A method for monitoring and evaluating children's growth and development based on medical big data

Abstract

The present invention discloses a child growth and development monitoring and evaluation method based on medical big data, which relates to the field of medical health and child growth and development monitoring technology. The method comprises the following steps: Step 1: Read the child's height data, weight data and corresponding collection time points and diagnosis records from the databases of at least two medical institutions respectively. The child growth and development monitoring and evaluation method based on medical big data, by establishing a fine piecewise linear benchmark growth interval, through quantitative difference coefficient and multiple threshold detection, effectively reduces data redundancy and misjudgment, and utilizes dynamic monitoring intervals and graded intervention mechanisms to achieve early identification and timely intervention of potential abnormalities. Through regular review and adaptive adjustment, the benchmark model and threshold can be continuously optimized, thereby overcoming the shortcomings of incomplete data, rigid monitoring frequency and difficulty in timely capturing abnormalities in traditional methods, so that child growth and development monitoring has accuracy, flexibility and scalability.

Description

Technical Field

[0001] The present invention relates to the technical field of medical health and child growth and development monitoring, and specifically to a child growth and development monitoring and evaluation method based on medical big data. Background Art

[0002] As children's health has received great attention from society and families, it has become a common practice to regularly monitor their height, weight and other physiological indicators to assess whether their growth and development are up to standard. However, in traditional methods, the limitations of data from a single medical institution or region make it difficult to comprehensively and accurately assess the outlier risk of children at different growth stages. In recent years, medical big data technology has gradually emerged, integrating large-scale dynamic data from multiple medical institutions to provide wider coverage and more timely support for children's growth monitoring. How to unify the format and quality screen heterogeneous data from different institutions, and then construct a statistically representative benchmark growth distribution, and compare and detect abnormalities in the continuous growth sequence of individual children has become a technical problem that needs to be urgently solved in the clinical and public health fields. Summary of the invention

[0003] 1. Technical issues to be resolved

[0004] In view of the shortcomings of the existing technology, the present invention provides a method for monitoring and evaluating children's growth and development based on medical big data, which solves the problem of how to unify the format and quality screen the heterogeneous data from different institutions, and then construct a statistically representative benchmark growth distribution to compare and detect abnormalities in the continuous growth sequences of individual children.

[0005] (II) Technical solution

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for monitoring and evaluating children's growth and development based on medical big data, the method comprising the following steps:

[0007] Step 1: Read the child's height data, weight data, and corresponding collection time points and diagnosis records from the databases of at least two medical institutions; match the data according to the child's unique identifier, remove data records whose collection time differs by more than a preset threshold, and remove isolated records that cannot be matched;

[0008] Step 2: Sort the height data and weight data of each child in chronological order, and add a timestamp corresponding to the collection time to each data point. If the height data or weight data corresponding to a timestamp is missing, fill it in using a linear interpolation algorithm; wherein the linear interpolation algorithm calculates interpolation based on the data points of the adjacent timestamps before and after, and the calculation formula is:

[0009] ;

[0010] Among them, X1 and X2 represent the measurement values ​​of the adjacent moments, t1 and t2 represent the corresponding timestamps, t represents the interpolation timestamp, and X interp Represents the interpolation result;

[0011] Step 3: Group the target population by age and gender, calculate the mean and standard deviation of the height and weight data of each group of samples, and obtain μ h , σ h , and μ w , σ w ; After eliminating pathological records, the height benchmark curve and weight benchmark curve of the corresponding age group were obtained using piecewise linear fitting method;

[0012] Step 4: At each monitoring time point t i , let the child's height be H i , weight is recorded as W i , respectively, with the reference value H of the reference curve ref，i With W ref，i Compare and calculate the growth difference coefficient; the calculation formula is as follows:

[0013] ;

[0014] ;

[0015] Among them, σ h and σ w are the standard deviation of height and weight for the corresponding age groups;

[0016] Step 5: D is the growth coefficient of the same child at multiple consecutive time points. h，i , D w，i Perform an overall analysis to determine whether any of the following conditions are met, including: Condition 1: There is at least one time point |D h，i ∣ or ∣ D w，i | is greater than the first threshold value T1, and the difference between adjacent time points is greater than the second threshold value T2; Condition 2: Within the preset monitoring window Δt, three consecutive test results are beyond the confidence interval of the reference curve; If any of the conditions is met, it is determined to be suspected growth abnormality;

