Physiological data-based elderly health suggestion generation method and system

Abstract

The invention provides an elderly health suggestion generation method and system based on physiological data, and relates to the technical field of smart pension and health monitoring, and the method comprises the steps: obtaining multi-dimensional physiological data of a to-be-detected elderly; preprocessing the multi-dimensional physiological data to obtain preprocessed data; carrying out personalized physiological baseline fitting on the preprocessed data to obtain a personalized physiological baseline; acquiring real-time multi-dimensional physiological data; according to the personalized physiological baseline and the real-time multi-dimensional physiological data, a linkage early warning result is obtained through multi-index linkage abnormal early warning; health risk grade division is carried out on the linkage early warning result, and the health risk grade of the to-be-tested old person is determined; and according to the health risk level, combining with the chronic disease condition of the to-be-tested old person to generate a health suggestion. According to the invention, precise grading and dynamic suggestion generation of the health risk of the elderly are realized.

Description

Technical Field

[0001] This invention relates to the field of smart elderly care and health monitoring technology, and in particular to a method and system for generating health recommendations for the elderly based on physiological data. Background Technology

[0002] With the increasing aging of the global population, health monitoring and management of the elderly has become an important social issue. The decline in physical function among the elderly is often accompanied by multiple chronic diseases, and their health status is the result of the combined effects of various physiological indicators. An abnormality in any single indicator may be an early signal of systemic risk. Therefore, there is an urgent need to monitor and manage the health of the elderly, thereby achieving a shift from "passive treatment" to "proactive prevention."

[0003] Currently, existing smart elderly care health monitoring technologies mainly fall into two categories. One category consists of single-indicator early warning systems based on fixed thresholds, such as triggering an alarm when heart rate or blood pressure exceeds a preset general safety range. The other category comprises more advanced systems capable of performing simple statistical analysis on historical data, generating trend charts, and allowing family members or doctors to view the data remotely.

[0004] However, existing technologies employ a uniform, fixed threshold warning model, which easily overlooks the significant individual differences among elderly individuals. This may lead to frequent false alarms for those with special physical conditions or underreporting of slowly deteriorating risks. Secondly, existing methods typically view various physiological indicators in isolation, lacking analysis of the correlations between indicators, thus failing to provide accurate basis for health risk intervention. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method for generating health recommendations for the elderly based on physiological data. This method addresses the shortcomings of existing technologies that use a uniform, fixed threshold warning model, which easily overlooks the significant individual differences among elderly individuals, potentially leading to frequent false alarms for those with special physical conditions or underreporting of slowly deteriorating risks. Furthermore, existing methods typically view various physiological indicators in isolation, lacking analysis of the interrelationships between indicators, thus failing to provide accurate evidence for health risk intervention.

[0006] A first aspect of this invention provides a method for generating health recommendations for the elderly based on physiological data, comprising: S1: Obtain multidimensional physiological data of the elderly person to be tested; S2: Preprocess the multidimensional physiological data to obtain preprocessed data; S3: Perform personalized physiological baseline fitting on the preprocessed data to obtain a personalized physiological baseline; S4: Acquire real-time multidimensional physiological data; S5: Based on personalized physiological baselines and real-time multi-dimensional physiological data, a linked early warning result is obtained through multi-indicator linkage abnormality warning; S6: Classify the health risk level of the linkage early warning results and determine the health risk level of the elderly to be tested; S7: Generate health recommendations based on the health risk level and the chronic disease status of the elderly person being tested.

[0007] A second aspect of this invention provides a health recommendation generation system for the elderly based on physiological data, comprising: a processor and a memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the method for generating health recommendations for the elderly based on physiological data as described in the first aspect.

[0008] A third aspect of the present invention provides a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the method for generating health recommendations for the elderly based on physiological data as described in the first aspect.

[0009] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, the baseline fitted based on personal historical data can dynamically reflect the unique physiological characteristics of an individual, significantly reducing false alarms and missed alarms caused by special physical conditions; at the same time, by comprehensively analyzing the linkage changes of multi-dimensional physiological indicators, potential systemic health risks can be captured more accurately, thus providing a more accurate basis for achieving early and precise health risk intervention. Attached Figure Description

[0010] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0011] Figure 1 This is a flowchart illustrating a method for generating health recommendations for the elderly based on physiological data, provided in an embodiment of the present invention.

[0012] Figure 2 This is a schematic diagram of a method for generating health recommendations for the elderly based on physiological data, provided in an embodiment of the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0014] The method for generating health recommendations for the elderly based on physiological data, provided by the present invention, will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0015] Reference manual attached Figure 1 The diagram illustrates a flowchart of a method for generating health recommendations for the elderly based on physiological data, provided by an embodiment of the present invention.

[0016] This invention provides a method for generating health recommendations for the elderly based on physiological data, which may include the following steps: S1: Obtain multi-dimensional physiological data of the elderly to be tested.

