An intelligent medical system for respiratory monitoring in clinical nursing

Abstract

The present invention relates to the field of respiratory monitoring technology, and specifically to an intelligent medical system for clinical nursing respiratory monitoring. The present invention first obtains the basic medical record information of all patients under respiratory monitoring, and the monitoring curves under each monitoring indicator; according to the similarity of the basic medical record information, all patient groups are obtained; the periodic characteristics of each monitoring curve are analyzed to obtain its degree of instability; and then the fluctuation period and fluctuation range of the respiratory monitoring data under each monitoring indicator in each patient group are obtained in combination with the autocorrelation and fluctuation amplitude of all monitoring curves; finally, respiratory monitoring warning is performed on the patients under respiratory monitoring. Based on the idea that the fluctuation changes of the respiratory monitoring data of patients with similar basic signs are also similar, the present invention classifies all patients, and combines the periodic characteristics of normal breathing to determine the monitoring warning reference of the respiratory monitoring data under each monitoring indicator in each patient category, thereby improving the respiratory monitoring effect for patients with different physical signs.

Description

Technical Field

[0001] The present invention relates to the technical field of respiratory monitoring, and in particular to an intelligent medical system for clinical nursing respiratory monitoring. Background Art

[0002] Respiratory monitoring is an important vital sign monitoring indicator in clinical care. It can help evaluate the patient's respiratory status and oxygenation, detect potential problems in a timely manner, and guide nursing and treatment measures. Respiratory monitoring is crucial to protecting the patient's life and health. At present, when patients are monitored for respiratory failure, monitoring instruments such as ventilators and blood oximeters are usually used to continuously monitor the patient's heart rate, blood pressure, respiratory rate, blood oxygen saturation and other respiratory monitoring data. When any respiratory monitoring data deviates from the corresponding normal parameter range, the monitoring instrument will sound an alarm, thereby reminding medical staff to respond quickly.

[0003] However, due to the different health conditions and monitoring importance of different patients, directly using the same normal parameter range for all patients may lead to ignoring the changes in the respiratory monitoring data of some seriously ill patients, thereby delaying the discovery of important diseases. It may also lead to excessive sensitivity to changes in the respiratory monitoring data of patients with milder conditions and frequent false alarms, thereby affecting the respiratory monitoring effect of patients with different physical conditions. Summary of the invention

[0004] In order to solve the technical problem that the existing technology has poor respiratory monitoring effect on patients with different physical conditions, the purpose of the present invention is to provide an intelligent medical system for clinical nursing respiratory monitoring, and the technical solution adopted is as follows:

[0005] The present invention provides an intelligent medical system for clinical nursing respiratory monitoring, the system comprising:

[0006] The data acquisition module is used to obtain the basic medical record information of all patients under respiratory monitoring, and obtain the monitoring curve of each patient under historical respiratory monitoring under each monitoring indicator, and the real-time monitoring data of patients under respiratory monitoring under each monitoring indicator;

[0007] A data analysis module is used to classify all historical respiratory monitoring patients according to the similarity of the basic medical record information between different historical respiratory monitoring patients to obtain all patient groups; obtain all periodic sub-segments in the monitoring curve according to the autocorrelation of each monitoring curve; obtain the instability of each monitoring curve according to the similarity between different periodic sub-segments in each monitoring curve; in each patient group, under each monitoring indicator, according to the autocorrelation and fluctuation amplitude of all the monitoring curves, combined with the instability of each monitoring curve, obtain the fluctuation period and fluctuation range of the respiratory monitoring data;

[0008] The intelligent monitoring module is used to determine a reference group of patients to be monitored from all patient groups based on the basic medical record information of the patients to be monitored; under each monitoring indicator, according to the fluctuation changes of the real-time monitoring data of the patients to be monitored, combined with the fluctuation period and the fluctuation range of the respiratory monitoring data under each monitoring indicator in the reference group, a respiratory monitoring warning is given to the patients to be monitored.

[0009] Furthermore, the basic medical record information includes surgical record data and outpatient physical examination data; the surgical record data at least includes the type of surgery and the anesthesia method; the outpatient physical examination data includes basic physical sign parameters under each physical sign indicator and all previous medical history.

[0010] Furthermore, the method for obtaining the patient group includes:

[0011] According to the similarity of the operation record data between different historical respiratory monitoring patients and the similarity of the basic physical sign data, obtaining the physical sign similarity coefficient between different historical respiratory monitoring patients;

[0012] All patients with historical respiratory monitoring are clustered based on the physical sign similarity coefficient, and the clustering results corresponding to all clusters are used as all patient groups.