[0017] Step 6: When suspected growth abnormalities occur, schedule monitoring cycles based on the abnormality type, including:

[0018] Shorten the length of the follow-up monitoring cycle so that the next collection of samples from the same child is earlier than originally planned;

[0019] When the same child has two consecutive abnormalities, the abnormality information is sent to the parents and designated pediatrician to trigger further offline evaluation;

[0020] Step 7: Write each monitoring result into the audit log, and synchronize the growth difference coefficient and the judgment result to the medical big data platform; after the preset period, recheck the statistical distribution of the growth difference coefficient of all children, and correct the baseline curve and the values ​​of the first threshold T1 and the second threshold T2 to continuously improve the judgment accuracy.

[0021] Preferably, the elimination of pathological records in step 3 includes:

[0022] Data on children diagnosed with endocrine diseases were retrieved from the diagnostic records and excluded from the subsequent calculation of means and standard deviations;

[0023] Data from children who had received high-dose steroid therapy within the last three months were identified and removed from the sample set or stored separately for specialized pathological growth model analyses.

[0024] Preferably, the step 4 further comprises:

[0025] Set a fixed monitoring time interval Δt. When the adjacent difference of the acquisition time exceeds Δt, generate a vacancy record and execute the interpolation algorithm in step 2.

[0026] By comparing the interpolation result with the reference curve, a more complete and continuous sequence of individual growth difference coefficients is obtained.

[0027] Preferably, when performing overall analysis in step 5, the following logical judgment is performed:

[0028] Calculate the change in the coefficient of difference between adjacent time points. The calculation formula is as follows:

[0029] ;

[0030] ;

[0031] When ΔD h，i or ΔD w，i When the pre-defined sudden increase or decrease threshold T2 is exceeded, the time point is marked as abnormal fluctuation;

[0032] If abnormal fluctuations occur at least twice in three consecutive tests, it is directly judged as suspected growth abnormality.

[0033] Preferably, the number of the multiple consecutive time points ranges from 3 to 6. When 3 to 6 monitoring values ​​in the child's growth difference coefficient sequence continuously exceed ±T1, it is judged as suspected growth abnormality and offline review is triggered as a priority.

[0034] Preferably, the initial values ​​of the first threshold T1 and the second threshold T2 are determined respectively in the following manners:

[0035] After counting all samples of the target group, the fluctuation range within the standard deviation of 1.5σ or 2σ is selected as the candidate value, and the comprehensive judgment accuracy is calculated;

[0036] The first threshold T1 and the second threshold T2 are iteratively updated until the difference between the abnormality detection rate and the false positive rate meets a predetermined condition.

[0037] Preferably, the alarm and dynamic intervention in step 6 further includes:

[0038] When suspected growth abnormality occurs, the time interval for the next monitoring is automatically shortened to 50% of the original interval or other preset ratio;

[0039] If the same child is judged to be abnormal within two monitoring intervals, the system will send a reminder request to the pediatric diagnosis and treatment center and push a detailed assessment report to the parents;

[0040] When the same child showed abnormal status in three consecutive monitorings, it triggered a specialist consultation and included the child in the key follow-up file.

[0041] A child growth and development monitoring and evaluation system based on medical big data, the system comprising:

[0042] A data aggregation module for receiving and integrating the height, weight and diagnosis records of children transmitted from at least two medical institutions;

[0043] A time series processing module, used to execute step 2 of the child growth and development monitoring and evaluation method based on medical big data, and perform time series and interpolation calculation on the growth data of each child;

[0044] A benchmark construction module, used to execute step 3 of the child growth and development monitoring and evaluation method based on medical big data, and to construct a benchmark growth interval for the target population;

[0045] A coefficient of difference calculation module, used to execute step 4 of the child growth and development monitoring and evaluation method based on medical big data, and calculate the growth coefficient of difference for the child's individual data;

[0046] An abnormality determination module, used to execute step 5 of the child growth and development monitoring and evaluation method based on medical big data, and identify abnormal trends according to the difference coefficient and a preset threshold;

[0047] The intervention strategy module is used to execute step 6 of the child growth and development monitoring and evaluation method based on medical big data, and to issue an alarm and adjust the subsequent monitoring strategy when it is determined to be suspected abnormal.

