Peritoneal dialysis personalized prescription recommendation method and system based on mapping knowledge domain
By constructing a peritoneal dialysis knowledge graph to integrate multi-source data and combining graph traversal and reasoning, personalized peritoneal dialysis prescriptions are generated, solving the problem that personalization and standardization are difficult to achieve in existing methods and improving the accuracy and reliability of prescriptions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- THE FIRST AFFILIATED HOSPITAL OF SUN YAT SEN UNIV
- Filing Date
- 2025-12-09
- Publication Date
- 2026-05-01
AI Technical Summary
Current methods for developing peritoneal dialysis prescriptions rely on physician experience or fixed formulas, making it difficult to achieve personalization and standardization, and failing to effectively integrate multi-source data, resulting in a lack of credibility in the recommendations.
A peritoneal dialysis knowledge graph is constructed, integrating multi-source clinical data. Personalized prescriptions are generated by calculating the combination of elements in a personalized prescription through graph traversal and rule-based reasoning.
It achieves a balance between personalized and standardized recommendations, improves the accuracy and reliability of prescriptions, and solves the problems of subjective physician experience and insufficient utilization of unstructured data.
Smart Images
Figure CN121964047A_ABST
Abstract
Description
A knowledge graph-based method and system for personalized peritoneal dialysis prescription recommendation Technical Field
[0001] This invention relates to the field of medical information technology, and in particular to a method and system for recommending personalized prescriptions for peritoneal dialysis based on knowledge graphs. Background Technology
[0002] Peritoneal dialysis (PD) is a commonly used alternative treatment for patients with end-stage renal disease (ESRD), and its application is widespread. However, existing prescription formulation methods have several shortcomings. Currently, most prescription designs rely on clinicians' experience or estimation methods based on fixed formulas, making it difficult to achieve a balance between personalization and standardization. Specifically: First, the subjectivity of physicians' experience leads to individual differences in prescription design, making it difficult to standardize. Second, traditional formulas or rule models only consider some clinical parameters and cannot fully adapt to complex individual differences and dynamic changes in disease course. Third, clinical medical records contain a large amount of unstructured text information (such as follow-up records and nursing logs), and existing methods cannot effectively integrate this multi-source data, resulting in insufficient information utilization. Fourth, existing intelligent recommendation methods are mostly "black box" outputs, lacking medical knowledge constraints, making it difficult for physicians to trust the recommendations.
[0003] Therefore, there is an urgent need to provide a technical solution to address the above problems. Summary of the Invention
[0004] To address the aforementioned technical problems, this invention provides a method and system for recommending personalized peritoneal dialysis prescriptions based on knowledge graphs.
[0005] In a first aspect, the present invention provides a method for personalized prescription recommendation for peritoneal dialysis based on a knowledge graph. The technical solution of the method is as follows: constructing a peritoneal dialysis knowledge graph, wherein the peritoneal dialysis knowledge graph includes medical entities, entity attributes, and entity relationships; acquiring multi-source clinical data of the target patient, wherein the multi-source clinical data includes structured data and unstructured text data; normalizing the structured data and associating the normalized structured data with the corresponding medical entities in the peritoneal dialysis knowledge graph; extracting entities and attributes from the unstructured text data to obtain text entities and text attributes, and matching the text entities and text attributes with the corresponding medical entities in the peritoneal dialysis knowledge graph; based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph, performing graph traversal or rule reasoning to calculate the combination of prescription elements compatible with the clinical state of the target patient, thereby obtaining a candidate set of prescription elements; and generating a personalized prescription recommendation for the target patient based on the candidate set of prescription elements.
[0006] The beneficial effects of the knowledge graph-based personalized prescription recommendation method for peritoneal dialysis of the present invention are as follows: The method of the present invention integrates multi-source data by constructing a peritoneal dialysis knowledge graph, and combines graph traversal and reasoning to solve the problems of subjective doctor experience, simplified formulas and insufficient utilization of unstructured data, thereby achieving a balance between personalized and standardized recommendations and improving the accuracy and reliability of prescriptions.
[0007] Based on the above scheme, the knowledge graph-based personalized prescription recommendation method for peritoneal dialysis of the present invention can be further improved as follows.
[0008] In one alternative approach, the step of constructing a peritoneal dialysis knowledge graph includes: extracting medical entities and entity relationships from medical literature and clinical guidelines, defining corresponding entity attributes for the medical entities, and storing the medical entities, entity attributes, and entity relationships in a graph structure to generate the peritoneal dialysis knowledge graph.
[0009] In one alternative approach, the step of acquiring multi-source clinical data of the target patient includes: acquiring the structured data and the unstructured text data of the target patient; wherein the structured data includes: laboratory test indicators and physiological parameters, and the unstructured text data includes: follow-up records and nursing logs.