[0013] Furthermore, the method for obtaining the physical sign similarity coefficient includes:

[0014] A binary array is constructed according to the surgical type and anesthesia method of each historical respiratory monitoring patient, and the binary array is used as surgical record data; the Jaccard coefficient between the binary arrays of different historical respiratory monitoring patients is used as the surgical correlation between the corresponding different historical respiratory monitoring patients;

[0015] According to the parameter difference between the basic physical sign parameters under all the same physical sign indicators, the physical sign consistency between the different historical respiratory monitoring patients is obtained; the parameter difference is negatively correlated with the physical sign consistency;

[0016] Constructing a multivariate array according to all the past medical histories of each patient with historical respiratory monitoring, and using the Jaccard coefficient between the multivariate arrays of different patients with historical respiratory monitoring as the medical history correlation between the corresponding different patients with historical respiratory monitoring;

[0017] A sign similarity coefficient is obtained according to the surgical correlation, the sign fit and the medical history correlation between patients with different historical respiratory monitoring, and the surgical correlation, the sign fit and the medical history correlation are all positively correlated with the sign similarity coefficient.

[0018] Furthermore, the method for obtaining the periodic sub-segment includes:

[0019] The period of each monitoring curve is obtained based on the autocorrelation function, and the period is used as the length of each period subsegment in the corresponding monitoring curve. The period is divided iteratively from the starting point of each monitoring curve to obtain all period subsegments of each monitoring curve.

[0020] Furthermore, the method for obtaining the degree of instability includes:

[0021] In each of the monitoring curves, the average values ​​of the DTW distances between all different period subsegments are normalized, and the normalized result is used as the instability degree of each of the monitoring curves.

[0022] Furthermore, the method for obtaining the fluctuation period and fluctuation range includes:

[0023] Under each monitoring indicator, all stable monitoring curves are screened out from all the corresponding monitoring curves in each patient group according to the instability degree;

[0024] In each of the patient groups, under each monitoring index, the average of the periods of all the stable monitoring curves is used to obtain the fluctuation period of the respiratory monitoring data;

[0025] In each of the patient groups, under each monitoring index, the maximum monitoring data in all the stable monitoring curves is used as the upper limit of the fluctuation range, and the minimum monitoring data in all the stable monitoring curves is used as the lower limit of the fluctuation range to obtain the fluctuation range.

[0026] Furthermore, the method for obtaining the stable monitoring curve includes:

[0027] In each of the patient groups, under each monitoring indicator, all the monitoring curves whose instability is less than a preset threshold are regarded as stable monitoring curves.

[0028] Furthermore, the method for obtaining the reference group includes:

[0029] The physical sign similarity coefficient between the patient to be monitored for respiratory failure and the corresponding cluster center of each patient group is used as a matching parameter; the patient group corresponding to the minimum matching parameter is used as a reference group of the patient to be monitored for respiratory failure.

[0030] Furthermore, the method for performing respiratory monitoring early warning on a patient undergoing respiratory monitoring includes:

[0031] Under each monitoring index, the change cycle of the real-time monitoring data of the patient to be monitored for respiratory function is obtained based on the autocorrelation function;

[0032] If the real-time monitoring data of the patient to be monitored for respiratory failure under any monitoring indicator exceeds the fluctuation range of the respiratory monitoring data under the corresponding monitoring indicator in the reference group, or the period difference between the change period and the fluctuation period of the real-time monitoring data of the patient to be monitored for respiratory failure under any monitoring indicator is greater than a preset period threshold, a respiratory monitoring warning will be issued to the patient to be monitored for respiratory failure.

[0033] The present invention has the following beneficial effects:

[0034] The present invention obtains the basic medical record information of all respiratory monitoring patients through a data acquisition module, and obtains the monitoring curve of each historical respiratory monitoring patient under each monitoring index, and the real-time monitoring data of the patients to be respiratory monitoring under each monitoring index, so as to prepare data for the subsequent analysis of patients with similar signs and determine the normal fluctuation reference of the corresponding monitoring data of each similar patient group; in the data analysis module, all historical respiratory monitoring patients are classified according to the similarity of the basic medical record information between different historical respiratory monitoring patients, and all patient groups are obtained. The physical fitness of all historical respiratory monitoring patients in each patient group is roughly similar, so that the corresponding monitoring warning reference of each patient group can be determined according to the respiratory monitoring situation of all patients with similar signs in each patient group; then, according to the autocorrelation of each monitoring curve, all periodic subsegments of the monitoring curve are obtained; according to each The similarity between different periodic sub-segments in the monitoring curves is used to obtain the instability of each monitoring curve. The instability reflects whether the monitoring curve has normal periodic fluctuation characteristics, so as to facilitate the subsequent screening of normal patient groups and accurately determine the monitoring warning reference; in each patient group, under each monitoring index, according to the autocorrelation and fluctuation amplitude of all monitoring curves, combined with the instability of each monitoring curve, the fluctuation period and fluctuation range of the respiratory monitoring data are obtained; finally, in the intelligent monitoring module, according to the basic medical record information of the patients to be monitored, the reference group of patients to be monitored is determined from all patient groups; under each monitoring index, according to the fluctuation change of the real-time monitoring data of the patients to be monitored, combined with the fluctuation period and fluctuation range of the respiratory monitoring data under each monitoring index in the reference group, the patients to be monitored are given a respiratory monitoring warning. The present invention is based on the idea that different patients have different basic signs, and the fluctuation change of the respiratory monitoring data of patients with similar basic signs is also similar, classifies all patients, and determines the fluctuation range and fluctuation period corresponding to the respiratory monitoring data under each monitoring index in each patient category in combination with the periodic change characteristics of the normal respiratory monitoring data in each patient category, thereby providing a monitoring warning reference and improving the respiratory monitoring effect of patients with different physical signs. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 A system block diagram of an intelligent medical system for clinical nursing respiratory monitoring provided by an embodiment of the present invention;