[0048] An electronic device comprises at least one processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes any of the above-mentioned methods for monitoring and evaluating children's growth and development based on medical big data.

[0049] Preferably, a computer-readable storage medium stores a computer program, which, when executed by a processor, implements any of the methods for monitoring and evaluating child growth and development based on medical big data.

[0050] (III) Beneficial effects

[0051] The present invention provides a method for monitoring and evaluating children's growth and development based on medical big data. It has the following beneficial effects:

[0052] This child growth and development monitoring and evaluation method based on medical big data has established a refined piecewise linear benchmark growth interval based on the integration and time series processing of multi-source medical data. It has effectively reduced data redundancy and misjudgment by quantifying the difference coefficient and multiple threshold detection, and has achieved early identification and timely intervention of potential abnormalities by using dynamic monitoring intervals and graded intervention mechanisms. Through regular review and adaptive adjustment, the benchmark model and threshold can be continuously optimized, thus overcoming the shortcomings of incomplete data, rigid monitoring frequency, and difficulty in timely capturing abnormalities in traditional methods, making child growth and development monitoring accurate, flexible, and scalable. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of the child growth and development monitoring and evaluation method based on medical big data of the present invention;

[0054] Figure 2 This is a schematic diagram of the framework of a child growth and development monitoring and evaluation system based on medical big data. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] See also Figure 1 and Figure 2 The present invention provides a technical solution: a method for monitoring and evaluating children's growth and development based on medical big data, the method comprising the following steps:

[0057] Step 1: Read the height data, weight data, and corresponding collection time points and diagnosis records of children from the databases of at least two medical institutions; match the data according to the child's unique identifier to ensure the consistency of the data source; pre-set the allowable range of collection time difference. If it exceeds the range, the data will be considered invalid and eliminated to avoid introducing time mismatch errors in subsequent calculations; remove data records whose collection time difference exceeds a preset threshold, and eliminate isolated records that cannot be matched;

[0058] Step 2: Sort the height data and weight data of each child in chronological order, and add a timestamp corresponding to the collection time to each data point. If the height data or weight data corresponding to a timestamp is missing, fill it in using a linear interpolation algorithm; the linear interpolation algorithm calculates interpolation based on the data points of the adjacent timestamps before and after, and the calculation formula is:

[0059] ;

[0060] Among them, X1 and X2 represent the measurement values ​​of the adjacent moments, t1 and t2 represent the corresponding timestamps, t represents the interpolation timestamp, and X interp Represents the interpolation result;

[0061] In the specific implementation process, the matched data items are sorted in time and a corresponding timestamp is added to each data point. If a timestamp only has one of the data of height or weight, an interpolation algorithm is executed. The interpolation algorithm uses the values ​​of the adjacent valid data points to perform linear interpolation or other explicit interpolation operations. This step can obtain a complete sequence of height and weight aligned in time order for each child.

[0062] Step 3: Group the target population by age and gender, calculate the mean and standard deviation of the height and weight data of each group of samples, and obtain μ h , σ h , and μ w , σ w ; After eliminating pathological records, the height benchmark curve and weight benchmark curve of the corresponding age group were obtained using piecewise linear fitting method;

[0063] In the specific implementation process, based on the distribution of the target population at the same age and gender, the mean and standard deviation of the two indicators of height and weight are statistically analyzed, and the calculation is performed after eliminating pathological data. The baseline curve of height and weight is constructed using the piecewise linear fitting method. The curve corresponds to different linear fitting intervals in different age groups or different height and weight intervals to improve the adaptability to diversified data. This baseline curve can be regarded as a reference for the normal growth range, and the individual child's measurement values ​​will be compared with the curve later;

[0064] Step 4: At each monitoring time point t i , let the child's height be H i , weight is recorded as W i , respectively, with the reference value H of the benchmark curve ref，i With W ref，i Compare and calculate the growth difference coefficient; the calculation formula is as follows:

[0065] ;

