An artificial intelligence-based medical instrument registration whole-process intelligent collaborative management method and system

Abstract

This invention provides an intelligent collaborative management method and system for the entire medical device registration process based on artificial intelligence. It primarily classifies and determines the medical devices to be registered based on device descriptions, pre-trained machine learning models, and a pre-built database, generating an executable registration strategy and a list of registration documents. The database stores device classification types, guidelines, registration standards, competitor information, and regulations. Registration projects are created based on the registration document list, transforming it into traceable task items with logical dependencies and status attributes, forming a visualized initial project plan for user completion. The execution process and results of each task item are monitored for compliance and status, ultimately generating a compliant registration document. This invention achieves automated, intelligent, and collaborative management of the registration process, systematically improving the efficiency, accuracy, and certainty of medical device registration.

Description

Technical Field

[0001] This invention relates to the fields of medical information technology and medical device compliance management, specifically to an intelligent collaborative management method and system for the entire medical device registration process based on artificial intelligence. Background Technology

[0002] While the regulatory systems of major global medical device markets (including China's NMPA, the US FDA, and the EU CE) share a common core principle of ensuring product safety and efficacy, they differ in specific implementation aspects such as classification rules, approval pathways, and document formats. When developing registration strategies, companies need not only to achieve deep compliance in individual markets but also to collaborate efficiently across the globe to balance speed and cost. As regulations in various countries continue to evolve and update, companies need not only to passively follow changes but also to possess forward-looking compliance prediction and planning capabilities.

[0003] Successfully registering a medical device requires collaboration across multiple departments and specialties, including regulatory, quality, clinical medicine, and R&D technologies. Currently, the industry commonly faces the following pain points:

[0004] 1. The registration preparation period has been lengthened.

[0005] The creation of medical device registration documents is a systematic project spanning the entire product lifecycle. Many key documents (such as risk management records and availability traceability documents) should ideally be generated and updated in real-time during R&D and pilot production. However, due to development delays, insufficient manpower, and inadequate registration planning, companies often only begin "reverse engineering" after registration has commenced. This not only multiplies the workload but also leads to data gaps and logical inconsistencies due to the passage of time, creating potential problems for subsequent reviews and significantly prolonging the process.

[0006] 2. High friction costs in internal collaboration and external communication

[0007] Internally, inconsistent R&D and registration standards led to repeated revisions of documents, resulting in lengthy processing times. Externally, collaborations with testing institutions and clinical centers often resulted in passive waiting and rework due to unfamiliarity with the processes, becoming an uncontrollable factor in project progress. Key decisions were discussed but not resolved due to a lack of a clear framework, further delaying project progress.

[0008] 3. Medical device registration faces a talent shortage.

[0009] Senior specialists command high annual salaries, while small and medium-sized enterprises have insufficient staffing and professional coverage, and frequent talent turnover leads to the loss of knowledge assets. At the same time, many market registrations rely on repetitive labor and lack efficient reuse tools, further exacerbating resource waste and cost pressures.

[0010] 4. Risks throughout the entire process are not visible.

[0011] Enterprises lack the ability to proactively monitor and provide early warnings for key risks such as schedule delays, resource conflicts, and compliance deviations, and can only respond passively, resulting in low project controllability.

[0012] While current market tools offer single functions such as regulatory retrieval or form management, they generally lack the systematic capability to deeply integrate intelligent planning, task-driven approaches, collaborative execution, and real-time risk monitoring throughout the entire product lifecycle. Therefore, there is an urgent need to build an integrated intelligent platform to improve efficiency, ensure precise control, and predict risks throughout the entire registration process. Summary of the Invention

[0013] In order to overcome the shortcomings of the existing technology, the purpose of this invention is to provide an intelligent collaborative management method and system for the entire process of medical device registration based on artificial intelligence.

[0014] To achieve the above-mentioned objectives of this invention, this invention provides an intelligent collaborative management method for the entire medical device registration process based on artificial intelligence, comprising the following steps:

[0015] Receive device descriptions from users for the medical devices they wish to register;

[0016] Based on the device description, pre-trained machine learning model, and pre-built database, the medical devices to be registered are classified and determined, generating an executable registration strategy and a list of registration materials; the database stores device classification types, guidelines, registration standards, competitor information, and regulations.

