Where m is the number of codes of the i-th dimension of the indication information to be compared; b k is 0 or 1, b k =1 indicates that the kth code of the i-th dimension of the indication information to be compared is equal to or belongs to one of the codes of the i-th dimension to be compared with the indication information to be replaced; b k =0 indicates that the kth code of the i-th dimension of the indication information to be compared is irrelevant to all the codes of the i-th dimension to be compared with the indication information to be replaced.

[0090] To facilitate understanding of multiple codes under one dimension, an example is used for explanation. Assuming that there is only one disease that a certain drug can treat in an indication information, then there is only one code corresponding to the disease under the disease dimension in the indication information; assuming that there are n diseases that a certain drug can treat in an indication information, then there are n codes corresponding to these n diseases under the disease dimension in the indication information.

[0091] Assuming that the i-th dimension is disease, calculate X i When the disease dimension of the indication information to be compared is analyzed, each code under the disease dimension of the indication information to be compared needs to be analyzed in turn with each code in the indication information to be replaced, to analyze whether the correlation with one of the codes is equal to or belongs to; finally, the number of codes with the correlation equal to or belongs to is counted, which is X i .

[0092] In this way, the similarity can be accurately quantified. The above formula comprehensively considers multiple dimensions and their weight coefficients, as well as the number of codes that are equal to or belong to each dimension, so that the similarity between the indication information to be replaced and other information in the database can be accurately quantified. This precise quantification capability helps to more accurately evaluate the correlation and differences between different indication information in scenarios such as medical decision-making and drug development. The n dimensions involved in the formula make the similarity calculation more comprehensive and detailed. This multi-dimensional comparative analysis helps to discover the potential connections and common features between different indication information, and provide more comprehensive information support for medical decision-making. Based on the precise quantified similarity results, this technical solution can support intelligent recommendation and decision-making functions. For example, in scenarios such as drug replacement and treatment plan selection, the indication information or treatment plan that best meets the patient's needs can be automatically recommended based on the similarity calculation results, thereby improving the scientificity and accuracy of medical decision-making. In addition, the weight coefficient can be flexibly adjusted. The weight coefficient ai in the formula can be flexibly adjusted according to actual needs, so as to highlight the importance of certain key dimensions in similarity calculation. This flexibility enables the method to be applicable to different application scenarios and decision-making needs, and improves the pertinence and practicality of the similarity calculation results.

[0093] The fourth step is to make similar recommendations for indication information based on the similarity calculation results. In the specific implementation, the results of the similarity calculation are sorted from large to small, and the first P indication information in the sorting are used as recommended entities. In this way, by sorting and selecting the first P indication information, the technical solution can provide users with accurate recommendation results. These recommended entities have a high degree of similarity with the indication information to be replaced. Therefore, in scenarios such as drug replacement and treatment plan selection, they can provide users with valuable reference information and support more scientific medical decisions. When faced with a large amount of indication information, it is often difficult to quickly find the information that best meets the needs. Through the sorting and recommendation functions of this method, users can quickly lock in the first P most relevant indication information, thereby greatly improving decision-making efficiency and reducing unnecessary search and comparison time.

[0094] By encoding the indication information to be replaced and comparing it with the information in the indication coding database, this method can accurately find similar indication information. This precise matching capability helps to quickly find the indication information that best meets the patient's needs in scenarios such as drug replacement and treatment plan selection, and improve the scientificity and accuracy of medical decision-making. Thereby improving the quality of medical decision-making. The similarity calculation and usage method provided by this method can provide more comprehensive and accurate information support for medical decision-making. By comparing the similarity of different indication information, doctors can more scientifically select treatment plans, drug replacements, etc., and improve the quality and effect of medical decision-making. In addition, this method allows users to flexibly set the dimensions to be compared and the weight coefficients of each dimension according to actual needs. This flexibility enables the method to be applicable to different application scenarios and decision-making needs, ensuring the pertinence and practicality of the similarity calculation results. This method can also realize the integration and utilization of indication information. By calculating the similarity, this method can integrate and classify the indication information to form a clearer and more orderly information structure. This information integration capability helps researchers and decision makers better understand and utilize indication information and tap its potential value.

[0095] In summary, this method provides more scientific, accurate and efficient information support for medical decision-making, drug research and development, and supervision through precise matching and recommendation, flexible setting of comparison dimensions and weights, efficient calculation and rapid response, information integration and utilization, and improvement of the quality of medical decision-making, and promotes the intelligent and precise development of the medical industry.

[0096] Embodiment 3

[0097] In order to better understand the method for retrieving and recommending indication information in the present invention, a specific example is used for explanation.

[0098] 1. Usage scenarios and requirements

[0099] Scenario: A large tertiary hospital is developing a treatment plan for a child with a rare heart disease. Since the disease is relatively rare, the doctor needs to refer to a large amount of relevant indication information to find the most suitable treatment plan for the patient.

[0100] Requirement: Doctors want to use the existing indication coding database to quickly find the top five indications that are most similar to the patient's condition, so as to refer to their treatment plans and drug selection.

[0101] 2. Specific use process

[0102] Data preparation:

[0103] First, collect the patient's detailed medical information, including symptoms, signs, test results, etc.

[0104] Input this information into the indication information classification and coding system for coding.

[0105] Set comparison dimensions and weights:

[0106] Doctors set the dimensions that need to be compared based on the patient's specific situation, such as disease name, symptoms, cause, biomarkers, etc.

[0107] Assign a weight coefficient to each dimension to reflect its importance in the selection of treatment options. For example, for this patient with a rare heart disease, the disease name and symptoms may have a higher weight, while the biomarker may have a lower weight.