[0066] ;

[0067] Among them, σ h and σ w are the standard deviation of height and weight for the corresponding age groups;

[0068] In the specific implementation process, for each monitoring time point t i , with H i Indicates the height at this time, W i Indicates the weight at this time, H ref，i Represents the baseline curve at time point t i The corresponding reference height, W ref，i Represents the baseline curve at time point t i Corresponding reference weight; introducing standard deviation σ h and σ w Respectively represent the standard deviation of height and weight of the target population, and the growth difference coefficient of individual children is calculated by the above growth difference coefficient calculation formula;

[0069] Step 5: D is the growth coefficient of the same child at multiple consecutive time points. h，i , D w，i Perform an overall analysis to determine whether any of the following conditions are met, including: Condition 1: There is at least one time point |D h，i ∣ or ∣ D w，i | is greater than the first threshold value T1, and the difference between adjacent time points is greater than the second threshold value T2; Condition 2: Within the preset monitoring window Δt, three consecutive test results are beyond the confidence interval of the reference curve; If any of the conditions is met, it is determined to be suspected growth abnormality;

[0070] In the specific implementation process, the growth difference coefficient D collected at multiple time points h，i , D w，i After being stored, it is necessary to determine whether the specified abnormal conditions are met; two thresholds T1 and T2 are set to detect abnormalities; when |D h，i ∣ or ∣ D w，iIf the value is greater than T1 and the difference between adjacent time points is greater than T2, it is considered as abnormal fluctuation; or if two of the three consecutive measurements are outside the reference distribution range, it is considered as suspected developmental abnormality;

[0071] Step 6: When suspected growth abnormalities occur, schedule monitoring cycles based on the abnormality type, including:

[0072] Shorten the length of the follow-up monitoring cycle so that the next collection of samples from the same child is earlier than originally planned;

[0073] When the same child has two consecutive abnormalities, the abnormality information is sent to the parents and designated pediatrician to trigger further offline evaluation;

[0074] In the specific implementation process, after an abnormality occurs, the interval for the next monitoring will be shortened to strengthen the follow-up tracking of children suspected of abnormalities. If the same child is judged to be abnormal in two consecutive monitorings, a warning message will be sent to the parents through the online medical system, and a follow-up consultation recommendation will be sent to the pediatrician;

[0075] Step 7: Write each monitoring result into the audit log, and synchronize the growth difference coefficient and the judgment result to the medical big data platform; recheck the statistical distribution of the growth difference coefficient of all children after the preset period, and correct the values ​​of the baseline curve and the first threshold T1 and the second threshold T2 to continuously improve the judgment accuracy;

[0076] During the specific implementation process, each monitoring data is written into the log for future retrieval and traceability, and synchronized to the medical big data platform. The growth difference coefficient distribution of all children is regularly counted, and the baseline curve and the values ​​of thresholds T1 and T2 are dynamically adjusted to improve the accuracy and stability under the measured data.

[0077] Pathological records excluded in step 3 include:

[0078] Data on children diagnosed with endocrine diseases were retrieved from the diagnostic records and excluded from the subsequent calculation of means and standard deviations;

[0079] Data from children who had received high-dose steroids within the last three months were identified and removed from the sample set or stored separately for specialized pathological growth model analyses;

[0080] In the specific implementation process, the elimination of pathological records includes identifying children diagnosed with endocrine diseases and children receiving high-dose hormone therapy, and removing their corresponding data from the sample or storing them separately. By searching the disease codes in the medical diagnosis records, we can find the types of diseases that are determined to seriously affect the growth pattern. In the case of high-dose hormone therapy (such as high-dose glucocorticoids in drug prescriptions), the relevant data will be excluded from the large sample cohort of healthy children because it may interfere with the normal growth curve of children.

[0081] Step 4 further includes:

[0082] Set a fixed monitoring time interval Δt. When the adjacent difference of the acquisition time exceeds Δt, generate a vacancy record and execute the interpolation algorithm in step 2.

