Behavioral support nudge generation and management technology based on spatial pattern progression prediction of glaucoma visual field defects.

Abstract

Regarding visual field defects caused by glaucoma, this invention provides a nudge generation system, behavioral change support method, and program that analyze visual field information previously acquired at medical institutions and retinal structure information obtained by optical coherence tomography, and continuously predict the spatial patterns of current and future visual field defects using a time-series model. [Solution] The system comprises an analysis unit 30 and a prediction unit 40 that grasp the current and future visual field impairment patterns from past data; an action context information acquisition unit 50 that acquires location information, movement information, and environmental information such as illuminance and weather; a life risk estimation unit 60 that compares the visual field impairment prediction results with the action context information to estimate individual risks in daily activities such as walking, climbing stairs, and driving; a nudge generation unit 70 that generates and presents nudges at an appropriate timing and in an appropriate manner based on the estimated risks; and a nudge presentation control unit 80. The nudge generation unit visualizes future risks due to non-compliance with treatment and provides treatment adherence support to motivate continued use of eye drops and hospital visits.

Description

Technical Field

[0001] The present invention relates to visual field disorders caused by glaucoma, and particularly to analyzing visual field information acquired in the past at medical institutions and retinal structure information (hereinafter referred to as "OCT structure information") by an optical coherence tomography (hereinafter referred to as "OCT"), and continuously predicting the spatial patterns of current and future visual field disorders using a time series model.

[0002] Furthermore, by integrating the prediction results with behavior context information (such as movement status and daily behavior patterns) and environmental information (such as weather, time of day, illuminance, etc.) for each patient, and estimating individual risks in daily behaviors such as walking, driving, ascending and descending stairs, etc., the present invention relates to a behavior modification support system and method for generating and presenting nudges that prompt attention calls and encourage compliance with treatment according to the risks.

Background Art

[0003] In Japan, it is estimated that about 5% (1 in 20) of the population aged 40 and over has glaucoma. Glaucoma is the leading cause of blindness in adults, and the number of people certified as visually impaired due to visual field disorders caused by glaucoma reaches about 16 times that of the blind. There are far more patients who have visual field disorders but have not received certification as visually impaired.

[0004] Visual field disorders caused by glaucoma have been reported to increase life risks such as falls during walking or ascending and descending stairs, and accidents during vehicle driving, and are an important issue not only in public health but also socially.

[0005] The progression of visual field disorders caused by glaucoma is slow, and in many cases, it starts from the peripheral part of the visual field and progresses in a form that avoids the fixation point, and the central visual field tends to be relatively maintained until the end stage. Furthermore, due to the mutual complementation by binocular vision and the visual compensation function of the brain, patients themselves tend to be less aware of abnormalities.

[0006] Therefore, glaucoma patients often act without realizing that they do not have sufficient vision in certain areas of their field of vision necessary for safely performing daily activities such as walking, climbing stairs, and driving.

[0007] Furthermore, because these lifestyle-related risks are often not recognized, patients may not fully grasp the importance of treatment, which can lead to decreased adherence to eye drops and treatment discontinuation.

[0008] In clinical practice, visual field testing using devices such as the Humphrey automated perimeter and OCT examinations are regularly performed as standard tests for glaucoma and are used for diagnosis and treatment planning. However, mechanisms for utilizing the results of these tests to support patients' behavior, such as avoiding risks in their daily lives and adhering to treatment, are not yet fully established.

[0009] Visual field defects have two distinct characteristics: "individualized" and "chronic, progressive central nervous system dysfunction," each presenting its own set of challenges.

[0010] First, individuality refers to the fact that the spatial pattern of visual field defects (whether the affected area is upward or downward, left or right, distance from the fixation point, extent and severity of the defect, etc.) differs from patient to patient. In addition, since patients' daily living and occupational behaviors are also individual, the impact of visual field defects on daily activities is highly individualized. However, there is no established support technology that can estimate the risks that individual visual field defect patterns pose to daily activities and encourage risk avoidance behaviors in real time while patients are performing daily activities such as walking, driving, and climbing stairs.

[0011] Next, chronic progressive central nervous system dysfunction refers to a condition in which central nervous system dysfunction progresses slowly throughout a person's life. Here, visual field testing evaluates one of the central nervous system functions, visual function, and is a type of psychophysical test that requires the subject's response. Because it is a time-consuming test to ensure accuracy and reproducibility, it has the unique characteristic of being difficult to repeat as easily as vital sign measurements. For this reason, in clinical practice, visual field testing is performed once every 4 months to 1 year (average 6 months), and the data obtained is limited to information from discrete points in the past.

[0012] On the other hand, while home-use visual field testing equipment is being researched and developed, its accuracy and reproducibility have not been fully established. Furthermore, challenges such as time constraints and the burden on the test subject remain unresolved, making it unsuitable for repeated daily use. Currently, it is only expected to serve as a supplement to visual field testing performed at medical institutions. Therefore, there are limitations to reflecting the ever-changing state of the visual field in real time.

[0013] While visual field test data from medical institutions is inevitably discrete, various statistical methods are used to analyze it, making it possible to predict future progression of visual field defects based on past trends. This is effectively utilized in clinical practice as an indicator for determining treatment strategies. However, there are no known examples of applying this visual field progression prediction technology, which uses past data to predict a patient's current visual field state, to real-time risk avoidance in daily life.

[0014] In recent years, advancements in behavioral science and information and communication technology (ICT) have led to the proposal of systems that present nudges to encourage behavioral change according to the user's behavioral factors. Patent Document 1 discloses a behavioral support technology that analyzes behavioral change techniques as constituent elements and generates messages based on psychological factors. Patent Document 2 discloses a nudge generation technology that analyzes the user's behavioral factors (attitudes, subjective norms, self-efficacy, etc.) and selects and presents messages according to their psychological state. However, both of these are limited to general behavioral change support and lack a framework for handling medical data or disease-specific physiological risks.

[0015] Meanwhile, in the medical field, technologies have been proposed that apply artificial intelligence (AI) and ICT to understand patients' behavior at home and enable intervention when risks occur. Patent document 3 discloses a system that uses AI, wearable sensors, and smart devices to analyze daily life behaviors (such as decreased activity, prolonged stays in the toilet, forgetting to take medication, and delays in climbing stairs), and provides notifications, emergency calls, or two-way communication to intervene when abnormalities are detected. Patent document 4 also discloses a technology that monitors medication compliance and provides reminders and notifications to prevent missed doses and duplicate medication, as well as providing information to healthcare professionals.

[0016] However, none of these technologies address visual field defects, which are a type of chronic, progressive central nervous system dysfunction, and they lack a framework for predicting current and future visual field conditions from past examination data. Furthermore, no technology has been disclosed that integrates diverse visual field defect patterns for each patient with daily behavioral risks in real time and presents nudges. [Prior art documents] [Patent Documents]

[0017] [Patent Document 1] Patent 7357418 Public Relations

[0018] [Patent Document 2] Japanese Patent Publication No. 2025-74283 Public Relations

[0019] [Patent Document 3] US20120092156A1 (US Published Patent)

[0020] [Patent Document 4] US20250118425A1 (US Published Patent) [Summary of the Invention] [Problems to be Solved by the Invention]

[0021] As described above, due to the characteristic that glaucoma patients are less likely to be aware of visual field impairment, they have risks in daily life activities and a risk of reduced treatment adherence. It is desirable not only to receive medical treatment at a medical institution, but also to promote behavioral changes by recognizing risks in life activities and motivating treatment in daily life.

[0022] However, none of the behavioral change support technologies disclosed in Patent Document 1 and Patent Document 2 can solve the problems of life activity risks and reduced treatment adherence caused by visual field impairment that is difficult to perceive and highly individualized like glaucoma. Also, in the remote monitoring technologies utilizing AI and ICT disclosed in Patent Document 3 and Patent Document 4, although mechanisms for risk detection, notification, and intervention in daily life have been proposed, even if these are applied to glaucoma patients, due to the reasons described below, the problems specific to these patients cannot be solved.

[0023] First, the visual field impairment in glaucoma is diverse, and the impact on the safety of the daily behavior patterns of patients, who are originally highly diverse, is highly individualized. However, the risk detection systems and medication notifications proposed in prior patents are centered around uniform notifications and cannot reflect this individuality in glaucoma.

[0024] Next, as a problem specific to glaucoma, as described above, since visual field impairment is a chronic progressive central functional disorder, there is only visual field examination data obtained at discrete points in the past, and it does not directly represent the current visual field state. Therefore, there is a problem that the above-described technology cannot accurately evaluate the real-time life behavior risks faced by patients.

[0025] There are technologies for predicting the progression of future visual field impairment from past data obtained in medical institutions. However, these are only predictions for use in determining treatment intensification in medical treatment, and technology development has not been carried out with the idea of estimating the patient's changing current visual field state in real time. That is, there is no mechanism like the present invention that continuously and immediately estimates the patient's current visual field state, integrates the result with life behavior context information, evaluates the specific behavior risks in daily life, and presents a nudge on the spot.

[0026] Furthermore, although visual field examination is an examination conducted under certain conditions in a medical institution, the environments in which patients are placed in daily life are diverse. Since the central function of the visual field has the characteristic that its function deteriorates in the dark, the actual visual field state is affected by environmental factors such as weather, time of day, and illuminance, and it is necessary to reflect such environmental information.

[0027] On the other hand, in the case of medication behavior such as eye drop treatment, although there are technologies for recording and notifying compliance status as in Patent Document 2, there is no technology for visualizing and motivating patients for the disadvantages that may occur in the future due to non-compliance with medication (such as an increased risk of falls and accidents due to the progression of visual field impairment and the risk of blindness), leaving the problem that glaucoma patients cannot realize the importance of treatment, leading to a decrease in adherence and treatment interruption.

[0028] To solve the above problems, the present invention aims to support patient safety and treatment continuation by continuously predicting current and future visual field defects using a time-series model based on past visual field data, integrating the individuality of each patient's visual field defect with behavioral context and environmental information to estimate individualized current and future life risks, and generating and presenting nudges in real time based on the results to encourage behavioral changes such as risk avoidance and adherence to treatment. [Means for solving the problem]

[0029] To achieve the above objective, the present invention relates to a nudge generation system that estimates the current and future visual field state using past visual field data and OCT structural information obtained at a medical institution, integrates the estimation results with the patient's individual behavioral context information and environmental information to calculate individualized life risks, and further generates nudges for risk avoidance and treatment encouragement based on the results, and presents them to the patient terminal at an appropriate time. The system comprises a visual field information acquisition unit 10 that acquires the patient's examination data from the medical institution's visual field testing equipment and OCT testing equipment, or electronic medical records, via a patient user terminal such as a mobile phone on which the system's dedicated application is installed; a server connected to the user terminal via a wide-area network that collects the acquired examination data via the user terminal, and a series of analysis units within the server that process the data to automatically analyze the spatial patterns of visual field defects and predict their progression (before visual field information acquisition unit 10). The system is characterized by comprising a processing unit 20, a visual field defect analysis unit 30, and a visual field defect progression prediction unit 40, and a behavior context information acquisition unit 50 within the server that continuously collects behavioral information of the patient from sensors of various user terminals used by the patient, such as mobile phones, smartwatches, smart glasses, and car navigation screens, which are connected to the server via a wide-area network; a lifestyle risk estimation unit 60 within the server that estimates the risks in the patient's daily life by comparing the information from the analysis unit with the information from the behavior context information acquisition unit 50; a nudge generation unit 70 that generates nudges according to the estimated risks; and a nudge presentation control unit 80 that selects the nudge method, timing, and user terminal to perform the nudge based on the information from the behavior context information acquisition unit 50; an overall management and integration unit 90 that oversees the data and processing flow of the entire system; and a recording, analysis, and improvement unit 100 that records and analyzes all system activities and performs continuous improvement.

[0030] The components of the present invention will be described in detail below.

[0031] Visual field information acquisition unit 10 In the behavioral support nudge generation technology according to the present invention, the visual field information acquisition unit 10 is a unit that acquires visual field data and OCT data from visual field testing equipment, OCT testing equipment, electronic medical records, etc. In the behavioral support nudge generation technology according to the present invention, the visual field information acquisition unit 10 includes a measuring instrument interface 11 for connecting to measuring instruments such as visual field testing equipment and OCT devices, and a cloud transfer buffer 16 for temporarily storing acquired data and transmitting it to the cloud according to the communication status. In the behavioral support nudge generation technology according to the present invention, it is preferable that the visual field information acquisition unit 10 includes a data collection control module 12 for controlling the timing and amount of data reception in order to ensure the quality of acquired data, a communication encryption module 13 for securely sending and receiving patient information and examination data, an anonymization processing module 14 for removing or converting identification information in order to protect the patient's personal information, and a metadata assignment module 15 for recording additional information such as examination date and time, measurement conditions, and equipment identifier in the acquired data.

