CN-122025085-A - Diagnosis and treatment thinking dynamic simulation training method and system based on computable physiological model

CN122025085ACN 122025085 ACN122025085 ACN 122025085ACN-122025085-A

Abstract

The invention relates to the technical field of medical simulation training, in particular to a diagnosis and treatment thinking dynamic simulation training method and system based on a computable physiological model, comprising the steps of instantiating a virtual patient model defined by a physiological state vector, a clinical manifestation vector and static parameters; receiving a diagnosis and treatment operation instruction of a user, solving the influence of dynamic calculation operation of a coupled differential equation set on a physiological state vector according to a medical knowledge rule base, generating a corresponding clinical expression vector according to an updated physiological state through a mapping function, converting the clinical expression into a multi-mode feedback signal to be output, forming an interactive training cycle, and carrying out multi-dimensional quantitative evaluation based on a complete operation sequence and a final state of a model. The system takes the computable model as a core, realizes continuous dynamic evolution and response of cases, and can effectively train and evaluate the high-order diagnosis and treatment thinking capability of clinicians.

Inventors

HU QUN
LIAO GUANGSHENG
LIAO KANGSHENG
CHEN RUIHONG
WEI WEIFENG

Assignees

深圳大学附属华南医院

Dates

Publication Date: 20260512
Application Date: 20260123

Claims (9)

1. A diagnosis and treatment thinking dynamic simulation training method based on a computable physiological model is characterized by comprising the following steps: s1, initializing a case model, namely, instantiating a virtual patient model according to the selected case type, wherein the state at the simulation time t is represented by a physiological state vector Vector of clinical manifestation Static parameter vector Co-definitions in which To describe the hidden state variable set of the internal pathophysiological processes: S2, diagnosis and treatment operation interaction, namely receiving diagnosis and treatment operation instructions input by a user The instructions include at least one of a consultation query, a physical examination, an assisted examination application, or a treatment regimen; S3, model driving state calculation, namely calculating the influence of the diagnosis and treatment operation instruction a on the virtual patient model according to a preset medical knowledge rule base, and updating the physiological state vector by solving the following coupling differential equation set To the point of Wherein, the As a function of the natural course of the disease, In order to diagnose the response function of the intervention, Is a preset medical rule; s4, dynamic clinical manifestation mapping according to the updated physiological state vector Mapping functions by predefined physiological clinical manifestations Calculating to generate corresponding clinical expression vector The mapping relation is as follows: Wherein, the Random vectors for modeling individual differences and inspection noise; S5, generating a multi-modal response, and carrying out vector analysis on the clinical manifestation The result data related to the diagnosis and treatment operation instruction a is converted into corresponding text, voice, image or physical feedback signals and is output to a user interface or an entity simulator; s6, iterating the training loop, namely repeatedly executing S2-S5 by taking the updated virtual patient model state as a new current state to form an interactive training loop; And S7, quantitatively evaluating, namely calculating multi-dimensional quantitative scores for user diagnosis and treatment logic and decision by comparing preset standard diagnosis and treatment paths based on the complete user operation sequence and the final state of the virtual patient model.
2. The method for dynamically modeling clinical thinking based on a computable physiological model according to claim 1, wherein the physiological state vector H (t) at least comprises a state variable H circ (t) representing the circulatory system, a state variable H resp (t) of the respiratory system and a state variable H inflam (t) of the inflammatory response system, and the F natural term of the coupled differential equation set encodes the pathophysiological coupling relationship between the state variables.
3. The method for dynamically simulating training the diagnosis and treatment thinking based on the computable physiological model according to claim 1, wherein when the diagnosis and treatment operation instruction a is a treatment instruction for administering a specific drug, the influence of the intervention response function F intervention on the inflammatory response system state variable H inflam (t) is realized by integrating the following pharmacodynamic models: Wherein B (t) is a pathogen carrier sub-variable in H inflam (t), C serum (t) is a blood drug concentration calculated by drug administration dosage and route through a pharmacokinetic model, and E max 、EC 50 and H are preset pharmacodynamic parameters for the drug.
4. The method for dynamically simulating training the medical thought based on the computable physiological model according to claim 1, wherein the step S5 is specific to a auscultation physical examination instruction, and the method specifically comprises the steps of calling an acoustic synthesis engine based on physical modeling according to the lung sign parameters in the clinical manifestation vector C (t+Deltaτ), generating a respiratory sound digital audio signal with corresponding pathological characteristics, and outputting the respiratory sound digital audio signal through a spatial audio rendering technology.
5. The method for dynamically simulating training clinical thinking based on the computable physiological model according to claim 1, wherein in the step S5, when the feedback carrier is a physically simulated person, the method further comprises querying a sign-driving signal mapping table according to sign data in the clinical manifestation vector C (t+Deltaτ) to generate a lower control command for controlling the physically simulated person internal executor to generate a corresponding physical sign.
6. The method for dynamically simulating training the diagnosis and treatment thinking based on the computable physiological model according to claim 1, wherein the physiological-clinical manifestation mapping function O (-) is realized through a medical knowledge graph, and the side weights of the knowledge graph define quantitative association relations between physiological state variables and clinical symptoms, physical signs and laboratory examination abnormal values and trigger thresholds.
7. The method for dynamically simulating training of clinical thinking based on computable physiological model according to claim 1, wherein in step S7, the evaluation of the time sequence of clinical logic is performed by calculating the edit distance between the user operation sequence and the standard clinical path sequence, and the evaluation of the accuracy of diagnosis is performed by comparing the matching degree between the user diagnosis conclusion and the preset diagnosis.
8. The method for dynamically simulating and training diagnosis and treatment thinking based on the computable physiological model according to claim 1, wherein the method further comprises real-time risk early warning, and the model calculation engine triggers early warning signals and is used as implicit critical sign information to be integrated into subsequent clinical manifestation mapping and feedback if any state variable is monitored to exceed a preset safety threshold in the solving process.
9. A system for dynamically simulating training of medical thought for implementing the method according to any one of claims 1-8, comprising: The model initialization module is used for loading the case template and generating an initialized virtual patient model knowledge rule base and storing the medical knowledge rule base Standard diagnosis and treatment path and physiological-clinical manifestation mapping function ; A model calculation module for executing step S3, integrated with a numerical solver, to process the coupled differential equation set; The interaction processing module is used for receiving user instructions and converting the user instructions into diagnosis and treatment operation instructions ; The feedback rendering module is used for executing the steps S4 and S5, and generating and outputting a multi-mode feedback signal; The evaluation analysis module is used for executing the step S7 and generating an evaluation report; and the system control module is used for coordinating the work of each module and managing the training circulation flow.