[0083] By comparing the interpolation results with the benchmark curve, a complete and continuous sequence of individual growth difference coefficients is obtained;

[0084] In the specific implementation, during the time series structuring, interpolation is performed on the blank periods where the difference between adjacent time points is too large, and the filling result is compared with the reference curve to obtain a continuous difference coefficient sequence;

[0085] The monitoring time interval Δt is specified. If the collection interval between two adjacent records is greater than Δt, it is considered that there is a period of no collection in between.

[0086] After using the linear interpolation method to fill the data for this period, it is compared with the benchmark curve one by one to generate a complete and continuous list of difference coefficients.

[0087] When performing overall analysis in step 5, perform the following logical judgment:

[0088] When performing overall analysis, the difference between adjacent time points ΔD is calculated. h，i With ΔD w，i , if it exceeds the threshold, it is considered as abnormal fluctuation, and if abnormal fluctuation occurs twice or more in three consecutive tests, it can be directly judged as suspected abnormality; at each time point, calculate:

[0089] ;

[0090] ;

[0091] When ΔD h，i or ΔD w，i When it exceeds T2, the time point is marked as a fluctuation abnormality; if more than two fluctuation abnormalities occur in three consecutive tests, the system directly gives a suspected abnormality conclusion.

[0092] The number of consecutive multiple time points is between 3 and 6. When at least 3 to 6 monitoring values ​​exceed ±T1, offline review is triggered first; this limits the specific range of "consecutive multiple times" of monitoring time points, and once repeated abnormal results appear within this range, a high-level abnormal warning and offline intervention are automatically triggered logically. This is a consideration of actual clinical scenarios to avoid overdiagnosis due to an occasional measurement error.

[0093] The initial values ​​of the first threshold T1 and the second threshold T2 are determined in the following manners:

[0094] After counting all samples of the target group, the fluctuation range within the standard deviation of 1.5σ or 2σ is selected as the candidate value, and the comprehensive judgment accuracy is calculated;

[0095] Iteratively update the first threshold T1 and the second threshold T2 until the difference between the abnormal detection rate and the false positive rate meets the predetermined conditions; in the specific implementation process, the initial values ​​of the first threshold T1 and the second threshold T2 are based on the overall statistics of the population, and the difference between the abnormal detection rate and the false positive rate is iteratively optimized; wherein, the overall statistics of the population refer to the initial thresholds obtained based on the mean, standard deviation and fluctuation range of large-scale children's data. Both the abnormal detection rate and the false positive rate can be statistically compared with confirmed real abnormal cases and non-abnormal cases. When optimizing iteratively, the goal is to minimize the false positive rate and maintain a high detection rate, and finally select T1 and T2 that meet clinical requirements.

[0096] The alarm and dynamic intervention in step 6 also include:

[0097] When suspected growth abnormality occurs, the time interval for the next monitoring is automatically shortened to 50% of the original interval or other preset ratio;

[0098] If the same child is judged to be abnormal within two monitoring intervals, the system will send a reminder request to the pediatric diagnosis and treatment center and push a detailed assessment report to the parents;

[0099] When the same child shows abnormal status in three consecutive monitorings, it triggers a specialist consultation and the child is included in the key follow-up file;

[0100] In the specific implementation process, during the alarm and dynamic intervention process, if the same child has two consecutive abnormalities, an alarm will be sent to the parents and pediatricians; if three consecutive abnormalities occur, it will directly trigger an expert consultation and be included in the key tracking file. After identifying the "suspected abnormalities" in the previous step, the program further increases attention to serious abnormalities through graded alarms. When three consecutive abnormalities occur, the highest level of treatment is used, and it is recommended that an expert team of medical staff immediately intervene to analyze their growth and development to rule out potential diseases or serious abnormalities.

[0101] A child growth and development monitoring and evaluation system based on medical big data, the system includes a data aggregation module, a time series processing module, a benchmark construction module, a difference coefficient calculation module, an abnormality judgment module, and an intervention strategy module; wherein the data aggregation module is used to receive and integrate the child height, weight and diagnosis records transmitted by at least two medical institutions, mark the records from different sources and archive them uniformly; the time series processing module is used to execute step 2 of the child growth and development monitoring and evaluation method based on medical big data, and perform time series and interpolation calculation on the growth data of each child; the benchmark construction module is used to execute step 3 of the child growth and development monitoring and evaluation method based on medical big data, and construct a benchmark growth interval for the target population; the difference coefficient calculation module is used to execute step 4 of the child growth and development monitoring and evaluation method based on medical big data, and calculate the growth difference coefficient for the individual data of the child, that is, calculate ΔD successively h，i With ΔD w，i The abnormality determination module is used to execute step 5 of the child growth and development monitoring and evaluation method based on medical big data, and identify abnormal trends according to the difference coefficient and the pre-set threshold value; the intervention strategy module is used to execute step 6 of the child growth and development monitoring and evaluation method based on medical big data, and issue an alarm and adjust the subsequent monitoring strategy when it is determined to be a suspected abnormality.