[0032] Field of view information preprocessing unit 20 In the behavioral support nudge generation technology according to the present invention, the pre-processing unit 20 for visual field information etc. is a unit that prepares the visual field data and OCT structure information transmitted from the visual field information etc. acquisition unit 10 into a format suitable for analysis. In the behavioral support nudge generation technology according to the present invention, the pre-processing unit 20 for visual field information etc. includes a data acceptance / format conversion module 21 that accepts acquired data and standardizes data formats between different devices, a coordinate system unification / binocular integration module 27 that integrates different coordinate systems and integrates measurement results from both eyes, and an output interface 29 that outputs the standardized pre-processed data to the next process. In the behavioral support nudge generation technology according to the present invention, the preprocessing unit 20 for field of view information, etc. preferably comprises a quality and reliability scoring module 22 that evaluates the reliability of the data and assigns a quality score, a noise reduction module 23 that removes measurement noise and outliers, a missing value interpolation module 24 that interpolates missing values ​​using statistical or machine learning methods, a spatial interpolation processing module 25 that interpolates spatially discrete data points to generate continuous data, a sensitivity normalization and calibration module 26 that corrects variability in sensitivity between measuring instruments and unifies them to a standard sensitivity scale, and a time alignment and longitudinal correction module 28 that aligns time series data and corrects differences in inspection intervals and measurement conditions.

[0033] Visual field defect analysis unit 30 In the behavioral support nudge generation technology according to the present invention, the visual field defect analysis unit 30 is a unit that analyzes pre-processed data transmitted from the visual field information pre-processing unit 20 and extracts spatial patterns and features of the patient's visual field defects. In the behavioral support nudge generation technology according to the present invention, the visual field defect analysis unit 30 comprises a visual field defect extraction module 31 that extracts defective areas from pre-processed data, a spatial feature generation module 32 that quantifies the location, extent, and severity of the extracted visual field defects, a progression vector extraction module 33 that compares and analyzes all past data to form progression direction vectors in each part of the visual field that show the distribution of the patient's visual field sensitivity decline and the direction and amount of change, and an output interface 36 for transmitting the analysis results to the visual field defect progression prediction unit 40 and the life risk estimation unit 60. In the behavioral support nudge generation technology according to the present invention, it is preferable that the visual field defect analysis unit 30 comprises a structure-function correspondence map generation module 34 that associates OCT structure information with visual field data, and a result integration module 35 that integrates multiple analysis results to generate a final analysis output.

[0034] Visual field defect progression prediction unit 40 In the behavioral support nudge generation technology according to the present invention, the visual field impairment progression prediction unit 40 estimates the current visual field state and predicts future visual field impairment progression patterns based on the progression direction vector generated by the visual field impairment analysis unit 30. In the behavioral support nudge generation technology according to the present invention, the visual field impairment progression prediction unit 40 includes a time series analysis module 41 that takes a progression direction vector extracted from past data as input and predicts spatial changes in visual field impairment on a time axis from the present to the future, a predictive visual field map generation module 42 that reconstructs the visual field impairment distribution based on the time series prediction results and generates a predicted visual field map from the present to the future, and an output interface 45 that transmits the prediction results to the visual field impairment analysis unit 30 and the life risk estimation unit 60. In the behavioral support nudge generation technology according to the present invention, it is preferable that the visual field impairment progression prediction unit 40 includes a structure-function integrated prediction module 43 that integrates OCT structure information and visual field function data to improve prediction accuracy, and an uncertainty evaluation module 44 that assigns reliability and error range to the prediction results.

[0035] Behavioral context information acquisition unit 50 In the behavioral support nudge generation technology according to the present invention, the behavioral context information acquisition unit 50 is a unit that collects behavior-related information from a user terminal or an external database and generates a behavioral pattern model in order to estimate the patient's daily life and behavioral patterns. In the behavioral support nudge generation technology according to the present invention, the behavioral context information acquisition unit 50 comprises a patient declaration information acquisition module 51 that accepts pre-inputs and questionnaires from the patient, a location information acquisition module 52 that collects location information from the GPS of the user terminal, etc., a motion analysis module 53 that detects walking, stair climbing and driving movements using an acceleration sensor and a gyroscope sensor, an environmental information acquisition module 54 that acquires environmental information such as illuminance, weather, time of day, and season from sensors and an external database as factors that affect visual function, an activity log generation module 55 that generates basic data for behavioral analysis and risk estimation, a behavioral pattern model generation module 56 that integrates the acquired behavioral information and environmental information, learns each patient's lifestyle rhythm and behavioral tendencies and generates a behavioral pattern model, and an output interface 58 that transmits the acquired information to the lifestyle risk estimation unit 60 and the nudge presentation control unit 80. In the behavioral support nudge generation technology according to the present invention, it is preferable that the behavioral context information acquisition unit 50 includes a privacy protection module 57 that anonymizes and encrypts patient personal information to protect it.

[0036] Life Risk Estimation Unit 60 In the behavioral support nudge generation technology according to the present invention, the life risk estimation unit 60 comprises a behavioral scenario mapping module 61 that maps typical life behavioral scenarios based on behavioral context, a risk area matching module 62 that matches visual field impairment patterns with risk areas related to daily life, and an output interface 67 that transmits the estimation results to the nudge generation unit 70. In the behavioral support nudge generation technology according to the present invention, it is preferable that the life risk estimation unit 60 comprises a dynamic risk scoring module 63 that numerically weights the risks estimated in accordance with behavior, a temporal risk fluctuation analysis module 64 that corrects the risk by considering behavioral patterns in specific time periods, an environmental factor risk correction module 65 that considers external environmental factors such as rain, nighttime, and fog, estimates and corrects for decreased visual field sensitivity and fluctuations in contrast sensitivity, and dynamically recalculates the risk evaluation for each life behavioral scenario, and a high-risk behavior extraction module 66 that extracts behaviors with a particularly high degree of danger from the estimated ones and makes them targets for warning.

[0037] Nudge generation unit 70 In the behavioral support nudge generation technology according to the present invention, the nudge generation unit 70 is a unit that generates behavioral support nudges to be presented to the patient based on the estimation results of the lifestyle risk estimation unit 60. In the behavioral support nudge generation technology according to the present invention, the nudge generation unit 70 includes a nudge content generation module 71 that selects nudge content to be presented according to the estimated risk, a treatment adherence support module 75 that predicts the future progression of visual field impairment and the increase in lifestyle risks due to non-compliance with eye drop application, visualizes the importance of treatment behavior by comparing it with the case of compliance, and generates nudges that strengthen and support the continuation of treatment, and an output interface 76 that transmits the generated nudges to the nudge presentation control unit 80. In the behavioral support nudge generation technology according to the present invention, it is preferable that the nudge generation unit 70 includes a timing optimization module 72 that optimizes the timing of nudge presentation in accordance with the behavioral context, a priority / urgency evaluation module 73 that determines priority and urgency according to the severity of the risk, and a learning / optimization module 74 that updates the nudge generation algorithm based on the presentation history and the patient's response.

[0038] Nudge presentation control unit 80 In the behavioral support nudge generation technology according to the present invention, the nudge presentation control unit 80 controls the timing, method, and priority of presenting nudges to various user terminals (smartphones, smartwatches, smart glasses, in-car displays, voice assistants, etc.) based on the nudges generated by the nudge generation unit 70 and behavioral and environmental information obtained from the behavioral context information acquisition unit 50. In the behavioral support nudge generation technology according to the present invention, the nudge presentation control unit 80 includes a presentation scheduling module 81 that adjusts the time and frequency of nudge presentation, a presentation selection context adaptation module 82 that adapts the presentation content and method according to the patient's behavioral context, situation, and the state of visual function in the predicted visual field map, a treatment motivation optimization presentation module 86 that optimizes presentation to enhance treatment motivation and continuous motivation, and an output interface 87 that delivers the generated nudges to the user terminal. In the behavioral support nudge generation technology according to the present invention, the nudge presentation control unit 80 preferably comprises an urgency priority presentation module 83 that prioritizes the presentation of nudges with high urgency, a multi-channel distribution module 84 that distributes nudges across multiple channels such as smartphones, smartwatches, and in-vehicle devices, and a history and feedback management module 85 that manages presentation history and patient feedback and reflects them in future presentation methods.

[0039] Overall management / integration department 90 In the behavioral support nudge generation technology according to the present invention, the overall management and integration unit 90 is a unit that oversees the processing flow and data flow of the entire system and controls each unit to operate in cooperation with each other. In the behavioral support nudge generation technology according to the present invention, the overall management and integration unit 90 includes a processing scheduling module 91 that manages the processing order and priority of each unit, a data synchronization and communication management module 92 that maintains data consistency between different modules and between servers and terminals, and a fail-safe error handling module 93 that safely stops and restores processing in the event of an exception to ensure the stability of the entire system. Furthermore, it is preferable that the overall management and integration unit 90 includes a system security monitoring function, a data backup function, an access rights management function, and so on.

[0040] Recording, Analysis, and Improvement Department 100 In the behavioral support nudge generation technology according to the present invention, the recording, analysis, and improvement unit 100 systematically records and analyzes the presented nudges and the user's reactions and behavioral changes, and uses the results to improve the entire system and for clinical application. In the behavioral support nudge generation technology according to the present invention, the recording, analysis, and improvement unit 100 includes a presentation history recording module 101 that records the content, timing, terminal, etc., of the presented nudges; a reaction / behavioral change recording module 102 that records the user's reactions and changes in daily behavior; and an effect analysis module 103 that statistically and machine-learningly analyzes the accumulated data to evaluate the nudge effect. In the behavioral support nudge generation technology according to the present invention, it is preferable that the recording, analysis, and improvement unit 100 includes a model improvement support module 104 that uses the results of the effect analysis to help improve the algorithm and presentation method, a medical and research information sharing module 105 that generates reports for doctors and researchers that visualize treatment compliance rates and accident risk reduction effects, and a family information sharing module 106 that provides the family with information such as the patient's eye drop application status, nudge presentation history, and lifestyle risk detection results in an appropriate format. [Effects of the Invention]

[0041] According to the present invention, by analyzing past visual field data and OCT structural information, it becomes possible to predict the spatial patterns of current and future visual field defects.

[0042] This system applies periodic visual field test data and OCT structural information, traditionally used in medical institutions to determine treatment plans, to estimate the current visual field state in daily life in real time, providing a foundation for extracting individualized, on-the-spot risks for each patient.

[0043] Furthermore, the present invention integrates the current visual field impairment estimation results with the patient's behavioral context information and environmental information (illumination, weather, time of day, etc.) to individually identify risks in line with specific behavioral scenarios such as walking, descending stairs, and driving a vehicle, and to assess the dangers the patient faces in real life in real time. In addition, by considering environmental information, it becomes possible to estimate risks that reflect fluctuations in visual state under real environmental conditions, and the accuracy of risk assessment in response to environmental changes such as outdoors or at night can be improved.

[0044] This integrated processing goes beyond simple visual field prediction and notification, enabling the presentation of nudges tailored to individual predicted risks and surrounding environmental conditions at the appropriate time and through the right means. This allows patients to intuitively understand their visual field impairment and the risks to their daily lives, thereby encouraging behavioral change.

[0045] Furthermore, by visualizing the predicted future worsening of visual field impairment and the specific disadvantages associated with it (such as difficulty descending stairs or visual errors while driving) if non-adherence to eye drop administration continues, patients can understand the causal relationship between their treatment behavior and future behavioral risks, thereby improving their motivation to continue treatment in a way that was not possible before.

