Method for testing and evaluating safety effectiveness of digital therapy product for mental diseases

Abstract

The invention discloses a method for testing and evaluating safety effectiveness of a digital therapy product for mental diseases, which relates to the technical field of mental diseases and comprises the following steps: S1, evaluating objective indexes of the digital therapy product; s2, testing a man-machine interaction feedback mode; s3, mapping measurement of product output content and patient feedback; and S4, performing real-time adjustment and optimization. According to the method for testing and evaluating the safety effectiveness of the mental disease digital therapy product, aiming at the mental disease digital therapy product, a special testing and evaluating system is constructed, objective indexes and output contents of the digital therapy product are tested, and in the interaction process of a patient and the product, the safety effectiveness of the product is tested. By detecting data such as micro-expression, electroencephalogram signals and eye movement signals of a user in real time, feedback of a patient is obtained in real time, corresponding relation analysis is conducted on the feedback and output content of a digital therapy product, therefore, the treatment effect of the product can be accurately evaluated, and the accuracy of mental disease diagnosis and treatment is improved.

Description

Technical Field

[0001] This invention relates to the field of mental illness technology, specifically to a method for testing and evaluating the safety and effectiveness of digital therapy products for mental illness. Background Technology

[0002] The lifetime prevalence of various mental illnesses among adults in my country is as high as 16.60%, with depressive disorders, schizophrenia, and dementia ranking among the top three. Due to insufficient awareness and social stigma, the diagnosis and treatment rates for mental illnesses in my country are low, with approximately 92% of mental illness patients never receiving diagnosis or treatment. Most mental illnesses do not rely on instrumental examinations or biochemical tests; routine diagnosis and treatment depend on communication between doctors and patients, with doctors observing the patient's behavior and expressions, relying on their strong theoretical knowledge and practical experience to make judgments. Therefore, the probability of misdiagnosis and missed diagnosis is extremely high.

[0003] Digital therapy refers to software-driven, evidence-based disease interventions for patients or healthcare institutions, used for disease prevention, management, and treatment. Digital therapy primarily targets mental and psychological disorders, endocrine-metabolic dysfunctions, motor-nervous system disorders, and visual impairments, utilizing evidence-based medicine theories such as cognitive behavioral therapy from psychology, biofeedback therapy from neuroscience, exercise therapy from rehabilitation medicine, and pharmacokinetics from pharmacology to provide digital disease interventions for patients.

[0004] Digital therapy products for mental illness utilize electronic devices such as computers and wearable devices, employing technologies like extended reality, artificial intelligence, and the Internet of Things to assess and intervene in mental illnesses or symptoms. Compared to traditional methods, they have been proven to have better diagnostic and treatment effects for various mental disorders. However, my country currently lacks specific methods and tools for evaluating the safety and effectiveness of such products. Traditional testing methods do not consider the differences in cognitive intervention, patient-device interaction, and the specific characteristics of the target population, thus the safety and effectiveness of these products cannot be guaranteed. Summary of the Invention

[0005] The purpose of this invention is to provide a method for testing and evaluating the safety and effectiveness of digital therapy products for mental illnesses, in order to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for testing and evaluating the safety and effectiveness of digital therapy products for mental illness, comprising the following steps: S1. Objective evaluation of digital therapy products: Safety indicators are tested for the three main testing modules involved in digital therapy for mental illness: virtual reality, audio and video interaction, and functionality. S2. Human-computer interaction feedback test: By capturing micro-expressions, electroencephalogram (EEG) signals, and eye movement signals, the patient's emotional state is analyzed to determine whether the patient uses the treatment product as planned and whether there are any distractions that could reduce the effectiveness of the treatment. S3. Measurement of the mapping between product output content and patient feedback: Construct a product evaluation system by analyzing the product's real-time data output and patient feedback information in real time; S4. Real-time adjustment and optimization: Based on the evaluation results, adjust the prescription of digital therapy products in real time (such as changing training content, difficulty, etc.) and scenario design (the product development team can optimize the scenario design, interaction methods, etc. of the product based on the evaluation results to improve the therapeutic effect and user experience of the product).

