A multi-functional guidance display device for hospitals

Abstract

This invention relates to the field of medical data processing technology, specifically to a multifunctional guidance and display device for hospitals, which performs the following steps: acquiring physiological data and normal data for each type of physiological parameter of the patient to be monitored; obtaining a deviation baseline based on the differences between the physiological data and normal data of each type of physiological parameter at different times of the day and the differences between the physiological data at each time and adjacent historical times; obtaining time nodes based on the differences between the deviation baselines of each type of physiological parameter at the same time on different days and the clustering of differences corresponding to the same time on different days; obtaining behavioral pattern time periods based on the differences in the changing trends of physiological data of each type of physiological parameter before and after the time nodes of each day; and monitoring and displaying abnormalities of physiological parameters based on the deviation baseline and behavioral pattern time periods. This invention avoids misjudging regular normal data as abnormal, thus improving the accuracy of visualization results.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and more specifically to a multifunctional guidance display device for hospitals. Background Technology

[0002] Patient status monitoring systems in hospitals are an important part of modern medical care. With the advancement of technology, medical equipment is gradually developing towards intelligence and digitalization. By integrating sensors and monitoring devices, the system can track patients' health status in real time, such as heart rate, blood pressure, body temperature, and blood oxygen, and organize and visualize this health status data. This can effectively help medical staff to obtain information intuitively and clearly, understand changes in patients in a timely manner, and respond accordingly.

[0003] When implementing multi-functional guidance display devices in hospitals, it is often necessary to extract abnormal data during the monitoring process to facilitate medical staff in quickly obtaining patient status. Current technologies typically only extract data exhibiting abnormal fluctuations or numerical anomalies. However, patient behavior can also cause some data to show suspicious changes, leading to misidentification and resulting in poor accuracy of the visualization results. Summary of the Invention

[0004] To address the technical problem of existing methods easily leading to misidentification and resulting in poor accuracy of visualization results, the present invention aims to provide a multifunctional guidance display device for hospitals. The specific technical solution adopted is as follows:

[0005] A multi-functional guidance display device for hospitals includes a memory and a processor, the processor executing a computer program stored in the memory to perform the following steps:

[0006] Acquire physiological data for each type of physiological parameter of the patient to be monitored at different times of the day, as well as the normal data for the corresponding category;

[0007] Based on the differences between the physiological data and normal data of each type of physiological parameter at different times of the day, and the differences between the physiological data at each time and the adjacent historical time, the deviation base of each type of physiological parameter at each time of the day is obtained.

[0008] Based on the differences in deviations of each type of physiological parameter from the baseline at the same time on different days, and the clustering of differences corresponding to the same time on different days, time nodes are obtained by filtering the moments within a day.

[0009] Based on the differences in the changing trends of physiological data for each type of physiological parameter before and after each time point in the day, all time points are filtered to obtain the time periods of behavioral patterns.

[0010] Based on the deviation of each type of physiological parameter from the baseline at each time of day and the time period of behavioral patterns, abnormalities in physiological parameters are monitored and displayed.

[0011] Preferably, the step of filtering time points within a day based on the differences in deviations of each type of physiological parameter at the same time on different days, and the clustering of differences corresponding to the same time on different days, specifically includes:

[0012] For any class of physiological parameters;

[0013] Based on the differences between the deviation base values ​​at the same time on different days, and combined with the deviation base values ​​at each time within the time neighborhood of the same time, the difference components at each time in each day are obtained;

[0014] Based on the difference distance between the difference components of different days at the same time, clustering is performed on all days corresponding to the same time to obtain the number of days cluster for each time.

[0015] Based on the maximum number of days contained in all day clusters at each time point, and the deviation from the baseline for each day in the day cluster corresponding to the maximum value, the abnormal pattern evaluation value for each time point is obtained.

[0016] Based on the evaluation value of the abnormal patterns at each moment, all moments are filtered to obtain time nodes.

[0017] Preferably, the step of obtaining the difference component for each moment of each day based on the difference between the deviation base values ​​of different days at the same time, combined with the deviation base values ​​of each moment within the time neighborhood of the same moment, specifically includes:

[0018] Record any day as the base day, any time as the target time, and other days other than the base day as reference days. Record each time within the time neighborhood of the target time of the reference day as the control time.

[0019] The data difference coefficient between the benchmark date and the arbitrary reference date is determined based on the minimum difference between the deviation base of the benchmark date at the target time and the deviation base of any reference date at each control time.

[0020] Based on the time interval between the reference time corresponding to the minimum difference between the target time and any reference day, the time difference coefficient between the base day and the arbitrary reference day is determined.

[0021] Based on the balance of the product between the data difference coefficients and time difference coefficients between the base date and each reference date, the difference component of the target time in the base date is determined.

[0022] Preferably, the step of obtaining the anomaly evaluation value for each moment based on the maximum number of days contained in all day clusters at each moment, and the deviation from the baseline for each day in the day cluster corresponding to the maximum value, specifically includes:

[0023] In the clustering results at the target time, the cluster corresponding to the maximum number of days included is taken as the feature cluster;

[0024] The proportion of days contained in the feature cluster is used as the first coefficient, the mean of the deviation of all days in the feature cluster from the base at the target time is used as the second coefficient, and the negative correlation coefficient of the mean of the difference components of all days in the feature cluster at the target time is used as the third coefficient.

[0025] The product of the first, second, and third coefficients is normalized to obtain the abnormal pattern evaluation value at the target time.

