A follow-up method after diagnosis and treatment, a computer device, and a program product

Abstract

The specification provides a follow-up method after diagnosis and treatment, comprising: generating a target follow-up schedule for a patient according to a target follow-up plan and a first time input by the patient; wherein the first time is the time when the patient receives diagnosis and treatment services, or the time when a preset physiological behavior occurs; the target follow-up plan is determined by a doctor client of a first account and pushed to a patient client of a second account; the target follow-up schedule includes a first schedule, and the first schedule includes a first task time and a first follow-up task; within a time window related to the first task time, a first reminder about the first follow-up task is pushed to the patient, and a first feedback content submitted by the patient in response to the first reminder is received; through an evaluation model, a risk level of the patient is determined based on the first feedback content; according to the risk level, a first follow-up conclusion is generated in the second account of the patient and / or under the first account.

Description

Technical Field

[0001] This specification relates to the field of artificial intelligence technology, and in particular to a follow-up method, computer device and program product after diagnosis and treatment. Background Technology

[0002] After a patient visits a hospital for treatment such as consultation or surgery, the hospital generally needs to follow up with the patient. This follow-up includes reminding the patient to have a follow-up examination, inquiring about the patient's condition, and informing the patient of nursing care methods. These follow-up measures are used to ensure the patient's health after treatment.

[0003] In existing technologies, hospitals typically need to hire additional physician assistants to follow up with patients. These assistants create group chats that include both patients and doctors, and then conduct follow-up visits based on the follow-up plan determined by the doctor. This increases the hospital's cost for patient follow-up. Summary of the Invention

[0004] In view of the above, one or more embodiments of this specification provide a follow-up method, computer device, and program product for post-diagnosis and treatment.

[0005] According to a first aspect of one or more embodiments of this specification, a follow-up method after diagnosis and treatment is provided, comprising:

[0006] Based on the target follow-up plan and the first time input by the patient, a target follow-up schedule is generated for the patient; wherein, the first time is the time when the patient receives medical services or the time when a preset physiological behavior occurs; the target follow-up plan is determined by the doctor's client of the first account and then pushed to the patient's client of the second account; the target follow-up schedule includes a first schedule arrangement, which includes a first task time and a first follow-up task;

[0007] Within a time window related to the first task time, a first reminder about the first follow-up task is pushed to the patient, and the first feedback content submitted by the patient in response to the first reminder is received;

[0008] The patient's risk level is determined based on the first feedback content using an evaluation model.

[0009] Based on the risk level, a first follow-up conclusion is generated under the patient's second account and / or first account. According to a second aspect of one or more embodiments of this specification, a post-treatment follow-up device is provided, comprising:

[0010] The schedule generation module is used to generate a target follow-up schedule for the patient based on the target follow-up plan and the first time input by the patient; wherein, the first time is the time when the patient receives medical services or the time when a preset physiological behavior occurs; the target follow-up plan is determined by the doctor's client of the first account and then pushed to the patient's client of the second account; the target follow-up schedule includes a first schedule arrangement, which includes a first task time and a first follow-up task;

[0011] The reminder module is used to push a first reminder about the first follow-up task to the patient within a time window related to the first task time, and to receive the first feedback content submitted by the patient in response to the first reminder;

[0012] The risk level determination module is used to determine the patient's risk level based on the first feedback content using an evaluation model.

[0013] The follow-up conclusion generation module is used to generate a first follow-up conclusion based on the risk level under the patient's second account and / or first account.

[0014] According to a third aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores computer instructions thereon, which, when executed by a processor, implement the follow-up method after diagnosis and treatment as described in the first aspect of the embodiments of this specification.

[0015] According to a fourth aspect of the embodiments of this specification, a computer device is provided, the computer device comprising:

[0016] processor;

[0017] Memory used to store processor-executable instructions;

[0018] The processor executes the executable instructions to implement the post-diagnosis follow-up method as described in the first aspect of the embodiments of this specification.

[0019] According to a fifth aspect of the embodiments of this specification, a computer program product is provided, which, when executed by a processor, implements the post-diagnosis follow-up method as described in the first aspect of the embodiments of this specification.

[0020] In one or more embodiments of this specification, a target follow-up schedule is generated tailored to the patient based on their input. The schedule is then updated to include the scheduled tasks, prompting the patient to perform the corresponding follow-up tasks. During the follow-up process, the patient can provide initial feedback to the client. The client can then use an evaluation model to determine the patient's risk level based on this initial feedback and display the follow-up results to the patient and / or doctor based on the risk level.

[0021] The follow-up schedule generated using the above method only requires the doctor to push the patient's follow-up plan to the patient through the doctor's client, eliminating the need for human intervention such as a doctor's assistant, thereby reducing the cost of follow-up. Furthermore, this method can automatically determine the risk level and generate follow-up conclusions based on patient feedback, helping doctors and / or patients to understand the patient's health condition more conveniently and quickly.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this specification. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this specification and, together with the description, serve to explain the principles of this specification.

[0024] Figure 1 This is a diagram illustrating how a doctor's client application sets up group tags and follow-up plans for patients.

[0025] Figure 2 This is a diagram illustrating how a patient joins a follow-up program on a client-side basis.

[0026] Figure 3 This is a schematic diagram of an interface for patients to fill in the time on a client-side application.