[0037] Figure 2 A flow chart of a method for obtaining a physical sign similarity coefficient provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, features and effects of an intelligent medical system for clinical nursing respiratory monitoring proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0039] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0040] The specific scheme of an intelligent medical system for clinical nursing respiratory monitoring provided by the present invention is described in detail below with reference to the accompanying drawings.

[0041] See also Figure 1 , which shows a system block diagram of an intelligent medical system for clinical nursing respiratory monitoring provided by an embodiment of the present invention. The system includes a data acquisition module 101, a data analysis module 102 and an intelligent monitoring module 103.

[0042] The data acquisition module 101 is used to obtain the basic medical record information of all respiratory monitoring patients, and obtain the monitoring curve of each historical respiratory monitoring patient under each monitoring indicator, and the real-time monitoring data of the patient to be respiratory monitored under each monitoring indicator.

[0043] In one embodiment of the present invention, basic case information of all respiratory monitoring patients who need respiratory monitoring during postoperative clinical care is first obtained from the hospital database, wherein the basic medical record information includes surgical record data and outpatient physical examination data; the surgical record data at least includes the surgical type and anesthesia method of the respiratory monitoring patient; the outpatient physical examination data includes the basic physical sign parameters under each physical sign indicator and all previous medical history; further, the monitoring curve of each historical respiratory monitoring patient under each monitoring indicator during the historical clinical care process is obtained, and the real-time monitoring data of the current patient to be monitored under each monitoring indicator is collected and monitored in real time; the above data are obtained for subsequent analysis of patients with similar signs, and data preparation is made for determining the normal fluctuation reference of the corresponding monitoring data of each similar patient group, so as to improve the respiratory monitoring effect of the patients to be monitored.

[0044] The surgical record data needs to be retrieved from the database and analyzed by analyzing the surgical record forms of patients under respiratory monitoring. The types of surgery include different risk levels such as Class I risk, Class II risk and Class III risk, which are determined by relevant medical staff based on the difficulty of the surgery; anesthesia methods include local anesthesia, general anesthesia, etc.

[0045] Outpatient physical examination data needs to be retrieved from the database and analyzed by retrieving the outpatient physical examination record form of respiratory monitoring patients. The physical sign indicators at least include basic information indicators such as blood pressure, heart rate, body temperature, blood oxygen, height, weight, etc., and the specific blood pressure values, heart rate values, etc. are used as the basic physical sign parameters under the corresponding physical sign indicators; past medical history needs to extract important keywords from the information recorded in the past medical history column in the outpatient physical examination record form. Specifically, the jieba word segmentation tool is used to segment the information recorded in the past medical history column, and stop words are removed to obtain all past medical history segmentations, such as heart disease, diabetes, kidney stones, etc.

[0046] The monitoring curve needs to be collected through a multi-parameter monitor, where the monitoring indicators include at least heart rate, tidal value, respiratory rate, blood oxygen saturation, etc., to obtain multiple monitoring curves collected and recorded by the multi-parameter monitor; at the same time, it is also necessary to collect monitoring data of patients awaiting respiratory monitoring in real time; the monitoring curve of patients with historical respiratory monitoring is the monitoring curve during the complete monitoring process, while the real-time monitoring data of patients awaiting respiratory monitoring is the monitoring curve collected up to the current moment.

[0047] It should be noted that in other embodiments of the present invention, the implementer may also obtain surgical record data of other dimensions or quantities, such as anesthesia dosage or operation duration, to construct surgical record data; may also obtain other types or quantities of vital sign indicators and collect corresponding basic vital sign parameters; may also use deep learning algorithms to extract medical history segmentation, which and jieba segmentation are both existing technologies well known to technicians in this field and will not be described in detail here; the implementer may also obtain other types or quantities of monitoring indicators and collect corresponding monitoring curves, and may also use other monitoring instruments to collect monitoring curves under each monitoring indicator, which are all existing technologies and will not be described in detail here.

