Medical room health assessment and intervention decision system based on health data analysis

Abstract

This invention relates to the field of data analysis technology, specifically to a health assessment and intervention decision-making system for medical clinics based on health data analysis. The system includes a signal acquisition module, a data processing module, a deviation identification module, a decision triggering module, and an assessment archiving module. In this invention, multi-source sensor signals such as heart rate, blood pressure, and body temperature are continuously acquired and cached. Abnormal data is filtered out by setting abnormal intervals. Effective data is averaged and smoothed to construct time-series features of individual health parameters. Consistency comparisons of multiple trend directions are performed to identify unstable fluctuation segments. Key segments potentially indicating risk are extracted using time window standards. These segments are matched with preset intervention standards to complete intervention behavior delivery and feedback archiving. This makes health assessments more timely and trend-aware, and intervention strategies more targeted and individualized, significantly improving the intelligence level and practical application value of health assessments in medical clinics.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a health assessment and intervention decision-making system for medical clinics based on health data analysis. Background Technology

[0002] Data analytics technology involves the collection, cleaning, modeling, analysis, and mining of massive, multidimensional, and dynamic data to uncover potential patterns, trends, and value. It is widely applied in various industries such as healthcare, finance, industry, and transportation, including structured data processing, statistical analysis, visualization, intelligent prediction, and decision support. Particularly in the healthcare field, data analytics technology can be used for continuous monitoring of individual health status, health risk prediction, and intervention recommendation formulation, serving as a crucial supporting technology for achieving precision health management. Among these, a health assessment and intervention decision-making system in a medical clinic refers to a system set up in health service locations such as schools, enterprises, or communities to conduct preliminary assessments of the health status of visiting individuals and provide intervention recommendations. This typically involves collecting an individual's vital signs or basic health data using basic equipment such as thermometers, blood pressure monitors, and blood glucose meters. Medical personnel then manually analyze the data based on paper records or computerized registration forms and provide individual health recommendations based on past experience or basic health knowledge.

[0003] In current health assessment processes at medical clinics, the process largely relies on basic equipment to collect vital signs such as body temperature and blood pressure, which are then manually analyzed by medical staff. This approach has significant shortcomings in terms of the continuity and precision of data processing. Manual judgment based on experience is easily influenced by subjectivity and makes it difficult to capture the dynamic changes in an individual's health status in a timely manner, resulting in delayed and inaccurate assessment results. At the same time, the data is usually in the form of paper records or static registration forms, lacking systematic integration of health information across multiple time periods and cycles. This is not conducive to identifying potential health risk signals, limiting the pertinence and timeliness of intervention recommendations, and consequently affecting the effectiveness of health management and decision support capabilities. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a health assessment and intervention decision-making system for medical clinics based on health data analysis. The technical solution is as follows: On the one hand, a health assessment and intervention decision-making system for medical clinics based on health data analysis is provided. This system includes: The signal acquisition module acquires heart rate, blood pressure and body temperature sensor signals from the user's mobile terminal, monitors the continuity of the data signals and marks them with timestamps before caching them to obtain physiological data monitoring records; The data processing module performs screening based on the physiological data monitoring records, sets abnormal heart rate intervals, abnormal blood pressure intervals, and abnormal body temperature intervals, performs average smoothing on the remaining data sequences, constructs the current user's continuous feature trend, and generates a health performance trajectory. The deviation identification module determines the fluctuation segments by judging the consistency of parameter direction based on the health performance trajectory, compares the start and end time periods of all fluctuation segments with the set reference standard time window, extracts the performance segments that are greater than the position limit of the set reference standard time window, and generates a risk tendency indicator. The decision triggering module compares the performance paragraphs in the risk tendency identifier with the defined intervention template standards, selects the corresponding intervention behavior content items, forms a task queue and synchronizes it to the user's mobile terminal, and generates a health intervention prompt document. The assessment and archiving module collects user feedback records and behavior performance based on the health intervention prompts, categorizes and tags the intervention content, and generates a health assessment and intervention decision file for the clinic.

[0005] As a further aspect of the present invention, in the process of constructing the current user's continuous feature trend, the data point sequence after processing each indicator is read, and dynamic arrangement blocks of heart rate, blood pressure, and body temperature are formed in chronological order. The direction of change is calculated by selecting three adjacent sets of data in each block.

[0006] As a further aspect of the present invention, the parameter direction consistency judgment is performed by extracting three consecutive change sequences of heart rate, blood pressure, and body temperature, calculating the direction of difference between the start and end values ​​of each sequence, marking it as rising or falling, and matching adjacent two direction labels in turn. If the direction labels of the three sequences are consistent, it is determined that the direction is stable; if inconsistent labels appear, it is determined that the direction fluctuates.

[0007] As a further aspect of the present invention, the physiological data monitoring record includes digitized heart rate signal sequences, blood pressure signal sequences, body temperature signal sequences and corresponding timestamps; the health performance trajectory includes smoothed heart rate trend sequences, blood pressure trend sequences, and body temperature trend sequences; the risk tendency identifier includes directional fluctuation markers, fluctuation segment time intervals, and over-limit performance segment time markers; the health intervention prompt document includes intervention behavior content items, task queue instructions, and user terminal synchronization information; and the medical room health assessment intervention decision file specifically includes intervention behavior performance records, user feedback data, and intervention label classification information.