[0027] Figure 4 This is a flowchart of a follow-up method after diagnosis and treatment.

[0028] Figure 5 This is a diagram illustrating how a patient's client sends a first notification.

[0029] Figure 6 This is a diagram illustrating how a patient client displays risk level assessments and / or follow-up recommendations to patients.

[0030] Figure 7 This is a schematic diagram of a control used by a patient client to interact with both the patient and doctor clients.

[0031] Figure 8 It is a hardware structure diagram of a computer device. Detailed Implementation

[0032] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.

[0033] It should be noted that the steps of the corresponding methods are not necessarily performed in the order shown and described in this specification in other embodiments. In some other embodiments, the methods may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be broken down into multiple steps in other embodiments; and multiple steps described in this specification may be combined into a single step in other embodiments.

[0034] This specification covers the use of AI agents. AI agents are generally built upon Large Language Models (LLMs). A Large Language Model, also simply called a large model, is a natural language processing model based on deep learning techniques. Its parameter count typically ranges from billions to hundreds of billions or even higher, possessing powerful language understanding and generation capabilities. Large Language Models can employ the Transformer architecture or its variants (such as GPT, BERT, etc.). This architecture utilizes an attention mechanism to globally model sequential data, efficiently handling long-distance dependencies and thus performing exceptionally well in natural language tasks. Large Language Models learn the statistical features and semantic relationships of language through pre-training on large-scale corpora, giving them good generalization capabilities. The core capabilities of Large Language Models include, but are not limited to: understanding contextual semantics, generating coherent and grammatically correct text, performing logical reasoning, and handling multi-task scenarios. Their usage typically includes two modes: direct inference and fine-tuning. In direct inference mode, the user guides the Large Language Model to generate specific outputs by designing prompts. Cue words can be task descriptions or instructions in text form, used to stimulate the semantic understanding and generation capabilities of large language models. In fine-tuning mode, large language models are further trained on small-scale datasets in specific domains to optimize their performance on specific tasks. The powerful generalization ability and flexibility of large language models make them an important tool in the field of artificial intelligence, providing efficient and accurate solutions for automated text generation and understanding.

[0035] In some embodiments, large language models can also understand and generate data from other modalities (such as visual and audio data). In this case, large language models can also be called multimodal large language models (MLLMs). MLLMs provide a richer and more natural interactive experience by integrating multiple types of input and output, such as text, images, and sound. The core advantage of MLLMs lies in their ability to process and understand information from different modalities and fuse this information to complete complex tasks. For example, MLLMs can analyze an image and generate descriptive text, or generate a corresponding image based on a text description. This cross-modal understanding and generation capability makes MLLMs widely applicable across multiple fields.

[0036] It should be noted that the key technologies of large language models can be found in the detailed description in the paper "A Survey of Large Language Models" (paper number: arXiv:2303.18223v16, published on March 11, 2025, public link: https: / / doi.org / 10.48550 / arXiv.2303.18223), and will not be repeated here.

[0037] In the embodiments described in this specification, the AI ​​agent can be deployed on a server or a client. When deployed on a client, it can be a lightweight, compressed agent.

[0038] As mentioned earlier, to reduce follow-up costs, this specification proposes a post-treatment follow-up method applied to the patient client. The patient client receives the time when the patient received medical services or a preset physiological behavior occurred (hereinafter referred to as the first time), and generates a follow-up schedule (hereinafter referred to as the target follow-up schedule) for the patient based on the follow-up plan pushed by the doctor client (hereinafter referred to as the target follow-up plan) and the first time. The follow-up schedule includes several schedule arrangements that include task times and follow-up tasks.

[0039] Follow-up plans typically include schedules at specific points in time after a patient's initial visit. For example, it might include a follow-up examination for item A on day 4 post-visit and a temperature check on day 8. To generate a follow-up schedule tailored to a specific patient, the required appointment times in the follow-up plan can be determined by obtaining the initial data, thus generating a schedule specific to that patient. For instance, if a patient enters their appointment date as January 1st, and the follow-up plan includes a follow-up examination for item A on day 4 post-visit, then the generated follow-up schedule for that patient could include a follow-up examination for item A on January 5th.

[0040] Thus, in one or more embodiments of this specification, a target follow-up schedule is generated tailored to the patient based on their input, and the patient is reminded to perform the corresponding follow-up tasks according to the task times arranged in the target follow-up schedule. During the completion of the follow-up tasks, the patient can provide initial feedback to the client. The client can determine the patient's risk level based on the initial feedback using an evaluation model, and display the follow-up results to the patient and / or doctor based on the risk level.

[0041] The follow-up schedule generated using the above method only requires the doctor to push the patient's follow-up plan to the patient through the doctor's client, eliminating the need for human intervention such as a doctor's assistant, thereby reducing the cost of follow-up. Furthermore, this method can automatically determine the risk level and generate follow-up conclusions based on patient feedback, helping doctors and / or patients to understand the patient's health condition more conveniently and quickly.

[0042] The following section will provide a detailed description of one of the follow-up methods for post-treatment care as illustrated in this instruction manual.

[0043] In one alternative implementation, the method can be accomplished through an application with chat functionality. Accordingly, during or after treatment, patients can add the doctor's contact information via scanning a QR code, searching keywords, or other methods. The doctor can then use the doctor's client to determine a target follow-up plan for the patient. The doctor performs actions on the doctor's client (e.g., selecting the target follow-up plan and patient, and clicking the send button) to push the target follow-up plan to the patient's client.