[0048] The data analysis module 102 is used to classify all historical respiratory monitoring patients according to the similarity of basic medical record information between different historical respiratory monitoring patients to obtain all patient groups; obtain all periodic sub-segments in the monitoring curve according to the autocorrelation of each monitoring curve; obtain the instability of each monitoring curve according to the similarity between different periodic sub-segments in each monitoring curve; in each patient group, under each monitoring indicator, according to the autocorrelation and fluctuation amplitude of all monitoring curves, combined with the instability of each monitoring curve, obtain the fluctuation period and fluctuation range of the respiratory monitoring data.

[0049] Considering that the physical fitness of different patients may be similar, their postoperative nursing recovery process may be relatively similar, and the corresponding respiratory monitoring data changes may also be similar, and the basic medical record information roughly reflects the physical fitness of each respiratory monitoring patient; therefore, after using the data acquisition module 101 to obtain the basic medical record information of all respiratory monitoring patients, all historical respiratory monitoring patients can be classified according to the similarity of the basic medical record information between different historical respiratory monitoring patients, and all patient groups can be obtained. The physical fitness of all historical respiratory monitoring patients in each patient group is roughly similar, so that the corresponding monitoring warning reference for each patient group can be determined based on the respiratory monitoring situation of all patients with similar signs in each patient group.

[0050] Preferably, in one embodiment of the present invention, considering that the more similar the surgical record data is, the more similar the difficulty of postoperative nursing monitoring is, the more similar the patient's recovery process is, and the more similar the basic physical sign data is, the more similar the patient's basic physical fitness is, and the more likely the postoperative recovery progress is to be similar; combining the two can further accurately evaluate respiratory monitoring patients with similar signs; considering that clustering can cluster all similar data into one cluster; based on this, the method for obtaining patient groups includes:

[0051] According to the similarity of surgical record data and basic sign data between different historical respiratory monitoring patients, the physical sign similarity coefficient between different historical respiratory monitoring patients was obtained; all historical respiratory monitoring patients were clustered based on the physical sign similarity coefficient, and the clustering results corresponding to all clusters were used as all patient groups.

[0052] As an example, after obtaining the physical sign similarity coefficients between patients with different historical respiratory monitoring, considering that the physical sign similarity coefficient reflects the degree of physical sign similarity between patients with different historical respiratory monitoring, the more similar they are, the more they should be clustered into one patient group. Therefore, the physical sign similarity parameters are first negatively correlated and mapped, such as performing a reciprocal operation or mapping the physical sign similarity parameters to an exponential function exp(-x) with a natural constant e as the base, and the negative correlation mapping result is used as the metric distance between patients with different historical respiratory monitoring. Then, the K-means clustering algorithm is used for clustering, and the number of clusters is obtained by the elbow method to obtain all clusters, where one cluster corresponds to one patient group. It should be noted that a very small positive parameter should be added when performing a reciprocal operation to avoid meaninglessness caused by a denominator of 0. Implementers may also use other negative correlation mapping methods or clustering algorithms, which are all existing technologies as well as the K-means algorithm and the elbow method and will not be described in detail here.

[0053] In a preferred embodiment of the present invention, the method for obtaining the physical sign similarity coefficient includes:

[0054] See also Figure 2 , which shows a flow chart of a method for obtaining a physical sign similarity coefficient provided by an embodiment of the present invention, specifically comprising:

[0055] Step S201, constructing a binary array according to the surgical type and anesthesia method of each historical respiratory monitoring patient, and using the binary array as surgical record data; using the Jaccard coefficient between the binary arrays of different historical respiratory monitoring patients as the surgical correlation between the corresponding different historical respiratory monitoring patients.

[0056] As an example, the binary array form of the surgical record data of a patient with historical respiratory monitoring is specifically ["Class I risk", "local anesthesia"]. Considering that the Jaccard coefficient can be used to measure the similarity of two sets, the Jaccard coefficient between the binary arrays of different historical respiratory monitoring patients is used as the surgical correlation between different historical respiratory monitoring patients. The surgical correlation reflects the monitoring similarity between different historical respiratory monitoring patients from the perspective of surgical record data.

[0057] It should be noted that the implementer may also use the Tanimoto coefficient or assign numerical labels to the text information in the binary array and then calculate the Euclidean distance or cosine similarity, etc., which and the Jaccard coefficient are both existing technologies and will not be described in detail here.

[0058] Step S202, among different historical respiratory monitoring patients, according to the parameter difference between the basic physical sign parameters under all the same physical sign indicators, the physical sign consistency between different historical respiratory monitoring patients is obtained; the parameter difference is negatively correlated with the physical sign consistency.

[0059] Considering that the smaller the difference in basic physical sign parameters under each physical sign indicator, the more similar the physical constitution of the historical respiratory monitoring patients is; the physical sign fit reflects the monitoring similarity between different historical respiratory monitoring patients from the perspective of basic physical sign information in the outpatient physical examination records.