[0017] The multidimensional physiological data refers to three categories of data indicators: basic vital signs indicators, core chronic disease monitoring indicators, and lifestyle-related indicators. Specific indicators of basic vital signs include heart rate (resting / active) and blood oxygen saturation, monitored in real-time (using a fitness tracker), with the core function of reflecting cardiopulmonary function and providing early warning of acute risks (such as a sudden drop in heart rate). Core body temperature is monitored once per hour (using a fitness tracker), with the core function of identifying infection (such as fever) and hypothermia risk. Resting respiratory rate is monitored once every 5 minutes (using millimeter-wave radar), with the core function of assisting in the assessment of respiratory dysfunction (such as acute exacerbations of COPD). Specific indicators of core chronic disease monitoring include blood pressure (systolic blood pressure SBP / diastolic blood pressure DBP), monitored once each in the morning and before bedtime (using a wrist blood pressure monitor), with the core function of hypertension management and providing early warning of stroke and myocardial infarction risks. Blood glucose (fasting / 2 hours postprandial / bedtime), monitoring frequency / scenario is fasting + real-time postprandial (continuous glucose monitoring), its core function is the core of diabetes management, warning of hypoglycemia and ketosis risk. Glycated hemoglobin (HbA1c), monitoring frequency / scenario is once / 3 months (external testing module), its core function is to reflect long-term blood glucose control and assist in trend and risk assessment. Specific indicators related to lifestyle status include sleep structure (deep sleep / light sleep / REM sleep / number of awakenings), monitoring frequency / scenario is continuous nighttime monitoring (wristband + radar), its core function is to correlate with immunity and blood pressure fluctuations, and to warn of sleep apnea. Daytime activity level (steps / activity duration), monitoring frequency / scenario is real-time (wristband), its core function is to help distinguish whether "abnormal physiological indicators" are caused by activity.

[0018] S2: Preprocess the multidimensional physiological data to obtain preprocessed data.

[0019] Preprocessing refers to the technical operations of initially processing and organizing raw physiological data to form a complete, standardized, and reliable preprocessed data sequence.

[0020] In this embodiment of the invention, data cleaning, format standardization, and outlier correction effectively improve the integrity and reliability of the original physiological data, laying a high-quality data foundation for subsequent accurate analysis.

[0021] In one possible implementation, S2 specifically includes sub-steps S201 to S206: S201: Standardize the format of multidimensional physiological data.

[0022] Specifically, data from various devices is received through the edge computing gateway and uniformly encapsulated as "timestamp (accurate to the second) + device ID + indicator type + original value + data quality label (such as "high / medium / low", which is determined by the device itself to determine the signal strength), for example: 2024-05-20 08:30:15|wristband 001|heart rate|82 beats / min|high quality.

[0023] S202: Perform timeline synchronization processing on the data after it has been formatted into a unified format.

[0024] Among them, time axis synchronization processing refers to the process of unifying physiological data from multiple monitoring devices with different sampling frequencies and initial timestamps into the same standard time coordinate system through techniques such as interpolation, alignment and resampling.

[0025] Specifically, based on the local clock of the edge gateway, timestamp calibration is performed on data from multiple devices at the same time (within ±3 seconds) to avoid “misinterpretation of cross-time period data” caused by device clock deviation (such as judging blood pressure at 8:30 and heart rate at 8:35 as data from the same time period).

[0026] S203: Perform data cleaning on the timeline-synchronized data to obtain cleaned data.

[0027] Data cleaning refers to the process of detecting, identifying, and processing raw datasets to eliminate errors, incompleteness, inconsistencies, or irrelevant parts, thereby improving data quality.

[0028] In one possible implementation, S203 specifically includes sub-steps S2031 and S2032: S2031: Perform missing value repair on the timeline-synchronized data. If the missing data duration is longer than the preset duration, it will be repaired using a long-term missing value repair method. If the missing data duration is less than or equal to the preset duration, it will be repaired using a short-term missing value repair method.

[0029] The specific method for repairing long-term missing data is as follows: when the duration of data loss is a critical indicator, a device reconnection reminder is triggered. When the duration of data loss is not a critical indicator, the duration of data loss is marked as an invalid field.

[0030] The short-term missing data repair method specifically uses a linear interpolation algorithm:

[0031] in, express t Duration of missing data at any given moment express t The data value at time -1 is the data value from the previous valid time. express tThe data value at time +1 is the data value at the next valid time. T Indicates the duration of the missing information. express t Time and t -1 is the time interval.

[0032] Linear interpolation is a mathematical method used to estimate the value of an intermediate point between two known data points. Its core idea is to assume that the change between these two known points is linear, meaning it changes uniformly along a straight line.

[0033] It should be noted that those skilled in the art can set the preset duration according to actual needs, and this invention does not limit this.

[0034] S2032: After the missing values ​​are repaired, outlier filtering is performed to obtain cleaned data.

[0035] Outlier filtering refers to the data processing procedure that identifies and removes obviously erroneous or impossible extreme values ​​from the original data within a pre-defined reasonable range that conforms to physiological common sense or statistical laws.

[0036] Specifically, adopting "3" σ A dual mechanism of "criteria + physiological rationality verification": first, the mean of the historical data of the indicator is calculated. μ and standard deviation σ Preliminary Filtering | x - μ |>3 σ The numerical values, combined with the physiological limits of the elderly (such as heart rate ≤30 beats / min or ≥180 beats / min, systolic blood pressure >220), are considered. mmHg All of these are physiologically unreasonable values. Finally, outliers were identified and replaced with "the previous valid value + smoothing coefficient" (the smoothing coefficient is 0.8 to avoid data mutation).