[0007] Furthermore, the digital therapy product interacts with patients through a digital training program to improve symptoms. The therapeutic effect of the product depends on the rationality of the digital prescription and whether the patient completes the training as required by the prescription.

[0008] Furthermore, in step S1, the security indicator test specifically includes: 1) Virtual Reality Module Testing Verify the safety and reliability of virtual reality (VR) technology in terms of immersion, interactivity, and multi-sensory perception, avoiding patient discomfort or secondary harm; the test content includes: Hardware safety check: Ensure that VR devices (such as headsets and controllers) meet standards for electromagnetic radiation and wearing comfort; Scene content evaluation: Testing the rationality of the scene in the virtual environment; Physiological indicator monitoring: The patient's physiological state is monitored in real time through sensors such as heart rate and skin conductance to ensure that the immersive experience does not cause excessive anxiety or physiological abnormalities; Interaction logic verification: Check whether the user's interaction feedback with the virtual environment is natural; 2) Audio and video interaction module testing Ensure the security of audio and video content during interaction, including psychological compatibility and technical stability; the test content includes: Audio safety testing: assessing whether background music, voice guidance, and other content are appropriate for the treatment scenario; Video content review: Check whether the colors and rhythm of the video are suitable for the patient's psychological state; Interactive response testing: Verify the accuracy and latency of voice recognition or touch interaction to ensure that commands are executed precisely; Fatigue test: Simulate long-term use scenarios to assess the psychological fatigue caused by the complexity of interaction, and thus determine whether the operation steps need to be simplified. 3) Functional module testing Verify the clinical effectiveness and operational safety of core therapeutic functions (such as cognitive training and behavioral intervention); the test content includes: Algorithm logic verification: Testing whether AI-driven intervention programs (such as CBT modules) conform to evidence-based medicine standards; Data security testing: Ensure that patient data is encrypted during collection, transmission and storage, and complies with medical data compliance requirements (GDPR or China's Data Security Law). Clinical efficacy control: The efficacy of the functional modules was verified through randomized controlled trials (RCTs). Anomaly handling test: Simulate scenarios such as network outages and equipment failures to check whether the system can safely interrupt or resume the treatment process.

[0009] Furthermore, step S2 includes the following sub-steps: S21. Analysis of patient emotional changes using electroencephalogram (EEG) signals: Baseline data acquisition: The patient wears the head-mounted EEG device correctly, ensuring good contact between the device electrodes and the scalp. Record the patient's EEG waveforms (such as alpha waves and beta waves) at rest before treatment begins to establish a personalized baseline. Emotion feature extraction and real-time decoding: Fourier transform is used to extract the energy of frequency bands such as theta waves (4-8Hz, anxiety-related) and beta waves (13-30Hz, tension-related) to quantify emotional fluctuations. A pre-trained LSTM model is used to map EEG features to emotion labels (such as relaxation, distraction, anxiety) and output the emotional state in real time. Dynamic intervention strategy: When a persistent abnormality is detected (such as theta wave energy exceeding the threshold), the treatment content is automatically adjusted (such as simplifying the difficulty of the VR scene) or voice guidance is triggered; S22, Eye Movement Signal Monitoring Treatment Quality: Basic eye movement parameter acquisition: The duration and trajectory of the patient's gaze on a specific area of ​​the screen (such as a treatment instruction button) are tracked using an infrared eye tracker to analyze attention concentration and monitor changes in pupil diameter in real time. Pupil dilation reflects increased cognitive load, while pupil constriction reflects fatigue or distraction. Treatment adherence assessment: When a patient gazes at a non-treatment-related area (such as the edge of an interface) for more than a preset proportion of time, it is considered distraction. At the same time, eye-tracking thermography is used to verify whether the patient has completed the key steps of the procedure (such as target clicking in cognitive training). Real-time feedback and correction: When eye movement data indicates that the operation is not performed as required (such as not focusing on the core content for a long time), a voice prompt is triggered (such as "Please return to the task area"). S23. Micro-expression capture assists in emotion verification: Facial Action Unit (AU) Recognition: Capture subtle facial expressions (such as raised eyebrows and tightened corners of the mouth) using a camera, use a convolutional neural network to identify AU combinations (such as AU4 frowning to indicate anxiety), and perform emotional conflict detection with EEG / eye movement data. Before and after treatment: Record changes in micro-expressions during treatment and assess the effect of mood improvement (e.g., increased frequency of smiles in depressed patients).

