An ePRO scale real-time consistency verification method and device, a terminal device and a computer readable storage medium

Abstract

The application discloses an ePRO scale real-time consistency verification method, device and equipment and a computer readable storage medium, and belongs to the technical field of medical informatization. The method comprises the following steps: acquiring physiological data by calling corresponding sensors according to the type of the question and extracting characteristic values; in walking or dexterity verification, the inconsistency degree of the subjective description and the objective characteristic value is calculated based on a verification table, and if the condition is met, an abnormality is fed back; in cognitive verification, cognitive fatigue or operation confusion state is identified based on the characteristic value and feedback is given. Through multi-modal implicit perception and classification verification, the application effectively solves the problems that the subjective description is disconnected with the objective state, real-time verification is lacked and answer abnormality is difficult to identify.

Description

Technical Field

[0001] This invention relates to the field of medical information technology, and in particular to a method, system, electronic device, and computer-readable storage medium for real-time consistency verification of the ePRO scale based on multimodal intelligent terminal sensor data. Background Technology

[0002] In the field of healthcare and physiological function assessment, the electronic patient-reported outcomes (ePRO) scale is widely used to assess users' health status. When completing the scale, users typically need to select subjective descriptions that reflect their motor control or cognitive abilities for subsequent analysis.

[0003] However, existing ePRO scale systems suffer from a serious disconnect between subjective and objective factors. Due to users' psychological adaptability or other subjective factors, their selected subjective state descriptions often fail to accurately reflect their objective physiological function. Furthermore, existing verification methods lack differentiated sensor configuration mechanisms for different item types (such as cognitive, gait, and dexterity tests), and there is no rigorous logical mapping between sensor data and the scale's item levels. This makes it impossible to use objective feature values ​​to perform real-time, quantitative accuracy verification of subjective descriptions. In addition, existing technologies struggle to identify abnormal states such as cognitive fatigue or operational confusion that are inconsistent with the user's response behavior in real time during the answering process, failing to provide anomaly alerts at the data entry point, thus reducing the objectivity and reliability of ePRO scale data. Summary of the Invention

[0004] This invention provides a real-time consistency verification method for the ePRO scale, which can solve the problems in the prior art such as the disconnect between the user's subjective state description and objective physiological function state, the lack of a targeted real-time verification mechanism, and the difficulty in identifying abnormal responses in real time.

[0005] An embodiment of the present invention provides a method for real-time consistency verification of the ePRO scale, comprising: The system obtains the question type of the current item in the ePRO scale, selects the corresponding sensor based on the question type, and synchronously collects sensor data reflecting the user's objective physiological function status based on the selected sensor while the user answers the current question. The question types include: cognitive verification, walking verification, and dexterity verification. Feature extraction is performed on the sensor data to obtain real-time feature values; When the question type is walking verification or dexterity verification, the real-time subjective state description selected by the user when answering the current question to reflect the motion control ability is obtained; based on the verification table corresponding to the question type, the expected feature value range corresponding to the real-time subjective state description is determined, the degree of deviation of the real-time feature value from the expected feature value range is calculated, and the degree of inconsistency is obtained; when the degree of inconsistency meets the preset triggering condition, an abnormal prompt is fed back. When the question type is cognitive verification, it is determined whether the real-time feature value meets the preset cognitive anomaly judgment condition; if it does, an anomaly prompt is fed back; the cognitive anomaly judgment condition is used to identify whether the user has cognitive fatigue or operational confusion that is inconsistent with the answering behavior.

[0006] Furthermore, the step of selecting the corresponding sensor according to the question type, and simultaneously collecting sensor data reflecting the user's objective physiological functional state according to the selected sensor while the user answers the current question, includes: When the question type is a dexterity test, the accelerometer and gyroscope are used to collect acceleration and angular velocity signals reflecting hand movements, and the touchscreen sensor is used to collect touch trajectory. When the question type is a walking test, the barometric pressure sensor is used to collect height change data, the pedometer is used to collect step count data, and historical gait statistics cached within a preset time period are obtained. The historical gait statistics include swing time records for each of the two limbs. When the question type is a cognitive test, the touchscreen sensor is invoked to collect the touchscreen click timestamp data during the user's answering process.