[0060] As an example, the formula for calculating the physical sign fit is: ;in, For the The history of respiratory monitoring patients and The consistency of physical signs between patients with historical respiratory monitoring; is the serial number of the physical sign indicator; is the total number of physical sign indicators; For the Patients with history of respiratory monitoring in Basic physical sign parameters under individual sign indicators; For the Patients with history of respiratory monitoring in Basic physical sign parameters under individual sign indicators; is the absolute value symbol; For the The history of respiratory monitoring patients and Patients with history of respiratory monitoring in Differences in parameters under individual sign indicators; In order to preset the minimum positive parameter, this example takes 0.1 to ensure that the denominator is meaningful without affecting the calculation results.

[0061] In the calculation formula of the physical sign fit, the parameter difference between the basic physical sign parameters under each physical sign indicator is added to the preset minimum positive parameter and then an inverse operation is performed, so that the smaller the parameter difference, the greater the physical sign fit; in other examples, implementers can also adopt other negative correlation mapping methods, such as taking the parameter difference as the x in the exponential function exp (-x) with the natural constant e as the base, which will not be repeated here.

[0062] Step S203, constructing a multivariate array according to all the past medical histories of each historical respiratory monitoring patient, and using the Jaccard coefficient between the multivariate arrays of different historical respiratory monitoring patients as the medical history correlation between the corresponding different historical respiratory monitoring patients.

[0063] Considering that the more similar the past medical history is, the more similar the physical constitution of the historical respiratory monitoring patients is; and considering that the Jaccard coefficient can be used to measure the similarity of two sets, a multivariate array is constructed based on the past medical history. For example, the multivariate array form of the past medical history of a historical respiratory monitoring patient is specifically ["diabetes", "kidney stones", "heart disease"], where the form of the multivariate array corresponding to different historical respiratory monitoring patients may be different, that is, the number of past medical history segmentations in the multivariate array is different; the Jaccard coefficient between the multivariate arrays of different historical respiratory monitoring patients is used as the medical history correlation between the corresponding different historical respiratory monitoring patients. Medical history correlation reflects the monitoring similarity between different historical respiratory monitoring patients from the perspective of past medical history in outpatient physical examination records.

[0064] It should be noted that the implementer may also use the Tanimoto coefficient or assign numerical labels to the text information in the multivariate array and then calculate the Euclidean distance or cosine similarity, etc., which and the Jaccard coefficient are both existing technologies and will not be described in detail here.

[0065] Step S204, obtaining a physical sign similarity coefficient according to the surgical correlation, physical sign consistency and medical history correlation between patients with different historical respiratory monitoring. The surgical correlation, physical sign consistency and medical history correlation are all positively correlated with the physical sign similarity coefficient.

[0066] As an example, the surgical relevance, physical sign fit and medical history relevance between different historical respiratory monitoring patients are multiplied and combined to obtain the physical sign similarity coefficient; the physical sign similarity coefficient combines the three to reflect the monitoring similarity between different historical respiratory monitoring patients, thereby facilitating the subsequent classification of all historical respiratory monitoring patients. In other examples, the implementer may also combine the three by basic mathematical operations such as addition or weighted summation or related mapping methods, which will not be repeated here.

[0067] At this point, all patient groups are obtained.

[0068] Considering that under normal circumstances, during the monitoring process of respiratory monitoring patients, their breathing and heartbeat have certain regularities, the fluctuation changes of the corresponding monitoring curve should also have certain fluctuation regularities and periods. If the fluctuation is disordered, it means that the vital signs of the respiratory monitoring patient may be abnormal; therefore, each monitoring curve can be segmented according to the period, and each segment corresponds to a curve sub-segment of periodic fluctuation. By analyzing the similarities between all curve sub-segments, it can be determined whether the monitoring curve has periodic fluctuation characteristics, which makes it easier to know the abnormal fluctuation of the monitoring curve.

[0069] Considering that autocorrelation can help identify periodic patterns in time series, the embodiment of the present invention obtains all periodic sub-segments in the monitoring curve according to the autocorrelation of each monitoring curve; further, according to the similarity between different periodic sub-segments in each monitoring curve, the instability of each monitoring curve is obtained. Each periodic sub-segment approximately represents a periodic fluctuation curve segment in the monitoring curve, and the more similar the different periodic sub-segments in each monitoring curve are, the more obvious the periodic characteristics of the monitoring curve are, and the smaller the instability is; the instability reflects whether the monitoring curve has normal periodic fluctuation characteristics.

[0070] Preferably, in one embodiment of the present invention, the method for acquiring the periodic sub-segment includes:

[0071] The period of each monitoring curve is obtained based on the autocorrelation function, and the period is used as the length of each period sub-segment in the corresponding monitoring curve. The period is iteratively divided from the starting point of each monitoring curve to obtain all period sub-segments.