[0026] Preferably, the step of filtering all moments based on the abnormal pattern evaluation value at each moment to obtain time nodes specifically includes:

[0027] The time point corresponding to the abnormal pattern evaluation value being greater than or equal to the preset pattern threshold is taken as the time node.

[0028] Preferably, the step of filtering all time points to obtain behavioral pattern time periods based on the differences in the changing trends of physiological data for each type of physiological parameter before and after each daily time point specifically includes:

[0029] For any type of physiological parameter, any time point is selected as the chosen time point. The first fitted line is obtained by performing linear fitting on the physiological data before the selected time point each day, and the second fitted line is obtained by performing linear fitting on the physiological data after the selected time point each day.

[0030] The mutation factor at the selected time node is obtained based on the difference in the slope between the first and second fitted lines each day, and the deviation base of the first and second fitted lines at the corresponding time points.

[0031] The normalized sum of the mutation factors of all days at the selected time point is taken as the mutation coefficient of the selected time point.

[0032] Based on the mutation coefficient of each time point, all time points are filtered to obtain the time periods of behavioral patterns.

[0033] Preferably, the step of obtaining the mutation factor at a selected time node each day based on the difference in the slope between the first and second fitted lines each day, and the deviation base of the first and second fitted lines at corresponding times, specifically includes:

[0034] For any given day, the normalized value of the angle between the first fitted line and the second fitted line is taken as the first feature difference.

[0035] The absolute value of the difference between the mean deviation from the baseline at each time corresponding to the first fitted line and the mean deviation from the baseline at each time corresponding to the second fitted line is used as the second feature difference.

[0036] The mutation factor at a selected time point on any given day is obtained by calculating the product of the first feature difference and the second feature difference.

[0037] Preferably, the step of filtering all time nodes based on the mutation coefficient of the time nodes to obtain the behavioral pattern time period specifically includes:

[0038] The average mutation coefficient of all time nodes within the time neighborhood of each time node is used as the start-up evaluation value of each time node. The time neighborhood of the time node whose start-up evaluation value is greater than or equal to the preset start-up threshold is used as the behavioral pattern time period.

[0039] Preferably, the step of monitoring and displaying abnormalities in physiological parameters based on the deviation of each type of physiological parameter from the baseline at each time of day and the time period of behavioral patterns specifically includes:

[0040] For any type of physiological parameter, when the deviation from the baseline at any time of day is greater than or equal to the preset abnormal threshold, and that time does not belong to the period of behavioral regularity, the abnormal physiological parameter will be displayed at that time.

[0041] Preferably, the step of obtaining the deviation baseline for each type of physiological parameter at each time of day based on the differences between physiological data and normal data at different times of day for each type of physiological parameter, and the differences between physiological data at each time point and adjacent historical times, specifically includes:

[0042] For any type of physiological parameter, the difference between the physiological data at each time of day and the corresponding normal data is taken as the first difference coefficient, and the ratio between the difference between the physiological data at each time of day and the physiological data at the previous time and the time interval is taken as the second difference coefficient; the normalized result of the product between the first difference coefficient and the second difference coefficient is calculated to obtain the deviation base of the arbitrary type of physiological parameter at each time of day.

[0043] The embodiments of the present invention have at least the following beneficial effects:

[0044] This invention first collects various physiological data, analyzes the differences between these physiological data and normal data of the same type under normal conditions, and combines the differences between physiological data at each moment and adjacent historical moments to comprehensively assess the deviation of each type of physiological parameter at each moment of the day, characterizing the degree of deviation of physiological data from the normal state. Then, it analyzes the differences in deviation from the baseline and the clustering of differences at the same moment on different days, which can measure the data deviation patterns at different time points throughout the day, filtering out the more regular moments to obtain time nodes. Furthermore, by analyzing the changing trends of physiological data before and after each time node, it more accurately analyzes the regular physiological data performance, further filtering out the time ranges of regular daily behaviors, i.e., regular time periods. Finally, by using the deviation baseline and regular time periods of behavior together, it monitors abnormalities in physiological data at each moment. This method avoids misjudging regular normal data as abnormal, improves the accuracy of visualization results, achieves segmented visualization of monitoring information, ensures the focus of abnormal data, and enables medical staff to obtain monitoring information more efficiently. Attached Figure Description

[0045] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart of the steps of a multifunctional guidance display method for hospitals provided by the present invention;

[0047] Figure 2 This is a flowchart of the steps in the method for obtaining behavioral pattern time periods provided by the present invention;

[0048] Figure 3 This is a flowchart of the steps of the method for obtaining differential components provided by the present invention;

[0049] Figure 4 This is a flowchart of the steps in the method for obtaining behavioral pattern time periods provided by the present invention;

[0050] Figure 5 This is a partial schematic diagram of the first and second fitted lines for the time nodes provided by the present invention;

[0051] Figure 6 This is a schematic diagram of the display panel of a multifunctional guidance display device for hospitals provided by the present invention. Detailed Implementation

[0052] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following detailed description, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effects of a multifunctional guidance display device for hospitals based on the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0053] 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 invention pertains.

[0054] The following describes in detail, with reference to the accompanying drawings, a specific solution for a multifunctional guidance display device for hospitals provided by the present invention. Specifically, a multifunctional guidance display device for hospitals includes a memory and a processor, wherein the processor executes a computer program stored in the memory to implement the steps of a multifunctional guidance display method for hospitals.