[0102] An electronic device, which can be a hospital's backend server or cloud computing platform, or a local machine equipped with sufficient computing power, includes at least one processor and a memory, wherein a computer program is stored in the memory. When the computer program is executed by the processor, the processor executes any of the above-mentioned child growth and development monitoring and evaluation methods based on medical big data. When executing this method, the processor executes the corresponding algorithm in the order of steps 1 to 7, and cooperates with database reading and writing operations.

[0103] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, it implements any one of the child growth and development monitoring and evaluation methods based on medical big data. In the specific implementation process, the storage medium can be a hard disk, a flash memory or a CD, which contains executable code for executing the above-mentioned monitoring and evaluation method, and can complete all the functions of the present invention after deployment.

[0104] It should be further explained that in the specific implementation process: first, the system environment and hardware configuration are set: for the medical big data platform: it is equipped with a high-performance database and distributed computing framework to collect and process data from multiple hospitals; for the data interface of multiple hospitals: each hospital regularly uploads children's growth data through a secure network connection, and the data format complies with the unified specifications agreed in advance; for the server side: it includes at least one computing node with a quad-core processor, 16GB of memory, and a hard disk with a storage capacity of 2TB; the server side runs the monitoring and evaluation program described below; for the client side: the doctor's workstation and the parent's mobile application are used to view the monitoring results and abnormal reminder information;

[0105] The medical big data platform extracted physical examination data from the databases of the two hospitals over the past two years, including height and weight information of at least 5,000 children, with collection dates and diagnosis records. The system matched data from different sources by the child’s ID number or the unique hospital visit number. If a record could not be matched to any registered child’s identity, it would be deleted. According to the pre-defined rules, if two or more repeated measurements of the same child occurred on the same day, the system would only retain the measurement record closest to the noon period, and the other records would be considered redundant and discarded.

[0106] The records of children with endocrine diseases (disease codes in the range of E00-E07) or high-dose hormone treatment in the diagnosis field are identified and marked by the system; the height and weight data of these marked children are stored separately in another database table and no longer participate in the statistical calculation of the normal baseline curve; after statistics, after eliminating pathological or severely intervened data, the remaining health development data of about 4,500 children meet the requirements of benchmark construction;

[0107] The data of each child are arranged in ascending order by the date of collection to obtain a sequence with equally or irregularly spaced timestamps; if the interval between two consecutive measurements is greater than 90 days, several virtual timestamps are inserted in this time period and linear interpolation is performed so that each child can eventually obtain a continuous time series with no missing values ​​between height and weight;

[0108] The children were classified into 24 subgroups according to age (1 to 12 years old) and gender (male and female). Each subgroup contained a large number of children's measurement series. In each subgroup, the mean values ​​of height and weight were calculated. h and μ w , and the standard deviation σ h and σ w ; Use piecewise linear fitting to draw the height and weight baseline curves corresponding to 1 to 12 years old for each subgroup. The curves may use different linear function segments in different age ranges to improve fitting accuracy;

[0109] Analyze a child A who has recently uploaded data on the platform; assume that A has added 5 measurement records recently, including height (103cm, 105cm, 107cm, 110cm, 113cm) and weight (17kg, 18.5kg, 19kg, 20kg, 20.5kg), and the corresponding time intervals are all about 3 months;

[0110] On the subgroup baseline curve corresponding to the age and gender of the child A, the system searches for the corresponding reference height H ref，i and reference weight W ref，i , and based on σ h , σ w The measured values ​​were used to calculate the coefficient of variation;