[0072] It should be noted that obtaining the period of the curve of the autocorrelation function is an existing technology well known to those skilled in the art and will not be elaborated here; the period of each monitoring curve may be different, starting from the starting point of each monitoring curve, the first period sub-segment of the corresponding period length is obtained, and then the next data point at the end point of the first period sub-segment is used as the starting point to obtain the second period sub-segment of the corresponding period length, and it is continuously iterated until all period sub-segments are obtained; the last period sub-segment may not reach the corresponding period length due to the end of the monitoring task, which is a normal situation, and the last period sub-segment can be discarded or retained according to the actual situation without affecting the similarity measurement between subsequent period sub-segments.

[0073] Preferably, in one embodiment of the present invention, considering that a dynamic time warping (DTW) algorithm can be used to evaluate the difference between different periodic subsegments, the method for obtaining the degree of instability includes:

[0074] In each monitoring curve, the mean of the DTW distances between all different period subsegments is normalized, and the normalized structure is used as the instability degree of each monitoring curve.

[0075] As an example, the instability degree is calculated as: ;in, For the Patients with history of respiratory monitoring in The instability of the monitoring curve under each monitoring indicator; is an S-type function; For the Patients with history of respiratory monitoring in The serial number of the periodic sub-segment in the monitoring curve under each monitoring indicator; For the Patients with history of respiratory monitoring in Among the monitoring curves under the monitoring indicators, except for the The serial numbers of other periodic sub-segments outside the periodic sub-segment; For the Patients with history of respiratory monitoring in In the monitoring curve under the monitoring indicators, the The period sub-segment and DTW distance between periodic sub-segments; For the Patients with history of respiratory monitoring in The total number of periodic subsegments in the monitoring curve under a monitoring indicator.

[0076] In the calculation formula of the instability, the DTW distance between every two different periodic sub-segments is averaged, and then the average is mapped to the S-type function for normalization. The larger the normalized value, the greater the difference between different periodic sub-segments in the monitoring curve, the weaker the periodic characteristics, and the greater the instability. In other examples, the implementer can also use other difference measurement methods such as Euclidean distance instead of DTW distance, and can also use other methods such as linear normalization for normalization, which are well-known technologies like DTW distance and will not be repeated here.

[0077] Considering that in each patient group, the fluctuation of the monitoring curve under each monitoring indicator should be similar, if the fluctuation of the monitoring curve of a respiratory monitoring patient deviates from the corresponding reference value of the group, it can be considered that the respiratory monitoring patient deviates from the group level of similar patient groups, and the monitoring curve can be considered to fluctuate abnormally; considering that there may be historical respiratory monitoring patients with abnormal monitoring in each patient group, that is, the instability of their monitoring curves is relatively high, and their corresponding monitoring curves have a low reference value for determining the group level of the fluctuation of the monitoring curve under each monitoring indicator in the patient group, and even affect the monitoring effect; therefore, in each patient group, under each monitoring indicator, the embodiment of the present invention obtains the fluctuation period and fluctuation range of the respiratory monitoring data according to the autocorrelation and fluctuation amplitude of all monitoring curves, combined with the instability of each monitoring curve; the fluctuation range and fluctuation period can be considered as the reference group level corresponding to the historical respiratory monitoring patient group with normal monitoring in the patient group.

[0078] Preferably, in one embodiment of the present invention, considering that the greater the instability, the more abnormal the respiratory monitoring result of the historical respiratory monitoring patient is, so it should be eliminated, so as to obtain a stable monitoring curve corresponding to all normal respiratory monitoring results; considering that the stable monitoring curve represents the group level corresponding to all normal respiratory monitoring results in the patient group, the maximum fluctuation range and average fluctuation period under each monitoring indicator can be further determined by the group level, which is used as a reference for respiratory monitoring; based on this, the method for obtaining the fluctuation period and fluctuation range includes:

[0079] Under each monitoring indicator, all stable monitoring curves are screened out from all corresponding monitoring curves in each patient group according to the degree of instability; in each patient group, under each monitoring indicator, the average of the periods of all stable monitoring curves is used to obtain the fluctuation period of respiratory monitoring data; in each patient group, under each monitoring indicator, the maximum monitoring data in all stable monitoring curves is used as the upper limit of the fluctuation range, and the minimum monitoring data in all stable monitoring curves is used as the lower limit of the fluctuation range to obtain the fluctuation range.

[0080] As a preferred example, in each patient group, under each monitoring indicator, all monitoring curves with an instability level less than a preset threshold are used as stable monitoring curves; the preset threshold is set to 0.5, and the implementer can also fine-tune it according to the actual situation, but it is necessary to ensure that in each patient group, under each monitoring indicator, a monitoring curve is retained as a stable monitoring curve to avoid the subsequent inability to obtain the fluctuation range and fluctuation cycle; and further obtain the fluctuation cycle and fluctuation range.