[0055] Please see Figure 1 The diagram illustrates a flowchart of a multifunctional guidance display method for hospitals according to an embodiment of the present invention, the method comprising the following steps:

[0056] Step S100: Obtain physiological data of each type of physiological parameter of the patient to be monitored at different times of the day and the normal data of the corresponding category.

[0057] It should be noted that the main purpose of this invention is to collect various physiological parameters of a patient under monitoring in real time, and then monitor whether there are any abnormalities at any time of day. The time when there are abnormalities is marked and displayed to assist medical staff in obtaining patient information.

[0058] It should be understood that the implementer can acquire the categories of physiological parameters according to the specific real-time scenario. Various physiological parameters may include patient body temperature, heart rate, blood oxygen saturation, and blood glucose concentration, etc. The acquisition methods for various physiological parameters are well-known technologies, such as infrared thermometry, electrocardiogram monitoring, pulse oximeter, and blood glucose meter, etc., for data acquisition.

[0059] Specifically, data collection of various physiological parameters is carried out at fixed times every day, that is, the time interval between two adjacent times is equal, to obtain the normal value or normal range of each type of physiological parameter under normal conditions. This is a well-known technique and will not be elaborated on here. For example, the normal adult's (axillary) body temperature is generally maintained at 36℃~37℃; the normal adult's heart rate is 60~100 beats / minute; the normal adult's blood oxygen saturation is maintained at 95%~100%; and the normal adult's fasting blood glucose is maintained at 3.9~6.1mmol / L.

[0060] More specifically, in this embodiment, each type of physiological parameter is analyzed separately, meaning the analysis process for different categories of physiological parameters is exactly the same. However, the timing of physiological data collection may differ. Considering the different monitoring cycles for different categories of physiological parameters, different data collection times are adaptively set. For example, heart rate and blood oxygen saturation are generally monitored continuously with short time intervals between adjacent moments, collecting heart rate and blood oxygen saturation data every minute. Blood glucose and body temperature data are generally monitored continuously on an hourly basis, collecting blood glucose concentration and body temperature data every hour.

[0061] It should be understood that the implementer can set the timing for collecting physiological data for each type of physiological parameter according to the specific implementation scenario. Thus, for each category of physiological parameter, one physiological data point is collected at the same time every day, and normal data for each category of physiological parameter under normal conditions is also obtained. This embodiment uses normal data from an adult as an example for illustration.

[0062] Step S200: Based on the differences between the physiological data and normal data of each type of physiological parameter at different times of the day, and the differences between the physiological data at each time and the adjacent historical time, obtain the deviation base of each type of physiological parameter at each time of the day.

[0063] When visualizing patient physiological data, in order to enable medical staff to understand the situation in a timely manner, in addition to a complete overview of various information, it is often necessary to extract potentially abnormal data and present key summaries of the abnormal data. Therefore, the first step is to obtain the data deviation of various physiological parameters based on the numerical fluctuations of the patient's physiological data, providing a data foundation for further analysis of data deviations and anomalies.

[0064] Based on this, by comparing the differences between the physiological data of each type of physiological parameter at each time of day and the normal data of the corresponding category under normal conditions, and combining the data change rate of the physiological data at each time, the data deviation of the physiological data at each time is evaluated to reflect the abnormal data state at the corresponding time.

[0065] It should be noted that the feature analysis process is the same for each type of physiological parameter; therefore, an example of any one type of physiological parameter will be used here. Specifically, for any one type of physiological parameter, the difference between the physiological data at each time of day and the corresponding normal data is used as the first difference coefficient, and the ratio between the difference between the physiological data at each time of day and the physiological data at the adjacent previous time and the time interval is used as the second difference coefficient. The normalized result of the product between the first and second difference coefficients is used to obtain the deviation base of the aforementioned physiological parameter at each time of day.

[0066] As a concrete example, for any type of physiological parameter, taking any time of day as an example, the method for obtaining the deviation from the baseline can be expressed by the formula:

[0067]

[0068] in, This represents the deviation from the baseline at time t on day d. This represents the physiological data at time t on day d. This indicates normal data. This represents the physiological data at time t-1 on day d. Indicates the time interval between adjacent moments. It is a linear normalization function.

[0069] When the normal data for a certain category of physiological parameters is a data range with upper and lower boundary values, the first difference coefficient... Take the maximum value between the calculation results of the upper and lower boundaries. The value of the first difference coefficient can be specifically expressed as... , This represents the upper boundary value of the normal range for the corresponding category of physiological parameters. This represents the lower boundary value of the normal range for the corresponding category of physiological parameters.

[0070] It should be noted that, regarding blood glucose data, considering that the normal range of blood glucose is time-dependent (significantly affected by food intake), in this step, the deviation of physiological data can be analyzed first using the normal fasting range as a benchmark, and then the postprandial state can be distinguished through correlation analysis between the data. It should be further noted that when time t is the first time of the day, the second difference coefficient is calculated using the last time of the adjacent previous day.

[0071] First difference coefficient This reflects the deviation of physiological data from the normal range at each moment of the day, and thus reflects the relative abnormality of physiological data from the normal state at each moment. The second difference coefficient... It reflects the instantaneous rate of change of each moment compared to historical data. The larger the value of the first difference coefficient and the larger the value of the second difference coefficient, the greater the deviation of the physiological data at the corresponding moment from the normal state, and the greater the instantaneous rate of change. This indicates a greater possibility that the physiological data at that moment is abnormal, and the larger the deviation from the baseline value.

[0072] Thus, by using the same method, we can obtain the deviation base of each type of physiological parameter at each time of day, which represents the degree to which the physiological data at the corresponding time deviates from the normal state and the probability of abnormality in the physiological data at the corresponding time.