[0008] As a further aspect of the present invention, the signal acquisition module includes: The physiological signal receiving submodule acquires heart rate, blood pressure and body temperature sensor signals output by the mobile terminals of registered users in the medical room, divides the signal channels according to the sensor type, judges and removes abnormal waveforms based on the signal amplitude, caches and classifies the remaining signals, and generates the original physiological signal cache set. The digital signal conversion submodule samples the analog signals from each channel in the original physiological signal buffer set according to the set sampling frequency and quantization precision, obtains continuous digital signals, adds timestamps and arranges them in a unified order, and generates a digital signal sequence with timestamps. The monitoring record generation submodule determines the signal continuity based on the timestamp order of various signal segments in the timestamped digital signal sequence, filters signal segments that meet the set time interval conditions, numbers and caches them, and generates physiological data monitoring records.

[0009] As a further aspect of the present invention, the data processing module includes: The signal sorting submodule acquires the heart rate, blood pressure and body temperature signal sequences from the physiological data monitoring records, sorts the signal sequences in ascending order according to the recording timestamp field, reads the adjacent data segment numbers in the recording order and performs a number consistency check, removes records with missing timestamps and archives and caches the remaining signal values ​​to generate a time series signal mapping table. The anomaly removal submodule sets abnormal intervals for heart rate, blood pressure and body temperature based on the heart rate, blood pressure and body temperature values ​​recorded in the time series signal mapping table. It then performs a screening operation on the signal values ​​that fall into the abnormal intervals and performs a three-point moving average processing on the remaining data segment after screening to generate a smoothed signal sequence. The trend generation submodule extracts three adjacent consecutive data sequences based on heart rate, blood pressure and body temperature for each processed data segment in the smoothed signal sequence and arranges them into time-series blocks. Based on the three data segments in each group, it calculates the change difference sign set and then counts the proportion of the same sign to determine the direction of the change trend. It integrates the change trends of various indicators to generate a health performance trajectory.

[0010] As a further aspect of the present invention, the deviation identification module includes: The trend direction recognition submodule extracts three continuous change sequences for each indicator in chronological order based on the continuous trend of the central rate, blood pressure and body temperature of the health performance trajectory. It calculates the difference between the start value and the end value of each change sequence. If the difference is greater than zero, it is marked as rising; if the difference is less than zero, it is marked as falling. It reads two adjacent direction labels in sequence for matching. If the three labels are consistent, it is set as stable direction; otherwise, it is set as fluctuating direction, generating a trend change consistency label sequence. The fluctuation segment extraction submodule extracts all segments marked as directional fluctuations in the trend change consistency marker sequence, extracts the start and end timestamps in the health performance trajectory according to the corresponding signal type and number, constructs a fluctuation segment time interval index sequence, compares the time span of each fluctuation segment with the set standard time window, filters segments whose duration exceeds the standard time window, and generates an overtime fluctuation time interval set. The risk label generation submodule extracts the start and end times and signal type from each record in the timeout fluctuation time interval set, and constructs a feature matrix by superimposing the signal fluctuation frequency and duration parameters. Based on whether the duration of continuous fluctuation of a single channel in the feature matrix exceeds the standard threshold, it makes a logical judgment, obtains the trend segment and corresponding signal label that meet the conditions, and generates a risk tendency label.

[0011] As a further aspect of the present invention, the decision triggering module includes: The intervention matching submodule obtains the segment in the risk tendency identifier, reads the signal type, abnormal characteristics, duration and timestamp information, matches each field with the defined intervention template standard, determines the intervention level to which the signal belongs and selects the corresponding template content item, adds the successfully matched record to the control queue and generates an intervention behavior task item set; The task synchronization submodule maps the pending task content in the intervention behavior task item set to the medical room terminal platform according to the task number, constructs a terminal task stack and sorts it according to time priority, forwards the task content to the corresponding user mobile terminal, and generates a synchronous push task structure. The notification form generation submodule fills in the intervention type, intervention measures, execution suggestions and task issuance time based on the task number, intervention behavior name, execution time and corresponding user identifier in the synchronous push task structure, and merges them into the corresponding node under the user push history index to generate a health intervention notification form.

[0012] As a further aspect of the present invention, the evaluation archiving module includes: The behavior record collection submodule obtains the intervention behavior content items recorded in the health intervention prompt document, reads the task number, intervention measures, issuance time and user identifier, collects feedback content and behavior performance status, and generates a user intervention response record set; The performance status labeling submodule reads each task record in the user intervention response record set, performs Boolean labeling on the performance status, binds the completed label if the status is executed, binds the incomplete label if the status is empty or exceeds the execution period, classifies and groups the intervention content field, performs a one-to-one correspondence operation between intervention content and label, and generates an intervention performance label mapping table. The archive generation submodule writes all information into a structured storage unit based on the task number, user ID, tag field, and intervention category recorded in the intervention performance tag mapping table, constructs intervention records at the user level, completes archiving, and generates a health assessment intervention decision archive for the medical room.