[0010] Furthermore, in step S23, the emotion conflict detection involves ensuring that the timestamps of the EEG, eye movement, and micro-expression data are consistent. The micro-expression data is compared with the EEG / eye movement data. If the results are consistent, the emotion verification is deemed successful. If the results are contradictory (e.g., the EEG shows relaxation but the micro-expression shows anxiety), manual review is required.

[0011] Furthermore, step S3 includes the following sub-steps: S31. Data Collection: Collect real-time data generated during the use of digital therapy products, such as rendering frame rate, response time, video synchronization, etc., as well as patient feedback information, including subjective feelings and symptom improvement. S32. Mapping Analysis: Perform time-series correspondence analysis between real-time product data and patient feedback information to study the impact of product characteristics on patient treatment outcomes. S33. Evaluation System Construction: Based on the mapping analysis results, construct a comprehensive evaluation system for digital therapy products.

[0012] Furthermore, in step S32, the data mapping analysis, based on time series analysis and correlation analysis, uses statistical methods to study the association between product characteristics and patient treatment effects. Specifically: Data preprocessing: Real-time data from the product side (such as frame rate per second, API response time) and patient feedback data (such as daily symptom scores, treatment task completion rate) are aligned with a unified time axis to construct a time-series correspondence matrix. For response time loss caused by network interruption, forward imputation combined with the mean of data from similar patients is used for interpolation. Temporal correlation modeling: The DTW algorithm is used to handle the temporal misalignment between patient feedback delay and product performance fluctuation, and a vector autoregression model is constructed to verify whether the increase in product delay significantly leads to the deterioration of patient symptom scores; Multidimensional correlation analysis: The linear correlation coefficient between product performance indicators (frame rate stability, video synchronization error) and patient outcome indicators (PHQ-9 depression scale score reduction rate) was calculated to identify strongly correlated dimensions. For non-normally distributed data (such as patients' daytime mood fluctuations), Spearman rank correlation analysis was used to analyze the monotonic relationship between product response time and treatment adherence.

[0013] Furthermore, in step S33, the evaluation dimensions of the comprehensive evaluation system include the product's own characteristics, the rationality of the digital prescription issued by the doctor, and the standardization of the patient's usage process, and the evaluation methods for each dimension are as follows: Product performance evaluation: The product's performance and stability are evaluated by testing parameters such as rendering frame rate, response time, and video synchronization. Digital prescription rationality assessment: Based on the patient's condition and feedback, assess whether the digital prescriptions issued by the doctor are reasonable and effective; Patient usage process standardization assessment: By analyzing patients' usage records and feedback, we assess whether patients have completed the training as required by the prescription, as well as the standardization and compliance during the usage process.

[0014] Furthermore, the evaluation system is constructed using the Analytic Hierarchy Process (AHP) to synthesize the evaluation results from the three dimensions, forming an overall evaluation result for the digital therapy product. The CR value (consistency ratio) requirement is <0.1. The weight allocation of each indicator is determined through expert scoring as follows: Product characteristic = 0.45; Technical reliability = 0.6 x 0.45 = 0.27; Clinical fit = 0.4 x 0.45 = 0.18; Prescription rationality = 0.35; Use specification = 0.20.