[0007] Further, the step of extracting features from the sensor data to obtain real-time feature values ​​includes: When the question type is agile verification, the acceleration signal and angular velocity signal are converted to the frequency domain and a normalized Fourier amplitude spectrum curve is calculated; the negative arc length of the normalized Fourier amplitude spectrum curve is taken as the spectral arc length; time-frequency analysis is performed on the acceleration signal to obtain the power spectral density, and the ratio of the energy of the preset target frequency band to the total energy of the entire frequency band is calculated to obtain the intention tremor index; and based on the touch trajectory, the distance between the touch point and the center of the target area is calculated to obtain the click deviation, and linear fitting analysis is performed on the touch trajectory to obtain the touch trajectory linearity; the spectral arc length, power spectral density, intention tremor index, click deviation, and touch trajectory linearity are taken as the real-time feature values; When the question type is walking verification, based on the historical gait statistical characteristics, the ratio of the standard deviation to the mean of the continuous gait cycle is calculated to obtain the stride length variability; based on the swing time records of each side of the limbs in the historical gait statistical characteristics, the absolute value of the difference between the swing time of the left limb and the swing time of the right limb is calculated, and the ratio of the absolute value of the difference to the mean of the swing time of both sides is calculated to obtain the asymmetry index; based on the height change data, the difference is calculated to obtain the height change amount, and based on the step count data, the number of steps within the preset monitoring time is counted to obtain the short-term step count; stride length variability, asymmetry index, height change amount, and short-term step count are used as the real-time feature values; When the question type is cognitive verification, based on the touch screen click timestamp data, the time interval between two adjacent keystrokes and the duration of a single keystroke from pressing to releasing are calculated to obtain the keystroke flight time and key hold time; the keystroke flight time and key hold time are used as the real-time feature values.

[0008] Furthermore, the verification table includes: a health status level, an expected subjective status description corresponding to each health status, an expected feature value range, and a trigger prompt condition; wherein, the trigger prompt condition is used to define the judgment logic for triggering an abnormal prompt when the real-time feature value deviates from the expected feature value range; The step of determining the expected feature value range corresponding to the real-time subjective state description based on the verification table corresponding to the question type includes: The real-time subjective state description is matched with each expected subjective state description in the verification table, and the expected subjective state description that matches the real-time subjective state description is taken as the target subjective state description. The expected feature value range corresponding to the subjective state description of the target is taken as the expected feature value range corresponding to the real-time subjective state description.

[0009] Further, the degree to which the real-time feature value deviates from the expected feature value range is calculated to obtain the degree of inconsistency. When the degree of inconsistency meets a preset trigger condition, an anomaly prompt is fed back, including: The real-time feature values ​​are converted into normalized objective functional scores using the Z-Score standardization method. The health status level corresponding to the real-time subjective state description is normalized into a numerical value to obtain the subjective function score. Calculate the absolute value of the difference between the subjective function score and the objective function score, and use the absolute value as the degree of inconsistency. When the degree of inconsistency is between the first warning threshold and the second warning threshold of the preset triggering conditions, it is determined that the soft conflict type prompt triggering condition in the preset triggering conditions is met, and a confirmation prompt box pops up on the interactive interface; wherein, the first warning threshold is less than the second warning threshold; When the degree of inconsistency is greater than the second warning threshold, it is determined that the hard conflict type prompt trigger condition in the preset trigger conditions is met, and the current submission operation is blocked or the current data is marked as untrustworthy in the interactive interface.

[0010] Furthermore, when the question type is cognitive verification, determining whether the preset cognitive anomaly judgment conditions are met based on the real-time feature value includes: Based on all key travel time and key hold time during the user's response process, the average key travel time, average key hold time, key travel time variance, and key hold time variance are calculated. When the average keystroke flight time or the average key hold time exceeds a preset fatigue threshold, it is determined that the user is experiencing cognitive fatigue. When the keystroke flight time variance or the key hold time variance exceeds a preset stability threshold, it is determined that the user is operating in a disorderly manner. When it is determined that the user is experiencing cognitive fatigue or the operation is chaotic, the preset cognitive abnormality determination conditions are met.

[0011] Furthermore, this embodiment also includes: Compare the real-time feature values ​​with the current historical mean corresponding to the individual feature baseline; If the absolute difference between the real-time feature value and the current historical mean is greater than the product of the current historical standard deviation corresponding to the personal feature baseline and the preset statistical coefficient, it is determined that the user has experienced a sudden change in status, and an abnormality prompt is given. Otherwise, the personal characteristic baseline is updated based on the real-time feature values; In the initial stage, a baseline of individual characteristics is established based on the characteristic value data of the nervous system disease group; Each time a user completes a valid ePRO scale response, the feature value data of the valid response is collected, and the number of valid responses is accumulated. When the number of valid responses reaches a preset number, the personal feature baseline is updated based on the feature value data of all valid responses. Based on the updated personal feature baseline, the updated historical mean and the updated historical standard deviation are calculated.

[0012] Another embodiment of the present invention provides an ePRO scale real-time consistency verification device, including: a data acquisition module, a feature extraction module, a motion verification module and a cognitive verification module; The data acquisition module is used to obtain the question type of the current item in the ePRO scale, select the corresponding sensor according to the question type, and synchronously collect sensor data reflecting the user's objective physiological function status according to the selected sensor while the user answers the current question; wherein, the question type includes: cognitive verification, walking verification, and dexterity verification. The feature extraction module is used to extract features from the sensor data to obtain real-time feature values; The motion verification module is used to obtain the real-time subjective state description that reflects the user's motion control ability when answering the current question, when the question type is walking verification or dexterity verification; based on the verification table corresponding to the question type, determine the expected feature value range corresponding to the real-time subjective state description, calculate the degree of deviation of the real-time feature value from the expected feature value range, obtain the degree of inconsistency, and when the degree of inconsistency meets the preset triggering condition, provide feedback of the abnormality prompt. The cognitive verification module is used to determine whether the real-time feature value meets the preset cognitive anomaly judgment conditions when the question type is cognitive verification; if it does, an anomaly prompt is fed back; the cognitive anomaly judgment conditions are used to identify whether the user has cognitive fatigue or operational confusion that is inconsistent with the answering behavior.