[0013] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following: By continuously acquiring and caching multi-source sensor signals such as heart rate, blood pressure, and body temperature, and combining this with a set abnormal interval to screen out abnormal data, the effective data is averaged and smoothed to construct time series features of individual health parameters. Multiple trend direction consistency comparisons are performed to identify unstable fluctuation segments, and key segments that may indicate risk are extracted by combining time window standards. By matching these segments with preset intervention standards, intervention behavior is pushed out and feedback is archived. This makes health assessment more timely and trend-aware, and intervention strategies more targeted and individualized, significantly improving the intelligence level and practical application value of health assessment in medical clinics. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0015] Figure 1 This is a schematic diagram of a health assessment and intervention decision-making system for a medical clinic based on health data analysis, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the signal acquisition module of the present invention; Figure 4 This is a flowchart of the data processing module of the present invention; Figure 5 This is a flowchart of the deviation identification module of the present invention; Figure 6 This is a flowchart of the decision triggering module of the present invention; Figure 7 This is a flowchart of the evaluation and archiving module of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0017] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0018] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0019] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0020] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0021] This invention provides a health assessment and intervention decision-making system for medical clinics based on health data analysis, such as... Figure 1-2 The diagram shown illustrates a health assessment and intervention decision-making system for a medical clinic based on health data analysis. The system includes: The signal acquisition module acquires the heart rate sensor signal, blood pressure sensor signal, and body temperature sensor signal output from the mobile terminal of the registered user in the medical room, calls the analog-to-digital conversion module to convert the signal into a digital signal, monitors the continuity of the data signal, marks the timestamps on each signal, and then caches them to obtain physiological data monitoring records. The data processing module reads and sorts the previous and next records based on the signal sequence in the physiological data monitoring records. It performs screening operations by setting abnormal heart rate, abnormal blood pressure, and abnormal body temperature intervals. After removing abnormal data, it performs average smoothing on the remaining data sequence and integrates all parameter records to construct the continuous feature trend of the current user (reading the data point sequence after processing each indicator, forming dynamic arrangement blocks of heart rate, blood pressure, and body temperature in chronological order, and calculating the direction of change of the three adjacent sets of data in each block respectively), and generates a health performance trajectory. The deviation identification module judges the consistency of the direction of the continuous characteristics of heart rate, blood pressure and body temperature in the health performance trajectory with the same performance change trend in the last three cycles (extracting three continuous change sequences of heart rate, blood pressure and body temperature, calculating the direction of difference between the start value and the end value of each sequence, marking it as rising or falling, and matching adjacent two direction labels in turn. If the direction labels of the three sequences are consistent, it is determined that the direction is stable. If the labels are inconsistent, it is determined that the direction is fluctuating). It identifies the fluctuation segments, compares the start and end time periods of all fluctuation segments with the set reference standard time window, extracts the performance segments that are larger than the position limit of the set reference standard time window, and generates risk tendency indicators. The decision triggering module compares the performance segments in the risk tendency identifier with the defined intervention template standards, selects the corresponding intervention behavior content items, forms a task queue through the medical room terminal platform and synchronizes it to the user's mobile terminal, and generates a health intervention prompt document. The assessment and archiving module collects user feedback records and behavior performance information based on the intervention behavior content items recorded in the health intervention prompt form, performs intervention record annotation on the clinic terminal platform, completes the classification and tag binding of intervention content, and forms archived records to generate a clinic health assessment intervention decision file.

[0022] Physiological data monitoring records include digitized heart rate signal sequences, blood pressure signal sequences, body temperature signal sequences, and corresponding timestamps. Health performance trajectories include smoothed heart rate trend sequences, blood pressure trend sequences, and body temperature trend sequences. Risk tendency indicators include directional fluctuation markers, fluctuation segment time intervals, and time markers for exceeding performance limits. Health intervention prompts include intervention behavior content items, task queue instructions, and user terminal synchronization information. The health assessment intervention decision-making file in the medical office specifically includes intervention behavior performance records, user feedback data, and intervention label classification information.

[0023] Specifically, such as Figure 2 , 3 As shown, the signal acquisition module includes: The physiological signal receiving submodule acquires heart rate, blood pressure and body temperature sensor signals output by the mobile terminals of registered users in the medical room, divides the signal channels according to the sensor type, judges and removes abnormal waveforms based on the signal amplitude, caches and classifies the remaining signals, and generates the original physiological signal cache set. When acquiring heart rate, blood pressure, and body temperature sensor signals output from the mobile terminals of registered users in the medical clinic, it is first necessary to establish a mapping relationship between signal types and channel numbers for subsequent retrieval and caching. Typically, heart rate, blood pressure, and body temperature signals are mapped to channels numbered CH01, CH02, and CH03, respectively. After each type of signal is received, amplitude verification is performed based on the sensor's output voltage range. For example, the normal amplitude range for a heart rate signal is 0.3V to 3.0V. If a signal segment with an amplitude below 0.3V or above 3.0V is detected, it is marked as an interference signal and excluded. If a sampled heart rate waveform sequence is [0... [0.25V, 0.28V, 0.32V, 2.8V], the first two will be discarded, leaving only 0.32V and 2.8V as valid signal segments. Furthermore, each signal type corresponds to a different reference frequency band; the blood pressure signal frequency should be between 0.3 and 1.5 Hz. When the sampling frequency of a blood pressure signal segment, after Fast Fourier Transform, yields a dominant frequency of 0.2 Hz, it is determined to be an abnormal segment and removed from the buffer. After eliminating interference signals through amplitude-frequency joint filtering, the remaining valid signals are paired with channel numbers and buffered. The buffer structure uses a two-level array, with each array unit recording the channel number, initial timestamp value, and voltage value. An example is shown in Table 1 below. Table 1 Physiological signal channel parameter settings As shown in Table 1, different signal types are set with abnormal judgment thresholds based on amplitude and frequency range. Real-time comparison and screening processes are performed in the hardware sampling module. After channel number matching and cache management, the original physiological signal cache set is finally obtained.