[0017] A registration project is created based on the registration information list, which is then transformed into traceable task items with logical dependencies and status attributes, forming a visual initial project plan for users to fill in. The execution process and results of each task item are monitored for compliance and status, and finally a compliant registration document is generated.

[0018] Optionally, the steps for generating an executable registration strategy and registration data list include:

[0019] Classification prediction: Based on the regulations and rules in the database and the machine learning model, at least one alternative classification is matched for the device description, so that the user can select the classification information of the medical device to be registered from the alternative classifications;

[0020] Backend guidance matching: Based on the device description and classification information, at least one alternative guidance and its corresponding rules that match the medical device to be registered are retrieved from the database.

[0021] Relevant standard matching: Based on the device description, classification information, and matched guidelines, candidate registration standards are retrieved from the database; the characteristics of the candidate registration standards are compared with the device characteristics to match the registration standards for the medical device to be registered.

[0022] Registration path prediction: Based on the device description and classification information, and in accordance with the national standards for determining registration paths, the machine learning model is invoked to predict the registration path;

[0023] Clinical evaluation path prediction: Based on device description and classification information, similar devices to the medical device to be registered are retrieved, and the clinical evaluation paths of the similar devices are obtained. Combining the device description and the clinical evaluation paths of the similar devices, the machine learning model is called to predict the registration path of the medical device to be registered.

[0024] Competitive analysis: Search for similar medical devices, compile information on similar medical devices, and provide a competitive analysis;

[0025] Based on the above classification prediction, background guidance principle matching, relevant standard matching, registration path prediction, clinical evaluation path prediction, and competitor analysis results, a machine learning model is invoked to generate a registration strategy and a list of registration materials.

[0026] Optionally, the initial visualization project may include a Gantt chart and / or a task board.

[0027] Optionally, compliance monitoring includes automatically identifying inconsistencies in key fields.

[0028] Optionally, during status monitoring, the system can listen to the status changes and file content of each task item in real time, dynamically calculate the project health and risk index based on preset business rules and historical project models, and proactively issue risk warnings and notifications for potential risks.

[0029] Optional steps may also include integration with external enterprise systems:

[0030] Product data is synchronized from the product lifecycle management or enterprise resource planning system via a standardized API interface, and then returned to the product lifecycle management or enterprise resource planning system as a design history file after the compliance registration document is approved.

[0031] Optionally, it also includes initiating an AI pre-screening step:

[0032] Based on historical data from supplementary cases, a simulated review is conducted on the compliance registration documents to be submitted, and a report on potential weaknesses is generated.

[0033] This application also provides an intelligent collaborative management system for the entire medical device registration process based on artificial intelligence, including:

[0034] The intelligent strategy engine receives the device description from the user for the medical device to be registered, and classifies and determines the medical device to be registered based on the pre-trained machine learning model and the pre-built database, generating an executable registration strategy and a list of registration materials.

[0035] The project management hub is used to receive the registration data list and transform it into traceable task items with logical dependencies and status attributes, forming a visual initial project plan.

[0036] The compliance execution workbench provides a context-sensitive execution environment for each task and performs compliance audits on documents received from the project management hub.

[0037] The status monitoring and risk warning module is used to monitor the status changes of task items and file content in real time, and to provide proactive risk warnings for potential risks based on preset business rules and historical project models.

[0038] Optionally, a unified data service layer is also included, which stores project metadata, task networks, and document versions, and provides real-time, consistent data access and event-driven interfaces to the intelligent strategy engine, project management hub, compliance execution workbench, status monitoring and risk warning modules through a data bus.

[0039] Optionally, it also includes an expert service docking module, which provides an interface for one-click connection to the platform's built-in expert network for online consultation or commissioned services when a risk warning is triggered or a user requests it.

[0040] Optionally, the unified data service layer is connected to the enterprise knowledge base, storing the entire process data, document templates, and approval records during the registration process in the enterprise knowledge base.