[0073] Step S300: Based on the differences between the deviation base values ​​of each type of physiological parameter at the same time on different days, and the clustering of differences corresponding to the same time on different days, the time points within a day are filtered to obtain time nodes.

[0074] In actual monitoring, some of the patient's behaviors may also cause changes in physiological indicators, such as an increase in heart rate after exercise or an increase in blood sugar after eating. The increase in blood sugar at this time is a normal physiological phenomenon and should be different from the standard of normal range of blood sugar indicators in other states. Therefore, the fluctuations in data of these normal physiological changes should be specially marked rather than directly classified as abnormal.

[0075] In hospital nursing monitoring, patients' lives are relatively regular. For example, behaviors such as medication, sleep, eating, and exercise rehabilitation often occur at similar times of the day. Therefore, the changes in physiological indicators caused by these behaviors also show certain regularities. Thus, this step analyzes the patterns of physiological indicator changes between different days in order to extract the key time points that affect the changes in physiological indicators.

[0076] Based on this, the abnormality pattern at each time point is evaluated by comparing the regularity of the abnormal baseline change trend across different days. Taking blood glucose as an example, if a patient eats at similar times each day, and blood glucose gradually rises after eating, approaching the blood glucose threshold, the abnormal baseline will gradually increase during this process. If, through comparison over multiple days, similar blood glucose elevation characteristics are observed at similar times, then this abnormal process is regular and may be due to changes in related data caused by lifestyle behaviors.

[0077] As a concrete example, the methods for obtaining the time points for various physiological parameters are exactly the same. Here, we will use physiological data under any one type of physiological parameter and deviations from the baseline as an example for illustration. For example... Figure 2 As shown, the method for obtaining the time node can be implemented by steps S301 to S304.

[0078] Step S301: Based on the differences between the deviation base values ​​of different days at the same time, and combined with the deviation base values ​​of each time within the time neighborhood of the same time, the difference component of each time in each day is obtained.

[0079] Specifically, such as Figure 3 As shown, the method for obtaining the difference component can be implemented by steps S3011 to S3014.

[0080] Step S3011: Record any day as the base day, record any time as the target time, record other days besides the base day as reference days, and record each time within the time neighborhood of the target time of the reference day as the control time.

[0081] In this embodiment, day d is taken as the base day and time t is taken as the target time. It should be understood that the physiological data and deviation from the base corresponding to time t can be obtained every day, which means the data distribution of the same time on different days.

[0082] Considering that patients in hospitals have regular daily routines, but not necessarily perform their activities at the exact same time every day, the feature analysis process in this embodiment takes into account the allowable time offset range. That is, it performs joint analysis on the distribution of physiological data from multiple moments within the time neighborhood of each moment. In this embodiment, the allowable time offset range is 30 minutes long. Specifically, the time neighborhood range of each moment is defined as the time range from 30 minutes before the moment to 30 minutes after the moment. In other embodiments, the implementer can set it according to the specific implementation scenario.

[0083] Furthermore, in order to facilitate the observation of patients' daily routines, the number of days for routine feature analysis should be at least 7 days, that is, at least 6 reference days in addition to the baseline day. Within each reference day, the control times of the t-th time and its time neighborhood range can be obtained.

[0084] Step S3012: Based on the minimum difference between the deviation base of the benchmark date at the target time and the deviation base of any reference date at each control time, determine the data difference coefficient between the benchmark date and the arbitrary reference date.

[0085] Specifically, by comparing the deviation of the benchmark date from the target time with the deviation of each reference date within the time neighborhood of the target time, the moment with the smallest data difference can be located within the allowable time offset range, so as to analyze the difference in data fluctuation between the benchmark date and the reference date within the approximate time range of the target time.

[0086] As a specific example, taking the i-th reference day as an example, the absolute value of the difference between the deviation base of the benchmark day at the target time and the deviation base of each control time within the time neighborhood of the i-th reference day at the target time is calculated. The minimum value of the absolute value of the difference corresponding to all control times of the i-th reference day at the target time is determined as the data difference coefficient between the benchmark day and the i-th reference day at the target time.

[0087] The data difference coefficient reflects the data fluctuations of the base date and reference date within a similar time range. The smaller the value, the more similar the life behaviors of the base date and reference date are within a similar time range, resulting in smaller fluctuations in the deviation of physiological data.

[0088] Step S3013: Based on the time interval between the reference time corresponding to the minimum difference between the target time and any reference day, determine the time difference coefficient between the base day and the arbitrary reference day.

[0089] Specifically, the reference time of the i-th reference day corresponding to the minimum absolute value of the difference between all reference times within the time neighborhood of the i-th reference day and the target time is taken as the matching time of the target time. The matching time represents the time point within the time neighborhood of the i-th reference day that is closest to the data fluctuation characteristics of the target time of the benchmark day.

[0090] The ratio of the interval between the matching time and the target time to the total duration of the time neighborhood is used as the delay time percentage. The sum of the delay time percentage and the value of 1 is used as the time difference coefficient between the base day and the i-th reference day at the target time. The delay time percentage is incremented by 1 to prevent the time difference coefficient from being zero when the matching time is the target time, thus avoiding any impact on the feature analysis results of the data difference coefficient when the value is zero.