[0037] In this embodiment of the invention, by differentiating between long-term and short-term missing data and using a dual outlier filtering mechanism, the accuracy and physiological rationality of data quality are effectively improved while ensuring data continuity.

[0038] S204: By using sensor-differentiated noise reduction, the cleaning data is processed to obtain noise-reduced data.

[0039] Among them, noise reduction processing refers to the process of using digital signal processing technology to identify and separate target physiological signals and noise components from the collected raw physiological data, thereby improving signal quality and highlighting effective information.

[0040] In one possible implementation, S204 specifically includes sub-steps S2041 to S2043: S2041: For heart rate or blood oxygen data from the wristband, a Kalman filter algorithm is used for noise reduction. State prediction:

[0041] Covariance prediction:

[0042] Kalman gain:

[0043] Status Update:

[0044] in, express k Value after noise reduction at any time express k The value after noise reduction at time -1 express k Original measurement value at time, A Represents the state transition matrix. H Represents the observation matrix. Q Indicates process noise. R Indicates measurement noise. express k Predicted state value at time of day B Represents the control input matrix. express k Predict the state covariance matrix at each time step. express k The optimal estimated covariance matrix at time -1 express k Time-based Kalman gain; Among them, the Kalman filter algorithm is a highly efficient optimal recursive data processing algorithm. It does not simply smooth the data, but rather performs a "prediction-correction" of the current state based on the dynamic model of the system.

[0045] Furthermore, the state transition matrix is ​​set to 1.0, the observation matrix to 1.0, the process noise to 0.05, and the measurement noise to 0.1. The control input matrix is ​​used to quantify the impact of control input on physiological data; without control input, B It can be set to 0. The predicted state covariance matrix describes the uncertainty of the predicted state value; the larger the value, the less reliable the prediction. The larger the Kalman gain, the more confident the measured value is, and vice versa.

[0046] S2042: For continuous glucose monitoring data, a moving average filtering algorithm is used for noise reduction.

[0047] in, express kBlood sugar levels after noise reduction express v Original value at time, K Indicates the total duration.

[0048] The moving average filtering algorithm is a very simple and intuitive method for smoothing and denoising. Its principle is to define a fixed-length "time window" and then use the arithmetic mean of all data points within this window to represent the data value at the center of the window. Then, the window is slid forward one time point, and the average is calculated repeatedly, and so on.

[0049] S2043: For blood pressure monitor data, a median filtering algorithm is used for noise reduction.

[0050] The median filtering algorithm is a classic nonlinear signal processing technique used to effectively remove impulse noise and outliers from data. Its core principle is quite intuitive: instead of using the average, it moves a sliding window across the data sequence, sorting all data points within the window by value, and finally taking the median, which is located in the exact middle of the sorted values, as the output value at the center of the window.

[0051] Specifically, the median filtering algorithm uses a window size of 3 measurements to exclude single abnormal fluctuations (such as high blood pressure caused by an overly loose cuff).

[0052] In this embodiment of the invention, by adaptively matching the optimal noise reduction algorithm (Kalman filtering, moving average, median filtering) based on the data characteristics of different physiological signals (such as the dynamics of heart rate, the stability of blood glucose, and the susceptibility of blood pressure to interference), various noise interferences are accurately filtered out while effectively preserving the true physiological change trend.

[0053] S205: Perform data standardization processing on the noise reduction data to obtain standardized data.

[0054] Standardization is a process of transforming raw data of different scales, dimensions, and ranges into a specific standard scale through mathematical transformation.

[0055] Specifically, standardized and unified units / scopes for indicators (such as blood pressure). mmHg ",blood sugar" mmol / L The formula is as follows: (Heart rate in "beats / minute"), and the values ​​are mapped to the "0-1" interval (for easier subsequent linkage analysis).

[0056] in, x Physiological metrics representing the noise reduction dataset and These represent the minimum and maximum values ​​of the physiologically reasonable range of physiological indicators, respectively.

[0057] S206: Perform cross-device consistency verification on standardized data to obtain preprocessed data.

[0058] Consistency verification refers to the process of checking for logical contradictions or conflicts within data records, between different records, or between data and predefined business rules during the data preprocessing stage.

[0059] Specifically, cross-device consistency checks compare data of the same metric from different devices (such as heart rate on a wristband versus heart rate on a radar device). If the deviation is greater than 10%, a "device fault alert" is triggered. If the deviation is less than or equal to 10%, the weighted average (0.7 for high-quality data and 0.3 for low-quality data) is taken as the final value.

[0060] Furthermore, the preprocessed data includes "timestamp + indicator type + standardized value + data confidence (0-1, determined by the cleaning / denoising effect)" for baseline fitting.

[0061] Furthermore, the preprocessed data is stored in a cloud-based health database, which stores the elderly’s historical physiological data (such as blood pressure and blood sugar records for one year), personalized physiological baseline parameters, and health records (such as whether they have hypertension or diabetes). Encrypted storage is used to protect data privacy.