Description

Diagnosis and treatment thinking dynamic simulation training method and system based on computable physiological model Technical Field The invention relates to the technical field of medical simulation training, in particular to a diagnosis and treatment thinking dynamic simulation training method and system based on a computable physiological model. Background Medical simulation training is a key means for culturing the clinical thinking and decision making ability of clinicians. The current mainstream training regimen relies on high-fidelity simulation of Humans (HPS) and computer case simulation (CBS). High-fidelity simulators are capable of providing realistic physical signs and operating environments, but the development of their case scenarios typically relies on preprogrammed, fixed or limited-branch scripts. The training process is essentially advanced according to a given scenario, and the physiological response of the simulated human is a preset result, rather than being dynamically evolved from the underlying continuous pathophysiological process. This makes the training scenario highly repeatable, but not dynamic and unpredictable, and the learner can easily memorize the scenario rather than actually master the reasoning and disposition ability to deal with complex, varied conditions. Computer case simulation can cover more disease cases and is less costly, but its interaction logic is mostly based on state-jump models or simple branched narrative trees. The system jumps to the preset next plot node according to the selection of the user, and a physiological model capable of continuously and quantitatively simulating the interaction of various systems of the human body is lacking in the system. Thus, the response of the system to a trainee's irregular or erroneous intervention is often one of several possibilities that the developer has previously conceived, failing to produce an unexpectedly yet in-the-mind dynamic result based on first principles of nature (e.g., pharmacokinetics, hemodynamic equations). This makes training depth limited, difficult to truly simulate the complexities and uncertainties faced in clinical decisions. More central is that prior art solutions generally treat cases as a set of static symptoms, signs and examination results, or a limited scenario network, rather than a dynamic system driven by hidden pathophysiological states with inherent evolution laws. This results in simulated training not faithfully reflecting the natural progression of the disease in the real world, therapeutic intervention effects, and non-linear, time-varying consequences of the interaction of the two. For example, the same antibiotic, in a virtual patient with different kidney functions and different sites of infection, should have its blood concentration profile and efficacy as a result of dynamic calculations, rather than a fixed active or inactive label. The lack of the bottom layer model makes it difficult for the existing system to perform depth evaluation based on the physiological mechanism level on the diagnosis and treatment logic of the trainee. In summary, the existing medical simulation training technology lacks a unified, computable and continuously evolutionary internal physiological model, so that the existing medical simulation training technology has the fundamental limitations of case evolution scripting, physiological response statics, intervention feedback discretization and the like. This severely constrains its effectiveness in high-order clinical thinking ability, especially in dynamic decision and real-time regulatory capability culture when dealing with complex critical conditions. Therefore, a novel diagnosis and treatment thinking simulation training method based on a mechanism is developed, wherein a computable physiological model is used as a driving core, continuous evolution of states, dynamic response generation and evaluation can be realized, and the novel diagnosis and treatment thinking simulation training method has important theoretical significance and urgent practical requirements. Disclosure of Invention The invention aims to provide a diagnosis and treatment thinking dynamic simulation training method and system based on a computable physiological model, which drive the continuous and autonomous evolution of the pathophysiological state of a virtual patient through a coupled differential equation set, and map and generate a dynamic consistent multi-mode clinical manifestation, thereby realizing the deep simulation and quantitative evaluation of the mechanization of a diagnosis and treatment decision process. In order to achieve the technical purpose and the technical effect, the invention is realized by the following technical scheme: a diagnosis and treatment thinking dynamic simulation training method based on a computable physiological model comprises the following steps: s1, initializing a case model, namely, instantiating a virtual patient model ac