[0044] In another alternative implementation, the method can be implemented using an AI agent. The doctor can first activate the AI ​​avatar function, generating an AI agent representing the doctor. During or after treatment, patients can report to the doctor by scanning a QR code or other means. The doctor can manually receive reported patients through a doctor's client, or the doctor's client can automatically receive reported patients. After the reported patient is received by the doctor, the patient can converse with the doctor's activated AI avatar.

[0045] Similar to the methods described above, doctors can use a doctor's client to determine a target follow-up plan for a patient, and then the doctor's client can push the target follow-up plan to the patient's client.

[0046] Determining the target follow-up plan can be achieved by the physician selecting a plan that meets the patient's needs from multiple pre-set follow-up plans by the current medical team. Alternatively, the target follow-up plan can be a new plan manually generated by the physician, or a follow-up plan obtained by modifying an existing one.

[0047] Furthermore, in an alternative implementation, doctors can label patients with notes or annotations related to their medical history, such as their illness or the surgery they underwent. Figure 1 As shown, doctors can view patient information on the doctor's client and set patient tags within the patient information viewing interface. Based on the patient information, the AI ​​agent deployed or connected to the doctor's client can also generate recommended grouping tags. For example... Figure 1 Based on the patient's symptoms, the AI ​​agent recommended the tags "Appendicitis Treatment" and "Post-Appendicitis Surgery." In the accompanying diagrams, rounded corners represent clickable controls. Doctors can directly click the AI-recommended tags to group patients. Doctors can also click the "Add Tag" control to access the tag selection interface and choose the appropriate tag for the patient from existing options; or, after clicking the "Add Tag" control, doctors can manually enter the tag to add for the patient.

[0048] exist Figure 1 The interface shown also allows you to set patient follow-up plans. The doctor's client can sort multiple preset follow-up plans based on the relevance between the patient's tags and these plans. This relevance can be determined based on the correspondence between other patients' tags and their follow-up plans. For example, if most patients with the same tags as this patient are assigned to follow-up plan A, then follow-up plan A can be ranked higher.

[0049] exist Figure 1 In this context, the follow-up plan recommended by the AI ​​agent is determined based on the correlation between the labels recommended by the AI ​​agent and multiple pre-set follow-up plans. Figure 1 In the interface shown, doctors can click on the follow-up plan recommended by the AI ​​agent to select a follow-up plan for the patient. Alternatively, doctors can click on the "Add Follow-up Plan" control to select a preset follow-up plan or generate a new follow-up plan, thereby determining the follow-up plan for the patient.

[0050] Click on the doctor Figure 1 After submitting the "Submit" control, the doctor's client can determine that the follow-up plan added by the doctor for the patient is the target follow-up plan and push the target follow-up plan to the patient's client.

[0051] After the doctor's client pushes the target follow-up plan to the patient's client, the patient's client can be notified of the receipt of the follow-up plan through application reminders, or the server of the application used by the patient and doctor can be notified through SMS reminders, etc., and the patient can complete the date and other information to generate a target follow-up schedule for that patient.

[0052] Specifically, after the patient opens the application as prompted, the patient client can display an interface card, hereinafter referred to as the first card, in the chat interface between the patient and the doctor (including the chat interface with the doctor and the doctor's AI avatar). This first card may include information about the target follow-up plan and a first control, which indicates joining the target follow-up plan. Alternatively, if the patient automatically joins the target follow-up plan, the first control indicates displaying the target follow-up plan. When the patient clicks the first control, an interface for filling in the time can be displayed in the patient client. This time, referred to as the first time in the context, is typically the time the patient receives medical services or the time of the physiological behavior instructed by the doctor. This first time can then serve as the start time of the follow-up plan.

[0053] In one optional implementation of the above method, the patient needs to manually join the target follow-up plan pushed by the doctor's client. This can be done as follows: Figure 2 As shown, in the chat interface for communication between the patient and the doctor, a first card 210 is displayed. Besides the first card, the chat interface may also include Dr. Li Si's personal profile, input boxes, etc. The first card may include a first control 220 and an introduction to the target follow-up plan. The patient can check if the information in the target follow-up plan matches their own situation; if so, they can click the first control 220 to join the target follow-up plan.

[0054] Regarding the display time of the first card, it can be displayed within a preset time period after the doctor pushes the target follow-up plan to the doctor's client, provided that the patient has not set an initial display time. For example, it can be displayed within 7 days after the doctor issues the target follow-up plan, provided that the patient has not set an initial display time.

[0055] After the patient clicks the first control, in one optional implementation, an interface displaying detailed information about the target follow-up plan can be shown. This interface also displays a "Confirm Join" control; clicking this control displays an interface for entering the time period. In another optional implementation, after the patient clicks the first control, the patient client can directly display an interface for entering the time period, where the patient can enter the first time period. The time period entry interface is as follows: Figure 3 As shown, the phrase "Please confirm the surgery date" can be changed according to actual needs. For example, if the follow-up plan is a post-treatment follow-up plan, then the content here can be "Please confirm the appointment date".