[0062] In this embodiment of the invention, through multi-level processing including format unification, time axis synchronization, data cleaning, sensor-specific noise reduction, standardization, and consistency verification, high-quality, highly comparable, and logically consistent standardized physiological data are constructed, laying a reliable data foundation for subsequent accurate fitting of personalized baselines and multi-indicator linkage analysis.

[0063] S3: Perform personalized physiological baseline fitting on the preprocessed data to obtain a personalized physiological baseline.

[0064] Personalized physiological baseline fitting refers to the process of calculating a series of dynamic reference ranges that can represent the "normal" health status of a specific individual, based on long-term historical physiological data and through statistical modeling or machine learning methods.

[0065] For example, the fitting is broken down by "indicator sub-type + monitoring scenario". For instance: Blood pressure: Fit four baselines: "morning resting systolic blood pressure baseline", "bedtime resting systolic blood pressure baseline", "morning resting diastolic blood pressure baseline", and "bedtime resting diastolic blood pressure baseline" (because elderly people's blood pressure has a diurnal rhythm, morning blood pressure is usually higher than bedtime blood pressure). Blood glucose: Fit three baselines: "fasting blood glucose baseline (6:00-7:00)", "2-hour postprandial blood glucose baseline (12:00-13:00 / 18:00-19:00)", and "bedtime blood glucose baseline (21:00-22:00)". Heart rate: Fit two baselines: "resting heart rate baseline (morning)" and "activity heart rate baseline (daytime)".

[0066] It should be noted that the individual feature weighted fusion algorithm breaks through the traditional static baseline of "fixed mean ± standard deviation" and combines the individual health characteristics of the elderly (age, chronic diseases, medication) and long-term data trends to construct a dynamically adjustable personalized baseline.

[0067] In this embodiment of the invention, by subdividing indicator types and monitoring scenarios and combining them with individual characteristics through weighted fusion, a dynamically adjustable personalized physiological baseline is constructed for each elderly person. This breaks through the limitations of traditional static thresholds and lays the core basis for subsequent accurate identification of health risks that deviate from their normal state.

[0068] In one possible implementation, S3 specifically includes sub-steps S301 to S305: S301: Perform data annotation on the preprocessed data to obtain initial baseline data with feature labels.

[0069] Data annotation refers to the process of adding meaningful labels or annotations to raw data (such as text, images, audio, video, or sensor data) so that it becomes information that machine learning models can understand and learn.

[0070] Specifically, the data collection period is four consecutive weeks (covering one complete physiological cycle to avoid the impact of short-term fluctuations). For the daily collected "detailed indicator-scenario" data, corresponding individual status labels are attached, such as: chronic disease labels: hypertension (1=yes, 0=no) and diabetes (1=yes, 0=no); medication labels: antihypertensive drugs (1=regularly taken, 0=not taken) and hypoglycemic drugs (1=regularly taken, 0=not taken); lifestyle habit labels: daily exercise (1=≥30 minutes, 0=<30 minutes) and alcohol consumption (1=drank alcohol that day, 0=drank alcohol).

[0071] S302: Calculate the individual feature weights of the initial baseline data to form an individual feature weight vector.

[0072] Among them, the individual feature weight is a quantified value used to represent the degree of influence of different personal attributes (such as age, medical history, and lifestyle habits) on the final calculation result.

[0073] Specifically, the Analytic Hierarchy Process (AHP) was used to determine the weight of each feature on the baseline. First, a judgment matrix was constructed, and geriatricians were invited to rate the importance of five features: age, hypertension, diabetes, medication, and exercise (using a 1-9 scale). For example, "Hypertension is more important to the baseline blood pressure than age," and a score of 3 was assigned. The consistency of the judgment matrix was then tested. CR After setting the threshold to <0.1), the weights of each feature are calculated. Example weights are as follows: Age (yes=1) has a weight of 0.15 on the baseline blood glucose and 0.10 on the baseline blood pressure. Hypertension (yes=1) has a weight of 0.35 on the baseline blood glucose and 0.05 on the baseline blood pressure. Diabetes (yes=1) has a weight of 0.05 on the baseline blood glucose and 0.40 on the baseline blood pressure. Regular medication use (yes=1) has a weight of 0.30 on the baseline blood glucose and 0.35 on the baseline blood pressure. Daily exercise (yes=1) has a weight of 0.15 on the baseline blood glucose and 0.10 on the baseline blood pressure.

[0074] S303: Based on the initial baseline data and individual feature weight vectors, calculate the baseline core parameters, which include the weighted mean and weighted standard deviation. The weighted mean is as follows:

[0075] in, express t Preprocessed data at any given time, express t Time of the first i Label values ​​of individual characteristics, Indicates the first i The weights of individual characteristics, n Indicates the total number of valid data. This represents the weighted average.

[0076] The weighted standard deviation is as follows:

[0077] in, This represents the weighted standard deviation.

[0078] The weighted average refers to assigning a "weight" representing the importance or reliability of each data point when calculating the average, resulting in an average that takes into account the importance of the data points.