[0013] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the steps of the ePRO scale real-time consistency verification method of the present invention.

[0014] Another embodiment of the present invention provides a computer-readable storage medium item, including: a stored computer program, wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute the real-time consistency verification method of the ePRO scale of the present invention.

[0015] The embodiments of the present invention have the following beneficial effects: This invention proposes a real-time consistency verification method for the ePRO scale. First, the question type of the current item (including cognitive verification, walking verification, and dexterity verification) is obtained, and the corresponding sensor is dynamically selected to collect sensor data reflecting the user's objective physiological functional state and extract real-time feature values. During walking or dexterity verification, the expected feature value range corresponding to the user's real-time subjective state description is determined based on the verification table. The degree of inconsistency is calculated by measuring the deviation of the real-time feature value from the expected feature value range, and an anomaly prompt is fed back according to preset trigger conditions. During cognitive verification, real-time feature values ​​are used to identify whether the user has cognitive fatigue or operational confusion inconsistent with their answering behavior, and an anomaly prompt is fed back. This application dynamically selects sensors for different question types. Synchronous data collection ensures the targetedness and accuracy of objective physiological feature extraction from the technical source. In particular, this application constructs a classification verification mechanism: for walking and dexterity verification, by introducing a pre-set verification table with expected feature value ranges, it realizes the quantitative comparison between subjective state descriptions and objective real-time feature values, which can accurately identify data deviations caused by the disconnect between subjective and objective factors; for cognitive verification, by judging real-time feature values, it can effectively capture the hidden cognitive fatigue or operational confusion of users during the answering process. By integrating the verification process throughout the entire user answering process, calculating the degree of inconsistency in real time and judging cognitive abnormalities, this application puts data verification at the answering stage, significantly improving the objectivity and accuracy of ePRO scale data. Attached Figure Description

[0016] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a real-time consistency verification method for an ePRO scale provided in an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the device for a real-time consistency verification method for an ePRO scale provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0021] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0024] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0025] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0026] To address the problems in existing technologies such as the disconnect between users' subjective state descriptions and objective physiological function states, the lack of targeted real-time verification mechanisms, and the difficulty in identifying abnormal responses in real time, an embodiment of the present invention provides a real-time consistency verification method for the ePRO scale, comprising: Step S1: Obtain the question type of the current item in the ePRO scale, select the corresponding sensor according to the question type, and simultaneously collect sensor data reflecting the user's objective physiological function status according to the selected sensor while the user answers the current question; wherein, the question type includes: cognitive verification, walking verification and dexterity verification. In a preferred embodiment, the step of selecting the corresponding sensor according to the question type and synchronously collecting sensor data reflecting the user's objective physiological functional state according to the selected sensor when the user answers the current question includes: when the question type is a dexterity test, calling an accelerometer and a gyroscope to collect acceleration and angular velocity signals reflecting hand movements, and calling a touch screen sensor to collect touch trajectory; when the question type is a walking test, calling a barometric pressure sensor to collect altitude change data, calling a pedometer to collect step count data, and obtaining historical gait statistical features cached within a preset time period; wherein, the historical gait statistical features include swing time records corresponding to each of the two limbs; When the question type is a cognitive test, the touchscreen sensor is invoked to collect the touchscreen click timestamp data during the user's answering process.

[0027] In illustrative terms, the ePRO scale response process described in this embodiment refers to the process by which a patient, in an uncontrolled environment (such as a home environment), reads the questions of the HUI3 (Health Utilities Index Mark 3) general multi-attribute health status classification system displayed on a smart terminal and clicks on the options. The system backend adopts a context-aware implicit data acquisition strategy, that is, during the natural interaction of the user browsing questions and making selections, the sensor data stream related to the current question type is automatically triggered, without requiring the user to execute additional standardized physical fitness test instructions.

[0028] For dexterity verification, the application scenario corresponds to the questions in the ePRO scale that assess hand and finger function: When a user reads such questions on the device and prepares to click on the screen option, the system background captures the minute tremor signals of the user's hand at the moment of hovering and clicking through high-frequency sampling by the accelerometer and gyroscope. These signals are the acceleration and angular velocity signals reflecting hand movement, which are used for subsequent analysis to determine whether there is cerebellar intention tremor specific to neurodegenerative diseases (such as leukoencephalopathy). At the same time, the complete path of the finger from touching the screen to leaving the screen recorded by the touch screen sensor is the touch trajectory. This data reflects the accuracy and motor coordination of the user's click operation.