[0046] Thus, the present invention interacts visual field progression prediction, behavioral context information, and environmental information to assess lifestyle behavior risks, thereby enabling glaucoma patients to become aware of lifestyle behavior risks and the disadvantages of non-adherence to eye drop use, which were previously difficult to predict. Furthermore, according to the present invention, the content and timing of nudge presentations can be dynamically optimized according to the visual field state and behavioral context, and by considering changes in environmental conditions (illuminance, weather, time of day, etc.), it is possible to improve the accuracy of risk estimation and the timeliness of warning presentations in real-life environments. This promotes behavioral change and has effects such as reducing falls and traffic accidents, preventing blindness, maintaining quality of life, and reducing the social burden. [Brief explanation of the drawing]

[0047] [Figure 1] This is a diagram showing the overall system configuration of the present invention. [Figure 2] This diagram shows an example configuration of the field of view information acquisition unit 10. [Figure 3] This figure shows an example of the configuration of the pre-processing unit 20 for field of view information, etc. [Figure 4] This diagram shows an example configuration of the visual field defect analysis unit 30. [Figure 5] This figure shows an example configuration of the visual field impairment progression prediction unit 40. [Figure 6] This figure shows an example configuration of the behavioral context information acquisition unit 50. [Figure 7] This figure shows an example of the configuration of the lifestyle risk estimation unit 60. [Figure 8] This figure shows an example configuration of the nudge generation unit 70. [Figure 9] This figure shows an example configuration of the nudge suggestion control unit 80. [Figure 10] This diagram shows an example configuration of the overall management and integration unit 90. [Figure 11] This diagram shows an example configuration of the Recording, Analysis, and Improvement Unit 100. [Figure 12] These are examples of multiple visual fields that show spatial patterns of visual field defects. [Figure 13]This figure shows an example of a nudge presentation in a situation where a person is descending stairs, as an example of the application of the present invention (Example 1). [Figure 14] This figure (Example 2) shows several examples of nudge presentations before and during operation as an application example of the present invention. [Figure 15] As an example of the application of the present invention, Figure 14(a) shows an example of a nudge presentation method on the user terminal screen before operation starts (Example 2). [Figure 16] As an example of the application of the present invention, the diagram shows a future prediction of the progression pattern of visual field impairment, and as an example of its impact on daily life, it shows the appearance of a risk of not seeing pedestrians when turning right in a car, and is an example of a nudge that encourages diligent use of eye drops. (Example 3) [Figure 17] As an example of applying the present invention, this figure shows a diagram illustrating the predicted progression pattern of visual field impairment after 5 years, and illustrates the impact on daily life, specifically a decrease in visibility when descending stairs. It also illustrates a target state after 5 years in which the progression of visual field impairment is slowed and visibility when descending stairs is maintained, thereby illustrating an example of a nudge to encourage diligent use of eye drops (Example 3). [Figure 18] As an example of the application of the present invention, this is a conceptual diagram showing how information is shared bidirectionally between the patient, family, and home healthcare team in home healthcare settings and this system (Example 4). [Modes for carrying out the invention]

[0048] The present invention provides a system and method for generating lifestyle behavioral nudges based on the progression of visual field defects. This system quantitatively analyzes and predicts the current and future spatial patterns of visual field defects in patients with visual field defects, estimates individual risks in daily life based on these predictions, and then presents nudges (warning or behavior-encouraging messages) to the user terminal at a timing and with content appropriate to those risks. This supports behavioral change and continuation of treatment for patients and contributes to the prevention of blindness.

[0049] The configuration, processing flow, and examples of embodiments of the system according to the present invention will be described in detail below.

[0050] <Overall System Configuration Diagram> Figure 1 is a block diagram showing an example of the overall configuration of the field of view nudge generation system according to the present invention. This system consists of the following multiple components: Visual field information acquisition unit 10 Field of view information preprocessing unit 20 Visual field defect analysis unit 30 Visual field defect progression prediction unit 40 Behavioral context information acquisition unit 50 Life Risk Estimation Unit 60 Nudge generation unit 70 Nudge presentation control unit 80 Overall management / integration department 90 Recording, Analysis, and Improvement Department 100

[0051] These components may be located within a single integrated server, or they may be distributed as a cloud-based system, interacting with each other over a network.

[0052] Figure 1 shows a cloud-based system. Visual field and OCT examination data from patients examined at clinics, etc., are continuously collected on the cloud via the visual field information acquisition unit 10 in a form that does not identify individuals (anonymized / pseudonymized). After preprocessing of the data in the visual field information preprocessing unit 20, the visual field defect analysis unit 30 analyzes past visual field examination data and OCT structural information to determine past progression trends. Based on the analysis results, the visual field defect progression prediction unit 40 estimates the current spatial pattern of visual field defects and predicts the future state of visual field defects, and sends this information to the life risk estimation unit 60.

[0053] Changes in OCT structural data are analyzed by the visual field defect analysis unit 30 and used to improve the accuracy of predicting the progression of visual field defects.

[0054] Meanwhile, the behavioral context information acquisition unit 50 collects lifestyle and work information declared by the patient, as well as location information (GPS, Wi-Fi positioning, beacon signals, etc.) obtained from the user terminal, and sends it to the lifestyle risk estimation unit 60 on the cloud.

[0055] In the lifestyle risk estimation unit 60, the spatial pattern of the patient's visual field impairment and its progression risk are compared with their lifestyle activities to extract the risk.

[0056] Depending on the identified risks, the nudge generation unit 70 generates nudges that encourage behavioral changes to avoid those risks.

[0057] The nudge presentation control unit 80 has a function to optimize the timing and format of providing nudges according to the user's situation.

[0058] The overall management and integration unit 90 can centrally manage data transfer between each component, adjustment of processing schedules, communication management, and synchronization of model updates.

[0059] The recording, analysis, and improvement unit 100 verifies all activities and effects performed in the system of the present invention, enabling future model improvements.

[0060] This system continuously updates current estimates and future predictions based on past discrete examination data, and corrects them with environmental information such as illuminance, weather, and time of day, thereby reflecting the ever-changing areas of visual field impairment and lifestyle risks in real time.

[0061] The role of each component is described in detail below.

[0062] <Visual field information etc. acquisition unit 10> The visual field information acquisition unit 10 (Figure 2) has the function of acquiring past visual field information and OCT structure information from multiple measurement means, such as visual field measurement equipment and fundus imaging equipment in medical institutions, mobile terminal applications, and also home-use simple visual field testing devices and simple OCT devices. The field of view information acquisition unit 10 (Figure 2) is equipped with interface processing to absorb differences in equipment and measurement protocols, and aims to safely and efficiently transfer acquired data in its original format to the field of view information preprocessing unit 20 on the cloud. Specifically, during data collection, patient identification information is anonymized, communication paths are encrypted, and supplementary information such as acquisition date and time and measurement conditions are added. Furthermore, it includes device connection drivers and API calling mechanisms to support measurement equipment from different manufacturers and models. Furthermore, the visual field information acquisition unit 10 according to the present invention is not limited to clinical testing equipment in medical institutions, but may also include a configuration that connects to a simple device that can be used by patients at home to acquire data. This makes it possible to transmit data acquired at home to a cloud server for analysis, enabling processing equivalent to that of a medical institution. The visual field information acquisition unit 10 (Figure 2) according to this embodiment includes a measurement device interface module 11 (driver for connecting to visual field measurement devices and OCT, communication API for fundus cameras and mobile terminal applications, etc.), a data acquisition control module 12 (measurement start / end trigger control, acquisition of measurement conditions [light intensity, stimulus pattern, etc.], etc.), a communication encryption module 13 (data transfer encryption by SSL / TLS, secure key management mechanism, etc.), an anonymization processing module 14 (removal or encoding of patient identification information, secure management of ID conversion table, etc.), a metadata assignment module 15 (addition of acquisition date and time, measurement conditions, device ID, etc.), and a transfer buffer 16 (temporary storage area for original format data, high-speed and reliable transfer to the visual field information preprocessing unit 20 on the cloud).

[0063] <Field of View Information Preprocessing Unit 20> The field of view information preprocessing unit 20 (Figure 3) converts the original field of view information received from the field of view information acquisition unit 10 into a unified format for analysis and performs preprocessing to improve the accuracy of the analysis. This preprocessing includes unifying coordinate systems, correcting the position of measurement points, interpolating missing values, normalizing measurement values, denoising, and assigning confidence scores. Furthermore, for data acquired at multiple points in time, time synchronization processing will be performed to ensure time-series consistency, correction for differences in measurement conditions will be carried out, and a unified format will be implemented to ensure compatibility with the analysis model. This makes it possible to input data obtained from different measuring instruments and measurement conditions into the subsequent visual field defect analysis unit 30 in a consistent format. The pre-processing unit 20 (Figure 3) for field of view information, etc., according to this embodiment includes a data reception / format conversion module 21 (converts output from different models [numerical values / images / DICOM / proprietary binary] to a common format), a quality evaluation / reliability scoring module 22 (calculates inspection reliability and flags it by scoring using field of view: Fixation Loss / FP / FN, reaction time; OCT: Signal Strength, artifact detection, segmentation accuracy), a noise reduction module 23 (reduces measurement noise using spatial smoothing, median, wavelets, etc.), a missing value interpolation module 24 (fills in missing points with spatial interpolation [bilinear interpolation, Gaussian interpolation, etc.] or time series prediction), a spatial interpolation processing module 25 (performs spatial interpolation processing between measurement points using geographic statistical methods such as inverse distance weighting [IDW], thin plate spline interpolation, and kriging), and a sensitivity normalization / calibration module 26 (age / pupil diameter / instrument scale correction, different methods are dB). It includes a mapping to a scale, a coordinate system unification and binocular integration module 27 (unifies horizontal and vertical axes with a fixation point reference, corrects mirror image relationships between the left and right eyes, corrects minute deviations such as head position angle, generates a binocular visual field map [minimum value / weighted minimum / visibility probability model, etc.], OCT structure-function matching), a time matching and longitudinal correction module 28 (restandardizes longitudinal data spanning measurement dates and equipment changes [reprojection to the same reference system]), and an output interface 29 (transfers all processed data to the visual field defect analysis unit 30 and the visual field defect progression prediction unit 40).

[0064] <Visual Field Impairment Analysis Unit 30> The visual field defect analysis unit 30 (Figure 4) has the function of quantitatively analyzing the spatial characteristics and progression trends of visual field defects, taking visual field data provided by the visual field information preprocessing unit 20 and, if necessary, integrated structural index data such as OCT as input. The visual field defect analysis unit 30 according to this embodiment may include the following multiple processing modules. 1. Visual field defect detection module 31 (a) From the interpolated and normalized visual field sensitivity map, the impaired region is extracted using a predetermined threshold (difference from a healthy reference / absolute dB threshold). (i) Identify clinically meaningful "areas" using two-dimensional labeling or clustering (k-means / DBSCAN, etc.), and suppress noisy small regions. (c) The smoothness and irregularity of the obstacle boundary are indicated by a simple morphological index (boundary length / area ratio, fractal-like index, etc.). 2. Spatial Feature Generation Module 32 (a) Summarize the location of the impairment (upper / lower / nasal / temporal / distance from the center), its extent and connectivity, the degree of influence of the vicinity of 5° from the center, etc., and convert them into feature vectors. (i) Standard clinical indicators such as MD / PSD / VFI are also included and organized to a level of detail usable in the visual field defect progression prediction unit 40 and the lifestyle risk estimation unit 60. 3. Progression vector extraction module 33 (a) Compare field sensitivity maps from multiple past points in time and extract the direction, speed, and trend of sensitivity decline as vectors. (i) Direction vectors are generated for each part of the visual field and output as indicators that summarize short- to medium-term changes. (c) Since the visual field defect progression prediction unit 40 is responsible for future time-series estimation and spatial reconstruction, this module is limited to the stage of extracting the change trends (initial conditions) that form the basis of the prediction. 4. Structure-Function Correspondence Map Generation Module 34 (a) Based on the coordinate system standardized in the preprocessing, abnormal distributions of OCT structural information indicators such as retinal nerve fiber layer thickness, macular structure, or optic disc shape are superimposed with visual field defects, and the degree of correspondence (strength of correlation, suggestion that structural changes precede functional changes) is tagged. (i) No lifestyle risk assessment is performed, but the information is added as flag information to reinforce the progression prediction of the visual field impairment progression prediction unit 40. 5. Results Integration Module 35 (a) The entire set of output (fault mask, spatial feature vector, progress signal summary, structure-function tag) is consolidated into a standardized format. 6. Output Interface 36 (a) Output destinations: Visual field impairment progression prediction unit 40... Future prediction model input (spatial features + change summary + OCT flag), Life risk estimation unit 60... Spatial summary values ​​(regional weights, etc.) for overlaying with behavioral scenarios.

[0065] <Visual field impairment progression prediction section 40> The visual field defect progression prediction unit 40 (Figure 5) takes the progression direction vector generated by the visual field defect analysis unit 30 as input, predicts the spatial changes in the visual field defect over a time axis from the present to the future, and has the function of qualitatively and quantitatively estimating the current and future visual field state. This configuration makes it possible to perform dynamic predictions that take into account the direction and speed of progression, rather than simply analyzing past data. The visual field defect progression prediction unit 40 according to this embodiment may include the following processing modules. 1. Time Series Analysis Module 41 (a) Using the direction vector extracted by the visual field defect analysis unit 30 as an initial condition, the future changes in sensitivity values ​​at each measurement point in the visual field are mathematically predicted. (i) Calculate the rate of change in sensitivity (dB / year) at each measurement point using linear regression, exponential regression, mixed-effects models, etc., and quantify the rate of progression. (c) To address situations with a small number of measurements or the presence of outliers, Bayesian estimation or Kalman filtering is applied to suppress errors and improve prediction accuracy. 2. Predictive field of view map generation module 42 (a) Based on the predicted future sensitivity values ​​for each measurement point obtained by the time series analysis module 41, the changes along the time axis are interpolated and extended to generate numerical data of field sensitivity at each present and future point in time. This obtains a future sensitivity data series that maintains the continuity of changes over time. (i) The future sensitivity data obtained above is used to reconstruct a future sensitivity distribution that reflects the direction and intensity of sensitivity changes across the entire visual field by referencing the direction vector obtained by the visual field defect analysis unit 30 and correcting the spatial consistency between each part. (c) To take into account spatial correlations and structural dependencies, spatial prediction models such as Gaussian process regression, Markov random field, cellular automaton models, and convolutional neural networks (CNNs) can be applied. (e) These methods allow for the reconstruction and visualization of the progression of visual field defects across multiple scales—short-term, medium-term, and long-term—providing a basis for behavioral risk assessment in the life risk estimation unit 60. 3. Structure-Function Integration Prediction Module 43 (a) The rate of change in structural elements such as retinal nerve fiber layer thickness, macular structure, or optic disc shape, along with the rate of change in visual field sensitivity, will be analyzed using multivariate regression analysis, machine learning models, or Bayesian estimation methods. (i) Areas where structural changes precede functional decline are tagged as high-risk areas and reflected in future vision predictions. (c) The rate of change over time of OCT structural information such as retinal nerve fiber layer thickness, macular structure, or optic nerve head shape is calculated, and the rate of change in visual field sensitivity is corrected based on this rate of change to further improve prediction accuracy. 4. Uncertainty Assessment Module 44 (a) Calculate the uncertainty by determining the confidence interval (e.g., 95% confidence interval) and predictive distribution for the predicted values. (i) If the confidence level is low, the nudge generation unit 70 will use it to make action suggestions such as "recommend retesting" or "shorten the monitoring interval." 5. Output Interface 45 (a) Generate a predictive visual field map of visual field defects and transmit the related metadata to the life risk estimation unit 60.