[0056] The interface displaying detailed information about the target follow-up plan can include the following: follow-up name, doctor information, follow-up instructions, doctor's message, patient and follow-up task content. The follow-up instructions can include a brief introduction to the follow-up to help the patient quickly understand its purpose. For example, the follow-up instructions could be, "The postoperative management plan is a series of tasks designed for you by your doctor. After each task is completed, the doctor will provide corresponding guidance as needed." Doctor's message can be a text or voice message from the doctor. Correspondingly, the doctor's client can provide a channel for doctors to leave messages when issuing the target follow-up plan. The follow-up task content can include the nodes of each follow-up task. For example, the postoperative follow-up plan could include, "On the first day after surgery at 8:00 AM, have a blood routine check and come to my clinic for a follow-up appointment," or "On the 7th day after surgery, fill out the postoperative condition survey form," etc.

[0057] also, Figure 2 The control 230 in the "Health Center" can also be used to view the follow-up plans that the patient has currently joined and the corresponding generated follow-up schedules, as well as to view the follow-up plans that the patient is waiting to join. After the patient clicks on control 230, a list of the follow-up plans that have been joined and those that are waiting to be joined will be displayed. The patient can then view the details of the follow-up plans on this page and join a follow-up plan.

[0058] In another alternative implementation, the patient client can automatically join the target follow-up plan pushed by the doctor's client. Accordingly, the first control 220 can represent and display detailed information about the target follow-up plan. After the patient clicks the first control, the patient client can be redirected to the detailed schedule interface of the target follow-up plan, and then, after clicking the interface to confirm the schedule, be redirected to the interface for filling in the time. Alternatively, after the patient clicks the first control, the patient client can directly redirect to the interface for filling in the time.

[0059] Next, we will combine Figure 4 This section explains the patient's client-side processing method after the interface displays the entered time. For example... Figure 4 As shown, this instruction manual provides a follow-up method after diagnosis and treatment, including the following steps:

[0060] Step 410: Generate a target follow-up schedule for the patient based on the target follow-up plan and the first time input by the patient.

[0061] Wherein, the first time is the time when the patient receives medical services or the time when a preset physiological behavior occurs; the target follow-up plan is determined by the doctor's client of the first account and then pushed to the patient's client of the second account; the target follow-up schedule includes a first schedule, which includes a first task time and a first follow-up task;

[0062] First, let's clarify the concept of "first time." "First time" refers to the patient's... Figure 3 The interface shown displays the time entered. Correspondingly, the patient client can receive the first time entered by the patient.

[0063] The time granularity for the first moment can be specific to hours or minutes, or it can be as follows: Figure 3 The time shown is specific to the date. The specific format of the first time depends on the needs of the follow-up plan. For example, if the follow-up plan requires reminding patients to take their medication multiple times a day, then the first time can be specified down to the hour or minute. If the follow-up plan does not have strict time requirements, then the first time can be in the form of a date.

[0064] The specific type of time data collected as the "first time" can be set according to the needs of the follow-up plan. The follow-up plan includes the schedule after a specific time point (in other words, the "first time" can be the start time of the follow-up plan). For example, if the follow-up plan includes what to do on the first day after surgery and what to do on the second day, then the "first time" can be the surgery time.

[0065] Regarding the specific content of the "first time" (or "first time"), the time when a patient receives medical services can include the time of surgery, admission, discharge, medication, consultation, completion of radiotherapy / chemotherapy, and completion of targeted therapy. Different types of "first time" can correspond to different types of follow-up plans. For example, the time of surgery can correspond to the postoperative follow-up schedule, the time of admission can correspond to the preoperative indication follow-up schedule, and the time of discharge, completion of radiotherapy / chemotherapy, completion of targeted therapy, and consultation can correspond to the post-diagnosis follow-up schedule.

[0066] The timeframe for the anticipated physiological behavior can include the date of the last menstrual period (WMS) and the date of birth. The date of the last menstrual period can correspond to follow-up plans related to menstruation, such as preoperative indication follow-up plans. For example, a hysterosalpingography (HSG) procedure is performed 3-7 days after menstruation ends, so the preoperative indication follow-up schedule can be determined based on the date of the last menstrual period. The date of birth can correspond to postoperative follow-up plans, such as postoperative follow-up plans for newborns. These plans include educating parents on proper feeding practices, jaundice monitoring, and sleep safety, as well as reminding parents to vaccinate their newborns.

[0067] In addition, the first time period filled in by the patient can include one or more of the aforementioned time periods. When the first time period includes multiple types of time, one of these times can be the start date of the follow-up plan, while others are used to assist in scheduling the follow-up. For example, the first time period can simultaneously include the last menstrual period and the surgery date to determine the preoperative indications and follow-up schedule for hysterosalpingography (HSG).

[0068] After explaining the initial steps, we will now explain how to generate the target follow-up schedule.

[0069] As mentioned above, the target follow-up schedule includes a first schedule. In an optional implementation, the target follow-up schedule may include several schedules, each schedule including a task time and a follow-up task. The first schedule is one of the several schedules.