[0015] Furthermore, in step S4, a digital prescription adjustment is triggered by a preset threshold. The adjustment is based on patient feedback (using real-time data such as micro-expressions, EEG signals, and eye movement signals to determine whether the current prescription triggers negative emotions, such as excessive tension) and efficacy data (comparing symptom scores before and after treatment, such as the PHQ-9 depression scale; if improvement is insufficient, the intervention intensity is increased). The adjustment strategies include, but are not limited to: Content difficulty: If a patient fails to complete the interactive task three times in a row, the task complexity will be automatically reduced and the number of dialogue rounds will be decreased. Intervention frequency: If the EEG signal shows a decrease in alpha waves (indicating insufficient relaxation), increase the frequency of mindfulness training, such as from once a day to twice a day; Prescription type: If eye tracking reveals that the patient’s pupil dilation lasts for more than 5 seconds during virtual exposure therapy, it is determined that the scene reaction is excessive, and the system is switched to progressive exposure (starting from a low-altitude scene).

[0016] This invention provides a method for testing and evaluating the safety and effectiveness of digital therapy products for mental illness, which has the following beneficial effects: This invention targets digital therapy products for mental illnesses. By constructing a specialized testing and evaluation system, it assesses the objective indicators and output content of the digital therapy products. During patient interaction with the product, it monitors user micro-expressions, electroencephalogram (EEG) signals, and eye-tracking signals in real time to obtain patient feedback. This feedback is then correlated with the output content of the digital therapy product, allowing for accurate evaluation of the product's therapeutic effect and improving the accuracy of mental illness diagnosis and treatment. Furthermore, by analyzing patient feedback in real time and adjusting the prescription and scenario design of the digital therapy product based on the evaluation results, this method better meets the specific needs of patients and further enhances treatment effectiveness. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the steps of a method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to the present invention. Figure 2 This is a logical diagram illustrating the safety and effectiveness testing and evaluation method for a digital therapy product for mental illness according to the present invention. Figure 3 This is a schematic diagram of a digital therapy product module for a method of testing and evaluating the safety and effectiveness of a digital therapy product for mental illness, as described in this invention. Figure 4 This is a schematic diagram of the operation process of the safety and effectiveness testing and evaluation method for a digital therapy product for mental illness according to the present invention; Figure 5 This is a schematic diagram of the digital therapy process for mental illness, which is a method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to the present invention. Detailed Implementation

[0018] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0019] like Figures 1-5 As shown, a method for testing and evaluating the safety and effectiveness of digital therapy products for mental illness includes the following steps: S1. Objective Evaluation of Digital Therapy Products: Safety indicators are tested across three main testing modules related to digital therapies for mental illnesses: virtual reality, audio-visual interaction, and functionality. Digital therapy products interact with patients through digital training programs to improve symptoms; the therapeutic effect depends on the rationality of the digital prescription and whether the patient completes the training as required by the prescription.

[0020] 1) Virtual Reality Module Testing Verify the safety and reliability of virtual reality (VR) technology in terms of immersion, interactivity, and multi-sensory perception, avoiding patient discomfort or secondary harm; the test content includes: Hardware safety check: Ensure that VR devices (such as headsets and controllers) meet standards for electromagnetic radiation and wearing comfort; Scene content evaluation: Test the rationality of the scene in the virtual environment. For example, traumatic scenes in exposure therapy need to be presented in a graded manner to avoid directly triggering strong stress reactions in patients. Physiological indicator monitoring: The patient's physiological state is monitored in real time through sensors such as heart rate and skin conductance to ensure that the immersive experience does not cause excessive anxiety or physiological abnormalities; Interaction logic verification: Check whether the user's interaction feedback with the virtual environment is natural, such as whether the action delay or visual deviation is within the safety threshold.

[0021] 2) Audio and video interaction module testing Ensure the security of audio and video content during interaction, including psychological compatibility and technical stability; the test content includes: Audio safety testing: Evaluate whether background music, voice guidance, and other content are suitable for the therapeutic scenario. For example, sleep aid audio should avoid high-frequency noise or abrupt sound effects to prevent causing irritability. Video content review: Check whether the color and rhythm of the video are suitable for the patient's psychological state. For example, patients with anxiety should avoid high-frequency flickering images. Interactive response testing: Verify the accuracy and latency of voice recognition or touch interaction to ensure that commands are executed precisely; Fatigue test: Simulate long-term use scenarios to assess the psychological fatigue caused by the complexity of interaction, and thus determine whether the operation steps need to be simplified.