[0041] The beneficial effects of this invention are:

[0042] This invention integrates intelligent planning, task-driven processes, collaborative execution, and real-time risk monitoring, solving problems such as fragmented registration processes, low efficiency, unstable quality, and uncontrollable risks in existing technologies. It achieves automated, intelligent, and collaborative management of the registration process, systematically improving the efficiency, accuracy, and certainty of medical device registration. Specifically, this is reflected in the following aspects:

[0043] 1. Achieve "Compliance by Design": By making registration requirements a dynamic task in advance, compliance evidence is generated synchronously during the development process, eliminating the need for "post-construction" from the source, which can shorten the overall cycle by more than 30%.

[0044] 2. Reconstruct the collaboration model: Transform the data list into a visual and traceable task flow, break down departmental silos, and achieve transparent and controllable cross-departmental collaboration.

[0045] 3. Improve data quality and first-time pass rate: Through intelligent guidance, template filling and automated consistency checking, systematically reduce human error and reduce "supplementation" caused by low-level problems.

[0046] 4. Transform passive risk management into proactive risk management: Through real-time data monitoring and model prediction, issue early warnings before risks occur, support adaptive adjustments, and improve project success rates.

[0047] 5. Accumulate organizational wisdom: Transform project experience into reusable and iterative digital assets, build a unique "compliance knowledge base" for the enterprise, and reduce reliance on core talent.

[0048] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0049] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0050] Figure 1 This is a flowchart illustrating the present invention;

[0051] Figure 2 This is a schematic diagram illustrating the principle of registration strategy generation. Detailed Implementation

[0052] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0053] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0054] Example 1

[0055] like Figure 1As shown, this application provides an embodiment of an intelligent collaborative management method for the entire medical device registration process based on artificial intelligence. In this embodiment, medical devices include not only medical devices in a broad sense, but also in vitro diagnostic reagents. Before implementing this embodiment, the system is pre-trained with machine learning models, including pre-trained models based on the Transformer architecture, such as the LLM large language model; a database is pre-built, which stores device classification types, guidelines, registration standards, competitor information, regulations, and rules. The main steps include registration strategy generation and project management, as detailed below:

[0056] Users enter a description of the medical device to be registered. The device description includes the device name and the target market. For example, the device name is "cardiac pacemaker," and the target markets are China and the United States.

[0057] The system receives a device description from a user for a medical device to be registered, and classifies the device based on this description, a pre-trained machine learning model, and a pre-built database, generating an executable registration strategy and a list of registration materials. For example... Figure 2 As shown, this step includes several steps: predicting device classification, matching guidelines and relevant standards, providing registration and clinical evaluation pathways, and generating a registration strategy report based on competitor analysis. Details are as follows:

[0058] Classification prediction: Based on regulatory rules in the database and machine learning models, at least one alternative classification is matched for the device description, allowing the user to select the classification information for the medical device to be registered from these alternative classifications. For example, if the device description is "cardiac pacemaker" and the target markets are China and the United States, the matched alternative classifications would be:

[0059] 12-01-01, Active Implantable Devices - Cardiac Rhythm Management Devices - Implantable Cardiac Pacemakers, Category III (Confidence Level: 0.85). Typically consists of an implantable pulse generator and a torque wrench. Electrical pulses are applied to specific areas of the patient's heart via pacing electrodes. Used to treat chronic arrhythmias. Resynchronization therapy pacemakers can also be used to treat heart failure.

[0060] According to the explanation: the term 'cardiac pacemaker' in the context of medical device regulation defaults to the implantable type; this entry matches all the core characteristics: implantable pulse generator + electrode lead structure, used for long-term treatment of chronic arrhythmias, references the dedicated mandatory standard GB 16174.2-2015 and registration guidelines, and has a significantly higher confidence level.

[0061] 12-01-04, Active Implantable Devices - Cardiac Rhythm Management Devices - Implantable Cardiac Pacing Electrode Leads, Management Category: III (Confidence Level: 0.79). Typically consists of electrode leads and accessories. Used in conjunction with an implantable cardiac pacemaker to treat chronic arrhythmias.

[0062] According to the instructions: It is suitable for temporary external pacing scenarios (such as postoperative transition or emergency). The structure is a non-implantable pulse generator, which does not conform to the conventional meaning of 'cardiac pacemaker'; it is only applicable when the user explicitly specifies 'temporary' or 'external'.