[0024] The digital signal conversion submodule performs sampling processing on analog signals from each channel in the original physiological signal buffer according to the set sampling frequency and quantization precision. After obtaining continuous digital signals, it adds timestamps and arranges them in a unified order to generate a digital signal sequence with timestamps. Each type of signal in the original physiological signal buffer is sent to the analog-to-digital conversion module. The sampling frequency and quantization precision parameters need to be set separately. During the conversion process, each analog signal sampling point is digitally encoded. Taking the heart rate signal as an example, its sampling frequency is 250Hz, collecting 250 sample points per second. With 12-bit quantization precision, the analog voltage value from 0 to 3.3V is divided into 4096 levels, with each level having a step size of 3.3 ÷ 4096 ≈ 0.000805V. If a sampling point is 1.2V, the corresponding digital value is 1.2 ÷ 0.000805 ≈ 1490. Similarly, the blood pressure signal sampling frequency is 200Hz, and the body temperature signal is 100Hz, using 12-bit and 10-bit precision respectively. The converted values ​​can be calculated according to the corresponding resolution. After conversion, to facilitate data synchronization, each digital signal needs to be timestamped in a uniform format. The timestamp interval is set to 4ms for heart rate, 5ms for blood pressure, and 10ms for body temperature, as detailed in Table 2. Table 2 Analog-to-Digital Conversion Sampling and Quantization Parameters As shown in Table 2, each signal type is matched with the corresponding sampling and precision settings according to its physiological characteristics. After the timestamps of all data points are generated, they are merged into a unified data structure according to the sampling order. Each record contains a digital value, signal type, timestamp and channel number. After merging, the output is a digital signal sequence with timestamps.

[0025] The monitoring record generation submodule is based on various signal segments in the time-stamped digital signal sequence. It judges the signal continuity according to the time stamp order, filters signal segments that meet the set time interval conditions, numbers and caches them, and generates physiological data monitoring records. After obtaining the timestamped digital signal sequence, the system needs to filter its temporal continuity to determine whether it constitutes a continuous and valid monitoring segment. The continuity judgment is based on the difference between the timestamps. If the time interval between any two adjacent data points is greater than a preset threshold, it is considered an interrupted segment and is removed. For example, the maximum time interval for heart rate signals is set to 1000ms. If the timestamp of data point A is 10000ms and that of B is 11200ms, the interval is 1200ms, which exceeds the threshold, and the entire segment is invalid. The threshold for body temperature signals is set to 5000ms. The specific parameter settings are listed in Table 3 below. Table 3 Continuity Determination Time Parameter Table As shown in Table 3, the system judges the time interval of various signal data segments according to the threshold. Those that meet the conditions are marked according to the numbering rules. For example, continuous heart rate data within 60 seconds, such as HR_SEQ_1 and HR_SEQ_2, etc. Each record contains a number, start and end time, channel number and corresponding numerical sequence, which are finally combined to form a physiological data monitoring record.

[0026] Specifically, such as Figure 2 , 4 As shown, the data processing module includes: The signal sorting submodule acquires the heart rate, blood pressure and body temperature signal sequences from physiological data monitoring records, sorts the signal sequences in ascending order according to the recording timestamp field, reads the adjacent data segment numbers in the recording order and performs a numbering consistency check, removes records with missing timestamps and archives and caches the remaining signal values, and generates a time series signal mapping table. When acquiring heart rate, blood pressure, and body temperature signal sequences from physiological data monitoring records, the data is first classified according to the signal type identified in the record fields. Heart rate, blood pressure, and body temperature record sets are then extracted based on the signal type. After extraction, the timestamp field of each record in each set is read. A sequential linked list structure is established by sorting the timestamp values ​​in ascending order. After sorting in ascending order, the consistency of the timestamp differences between adjacent records is checked. If an abnormal change occurs in the timestamp difference, the record is marked as invalid and removed. For example, if the current record's timestamp is 10020 milliseconds and the previous one is 10000 milliseconds, the difference is 20 milliseconds, which is within the stable range. If the difference is greater than 1000 milliseconds, it is considered a data interruption segment and removed. After clearing abnormal records, the records are re-archived according to the signal number field to form a multi-signal cache table arranged chronologically. Each row in this cache structure includes signal type, channel number, timestamp, and numerical fields, with the timestamp as the primary key index. This cache table enables sequential backtracking and retrieval based on the time axis, ultimately generating a time series signal mapping table.

[0027] The anomaly removal submodule is based on the heart rate, blood pressure and body temperature values ​​recorded in the time series signal mapping table. It sets abnormal intervals for heart rate, blood pressure and body temperature respectively, and performs a screening operation on the signal values ​​that fall into the abnormal intervals. Then, it performs three-point moving average processing on the remaining data segment after screening to generate a smoothed signal sequence. After extracting the signal values ​​recorded in the time series signal mapping table, an outlier screening operation needs to be performed to remove outliers. For heart rate signals, thresholds are set to less than 40 beats per minute or greater than 180 beats per minute; for blood pressure signals, systolic blood pressure below 70 mmHg or above 180 mmHg is considered abnormal; and for body temperature signals, below 35 degrees Celsius or above 40 degrees Celsius are considered invalid records. For example, if the heart rate sequence is 85, 190, 78, 42, then 190 and 42 are both outside the normal threshold range and are removed, only 85 and 78 are retained. Similarly, if the blood pressure values ​​are 125, 185, 110, 65, then 185 and 65 are not considered abnormal. Within a reasonable range, values ​​of 125 and 110 are retained. If the body temperature signals are 36.5, 41.2, 37.1, and 34.8, then 41.2 and 34.8 are removed, leaving two values. After filtering, an average smoothing operation is performed on the remaining signals. A weighted average is calculated using a sliding window of three adjacent data points. If the current sliding window contains values ​​of 78, 85, and 88, the moving average result is (78+85+88) / 3=83.7. A non-overlapping sliding window is used for traversal, with a fixed window step size of 1. The average value is output as the new sequence point. Segments with fewer than three points within the sliding window are not included in the calculation sequence. Example processing results are shown in Table 4 below. Table 4. Example data table for outlier data screening and moving average. As shown in Table 4, the result of the moving average processing is based on the effective signal after elimination, and the output is a smoothed signal sequence after traversal and smoothing calculation.