[0022] 3) Functional module testing Verify the clinical effectiveness and operational safety of core therapeutic functions (such as cognitive training and behavioral intervention); the test content includes: Algorithm logic verification: Testing whether AI-driven intervention programs (such as CBT modules) conform to evidence-based medicine standards; Data security testing: Ensure that patient data is encrypted during collection, transmission and storage, and complies with medical data compliance requirements (GDPR or China's Data Security Law). Clinical efficacy comparison: The efficacy of functional modules is verified through RCTs. For example, the social anxiety treatment module requires comparison of patient behavior data before and after virtual exposure. Anomaly handling test: Simulate scenarios such as network outages and equipment failures to check whether the system can safely interrupt or resume the treatment process.

[0023] S2. Human-computer interaction feedback test: By capturing micro-expressions, EEG signals, and eye movement signals, the patient's emotional state is analyzed to determine whether the patient uses the treatment product as planned and whether there are any distractions that could reduce the effectiveness of the treatment.

[0024] S21. Analysis of patient emotional changes using electroencephalogram (EEG) signals: Baseline data acquisition: The patient wears the head-mounted EEG device correctly, ensuring good contact between the device electrodes and the scalp. Record the patient's EEG waveforms (such as alpha waves and beta waves) at rest before treatment begins to establish a personalized baseline.

[0025] Emotion feature extraction and real-time decoding: Fourier transform is used to extract the energy of frequency bands such as theta waves (4-8Hz, anxiety-related) and beta waves (13-30Hz, tension-related) to quantify emotional fluctuations. A pre-trained LSTM model is used to map EEG features to emotion labels (such as relaxation, distraction, anxiety) and output the emotional state in real time.

[0026] Dynamic intervention strategy: When a persistent abnormality is detected (such as theta wave energy exceeding the threshold), the treatment content is automatically adjusted (such as simplifying the difficulty of the VR scene) or voice guidance is triggered.

[0027] S22, Eye Movement Signal Monitoring Treatment Quality: Basic eye movement parameter acquisition: The duration and trajectory of the patient's gaze on a specific area of ​​the screen (such as a treatment instruction button) are tracked using an infrared eye tracker to analyze attention concentration and monitor changes in pupil diameter in real time. Pupil dilation reflects increased cognitive load, while pupil constriction reflects fatigue or distraction.

[0028] Treatment adherence assessment: When a patient gazes at a non-treatment-related area (such as the edge of an interface) for more than a preset proportion of time, it is considered distraction. At the same time, eye-tracking thermography is used to verify whether the patient has completed key steps of the procedure (such as target clicking in cognitive training).

[0029] Real-time feedback and correction: When eye movement data indicates that the operation is not performed as required (such as not focusing on the core content for a long time), a voice prompt is triggered (such as "Please return to the task area").

[0030] S23. Micro-expression capture assists in emotion verification: Facial Action Unit (AU) Recognition: The system captures subtle facial expressions (such as raised eyebrows and tightened lips) using a camera, identifies AU combinations (such as AU4 frowning to indicate anxiety) using a convolutional neural network, and performs emotion conflict detection with EEG / eye movement data to ensure that the timestamps of EEG, eye movement, and micro-expression data are consistent. The system compares the micro-expression data with the EEG / eye movement data. If the results are consistent, the emotion verification is considered successful. If the results are contradictory (such as EEG showing relaxation but micro-expression showing anxiety), manual review is required.

[0031] Before and after treatment: Record changes in micro-expressions during treatment and assess the effect of mood improvement (e.g., increased frequency of smiles in depressed patients).

[0032] S3. Measurement of the mapping between product output content and patient feedback: By analyzing the product's real-time data output and patient feedback information, a product evaluation system is constructed.

[0033] S31. Data Collection: Collect real-time data generated during the use of digital therapy products, such as rendering frame rate, response time, video synchronization, etc., as well as patient feedback information, including subjective feelings and symptom improvement.