[0091] The time difference coefficient reflects the time interval between the closest data fluctuation states on the benchmark date and the reference date within their respective time ranges. A larger value indicates a greater time offset between similar or nearly identical data fluctuation states on different days, suggesting a temporal deviation in the patient's behavioral habits, requiring a larger time weighting for difference calculations. Conversely, a smaller value indicates a smaller time offset between similar or nearly identical data fluctuation states on different days, suggesting more regularity in the patient's behavioral habits, requiring a smaller time weighting for difference calculations.

[0092] Step S3014: Based on the balance of the product between the data difference coefficient and the time difference coefficient between the base date and each reference date, determine the difference component of the target time in the base date.

[0093] As a concrete example, the mean of the products of the base date and all reference dates reflects the equilibrium situation. More specifically, the difference component of the target time in the base date can be expressed by the formula:

[0094]

[0095] in, This represents the difference component at the target time within the base date, where d represents the d-th day and t represents the t-th time of the day; This represents the deviation from the baseline at time t on day d. Indicates the number of reference days. This represents the deviation from the baseline at time n within the time neighborhood of time t in the i-th reference day. This represents the temporal neighborhood range at time t. This represents the time interval between the matching time and the target time in the i-th reference day. The time length representing the time neighborhood; This represents the function that takes the minimum value. It is a linear normalization function. The coefficient of variation is the number of data points. This is the time difference coefficient.

[0096] The difference component reflects the equilibrium difference between the data deviation on the base date at the target time and the data deviation at similar time points on other days. A smaller value indicates a more similar fluctuation between the deviation base on the base date at the target time and the data analysis results of other days, suggesting a greater likelihood of a regularity in the data deviation on the base date at the target time. Conversely, a larger value indicates a greater difference between the deviation base on the base date at the target time and the data analysis results of other days, suggesting a lower likelihood of a regularity in the data deviation on the base date at the target time, and a greater likelihood that the abnormal deviation is a genuine anomaly.

[0097] Step S302: Based on the difference distance between the difference components of different days at the same time, cluster all the days corresponding to the same time to obtain the number of days cluster for each time.

[0098] Specifically, taking the target time as an example, the Euclidean distance between the difference components of any two different days at the target time is calculated as the distance metric for clustering. A clustering algorithm is then used to cluster all days, resulting in the day clustering result for the target time, which includes multiple day clusters. Within the same day cluster, each day exhibits similar differences in its components at the target time.

[0099] Among them, clustering algorithms can be AP clustering algorithm, DBSCAN clustering algorithm, etc., and implementers can choose according to the specific implementation scenario. The clustering process is a well-known technology and will not be described in detail here.

[0100] Therefore, it should be understood that each moment in a day corresponds to a clustering result. The clustering result corresponding to a moment reflects the similarity of data distribution over a certain number of days, providing a data basis for subsequent assessment of the probability that each moment in a day belongs to the patient's regular activity time.

[0101] Step S303: Based on the maximum number of days contained in all day clusters at each time point, and the deviation base corresponding to each day in the day cluster corresponding to the maximum value, obtain the abnormal pattern evaluation value for each time point.

[0102] Specifically, taking the target time as an example, the clustering results at the target time include several day clusters. Each day cluster contains several days with similar differential components at the target time. When the physiological data at the target time has regularity, the clustering results at the target time will show a cluster with a small mean of differential components and a large proportion of days in the day cluster, as well as several other outliers. That is, the physiological data at the target time shows similar deviation characteristics on most days. For example, patients show relatively similar data deviation characteristics at 12 o'clock on most days (and within half an hour before and after 12 o'clock).

[0103] Based on this, the regularity of the target time is evaluated by the cluster distribution characteristics formed by the distribution of the differential components in the clustering results at the target time.

[0104] The first step is to select the cluster corresponding to the maximum number of days in the clustering results at the target time as the feature cluster.

[0105] Feature clusters reflect the clusters with the largest proportion of similar difference components in the clustering results at the target time. The more days a feature cluster contains, the smaller the difference component values ​​of all days at the target time, indicating that the regularity reflected at the target time is stronger.

[0106] The second step is to use the proportion of days contained in the feature cluster as the first coefficient, the mean of the deviation of all days in the feature cluster from the base at the target time as the second coefficient, and the negative correlation coefficient of the mean of the difference components of all days in the feature cluster at the target time as the third coefficient. The product of the first, second and third coefficients is normalized to obtain the abnormal pattern evaluation value at the target time.

[0107] As a specific example, this embodiment takes the t-th moment of a day as the target moment. The method for obtaining the abnormal pattern evaluation value of the target moment can be expressed by the formula:

[0108]

[0109] in, The evaluation value representing the abnormal pattern at the target time. This represents the total number of days contained in the largest cluster in the clustering results at time t, which is also the total number of days contained in the feature clusters in the clustering results at time t. Indicates the total number of days; Let represent the mean of the difference components of all days at time t contained in the feature cluster at time t in the clustering results at time t. This represents the mean deviation from the cardinality of all days included in the feature cluster at time t in the clustering results at time t. It is a linear normalization function.

[0110] The first coefficient reflects the proportion of days with the largest cluster in the clustering results at the target time. The larger the value, the more days the difference components are close at the target time, and the greater the possibility that the deviation characteristics of the data at the target time have a regularity.

[0111] The second coefficient reflects the balanced performance of the deviation characteristics of the largest cluster in the clustering results at the target time. The larger the value, the greater the possibility of anomalies at the target time. The second coefficient is added to exclude normal times when no data anomalies occur.