[0081] In other examples, the implementer may first determine the maximum monitoring data in each stable monitoring curve under each monitoring indicator in each patient group, and then use the mode or mean of the maximum monitoring data in all stable monitoring curves as the upper limit of the fluctuation range, and similarly obtain the lower limit, and then obtain the fluctuation range; the mode or median of the periods of all stable monitoring curves may also be used as the fluctuation period, which will not be elaborated here.

[0082] So far, the fluctuation range and fluctuation period of the respiratory monitoring data under each monitoring indicator in each patient group have been obtained.

[0083] The intelligent monitoring module 103 is used to determine a reference group of patients to be monitored from all patient groups based on the basic medical record information of the patients to be monitored; under each monitoring indicator, according to the fluctuation changes of the real-time monitoring data of the patients to be monitored, combined with the fluctuation period and fluctuation range of the respiratory monitoring data under each monitoring indicator in the reference group, a respiratory monitoring warning is given to the patients to be monitored.

[0084] Considering that the physical fitness of respiratory monitoring patients with similar basic medical record information may be similar, the corresponding respiratory monitoring data changes may also be similar, so the reference group of patients to be monitored can be determined from all patient groups based on the basic medical record information of the patients to be monitored. The reference group is a patient group with the most similar physical fitness to the patients to be monitored, so that it is convenient to use the fluctuation range and fluctuation cycle of the respiratory monitoring data under each monitoring indicator in each patient group in the future to carry out respiratory monitoring warning for the patients to be monitored.

[0085] Preferably, in one embodiment of the present invention, considering that the cluster center represents the average level of all data points in the cluster, if the basic medical record information of the patient to be monitored for respiratory infection is more similar to the basic medical record information corresponding to the cluster center, it means that the patient is more likely to belong to the patient group corresponding to the cluster; based on this, the method for obtaining the reference group includes:

[0086] The physical sign similarity coefficient between the patients to be monitored for respiratory failure and the corresponding cluster center of each patient group is used as the matching parameter; the patient group corresponding to the minimum matching parameter is used as the reference group of patients to be monitored for respiratory failure.

[0087] In other embodiments of the present invention, the implementer may also use the average of the physical sign similarity coefficients between the patient to be monitored for respiratory failure and all the patients with historical respiratory failure in each patient group as a matching parameter to screen the reference group.

[0088] It should be noted that the physical sign similarity coefficient between the patient to be monitored for respiratory distress and the corresponding cluster center of each patient group is consistent with the method described in steps S201-S204, and will not be described in detail here.

[0089] Preferably, in one embodiment of the present invention, the method for performing respiratory monitoring early warning on a patient undergoing respiratory monitoring comprises:

[0090] Under each monitoring index, the change cycle of the real-time monitoring data of the patient to be monitored for respiratory function is obtained based on the autocorrelation function;

[0091] If the real-time monitoring data of the patient to be monitored for respiratory failure under any monitoring indicator exceeds the fluctuation range of the respiratory monitoring data under the corresponding monitoring indicator in the corresponding reference group, or the period difference between the change period and the fluctuation period of the real-time monitoring data of the patient to be monitored for respiratory failure under any monitoring indicator is greater than the preset period threshold, a respiratory monitoring warning will be issued to the patient to be monitored for respiratory failure.

[0092] As an example, considering that the real-time monitoring data of the patient to be monitored for respiration is the monitoring curve collected up to the current moment, the variation period of the real-time monitoring data of the patient to be monitored for respiration under each monitoring index is firstly obtained based on the autocorrelation function, wherein the period of the curve obtained by the autocorrelation function is an existing technology well known to those skilled in the art and will not be described in detail here;

[0093] Then, based on the fluctuation range and fluctuation cycle of the respiratory monitoring data under each monitoring indicator in the reference group, the real-time monitoring data of the patient undergoing respiratory monitoring can be given an intelligent monitoring warning; the preset cycle threshold is set to 3, and the implementer can also set it by himself. Under any monitoring indicator, when the amplitude of the real-time monitoring data of the patient undergoing respiratory monitoring exceeds the fluctuation range, or when the deviation of the change cycle of the real-time monitoring data from the fluctuation cycle is greater than 3, if at least one of these two situations is met, an intelligent monitoring warning will be given to remind relevant medical personnel to conduct key analysis and evaluation, so as to formulate a new nursing plan to ensure the life safety of each patient undergoing respiratory monitoring.