[0029] For walking verification, the application scenario corresponds to the questions in the ePRO scale that assess mobility or assistive device dependence. Considering that users may be stationary while answering questions, in addition to using barometric pressure sensors and pedometers for real-time posture change detection (such as determining whether the user stands up or walks during the question-answering process), the system also focuses on retrieving long-term data recorded in the mobile terminal coprocessor at low power consumption. Among them, the historical gait statistical characteristics refer to the walking parameters automatically cached by the device in the background over the past 24 to 48 hours, while the swing time records corresponding to each side of the limbs specifically refer to the duration of the left and right legs in the air swing phase during natural walking. This record is a key physiological indicator for assessing whether patients with peroneal muscular atrophy (CMT) have foot drop or bilateral muscle asymmetry.

[0030] For cognitive verification, the application scenario corresponds to the questions in the ePRO scale that assess memory, thinking, and learning abilities. In this scenario, the system focuses on keystroke dynamics. The touchscreen click timestamp data refers to the millisecond-level precise moments recorded by the system for each time a user's finger presses down and lifts up the screen. These timestamps are used to subsequently calculate keystroke intervals and keystroke durations, thereby keenly reflecting the user's reaction speed (whether there is cognitive fatigue) and the stability of motor control (whether there is operational confusion) during information processing. This achieves an objective mapping of the user's macroscopic cognitive state through micro-interaction behavior.

[0031] By dynamically calling specific sensor combinations for different question types, "targeted perception" and "implicit acquisition" of physiological signals can be achieved. Specifically, without increasing the user's additional testing burden, this method can accurately capture pathological features highly related to the current assessment dimension based on the answer context (such as hand tremors and click accuracy in dexterity testing, standing and gait stability features in walking testing, and reaction speed features in cognitive testing). This ensures the sensitivity of physiological function status assessment from the data source, laying a reliable data foundation for subsequent accurate verification that is consistent with both subjective and objective criteria.

[0032] Step S2: Extract features from the sensor data to obtain real-time feature values; In a preferred embodiment, the step of extracting features from the sensor data to obtain real-time feature values ​​includes: When the problem type is agility verification, the acceleration and angular velocity signals are converted to the frequency domain and the normalized Fourier amplitude spectrum curves are calculated; the negative value of the arc length of the normalized Fourier amplitude spectrum curve is taken as the spectral arc length (SPARC); the calculation formula is as follows: ; in, In this embodiment, the cutoff frequency is preferably set to 20Hz to cover the main frequency band of human free movement; The normalized Fourier amplitude spectrum is used to evaluate the smoothness of motion by measuring the complexity of the velocity spectrum curve. It is insensitive to the amplitude and duration of motion and has extremely high robustness. The acceleration signal is subjected to gravity removal processing and high-pass filtering. The power spectral density (PSD) is calculated using short-time Fourier transform (STFT), and the ratio of the energy in the preset target frequency band to the total energy across the entire frequency band is calculated to obtain the intention tremor index (ITI). The calculation formula is as follows: ; Among them, the target frequency band [ , The optimal frequency is 3-5Hz to accurately separate the cerebellar tremor component; the full frequency band [ , The preferred frequency is 1-20Hz, used to quantify the proportion of tremor energy in complex movements.

[0033] Furthermore, based on the touch trajectory, the distance between the touch point and the center of the target area is calculated to obtain the click deviation, and the touch trajectory is linearly fitted to obtain the touch trajectory linearity; the spectral arc length, power spectral density, intention tremor index, click deviation, and touch trajectory linearity are used as the real-time feature values. When the question type is walking verification, the stride duration variability (STV) is obtained by calculating the ratio of the standard deviation (SD) to the mean (Mean) of the continuous gait cycles based on the historical gait statistical characteristics; the specific calculation formula is as follows: ; Based on the swing time records of each limb in the historical gait statistics (left limb is set as...), The right side is set as The absolute value of the difference between the swing time of the left limb and the swing time of the right limb is calculated, and the ratio of the absolute value of the difference to the average swing time of both limbs is calculated to obtain the Asymmetry Index (ASI); the specific calculation formula is as follows: ; This indicator is used to assess the balance of bilateral limb movement. An ASI > 10% suggests significant unilateral muscle weakness or pain avoidance behavior.

[0034] The height change is calculated based on the difference in the height change data, and the short-term step count is obtained by counting the number of steps within a preset monitoring period based on the step count data; the step time variability, asymmetry index, height change, and short-term step count are used as the real-time feature values. When the question type is cognitive verification, based on the touch screen click timestamp data, the time interval between two adjacent keystrokes and the duration of a single keystroke from pressing to releasing are calculated to obtain the keystroke flight time and key hold time; the keystroke flight time and key hold time are used as the real-time feature values.

[0035] Specifically, when performing dexterity verification on simulated patients with significant cerebellar intention tremor, the corresponding algorithm was selected to process the collected acceleration and angular velocity signals. The calculated intention tremor index for this user was 0.42, indicating that the energy proportion in a specific frequency band far exceeded the normal range. At the same time, the calculated spectral arc length was -3.1, qualitatively reflecting severe coordination impairment and non-smoothness in the movement trajectory. In the scenario of simulating agent fraud, if a healthy person answers on behalf of a severely restricted patient, the calculated intention tremor index is only 0.02, and the linearity of the touch trajectory is greater than 0.95, showing extremely high movement stability and a near-straight-line movement trajectory. This provides core parameter support for subsequent identification of fraudulent behavior with "overly perfect data".