[0070] A target follow-up plan is a follow-up plan determined by the doctor for the patient. This plan is essentially a template, including the schedule for a certain period after a specific time (the initial follow-up). The follow-up tasks in a target follow-up plan are the same as those in a target follow-up schedule. The difference is that the schedule in a target follow-up plan does not correspond to specific times; its time is expressed in terms such as x days post-surgery or x days after treatment. In contrast, a target follow-up schedule includes specific task times, such as January 1, 2026. Correspondingly, the process of generating target follow-up tasks involves substituting the initial follow-up time into the target follow-up plan to obtain the specific task times corresponding to each follow-up task in the target follow-up schedule, thus generating the target follow-up schedule.

[0071] Regarding the implementation of this step, in one optional implementation, it can be performed by the patient client; in another optional implementation, the patient client can synchronize the target follow-up plan with the server in real time, and the server can generate the target follow-up schedule.

[0072] In one alternative implementation, if a patient has not yet joined the follow-up plan and has already missed the last scheduled time of the follow-up plan when the patient clicks the first control to join the follow-up plan, then the target follow-up schedule may not need to be generated.

[0073] In other words, the task time for the last follow-up task in the target follow-up plan can be determined based on the target follow-up plan and the first time entered by the user. The method for determining the task time is the same as the method for determining the task time of the target follow-up schedule. If the task time of the last follow-up task is later than the current time, the target follow-up schedule is generated. If the task time is earlier than the current time, a prompt can be sent to the patient indicating that they cannot join the follow-up plan. For example, the patient can be prompted that "This follow-up plan has expired and cannot be joined."

[0074] The target follow-up schedule may include several schedules, each including task time and follow-up tasks. Task time is a specific time, which can be specified down to the date or hour and minute; the specific format can be set by the doctor. Follow-up tasks are the tasks that the patient needs to complete. In an optional implementation, any schedule may include at least one of the following follow-up tasks: browsing patient education articles, completing a questionnaire, undergoing a follow-up examination and uploading the examination content and results, and viewing care reminders.

[0075] This includes browsing patient education articles, which are written by healthcare professionals to help patients better understand relevant medical knowledge and how to properly handle postoperative emergencies. Completing a survey allows patients to fill out a pre-prepared questionnaire through the patient app, reflecting their current health status. Follow-up appointments and completion of results involve patients visiting any hospital (not necessarily the one where the follow-up plan was issued) for a follow-up examination and uploading the results to the patient app via photo or manual input. Viewing care reminders can be proactively pushed to check on patients for any discomfort.

[0076] Furthermore, after generating the target follow-up schedule for this patient, in an optional implementation, such as Figure 2 As shown, a second card 240 can be displayed, indicating that the patient has joined the follow-up program.

[0077] After generating the target follow-up schedule for this patient, in another alternative implementation, if there is Figure 2 The "Health Management" control 230 shown can be used by the patient client to show how to view the target follow-up schedule through the control 230, thus guiding the patient to view the target follow-up schedule through the control 230.

[0078] Step 420: Within the time window related to the first task time, push a first reminder about the first follow-up task to the patient, and receive the first feedback content submitted by the patient in response to the first reminder.

[0079] This section uses the push method for the first schedule as an example to explain the specific methods for follow-up. It's understandable that if the target follow-up schedule includes multiple schedules, the processing methods for other schedules are the same as those for the first schedule, and will not be repeated here.

[0080] The initial notification can be sent via SMS or through a push notification in the patient-doctor chat interface. For example, in one optional implementation, it can be done as follows: Figure 5 The chat interface shown pushes a third card (510) to the patient. This third card includes the first follow-up task. The third card also includes controls to indicate whether the patient has completed the corresponding follow-up task, such as... Figure 5 The "Go to Complete" control allows patients to perform corresponding follow-up tasks, such as... Figure 5 In this example, the patient's client can be redirected to a page for uploading follow-up examination content and results, where the patient can upload examination content and result images.

[0081] In addition, the third card can be displayed within a preset time period after the first task's completion time, if the patient has not completed the first task. The third card may not be displayed after the preset time period to prevent user misunderstanding.

[0082] In other words, the process of pushing the first reminder in step 420 can specifically include: calculating the first time difference based on the current system time and the first task time; when the first time difference falls within a preset first time window, pushing the first reminder about the first follow-up task to the patient.

[0083] Figure 5 The example given is that the first follow-up task has only one item, which is a review and completion of the review content and results. It's understandable that if the first follow-up task is any other follow-up task, the third card and... Figure 5 Similar, but the content is different.

[0084] For example, if the first follow-up task is to fill out a questionnaire, the message "The doctor reminds you to have a follow-up examination and fill in the results" in line 510 can be changed to "The doctor reminds you to fill out a questionnaire." The specific follow-up content could be: filling out a questionnaire. The "Go to Complete" control could be specifically a "Go to Fill Out" control. After the patient clicks the "Go to Fill Out" control, they will be redirected to the questionnaire filling interface corresponding to this follow-up task, where they can fill out the questionnaire to complete the follow-up task.

[0085] If the first follow-up task is to view the care reminder, the message "The doctor reminds you to have a follow-up examination and fill in the results" in question 510 can be changed to "The doctor has sent you care content." The corresponding follow-up content could include: "How are you feeling now since discharge? Have you experienced fever, abdominal pain, or bloating? Please contact me immediately if you have any questions." The "Go to Complete" control can specifically be "Fill in and synchronize with doctor." This control is optional; the patient can click it or not. Clicking this control allows the patient to provide feedback on their current health status to the doctor.