[0066] <Behavior context information acquisition unit 50> The behavioral context information acquisition unit 50 (Figure 6) has the function of acquiring contextual information about the patient's behavioral state and living environment from multiple perspectives. Primarily, it enables the automatic detection of patient behaviors such as "walking," "climbing stairs," and "driving a car" based on objective data obtained from acceleration sensors, gyro sensors, GPS, indoor beacons, environmental sensors, and external weather services mounted on the user terminal. Furthermore, it enables the acquisition of environmental information such as illuminance, weather, and time of day that affect visual function from sensors and external databases, and uses this information for risk correction. Moreover, by feeding back information reported by the patient (e.g., eye drop application status, fall history, subjective response to nudge presentations, lifestyle habits, etc.) to this system, based on this sensor information, a more precise and individualized understanding of behavioral context is achieved. The behavioral context information acquisition unit 50 according to this embodiment may include the following processing modules. 1. Patient-reported information acquisition module 51 (a) During medical examinations or via a user terminal application, detailed information about daily life and occupation (e.g., frequency and time of driving, commute route, work environment, hobbies, movement patterns within the home, exercise habits, whether or not stairs are used, eye drop application status, experiences of falls or trips, subjective evaluations and responses to presented nudges) will be obtained. (i) This information will be updated as needed and used to understand long-term behavioral trends and changes in lifestyle, complementing contextual information that is difficult to obtain from sensor data. (c) Input can be made at any time, and multiple methods are supported, including free-response text, multiple-choice questionnaires, voice input, and an eye drop record interface. 2. Location information acquisition module 52 (a) The user's current location and movement status are obtained using GPS, Wi-Fi positioning, beacon signals, etc. (i) It supports both outdoor and indoor use, and high-precision positioning is performed indoors using BLE beacons and Wi-Fi triangulation. (c) From the acquired location data, geographical characteristics of daily activities (e.g., sections of road where cars are driven, walking routes, locations of stairs, dark areas, etc.) are estimated. 3. Motion analysis module 53 (a) The system utilizes motion sensors such as accelerometers, gyroscopes, and geomagnetic sensors installed in smartphones and wearable devices to determine the type of user activity (walking, running, cycling, climbing stairs, driving a vehicle, standing still, etc.). (i) For behavioral pattern recognition, machine learning models such as random forests, convolutional neural networks (CNNs), and recurrent neural networks (RNNs) can be applied. (c) The timing of behavioral changes (e.g., walking → running, sitting → standing) is detected in real time and used for behavioral scenario determination in the lifestyle risk estimation unit 60. 4. Environmental Information Acquisition Module 54 (a) Obtain location information via GPS from a smartphone or wearable device, and obtain weather data for that location from a weather service (e.g., Japan Meteorological Agency or weather API). (i) Use illuminance sensors, barometric pressure sensors, temperature and humidity sensors, etc., installed in smartphones and wearable devices to acquire information about the surrounding environment (brightness, weather, and changes in atmospheric pressure). (c) Identify activities that increase visual risk at night or in low-light conditions (e.g., driving at night, climbing stairs in the dark). (e) Taking into account the visual effects of weather, the system detects reduced visibility due to decreased illumination in cloudy or rainy weather, and visual impairment due to glare from sunlight in sunny weather. (e) Especially on clear days, the altitude and direction of sunlight are calculated from location information and time to estimate poor visibility due to backlighting or low angle of incidence (e.g., glare from the setting sun in the evening). (c) Changes in atmospheric pressure and weather can be analyzed in relation to changes in movement patterns and activity ranges, contributing to improved accuracy in determining behavioral scenarios in the lifestyle risk estimation unit 60. 5. Activity log generation module 55 (a) Location information, activity type, and environmental information are integrated chronologically and recorded as an activity context log. (i) The behavioral context log can include start time, end time, location, type of activity, environmental conditions, movement speed, and field of view requirements. (c) The log data will be used as foundational information when the lifestyle risk estimation unit 60 overlays important visual field areas for each behavioral scenario with a future predicted visual field impairment map. 6. Behavioral Pattern Model Generation Module 56 (a) Using the behavioral context logs obtained by the activity log generation module 55 and attribute information such as lifestyle habits, occupation, and activity time obtained from the patient-reported information acquisition module 51 as input, the system learns and models each patient's lifestyle rhythm and behavioral tendencies. (i) For the learning process, machine learning algorithms such as clustering, hidden Markov models (HMMs), recurrent neural networks (RNNs), and transformer-based time series analysis models are applied to quantify behavioral patterns by time of day, day of the week, and environmental conditions. (c) The obtained behavioral pattern model estimates what kind of behavior the patient tends to exhibit under specific conditions (e.g., going out at night, driving in the rain, outdoor activities on holidays, etc.), and uses the results as basic information for determining the behavioral context and assessing risk in the life risk estimation unit 60, forming a foundation for risk estimation and behavioral support that is in line with the actual living situation. (e) Furthermore, the behavioral pattern model is used to optimize "when" and "how" to present nudges in the nudge presentation control unit 80, thereby realizing real-time support that is in line with the behavioral context. (e) Continuously update new behavioral context logs and patient reports to enable the model to be updated to reflect changes in the patient's lifestyle (e.g., age, change in commuting style, job change, relocation, etc.). 7. Privacy Protection Module 57 (a) When implementing the patent, individual location data and activity history will be anonymized and encrypted, and the design will ensure that only the minimum amount of data necessary for the intended use is shared. (i) Remove personally identifiable information before sending to the cloud by employing on-device processing on the device or federated learning. 8. Output Interface 58 Activity type (walking, climbing stairs, driving, etc.) and environmental conditions (illuminance, weather, direction of sunlight, etc.) are estimated from acceleration sensors, GPS, illuminance sensors, etc. Flags related to activity risks such as nighttime, low-light conditions, and backlit conditions are assigned and output to the lifestyle risk estimation unit 60 and the nudge suggestion control unit 80.

[0067] <Life Risk Estimation Section 60> The lifestyle risk estimation unit 60 (Figure 7) integrates the current spatial pattern of visual field defects received from the visual field defect analysis unit 30 and the future spatial progression pattern received from the visual field defect progression prediction unit 40, along with behavioral and environmental information obtained from the behavioral context information acquisition unit 50, to estimate a specific and quantitative risk assessment tailored to the individual's current and future lifestyle. The lifestyle risk estimation unit 60 according to this embodiment may include the following processing modules. 1. Action Scenario Mapping Module 61 (a) The system analyzes location information, acceleration sensor data, activity logs, etc., obtained from the behavior context information acquisition unit 50 to classify and identify daily life behavior scenarios (e.g., driving a car, climbing stairs, crossing a pedestrian crossing, walking in a crowded environment, etc.). (i) Create a database in advance of the visual requirements characteristics corresponding to each behavioral scenario (field of view, importance of fixation and peripheral vision, need for scotopic vision, etc.). 2. Risk Area Matching Module 62 (a) The current visual field impairment pattern (obtained from the visual field impairment analysis unit 30) and the predicted future visual field impairment pattern (obtained from the visual field impairment progression prediction unit 40) are compared with the required characteristics of the behavioral scenario, and the overlap between the visual area required during the behavior and the visual field impairment area is analyzed. (i) If a visual field defect is likely to affect behavioral safety, mark that area and behavior as a risk. 3. Dynamic Risk Scoring Module 63 (a) Assign a score to the risk of each action (e.g., 0-100 points) and classify it as a risk level (low, medium, high). (i) Scoring is calculated by comprehensively evaluating the location and size of the affected area, the speed of progression, the presence or absence of overlap in the visual fields of both eyes, and environmental factors (brightness, congestion, etc.). 4. Time-based risk variation analysis module 64 (a) Combine diurnal variations, behavioral patterns during specific time periods, and predicted field of vision to assess risk for each time period. (i) For example, the risk should be adjusted by taking into account the increased visual load at specific times, such as driving at night or walking at dusk. 5. Environmental Factor Risk Adjustment Module 65 (a) By acquiring information on changes in weather, illuminance, and surrounding environment, and applying corrections to the risk score, a risk assessment that reflects the actual environment will be conducted. (i) For example, the risk of slipping when going up or down stairs in rainy weather, and the risk of reduced visibility when driving at night or in fog should be taken into consideration when revising the risk estimate. (c) Individually learn the field of view sensitivity and the correction curves for illumination and weather, and map them to the probability of visibility for each action to re-evaluate the risk. 6. High-risk behavior extraction module 66 (a) Among the estimated risks, behaviors that are particularly likely to lead to accidents, falls, or social limitations are extracted and sent to the subsequent nudge generation unit 70. (i) For example, specific incidents can be identified such as "difficulty detecting bicycles approaching from the left when driving a car at night" or "delay in seeing the lower steps when descending stairs." 7. Output Interface 67 (a) The generated risk assessment results are sent to the nudge generation unit 70 and used to generate appropriate messages that encourage risk avoidance, behavioral improvement, and treatment adherence.

[0068] <Nudge generation unit 70> The nudge generation unit 70 (Figure 8) has the function of generating messages (nudges) to encourage behavioral change, risk avoidance, and treatment adherence of the subject, based on current and future high-risk behavioral information received from the lifestyle risk estimation unit 60, and outputting them to the nudge presentation control unit 80. The nudge generation unit 70 according to this embodiment may include the following processing modules. 1. Nudge content generation module 71 (a) Analyze the types of high-risk behaviors (e.g., night driving, descending stairs, crossing pedestrian crossings, walking in crowded environments) and risk factors (location of the obstruction, speed of movement, environmental conditions), and select appropriate nudge actions. (i) For example, generate specific action suggestions such as, "Since your left-side visual field is deteriorating, be sure to check your left side carefully, especially at night when turning left at a crosswalk," and "Since you have a downward visual field impairment, use the handrail when descending stairs." 2. Timing Optimization Module 72 (a) The system references real-time data (location information, acceleration sensor data, time, illuminance information, etc.) from the behavior context information acquisition unit 50 and adds metadata that contributes to the optimization process in the nudge suggestion control unit 80. (i) For example, the nudge is immediately presented as soon as the user starts driving at night, or just before the user approaches a crosswalk, and is sent at the appropriate time when necessary, either immediately before or during an action. 3. Priority and Urgency Assessment Module 73 (a) Set nudge priorities according to the risk level and the danger of the behavior. (i) In cases of high urgency (e.g., immediate collision risk), an immediate notification will be issued using a combination of an alert sound and vibration. 4. Learning and Optimization Module 74 (a) Record and analyze user nudge responses (presence or absence of behavioral changes, success rate of risk avoidance, etc.) and continuously optimize the nudge content and presentation conditions. (i) By using distributed learning methods such as federated learning, it is possible to improve the model without directly sharing personal data. 5. Treatment Encouragement Support Module 75 (a) By comparing the results of the visual field impairment progression prediction unit 40 and the lifestyle risk estimation unit 60, the future progression of visual field impairment and the increase in lifestyle risks due to non-compliance with eye drop application are predicted, and by visualizing this compared to the case where compliance is observed, the importance of treatment behavior is made clear, and a nudge is generated to strengthen and support the continuation of treatment. (i) The presentation of the impact should be carried out in a way that allows the patient to intuitively understand their future (e.g., VR / AR simulation, color coding, superimposition of visual field impairment onto live-action footage). (c) This clarifies the significance and urgency of continuing treatment and improves motivation for diligent use of eye drops and regular check-ups. (e) The presentation method should take into consideration the patient's psychological burden and should be presented along with positive messages that lead to behavioral improvement. 6. Output Interface 76 (a) The generated nudge is sent to the nudge presentation control unit 80, and the optimal method is selected and executed in accordance with the behavioral context, such as the timing of the nudge notification, the format of the nudge presentation, and the selection of the user terminal to present the nudge. Thus, the nudge generation unit 70 • Generating appropriate content in response to high-risk behaviors. • Optimization of timing based on behavioral context • Selection of presentation format according to user status and device characteristics • Display control based on priority and urgency • Learning and improvement based on usage history • Support for encouraging treatment based on future changes in perspective The aim is to improve the safety of patients' daily lives by integrating these methods, prevent accidents and limitations in daily life caused by visual field defects, and maintain and enhance their motivation to continue treatment.