[0063] 12-01-11, Active Implantable Devices - Cardiac Rhythm Management Equipment - Cardiac Rhythm Management Programmable Devices, Management Category: II (Confidence Level: 0.76). Typically consists of a display unit, printing unit, programming unit, and software. Used for querying, programming, displaying data, or testing cardiac rhythm management devices such as implantable pacemakers and implantable cardioverter defibrillators.

[0064] According to the instructions: It is a key component of the pacing system and needs to be used in conjunction with the pacemaker unit. It does not have pulse generation function on its own. Since the user input did not mention the attributes of accessories such as 'electrodes' and 'leads', it is not the first choice.

[0065] Additional information: The user input was 'cardiac pacemaker,' a term that typically refers to 'implantable cardiac pacemaker' in clinical and regulatory contexts (e.g., GB 16174.2-2015 and CMDE guidelines explicitly classify it as a core product). Among the candidate entries, 'implantable cardiac pacemaker' (12-01-01) had the highest confidence level (0.854), significantly higher than other options (the second highest being temporary pacemaker at 0.786, a difference of 0.068 > 0.05), indicating no ambiguous competition. Its product description (including implantable pulse generator + electrode), intended use (treatment of chronic arrhythmias), and example product name (including 'implantable cardiac pacemaker') all perfectly match. Therefore, it can be clearly classified into this category.

[0066] The user confirms the selection of category 12-01-01 from the alternative categories.

[0067] Background guidance matching: Based on the device description and classification information, at least one alternative guidance and its corresponding rules that match the medical device to be registered are retrieved from the database. For example:

[0068] The category code needs to include "12-01-01".

[0069] The disease area is cardiology;

[0070] The device is an active implantable device.

[0071] The machine learning model is called to extract device features, such as "basic.product_category: cardiovascular implantable devices" and "classification.category_code: 12-01-01".

[0072] The execution rule determination engine compares the guideline rules with the instrument features and provides matching results, such as:

[0073] Technical Review Guidelines for Registration of Implantable Cardiac Pacemakers - 2016 Revised Edition (No. 21, 2016);

[0074] Technical Review Guidelines for Registration of the Service Life of Active Medical Devices (No. 23, 2019)

[0075] Technical Guidelines for Risk Assessment of Medical Device Products (2023 Revised Edition)

[0076] Relevant standard matching: Based on the device description, classification information, and matched guidelines, candidate registration standards are retrieved from the database; the rule-based judgment engine compares the features of the candidate registration standards with the device features to match the registration standard for the medical device to be registered. For example:

[0077] {'Standard System': 'YY / T', 'Standard Number': 316, 'Year of Publication': 2016, 'Full Name': 'Application of Medical Device Risk Management to Medical Devices'};

[0078] {'Standard System': 'GB / T', 'Standard Number': 42062, 'Year of Publication': 2022, 'Full Name': 'Application of Medical Device Risk Management to Medical Devices'};

[0079] {'Standard System': 'GB / T', 'Standard Number': 16886, 'Year of Publication': 2022, 'Full Name': 'Biological Evaluation of Medical Devices'}.

[0080] Registration path prediction: Based on device description and classification information, and in accordance with national standards for determining registration paths, the machine learning model is invoked to predict the registration path. For example:

[0081] {"summary": "This medical device is classified as Class 3 and is subject to product registration management, with approval handled by the National Medical Products Administration (NMPA)", "detailed_info": {"Management Category": 3, "Registration / Filing Type": "Registration", "Regulatory Authority": "National Medical Products Administration (NMPA)", "Detailed Description": "Class 3 medical devices are subject to product registration management, with approval handled by the National Medical Products Administration"}}.

[0082] Clinical evaluation path prediction: Based on device description and classification information, similar devices are retrieved, and their clinical evaluation paths are obtained. Combining the device description and the clinical evaluation paths of similar devices, the machine learning model is invoked to predict the registration path of the medical device to be registered. For example:

[0083] {"summary": "This medical device requires clinical evaluation", "detailed_info":{"output field": "clinical_eval", "clinical evaluation requirements": "clinical trials"}}.