[0112] The third coefficient is used because the value of each difference component is the result of normalization. This embodiment uses... As a negative correlation coefficient It reflects the balanced performance of the difference component of the largest cluster in the clustering results at the target time. The smaller the value, the larger the corresponding value of the third coefficient, indicating that more days at the target time show similar fluctuations, and the greater the possibility that the deviation of the data at the target time is regular.

[0113] Step S304: Based on the evaluation value of the abnormal patterns at each moment, filter all moments to obtain time nodes.

[0114] The abnormality pattern evaluation value characterizes the likelihood that an abnormal situation exists at the target time but is caused by regular daily life behaviors. The higher the abnormality pattern evaluation value at the target time, the more likely there may be a large abnormal fluctuation in physiological data at the target time. However, since regular daily life behaviors exist at the target time, this abnormal fluctuation is considered a normal, regular fluctuation.

[0115] The smaller the abnormal pattern evaluation value at the target time, the more likely there is a large abnormal fluctuation in physiological data at the target time. Since there is no regular daily behavior at the target time, this abnormal fluctuation is a real abnormal fluctuation.

[0116] Based on this, the time point corresponding to an abnormal pattern evaluation value greater than or equal to a preset pattern threshold is taken as the time node. In this embodiment, the pattern threshold is set to 0.6, but implementers can set it according to specific implementation scenarios. The time node represents the time points in a day when regular daily behaviors may occur; for example, there may be regular eating behavior at noon, and regular walking activities may occur at 7 pm.

[0117] Step S400: Based on the differences in the changing trends of physiological data for each type of physiological parameter before and after each daily time point, all time points are filtered to obtain the behavioral pattern time period.

[0118] Considering that the fluctuations in physiological data caused by different lifestyle behaviors often last for a period of time, time nodes are general representations of the time points where patterns exist, but cannot accurately represent the time points when behaviors begin. Therefore, it is necessary to conduct specific analysis on the data fluctuation patterns before and after each time node.

[0119] Regarding the starting point of a daily activity, physiological data will fluctuate after the activity begins, while the fluctuation is smaller before the activity begins. Therefore, the difference in physiological data fluctuation before and after the starting point of the activity is the greatest. For example, the heart rate is relatively stable and almost unchanged before exercise. Taking the start of exercise as the dividing point, the difference in the trend of heart rate change between the two periods before and after is the greatest.

[0120] As a concrete example, the methods for obtaining the time points for various physiological parameters are exactly the same. Here, we will use the data analysis results for any one type of physiological parameter as an example for illustration. For example... Figure 4 As shown, the method for obtaining the behavioral pattern time period can be implemented by steps S401 to S404.

[0121] More specifically, the data analysis process is the same for all time points. This embodiment uses any one time point as an example for illustration, that is, for any type of physiological parameter, any one time point is selected.

[0122] Step S401: Perform linear fitting on the physiological data before the selected time point each day to obtain a first fitted line, and perform linear fitting on the physiological data after the selected time point each day to obtain a second fitted line.

[0123] In this embodiment, a certain time length is set, and a first fitted line is obtained by performing linear fitting on the physiological data of all moments within 30 minutes before the selected time node of each day, and a second fitted line is obtained by performing linear fitting on the physiological data of all moments within 30 minutes after the selected time node of each day.

[0124] It should be understood that each selected time point corresponds to a first fitted line and a second fitted line. Furthermore, by analyzing the differences in data trends before and after the selected time point as the dividing point, a more accurate starting point for daily activities can be determined.

[0125] Step S402: Based on the difference in the degree of inclination between the first and second fitted lines each day, and the deviation base of the first and second fitted lines at corresponding times, obtain the mutation factor at the selected time node each day.

[0126] Specifically, for any given day, the normalized value of the angle between the first fitted line and the second fitted line is taken as the first feature difference; the absolute value of the difference between the mean deviation from the base at each time corresponding to the first fitted line and the mean deviation from the base at each time corresponding to the second fitted line is taken as the second feature difference; the product of the first feature difference and the second feature difference is calculated to obtain the mutation factor of the given day at the selected time node.

[0127] More specifically, taking any given day as an example, this embodiment uses the r-th time node as the selected time node. The method for obtaining the mutation factor at the selected time node on day d can be expressed by the formula:

[0128]

[0129] in, Let represent the mutation factor at the selected time point on day d, and r represent the r-th time point; This represents the angle between the first and second fitted lines at the r-th time node on day d. This represents the mean deviation from the base at all times on the first fitted line at the r-th time node on day d. This represents the mean deviation from the base at all times on the second fitted line at the r-th time node on day d.

[0130] The first feature difference reflects the difference between the changing trends of physiological data on the first and second fitted lines. The larger the value, the greater the difference in the changing trends between the two, and the greater the probability that the selected time point belongs to the time point of the start of life behavior. The second feature difference reflects the difference between the data deviations from the features on the first and second fitted lines. The larger the value, the greater the feature difference between the two, and the greater the probability that the selected time point belongs to the time point of the start of life behavior. At this time, the value of the mutation factor is larger.

[0131] Thus, the mutation factor characterizes the degree of mutation in physiological data showing significant changes starting from each time point of each day. The greater the degree of mutation, the more likely the time point is the starting point of daily life activities. Furthermore, by synthesizing the mutation performance of the same time point across all days, the characteristic mutation situation of each time point is accurately and comprehensively reflected.