[0010] In one optional approach, the step of normalizing the structured data and associating the normalized structured data with the corresponding medical entity in the peritoneal dialysis knowledge graph includes: performing missing value imputation, outlier correction, and numerical standardization on the structured data to obtain the normalized structured data; and associating the normalized structured data with the corresponding first medical entity in the peritoneal dialysis knowledge graph; wherein the first medical entity includes: a patient entity and a treatment parameter entity.
[0011] In one optional approach, the step of extracting entities and attributes from the unstructured text data to obtain text entities and text attributes, and matching the text entities and text attributes with corresponding medical entities in the peritoneal dialysis knowledge graph, includes: using named entity recognition technology to identify the text entities from the unstructured text data, and using attribute extraction technology to extract the text attributes from the unstructured text data; matching the text entities and text attributes with corresponding second medical entities in the peritoneal dialysis knowledge graph; wherein the second medical entity includes: symptom entities and disease entities.
[0012] In one optional approach, the step of performing graph traversal or rule-based reasoning based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph to calculate prescription element combinations compatible with the clinical state of the target patient and obtain a candidate set of prescription elements includes: taking the patient entity as the starting node, traversing the connected symptom entities, disease entities, and treatment parameter entities along the entity relationship path to obtain complete clinical feature information of the target patient; performing rule-based reasoning based on the complete clinical feature information and applying predefined medical rules; and selecting prescription element combinations compatible with the clinical state of the target patient based on the result of the rule-based reasoning to obtain the candidate set of prescription elements.
[0013] In one alternative approach, the step of generating a personalized prescription recommendation for the target patient based on the prescription element candidate set includes: scoring each combination of prescription elements in the prescription element candidate set; and generating the personalized prescription recommendation containing dialysis fluid composition, exchange cycle, and retention time based on the prescription element combination with the highest score.
[0014] Secondly, this invention provides a personalized peritoneal dialysis prescription recommendation system based on a knowledge graph. The system's technical solution includes: a construction module, an acquisition module, an association module, a matching module, an operation module, and a generation module. The construction module is used to: construct a peritoneal dialysis knowledge graph, which includes medical entities, entity attributes, and entity relationships. The acquisition module is used to: acquire multi-source clinical data of the target patient, including structured data and unstructured text data. The association module is used to: normalize the structured data and associate the normalized structured data with the peritoneal dialysis knowledge graph. The matching module is used to: extract entities and attributes from the unstructured text data to obtain text entities and text attributes, and match the text entities and text attributes with the corresponding medical entities in the peritoneal dialysis knowledge graph; the running module is used to: perform graph traversal or rule reasoning based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph to calculate the combination of prescription elements compatible with the clinical state of the target patient, and obtain a candidate set of prescription elements; the generation module is used to: generate personalized prescription recommendations for the target patient based on the candidate set of prescription elements.
[0015] The beneficial effects of the knowledge graph-based personalized prescription recommendation system for peritoneal dialysis of the present invention are as follows: The system of the present invention integrates multi-source data by constructing a peritoneal dialysis knowledge graph, and combines graph traversal and reasoning to solve the problems of subjective doctor experience, simplified formulas and insufficient utilization of unstructured data, thereby achieving a balance between personalized and standardized recommendations and improving the accuracy and reliability of prescriptions.
[0016] Thirdly, the technical solution of an electronic device of the present invention is as follows: it includes a memory, a processor, and a program stored in the memory and running on the processor, wherein when the processor executes the program, it implements the steps of the knowledge graph-based personalized prescription recommendation method for peritoneal dialysis of the present invention.
[0017] Fourthly, the present invention provides a computer-readable storage medium with the following technical solution: the computer-readable storage medium stores instructions, and when the computer-readable storage medium reads the instructions, it causes the computer-readable storage medium to perform the steps of the knowledge graph-based personalized prescription recommendation method for peritoneal dialysis of the present invention.
[0018] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description
[0019] The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of the invention. Throughout the drawings, the same reference numerals denote the same components. In the drawings: Figure 1 is a flowchart illustrating an embodiment of a knowledge graph-based personalized peritoneal dialysis prescription recommendation method according to the present invention; Figure 2 is a structural schematic diagram illustrating an embodiment of a knowledge graph-based personalized peritoneal dialysis prescription recommendation system according to the present invention; Figure 3 is a structural schematic diagram illustrating an embodiment of an electronic device according to the present invention. Detailed Implementation
[0020] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.
[0021] Figure 1 shows a flowchart of an embodiment of a knowledge graph-based personalized peritoneal dialysis prescription recommendation method provided by the present invention. This knowledge graph-based personalized peritoneal dialysis prescription recommendation method can be executed by electronic devices such as terminal devices or servers. The terminal device can be any fixed or mobile terminal such as user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, or wearable device. The server can be a single server or a server cluster composed of multiple servers. Any electronic device can implement the knowledge graph-based personalized peritoneal dialysis prescription recommendation method by having its processor call computer-readable instructions stored in its memory. As shown in Figure 1, the method includes the following steps: S1, constructing a peritoneal dialysis knowledge graph, which includes medical entities, entity attributes, and entity relationships.