[0079] The weighted standard deviation, based on the weighted mean, measures the dispersion (i.e., the magnitude of fluctuation) of a set of weighted data. It also considers the weight of each data point, thus more accurately reflecting the fluctuation after taking into account the importance of the data.

[0080] It should be noted that for each "sub-indicator-scenario" data point, the weighted mean and weighted standard deviation must be calculated as the baseline core parameters.

[0081] S304: Perform dynamic baseline calibration on the core baseline parameters.

[0082] in, Represents the smoothing coefficient. This represents the weighted average of the new data for the current month. This represents the weighted mean of the original baseline. This represents the weighted standard deviation of the new data for the current month. This represents the weighted standard deviation of the original baseline. This represents the weighted mean after dynamic calibration. This represents the weighted standard deviation after dynamic calibration.

[0083] Baseline dynamic calibration refers to the process by which the system continuously adjusts and updates its personalized physiological baseline based on the latest and most reliable physiological data, rather than doing so once.

[0084] Specifically, each month, the baseline is calibrated based on newly collected physiological data (200+ valid data points), i.e., real-time multidimensional physiological data. Exponential smoothing is used to balance the influence of historical and new data. The smoothing coefficient is 0.3, with new data having a 30% weight and historical data having a 70% weight.

[0085] S305: Determine the range of the personalized physiological baseline based on the dynamically calibrated core baseline parameters.

[0086] in, b This represents the risk coefficient.

[0087] Specifically, the range of the "safe fluctuation range," or personalized physiological baseline, is adjusted based on the elderly person's health risk level.

[0088] Furthermore, healthy elderly people b =1.2, elderly people with chronic diseases b =1.0, high-risk elderly b =0.8, the higher the risk, the narrower the range, and the more sensitive the early warning.

[0089] In this embodiment of the invention, a dynamic baseline fitting and calibration mechanism that integrates individual feature weights is used to construct a physiological benchmark interval that can accurately reflect the individual health status of the elderly and is adaptively updated, providing a highly personalized and sensitive judgment basis for abnormal early warning.

[0090] S4: Acquire real-time multidimensional physiological data.

[0091] S5: Based on personalized physiological baselines and real-time multi-dimensional physiological data, a linked early warning result is obtained through multi-indicator linkage abnormality warning.

[0092] It should be noted that traditional multi-indicator early warning only "simplely superimposes the number of abnormal indicators". This solution innovatively constructs an "indicator correlation matrix + linkage risk value" model, which combines physiological mechanisms (such as the fact that elevated blood sugar is easily accompanied by increased heart rate and elevated blood pressure is easily led to decreased blood oxygen) to achieve accurate early warning.

[0093] In this embodiment of the invention, by constructing an indicator correlation matrix and a linked risk value model, the limitation of simply superimposing the number of anomalies in traditional early warning is overcome, and multi-indicator collaborative risk perception based on physiological mechanisms is realized, which significantly improves the accuracy of anomaly early warning.

[0094] In one possible implementation, S5 specifically includes sub-steps S501 to S506: S501: Based on real-time multidimensional physiological data, the correlation strength between various indicators is calculated using the Pearson correlation coefficient; The specific formula for calculating the Pearson correlation coefficient is as follows:

[0095] in, express t Time indicators i The value, Indicators i Personalized baseline mean, Indicate t Time indicators j The value, Indicators j Personalized baseline mean, Indicators i and indicators j The correlation weight, with a value range of [-1, 1].

[0096] The Pearson correlation coefficient is a statistical indicator used to measure the strength and direction of the linear relationship between two variables.

[0097] It should be noted that before calculating the linkage strength, core linkage indicator pairs need to be screened. Based on geriatric medicine theory, indicator pairs with physiological correlations should be identified, such as: blood glucose (Glu) - heart rate (HR): blood glucose > 11. mmol / L At this time, heart rate tends to increase (metabolism accelerates). Blood pressure (SBP) - blood oxygen (SpO2): SBP > 160 mmHg At this time, SpO2 tends to decrease (vasoconstriction leads to insufficient oxygen supply). Body temperature (T) - heart rate (HR): When T > 38.5℃, HR tends to increase (HR increases by 10 beats / min for every 1℃ increase in body temperature).

[0098] Furthermore, a correlation matrix is ​​generated, as shown below (only core indicators are listed): Correlation between blood glucose and blood glucose is 1.0, with blood pressure is 0.3, with heart rate is 0.6, with blood oxygen is 0.2, and with body temperature is 0.4. Correlation between blood pressure and blood glucose is 0.3, with blood pressure is 1.0, with heart rate is 0.5, with blood oxygen is 0.7, and with body temperature is 0.3. Correlation between heart rate and blood glucose is 0.6, with blood pressure is 0.5, with heart rate is 1.0, with blood oxygen is 0.4, and with body temperature is 0.8. Correlation between blood oxygen and blood glucose is 0.2, with blood pressure is 0.7, with heart rate is 0.4, with blood oxygen is 1.0, and with body temperature is 0.3. The correlation coefficient between body temperature and blood glucose was 0.4, with blood pressure 0.3, with heart rate 0.8, with blood oxygen 0.3, and with body temperature 1.0.