[0028] The trend generation submodule extracts three adjacent continuous data segments according to heart rate, blood pressure and body temperature for each processed data segment in the smoothed signal sequence and arranges them into time-series blocks. Based on the three segments of each group, it calculates the change difference sign set and then counts the proportion of the same sign to determine the direction of the change trend. It integrates the change trends of various indicators and generates a health performance trajectory. Based on the signal point data of each type in the smoothed signal sequence, equal-length blocks are constructed in chronological order, with each block consisting of three segments. The first, middle, and last three consecutive signal sequences in the current signal block are read and named G1, G2, and G3 respectively. The mean of each group is recorded, and the sequence difference is determined. If G1, G2, and G3 correspond to heart rates of 78, 82, and 88 respectively, then the differences Δ1 = 82 - 78 = 4 and Δ2 = 88 - 82 = 6. Both differences are positive, and the trend is upward. If the differences are 78, 82, and 76... If Δ1 is positive and Δ2 is negative, it is considered a fluctuation. The system uses this to determine the type of symbol combination of adjacent changes, classifying them into three trends: rising, falling, or fluctuating. It further counts the number of each trend within a unit time period to obtain a change percentage matrix for each type of signal indicator. The matrix is ​​arranged by signal type as rows and change trend as columns, with corresponding quantity values ​​filled into the cells. Finally, the trend change states of the three types of signal indicators are integrated, and the block structure is merged according to time labels. The trend path in the continuous sequence is output and mapped to a time axis graphic sequence, ultimately outputting the health performance trajectory.

[0029] Specifically, such as Figure 2 , 5 As shown, the deviation recognition module includes: The trend direction recognition submodule is based on the continuous trend of the central rate, blood pressure and body temperature of the health performance trajectory. It extracts three continuous change sequences of each indicator in time order, calculates the difference between the start value and the end value of each change sequence, and marks it as rising if the difference is greater than zero and falling if the difference is less than zero. It reads two adjacent direction labels in turn for matching. If the three labels are consistent, it is set as stable direction; otherwise, it is set as fluctuating direction, and a trend change consistency label sequence is generated. Based on the continuous trend characteristics of heart rate, blood pressure, and body temperature in the health performance trajectory, each indicator signal sequence needs to be divided according to the sampling time, and three adjacent independent subsequences need to be extracted. Each subsequence is labeled with its sequence number and start and end timestamps. Then, the first and last data points of each subsequence are extracted, and their difference is calculated to determine the direction of change. If the end value is greater than the start value, the trend of that subsequence is marked as rising; otherwise, it is marked as falling. In the example, the start value of HR_SEQ_01 is 78, the end value is 85, the difference is 7, and the direction is marked as rising. Next, the direction markings of the three subsequences under the same signal type are judged for consistency. That is, the directions of HR_SEQ_01, HR_SEQ_02, and HR_SEQ_03 are compared sequentially to see if they are all the same. If all three subsequences are rising, the direction is consistent and marked as stable. If any set of directions is inconsistent, such as BP_SEQ_01 being falling and BP_SEQ_02 being rising, it is marked as directional fluctuation. Directional consistency is confirmed by labeling in units of three subsequences. The processing results are shown in the table below: Table 5 Example Data Table for Trend Direction Identification As shown in Table 5, the trend direction can be quickly classified and labeled as stable or fluctuating by comparing the first and last values. By batch processing the trend direction, the final output is a sequence of trend change consistency markers.

[0030] The fluctuation segment extraction submodule extracts all segments marked as directional fluctuations in the trend change consistency marker sequence, extracts the start and end timestamps in the health performance trajectory according to the corresponding signal type and number, constructs a fluctuation segment time interval index sequence, compares the time span of each fluctuation segment with the set standard time window, filters segments whose duration exceeds the standard time window, and generates an overtime fluctuation time interval set. Based on the various segments marked as directional fluctuations in the trend change consistency marker sequence, the start and end timestamps corresponding to these segments in the health performance trajectory are first read to construct a time interval structure with a unified format. This structure contains four fields: sequence number, signal type, start time, and end time. For example, if the start time of the BP_SEQ_01 fluctuation segment is 12:00 and the end time is 12:09:45, then the duration of this segment is 9 minutes and 45 seconds. Subsequently, each fluctuation segment time interval is compared item by item with a set reference standard time window. In this embodiment, the time interval is set as follows: The standard time window is 10 minutes. If the duration of a segment is no more than 600 seconds, the segment does not meet the set window length and is not included in the subsequent identification process. However, if the duration exceeds 600 seconds, the segment is selected and added to the candidate set. For example, 12:00–12:10:20 lasts for 620 seconds, which meets the timeout condition and is added to the timeout interval sequence. After batch processing all directional fluctuation segments, segments with a time span greater than the standard time window are output separately according to the signal type. The structure fields are unified to facilitate synchronization and matching with subsequent signal identifiers, and finally, the timeout fluctuation time interval set is obtained.