[0022] The peritoneal dialysis knowledge graph refers to a structured semantic network containing medical concepts and their relationships within the peritoneal dialysis field; for example, a database that organizes entities such as patients, diseases, symptoms, treatment parameters, and prescription elements and their relationships in a graph format. Medical entities refer to specific objects in the knowledge graph that represent key concepts in the medical field; for example, in a specific peritoneal dialysis case, medical entities may include a patient named "Patient A," the disease "End-stage renal disease," the symptom "edema," and the treatment parameter "ultrafiltration volume." Entity attributes refer to numerical or categorical information used to describe the specific characteristics of a medical entity; for example, the patient entity "Patient A" has attributes such as "Age: 55 years old" and "Gender: Male," and the treatment parameter entity "Ultrafiltration volume" has the attribute "Value: 500ml." Entity relationships refer to the associations established between medical entities under the constraints of medical knowledge; for example, in the knowledge graph, there is a "Has" relationship between the "Patient A" patient entity and the "End-stage renal disease" disease entity, and a "Clinical Manifestation" relationship between the "Edema" symptom entity and the "End-stage renal disease" disease entity.
[0023] S2. Obtain multi-source clinical data of the target patient, including structured data and unstructured text data.
[0024] The target patient refers to a specific individual requiring a personalized peritoneal dialysis prescription; for example, "Patient A" currently receiving peritoneal dialysis treatment in a hospital. Multi-source clinical data refers to a collection of patient medical information obtained from different sources; for example, including "Patient A's" laboratory test reports, physiological parameter records, follow-up interview records, and nursing observation logs. Structured data refers to standardized medical data with a fixed format and clearly defined fields; for example, "Patient A's" serum creatinine value "450 μmol / L" and blood pressure value "150 / 90 mmHg". Unstructured text data refers to free-text medical records without a preset format; for example, a doctor's note in a follow-up record stating, "The patient complains of decreased urine output today and mild edema in both lower extremities."
[0025] It should be noted that in this embodiment, the multi-source clinical data of the target patient is mapped into a vectorized representation, denoted as: ;in, Indicates age, Indicates gender. represents weight, This indicates the results of the peritoneal equilibration test (PET). Indicates serum albumin level, Indicates urine volume. Indicates ultrafiltration volume. This indicates blood pressure level.
[0026] S3. Normalize the structured data and associate the normalized structured data with the corresponding medical entities in the peritoneal dialysis knowledge graph.
[0027] The normalized structured data refers to structured data that has undergone standardization to eliminate the influence of dimensions; for example, converting serum creatinine value from "450 μmol / L" to a standardized fraction "2.8", and blood pressure value from "150 / 90 mmHg" to a standard normal distribution value "1.5".
[0028] S4. Extract entities and attributes from the unstructured text data to obtain text entities and text attributes, and match the text entities and text attributes with the corresponding medical entities in the peritoneal dialysis knowledge graph.
[0029] Here, text entities refer to specific medical concepts identified from unstructured text; for example, entities such as "decreased urine output" and "bilateral lower extremity edema" identified from follow-up records. Text attributes refer to entity feature information extracted from unstructured text; for example, attributes such as "location: bilateral lower extremities" and "severity: mild" of edema extracted from "mild bilateral lower extremity edema".
[0030] S5. Based on the associated and matched medical entities and their corresponding entity relationships in the peritoneal dialysis knowledge graph, perform graph traversal or rule-based reasoning to calculate the combination of prescription elements compatible with the clinical state of the target patient, and obtain a candidate set of prescription elements.
[0031] Graph traversal refers to the process of searching for relevant nodes in a knowledge graph along a specific path; for example, starting from the node "Patient A", the node "End-stage renal disease" is found along the "Has" relationship, and then the node "edema" is found along the "Clinical manifestations" relationship. Rule-based reasoning refers to the process of deducing based on predefined medical logic; for example, applying the rule "IF edema severity THEN recommends high-concentration glucose dialysate" for logical reasoning. Prescription element combination refers to the specific configuration of the various therapeutic elements constituting a peritoneal dialysis prescription; for example, a combination of elements including "dialysis glucose concentration: 2.5%", "daily exchange frequency: 4 times", and "retention time per exchange: 4 hours". Prescription element candidate set refers to the set of multiple prescription element combinations that meet the conditions obtained through calculation; for example, a set including the above prescription element combination and two other feasible combinations with different glucose concentrations and exchange frequencies.
[0032] S6. Generate a personalized prescription recommendation for the target patient based on the candidate set of prescription elements.
[0033] Personalized prescription recommendations refer to customized treatment plans generated for specific patients; for example, a complete prescription generated for "Patient A" that includes specific dialysis fluid components, exchange cycles, and retention time.