[0086] If the first follow-up task is to browse patient education articles, the message "The doctor reminds you to have a follow-up examination and fill in the results" in section 510 can be changed to "The doctor reminds you to browse xxx article". The corresponding follow-up content can include browsing xxx article. The "Go to Complete" control can be specifically a "Go to Browse" control. After the patient clicks the "Go to Browse" control, they will be redirected to the reading interface of the corresponding article.

[0087] In the above method, patients can send initial feedback to the patient client, such as follow-up examination details and results (i.e., examination result images), completed survey forms (i.e., indicator data), and / or feedback from patients regarding care reminders sent by doctors (i.e., symptom description text). Correspondingly, the patient client can receive the patient's initial feedback in response to the first reminder.

[0088] The examination result images can include examination reports containing the results obtained by the patient during hospital examinations. These reports can be in text format, such as a complete blood count report. They can also be a combination of text and images, such as an ultrasound report. For ease of subsequent processing, optical character recognition (OCR) methods can be used to recognize the text on the examination result images.

[0089] In one optional implementation, after receiving the first feedback content, the patient's client can push the entire first feedback content to the doctor's client. In another optional implementation, to reduce the doctor's workload, the patient's client or the server can analyze the first feedback content to determine whether there are any security risks, and push the first feedback content to the doctor's client if there are.

[0090] Step 430: Determine the patient's risk level based on the first feedback content using the evaluation model.

[0091] An evaluation model can be a pre-defined set of rules for assessing risk levels. For example, it could include evaluation rules for various types of initial feedback. By using an evaluation model, a user's risk level can be assessed more accurately, resulting in a more precise risk assessment outcome.

[0092] In terms of specific implementation, patients can input their initial feedback through a chat interface, which serves as a chat window between the patient and the doctor's AI avatar (i.e., an AI agent). The patient's client can deploy or connect to the AI ​​agent, and can automatically send a query to the AI ​​agent containing the patient's input and a first prompt word. This first prompt word can instruct the AI ​​agent (e.g., deployed on a server or doctor's end) to analyze the patient's input using a locally preset evaluation model.

[0093] This section will use a specific example to illustrate the method for determining risk levels; this method is not intended to limit the scope of this specification.

[0094] Specifically, in this method, the agent can evaluate multiple types of first feedback content separately, and generate a risk level based on the evaluation results of each type of first feedback content. For example, the agent uses an evaluation model to determine an indicator score for any indicator based on the corresponding scoring rules and the indicator value; the indicator value of the arbitrary indicator is either an arbitrarily entered indicator value or an indicator value identified through the examination result image; and / or, the agent performs semantic analysis on the symptom description to obtain a symptom score; and determines the patient's risk level based on the symptom score and / or the indicator scores corresponding to the several indicators.

[0095] As mentioned earlier, text can be identified from the examination result image, and the indicator values ​​of each indicator can be determined based on the identified text. Optionally, for examination result images that include both images and text, the text can be used as symptom description text and also input into the agent.

[0096] Specifically, any indicator The corresponding indicator score can be denoted as ,in Indicators The corresponding scoring rules, Indicators The corresponding indicator values. The agent can obtain the indicator scores for each indicator through the evaluation model. When designing the evaluation model, scoring rules adaptable to the data type of the indicator's value can be defined.

[0097] In one implementation method, the indicator value (hereinafter referred to as the first indicator for ease of distinction) is a numerical type, and the scoring rule corresponding to the first indicator can be designed as a piecewise function. When determining the score of the first indicator, the score can be determined by segmenting the indicator value according to the domain of the piecewise function. Taking a follow-up plan for appendicitis surgery as an example, and a complete blood count as an indicator, the indicator value can be the white blood cell (WBC) level, which can be used to determine whether the user is infected. The scoring rule for this indicator can be defined as follows:

[0098]

[0099] The scoring rule defines a segmented function with four domain segments. Based on the index value provided by the patient for the index, the domain segment in which the patient falls can be determined, thereby obtaining the index score corresponding to the index.

[0100] According to one implementation method, the indicator value (hereinafter referred to as the second indicator for ease of distinction) is of numerical type, and the scoring rule corresponding to the second indicator can be designed as a piecewise function. When determining the score of the second indicator, the indicator values ​​of the second indicator recorded by the patient in two first follow-up tasks can be compared to obtain the comparison result. Then, based on the segmentation of the comparison result within the domain of the piecewise function, the score of the second indicator is determined. Taking the indicator PSA as an example, during the execution of the follow-up plan, it is necessary to monitor the increase in indicator A. Therefore, the increase in indicator A can be obtained by subtracting the indicator value of indicator A provided by the patient in the previous follow-up task from the indicator value of indicator A provided by the patient in the current follow-up task. The scoring rule for this indicator can be defined as follows:

[0101]

[0102] This represents the increase in the value of indicator A. The domain of the piecewise function defined by this scoring rule includes three domain segments. Based on the increase in the indicator value provided by the patient for this indicator, the domain segment into which the patient falls can be determined, thereby obtaining the corresponding indicator score.

[0103] For data recorded in natural language text based on patient complaints, taking symptom descriptions as an example, the intelligent agent can perform semantic analysis on the symptom descriptions to obtain symptom scores. .