[0036] In the aforementioned feature extraction steps, a specific mathematical model is used to transform the raw sensor signals at the bottom layer into digital biomarker feature values ​​with clear physiological significance, achieving highly sensitive capture and quantification of subtle physiological dysfunctions. This method can not only accurately identify features such as pathological tremors or decreased smoothness of movement using frequency domain analysis, but also construct an objective physiological profile of the user from multiple dimensions by combining touch dynamics, gait variability, and keystroke characteristics. By normalizing, time-frequency transforming, and calculating deviations in the raw data, environmental noise and non-pathological motion interference can be effectively filtered out, ensuring that the extracted real-time feature values ​​can truly reflect the user's objective functional state. This provides a scientific and traceable data basis for the subsequent accurate quantification of the degree of inconsistency between subjective and objective factors and for identifying complex fraudulent behaviors such as concealing medical conditions or having agents fill out forms on behalf of the user.

[0037] Step S3: When the question type is walking verification or dexterity verification, obtain the real-time subjective state description selected by the user when answering the current question to reflect the motion control ability; based on the verification table corresponding to the question type, determine the expected feature value range corresponding to the real-time subjective state description, calculate the degree of deviation of the real-time feature value from the expected feature value range to obtain the degree of inconsistency, and when the degree of inconsistency meets the preset trigger condition, provide feedback on the abnormality prompt. Specifically, the verification table includes: health status level, expected subjective status description for each health status, expected feature value range, and trigger prompt conditions; wherein, the trigger prompt conditions are used to define the judgment logic for triggering an anomaly prompt when the real-time feature value deviates from the expected feature value range; The step of determining the expected feature value range corresponding to the real-time subjective state description based on the verification table corresponding to the question type includes: The real-time subjective state description is matched with each expected subjective state description in the verification table, and the expected subjective state description that matches the real-time subjective state description is taken as the target subjective state description. The expected feature value range corresponding to the subjective state description of the target is taken as the expected feature value range corresponding to the real-time subjective state description.

[0038] To illustrate, taking clever verification as an example, the specific contents of the verification table corresponding to the clever verification question are shown in the table below:

[0039] In one specific embodiment, firstly, based on the verification table, the user-selected subjective state description (corresponding to the health status level) is matched with the expected objective feature range, and the expected feature value range is compared with the real-time feature value to obtain a preliminary judgment result; in order to quantify the degree to which the real-time feature value deviates from the expected objective feature range, the following process is performed: First, the extracted real-time feature values ​​(including spectral arc length (SPARC), intention tremor index (ITI), step time variability (STV), and asymmetry index (ASI)) are standardized using the Z-Score standardization method to obtain the dimensionless standardized score corresponding to each real-time feature value: The standardized scoring formula for SPARC (Spectral Arc Length) is as follows: , The standardized scoring formula for the Intention Tremor Index (ITI) is as follows: , The standardized scoring formula for step-time variability (STV) is as follows: , The standardized scoring formula for the Asymmetry Index (ASI) is as follows: , in, and These are the historical mean and historical standard deviation for each real-time feature value of the current user. The historical mean measures the average level of the user's physiological indicator accumulated over a preset time period (or based on a baseline of a specific disease group), while the historical standard deviation measures the normal fluctuation range of the user's physiological indicator. The real-time feature values ​​are standardized to obtain the corresponding Z-score (i.e., standard score). This Z-score no longer represents the magnitude of the physical quantity but rather the degree to which the current state deviates from the individual's baseline. For example, if... A large amplitude indicates that the current vibration level far exceeds the user's usual fluctuation range, suggesting an anomaly. Subsequently, the sigmoid function is used to map the above standard scores to the 0-1 interval. In this interval, values ​​close to 0 represent "normal function / low deviation," and values ​​close to 1 represent "loss of function / extremely high deviation." This step is to transform the objective data into probability values ​​on the same scale as the subjective scores of the ePRO scale (also normalized to the 0-1 interval), so that the two can be subtracted. Then, a preset weight is combined. Perform weighted summation to construct an objective function score. The specific calculation formula is as follows: ; Among them, weight The sum is 1 and can be dynamically adjusted according to the question type; for example, when filling in the "dexterity check" question, the system considers hand function to be the core of the assessment, so it sets the weights ω1 (corresponding to spectral arc length / smoothness) and ω2 (corresponding to intention tremor) related to the hand to a higher value (e.g., each accounting for 0.5), while forcibly setting ω3 and ω4 related to the lower limbs to zero, in order to shield the noise interference of walking or sitting posture on the upper body assessment; when filling in the "walking check" question, a higher value is set. (Corresponding to step size variability) (Corresponding to the asymmetric index) weights (e.g., each accounting for 0.5), , Then set it to 0, ignoring hand features; At the same time, the system uses formulas The user-selected subjective levels are normalized into subjective function scores; finally, the scores are calculated using a formula. The absolute value of the difference between subjective and objective ratings is calculated to obtain the degree of inconsistency (inconsistency index I); Taking a scenario simulating a patient concealing their condition as an example, a subject with significant cerebellar tremor selected level 1 in the dexterity test due to psychological adaptability. The system determined the target subjective state as "completely normal," with the corresponding expected characteristic value range being "intention tremor index less than the threshold and smooth movement." Within seconds of the user's response, the real-time characteristic values ​​calculated by the backend were: Intention Tremor Index (ITI) equal to 0.42, and Spectral Arc Length (SPARC) equal to -3.1, significantly deviating from the expected range. The subjective score calculated using the above formula was 0, while the objective functional score was approximately 0.65, and the absolute value of the difference between the two, i.e., the degree of inconsistency, was 0.65. Since this value exceeded the preset warning threshold (e.g., 0.4), it was determined that the prompt triggering condition defined for level 1 in the verification table was met (i.e., significant tremor or motor incoordination was detected), thus a confirmation prompt box popped up in real time on the interactive interface to guide the user to correct the data. By constructing this closed-loop verification logic of "subjective description - expected range - objective characteristics," the quality of scale data can be controlled from the source. This method uses standardized processing and weighted scoring models to unify vague subjective feelings and complex physiological signals into comparable mathematical indicators, which can effectively quantify the adaptive bias caused by users' psychological factors. At the same time, by introducing a graded triggering mechanism, the system can correct unintentional misfills through gentle prompts and intercept serious data fraud through hard conflict mechanisms. Without increasing the user's additional testing burden, it can significantly improve the authenticity, objectivity, and logical consistency of ePRO scale data.