[0034] The technical solution in this embodiment integrates multi-source data by constructing a peritoneal dialysis knowledge graph, and combines graph traversal and reasoning to solve the problems of subjective doctor experience, simplified formulas, and insufficient utilization of unstructured data. It achieves a balance between personalized and standardized recommendations, and improves the accuracy and reliability of prescriptions.
[0035] In one optional approach, S1 specifically includes: extracting medical entities and entity relationships from medical literature and clinical guidelines, defining corresponding entity attributes for the medical entities, and storing the medical entities, entity attributes, and entity relationships in a graph structure to generate the peritoneal dialysis knowledge graph.
[0036] Medical literature refers to publicly published materials that record professional medical knowledge; for example, the relevant content on peritoneal dialysis prescriptions in the *Clinical Practice Guidelines for Peritoneal Dialysis* and the *Journal of Nephrology*. Clinical guidelines refer to normative documents on medical practice formulated by professional institutions; for example, the *Peritoneal Dialysis Operating Procedures* published by the Chinese Society of Nephrology. Graph structures refer to data storage formats composed of nodes and edges; for example, using the Neo4j graph database to store nodes such as patients, diseases, and symptoms, and their relationships.
[0037] It should be noted that, in addition to medical literature and clinical guidelines, medical entities can also be extracted from historical case data. These medical entities include, but are not limited to: ① patient characteristics (age, gender, weight, peritoneal transport type); ② treatment parameters (dialysis fluid glucose concentration, perfusion volume, number of exchanges); ③ clinical indicators (solute clearance rate Kt / V, serum albumin, urine output, ultrafiltration volume); ④ complications (peritonitis, peritoneal sclerosis, heart failure).
[0038] Among the above-mentioned optional approaches, a structured peritoneal dialysis knowledge graph can be constructed by integrating medical entities and relationships from medical literature and clinical guidelines, providing a knowledge foundation for subsequent data association and reasoning.
[0039] In one alternative approach, S2 specifically includes: acquiring the structured data and unstructured text data of the target patient; wherein the structured data includes: laboratory test indicators and physiological parameters, and the unstructured text data includes: follow-up records and nursing logs.
[0040] Laboratory test indicators refer to physiological parameters obtained through laboratory testing; for example, "Patient A's" serum creatinine "450 μmol / L" and blood urea nitrogen "15.2 mmol / L". Physiological parameters refer to measurable indicators reflecting physiological function; for example, "Patient A's" blood pressure "150 / 90 mmHg" and weight "65 kg". Follow-up records refer to medical documents generated from regular follow-up visits to patients; for example, written records of "Patient A's" recent changes in symptoms and quality of life. Nursing logs refer to daily observation information recorded by nursing staff; for example, nursing documents recording "Patient A's" daily fluid intake and output and edema status.
[0041] Among the above-mentioned optional methods, further efforts can be made to obtain structured and unstructured text data of target patients, integrate patient laboratory indicators, physiological parameters, follow-up records and nursing logs, and provide comprehensive multi-source data support for personalized prescription recommendations.
[0042] In one optional approach, S3 specifically includes: S31, performing missing value imputation, outlier correction, and numerical standardization on the structured data to obtain the normalized structured data.
[0043] Missing value imputation refers to the process of reasonably supplementing missing values in data; for example, using the mean of historical data of the same patient to impute the currently missing serum creatinine value. Outlier correction refers to correcting values that deviate significantly from the normal range; for example, correcting a clearly erroneous blood pressure value "300 / 150 mmHg" to a reasonable value inferred from historical data. Numerical standardization refers to data processing that converts numerical values to a uniform dimension; for example, converting various laboratory indicators into standard scores with a mean of 0 and a variance of 1.
[0044] S32. Associate the normalized structured data with the corresponding first medical entity in the peritoneal dialysis knowledge graph; wherein the first medical entity includes: patient entity and treatment parameter entity.
[0045] In this context, a patient entity refers to a knowledge graph node representing a specific patient; for example, a knowledge graph node containing basic information about "Patient A". A treatment parameter entity refers to a knowledge graph node representing treatment-related parameters; for example, a knowledge graph node representing the "ultrafiltration rate" parameter.
[0046] Among the above-mentioned optional methods, further improvements can be made to the quality and consistency of structured data through missing value imputation, outlier correction, and numerical standardization, ensuring accurate association with the knowledge graph.
[0047] In one optional approach, S4 specifically includes: S41, using named entity recognition technology to identify the text entities from the unstructured text data, and using attribute extraction technology to extract the text attributes from the unstructured text data.
[0048] Named entity recognition technology refers to technical methods for identifying medical concepts from text; for example, using the BERT model to identify entities such as "edema" and "decreased urine output" from follow-up records. Attribute extraction technology refers to technical methods for extracting entity features from text; for example, using rule matching to extract the "location" and "degree" attributes from "mild edema of both lower limbs".
[0049] S42. Match the text entity and the text attribute with the corresponding second medical entity in the peritoneal dialysis knowledge graph; wherein, the second medical entity includes: symptom entity and disease entity.