[0104] For example, with This refers to the symptom description provided by the patient, recorded in natural language. This represents n predefined symptom evaluation indicators, each symptom evaluation indicator Includes text description (e.g., bloating and vomiting, mild bloating) and a corresponding score. . The text vectorization function can be implemented based on models such as BERT, mapping text into a high-dimensional semantic vector. This represents a vector similarity calculation function, which can be based on cosine similarity, for example. Thus, symptom description... Symptom score The calculation process can be represented as follows:

[0105] Calculate symptom description With each symptom evaluation index Text description Semantic similarity:

[0106]

[0107] By determining the index k corresponding to the text description with the highest semantic similarity, the correlation with the symptom description is established. Most relevant symptom assessment indicators :

[0108]

[0109] in, This indicates that the parameter index that maximizes the function value is returned.

[0110] Obtain symptom evaluation indicators Corresponding score As a symptom description Corresponding symptom scores , shown as:

[0111]

[0112] Finally, after obtaining the patient's symptom score and the corresponding index scores of the aforementioned indicators, the agent can determine the patient's risk level accordingly.

[0113] If only one score is obtained, such as a score for only one symptom or an indicator, the risk level can be determined directly based on that score. For example, the patient's risk level can be determined based on a pre-defined correspondence between risk scores and risk levels.

[0114] If at least two scores are obtained, such as at least two symptom scores, or symptom scores and several indicator scores, in an optional implementation, the agent can determine the maximum value among the symptom scores and several indicator scores as the patient's risk score. , can be represented as:

[0115]

[0116] In another alternative implementation, given that at least two scores have been obtained, the agent can perform a weighted summation of the symptom scores and the scores of the plurality of indicators to obtain the patient's risk score. For example, given symptom scores and scores for several indicators, the method for determining R can be expressed as:

[0117]

[0118] in, It is the weight of the symptom score. It is the weight of the i-th indicator score, and all weights satisfy normalization, i.e. The weight of an indicator / symptom is positively correlated with the degree of association between the indicator / symptom description and the patient's disease. In other words, the closer the association between the indicator / symptom description and the patient's disease, the higher the predictive value of its indicator / symptom score, and the greater its corresponding weight. For example, for patients after appendicitis surgery, the weight of the indicator C-reactive protein is usually higher.

[0119] After obtaining R through any of the two examples above, the agent can determine the patient's risk level based on the patient's risk score R and the preset correspondence between risk scores and risk levels.

[0120] The risk score obtained by weighted summation can more comprehensively reflect the overall progress of the patient's condition and avoid the impact of accidental fluctuations of a single indicator on the assessment of the condition.

[0121] Step 440: Based on the risk level, generate a first follow-up conclusion under the patient's second account and / or first account.

[0122] The large language model can generate a text version of the follow-up conclusion based on the risk level assessment, so that patients and doctors can quickly understand the key points of the first feedback.

[0123] Regarding the presentation of follow-up results, in one optional implementation, after the follow-up results are obtained through analysis, they can be returned to the patient's client. The patient's client can then display the follow-up results to the patient in the chat interface to help the user determine whether their current physical condition is healthy and whether they need to go to the hospital for further medical treatment.

[0124] In another optional implementation, after the large language model analysis is completed, the follow-up results can be returned to the doctor's client. The doctor's client can then send a reminder to the doctor to check the follow-up results. This allows the follow-up results to be conveniently synchronized with the doctor, enabling the doctor to monitor the patient's health status.

[0125] Of course, the follow-up results can also be fed back to both the doctor and the patient so that they can understand the patient's physical condition.

[0126] Furthermore, regarding the method of presenting follow-up results, it can be determined which party to present the follow-up results to based on the patient's risk level.

[0127] For example, if the patient's risk level does not meet the preset warning conditions, such as Figure 6 As shown, risk level assessments and / or follow-up recommendations can be displayed to patients under a second account to alleviate their concerns about their health.

[0128] In addition to displaying the risk level assessment results, it can also show users that the first follow-up task has been completed. For example, Figure 5 As shown, if the user has not completed the follow-up task, the third card may include an "Incomplete" label; if the user has completed the follow-up task, such as... Figure 6 As shown, the "Incomplete" label can be changed to the "Completed" label.

[0129] Furthermore, for the two follow-up tasks of viewing care reminders and browsing patient education articles, users may not provide corresponding feedback. Therefore, if a care reminder is sent in the chat interface and the user views it, the "Incomplete" label can be directly changed to "Completed." For the follow-up task of browsing patient education articles, after the user clicks the "Go to Browse" control to view the article, the "Incomplete" label can be changed to "Completed."

[0130] When a patient's risk level meets the warning criteria, a warning alert can be sent to the doctor's client to synchronize the warning information with the doctor, and / or interactive controls can be displayed to allow the patient to contact the doctor for advice. Figure 7 As shown, Figure 7 The "Contact Doctor" control is the one that interacts with the doctor's client. Simultaneously, the doctor's client can also receive corresponding alerts and notifications.

[0131] In this way, sending early warning messages to doctors only when the patient's risk level is high can reduce the workload of doctors in confirming follow-up results.

[0132] In another alternative implementation, the patient may initiate a conversation with the doctor within the chat interface. In such conversations, the patient might mention issues that indicate a serious threat to their health. In such cases, to ensure the patient's well-being, they can report this problem to the doctor.