[0040] Step S4: When the question type is cognitive verification, determine whether the real-time feature value meets the preset cognitive anomaly judgment condition; if it does, provide an anomaly prompt; the cognitive anomaly judgment condition is used to identify whether the user has cognitive fatigue or operational confusion that is inconsistent with the answering behavior.

[0041] In a preferred embodiment, when the question type is cognitive verification, determining whether a preset cognitive anomaly judgment condition is met based on the real-time feature value includes: Based on all key travel time and key hold time during the user's response process, the average key travel time, average key hold time, key travel time variance, and key hold time variance are calculated. When the average keystroke flight time or the average key hold time exceeds a preset fatigue threshold, it is determined that the user is experiencing cognitive fatigue. When the keystroke flight time variance or the key hold time variance exceeds a preset stability threshold, it is determined that the user is operating in a disorderly manner. When it is determined that the user is experiencing cognitive fatigue or the operation is chaotic, the preset cognitive abnormality determination conditions are met. Specifically, when a user clicks to answer a cognitive verification question, the touchscreen sensor is used to collect the timestamp sequence of n touch events generated by the user's continuous clicks; let the pressing time of the i-th click be... The moment of lifting is The system calculates the keystroke flight time between the i-th and i+1-th clicks as follows: The duration of the button press for the i-th click is ; Based on the above keystroke flight time and key hold time, calculate the average keystroke flight time, average key hold time, keystroke flight time variance, and key hold time variance: The average keystroke flight time is: The average key press hold time is calculated using the same average formula; the formula for calculating the variance of key press flight time is: The variance of the key hold time is calculated using the same variance formula; the preset cognitive abnormality judgment conditions include: or When any one of the following exceeds a preset fatigue threshold, it is determined that the objectively reflected cognitive processing speed has slowed down significantly, indicating cognitive fatigue; when or If any one of the following exceeds the preset stability threshold, it is determined that the objectively reflected fine motion control capability fluctuates, indicating operational chaos. Specifically, taking a simulated patient who subjectively rated their memory as "normal" as an example, during their response, the system measured their average keystroke flight time to be 1200 milliseconds, while the preset fatigue threshold was 800 milliseconds. Simultaneously, the calculated variance of their keystroke flight time far exceeded the normal stability threshold. At this point, the system determined that although the user subjectively selected a normal level, their objective characteristic values ​​met the fatigue and confusion indicators in the cognitive abnormality judgment criteria. Therefore, the system determined that their objective cognitive dynamics state was inconsistent with their subjective description and provided an abnormality warning to the interactive interface.

[0042] By introducing this passive keystroke dynamics analysis method, cognitive functional states that are difficult to perceive directly can be transformed into mathematically measurable average and variance indices. Since the temporal characteristics of touch clicks are subconsciously driven physiological feedback, which is extremely difficult to fake, by monitoring the mean and fluctuations of flight time and hold time, it is possible to more sensitively capture the characteristics of the subject's actual cognitive processing speed decline or chaotic operation rhythm during the response process than subjective questionnaires. This judgment logic based on micro-behavioral data can effectively identify logical contradictions caused by cognitive decline, fatigue, or proxy completion, and significantly improve the data authenticity and objectivity of the ePRO scale in the cognitive dimension assessment without increasing the additional testing burden.