[0050] In this context, symptom entities refer to knowledge graph nodes that represent a patient's clinical manifestations; for example, a knowledge graph node representing the symptom of "edema." Disease entities refer to knowledge graph nodes that represent a specific disease; for example, a knowledge graph node representing the disease "end-stage renal disease."
[0051] Specifically, a large language model is used to parse unstructured text data from patients. Named entity recognition technology is then used to identify text entities from the unstructured text data using the large language model, and attribute extraction technology is employed to extract text attributes. The extracted set of text entities is denoted as... Each text entity To represent a medical concept, for example, from the follow-up record "patient complains of decreased urine output and mild edema in both lower extremities," we can identify "decreased urine output" and "edema in both lower extremities" as text entities; the extracted set of text attributes is denoted as... Each text attribute Features describing entities, such as extracting "location: bilateral lower limbs" and "severity: mild" as text attributes from "mild edema of both lower limbs"; constructing feature vectors based on the extracted text entities and text attributes. Each feature corresponds to a numerical representation of a text attribute; the text feature vector... Compared with existing patient feature vectors The data is fused to form an enhanced patient feature representation; text entities and text attributes are matched with corresponding second medical entities in the peritoneal dialysis knowledge graph, where the second medical entities include symptom entities and disease entities. For example, the text entity "edema" is matched with the symptom entity in the knowledge graph, and the text entity "end-stage renal disease" is matched with the disease entity in the knowledge graph.
[0052] In the above-mentioned optional methods, unstructured text data is further processed through named entity recognition and attribute extraction techniques to extract entities such as symptoms and diseases and their attributes, and then matched with knowledge graphs to fully explore patients' clinical information.
[0053] In one optional approach, S5 specifically includes: S51, taking the patient entity as the starting node, traversing the connected symptom entity, disease entity, and treatment parameter entity along the entity relationship path to obtain the complete clinical characteristic information of the target patient.
[0054] Among them, complete clinical feature information refers to a comprehensive description of the patient's clinical status after integrating multi-source data; for example, integrating all relevant information on "Patient A's" laboratory indicators, symptoms, and treatment history.
[0055] S52. Based on the complete clinical feature information, apply predefined medical rules to perform rule reasoning.
[0056] Among them, predefined medical rules refer to rules that are set in advance based on medical knowledge; for example, business rules such as "IF insufficient ultrafiltration AND edema THEEN increase dialysate glucose concentration".
[0057] S53. Based on the result of the reasoning according to the rules, select the combination of prescription elements that is compatible with the clinical status of the target patient to obtain the candidate set of prescription elements.
[0058] Specifically, based on the associated and matched medical entities and their corresponding entity relationships in the peritoneal dialysis knowledge graph, graph traversal and rule-based reasoning are performed. Starting with the patient entity as the initial node, the system traverses the connected symptom entities, disease entities, and treatment parameter entities along the entity relationship path to obtain complete clinical characteristic information of the target patient. Based on this complete clinical characteristic information, predefined medical rules are applied for logical reasoning. According to the results of the rule-based reasoning, combinations of prescription elements compatible with the target patient's clinical state are selected to obtain a candidate set of prescription elements. .
[0059] Among the above-mentioned optional methods, a combination of graph traversal and rule-based reasoning is further used to start from the patient entity and filter the combination of prescription elements that are compatible with the patient's state along the entity relationship path, thereby improving the accuracy and logic of the reasoning.
[0060] In one alternative approach, S6 specifically includes: S61, scoring each combination of prescription elements in the candidate set of prescription elements.
[0061] S62. Based on the highest-scoring combination of prescription elements, generate the personalized prescription recommendation that includes the dialysis fluid composition, exchange cycle, and retention time.
[0062] The components of the dialysis fluid refer to the chemical composition of the peritoneal dialysis fluid; for example, the specific components such as glucose concentration and calcium ion concentration. The exchange cycle refers to the regularity of peritoneal dialysis fluid replacement; for example, a specific schedule of "exchanging every 6 hours". The retention time refers to the duration the dialysis fluid remains in the peritoneal cavity; for example, a specific duration of "retention for 4 hours each time".
[0063] Specifically, for each combination of prescription elements in the candidate prescription element set Calculate utility function ,in Represents the feature vector of the target patient. This represents a knowledge graph of peritoneal dialysis. , , These are the weighting coefficients; solved by... Select the combination of prescription elements that yields the highest utility function value; generate a personalized prescription recommendation based on the selected combination of prescription elements, including dialysate composition, exchange cycle, and retention time. Indicates the combination of prescription elements A function for evaluating the treatment effect on the target patient P; This represents the function for assessing the tolerance of the target patient P to the treatment regimen; Represents the combination of prescription elements based on knowledge graph G. Risk assessment function; Indicates the combination of prescription elements The combined utility function relative to the target patient P and the knowledge graph G; This represents the optimization process of selecting the optimal combination of prescription elements from the candidate set by maximizing the utility function.