[0084] Competitive Analysis: This function retrieves similar medical devices, compiles information on them such as quantity, market launch date, and product characteristics, and provides a competitive analysis. Similar devices can be retrieved from a pre-built database or from external data.

[0085] Based on the above classification prediction, background guidance principle matching, relevant standard matching, registration path prediction, clinical evaluation path prediction, and competitor analysis results, a machine learning model is invoked to generate a registration strategy and a list of registration materials.

[0086] Once the registration strategy and registration materials checklist are generated, you can either export the report directly or proceed to the project management steps. The project management steps are as follows:

[0087] Based on the registration data checklist, a registration project is created, and the checklist items are broken down into traceable task items with logical dependencies and status attributes, forming a visual initial project plan for the user to fill in. Checklist items refer to the project-related lists in the registration data; one checklist may be transformed into several tasks, or several checklists may be merged into one task. For example, the registration data checklist may include a "Clinical Evaluation Report" item. The generated tasks might include: clinical trial access, clinical trial implementation (important milestones such as pre-clinical, during, and post-clinical trials), and clinical evaluation report writing. Alternatively, if the checklist lists several different risk-related documents, the corresponding generated project might only be risk management. Tasks include classification, testing, and document preparation. The visual initial project plan includes Gantt charts and / or task boards, such as writing product technical requirements and completing biocompatibility testing, and can be pre-filled based on the device description. User input includes, but is not limited to, modifying task details, adjusting dependencies, decomposing / merging tasks, estimating workload, assigning / changing responsible parties, uploading / updating deliverables, submitting for review, adding collaborators, updating progress status, revising timelines, marking risks / issues, adding comments, customizing filter views, generating progress reports, setting personal reminders, and exporting data. In this embodiment, logical dependencies refer to the objectively existing sequence and constraints between tasks, which are inherent logic determined by legal procedures, technical processes, and resource constraints. Status attributes include: progress, time, deliverables and quality, responsibility and risk, etc.

[0088] During project management, compliance and status monitoring are conducted on the processes and results of each task item, ultimately generating a compliance registration document.

[0089] Compliance monitoring includes automatically identifying inconsistencies in key fields. This means that during document completion, submission, or integration, inconsistencies in key fields are automatically identified and highlighted, for example, by marking them in red. Inconsistencies must be resolved before proceeding to the next step. Key fields include registrant information and product model.

[0090] During status monitoring, the system listens in real-time to changes in the status of each task and the content of its files. Based on preset business rules and historical project models, it dynamically calculates the project's health and risk index, and proactively issues risk warnings and notifications for potential risks. When a user modifies a task's status, the system checks the current task status and compares it with historical task statuses to calculate the current project's health and risk. When calculating the project's health and risk, it uses existing statistical learning and machine learning methods, based on the duration and dependencies of each task in historical projects, to estimate the current task's status and compare it with the actual status. In this embodiment, business rules refer to regulations, industry practices, and other searchable experiences. Risk warning types include progress delay warnings, resource conflict warnings, regulatory update impact warnings, and simulated review warnings based on historical supplementary cases. For example, if the "inspection report" task has not been uploaded by the deadline, a delay warning is sent to the responsible person. The system regularly scans for regulatory updates; if there are errata in the referenced standards, the relevant task responsible person is notified.

[0091] After all tasks are completed, all final files are integrated to generate a data package that meets the application requirements. All project data is archived to the enterprise knowledge base for intelligent recommendation of subsequent projects.

[0092] In one optional embodiment, an AI pre-review step is also included: a simulated review of the compliance registration documents to be submitted is conducted based on historical supplementary case data, and a potential weakness report is output.

[0093] In another optional embodiment, the step of integrating with external enterprise systems is also included: synchronizing product data from the product lifecycle management or enterprise resource planning system through a standardized API interface, and then returning the product lifecycle management or enterprise resource planning system as a design history file after the compliance registration document is approved.

[0094] Example 2

[0095] This invention provides an embodiment of an intelligent collaborative management system for the entire medical device registration process based on artificial intelligence. The system performs intelligent collaborative management of the entire medical device registration process based on the method described in Embodiment 1. In this embodiment, the system adopts a microservice architecture with a front-end and back-end separation. The front-end provides users with a unified user interaction module, including a web workbench and a display interface. The display interface is used to display information to the user, while the web workbench is used for user input, including a description of the medical device to be registered.