[0111] Set the first threshold value T1 = 2.0 and the second threshold value T2 = 1.0; if child A has |D at least twice in three consecutive measurements h，i ∣ Greater than 2.0, or the difference between adjacent time points is ΔD h，i If it exceeds 1.0, the system determines that A is at risk of significantly deviating from the benchmark;

[0112] If A only slightly exceeds the reference range in a single measurement, the system will consider it a temporary deviation and will not immediately trigger a high-level abnormality, but will automatically shorten the measurement interval to half of the original one during the next collection;

[0113] Suppose child A appears twice in a row | D h，i ∣ and ∣D w，i ∣If both are over 2.0, the system will judge it as a serious abnormality and automatically send a reminder to A's guardian on the mobile app; at the same time, the system background will organize A's historical data into charts and send them to the pediatrician's workstation, and the doctor can make further judgments based on the existing diagnostic information; if A is judged to be abnormal again, the system will trigger the expert consultation process, include A in the key tracking file, and recommend comprehensive medical examinations to rule out potential diseases;

[0114] The system generates an unchangeable log record for each new piece of data and each judgment result, and stores it in the audit database table for subsequent audit and tracing;

[0115] During regular maintenance, the system re-counts all collected data to detect the actual abnormal detection rate and false positive rate, and if necessary, h , σ w Or make small-scale optimizations on T1 and T2 to continuously improve the detection accuracy;

[0116] The method runs on the medical big data platform (server), and the platform exchanges data with the local server of each hospital through a dedicated line connection or a virtual private network; parents and pediatricians access the system interface or mobile application through electronic devices (mobile phones, tablets, computers) to obtain real-time monitoring results; the corresponding client program is installed on the device, and the client program and the cloud server transmit data using an encrypted protocol;

[0117] Starting from data collection, to time series processing, abnormality determination, intervention reminders, and subsequent data recording and strategy optimization, an organic growth and development monitoring system is formed; the system triggers medical intervention or expert consultation process when necessary (such as multiple abnormalities in the same child), and stores all test data and results in full, which is convenient for re-verification and correction of benchmarks;

[0118] By strictly controlling data quality during the cleaning and alignment stages and eliminating pathological data, the scientific nature of the baseline curve is effectively guaranteed; through difference coefficient calculation and threshold determination, refined detection of growth differences between individuals and groups is achieved to reduce false alarms and missed alarms; a dynamic intervention and graded alarm mechanism is adopted to optimize the allocation of medical resources while ensuring monitoring accuracy; and by regularly updating the baseline curve and threshold, the reliability and robustness of the algorithm are continuously improved through iteration.

[0119] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0120] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring and evaluating children's growth and development based on medical big data, characterized in that: The method comprises the following steps: Step 1: Read the child's height data, weight data, and corresponding collection time points and diagnosis records from the databases of at least two medical institutions; match the data according to the child's unique identifier, remove data records whose collection time differs by more than a preset threshold, and remove isolated records that cannot be matched; Step 2: Sort the height data and weight data of each child in chronological order, and add a timestamp corresponding to the collection time to each data point. If the height data or weight data corresponding to a timestamp is missing, fill it in using a linear interpolation algorithm; Step 3: Group the target population by age and gender, calculate the mean and standard deviation of the height and weight data of each group of samples, and obtain μ h , σ h , and μ w , σ w ; After excluding pathological records, the height baseline curve and weight baseline curve were obtained using piecewise linear fitting; Step 4: At each monitoring time point t i , let the child's height be H i , weight is recorded as W i , respectively, with the reference value H of the reference curve ref，i With W ref，i Compare and calculate the growth difference coefficient. The calculation formula is as follows: ； ； Among them, σ h and σ w are the standard deviation of height and weight for the corresponding age groups; Step 5: D is the growth coefficient of the same child at multiple consecutive time points. h，i , D w，i Perform an overall analysis to determine whether any of the following conditions are met, including: Condition 1: There is at least one time point |D h，i ∣ or ∣ D w，i | is greater than the first threshold value T1, and the difference between adjacent time points is greater than the second threshold value T2; Condition 2: Within the preset monitoring window Δt, three consecutive test results are beyond the confidence interval of the reference curve; If any of the conditions is met, it is determined to be suspected growth abnormality; Step 6: When suspected growth abnormalities occur, schedule monitoring cycles based on the abnormality type; Step 7: Write each monitoring result into the audit log, and synchronize the growth difference coefficient and judgment result to the medical big data platform.