[0064] Among the above-mentioned optional methods, the optimal combination of prescription elements can be further selected based on a scoring mechanism to generate a complete personalized prescription recommendation that includes dialysate composition, exchange cycle and retention time, thereby enhancing the reliability and practicality of the results.
[0065] To better illustrate the technical solution of this embodiment, the following complete example is used for explanation. Specifically: S10: Construct a peritoneal dialysis knowledge graph, extract medical entities and entity relationships from medical literature and clinical guidelines, define corresponding entity attributes for medical entities, and store medical entities, entity attributes, and entity relationships in a graph structure; for example, medical entities include a patient entity named Patient A, an end-stage renal disease entity, a symptom edema entity, and a treatment parameter ultrafiltration volume entity. Entity attributes include Patient A's age of 55 years and gender as male. Entity relationships include the "having" relationship between the patient entity and the disease entity, and the clinical manifestation relationship between the symptom entity and the disease entity; S20: Obtain multi-source clinical data of the target patient A, including structured data and Unstructured text data; structured data includes laboratory test indicators such as serum creatinine (450 μmol / L) and physiological parameters such as blood pressure (150 / 90 mmHg); unstructured text data includes follow-up records of the patient's complaints of decreased urine output and mild edema in both lower extremities, and nursing log records of daily fluid intake and output; S30: Normalize the structured data, including imputing missing values, correcting outliers, and standardizing numerical data to obtain normalized structured data; associate the normalized structured data with the corresponding first medical entity in the peritoneal dialysis knowledge graph, which includes the patient entity and the treatment parameter entity; S40: Extract entities and attributes from the unstructured text data, using name... Named entity recognition technology identifies text entities from unstructured text data and uses attribute extraction technology to extract text attributes; the resulting text entity set includes decreased urine output and bilateral lower extremity edema, and the text attribute set includes location (bilateral lower extremities) and severity (mild); the text entities and text attributes are matched with corresponding second medical entities in the peritoneal dialysis knowledge graph, which includes symptom entities and disease entities; S50: Based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph, graph traversal and rule reasoning are performed; starting with the patient entity as the starting node, the connected symptom entities, disease entities, and treatment parameter entities are traversed along the entity relationship path to obtain complete clinical feature information of the target patient A; based on the complete clinical feature information, the following is applied... Logical reasoning is performed using predefined medical rules, including IF insufficient ultrafiltration AND edema THEN increase dialysate glucose concentration. Based on the reasoning results, prescription element combinations compatible with the clinical state of target patient A are selected, resulting in a candidate set of prescription elements. S60: Each prescription element combination in the candidate set is scored, and a utility function value is calculated. The utility function is defined as a comprehensive evaluation of the prescription element combination relative to the target patient's feature vector and peritoneal dialysis knowledge graph. The prescription element combination with the highest score is selected by maximizing the utility function value. Based on the selected prescription element combination, a personalized prescription recommendation is generated, including a dialysate glucose concentration of 2.5%, an exchange cycle of 6 hours, and a retention time of 4 hours.
[0066] Figure 2 shows a schematic diagram of an embodiment of a knowledge graph-based personalized prescription recommendation system for peritoneal dialysis provided by the present invention. As shown in Figure 2, the system 200 includes: a construction module 201, an acquisition module 202, an association module 203, a matching module 204, a running module 205, and a generation module 206; the construction module 201 is used to: construct a peritoneal dialysis knowledge graph, which includes medical entities, entity attributes, and entity relationships; the acquisition module 202 is used to: acquire multi-source clinical data of the target patient, which includes structured data and unstructured text data; the association module 203 is used to: normalize the structured data and match the normalized structured data with the corresponding medical entities in the peritoneal dialysis knowledge graph. The matching module 204 is used to: extract entities and attributes from the unstructured text data to obtain text entities and text attributes, and match the text entities and text attributes with the corresponding medical entities in the peritoneal dialysis knowledge graph; the running module 205 is used to: perform graph traversal or rule reasoning based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph to calculate the combination of prescription elements compatible with the clinical state of the target patient, and obtain a prescription element candidate set; the generation module 206 is used to: generate personalized prescription recommendations for the target patient based on the prescription element candidate set.
[0067] In one alternative approach, the construction module 201 is specifically used to: extract medical entities and entity relationships from medical literature and clinical guidelines, define corresponding entity attributes for the medical entities, and store the medical entities, entity attributes, and entity relationships in a graph structure to generate the peritoneal dialysis knowledge graph.
[0068] In one optional manner, the acquisition module 202 is specifically used to: acquire the structured data and the unstructured text data of the target patient; wherein the structured data includes: laboratory test indicators and physiological parameters, and the unstructured text data includes: follow-up records and nursing logs.