[0069] <Nudge presentation control unit 80> The nudge presentation control unit 80 (Figure 9) combines various nudges generated by the nudge generation unit 70 with real-time behavioral and environmental information obtained from the behavioral context information acquisition unit 50, and has the function of controlling the presentation at the optimal timing, terminal, and format based on the user's current situation, urgency, device environment, and presentation history. The nudge suggestion control unit 80 according to this embodiment may include the following processing modules. 1. Presenting scheduling module 81 (a) The system references behavioral and environmental data obtained from the behavioral context information acquisition unit 50 to perform scheduling that is in line with the actual behavioral context. (i) Low-priority nudges should be presented in batches during times when users have more free time or during breaks to avoid information overload and stress. 2. Presenting Selection Context-Adapted Module 82 (a) In addition to real-time information obtained from the behavioral context information acquisition unit 50 (location, time, illuminance, movement speed, etc.) and device status (screen ON / OFF, voice enabled / disabled), the current state of visual function in the predicted visual field map output from the visual field impairment progression prediction unit 40 is referenced to determine whether or not to present the information and the appropriate presentation format (text display, voice guidance, vibration notification, etc.). (i) Target devices include smartphones, smartwatches, in-car displays, voice assistants, etc. Depending on the user's condition and preferences, the system will select a format appropriate to the usage environment, such as prioritizing voice output if the user's visual impairment is progressing, and using vibration notifications in combination in noisy environments. (c) In particular, in high-risk behavioral scenarios such as descending stairs, driving a car, and maintaining eye contact, audio presentation will be prioritized over text presentation to ensure immediacy and safety. 3. Urgency Prioritization Module 83 (a) If high-risk behavior or danger requiring immediate avoidance is detected, a nudge will be presented immediately, disregarding the normal schedule. (i) By combining multiple sensory channels such as alert sounds, strong vibrations, and screen flashing, reliable recognition is achieved. 4. Multi-channel distribution module 84 (a) Deliver nudges simultaneously or selectively to multiple devices, such as smartphones, smartwatches, in-car displays, smart glasses, and voice assistants. (i) Synchronize devices to prevent duplicate presentations or missed presentations. 5. History and Feedback Management Module 85 (a) The history of the nudges presented and the user's response (confirmation, ignoring, taking action) are recorded and fed back to the nudge generation unit 70 and the learning / optimization module. (i) Tune the presentation frequency and timing for each user to maximize effectiveness. 6. Treatment Motivation Optimization Presentation Module 86 (a) The content of the "Prediction of Future Visual Field Impairment Progression and its Impact on Daily Living Activities" created by the treatment adherence support module (nudge generation unit 70) will be distributed regularly and in a manner that takes into consideration the patient's psychology. (i) Example: Visualize the extent to which the predicted progression is suppressed by continuing treatment, and emphasize positive changes to encourage diligent use of eye drops and continued medical consultations. 7. Output Interface 87 (a) The output interface 87 outputs to the selected user terminal at the selected timing and using the selected presentation format.

[0070] <Overall Management / Integration Department 90> The overall management and integration unit 90 (Figure 10) has the function of comprehensively controlling each component of the system according to the present invention (field of view information acquisition unit 10, field of view information preprocessing unit 20, field of view impairment analysis unit 30, field of view impairment progression prediction unit 40, behavioral context information acquisition unit 50, lifestyle risk estimation unit 60, nudge generation unit 70, and nudge presentation control unit 80). This makes it possible to centrally manage data transfer between each component, adjustment of processing schedules, communication management, model update synchronization, etc. The overall management and integration unit 90 according to this embodiment may include the following modules. 1. Processing scheduling module 91 The system optimizes the overall processing load by controlling the execution timing of each component, data acquisition intervals, model update cycles, etc. For example, by executing high-load analysis processes during idle time, it achieves both real-time performance and power efficiency. 2. Data synchronization and communication management module 92 It manages data transmission and reception between user terminals and cloud servers, ensuring security through encrypted communication such as TLS. It also handles parameter transmission and reflection of learning results for distributed learning methods such as federated learning. 3. Fail-safe error handling module 93 In the event of abnormalities such as communication failure, sensor malfunction, or model loading failure, alternative processing and notifications are performed to ensure the continuity of the entire system's operation.

[0071] <Recording, Analysis, and Improvement Department 100> The recording, analysis, and improvement unit 100 (Figure 11) has the function of recording and accumulating the content of the nudge presented in the system of the present invention, the time of presentation, the format of presentation, the presentation device, the user's reaction, behavioral changes, and changes in visual field defects over time, and analyzing these. This enables verification of the effectiveness of the nudge, support for continued treatment, information sharing with medical and research institutions, and future model improvement. The recording, analysis, and improvement unit 100 according to this embodiment may include the following modules. 1. Presentation history recording module 101 The date and time of presentation, nudge content, presentation format (audio, text, etc.), and presentation device are recorded. 2. Response / Behavior Change Recording Module 102 Changes in user behavior after presentation (e.g., speed reduction, change of direction, evasive maneuvers, etc.) are acquired and recorded based on sensor data or self-reported information from the user. 3. Effectiveness Analysis Module 103 The success rate of risk avoidance through nudge presentation, the duration of behavioral improvement, etc., are quantitatively evaluated, and the results are fed back to the nudge generation unit 70 and the nudge presentation control unit 80. 4. Model Improvement Support Module 104 Federated learning and other distributed learning techniques continuously improve the predictive accuracy of the overall model without directly sharing personal data. 5. Information Sharing Module 105 for Medical and Research Use This system automatically generates reports including predictions of visual field impairment progression, assessments of lifestyle risks, and nudge effect analyses, and provides them to medical institutions, research institutions, or home healthcare teams. This supports the revision of treatment plans and the development of new intervention strategies. It also features two-way communication capabilities to receive observations from healthcare professionals, supplementary information on the patient's condition, and instructions for lifestyle support, and incorporate them into the system's analysis and nudge generation. 6. Family Information Sharing Module 106 The system organizes the patient's eye drop application status, nudge presentation history, and lifestyle risk detection results, and provides them to the family in an appropriate format. This allows the family to understand the patient's treatment status and lifestyle risks, and to support continued treatment through verbal encouragement, lifestyle support, and eye drop application confirmation. Furthermore, it features a two-way communication function that receives reports on lifestyle conditions from the family, confirmation of eye drop application, and feedback on patient behavior, and incorporates this information into the system.

[0072] To aid in understanding the present invention, specific examples of its use are shown below. For convenience, Figure 12 illustrates a visual field impairment pattern, and Figures 13 to 18 show specific examples of its use; however, these do not limit the technical scope of the present invention.

[0073] Figure 12 shows an example of a spatial pattern of visual field defects in the binocular visual field that is analyzed by the system according to the present invention. As shown in the figure, the location and extent of visual field defects differ from patient to patient, and there are cases where defects are present in the left visual field, in the lower right visual field, or where defects are spread throughout the entire lower visual field. The system according to the present invention analyzes these different spatial patterns qualitatively and quantitatively, and further predicts future spatial progression patterns based on the temporal changes in each visual field. By inputting the analysis results into the behavioral risk estimation unit 60, it enables individualized risk assessment for each patient.

[0074] Figure 13 illustrates an example of the application of the present invention, showing how a nudge is presented when descending stairs. (a) is a diagram showing the patient's current visual field test results, where there is extensive visual field impairment in the lower region, some of which extends to the fixation point. (b) is a diagram illustrating the visibility of the steps when the patient descends the stairs, with the visual field impairment pattern superimposed on the stair image to clearly show areas that are easy to see and areas that are difficult to see. It can be seen that a narrow area of ​​about one foot's width of the step is clearly visible, but a wide area is unclear. This is assessed as a state in which the risk of falling is high if the patient descends carelessly. (c) is a diagram showing how the patient receives a nudge from the user terminal when descending stairs. The person shown in white is a symbolic representation of the user terminal. The system according to the present invention precisely recognizes the details of the risks associated with descending stairs, and at the moment it detects the start of descent, it presents an audio nudge from a mobile user terminal such as a smartphone or smartwatch, stating something to the effect of, "It is safer to use the handrail, look from side to side to check your feet and the steps, and descend slowly," thereby preventing falls. In this way, the present invention integrates the spatial pattern of visual field impairment with the patient's behavioral context, allowing for the selection of appropriate nudge content and presentation format in high-risk situations inherent in daily life, such as descending stairs, thereby promoting accident prevention.

[0075] Figure 14 illustrates an example of the application of the present invention, using automobile driving as an example, and shows the manner in which nudges are presented according to the driving situation. (a) is a diagram showing the scene before driving begins, in which the patient has a visual impairment to the left. In this case, a nudge message such as "It's raining today. It's easy to overlook pedestrians and cyclists on the left, so it's reassuring to occasionally turn your attention to the left." is displayed on the in-car display, which acts as the user terminal. Since it is possible to focus on the display before driving begins, text display is selected as the presentation format. (b) is a diagram showing a scene in which the patient is driving, in which attention to the left side of the vehicle, including when turning left at an intersection, is particularly important. A nudge message such as "It's reassuring to pay attention to whether there are any bicycles or mopeds on the left." is presented by voice from a mobile terminal or an in-car display connected to a mobile terminal via Bluetooth. Since it is deemed dangerous to check text information while driving, voice is selected as the presentation format. (c) is an image diagram of turning right. The patient has a visual impairment to the right. A nudge message is presented audibly: "It is reassuring to be aware of whether there are pedestrians on the right side of the crosswalk. Turning your head will make it easier to see." (d) is a diagram showing a situation where the surroundings become dark due to sudden rain while driving, and the patient has impaired left-side vision. In this case, a nudge message is presented audibly: "It has become dark due to the sudden rain. It is difficult to notice bicycles and mopeds on the left. It is reassuring to slow down and focus your attention to the left." Thus, according to the present invention, nudge content is generated based on driving conditions, road conditions, and environmental information, and the presentation format is appropriately selected, thereby promoting the patient's avoidance of danger.

[0076] Figure 15 shows an example of a visual display that presents a nudge on the user terminal screen in a scenario such as Figure 14(a), as an application example of the present invention. Assume the patient has a visual field defect on the left side. On the left side of the screen, an example of a nudge is presented to alert the user to bicycles, motorcycles, kick scooters, etc., traveling on the left side of the vehicle. The text and illustration are presented together, stating, "It is safer to pay attention to bicycles, motorcycles, kick scooters, etc. on the left side. Turning your head will improve your vision on the left side where your field of vision is narrowed." On the right side of the screen, an example of a nudge is presented based on the weather forecast for the day, such as "Rain is forecast for the afternoon. It is especially safer to pay attention to the left side." In this way, the system according to the present invention promotes the patient's avoidance of danger by enabling intuitive and situation-appropriate warnings using visual representations on the user terminal screen, based on information obtained from the visual field defect analysis unit 30 (Figure 4), the behavioral context information acquisition unit 50 (Figure 6), and the environmental information acquisition module 54 (Figure 6).

[0077] Figure 16 shows an example of an application of the present invention for a patient with left visual field impairment who has not adequately received eye drop treatment. The figure shows a display that presents the predicted progression pattern of visual field impairment after 5 years based on time-series visual field test results, along with an example of an increased risk in daily life due to this impairment, and an example of a nudge to encourage treatment. (a) is the visual field test result showing the current state of visual field impairment, and (b) shows the predicted progression of visual field impairment after 5 years if the current treatment situation continues. (a) and (b) are displayed side by side on the user terminal screen. (c) and (d) are user terminal screen displays that present an example of the impact of the progression of the visual field impairment on the patient's daily life, showing a situation where it becomes difficult to see pedestrians approaching from the right side of a crosswalk when a car is turning right, and include nudge text to encourage treatment. (c) is an overhead view showing a vehicle turning right, and (d) is a diagram showing the situation where the visibility of pedestrians approaching from the right side of the crosswalk is reduced from the perspective of a vehicle turning right. (e) is a nudge statement that reads, "Currently, there is an area on the left that is difficult to see. If eye drops are not used sufficiently, it is predicted that the area of ​​difficulty in seeing will expand to the right in the future. For example, when turning right, pedestrians approaching from the right at a crosswalk may become unclear as shown in the diagram, potentially delaying their recognition." In this way, the present invention promotes motivation for patients to continue treatment by presenting the predicted changes in their visual field and the resulting disadvantages in daily life in an intuitively understandable form.