[0132] It should be noted that the angle between two straight lines generally ranges from 0° to 90°. Therefore, this embodiment uses the ratio of the included angle to 90° to reflect the normalized result of the included angle. In other embodiments, the implementer may also use other normalization methods for processing. For example... Figure 5 As shown, The angle between the two fitted lines is given by... Figure 5 It can be seen that the data change trend of the fitted line before the time node is relatively flat, while the data change trend of the fitted line after the time node is relatively fast. Therefore, the time node is more likely to be the time point when the life behavior begins.

[0133] Step S403: The normalized result of the sum of mutation factors of all days at the selected time node is used as the mutation coefficient of the selected time node.

[0134] Specifically, the normalization method can be the minimax normalization method, and the implementer can also choose according to the specific implementation scenario.

[0135] Step S404: Based on the mutation coefficient of the time nodes, filter all time nodes to obtain the behavioral pattern time period.

[0136] Since each day does not necessarily start at a specific moment, meaning the starting time of daily activities may vary, the starting range of activities is determined by time nodes within a certain time frame.

[0137] Specifically, the average mutation coefficient of all time nodes within the time neighborhood of each time node is used as the activation evaluation value of each time node. The time neighborhood of time nodes whose activation evaluation values ​​are greater than or equal to a preset activation threshold is defined as the behavioral pattern time period. The activation threshold is set to 0.7, which can be set by the implementer according to the specific implementation scenario.

[0138] In this embodiment, the time neighborhood range of each time node is a time period centered on each time node with a duration of 1 hour. For example, if the time node is 12 o'clock, then the time neighborhood range of that time node is from 11:30 to 12:30.

[0139] When the activation evaluation value of the selected time node is greater than or equal to the activation threshold, it indicates that the abrupt change coefficient of multiple time nodes within the time neighborhood of the selected time node is relatively balanced. This suggests that there is a high probability of regular life behaviors within the time neighborhood of the selected time node. Therefore, the time period corresponding to the time neighborhood of the selected time node is taken as the behavioral regularity time period.

[0140] When the evaluation value at the selected time point is less than the threshold, it indicates that the balance of the mutation coefficients of multiple time points within the time neighborhood of the selected time point is small, which in turn indicates that the possibility of regular life behavior within the time neighborhood of the selected time point is small. In this case, regular behavior is not marked.

[0141] Therefore, the behavioral pattern time period represents the time range during which regular behavior occurs. That is, within this time period, there may be behaviors that cause fluctuations in physiological data (such as eating causing a rise in blood sugar). Although data fluctuations occur during these time periods, they are normal physiological phenomena caused by behavior and should be displayed separately from abnormal data to avoid confusion with abnormal data and reduce the focus of abnormal data display.

[0142] Step S500: Based on the deviation of each type of physiological parameter from the baseline at each time of day and the behavioral pattern time period, monitor and display abnormalities of the physiological parameters.

[0143] In this embodiment, the display method for each type of physiological parameter is exactly the same. Here, we will use any one type of physiological parameter as an example for explanation. Physiological data is collected in real time at each moment of the day. The deviation from the baseline at each moment of the day is calculated according to the method in step S200. When the deviation from the baseline at each moment of the day is greater than or equal to a preset abnormal threshold, and that moment does not belong to a period of behavioral regularity, the physiological parameter at that moment is displayed as an anomaly. The abnormal threshold is set to 0.4, and the implementer can set it according to the specific implementation scenario.

[0144] By analyzing the deviations of each type of physiological parameter from the baseline and the time periods of behavioral patterns at each time of day, the data is visualized for moments exhibiting abnormalities. As a concrete example, ... Figure 6 As shown, the main content displayed can include: Area 1 Information Overview, including records of all complete physiological data; Area 2 Abnormal Summary, including records of abnormal moments that occurred in the past 24 hours; Area 3 Behavioral Records, including records of moments when data fluctuations were caused by behavior. The implementer can configure the displayed content according to the specific implementation scenario to show the various data required during abnormal moments.

[0145] More specifically, when the deviation from the baseline at a certain moment is greater than or equal to the anomaly threshold, it indicates a greater degree of deviation of the physiological data from the overall state at that moment, and thus a greater likelihood of an anomaly at that moment. If this moment falls within a regular behavior period, it suggests that the data fluctuation at that moment may be influenced by regular behavior, and can be categorized as a daily behavior record for display, rather than being displayed as an anomaly, i.e., displayed in area 3 above. If this moment does not fall within a regular behavior period, it indicates that the data deviation at that moment is a genuine anomaly fluctuation, and therefore, this moment is displayed as an anomaly, i.e., displayed in area 2 above. This display method can avoid misjudging regular, normal data as anomalies, thus preventing the focus on displaying abnormal data.

[0146] This visualizes the physiological data collected during hospital nursing monitoring, enabling relevant medical staff to quickly obtain the patient's physiological status and promptly detect any potential abnormalities, thus realizing a multifunctional guidance display device for hospital use.