[0069] In one optional manner, the association module 203 is specifically used to: impute missing values, correct outliers, and standardize the numerical values of the structured data to obtain the normalized structured data; and associate the normalized structured data with the corresponding first medical entity in the peritoneal dialysis knowledge graph; wherein the first medical entity includes: a patient entity and a treatment parameter entity.
[0070] In one optional manner, the matching module 204 is specifically used to: identify the text entity from the unstructured text data using named entity recognition technology, and extract the text attribute from the unstructured text data using attribute extraction technology; and match the text entity and the text attribute with the corresponding second medical entity in the peritoneal dialysis knowledge graph; wherein the second medical entity includes: symptom entity and disease entity.
[0071] In one optional manner, the running module 205 is specifically configured to: take the patient entity as the starting node, traverse the connected symptom entities, disease entities, and treatment parameter entities along the entity relationship path to obtain complete clinical feature information of the target patient; based on the complete clinical feature information, apply predefined medical rules to perform rule reasoning; and according to the result of the rule reasoning, filter out prescription element combinations that are compatible with the clinical state of the target patient to obtain the prescription element candidate set.
[0072] In one alternative approach, the generation module 206 is specifically configured to: score each combination of prescription elements in the candidate set of prescription elements; and generate the personalized prescription recommendation containing dialysate components, exchange cycle, and retention time based on the highest-scoring combination of prescription elements.
[0073] It should be noted that the beneficial effects of the knowledge graph-based personalized peritoneal dialysis prescription recommendation system 200 provided in the above embodiments are the same as those of the knowledge graph-based personalized peritoneal dialysis prescription recommendation method described above, and will not be repeated here. Furthermore, the system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the system can be divided into different functional modules according to the actual situation to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, and will not be repeated here.
[0074] The knowledge graph-based personalized peritoneal dialysis prescription recommendation system 200 of the present invention can be a computer program (including program code) running on a computer device. For example, the knowledge graph-based personalized peritoneal dialysis prescription recommendation system 200 of the present invention is an application software that can be used to execute the corresponding steps in the knowledge graph-based personalized peritoneal dialysis prescription recommendation method of the present invention.
[0075] In some embodiments, the knowledge graph-based personalized peritoneal dialysis prescription recommendation system 200 of the present invention can be implemented in a combination of hardware and software. As an example, the knowledge graph-based personalized peritoneal dialysis prescription recommendation system 200 of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the knowledge graph-based personalized peritoneal dialysis prescription recommendation method of the present invention. For example, the processor in the form of a hardware decoding processor can adopt one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.
[0076] The modules described in the embodiments of this invention can be implemented in software or hardware. The names of the modules are not, in some cases, limiting the scope of the module itself.
[0077] An electronic device according to an embodiment of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-mentioned knowledge graph-based personalized peritoneal dialysis prescription recommendation methods. That is, an electronic device according to an embodiment of the present invention may include, but is not limited to: a processor and a memory; the memory is used to store the computer program; the processor is used to execute the knowledge graph-based personalized peritoneal dialysis prescription recommendation method shown in any embodiment of the present invention by calling the computer program.
[0078] In one optional embodiment, an electronic device is provided, as shown in FIG3. The electronic device 4000 shown in FIG3 includes a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present invention.
[0079] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.
[0080] Bus 4002 may include a pathway for transmitting information between the aforementioned components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 4002 can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent bus 4002 in Figure 3, but this does not mean that there is only one bus or one type of bus.
[0081] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.
[0082] The memory 4003 stores application code (computer program) for executing the present invention, and its execution is controlled by the processor 4001. The processor 4001 executes the application code stored in the memory 4003 to implement the content shown in the foregoing method embodiments.
[0083] Among them, electronic devices can also be terminal devices. A terminal device can be any terminal device that can install applications and access web pages through applications, including at least one of smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, smart TVs, and smart in-vehicle devices.
[0084] It should be noted that the electronic device shown in Figure 3 is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.
[0085] An embodiment of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the above-mentioned knowledge graph-based personalized peritoneal dialysis prescription recommendation methods.
[0086] Alternatively, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.
[0087] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned knowledge graph-based personalized peritoneal dialysis prescription recommendation method.
[0088] Computer program code for performing the operations of this invention can be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, and C++, and conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0089] It should be understood that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0090] The computer-readable storage medium provided in this invention can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
[0091] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.
[0092] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.
[0093] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.
[0094] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this invention can be specifically implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, this invention can also be implemented as a computer program product contained in one or more computer-readable media, which includes computer-readable program code.
[0095] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.