[0031] The risk label generation submodule extracts the start and end time and signal type from each record in the timeout fluctuation interval set, and constructs a feature matrix by superimposing the signal fluctuation frequency and duration parameters. It makes a logical judgment based on whether the duration of continuous fluctuation of a single channel in the feature matrix exceeds the standard threshold, obtains the trend segment and corresponding signal label that meet the conditions, and generates a risk tendency label. For each record in the time interval set of time-out fluctuations, it is necessary to match the trend segment number of its corresponding signal in the health performance trajectory, and retrieve auxiliary parameters such as trend direction and change frequency of the corresponding position segment in its time series to construct a three-dimensional structured data matrix. The structural dimensions are signal type, trend stability, and duration. The duration field of each record is read and it is determined whether it exceeds the set threshold value. In this embodiment, the single-channel risk threshold is set to 60 minutes. That is, if the continuous fluctuation state of a certain signal lasts for more than 3600 seconds, the risk identification mechanism is triggered. For example, if the body temperature signal TP_SEQ_07 fluctuates continuously from 09:00 to 10:15 for a total of 75 minutes, it is determined to be an abnormal trend segment. Then, its number, start and end time and signal category fields are combined to construct a unified identification content, and a preset risk label such as RISK_FLAG_01 is assigned to it. The number result and label sequence are uniformly output in all segments that meet the conditions, and finally the risk tendency identification is obtained.

[0032] Specifically, such as Figure 2 , 6 As shown, the decision triggering module includes: The intervention matching submodule obtains the segment in the risk tendency identifier, reads the signal type, abnormal characteristics, duration and timestamp information, matches each field with the defined intervention template standard, determines the intervention level to which the signal belongs and selects the corresponding template content item, adds the successfully matched record to the control queue and generates an intervention behavior task item set; After obtaining the performance information from the risk propensity indicators, it is necessary to read the risk indicator number, signal type, abnormal performance description, and start and end time interval for each indicator. This information is then input into the standard structure of the intervention template as the matching basis in the task allocation process. Intervention content is matched based on signal type and abnormal pattern through field comparison. The intervention template is classified into levels according to multiple parameters such as abnormal characteristics and duration. For example, if the template specifies that a continuous upward trend lasts for more than 60 minutes, it belongs to Level II intervention. If the heart rate signal RISK001 meets this condition, the corresponding intervention measure "rest + reminder to retest" is selected. If the blood pressure signal fluctuates by more than 20 mmHg, Level I intervention is triggered. A body temperature exceeding 39.5°C for more than 10 minutes is classified as Level III intervention. The matching process will automatically select the most suitable intervention item according to the set priority, as shown in the table below: Table 6 Example Data Table of Intervention Behavior Matching As shown in Table 6, each risk manifestation is matched one-to-one with the corresponding intervention template field. After a successful match, the intervention measure number, content item and risk paragraph in the template are bound and written into the control queue, ultimately forming a set of intervention behavior task items.

[0033] The task synchronization submodule is based on the task content to be processed in the intervention behavior task item collection. It maps the task to the medical room terminal platform according to the task number, builds the terminal task stack and sorts it according to time priority. It forwards the task content to the corresponding user's mobile terminal and generates a synchronous push task structure. When processing task records based on intervention behavior task items, the task execution level needs to be filtered and sorted according to the task number order. First, the intervention measures, signal types, and user identification information in the task records are parsed. Then, the task scheduling structure table of the medical room terminal platform is read, and the access port is bound with the user identification as the retrieval field. For example, if the user number associated with task number RISK002 is USER005, then its instructions are bound to the USER005 terminal port, and a task push stack is generated. During the stack generation process, timestamp index and task execution priority fields are set. For example, the priority of level I intervention tasks is set to 1, and level III is set to 3. After being sorted in ascending order of priority, they are written into the synchronization index table. After the index table is established, the terminal synchronization module is started to push the task, and the task structure is packaged and sent to the mobile terminal cache of USER005. The structure contains task number, push time, intervention description, and prompt identification fields. After the task synchronization is completed, the synchronized push task structure is obtained.

[0034] The notification form generation submodule fills in the intervention type, intervention measures, execution suggestions and task issuance time based on the task number, intervention behavior name, execution time and corresponding user identifier in the synchronous push task structure, and merges them into the corresponding node under the user push history index to generate a health intervention notification form; Based on the task number and intervention behavior content in the synchronous push task structure, the template field structure is first read, and the health intervention prompt form content is generated according to the standard template format. The prompt form structure is filled with five content fields in sequence: user identifier, signal category, intervention level, suggested behavior, and intervention time. At the same time, a prompt form number and a unique hash index are set to identify repeated issuance records under the same intervention scenario. The generated prompt form is then stored in the medical room terminal platform database. After the prompt form is generated, the data is synchronously pushed to the user terminal interaction interface. The prompt fields will highlight the intervention instruction that needs to be executed, such as "Please drink water and lie still for ten minutes and then retest". A feedback confirmation field is attached at the end of the prompt form to record the user's confirmation status. After all fields are filled and archived, the final health intervention prompt form is formed.