[0147] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A multi-functional guidance display device for hospitals, comprising a memory and a processor, characterized in that, The processor executes the computer program stored in the memory to perform the following steps: Acquire physiological data for each type of physiological parameter of the patient to be monitored at different times of the day, as well as the normal data for the corresponding category; Based on the differences between the physiological data and normal data of each type of physiological parameter at different times of the day, and the differences between the physiological data at each time and the adjacent historical time, the deviation base of each type of physiological parameter at each time of the day is obtained. Based on the differences in deviations from the baseline for each type of physiological parameter at the same time on different days, and the clustering of differences corresponding to the same time on different days, time nodes are obtained by filtering the moments within a day, specifically including: For any type of physiological parameter; based on the differences in deviation from the baseline at the same time on different days, and combined with the deviation from the baseline at each time within the time neighborhood of the same time, the difference component at each time of each day is obtained; based on the difference distance between the difference components at the same time on different days, all days corresponding to the same time are clustered to obtain the number of days clusters at each time; based on the maximum number of days contained in all the number of days clusters at each time, and the deviation from the baseline corresponding to each day in the number of days cluster corresponding to the maximum value, the abnormal pattern evaluation value at each time is obtained; based on the abnormal pattern evaluation value at each time, all times are filtered to obtain time nodes; Based on the differences in the changing trends of physiological data for each type of physiological parameter before and after each daily time point, all time points are filtered to obtain the time periods of behavioral patterns, specifically including: For any type of physiological parameter, any time point is selected. A first fitted line is obtained by fitting a straight line to the physiological data before the selected time point each day, and a second fitted line is obtained by fitting a straight line to the physiological data after the selected time point each day. The mutation factor at the selected time point is obtained based on the difference in the slope between the first and second fitted lines each day, and the deviation from the baseline at the corresponding time points of the first and second fitted lines. The normalized sum of the mutation factors at the selected time points for all days is used as the mutation coefficient at the selected time point. Based on the mutation coefficients at the time points, all time points are filtered to obtain the behavioral pattern time period. Based on the deviation of each type of physiological parameter from the baseline at each time of day and the time period of behavioral patterns, abnormalities in physiological parameters are monitored and displayed.

2. The multifunctional guidance display device for hospitals according to claim 1, characterized in that, The method involves determining the difference component for each moment of each day based on the differences in deviation base values ​​across different days at the same time, combined with the deviation base values ​​for each moment within the time neighborhood of that same moment. Specifically, this includes: Record any day as the base day, any time as the target time, and other days other than the base day as reference days. Record each time within the time neighborhood of the target time of the reference day as the control time. The data difference coefficient between the benchmark date and the arbitrary reference date is determined based on the minimum difference between the deviation base of the benchmark date at the target time and the deviation base of any reference date at each control time. Based on the time interval between the reference time corresponding to the minimum difference between the target time and any reference day, the time difference coefficient between the base day and the arbitrary reference day is determined. Based on the balance of the product between the data difference coefficients and time difference coefficients between the base date and each reference date, the difference component of the target time in the base date is determined.

3. A multifunctional guidance display device for hospitals according to claim 2, characterized in that, The abnormal pattern evaluation value for each time moment is obtained based on the maximum number of days contained in all day clusters at each time moment, and the deviation from the base number for each day in the day cluster corresponding to the maximum value. Specifically, this includes: In the clustering results at the target time, the cluster corresponding to the maximum number of days included is taken as the feature cluster; The proportion of days contained in the feature cluster is used as the first coefficient, the mean of the deviation of all days in the feature cluster from the base at the target time is used as the second coefficient, and the negative correlation coefficient of the mean of the difference components of all days in the feature cluster at the target time is used as the third coefficient. The product of the first, second, and third coefficients is normalized to obtain the abnormal pattern evaluation value at the target time.

4. A multifunctional guidance display device for hospitals according to claim 3, characterized in that, The process of filtering all time points based on the abnormal pattern evaluation value at each time point to obtain time nodes specifically includes: The time point corresponding to the abnormal pattern evaluation value being greater than or equal to the preset pattern threshold is taken as the time node.

5. A multifunctional guidance display device for hospitals according to claim 1, characterized in that, The method of obtaining the mutation factor at a selected time node each day based on the difference in the slope between the first and second fitted lines each day, and the deviation base of the first and second fitted lines at corresponding times, specifically includes: For any given day, the normalized value of the angle between the first fitted line and the second fitted line is taken as the first feature difference. The absolute value of the difference between the mean deviation from the baseline at each time corresponding to the first fitted line and the mean deviation from the baseline at each time corresponding to the second fitted line is used as the second feature difference. The mutation factor at a selected time point on any given day is obtained by calculating the product of the first feature difference and the second feature difference.

6. A multifunctional guidance display device for hospitals according to claim 5, characterized in that, The process of filtering all time points based on their mutation coefficients to obtain behavioral pattern time periods specifically includes: The average mutation coefficient of all time nodes within the time neighborhood of each time node is used as the start-up evaluation value of each time node. The time neighborhood of the time node whose start-up evaluation value is greater than or equal to the preset start-up threshold is used as the behavioral pattern time period.

7. A multifunctional guidance display device for hospitals according to claim 1, characterized in that, The method monitors and displays abnormalities in physiological parameters based on their deviation from the baseline and behavioral patterns at each time of day, specifically including: For any type of physiological parameter, when the deviation from the baseline at any time of day is greater than or equal to the preset abnormal threshold, and that time does not belong to the period of behavioral regularity, the abnormal physiological parameter will be displayed at that time.

8. A multifunctional guidance display device for hospitals according to claim 1, characterized in that, The deviation baseline for each type of physiological parameter at each time of day is obtained based on the differences between physiological data and normal data at different times of the day for each type of physiological parameter, as well as the differences between physiological data at each time and adjacent historical times. Specifically, this includes: For any type of physiological parameter, the difference between the physiological data at each time of day and the corresponding normal data is taken as the first difference coefficient, and the ratio between the difference between the physiological data at each time of day and the physiological data at the previous time and the time interval is taken as the second difference coefficient; the normalized result of the product between the first difference coefficient and the second difference coefficient is calculated to obtain the deviation base of the arbitrary type of physiological parameter at each time of day.

Images