[0096] The backend comprises: an intelligent strategy engine, a project management hub, a compliance execution workbench, a status monitoring and risk warning module, a unified data service layer, and an expert service integration module. The intelligent strategy engine, acting as the "compliance brain," transforms device descriptions into executable registration strategies and registration document lists based on rules and machine learning models. The project management hub, acting as the "process engine," transforms static lists into dynamic, visual, and traceable task networks. The compliance execution workbench, acting as the "intelligent workbench," provides templates, pre-filling, and verification assistance within the task context, improving document preparation quality and efficiency. The status monitoring and risk warning module, acting as the "early warning radar," monitors the overall situation in real time, proactively identifying and issuing warnings of risks. The unified data service layer, acting as the "data hub," ensures seamless data flow and consistency between modules. A detailed description of each module follows.

[0097] The intelligent strategy engine connects to the user interaction module to receive the device description and classify the medical devices to be registered based on a pre-trained machine learning model and a pre-built database, generating an executable registration strategy and a list of registration materials. The registration strategy includes registration category, application path, estimated timeline, etc. The machine learning model in this embodiment includes a pre-trained model based on the Transformer architecture, such as the LLM large language model.

[0098] The project management hub receives the registration data list and automatically transforms it into a series of traceable task items with logical dependencies and status attributes through a task orchestration engine. This forms a visual initial project plan, which is displayed on the user interaction module. Tasks include categorization, verification, and document preparation. The visual initial project plan includes a Gantt chart and / or task Kanban board, and supports mandatory node control based on task logical dependencies. This means that tasks are logically dependent on each other, and the next task can only proceed after the previous one is completed.

[0099] The compliance execution workbench provides a context-sensitive execution environment for each task, including dynamically calling structured document templates, intelligent pre-population based on the project database, providing compliance checklists, and supporting document editing, collaborative review, and file submission. It also performs compliance audits on documents received from the project management center. In this embodiment, each task corresponds to documents that need to be prepared, and these documents have formatting and writing requirements. The compliance execution workbench generates templates for each task and pre-populates fixed content and content already determined by upstream tasks.

[0100] In this embodiment, compliance review is performed by the automatic consistency verification unit in the compliance execution workbench. Specifically, the automatic consistency verification unit automatically identifies inconsistencies in key fields during the execution process and results of each task item. Inconsistencies can be prominently marked, for example, in red, and mandatory consistency must be achieved before proceeding to the next step. Key fields include registrant information and product model.

[0101] The status monitoring and risk warning module is used to monitor the status changes and file content of task items in real time, and to proactively warn and notify potential risks based on preset business rules and historical project models.

[0102] The unified data service layer, acting as the system's data hub, stores project metadata, task networks, and document versions. Through a data bus, it provides real-time, consistent data access and event-driven interfaces to the intelligent strategy engine, project management center, compliance execution workbench, and status monitoring and risk warning modules. Connecting to the enterprise knowledge base, the unified data service layer stores all registration process data, document templates, and approval records, enabling intelligent recommendations and rapid project launch support. For example, when registering a new medical device, the system can retrieve historical similar devices based on the device information corresponding to the new project, and quickly determine and generate a template based on the prediction results and data from similar device registrations.

[0103] The expert service docking module provides an interface for one-click connection to the platform's built-in expert network for online consultation or commissioned services when a risk warning is triggered or a user requests it.

[0104] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0105] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for intelligent collaborative management of the entire medical device registration process based on artificial intelligence, characterized in that, Includes the following steps: Receive device descriptions from users for the medical devices they wish to register; Based on the device description, pre-trained machine learning model, and pre-built database, the medical devices to be registered are classified and determined, generating an executable registration strategy and a list of registration materials; the database stores device classification types, guidelines, registration standards, competitor information, and regulations. A registration project is created based on the registration information list, which is then transformed into traceable task items with logical dependencies and status attributes, forming a visual initial project plan for users to fill in. The execution process and results of each task item are monitored for compliance and status, and finally a compliant registration document is generated.