[0034] S32. Mapping Analysis: This step involves performing a time-series correlation analysis between real-time product data and patient feedback information to study the impact of product characteristics on patient treatment outcomes. In this step, data mapping analysis, based on time-series analysis and correlation analysis, uses statistical methods to study the association between product characteristics and patient treatment outcomes. Specific operations are as follows: Data preprocessing: Real-time data from the product side (such as frame rate per second, API response time) and patient feedback data (such as daily symptom scores, treatment task completion rate) are aligned with a unified time axis to construct a time-series correspondence matrix. For response time loss caused by network interruption, forward imputation combined with the mean of data from similar patients is used for interpolation. Temporal correlation modeling: The DTW algorithm is used to handle the temporal misalignment between patient feedback delay and product performance fluctuation. For example, the correlation between system lag the day before and patient treatment interruption the next day is analyzed, and a vector autoregression (VAR) model is constructed to verify whether the increase in product latency significantly leads to the deterioration of patient symptom scores. Multidimensional correlation analysis: The linear correlation coefficient between product performance indicators (frame rate stability, video synchronization error) and patient outcome indicators (PHQ-9 depression scale score reduction rate) was calculated to identify strongly correlated dimensions. For non-normally distributed data (such as patients' daytime mood fluctuations), Spearman rank correlation analysis was used to analyze the monotonic relationship between product response time and treatment adherence.

[0035] S33. Evaluation System Construction: Based on the mapping analysis results, a comprehensive evaluation system for digital therapy products is constructed. In this step, the evaluation dimensions of the comprehensive evaluation system include the product's inherent characteristics, the rationality of the digital prescriptions issued by doctors, and the standardization of patient usage. The evaluation methods for each dimension are as follows: Product performance evaluation: The product's performance and stability are evaluated by testing parameters such as rendering frame rate, response time, and video synchronization. Digital prescription rationality assessment: Based on the patient's condition and feedback, assess whether the digital prescriptions issued by the doctor are reasonable and effective; Patient usage process standardization assessment: By analyzing patients' usage records and feedback, we assess whether patients have completed the training as required by the prescription, as well as the standardization and compliance during the usage process.

[0036] In this embodiment, the evaluation system is constructed using the Analytic Hierarchy Process (AHP) to synthesize the evaluation results of the three dimensions mentioned above, forming an overall evaluation result for the digital therapy product. The CR value (Conformity Ratio) requirement is <0.1, and the weight allocation of each indicator is determined through expert scoring as follows: Product characteristic = 0.45; Technical reliability = 0.6 x 0.45 = 0.27; Clinical fit = 0.4 x 0.45 = 0.18; Prescription rationality = 0.35; Use specification = 0.20.

[0037] The calculation example for this embodiment is as follows: Product characteristic score: Technical reliability 8.2×0.27 + Clinical compatibility 7.5×0.18=3.56; Prescription rationality score: indication matching 9.0 × 0.14 + dosage rationality 8.5 × 0.105 + treatment course setting 8.0 × 0.105 = 3.58; Score for adherence to standards: Time compliance 8.8 × 0.10 + Operating standards 7.2 × 0.06 + Attention 6.5 × 0.04 = 1.48; Overall score = 3.56 + 3.58 + 1.48 = 8.62 / 10.

[0038] S4. Real-time adjustment and optimization: Based on the evaluation results, adjust the prescription of digital therapy products in real time (such as changing training content, difficulty, etc.) and scenario design (the product development team can optimize the scenario design, interaction methods, etc. of the product based on the evaluation results to improve the therapeutic effect and user experience of the product).

[0039] In this embodiment, digital prescription adjustments are triggered by a preset threshold. The adjustments are based on patient feedback (using real-time data such as micro-expressions, EEG signals, and eye-tracking signals to determine if the current prescription triggers negative emotions, such as excessive tension) and efficacy data (comparing symptom scores before and after treatment, such as the PHQ-9 depression scale; if improvement is insufficient, the intervention intensity is increased). Adjustment strategies include, but are not limited to: Content difficulty: If a patient fails to complete the interactive task three times in a row, the task complexity will be automatically reduced and the number of dialogue rounds will be decreased. Intervention frequency: If the EEG signal shows a decrease in alpha waves (indicating insufficient relaxation), increase the frequency of mindfulness training, such as from once a day to twice a day; Prescription type: If eye tracking reveals that the patient’s pupil dilation lasts for more than 5 seconds during virtual exposure therapy, it is determined that the scene reaction is excessive, and the system is switched to progressive exposure (starting from a low-altitude scene).