[0043] like Figure 2 As shown, another embodiment of the present invention also provides an ePRO scale real-time consistency verification device, including: a data acquisition module, a feature extraction module, a motion verification module and a cognitive verification module; The data acquisition module is used to obtain the question type of the current item in the ePRO scale, select the corresponding sensor according to the question type, and synchronously collect sensor data reflecting the user's objective physiological function status according to the selected sensor while the user answers the current question; wherein, the question type includes: cognitive verification, walking verification, and dexterity verification. The feature extraction module is used to extract features from the sensor data to obtain real-time feature values; The motion verification module is used to obtain the real-time subjective state description that reflects the user's motion control ability when answering the current question, when the question type is walking verification or dexterity verification; based on the verification table corresponding to the question type, determine the expected feature value range corresponding to the real-time subjective state description, calculate the degree of deviation of the real-time feature value from the expected feature value range, obtain the degree of inconsistency, and when the degree of inconsistency meets the preset triggering condition, provide feedback of the abnormality prompt. The cognitive verification module is used to determine whether the real-time feature value meets the preset cognitive anomaly judgment conditions when the question type is cognitive verification; if it does, an anomaly prompt is fed back; the cognitive anomaly judgment conditions are used to identify whether the user has cognitive fatigue or operational confusion that is inconsistent with the answering behavior.

[0044] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the real-time consistency verification method for the ePRO scale provided by any of the above-described method embodiments of the present invention.

[0045] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0046] Based on the above-described embodiment of the ePRO scale real-time consistency verification method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements an ePRO scale real-time consistency verification method according to any embodiment of the present invention.

[0047] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0048] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0049] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0050] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the ePRO scale real-time consistency verification method described in any of the above-described method embodiments of the present invention.

[0051] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0052] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for real-time consistency verification of an ePRO scale, characterized in that, include: The system obtains the question type of the current item in the ePRO scale, selects the corresponding sensor based on the question type, and synchronously collects sensor data reflecting the user's objective physiological function status based on the selected sensor while the user answers the current question. The question types include: cognitive verification, walking verification, and dexterity verification. Feature extraction is performed on the sensor data to obtain real-time feature values; When the question type is walking verification or dexterity verification, the real-time subjective state description selected by the user when answering the current question to reflect the motion control ability is obtained; based on the verification table corresponding to the question type, the expected feature value range corresponding to the real-time subjective state description is determined, the degree of deviation of the real-time feature value from the expected feature value range is calculated, and the degree of inconsistency is obtained; when the degree of inconsistency meets the preset triggering condition, an abnormal prompt is fed back. When the question type is cognitive verification, it is determined whether the real-time feature value meets the preset cognitive anomaly judgment condition; if it does, an anomaly prompt is fed back; the cognitive anomaly judgment condition is used to identify whether the user has cognitive fatigue or operational confusion that is inconsistent with the answering behavior.

2. The real-time consistency verification method for the ePRO scale as described in claim 1, characterized in that, The step of selecting a corresponding sensor based on the question type and synchronously collecting sensor data reflecting the user's objective physiological functional state based on the selected sensor while the user answers the current question includes: When the question type is a dexterity test, the accelerometer and gyroscope are used to collect acceleration and angular velocity signals reflecting hand movements, and the touchscreen sensor is used to collect touch trajectory. When the question type is a walking test, the barometric pressure sensor is used to collect height change data, the pedometer is used to collect step count data, and historical gait statistics cached within a preset time period are obtained. The historical gait statistics include swing time records for each of the two limbs. When the question type is a cognitive test, the touchscreen sensor is invoked to collect the touchscreen click timestamp data during the user's answering process.

3. The real-time consistency verification method for the ePRO scale as described in claim 2, characterized in that, The step of extracting features from the sensor data to obtain real-time feature values ​​includes: When the question type is agile verification, the acceleration signal and angular velocity signal are converted to the frequency domain and a normalized Fourier amplitude spectrum curve is calculated; the negative arc length of the normalized Fourier amplitude spectrum curve is taken as the spectral arc length; time-frequency analysis is performed on the acceleration signal to obtain the power spectral density, and the ratio of the energy of the preset target frequency band to the total energy of the entire frequency band is calculated to obtain the intention tremor index; and based on the touch trajectory, the distance between the touch point and the center of the target area is calculated to obtain the click deviation, and linear fitting analysis is performed on the touch trajectory to obtain the touch trajectory linearity; the spectral arc length, power spectral density, intention tremor index, click deviation, and touch trajectory linearity are taken as the real-time feature values; When the question type is walking verification, based on the historical gait statistical characteristics, the ratio of the standard deviation to the mean of the continuous gait cycle is calculated to obtain the stride length variability; based on the swing time records of each side of the limbs in the historical gait statistical characteristics, the absolute value of the difference between the swing time of the left limb and the swing time of the right limb is calculated, and the ratio of the absolute value of the difference to the mean of the swing time of both sides is calculated to obtain the asymmetry index; based on the height change data, the difference is calculated to obtain the height change amount, and based on the step count data, the number of steps within the preset monitoring time is counted to obtain the short-term step count; stride length variability, asymmetry index, height change amount, and short-term step count are used as the real-time feature values; When the question type is cognitive verification, based on the touch screen click timestamp data, the time interval between two adjacent keystrokes and the duration of a single keystroke from pressing to releasing are calculated to obtain the keystroke flight time and key hold time; the keystroke flight time and key hold time are used as the real-time feature values.