[0078] Figure 17 illustrates an application example of the present invention for a patient with right-downward visual field impairment who has not received sufficient eye drop treatment. It shows the predicted progression of visual field impairment after 5 years based on time-series visual field test results, and the resulting increase in daily life risks, as well as illustrating the effect of suppressing progression through diligent treatment. (a) is the visual field test result showing the patient's current state of visual field impairment, and (b) is the visual field test result showing the predicted progression of visual field impairment after 5 years if the current insufficient treatment continues. (c) is a diagram showing the target visual field state after 5 years that can be achieved if treatment is properly followed. These (a) to (c) are displayed side by side on a single user terminal screen. (d) is a diagram showing the current visibility of steps when descending stairs. By superimposing the patient's visual field impairment pattern onto the stair image, it illustrates how visibility is reduced in a part of the right side of the stairs, but visibility at the feet is maintained. (e) is a diagram showing the predicted decrease in the visibility of steps when descending stairs after 5 years if the condition remains inadequately treated, illustrating how the visibility of the feet decreases as the visual field defect progresses, increasing the risk of falling. (f) is a diagram showing that by diligently continuing treatment, the progression of the visual field defect can be suppressed after 5 years, and visibility when descending stairs can be maintained at a relatively good level. (g) is a nudge statement that reads, "Currently, there is a slightly difficult-to-see area in the lower right, but if the eye drops are not used sufficiently, it is expected that this will spread throughout the lower area after 5 years, making it difficult to see the steps when descending stairs. Let's aim to slow the progression by maintaining sufficiently low intraocular pressure so that you can still see the steps well after 5 years." These (d) to (g) are displayed side by side on a single user terminal screen. In this way, according to the present invention, by presenting the predicted progression of visual field defect if treatment is neglected and the resulting increase in risks in daily life, and by presenting the effect of suppressing progression by diligently continuing treatment as a target image, the motivation of patients to continue treatment is further strengthened.

[0079] Figure 18 is a conceptual diagram illustrating an application example of the present invention, showing how information sharing and resulting improvement in nudge generation can occur through bidirectional communication between the patient, family, and home healthcare team and this system in a home healthcare environment. The system according to the present invention presents patients with nudges related to daily life and treatment adherence via a nudge generation unit 70 (Figure 8) and a nudge presentation control unit 80 (Figure 9) through a user terminal. It also allows patients to input information such as eye drop application status, fall experiences, lifestyle habits, and responses to nudges via a patient-reported information acquisition module 51, and feeds this information back into the system, thereby enabling individual optimization of nudge generation. For families, information such as the patient's eye drop records, nudge presentation history, and lifestyle risk detection results are provided as notifications or reports via the family information sharing module 106 (Figure 11). The system also receives reports from families regarding the patient's living situation, eye drop confirmation results, and feedback on patient behavior, which are then analyzed and incorporated into nudge generation. This two-way communication provides families with motivations and methods for participating in treatment and lifestyle support. Furthermore, the home healthcare team is provided with information such as the patient's visual field defect status, prediction of visual field defect progression, eye drop adherence, nudge history, and lifestyle risk assessment via the medical and research information sharing module 105 (Figure 11). The system also receives observations, supplementary information, and lifestyle support instructions from doctors and nurses, and incorporates these into nudge generation and treatment support through bidirectional communication. This enables the home healthcare team to implement treatment and nursing care that is tailored to the actual situation. In this way, patients, families, and home healthcare teams communicate bidirectionally with the system of the present invention via their respective dedicated modules (patient-reported information acquisition module 51, family-oriented information sharing module 106, and medical / research-oriented information sharing module 105), and information relevant to daily life is continuously collected and reflected. As a result, nudge generation is improved and enhanced to reflect individuality and reality, further strengthening the effectiveness of safety and treatment continuation support in home healthcare.

[0080] These illustrations demonstrate that the system according to the present invention not only displays analysis results but also provides support directly related to specific behavioral situations in daily life. In other words, patients can receive warnings tailored to their own living situation, rather than relying on abstract numerical values ​​or diagnostic results, thereby effectively promoting accident prevention and treatment adherence. Figures 13 to 18 show examples of the present invention and do not limit the technical scope of the present invention thereto. [Examples]

[0081] The system according to this embodiment illustrates a configuration that detects the "risk of tripping and falling" during walking and stair climbing for patients with downward visual field defects in both eyes, and presents a warning nudge at an appropriate time.

[0082] The visual field information acquisition unit 10 (Figure 2) acquires visual field data, which is then processed by the visual field information preprocessing unit 20 (Figure 3) and the visual field impairment analysis unit 30 (Figure 4) to analyze the spatial pattern of visual field impairment and past progression trends. Based on these analysis results, the visual field impairment progression prediction unit 40 (Figure 5) extends the progression direction vector, which shows the direction and amount of change in visual field sensitivity decrease for each area, onto the time axis to generate a visual field sensitivity distribution (predicted visual field map) at the current time in real time and transmits it to the lifestyle risk estimation unit 60 (Figure 7). The behavioral context information acquisition unit 50 (Figure 6) detects that the patient is in a behavioral context such as "walking" or "starting to descend stairs" based on sensor information from the user terminal (accelerometer, location information, indoor beacon, etc.) and transmits it to the lifestyle risk estimation unit 60. The lifestyle risk estimation unit 60 compares the state of visual function in the predicted visual field map (i.e., the visual field sensitivity distribution and defective areas of each part) with behavioral context information to quantitatively estimate the "risk of tripping and falling" caused by obstacles and steps during walking, and poor recognition of steps when descending stairs. In particular, if a lower visual field defect is detected, poor recognition of steps and steps is quantified as a risk factor and reflected in the risk score. Furthermore, by referring to the patient's lifestyle rhythm and behavioral tendencies accumulated and learned by the behavioral pattern model generation module 56, behavioral patterns according to time of day, day of the week, and environmental conditions are reflected as correction elements, enabling dynamic and individualized risk estimation that is in line with actual living conditions. External environmental conditions such as weather and illuminance acquired by the environmental information acquisition module 54 (Figure 6) are reflected to dynamically correct the risk score.

[0083] Based on these analysis results, the nudge generation unit 70 (Figure 8) generates the following example nudges. • Before leaving for work: "(Alert + text) It is predicted that it will rain this afternoon, making it dark and difficult to see your footing. Take an umbrella, keep an eye on your footing, and use the handrail on stairs for safety." • When going out in the evening: "(Alert + text) As the sun sets, visibility is becoming poor. Check the condition of the road surface on sidewalks and use the handrail on stairs for greater safety." • When using the stairs at home: "(Audio) It's safer to use the handrail and descend slowly while checking your feet and the steps by looking from side to side." Figure 13 shows an example of how to present a nudge when using stairs at home. The nudge presentation control unit 80 (Figure 9) selects the appropriate means, such as voice, vibration, sound effects, or text display, depending on the situation, and provides immediate guidance in emergency situations.

[0084] This embodiment reduces the risk of falls during walking and stair climbing in patients with lower visual field impairment, thereby contributing to improved safety and quality of life. [Examples]

[0085] The lifestyle behavior nudge generation system based on visual field impairment in this embodiment exemplifies a scenario in which a patient with visual field impairment in both eyes drives a car.

[0086] In this system, the behavioral context information acquisition unit 50 (Figure 6) detects that the user is driving a car by utilizing GPS information installed on the user terminal and the connection status with the in-vehicle Bluetooth. This distinguishes the user from normal walking or using public transportation and recognizes that they are in a "car driving scenario" as their behavioral context. Visual field data acquired periodically by the visual field information acquisition unit 10 (Figure 2) is analyzed by the visual field information preprocessing unit 20 (Figure 3) and then by the visual field impairment analysis unit 30 (Figure 4). Here, the visual field impairment analysis unit 30 extracts progression direction vectors from past visual field data that show the direction and amount of change in visual field sensitivity for each area, and quantitatively understands the direction in which the patient's visual field impairment is progressing. The visual field impairment progression prediction unit 40 (Figure 5) takes the progression direction vectors as input and performs time-series analysis to estimate changes in visual field sensitivity. This generates a predicted visual field map that visualizes the sensitivity distribution in the present and future, and transmits the results to the life risk estimation unit 60 (Figure 7). The lifestyle risk estimation unit 60 (Figure 7) compares this predicted visual field map with the driving behavior context and calculates the risk of reduced visibility of the surrounding traffic environment (other vehicles, pedestrians, bicycles, motorcycles, traffic lights, signs, etc.) according to the spatial pattern of the visual field impairment. For example, if there is a visual field impairment on the left, the risk of reduced visibility of vehicles approaching from the left, bicycles, motorcycles, parked vehicles, road signs, and pedestrians on crosswalks when turning left is calculated qualitatively and quantitatively. Similarly, if there is a visual field impairment on the right, the risk of reduced visibility of vehicles approaching or overtaking from the right, pedestrians and bicycles crossing crosswalks on the right side when turning right, and road signs is calculated qualitatively and quantitatively. The quantification of risk is performed by mapping the "location, size, and degree of the visual field impairment area" with "safety requirements when turning left, turning right, entering intersections, etc." Furthermore, the system references each patient's driving habits (driving speed, commuting route, driving time, reaction tendencies under different lighting conditions, etc.) learned by the behavioral pattern model generation module 56 to add individuality to the risk prediction during driving. In addition, the risk assessment is corrected by reflecting the weather, illuminance, and angle of sunlight during driving, etc., acquired by the environmental information acquisition module 54 (Figure 6). This allows for dynamic risk estimation that is in line with the actual driving environment, such as the setting sun in the evening or reduced visibility in rainy weather.

[0087] The nudge generation unit 70 (Figure 8) generates the following nudges based on the above analysis results. An example of a patient with left visual field defects: • At the start of driving: "(In-vehicle display alert + text) It is forecast to rain today. It is safer to occasionally pay attention to the left side as it is easy to overlook pedestrians and cyclists on the left." (Figure 14(a) illustrates a specific example). The display on the in-vehicle display should preferably be in a form that is intuitively recognizable using visual representations, as illustrated in Figure 15. • When turning left at an intersection: "(Voice follows left turn signal) You are turning left at the intersection. For added safety, briefly turn your head to the left to check for pedestrians and cyclists." (See Figure 14(b) for an example). • After sunset: "(Voice) It's getting a little dark. Visibility to the left is likely to decrease during this time. Slow down at intersections and pay extra attention for safety." • In case of sudden rain: "(Audio) It's gotten dark due to the sudden rain. It's harder to notice bicycles and mopeds on the left. It's safer to slow down and pay attention to the left." (See Figure 14(d) for an example).

[0088] An example of a patient with right-field visual field defects: • When turning right at an intersection: "(Voice follows right turn signal) It's safer to be aware of whether there are pedestrians on the right side of the crosswalk. Turning your head will make it easier to see." (See Figure 14(c) for an example).

[0089] The nudge presentation control unit 80 (Figure 9) switches the presentation method according to the driving environment. For example, it links the vehicle navigation screen with a smartphone and presents a combination of alert sound + screen display or voice output. In addition, the urgency priority presentation module 83 (Figure 9) detects the timing just before a risk materializes, such as "just before an intersection" or "sudden weather change," and controls immediate presentation.

[0090] According to this embodiment, the risk of overlooking pedestrians, vehicles, etc., when a patient with left-sided or right-sided visual field impairment drives a car is reduced, thereby contributing to the prevention of traffic accidents and the improvement of safety in daily life. [Examples]

[0091] The visual field defect analysis unit 30 (Figure 4) analyzes visual field data from the past several years and extracts progression direction vectors that show the direction and amount of change in visual field sensitivity decline in each area. The visual field defect progression prediction unit 40 (Figure 5) takes the progression direction vectors as input and performs time-series analysis to model sensitivity changes from the past to the present and into the future, thereby generating a predicted visual field map of the spatial pattern of future visual field defects in years, 5 years, 10 years, etc. The predicted visual field map quantifies the depth, extent, direction and speed of sensitivity decline in each visual field area and outputs it in a format that can be visually presented to the patient.

[0092] In parallel with this, the lifestyle risk estimation unit 60 (Figure 7) acquires the patient's daily life and occupational behavioral patterns from the behavioral context information acquisition unit 50 (Figure 6). Furthermore, it refers to the behavioral pattern model learned by the behavioral pattern model generation module 56 and analyzes how the predicted progression of visual field defects will affect the patient's daily activities (e.g., driving, desk work, climbing stairs, etc.). For example, the reduced visibility of signs, pedestrians, bicycles, and motorcycles while driving, the reduced visibility of mouse pointers and text while working on a computer, and the reduced visibility of steps while climbing stairs are quantified as "risks for each behavioral scenario."