2. According to claim 1, a method for monitoring and evaluating children's growth and development based on medical big data is characterized by: Eliminating pathological records in step 3 includes: Data of children diagnosed with endocrine diseases were retrieved from diagnostic records, and data of children diagnosed with endocrine diseases were not included in the subsequent calculation of means and standard deviations; Data from children who had received high-dose steroid therapy within the last three months were identified and removed from the sample set or stored separately for specialized pathological growth model analyses.

3. A method for monitoring and evaluating children's growth and development based on medical big data according to claim 2, characterized in that: The step 4 further comprises: Set a fixed monitoring time interval Δt. When the adjacent difference of the acquisition time exceeds Δt, generate a vacancy record and execute the interpolation algorithm in step 2. The interpolation result is compared with the reference curve to obtain a complete and continuous sequence of individual growth difference coefficients.

4. The method for monitoring and evaluating children's growth and development based on medical big data according to claim 3 is characterized by: When performing overall analysis in step 5, the following logical judgment is performed: Calculate the change in the coefficient of difference between adjacent time points. The calculation formula is as follows: ； ； When ΔD h，i or ΔD w，i When the pre-defined sudden increase or decrease threshold T2 is exceeded, the time point is marked as abnormal fluctuation; If abnormal fluctuations occur at least twice in three consecutive tests, it is directly judged as suspected growth abnormality.

5. A method for monitoring and evaluating children's growth and development based on medical big data according to claim 4, characterized in that: The number of the multiple consecutive time points ranges from 3 to 6. When 3 to 6 monitoring values ​​in the child's growth difference coefficient sequence exceed ±T1 continuously, it is judged as suspected growth abnormality and offline review is triggered as a priority.

6. A method for monitoring and evaluating children's growth and development based on medical big data according to claim 5, characterized in that: The initial values ​​of the first threshold T1 and the second threshold T2 are determined in the following manners: After counting all samples of the target group, the fluctuation range within the standard deviation of 1.5σ or 2σ is selected as the candidate value, and the comprehensive judgment accuracy is calculated; The first threshold T1 and the second threshold T2 are iteratively updated until the difference between the abnormality detection rate and the false positive rate meets a predetermined condition.

7. A method for monitoring and evaluating children's growth and development based on medical big data according to claim 6, characterized in that: The alarm and dynamic intervention in step 6 also include: When suspected growth abnormality occurs, the time interval for the next monitoring is automatically shortened to 50% of the original interval or other preset ratio; If the same child is judged to be abnormal within two monitoring intervals, the system will send a reminder request to the pediatric diagnosis and treatment center and push a detailed assessment report to the parents; When the same child showed abnormal status in three consecutive monitorings, it triggered a specialist consultation and included the child in the key follow-up file.

8. A child growth and development monitoring and evaluation system based on medical big data, characterized in that: The system comprises: A data aggregation module for receiving and integrating the height, weight and diagnosis records of children transmitted from at least two medical institutions; A time series processing module, used to perform step 2 of the method according to claim 1, and perform time series and interpolation calculation on the growth data of each child; A benchmark construction module, used to execute step 3 of the method according to claim 1, to construct a benchmark growth interval for a target population; A coefficient of difference calculation module, used to perform step 4 of the method according to claim 1, and calculate the growth coefficient of difference for the individual data of the children; An abnormality determination module, used to execute step 5 of the method according to claim 1, and identify abnormal trends according to the difference coefficient and a preset threshold; The intervention strategy module is used to execute step 6 of the method according to claim 1, and issue an alarm and adjust the subsequent monitoring strategy when it is determined to be a suspected abnormality.

9. An electronic device, characterized in that: It includes at least one processor and a memory, in which a computer program is stored. When the computer program is executed by the processor, the processor executes the child growth and development monitoring and evaluation method based on medical big data according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the method for monitoring and evaluating children's growth and development based on medical big data according to any one of claims 1 to 7 is implemented.

Images