[0108] Similarity calculation:

[0109] The formula is used to calculate the similarity between the indication information to be replaced (i.e., patient disease information) and each indication information in the indication coding database.

[0110] The formula takes into account multiple dimensions and their weight coefficients, as well as the number of codes for which the association relationship is equal to or belongs to under each dimension.

[0111] Sorting and recommendation:

[0112] According to the similarity calculation results, the indication information in the indication coding database is sorted.

[0113] The top five indication information in the ranking are selected as recommended entities.

[0114] View recommended results:

[0115] The doctor reviews the recommended results and obtains information on the top five indications that are most similar to the patient's condition.

[0116] Read the treatment options and drug selection information carefully and formulate the final treatment plan based on the patient's specific situation.

[0117] 3. Principle and Effect Description

[0118] Principle: This technical solution builds an indication coding database based on the indication information classification coding method. By setting the comparison dimension and weight coefficient, the formula is used to calculate the similarity between the indication information to be replaced and each indication information in the database. The results of the similarity calculation are sorted and the first P indication information is selected as the recommended entity.

[0119] Effect:

[0120] Accurate recommendations: By accurately quantifying similarities, this technical solution can provide users with accurate recommendation results. In this example, the doctor can quickly find the top five indications with the highest similarity to the patient's symptoms, providing valuable reference for formulating treatment plans.

[0121] Improve decision-making efficiency: The sorting and recommendation functions enable doctors to quickly identify the most relevant indication information, reduce unnecessary search and comparison time, and improve decision-making efficiency.

[0122] Optimize user experience: Intuitive sorting and recommendation results help doctors understand recommendation information more quickly and optimize user experience.

[0123] Supporting personalized needs: By adjusting the comparison dimensions and weight coefficients, this technical solution can support personalized needs and provide users with recommendation results that are more in line with their actual conditions. In this example, the doctor set the comparison dimensions and weight coefficients according to the patient's specific situation to ensure the pertinence and practicality of the recommendation results.

[0124] In summary, this method plays an important role in the formulation of treatment plans for patients with rare heart diseases, provides doctors with accurate and efficient recommendation results, and supports more scientific medical decisions.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A classification and coding method for indication information, characterized in that: The following steps are involved: S0. Prepare standardized codes for the attribute characteristics of each dimension of indication information to form a standardized code database; wherein, for a dimension, its standardized code includes each code used in the dimension and the standard information corresponding to each code; a code under a dimension includes a dimension code, a level code, and an additional code field; and specify the association relationship between each code under the same dimension; S1. Obtain the indication information from each data source and construct the original indication information dataset; S2. Based on the dimensions included in the standardized coding, the attribute characteristic data of the acquired indication information is captured to obtain the attribute characteristic information of each dimension of the indication information; S3, normalizing the captured attribute feature information of each dimension to obtain standardized information of the attribute feature information of each dimension that meets the preset standard; S4. Use the standardized coding database constructed in S0 to encode the attribute characteristics of each dimension of each indication information, and store the coding results in a preset indication coding database.

2. The method for classifying and encoding indication information according to claim 1, characterized in that: In S0, the association relationships between codes in the same dimension include equal, belongs to and no relationship.

3. The method for classifying and encoding indication information according to claim 1, characterized in that: In S0, dimensions include disease name, symptoms, causes, biomarkers, pathological characteristics, clinical characteristics, level, use, applicable population, treatment status, and other custom dimensions.

4. The method for classifying and encoding indication information according to claim 3, characterized in that: In S1, data sources include public information from medical product regulatory agencies, clinical trials, medical product instructions, news, annual reports, diagnosis and treatment guidelines, literature, medical records, and diagnostic reports.

5. The method for classifying and encoding indication information according to claim 1, characterized in that: In S2, attribute feature data is captured manually or using AI algorithms based on machine learning.

6. A method for retrieving and recommending indication information, characterized in that: The indication coding database obtained by using the indication information classification coding method according to claim 2 comprises the following steps: The first step is to encode the indication information to be replaced according to the indication information classification encoding method according to claim 2; The second step is to set the dimensions to be compared and the weight coefficients of each dimension according to the needs; The third step is to calculate the similarity between the indication information to be replaced and each indication information in the indication coding database according to the set dimensions and the weight coefficients of each dimension; The fourth step is to make similar recommendations for indication information based on the similarity calculation results.

7. The method for retrieving and recommending indication information according to claim 6, characterized in that: In the third step, the similarity between the indication information to be replaced and the indication information in the indication coding database is calculated according to the following formula: In the formula, F represents the similarity between indication information; n is the number of dimensions to be compared; a i is the weight coefficient of the i-th dimension to be compared; Z i is the number of codes of the i-th dimension that need to be compared for the indication information to be replaced; X i The number of codes for which the correlation is equal to or belongs to in the i-th dimension for which the two indications need to be compared.

8. The method for retrieving and recommending indication information according to claim 7, characterized in that: X i The calculation formula is: Where m is the number of codes of the i-th dimension of the indication information to be compared; b k is 0 or 1, b k =1 indicates that the kth code of the i-th dimension of the indication information to be compared is equal to or belongs to one of the codes of the i-th dimension to be compared with the indication information to be replaced; b k =0 indicates that the kth code of the i-th dimension of the indication information to be compared is irrelevant to all the codes of the i-th dimension to be compared with the indication information to be replaced.

9. The method for retrieving and recommending indication information according to claim 6, characterized in that: In the fourth step, the results of similarity calculation are sorted from large to small, and the first P indication information in the sorting is used as the recommended entity.

Images