Claims
1. A method for personalized peritoneal dialysis prescription recommendation based on knowledge graph, characterized in that, include: A peritoneal dialysis knowledge graph is constructed, which includes medical entities, entity attributes, and entity relationships. Acquire multi-source clinical data of the target patient, including structured data and unstructured text data; normalize the structured data and associate the normalized structured data with the corresponding medical entities in the peritoneal dialysis knowledge graph; The unstructured text data is subjected to entity and attribute extraction to obtain text entities and text attributes. The text entities and text attributes are then matched with corresponding medical entities in the peritoneal dialysis knowledge graph. Based on the associated and matched medical entities and their corresponding entity relationships in the peritoneal dialysis knowledge graph, graph traversal or rule-based reasoning is performed to calculate the combination of prescription elements compatible with the clinical status of the target patient, thereby obtaining a candidate set of prescription elements. Based on the candidate set of prescription elements, a personalized prescription recommendation for the target patient is generated.
2. The method for personalized peritoneal dialysis prescription recommendation based on knowledge graphs according to claim 1, characterized in that, The steps for constructing the peritoneal dialysis knowledge graph include: extracting medical entities and entity relationships from medical literature and clinical guidelines, defining corresponding entity attributes for the medical entities, and storing the medical entities, entity attributes, and entity relationships in a graph structure to generate the peritoneal dialysis knowledge graph.
3. The method for personalized peritoneal dialysis prescription recommendation based on knowledge graphs according to claim 1, characterized in that, The step of acquiring multi-source clinical data of the target patient includes: acquiring the structured data and the unstructured text data of the target patient; wherein, the structured data includes: laboratory test indicators and physiological parameters, and the unstructured text data includes: follow-up records and nursing logs.
4. The method for personalized peritoneal dialysis prescription recommendation based on knowledge graphs according to claim 1, characterized in that, The step of normalizing the structured data and associating the normalized structured data with the corresponding medical entities in the peritoneal dialysis knowledge graph includes: performing missing value imputation, outlier correction, and numerical standardization on the structured data to obtain the normalized structured data; and associating the normalized structured data with the corresponding first medical entity in the peritoneal dialysis knowledge graph; wherein the first medical entity includes: a patient entity and a treatment parameter entity.
5. The method for personalized peritoneal dialysis prescription recommendation based on knowledge graphs according to claim 4, characterized in that, The step of extracting entities and attributes from the unstructured text data to obtain text entities and text attributes, and matching the text entities and text attributes with corresponding medical entities in the peritoneal dialysis knowledge graph, includes: using named entity recognition technology to identify the text entities from the unstructured text data, and using attribute extraction technology to extract the text attributes from the unstructured text data; matching the text entities and text attributes with corresponding second medical entities in the peritoneal dialysis knowledge graph; wherein the second medical entity includes: symptom entities and disease entities.
6. The method for personalized peritoneal dialysis prescription recommendation based on knowledge graphs according to claim 5, characterized in that, The step of calculating a candidate set of prescription elements by performing graph traversal or rule-based reasoning based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph, and performing graph traversal or rule-based reasoning, includes: starting with the patient entity as the starting node, traversing the connected symptom entities, disease entities, and treatment parameter entities along the entity relationship path to obtain complete clinical feature information of the target patient; performing rule-based reasoning based on the complete clinical feature information using predefined medical rules; and selecting the candidate set of prescription elements based on the result of the rule-based reasoning to obtain the combination of prescription elements compatible with the clinical state of the target patient.
7. The method for personalized peritoneal dialysis prescription recommendation based on knowledge graphs according to any one of claims 1 to 6, characterized in that, The step of generating a personalized prescription recommendation for the target patient based on the prescription element candidate set includes: scoring each combination of prescription elements in the prescription element candidate set; and generating the personalized prescription recommendation containing dialysis fluid composition, exchange cycle, and retention time based on the prescription element combination with the highest score.
8. A knowledge graph-based personalized prescription recommendation system for peritoneal dialysis, characterized in that, include: The system comprises a construction module, an acquisition module, an association module, a matching module, an operation module, and a generation module. The construction module is used to construct a peritoneal dialysis knowledge graph, which includes medical entities, entity attributes, and entity relationships. The acquisition module is used to acquire multi-source clinical data of the target patient, including structured and unstructured text data. The association module is used to normalize the structured data and associate the normalized structured data with corresponding medical entities in the peritoneal dialysis knowledge graph. The matching module is used to extract entities and attributes from the unstructured text data to obtain text entities and text attributes, and match the text entities and text attributes with corresponding medical entities in the peritoneal dialysis knowledge graph. The operation module is used to perform graph traversal or rule-based reasoning based on the associated and matched medical entities and corresponding entity relationships in the peritoneal dialysis knowledge graph to calculate prescription element combinations compatible with the clinical state of the target patient, obtaining a prescription element candidate set. The generation module is used to generate personalized prescription recommendations for the target patient based on the prescription element candidate set.
9. An electronic device, characterized in that, The electronic device includes a processor coupled to a memory storing at least one computer program, which is loaded and executed by the processor to enable the electronic device to implement the knowledge graph-based personalized prescription recommendation method for peritoneal dialysis as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which, when executed by a processor, implements the knowledge graph-based personalized prescription recommendation method for peritoneal dialysis as described in any one of claims 1 to 7.