[0040] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A method for testing and evaluating the safety and effectiveness of digital therapy products for mental illness, characterized in that, Includes the following steps: S1. Objective evaluation of digital therapy products: Safety indicators are tested for the three main testing modules involved in digital therapy for mental illness: virtual reality, audio and video interaction, and functionality. S2. Human-computer interaction feedback test: By capturing micro-expressions, electroencephalogram (EEG) signals, and eye movement signals, the patient's emotional state is analyzed to determine whether the patient is using the treatment product as planned. S3. Measurement of the mapping between product output content and patient feedback: Construct a product evaluation system by analyzing the product's real-time data output and patient feedback information in real time; S4. Real-time Adjustment and Optimization: Adjust the prescription and scenario design of digital therapy products in real time based on the evaluation results.

2. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 1, characterized in that, The digital therapy product interacts with patients through a digital training program to improve symptoms. The therapeutic effect of the product depends on the rationality of the digital prescription and whether the patient completes the training as required by the prescription.

3. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 1, characterized in that, In step S1, the security indicator test specifically includes: 1) Virtual Reality Module Testing Verify the safety and reliability of virtual reality technology's immersion, interactivity, and multi-sensory aspects; the test content includes: Hardware safety check: Ensure VR devices meet electromagnetic radiation and wearing comfort standards; Scene content evaluation: Testing the rationality of the scene in the virtual environment; Physiological indicator monitoring: Real-time monitoring of the patient's physiological status through sensors; Interaction logic verification: Check whether the user's interaction feedback with the virtual environment is natural; 2) Audio and video interaction module testing Ensure the security of audio and video content during interaction, including psychological compatibility and technical stability; the test content includes: Audio safety testing: assessing whether the background music and voice guidance content are appropriate for the treatment scenario; Video content review: Check whether the colors and rhythm of the video are suitable for the patient's psychological state; Interactive response test: Verify the accuracy and latency of voice recognition or touch interaction; Fatigue test: Simulate long-term usage scenarios to assess the psychological fatigue caused by the complexity of interaction; 3) Functional module testing Verify the clinical effectiveness and operational safety of the core therapeutic functions; the test content includes: Algorithm logic verification: Testing whether AI-driven intervention programs meet evidence-based medicine standards; Data security testing: Ensures that patient data is encrypted during collection, transmission, and storage, and complies with medical data compliance requirements; Clinical efficacy control: The efficacy of the functional modules was verified through randomized controlled trials (RCTs). Anomaly handling test: Simulate network outage and equipment failure scenarios to check whether the system can safely interrupt or resume the treatment process.

4. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 1, characterized in that, Step S2 includes the following sub-steps: S21. Analysis of patient emotional changes using electroencephalogram (EEG) signals: Baseline data acquisition: Record the patient's EEG waveforms at rest before treatment begins to establish a personalized baseline; Emotion feature extraction and real-time decoding: Fourier transform is used to extract the energy of frequency bands such as theta waves and beta waves to quantify emotional fluctuations, and an LSTM model is used to map EEG features to emotion labels and output emotional states in real time. Dynamic intervention strategy: When a persistent abnormality is detected, the treatment content is automatically adjusted or voice guidance is triggered; S22, Eye Movement Signal Monitoring Treatment Quality: Basic eye movement parameter acquisition: The duration and trajectory of the patient's gaze at a specific area of ​​the screen are tracked using an infrared eye tracker to analyze attention concentration and monitor changes in pupil diameter in real time. Pupil dilation reflects increased cognitive load, while pupil constriction reflects fatigue or distraction. Treatment compliance assessment: When a patient gazes at a non-treatment-related area for a predetermined proportion of time, it is considered distraction. At the same time, eye movement thermography is used to verify whether the patient has completed the key steps of the procedure. Real-time feedback and correction: When eye movement data indicates that the operation has not been performed as required, a voice reminder is triggered; S23. Micro-expression capture assists in emotion verification: Facial Action Unit (AU) recognition: Captures subtle facial expressions using a camera, identifies AU combinations using a convolutional neural network, and performs emotional conflict detection by combining EEG / eye-tracking data. Before and after treatment: Record changes in micro-expressions during treatment to assess the effect of mood improvement.

5. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 4, characterized in that, In step S23, emotion conflict detection involves ensuring that the timestamps of EEG, eye movement, and micro-expression data are consistent. The micro-expression data is compared with the EEG / eye movement data. If the results are consistent, the emotion verification is deemed successful. If the results are contradictory, manual review is required.

6. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 1, characterized in that, Step S3 includes the following sub-steps: S31. Data Collection: Collect real-time data generated during the use of digital therapy products, as well as patient feedback information; S32. Mapping Analysis: Perform time-series correspondence analysis between real-time product data and patient feedback information to study the impact of product characteristics on patient treatment outcomes. S33. Evaluation System Construction: Based on the mapping analysis results, construct a comprehensive evaluation system for digital therapy products.

7. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 6, characterized in that, In step S32, the data mapping analysis, based on time series analysis and correlation analysis, uses statistical methods to study the association between product characteristics and patient treatment effects. Specific operations are as follows: Data preprocessing: Real-time data from the product side and patient feedback data are aligned with a unified timeline to construct a time-series correspondence matrix. For response time loss caused by network interruption, forward imputation combined with the mean of data from similar patients is used for interpolation. Temporal correlation modeling: The DTW algorithm is used to handle the temporal misalignment between patient feedback delay and product performance fluctuation, and a vector autoregression model is constructed to verify whether the increase in product delay significantly leads to the deterioration of patient symptom scores; Multidimensional correlation analysis: Calculate the linear correlation coefficient between product performance indicators and patient outcome indicators, identify strongly correlated dimensions, and for non-normally distributed data, use Spearman rank correlation analysis to analyze the monotonic relationship between product response time and treatment adherence.

8. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 6, characterized in that, In step S33, the evaluation dimensions of the comprehensive evaluation system include the product's own characteristics, the rationality of the digital prescription issued by the doctor, and the standardization of the patient's use process, and the evaluation methods for each dimension are as follows: Product performance evaluation: The product's performance and stability are evaluated by testing its rendering frame rate, response time, and video synchronization parameters. Digital prescription rationality assessment: Based on the patient's condition and feedback, assess whether the digital prescriptions issued by the doctor are reasonable and effective; Patient usage process standardization assessment: By analyzing patients' usage records and feedback, we assess whether patients have completed the training as required by the prescription, as well as the standardization and compliance during the usage process.

9. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 8, characterized in that, The evaluation system is constructed using the analytic hierarchy process (AHP) to synthesize the evaluation results from the three dimensions, forming an overall evaluation result for the digital therapy product. The CR value is required to be <0.1, and the weight allocation of each indicator is as follows: Product characteristic = 0.45; Technical reliability = 0.6 x 0.45 = 0.27; Clinical fit = 0.4 x 0.45 = 0.18; Prescription rationality = 0.35; Use specification = 0.

20.

10. The method for testing and evaluating the safety and effectiveness of a digital therapy product for mental illness according to claim 1, characterized in that, In step S4, digital prescription adjustments are triggered by a preset threshold. The adjustment is based on patient feedback and efficacy data, and the adjustment strategies include, but are not limited to: Content difficulty: If a patient fails to complete the interactive task three times in a row, the task complexity will be automatically reduced and the number of dialogue rounds will be decreased. Intervention frequency: If the EEG signal shows a decrease in alpha waves, increase the frequency of mindfulness training; Prescription type: If eye tracking reveals that the patient's pupil dilation lasts for more than 5 seconds during virtual exposure therapy, it is determined that the scene reaction is excessive, and the system is switched to progressive exposure.

Images