[0099] S502: Based on real-time multi-dimensional physiological data, calculate the deviation of each indicator from the personalized physiological baseline for each individual indicator.

[0100] in, Indicates the first i The deviation of each indicator from the single indicator. Indicates the first i The weighted baseline mean of each indicator, Indicates the first i The weighted baseline standard deviation of each indicator. Indicates the first i The value of each indicator.

[0101] Among them, the deviation of a single indicator refers to a quantitative value used to measure the degree to which the real-time measurement value of a certain physiological indicator deviates from its personalized baseline normal range.

[0102] Furthermore, the greater the deviation of a single indicator, the higher the risk. A deviation greater than 0.5 for a single indicator is considered a significant deviation, while a deviation less than or equal to 0.5 is considered a minor deviation.

[0103] S503: Calculate the linkage risk value based on the linkage strength and the deviation of a single indicator.

[0104] Optionally, the formula for calculating the linked risk value is as follows:

[0105] in, R Indicates the linked risk value. m This indicates the total number of indicators involved in the linkage. d i and d j They represent the first i The deviation of the single indicator and the first indicator j The deviation of each indicator.

[0106] In this embodiment of the invention, by multiplying the deviation of each pair of indicators by their physiological correlation strength and then summing them, an upgrade from isolated anomaly judgment to comprehensive risk quantification is achieved, which can more accurately capture the systemic health risks implied by the synergistic anomalies of multiple indicators.

[0107] S504: Determine linkage anomalies based on linkage risk values.

[0108] Among them, the linkage anomaly determination refers to the system not viewing the anomaly of each physiological indicator in isolation, but comprehensively considering the abnormal combination patterns and physiological correlation strength among multiple indicators to ultimately determine whether to trigger an early warning decision process.

[0109] Specifically, thresholds are set based on the type of chronic disease the elderly person has. R th Healthy elderly people R th =0.3. Elderly individuals with a single chronic disease (e.g., hypertension only). R th =0.25. Elderly individuals with multiple chronic diseases (such as hypertension and diabetes). R th =0.2. If R > R th Furthermore, the abnormal indicators appeared simultaneously within a 10-minute time window (excluding misjudgment across time periods), and were initially determined to be "linked anomalies".

[0110] S505: Filter false alarms from abnormal judgment results to obtain health abnormal event alerts.

[0111] Among them, false alarm filtering refers to the process by which the system uses preset rules, context information or multi-indicator cross-validation mechanisms to identify and filter out invalid alarms triggered by unhealthy reasons (such as equipment errors, temporary activities, environmental interference).

[0112] In one possible implementation, S505 specifically involves: based on the anomaly determination result, accessing environmental data and device status, excluding unhealthy anomalies, and obtaining a health anomaly event alarm.

[0113] Specifically, it involves accessing environmental data (such as whether the person is exercising or eating) and device status (such as whether it is being worn correctly) to exclude "unhealthy abnormalities." For example, if "heart rate..." d i =0.6 + blood pressure d j =0.5", but simultaneously detected "activity level > 500 steps / 10 minutes", judged as "normal fluctuations caused by exercise", and filtered out the warning. If "blood sugar d i =0.7 + heart rate d j =0.4”, but simultaneously detected “within 30 minutes after the meal”, which was judged as “normal post-meal reaction” and filtered out as a warning.

[0114] In this embodiment of the invention, by introducing cross-validation of environmental context and device status, non-health risk anomalies caused by normal physiological activities or device problems can be effectively identified and filtered, significantly improving the accuracy and practicality of health warnings.

[0115] S506: Output the linkage warning result based on the alarm for abnormal health events.

[0116] Specifically, a "linked early warning report" is generated, which includes: abnormal indicator pairs (such as "abnormal correlation between blood glucose and heart rate"). Linked risk values ​​(such as...). R =0.32). Possible causes of the abnormality (e.g., "high blood sugar may lead to an increased heart rate").

[0117] In this embodiment of the invention, by quantifying the physiological correlation strength between indicators, calculating the standardized deviation from the baseline, and integrating the judgment of linked risk values ​​with environmental context filtering, a leap from isolated abnormality judgment to accurate perception of systemic health risks is achieved, significantly improving the specificity of early warning and clinical guidance value.

[0118] S6: Classify the health risk level of the linked early warning results and determine the health risk level of the elderly to be tested.

[0119] Specifically, different warning methods are triggered based on the risk level: Level I risk: An alert is sent simultaneously to family members, community doctors, and the 120 emergency medical services platform (including the elderly person's location and real-time physiological data). Level II risk: An alert is sent to family members and community doctors, along with intervention suggestions (such as "suggest taking hypoglycemic drugs"). Level III risk: A reminder is sent to family members, suggesting an appointment for a doctor's follow-up visit.

[0120] In this embodiment of the invention, by precisely linking the linkage early warning results with the graded response mechanism, a closed-loop management system from risk identification to graded intervention is achieved, ensuring that different levels of health risks can trigger matching early warning responses, significantly improving rescue efficiency and the level of precision in health management.

[0121] S7: Generate health recommendations based on the health risk level and the chronic disease status of the elderly person being tested.