[0133] Specifically, the patient client can receive the first question sent by the patient to the doctor client and determine the patient's risk level based on this question. The method for determining the risk level here can be the same as the method described earlier for determining the risk level based on the first feedback. If the determined risk level meets the warning conditions, a corresponding warning reminder message is sent to the doctor client. This makes it easier for doctors to detect the warning when the patient is unwell.

[0134] In an alternative implementation, where the chat interface is a chat interface between an AI avatar of the patient and the doctor, the AI ​​agent can not only determine the patient's risk level, but also generate a response to the patient's first question and display the response to the patient so that the patient can understand whether there is any risk to their current physical condition.

[0135] Corresponding to the embodiments of the foregoing methods, this specification also provides embodiments of the apparatus and the terminal to which it is applied.

[0136] This instruction manual describes a follow-up device for post-treatment care, comprising:

[0137] The schedule generation module is used to generate a target follow-up schedule for the patient based on the target follow-up plan and the first time input by the patient; wherein, the first time is the time when the patient receives medical services or the time when a preset physiological behavior occurs; the target follow-up plan is determined by the doctor's client of the first account and then pushed to the patient's client of the second account; the target follow-up schedule includes a first schedule arrangement, which includes a first task time and a first follow-up task;

[0138] The reminder module is used to push a first reminder about the first follow-up task to the patient within a time window related to the first task time, and to receive the first feedback content submitted by the patient in response to the first reminder;

[0139] The risk level determination module is used to determine the patient's risk level based on the first feedback content using an evaluation model.

[0140] The follow-up conclusion generation module is used to generate a first follow-up conclusion based on the risk level under the patient's second account and / or first account.

[0141] The specific implementation process of the functions and roles of each module in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0142] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of the solution in this specification according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0143] like Figure 8 As shown, Figure 8 A hardware structure diagram of a computer device is shown. The device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are internally connected to each other via the bus 1050.

[0144] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification. The processor implements the above-described methods by running executable instructions.

[0145] The memory 1020 for storing processor-executable instructions can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1020.

[0146] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0147] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0148] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0149] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0150] This specification also provides a computer program product that, when executed by a processor, implements the above-described post-diagnosis follow-up method.

[0151] This specification also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described post-diagnosis follow-up method.

[0152] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0153] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0154] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0155] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

Claims

1. A follow-up method after diagnosis and treatment, comprising: Based on the target follow-up plan and the first time input by the patient, a target follow-up schedule is generated for the patient; wherein, the first time is the time when the patient receives medical services or the time when a preset physiological behavior occurs; the target follow-up plan is determined by the doctor's client of the first account and then pushed to the patient's client of the second account; the target follow-up schedule includes a first schedule arrangement, which includes a first task time and a first follow-up task; Within a time window related to the first task time, a first reminder about the first follow-up task is pushed to the patient, and the first feedback content submitted by the patient in response to the first reminder is received; The patient's risk level is determined based on the first feedback content using an evaluation model. Based on the risk level, a first follow-up conclusion is generated under the patient's second account and / or first account.

2. The method according to claim 1, wherein, The step of generating a first follow-up conclusion based on the risk level under the patient's second account and / or first account includes: In response to the risk level not meeting the warning criteria, the patient is shown a risk level assessment and / or follow-up recommendations under the second account; In response to the risk level reaching the warning conditions, the system displays a control for interacting with the doctor's client to the patient under the second account, and / or displays the corresponding warning reminder information to the doctor under the first account.

3. The method according to claim 1, wherein, The first feedback content includes at least one of the following: symptom description text, entered indicator values ​​for each indicator, and examination result images.

4. The method according to claim 1, further comprising: Show the first card; The first card includes information about the target follow-up plan and a first control, which is used to indicate joining the target follow-up plan; In response to the patient's click on the first control, an interface for filling in the time is displayed to the patient.

5. The method according to claim 1, wherein, The step of generating a target follow-up schedule for the patient based on the target follow-up plan and the first time input by the patient includes: Based on the target follow-up plan and the first time, determine the first task time for the last follow-up task in the target follow-up plan; If the task time is later than the current time, the target follow-up schedule is generated.

6. The method according to claim 1, wherein, The first time includes at least one of the following: Surgery time, admission time, discharge time, last menstrual period time, medication time, consultation time, birth time, end time of radiotherapy and chemotherapy, and end time of targeted therapy.

7. The method according to claim 1, wherein, The first follow-up task includes at least one of the following: browsing patient education articles, filling out a questionnaire, reviewing and uploading the review content and results, and browsing and viewing care reminders.

8. The method according to claim 1, wherein, The step of sending a first reminder about the first follow-up task to the patient within a time window related to the first task time includes: Calculate the first time difference based on the current system time and the time of the first task; When the first time difference falls within the preset first time window, a first reminder about the first follow-up task is pushed to the patient.

9. The method according to claim 1, further comprising: Receive a first question sent by the patient to the doctor's client, and determine the patient's risk level based on the first question; When the risk level reaches the warning condition, a warning reminder message corresponding to the patient is sent to the doctor's client.

10. A computer program product that, when executed by a processor, implements the method as described in any one of claims 1-9.

11. A computer device, comprising: processor; Memory used to store processor-executable instructions; The processor implements the method as described in any one of claims 1-9 by executing the executable instructions.

Images