[0093] The treatment encouragement support module 75 (Figure 8) of the nudge generation unit 70 (Figure 8) presents the potential impact on the patient's future daily activities if visual field defects progress at a predicted pace, enabling intuitive understanding. Specifically, it visualizes the future decline in daily and occupational functional abilities using VR / AR experiential simulations, color-coded displays, or superimposed displays of visual field defects on live-action footage.

[0094] If the progression is slow and the future impact is minor, an affirmative nudge is generated, such as, "Your current treatment is working properly. The progression is being suppressed precisely because you are continuing to use the eye drops. Continuing this will help maintain safety in your daily activities in the future." On the other hand, if the progression is rapid and has a significant impact on daily activities, the future scenario of the predicted visual field map is visualized on a specific behavioral scenario, the current state of eye drop adherence is checked, and if there is room for improvement, a nudge is presented to encourage continued use of the eye drops. If adherence is already good, it is determined that the treatment effect may not be sufficient, and this information is shared with the attending physician via the medical / research information sharing module 105 (Figure 11), leading to consideration of additional eye drops or surgical treatment. Furthermore, the treatment motivation optimization presentation module 86 (Figure 9) of the nudge presentation control unit 80 (Figure 9) presents a positive future scenario of "progression suppression" and "maintenance of safety in daily activities" that is predicted if the treatment is improved or strengthened, thereby increasing the patient's motivation.

[0095] As a specific example, as shown in Figures 16(a) and 16(b), in a patient currently experiencing visual field defects localized to the left, vector analysis of the direction of progression suggests a tendency for the visual field defect to expand to the right. Based on this vector information, a predicted visual field map is generated that, if the current treatment continues, a significant decrease in sensitivity in the right visual field will occur after 5 years, resulting in a reduced visibility of pedestrians (e.g., schoolchildren) approaching from the right side of a crosswalk when turning right in a car, as shown in Figures 16(c) and 16(d). In this case, including a nudge such as, "Currently, there is a blind spot on your left. If eye drops are not used sufficiently, it is predicted that the blind spot will expand to the right in the future. For example, when driving and turning right, pedestrians approaching from the right side of a crosswalk may become unclear as shown in the figure, potentially leading to delayed recognition," allows the patient to understand future disadvantages and risks to their daily life that they are not yet aware of, thereby increasing their motivation to continue treatment.

[0096] In another example, as shown in Figure 17(a), in a patient currently experiencing a visual field defect localized to the lower right, similar analysis results show that if the current treatment continues, the visual field defect will expand to the entire lower area and reach the vicinity of the fixation point after 5 years, as shown in Figure 17(b). On the other hand, if treatment is properly followed, the progression of the visual field will be suppressed, and a good visual field state is presented as a target even after 5 years, as shown in Figure 17(c). In addition, to provide a sense of real-life experience, the visual field defect pattern is superimposed on a staircase image to reproduce visibility when descending stairs. Figure 17(d) illustrates the current field of vision when descending stairs, showing reduced visibility in a part of the right side, but the footing is clear. On the other hand, Figure 17(e) is a prediction after 5 years if treatment remains insufficient, visualizing the significant decrease in visibility of steps at foot level and the increased risk of falls. Furthermore, Figure 17(f) illustrates that if treatment is properly followed, the progression of the visual field will be suppressed, and visibility when descending stairs can be maintained at a relatively good level. In conjunction with these, a nudge statement such as, "Currently, there is a slightly blurry area in the lower right, but if eye drops are not used sufficiently, it is expected that this will spread throughout the lower area in 5 years, making it difficult to see steps when descending stairs. Let's make sure to continue eye drop treatment to slow the progression and aim to be able to see steps clearly in 5 years," is presented, allowing patients to intuitively understand the disadvantages of continuing insufficient treatment and the benefits of diligently following appropriate treatment.

[0097] Furthermore, the presentation scheduling module 81 (Figure 9) adjusts the presentation timing according to the detection of progression signs and lifestyle rhythms, and the medical / research information sharing module 105 (Figure 11) shares information with the attending physician. Examples of nudges include the following: • Eye drop reminder: "(Voice + alert) It's time to drop your eye drops. Maintaining a regular eye drop routine will protect your future vision." • Sharing with the attending physician: "(Text notification) Eye drop use is progressing well. I will share this information with your attending physician."

[0098] In this embodiment, patients can intuitively understand the future progression of visual field defects and their impact on daily life, and realize the significance of diligently following treatment. Furthermore, treatment discontinuation can be prevented through habit formation support using nudges. These factors contribute to suppressing the progression of progressive eye diseases such as glaucoma, and also contribute to optimizing treatment through collaboration with healthcare professionals. [Examples]

[0099] The system described in this embodiment exemplifies a configuration that facilitates treatment and provides lifestyle support for elderly patients and patients under care who have difficulty visiting hospitals, by collaborating with home healthcare teams and connecting the patient, family, and medical professionals. In particular, it is characterized by its ability to visualize and share current and future visual field impairment predictions, as well as estimated changes in lifestyle risks, among the three parties involved.

[0100] Similar to Examples 1-3, the visual field defect analysis unit 30 (Figure 4) extracts the spatial pattern and progression direction vector of visual field defects from past data, and based on these analysis results, the visual field defect progression prediction unit 40 (Figure 5) estimates the current spatial pattern in real time and further generates a future predicted visual field map. The lifestyle risk estimation unit 60 (Figure 7) quantifies behavioral risks in home life by referring to a behavioral pattern model generated by the contextual information acquisition unit 50 (Figure 6) based on the patient's behavioral context and living environment information. Similar to Examples 1-3, the nudge generation unit 70 (Figure 8) and nudge presentation control unit 80 (Figure 9) present the patient with nudges related to daily life and treatment adherence at optimized timings via a user terminal. • Example of a nudge for supporting daily living activities: "It's the stairs. Pay attention to the steps and use the handrail to descend them carefully, one step at a time." • Example of a nudge to encourage eye drop application: "(Voice + alert notification) It's time to apply your eye drops. Continuing to do so will help protect your vision next year." Furthermore, the visual field impairment progression prediction unit 40 (Figure 5) and the lifestyle risk estimation unit 60 (Figure 7) work together to present, through experiential simulations as in Example 3, the predicted visual field impairment in the future, such as two years and five years from now, and the associated disadvantages such as an increased risk of falls in daily life and reduced visibility when watching television. It also shows that the progression of visual field impairment can be suppressed and the disadvantages mitigated by continuing appropriate treatment. As a result, the patient can gain a concrete image that "if I continue using the eye drops, I can continue to act safely and watch television," thereby improving their sense of self-management.

[0101] Up to this point, it is similar to Examples 1 to 3, but the unique feature of Example 4 is the effective sharing of information presented to the patient with the family and the visiting medical team, as shown in Figure 18. First, information such as the patient's eye drop records, nudge presentation history, and lifestyle risk detection results is sent to the family as notifications or reports in a format edited and processed for family use via the family information sharing module 106 (Figure 11). The method of information sharing is selected according to the family structure, such as sending real-time notifications to family members living together, or sending periodic reports to family members living separately in the case of elderly people living alone. This allows the family to understand the patient's treatment status and lifestyle risks, and to gain opportunities to offer encouragement and support. • Eye drop records and reminder response status (e.g., a simple report is sent if eye drops are frequently forgotten) • Lifestyle risk information (e.g., alerts when fall risk is detected) Furthermore, families generally have little understanding of the extent of the patient's visual field impairment and the resulting problems in their daily lives. Therefore, visual information is shared that allows families to easily understand the current state of visual field impairment and its associated risks, as well as the predicted future state of visual field impairment and the resulting increase in risks. When a patient's mobility limitations increase or their lifestyle changes, such as becoming bedridden due to falls, the burden of support on the family inevitably increases. Families will understand that the patient's continued appropriate treatment is an effective way to prevent this burden from escalating, which will increase their motivation to support the patient, lead to improvements in the living environment (eliminating steps, adding more lighting, etc.), and encourage participation in treatment by monitoring whether the patient is administering eye drops according to the doctor's instructions, such as checking the liquid volume in the eye drop bottle.

[0102] Next, patient information is shared with visiting physicians and nurses via the medical and research information sharing module 105 (Figure 11), which transmits the information to the electronic medical record system for home healthcare. Visiting physicians and nurses can view the following information through its dashboard: • Current status and future predictions of visual field defects in patients • Eye drop adherence (implementation rate and forgetting patterns) • Nudge presentation history and response logs • Prediction of increased future life risks As a result, • Optimization of treatment plan (consideration of additional eye drops, laser treatment, or surgical treatment) is possible. • Customization of nudge content (e.g., prioritizing verbal nudges and increasing the number of nudges for patients with mild cognitive impairment). • Feedback received during visits (e.g., "frequent forgetting to apply eye drops in the evening," "increased risk of falling down stairs") enables more effective lifestyle guidance.

[0103] Furthermore, as shown in Figure 18, patients, families, and home healthcare teams communicate bidirectionally with the system according to the present invention via the patient-reported information acquisition module 51 (Figure 6), the family-oriented information sharing module 106 (Figure 11), and the medical / research-oriented information sharing module 105 (Figure 11). This allows for the collection and reflection of information that reflects the actual situation, leading to the refinement and sophistication of nudge generation. It should be noted that direct communication between patients and families, medical institutions and patients, and families and medical institutions may also occur in parallel as part of the reality of home healthcare, and these functions to complement the information sharing and nudge generation functions of this system.

[0104] In this embodiment, • Patients can visually understand future changes in their field of vision and lifestyle risks, and improve and maintain their treatment and lifestyle habits with the support of nudges. • Families can share the patient's current situation and future risks, gaining a sense of security and providing a starting point for concrete support actions and participation in treatment. • The visiting medical team can provide enhanced treatment and lifestyle guidance based on specific data. As a result, improved eye drop adherence, suppression of visual field impairment progression, and prevention of falls are expected, contributing to improved safety and quality of life in home healthcare. [Industrial applicability]

[0105] This invention is a medical safety technology that enables support for safe behaviors and health promotion in patients' daily lives, in addition to medical treatment at healthcare facilities. It is useful for supporting behavioral change, treatment continuation, and accident prevention for patients with visual field impairments, and can be widely applied to telemedicine, personalized medicine, and home monitoring in the field of ophthalmology. Furthermore, it can be used in conjunction with nursing homes, home care, visual rehabilitation, public transportation and automobile driver assistance systems, wearable devices and smartphone applications, and cloud-based digital health services. In addition, by applying it to health insurance and public health programs, government welfare policies, patient awareness and medical professional training, and epidemiological research and AI model development by research institutions, it can contribute to improving patients' quality of life, reducing the risk of falls and traffic accidents, improving treatment adherence, suppressing medical and care burdens, and even maintaining the productivity of society as a whole. [Explanation of Symbols]

[0106] 10...Visual field information etc. acquisition unit 11. Measuring instrument interface 12 … Data Acquisition Control Module 13… Communication encryption module 14 … Anonymization processing module 15 … Metadata assignment module 16 … Cloud transfer buffer 20 ... Field of view information and other preprocessing steps 21… Data reception and format conversion module 22… Quality and reliability scoring module 23… Noise reduction module 24... Missing Data Interpolation Module 25 … Spatial Interpolation Processing Module 26… Sensitivity Normalization and Calibration Module 27… Coordinate System Unit, Binocular Integrated Module 28… Time-aligned and longitudinal correction module 29 … Output Interface 30… Visual Field Impairment Analysis Department 31… Visual field defect detection module 32 … Spatial Feature Generation Module 33… Directional vector extraction module 34 … Structure-Function Correspondence Map Generation Module 35… Results Integration Module 36… Output Interface 40... Visual field defect progression prediction section 41… Time Series Analysis Module 42… Predictive field of view map generation module 43 … Structure-Function Integrated Prediction Module 44 … Uncertainty Assessment Module 45 … Output Interface 50 … Behavioral context information acquisition unit 51… Patient Declaration Information Acquisition Module 52 … Location information acquisition module 53… Motion analysis module 54 ... Environmental information acquisition module 55… Activity log generation module 56… Behavioral pattern model generation 57… Privacy protection module 58 … Output Interface 60… Life Risk Estimation Department 61… Action Scenario Mapping Module 62… Risk Area Matching Module 63… Dynamic Risk Scoring Module 64… Time-based risk variation analysis module 65… Environmental Factor Risk Adjustment Module 66… High-risk behavior extraction module 67 … Output Interface 70 ... Nudge generation unit 71 … Nudge Content Generation Module 72 … Timing Optimization Module 73… Priority and urgency assessment module 74… Learning and Optimization Module 75… Treatment adherence support module 76 … Output Interface 80 ... Nudge presentation control unit 81… Presented scheduling module 82 … Presenting Selection Context-Adapted Module 83… Urgency Prioritization Module 84… Multi-channel distribution module 85… History and Feedback Management Module 86… Treatment Motivation Optimization Presentation Module 87 … Output Interface 90 … Overall management / integration department 91 … Processing scheduling module 92 … Data synchronization and communication management module 93… Fail-safe error handling module 100 … Recording, Analysis, and Improvement Department 101… Presentation history recording module 102… Response / Behavior Change Recording Module 103… Effect Analysis Module 104 … Model Improvement Support Module 105… Information sharing module for medical and research purposes 106… Family-oriented information sharing module