[0122] For example, when blood sugar is abnormal, a message may be sent saying "Reduce refined carbohydrate intake and increase the proportion of vegetables"; when blood pressure is abnormal, a message may be sent saying "Avoid emotional excitement and monitor blood pressure twice a day".

[0123] Further, acute risk identification for elderly patients with hypertension. A smart bracelet recorded a real-time heart rate of 88 beats / min (consistent with the patient's baseline of 65-90 beats / min) → a wrist blood pressure monitor recorded a blood pressure of 180 / 110. mmHg (Exceeding baseline by 130-150 / 80-90) mmHg → Edge node preprocesses data and uploads it to the cloud → Real-time anomaly identification module marks "severely abnormal blood pressure single indicator" → Health risk grading module determines it as Level II risk → Response layer sends "Elderly person's blood pressure 180 / 110" message to family members mmHg "It is recommended to take antihypertensive medication immediately and rest." An alert is sent to the community doctor. After 30 minutes, the system reminds the elderly person to measure their blood pressure again. If it is still >160 / 100... mmHg The risk level has been upgraded to Level I, and 120 has been called.

[0124] Trend risk identification in elderly people with diabetes. Continuous glucose monitoring (CGM) collected 2-hour postprandial blood glucose levels for three consecutive weeks: Week 1 average 8.5. mmol / L The average for the second week was 9.2. mmol / L The average for week 3 was 10.1. mmol / L (The elderly person's baseline is 5.8-8.2) mmol / L → The trend analysis module, through linear regression, found that "the average weekly blood glucose level increased by 0.7-0.9." mmol / L → The health risk classification module determines it to be a Level III risk → The response layer sends a message to the family saying, "The elderly person's blood sugar has been rising for three consecutive weeks. It is recommended to make an appointment with an endocrinologist to adjust the medication." At the same time, it pushes dietary advice such as "Control the amount of staple food and do not exceed one small bowl of rice per meal."

[0125] In this embodiment of the invention, by deeply integrating health risk levels with individual chronic disease characteristics, a leap from general reminders to personalized, actionable health advice is achieved, effectively improving the accuracy of intervention measures and user compliance.

[0126] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, the baseline fitted based on individual historical data can dynamically reflect the unique physiological characteristics of an individual, significantly reducing false positives and false negatives caused by special physical conditions. Simultaneously, by comprehensively analyzing the interconnected changes of multi-dimensional physiological indicators, potential systemic health risks can be captured more accurately, thus providing a more accurate basis for early and precise health risk intervention.

[0127] Reference manual attached Figure 2 The diagram shows a structural schematic of a health advice generation system for the elderly based on physiological data, provided by an embodiment of the present invention.

[0128] This invention provides a health advice generation system 20 for the elderly based on physiological data, comprising: a processor 201 and a memory 202; The memory 202 stores programs or instructions that can run on the processor 201. When the program or instructions are executed by the processor 201, they implement the steps of the above-described method for generating health recommendations for the elderly based on physiological data and achieve the same technical effect. To avoid repetition, the present invention will not elaborate further.

[0129] It should be understood that the processor 201 in this embodiment of the invention may be a central processing unit (CPU), or it may be 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. The general-purpose processor may be a microprocessor or any conventional processor.

[0130] It should also be understood that the memory 202 in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DR RAM).

[0131] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0132] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0133] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0134] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0135] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0136] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0137] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0138] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0139] This invention provides a readable storage medium comprising: storing a program or instructions on the readable storage medium, wherein when the program or instructions are executed by a processor, the program or instructions implement the steps of the above-described method for generating health recommendations for the elderly based on physiological data, and can achieve the same technical effect. To avoid repetition, this invention will not elaborate further.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A method for generating health recommendations for the elderly based on physiological data, characterized in that, include: S1: Obtain multidimensional physiological data of the elderly person to be tested; S2: Preprocess the multidimensional physiological data to obtain preprocessed data; S3: Perform personalized physiological baseline fitting on the preprocessed data to obtain a personalized physiological baseline; S4: Acquire real-time multidimensional physiological data; S5: Based on the personalized physiological baseline and the real-time multi-dimensional physiological data, obtain the linkage warning result through multi-indicator linkage abnormal warning; S6: Classify the health risk level of the linked early warning results to determine the health risk level of the elderly person to be tested; S7: Based on the stated health risk level and the chronic disease status of the elderly person to be tested, generate health recommendations.

2. The method for generating health recommendations for the elderly based on physiological data according to claim 1, characterized in that, S2 specifically includes: S201: Standardize the format of the multi-dimensional physiological data; S202: Perform timeline synchronization processing on the data after it has been formatted into a unified format; S203: Perform data cleaning on the timeline-synchronized data to obtain cleaned data; S204: The cleaning data is denoised by sensor-differentiated denoising to obtain denoised data; S205: Perform data standardization processing on the noise reduction data to obtain standardized data; S206: Perform cross-device consistency verification on the standardized data to obtain the preprocessed data.