2. The intelligent collaborative management method for the entire medical device registration process based on artificial intelligence as described in claim 1, characterized in that, The steps to generate an executable registration policy and registration data list include: Classification prediction: Based on the regulations and rules in the database and the machine learning model, at least one alternative classification is matched for the device description, so that the user can select the classification information of the medical device to be registered from the alternative classifications; Backend guidance matching: Based on the device description and classification information, at least one alternative guidance and its corresponding rules that match the medical device to be registered are retrieved from the database. Relevant standard matching: Based on the device description, classification information, and matched guidelines, candidate registration standards are retrieved from the database; the characteristics of the candidate registration standards are compared with the device characteristics to match the registration standards for the medical device to be registered. Registration path prediction: Based on the device description and classification information, and in accordance with the national standards for determining registration paths, the machine learning model is invoked to predict the registration path; Clinical evaluation path prediction: Based on device description and classification information, similar devices to the medical device to be registered are retrieved, and the clinical evaluation paths of the similar devices are obtained. Combining the device description and the clinical evaluation paths of the similar devices, the machine learning model is called to predict the registration path of the medical device to be registered. Competitive analysis: Search for similar medical devices, compile information on similar medical devices, and provide a competitive analysis; Based on the above classification prediction, background guidance principle matching, relevant standard matching, registration path prediction, clinical evaluation path prediction, and competitor analysis results, a machine learning model is invoked to generate a registration strategy and a list of registration materials.

3. The intelligent collaborative management method for the entire medical device registration process based on artificial intelligence as described in claim 1, characterized in that, The initial visualization project includes a Gantt chart and / or a task board.

4. The intelligent collaborative management method for the entire medical device registration process based on artificial intelligence as described in claim 1, characterized in that, Compliance monitoring includes automatically identifying inconsistencies in key fields.

5. The intelligent collaborative management method for the entire medical device registration process based on artificial intelligence according to claim 1, characterized in that, During status monitoring, the system listens to the status changes and file content of each task item in real time, dynamically calculates the project health and risk index based on preset business rules and historical project models, and proactively issues risk warnings and notifications for potential risks.

6. The intelligent collaborative management method for the entire medical device registration process based on artificial intelligence according to claim 1, characterized in that, It also includes steps for integrating with external enterprise systems: Product data is synchronized from the product lifecycle management or enterprise resource planning system via a standardized API interface, and then returned to the product lifecycle management or enterprise resource planning system as a design history file after the compliance registration document is approved.

7. The intelligent collaborative management method for the entire medical device registration process based on artificial intelligence as described in claim 1, characterized in that, This also includes initiating an AI pre-screening process: Based on historical data from supplementary cases, a simulated review is conducted on the compliance registration documents to be submitted, and a report on potential weaknesses is generated.

8. An intelligent collaborative management system for the entire medical device registration process based on artificial intelligence, characterized in that, include: The intelligent strategy engine receives the device description from the user for the medical device to be registered, and classifies and determines the medical device to be registered based on the pre-trained machine learning model and the pre-built database, generating an executable registration strategy and a list of registration materials. The project management hub is used to receive the registration data list and transform it into traceable task items with logical dependencies and status attributes, forming a visual initial project plan. The compliance execution workbench provides a context-sensitive execution environment for each task and performs compliance audits on documents received from the project management hub. The status monitoring and risk warning module is used to monitor the status changes of task items and file content in real time, and to provide proactive risk warnings for potential risks based on preset business rules and historical project models.

9. The intelligent collaborative management system for the entire medical device registration process based on artificial intelligence as described in claim 8, characterized in that, It also includes a unified data service layer, which stores project metadata, task networks and document versions, and provides real-time and consistent data access and event-driven interfaces to the intelligent strategy engine, project management hub, compliance execution workbench and status monitoring and risk warning modules through a data bus.

10. The intelligent collaborative management system for the entire medical device registration process based on artificial intelligence as described in claim 8, characterized in that, It also includes an expert service docking module, which provides an interface for one-click connection to the platform's built-in expert network for online consultation or commissioned services when a risk warning is triggered or a user requests it.

Images