Claims

1. A nudge generation system for supporting the safety of daily activities and the continuation of treatment for patients with visual field defects, - A visual field information acquisition unit that acquires past visual field test data of patients obtained at medical institutions (hereinafter referred to as "visual field information") and structural image data of the fundus obtained by optical coherence tomography (OCT) (hereinafter referred to as "OCT structural information"), - A preprocessing unit for field of view information, etc., which prepares acquired field of view information and OCT structural information into a format suitable for analysis, - A visual field information analysis unit analyzes the spatial pattern of visual field defects and the trend of changes in said spatial pattern over time based on visual field information and OCT structural information acquired in the past, and generates a progression direction vector that shows the distribution of visual field defects in the patient and the direction and amount of change. A visual field defect progression prediction unit predicts the spatial distribution and sensitivity changes of current and future visual field defects over time based on the results of the analysis, and generates a predicted visual field map based on the results of the prediction. - A behavioral context information acquisition unit that acquires behavioral context information regarding the patient's behavioral state and living environment from sensor data installed in the user terminal, patient-reported information, and environmental information, and generates a behavioral pattern model. - A life risk estimation unit that compares the predictive field of view map, the behavioral pattern model, and the behavioral context information to estimate in real time the individual life behavioral risks that the patient may face now or in the future, - A nudge generation unit that generates nudges to encourage behavioral change based on the estimation results of the aforementioned lifestyle behavior risks, according to the urgency of the risk and the behavioral scenario, A nudge presentation control unit that dynamically optimizes the timing, frequency, and format of the nudge presentation, taking into consideration the type and priority of the nudge, as well as the patient's activity status and environmental information. - An overall management and integration unit that comprehensively manages the exchange of information, adjustment of processing schedules, and communication between each component: the visual field information acquisition unit, the visual field information preprocessing unit, the visual field information analysis unit, the visual field impairment progression prediction unit, the behavioral context information acquisition unit, the lifestyle risk estimation unit, the nudge generation unit, and the nudge presentation control unit. - A recording, analysis, and improvement unit that records and analyzes presented nudges, patient responses, behavioral changes, and changes in visual field defects over time, and continuously learns and improves the processing models of each of the following units: visual field information preprocessing unit, visual field information analysis unit, visual field defect progression prediction unit, behavioral context information acquisition unit, lifestyle risk estimation unit, nudge generation unit, and nudge presentation control unit. A nudge generation system characterized by comprising

2. A nudge generation system according to claim 1, characterized in that the visual field information and OCT structure information are acquired at a medical institution and then transmitted to a cloud analysis server, and the visual field information analysis unit and the visual field impairment progression prediction unit perform analysis and prediction processing on the cloud analysis server.

3. A nudge generation system according to claim 1 or 2, characterized in that the behavioral context information acquisition unit integrally processes sensor data indicating the patient's movement status, physical movements and surrounding environment, and patient-reported information including eye drop application status, lifestyle and occupational behavioral patterns and movement range to generate a behavioral pattern model.

4. A nudge generation system according to claim 1 or 2, characterized in that the life risk estimation unit integrates the predictive field of view map, the behavioral pattern model, and the behavioral context information, and estimates in real time the individual life behavioral risks that the patient may face now or in the future, taking into account the temporal changes, correlations, and importance of the predictive field of view map, the behavioral pattern model, and the behavioral context information, respectively.

5. A nudge generation system according to claim 1 or 2, characterized in that the visual field information analysis unit and the visual field impairment progression prediction unit correct changes in the direction or speed of progression that cannot be captured by estimation from past visual field information using OCT structure information acquired in the past, and improve the accuracy of the current and future predicted visual field maps based on the results of the correction.

6. A nudge generation system according to claim 1 or 2, characterized in that the behavioral context information acquisition unit acquires environmental information from an external information source or a user terminal sensor, incorporates the environmental information as an evaluation element of the lifestyle risk estimation unit, and dynamically corrects the lifestyle risk estimation result in accordance with environmental changes.

7. A nudge generation system according to claim 1 or 2, characterized in that the nudge generation unit generates different types of nudges, including approval nudges and warning nudges, based on the estimated results of the lifestyle behavior risk and the behavioral scenario, according to the urgency of the risk and the patient's condition.

8. A nudge generation system according to claim 1 or 2, characterized in that the nudge presentation control unit refers to the predicted field of view map and behavioral context information and executes presentation control logic that prioritizes audio presentation over text display in behavioral scenarios where suppression of eye gaze deviation is required, such as descending stairs, driving a car, or other similar actions.

9. A nudge generation system according to claim 1 or 2, characterized in that the nudge presentation control unit comprehensively considers the predicted field of view map, the behavioral pattern model, and the environmental information to dynamically optimize the timing, frequency, and means of presenting the nudge according to the patient's activity status, urgency, risk priority, and real-time surrounding environment.

10. A nudge generation system according to claim 1 or 2, characterized in that the nudge generation unit refers to the predicted visual field map and the behavioral pattern model, superimposes the patient's current and future visual field state onto a real-world image, and generates visual nudges that enable the patient to intuitively understand the effects of visual field impairment and take actions to avoid lifestyle risks or encourage treatment based on that understanding.

11. A nudge generation system according to claim 1 or 2, characterized in that the nudge is presented on the patient's user terminal.

12. A nudge generation system according to claim 1 or 2, characterized in that the nudge presentation control unit automatically selects a nudge presentation means considering the state of visual function in the current predicted field of view map and environmental information relating to noise and illuminance.

13. A nudge generation system according to claim 1 or 2, characterized in that the recording, analysis, and improvement unit records and analyzes presented nudges, patient responses, and actual changes in visual field defects in a time series, and updates and improves the nudge generation model and nudge presentation control logic using machine learning methods based on the results of the analysis.

14. A nudge generation system according to claim 1 or 2, characterized in that the recording, analysis, and improvement unit records and analyzes the history of presented nudges, patient responses, behavioral changes, and changes in visual field defects over time, and continuously improves each process, including analysis, estimation, generation, and presentation, using the results of the recording and analysis.

15. A nudge generation system according to claim 1 or 2, wherein the family information sharing module of the recording, analysis, and improvement unit organizes the patient's eye drop application status, lifestyle risk estimation results, and nudge presentation history, provides the organized eye drop application status, lifestyle risk estimation results, and nudge presentation history to the family, and the behavioral context information acquisition unit receives lifestyle status report data, eye drop confirmation input data, or observation data related to patient behavior transmitted from the family terminal, and has a bidirectional communication function that automatically reflects the received lifestyle status report data, eye drop confirmation input data, or observation data related to patient behavior to the lifestyle risk estimation unit, nudge generation unit, and nudge presentation control unit.

16. A nudge generation system according to claim 1 or 2, characterized in that the medical information sharing module of the recording, analysis, and improvement unit organizes the patient's visual field impairment progression prediction results, eye drop application status, nudge presentation history, and lifestyle risk estimation results and provides them to a medical institution or home healthcare team, and the behavioral context information acquisition unit receives finding data, supplementary information, or lifestyle support instruction data transmitted from a medical professional's terminal, and has a bidirectional communication function for medical collaboration that automatically reflects the received finding data, supplementary information, or lifestyle support instruction data to the visual field impairment progression prediction unit, lifestyle risk estimation unit, nudge generation unit, and nudge presentation control unit.

17. A nudge generation system according to claim 1 or 2, wherein data relating to at least one of the visual field impairment progression prediction results, behavioral context information, lifestyle risk estimation results, eye drop application status, and nudge presentation history is shared bidirectionally among multiple parties, the patient, family, and healthcare professionals via the behavioral context information acquisition unit, family information sharing module, and healthcare information sharing module, and the visual field impairment progression prediction unit, lifestyle risk estimation unit, nudge generation unit, and nudge presentation control unit cooperate with each other based on the visual field impairment progression prediction results, behavioral context information, lifestyle risk estimation results, eye drop application status, and nudge presentation history shared among multiple parties to generate an individually optimized nudge that is suitable for both the patient's actual living situation and medical risks.

18. A nudge generation system according to claim 1 or 2, characterized in that each of the processes of the pre-processing unit for visual field information, the analysis unit for visual field information, the prediction unit for the progression of visual field impairment, the acquisition unit for behavioral context information, the estimation unit for life risk, the nudge generation unit, and the nudge presentation control unit uses a machine learning algorithm to improve accuracy and improve the model.

19. A behavioral change support method in which a computer performs actions to support the safety of daily activities and the continuation of treatment for patients with visual field defects, (a) A process of acquiring past visual field test data of a patient obtained at a medical institution (hereinafter referred to as "visual field information") and structural image data obtained by optical coherence tomography (OCT) (hereinafter referred to as "OCT structural information"), (b) A process of preparing the acquired field of view information and OCT structural information into a format suitable for analysis. (c) A step of analyzing the spatial pattern of visual field defects and the trend of changes in said spatial pattern over time based on visual field information and OCT structural information acquired in the past, and generating a progression direction vector that shows the distribution of visual field defects in the patient and the direction and amount of change, (d) A step of predicting the spatial distribution and sensitivity changes of current and future visual field defects in a time series based on the results of the analysis, and generating a predicted visual field map obtained from the predicted results, (e) A step of acquiring behavioral context information regarding the patient's behavioral state and living environment from sensor data installed on the user terminal, patient-reported information, and environmental information, and generating a behavioral pattern model, (f) A step of real-time estimation of individual lifestyle risks that the patient may face now or in the future by comparing the predictive field map, the behavioral pattern model, and the behavioral context information, (g) A step of generating nudges to encourage behavioral change based on the estimation results of the lifestyle behavior risks, according to the urgency of the risk and the behavioral scenario, (h) A step of dynamically optimizing and presenting the timing, frequency, and format of the nudge, taking into consideration the type and priority of the nudge, as well as the patient's activity status and environmental information. (i) A process to record and analyze the presented nudges, patient responses, behavioral changes, and changes in visual field defects over time, and to continuously learn and improve the processing models used in each of the processes (b), (c), (d), (e), (f), (g), and (h), A method for supporting behavioral change, characterized by including [a specific element].

20. A method for supporting behavioral change according to claim 19, characterized in that the generation of the nudge in step (g) is carried out in a manner that specifically presents potential future disadvantages in daily life or occupation based on the progression of visual field impairment and the living environment.

21. A method for supporting behavioral change according to claim 19, characterized in that the acquisition of behavioral context information in step (e) is performed by integrally processing sensor data indicating the patient's behavioral state and living environment, and patient-reported information regarding the status of eye drop application, lifestyle / occupational behavioral patterns, or range of movement.

22. A method for supporting behavioral change according to claim 19, characterized in that the nudge in step (h) is presented by voice, and the voice nudge presents a warning in situations including when descending stairs, driving a car, or in behavioral scenarios that require suppression of gaze deviation.

23. A program for causing a computer to execute behavioral change support methods to support the safety of daily activities and the continuation of treatment for patients with visual field defects, (a) Process to acquire past visual field test data of patients obtained at medical institutions (hereinafter referred to as "visual field information") and structural image data of the fundus obtained by optical coherence tomography (OCT) (hereinafter referred to as "OCT structural information"), (b) Processing to prepare the acquired field of view information and OCT structural information into a format suitable for analysis. (c) A process that analyzes the spatial pattern of visual field defects and the trend of changes in said spatial pattern over time based on visual field information and OCT structural information acquired in the past, and generates a progression direction vector that shows the distribution of visual field defects in the patient and the direction and amount of change. (d) A process to predict the spatial distribution and sensitivity changes of current and future visual field defects over time based on the results of the analysis, and to generate a predicted visual field map obtained from the predicted results. (e) A process to acquire behavioral context information regarding the patient's behavioral state and living environment from sensor data installed on the user terminal, patient-reported information, and environmental information, and to generate a behavioral pattern model. (f) A process that estimates in real time individual lifestyle risks that the patient may face now or in the future by comparing the predictive field map, the behavioral pattern model, and the behavioral context information. (g) A process for generating nudges to encourage behavioral change based on the estimation results of the aforementioned lifestyle behavior risks, according to the urgency of the risk and the behavioral scenario. (h) A process for dynamically optimizing the timing, frequency, and format of presentation of the nudge, taking into consideration the type and priority of the nudge, as well as the patient's activity status and environmental information. (i) A process that records and analyzes the presented nudge, the patient's response, behavioral changes, and changes in visual field defects over time, and continuously learns and improves the processing model used in each of the processes (b), (c), (d), (e), (f), (g), and (h). A program characterized by causing a computer to execute something.

24. In the program described in claim 23, A program characterized in that the nudge generation in the process (g) includes an approval nudge or a warning nudge depending on the patient's behavior, and the type of approval nudge or warning nudge is selected according to the patient's situation.

Images