3. The method for generating health recommendations for the elderly based on physiological data according to claim 2, characterized in that, S203 specifically includes: S2031: Perform missing value repair on the data after timeline synchronization; when the missing data duration is greater than the preset duration, repair it using a long-term missing value repair method; when the missing data duration is less than or equal to the preset duration, repair it using a short-term missing value repair method. The long-term missing data repair method specifically involves: triggering a device reconnection reminder when the duration of data loss is a critical indicator; and marking the duration of data loss as an invalid field when the duration of data loss is a non-critical indicator. The short-term missing data repair method specifically includes: a linear interpolation algorithm. ;in, express t Duration of missing data at any given moment express t The data value at time -1 is the data value from the previous valid time. express t The data value at time +1 is the data value at the next valid time. T Indicates the duration of the missing information. express t Time and t -1 time interval; S2032: Perform outlier filtering on the data after missing value repair to obtain the cleaned data.

4. The method for generating health recommendations for the elderly based on physiological data according to claim 2, characterized in that, S204 specifically includes: S2041: For heart rate or blood oxygen data from the wristband, a Kalman filter algorithm is used for noise reduction. State prediction: Covariance prediction: Kalman gain: Status Update: ;in, express k Value after noise reduction at any time express k The value after noise reduction at time -1 express k Original measurement value at time, A Represents the state transition matrix. H Represents the observation matrix. Q Indicates process noise. R Indicates measurement noise. express k Predicted state value at time of day B Represents the control input matrix. express k Predict the state covariance matrix at each time step. express k The optimal estimated covariance matrix at time -1 express k Time-based Kalman gain; S2042: For continuous glucose monitoring data, a moving average filtering algorithm is used for noise reduction. ;in, express k Blood sugar levels after noise reduction express v Original value at time, K Indicates the total duration; S2043: For blood pressure monitor data, a median filtering algorithm is used for noise reduction.

5. The method for generating health recommendations for the elderly based on physiological data according to claim 1, characterized in that, S3 specifically includes: S301: The preprocessed data is labeled to obtain initial baseline data with feature labels; S302: Calculate the individual feature weights of the initial baseline data to form an individual feature weight vector; S303: Based on the initial baseline data and the individual feature weight vector, calculate the baseline core parameters, wherein the baseline core parameters include the weighted mean and the weighted standard deviation; The weighted mean is specifically: ;in, express t Preprocessed data at any given time, express t Time of the first i Label values ​​of individual characteristics, Indicates the first i The weights of individual characteristics, n Indicates the total number of valid data points. This represents the weighted mean; the weighted standard deviation is specifically: ;in, Indicates the weighted standard deviation; S304: Perform baseline dynamic calibration on the aforementioned baseline core parameters: ; ; in, Represents the smoothing coefficient. This represents the weighted average of the new data for the current month. This represents the weighted mean of the original baseline. This represents the weighted standard deviation of the new data for the current month. This represents the weighted standard deviation of the original baseline. This represents the weighted mean after dynamic calibration. This represents the weighted standard deviation after dynamic calibration; S305: Determine the range of the personalized physiological baseline based on the dynamically calibrated baseline core parameters: ;in, b This indicates the risk coefficient.

6. The method for generating health recommendations for the elderly based on physiological data according to claim 1, characterized in that, S5 specifically includes: S501: Based on the real-time multidimensional physiological data, calculate the correlation strength between each indicator using the Pearson correlation coefficient; The specific formula for calculating the Pearson correlation coefficient is as follows: ;in, express t Time indicators i The value, Indicators i Personalized baseline mean, Indicate t Time indicators j The value, Indicators j Personalized baseline mean, Indicators i and indicators j The correlation weight, with a value range of [-1, 1]; S502: Based on the real-time multi-dimensional physiological data, calculate the deviation of each indicator from the personalized physiological baseline: ;in, Indicates the first i The deviation of each indicator from the single indicator. Indicates the first i The weighted baseline mean of each indicator, Indicates the first i The weighted baseline standard deviation of each indicator. Indicates the first i The value of each indicator; S503: Calculate the linkage risk value based on the linkage strength and the deviation of the single index; S504: Based on the aforementioned linkage risk value, determine the linkage anomaly; S505: Filter false alarms from abnormal judgment results to obtain health abnormal event alerts; S506: Based on the health abnormality event alarm, output the linkage early warning result.

7. The method for generating health recommendations for the elderly based on physiological data according to claim 6, characterized in that, The specific formula for calculating the linked risk value is as follows: ;in, R Indicates the linked risk value. m This indicates the total number of indicators involved in the linkage. d i and d j They represent the first i The deviation of the single indicator and the first indicator j The deviation of each indicator.

8. The method for generating health recommendations for the elderly based on physiological data according to claim 6, characterized in that, Specifically, S505 is: Based on the anomaly determination results, access environmental data and device status, exclude unhealthy anomalies, and obtain the health anomaly event alarm.

9. A health advice generation system for the elderly based on physiological data, characterized in that, include: Processor and memory; The memory stores programs or instructions that can run on the processor, which, when executed by the processor, implement the steps of the method for generating health recommendations for the elderly based on physiological data as described in any one of claims 1 to 8.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the method for generating health recommendations for the elderly based on physiological data as described in any one of claims 1 to 8.

Images