4. The real-time consistency verification method for the ePRO scale as described in claim 3, characterized in that, The verification table includes: health status level, expected subjective status description for each health status, expected feature value range, and trigger prompt conditions; wherein, the trigger prompt conditions are used to define the judgment logic for triggering an anomaly prompt when the real-time feature value deviates from the expected feature value range; The step of determining the expected feature value range corresponding to the real-time subjective state description based on the verification table corresponding to the question type includes: The real-time subjective state description is matched with each expected subjective state description in the verification table, and the expected subjective state description that matches the real-time subjective state description is taken as the target subjective state description. The expected feature value range corresponding to the subjective state description of the target is taken as the expected feature value range corresponding to the real-time subjective state description.

5. The real-time consistency verification method based on the ePRO scale as described in claim 4, characterized in that, The degree of deviation of the real-time feature value from the expected feature value range is calculated to obtain the degree of inconsistency. When the degree of inconsistency meets a preset trigger condition, an anomaly prompt is fed back, including: The real-time feature values ​​are converted into normalized objective functional scores using the Z-Score standardization method. The health status level corresponding to the real-time subjective state description is normalized into a numerical value to obtain the subjective function score. Calculate the absolute value of the difference between the subjective function score and the objective function score, and use the absolute value as the degree of inconsistency. When the degree of inconsistency is between the first warning threshold and the second warning threshold of the preset triggering conditions, it is determined that the soft conflict type prompt triggering condition in the preset triggering conditions is met, and a confirmation prompt box pops up on the interactive interface; wherein, the first warning threshold is less than the second warning threshold; When the degree of inconsistency is greater than the second warning threshold, it is determined that the hard conflict type prompt trigger condition in the preset trigger conditions is met, and the current submission operation is blocked or the current data is marked as untrustworthy in the interactive interface.

6. The real-time consistency verification method based on the ePRO scale as described in claim 5, characterized in that, When the question type is cognitive verification, the system determines whether the preset cognitive anomaly judgment conditions are met based on the real-time feature values, including: Based on all key travel time and key hold time during the user's response process, the average key travel time, average key hold time, key travel time variance, and key hold time variance are calculated. When the average keystroke flight time or the average key hold time exceeds a preset fatigue threshold, it is determined that the user is experiencing cognitive fatigue. When the keystroke flight time variance or the key hold time variance exceeds a preset stability threshold, it is determined that the user is operating in a disorderly manner. When it is determined that the user is experiencing cognitive fatigue or the operation is chaotic, the preset cognitive abnormality determination conditions are met.

7. The real-time consistency verification method for the ePRO scale as described in claim 1, characterized in that, Also includes: Compare the real-time feature values ​​with the current historical mean corresponding to the individual feature baseline; If the absolute difference between the real-time feature value and the current historical mean is greater than the product of the current historical standard deviation corresponding to the personal feature baseline and the preset statistical coefficient, it is determined that the user has experienced a sudden change in status, and an abnormality prompt is given. Otherwise, the personal characteristic baseline is updated based on the real-time feature values; In the initial stage, a baseline of individual characteristics is established based on the characteristic value data of the nervous system disease group; Each time a user completes a valid ePRO scale response, the feature value data of the valid response is collected, and the number of valid responses is accumulated. When the number of valid responses reaches a preset number, the personal feature baseline is updated based on the feature value data of all valid responses. Based on the updated personal feature baseline, the updated historical mean and the updated historical standard deviation are calculated.

8. A real-time consistency verification device for an ePRO scale, characterized in that, include: Data acquisition module, feature extraction module, motion verification module, and cognitive verification module; The data acquisition module is used to obtain the question type of the current item in the ePRO scale, select the corresponding sensor according to the question type, and synchronously collect sensor data reflecting the user's objective physiological function status according to the selected sensor while the user answers the current question; wherein, the question type includes: cognitive verification, walking verification, and dexterity verification. The feature extraction module is used to extract features from the sensor data to obtain real-time feature values; The motion verification module is used to obtain the real-time subjective state description that reflects the user's motion control ability when answering the current question, when the question type is walking verification or dexterity verification; based on the verification table corresponding to the question type, determine the expected feature value range corresponding to the real-time subjective state description, calculate the degree of deviation of the real-time feature value from the expected feature value range, obtain the degree of inconsistency, and when the degree of inconsistency meets the preset triggering condition, provide feedback of the abnormality prompt. The cognitive verification module is used to determine whether the real-time feature value meets the preset cognitive anomaly judgment conditions when the question type is cognitive verification; if it does, an anomaly prompt is fed back; the cognitive anomaly judgment conditions are used to identify whether the user has cognitive fatigue or operational confusion that is inconsistent with the answering behavior.

9. A terminal device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the method as described in any one of claims 1-7.

Images