[0035] Specifically, such as Figure 2 , 7 As shown, the evaluation archiving module includes: The behavior record collection submodule obtains the intervention behavior content items recorded in the health intervention prompt document, reads the task number, intervention measures, issuance time and user ID, collects feedback content and behavior performance status, and generates a user intervention response record set. After obtaining the intervention behavior content items recorded in the health intervention prompt document, the intervention task number, signal type, intervention suggestion, and task issuance time fields must first be extracted. The task number and user identifier field are then combined and indexed. Feedback data is collected from the user's mobile terminal feedback database through this index. Key fields include "Executed," "Execution Time," "Feedback Time," and "Feedback Comments." Taking the RISK001 task as an example, user USER001's feedback time is 2024-11-10-10:15. The execution status is filled in as "Yes," and the execution time is noted as 10:10. A feedback note "Retested as required" is also attached. The system preprocesses the feedback data, checks the correctness of the time field format, and automatically fills empty fields with a "Not Reported" mark. All fields are uniformly written into a structured data table. See Table 7 for an example. Table 7. Example of Intervention Behavior Implementation Feedback Record As shown in Table 7, all records are categorized and stored according to task number and user number. Each piece of feedback data constitutes an independent record unit, forming a user intervention response record set.

[0036] The performance status labeling submodule reads each task record from the user intervention response record set, performs Boolean labeling on the performance status, and binds the completed label if the status is executed, and binds the incomplete label if the status is empty or exceeds the execution deadline. It also classifies and groups the intervention content field, performs a one-to-one correspondence operation between intervention content and label, and generates an intervention performance label mapping table. After obtaining the user intervention response record set, the status of the "executed" field of each record needs to be judged. If the value of this field is "yes", it is marked as "completed"; otherwise, it is marked as "incomplete". Then, the time difference between the task execution time and the task issuance time is calculated. If the feedback time exceeds 24 hours from the task issuance time and is still "no", the tag "overdue" is automatically added. For example, if RISK002 did not fill in the execution time and the feedback status is "no", the system judges it as "incomplete" and adds the "overdue" tag. Then, the intervention behavior content field of all completed records is read, and a mapping table between content and tags is established. The tag type field is an enumeration type, and the set value range includes "completed", "incomplete", "overdue", and "partially completed". The mapping structure uses a dictionary to bind the task number and the tag set to support subsequent archiving and classification operations. Finally, an intervention fulfillment tag mapping table is generated.

[0037] The archive generation submodule writes all information into a structured storage unit based on the task number, user ID, tag field, and intervention category recorded in the intervention performance tag mapping table, constructs intervention records at the user level, completes archiving, and generates a health assessment intervention decision archive for the medical office. Based on the user ID, intervention behavior classification field, and performance status label information recorded in the intervention performance label mapping table, the system reads the file management interface under the medical room terminal platform. It retrieves the corresponding intervention prompt document and task synchronization record through the task ID field. During the structure population process, the user ID and task ID are used as primary key indexes to construct a unified data archiving structure unit. Each unit contains fields including: intervention type, performance status, execution time, feedback content, task time, and label status. After being sorted and merged according to the intervention task time, these are embedded into the personal intervention file node under the user ID. The index fields and compressed storage format are set in the archiving template to form a continuous writing structure. After all fields are written, the data archiving submission operation is completed. The archiving operation completion flag is written to the log cache, ultimately generating the medical room health assessment intervention decision file.

[0038] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A health assessment and intervention decision-making system for clinics based on health data analysis, characterized in that, include: The signal acquisition module acquires heart rate, blood pressure and body temperature sensor signals from the user's mobile terminal, monitors the continuity of the data signals and marks them with timestamps before caching them to obtain physiological data monitoring records; The data processing module performs screening based on the physiological data monitoring records, sets abnormal heart rate intervals, abnormal blood pressure intervals, and abnormal body temperature intervals, performs average smoothing on the remaining data sequences, constructs the current user's continuous feature trend, and generates a health performance trajectory. The deviation identification module determines the fluctuation segments by judging the consistency of parameter direction based on the health performance trajectory, compares the start and end time periods of all fluctuation segments with the set reference standard time window, extracts the performance segments that are greater than the position limit of the set reference standard time window, and generates a risk tendency indicator. The decision triggering module compares the performance paragraphs in the risk tendency identifier with the defined intervention template standards, selects the corresponding intervention behavior content items, forms a task queue and synchronizes it to the user's mobile terminal, and generates a health intervention prompt document. The assessment and archiving module collects user feedback records and behavior performance based on the health intervention prompts, categorizes and tags the intervention content, and generates a health assessment and intervention decision file for the clinic.

2. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that: In the process of constructing the current user's continuous feature trend, the data point sequence after processing each indicator is read, and dynamic arrangement blocks of heart rate, blood pressure, and body temperature are formed in chronological order. The direction of change is calculated by selecting three adjacent sets of data in each block.

3. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that: The parameter direction consistency judgment is performed by extracting three consecutive change sequences of heart rate, blood pressure and body temperature, calculating the direction of difference between the start value and the end value of each sequence, and marking it as rising or falling. Adjacent direction labels are matched in turn. If the direction labels of the three sequences are consistent, it is determined that the direction is stable. If the labels are inconsistent, it is determined that the direction fluctuates.

4. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that: The physiological data monitoring records include digitized heart rate signal sequences, blood pressure signal sequences, body temperature signal sequences, and corresponding timestamps. The health performance trajectory includes smoothed heart rate trend sequences, blood pressure trend sequences, and body temperature trend sequences. The risk tendency identifier includes directional fluctuation markers, fluctuation segment time intervals, and over-limit performance segment time markers. The health intervention prompt document includes intervention behavior content items, task queue instructions, and user terminal synchronization information. The medical room health assessment intervention decision file specifically includes intervention behavior performance records, user feedback data, and intervention label classification information.

5. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that, The signal acquisition module includes: The physiological signal receiving submodule acquires heart rate, blood pressure and body temperature sensor signals output by the mobile terminals of registered users in the medical room, divides the signal channels according to the sensor type, judges and removes abnormal waveforms based on the signal amplitude, caches and classifies the remaining signals, and generates the original physiological signal cache set. The digital signal conversion submodule samples the analog signals from each channel in the original physiological signal buffer set according to the set sampling frequency and quantization precision, obtains continuous digital signals, adds timestamps and arranges them in a unified order, and generates a digital signal sequence with timestamps. The monitoring record generation submodule determines the signal continuity based on the timestamp order of various signal segments in the timestamped digital signal sequence, filters signal segments that meet the set time interval conditions, numbers and caches them, and generates physiological data monitoring records.

6. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that, The data processing module includes: The signal sorting submodule acquires the heart rate, blood pressure and body temperature signal sequences from the physiological data monitoring records, sorts the signal sequences in ascending order according to the recording timestamp field, reads the adjacent data segment numbers in the recording order and performs a number consistency check, removes records with missing timestamps and archives and caches the remaining signal values ​​to generate a time series signal mapping table. The anomaly removal submodule sets abnormal intervals for heart rate, blood pressure and body temperature based on the heart rate, blood pressure and body temperature values ​​recorded in the time series signal mapping table. It then performs a screening operation on the signal values ​​that fall into the abnormal intervals and performs a three-point moving average processing on the remaining data segment after screening to generate a smoothed signal sequence. The trend generation submodule extracts three adjacent consecutive data sequences based on heart rate, blood pressure and body temperature for each processed data segment in the smoothed signal sequence and arranges them into time-series blocks. Based on the three data segments in each group, it calculates the change difference sign set and then counts the proportion of the same sign to determine the direction of the change trend. It integrates the change trends of various indicators to generate a health performance trajectory.

7. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that, The deviation identification module includes: The trend direction recognition submodule extracts three continuous change sequences for each indicator in chronological order based on the continuous trend of the central rate, blood pressure and body temperature of the health performance trajectory. It calculates the difference between the start value and the end value of each change sequence. If the difference is greater than zero, it is marked as rising; if the difference is less than zero, it is marked as falling. It reads two adjacent direction labels in sequence for matching. If the three labels are consistent, it is set as stable direction; otherwise, it is set as fluctuating direction, generating a trend change consistency label sequence. The fluctuation segment extraction submodule extracts all segments marked as directional fluctuations in the trend change consistency marker sequence, extracts the start and end timestamps in the health performance trajectory according to the corresponding signal type and number, constructs a fluctuation segment time interval index sequence, compares the time span of each fluctuation segment with the set standard time window, filters segments whose duration exceeds the standard time window, and generates an overtime fluctuation time interval set. The risk label generation submodule extracts the start and end times and signal type from each record in the timeout fluctuation time interval set, and constructs a feature matrix by superimposing the signal fluctuation frequency and duration parameters. Based on whether the duration of continuous fluctuation of a single channel in the feature matrix exceeds the standard threshold, it makes a logical judgment, obtains the trend segment and corresponding signal label that meet the conditions, and generates a risk tendency label.

8. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that, The decision triggering module includes: The intervention matching submodule obtains the segment in the risk tendency identifier, reads the signal type, abnormal characteristics, duration and timestamp information, matches each field with the defined intervention template standard, determines the intervention level to which the signal belongs and selects the corresponding template content item, adds the successfully matched record to the control queue and generates an intervention behavior task item set; The task synchronization submodule maps the pending task content in the intervention behavior task item set to the medical room terminal platform according to the task number, constructs a terminal task stack and sorts it according to time priority, forwards the task content to the corresponding user mobile terminal, and generates a synchronous push task structure. The notification form generation submodule fills in the intervention type, intervention measures, execution suggestions and task issuance time based on the task number, intervention behavior name, execution time and corresponding user identifier in the synchronous push task structure, and merges them into the corresponding node under the user push history index to generate a health intervention notification form.

9. The health assessment and intervention decision-making system for clinics based on health data analysis according to claim 1, characterized in that, The evaluation archiving module includes: The behavior record collection submodule obtains the intervention behavior content items recorded in the health intervention prompt document, reads the task number, intervention measures, issuance time and user identifier, collects feedback content and behavior performance status, and generates a user intervention response record set; The performance status labeling submodule reads each task record in the user intervention response record set, performs Boolean labeling on the performance status, binds the completed label if the status is executed, binds the incomplete label if the status is empty or exceeds the execution period, classifies and groups the intervention content field, performs a one-to-one correspondence operation between intervention content and label, and generates an intervention performance label mapping table. The archive generation submodule writes all information into a structured storage unit based on the task number, user ID, tag field, and intervention category recorded in the intervention performance tag mapping table, constructs intervention records at the user level, completes archiving, and generates a health assessment intervention decision archive for the medical room.

Images