[0094] In summary, the present invention obtains the basic medical record information of all respiratory monitoring patients and the monitoring curves under each monitoring index through the data acquisition module; in the data analysis module, all patient groups are obtained according to the similarity of the basic medical record information; the periodic characteristics of each monitoring curve are analyzed to obtain its instability; in each patient group, under each monitoring index, according to the autocorrelation and fluctuation amplitude of all monitoring curves, combined with the instability of each monitoring curve, the fluctuation period and fluctuation range of the respiratory monitoring data are obtained; and then the respiratory monitoring patients are treated with respiratory monitoring warning in the intelligent monitoring module. The present invention is based on the idea that different patients have different basic signs and that the fluctuation changes of the respiratory monitoring data of patients with similar basic signs are also similar. All patients are classified, and the fluctuation range and fluctuation period corresponding to the respiratory monitoring data under each monitoring index in each patient category are determined in combination with the periodic change characteristics of the normal respiratory monitoring data in each patient category, thereby providing a monitoring warning reference and improving the respiratory monitoring effect for patients with different physical signs.

[0095] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An intelligent medical system for clinical nursing respiratory monitoring, characterized in that: The system comprises: The data acquisition module is used to obtain the basic medical record information of all patients under respiratory monitoring, and obtain the monitoring curve of each patient under historical respiratory monitoring under each monitoring indicator, and the real-time monitoring data of patients under respiratory monitoring under each monitoring indicator; A data analysis module is used to classify all historical respiratory monitoring patients according to the similarity of the basic medical record information between different historical respiratory monitoring patients to obtain all patient groups; obtain all periodic sub-segments in the monitoring curve according to the autocorrelation of each monitoring curve; obtain the instability of each monitoring curve according to the similarity between different periodic sub-segments in each monitoring curve; in each patient group, under each monitoring indicator, according to the autocorrelation and fluctuation amplitude of all the monitoring curves, combined with the instability of each monitoring curve, obtain the fluctuation period and fluctuation range of the respiratory monitoring data; The intelligent monitoring module is used to determine a reference group of patients to be monitored from all patient groups based on the basic medical record information of the patients to be monitored; under each monitoring indicator, according to the fluctuation changes of the real-time monitoring data of the patients to be monitored, combined with the fluctuation period and the fluctuation range of the respiratory monitoring data under each monitoring indicator in the reference group, a respiratory monitoring warning is given to the patients to be monitored.

2. An intelligent medical system for clinical nursing respiratory monitoring according to claim 1, characterized in that: The basic medical record information includes surgical record data and outpatient physical examination data; the surgical record data at least includes the type of surgery and the anesthesia method; the outpatient physical examination data includes the basic physical sign parameters under each physical sign indicator and all previous medical history.

3. An intelligent medical system for clinical nursing respiratory monitoring according to claim 2, characterized in that: The method for obtaining the patient group comprises: According to the similarity of the operation record data between different historical respiratory monitoring patients and the similarity of the basic physical sign data, obtaining the physical sign similarity coefficient between different historical respiratory monitoring patients; All patients with historical respiratory monitoring are clustered based on the physical sign similarity coefficient, and the clustering results corresponding to all clusters are used as all patient groups.

4. The intelligent medical system for clinical nursing respiratory monitoring according to claim 3, characterized in that: The method for obtaining the physical sign similarity coefficient includes: A binary array is constructed according to the surgical type and anesthesia method of each historical respiratory monitoring patient, and the binary array is used as surgical record data; the Jaccard coefficient between the binary arrays of different historical respiratory monitoring patients is used as the surgical correlation between the corresponding different historical respiratory monitoring patients; According to the parameter difference between the basic physical sign parameters under all the same physical sign indicators, the physical sign consistency between the different historical respiratory monitoring patients is obtained; the parameter difference is negatively correlated with the physical sign consistency; Constructing a multivariate array according to all the past medical histories of each patient with historical respiratory monitoring, and using the Jaccard coefficient between the multivariate arrays of different patients with historical respiratory monitoring as the medical history correlation between the corresponding different patients with historical respiratory monitoring; A sign similarity coefficient is obtained according to the surgical correlation, the sign fit and the medical history correlation between patients with different historical respiratory monitoring, and the surgical correlation, the sign fit and the medical history correlation are all positively correlated with the sign similarity coefficient.

5. The intelligent medical system for clinical nursing respiratory monitoring according to claim 1, characterized in that: The method for obtaining the periodic sub-segment includes: The period of each monitoring curve is obtained based on the autocorrelation function, and the period is used as the length of each period subsegment in the corresponding monitoring curve. The period is divided iteratively from the starting point of each monitoring curve to obtain all period subsegments of each monitoring curve.

6. The intelligent medical system for clinical nursing respiratory monitoring according to claim 1, characterized in that: The method for obtaining the degree of instability includes: In each of the monitoring curves, the average values ​​of the DTW distances between all different period subsegments are normalized, and the normalized result is used as the instability degree of each of the monitoring curves.

7. The intelligent medical system for clinical nursing respiratory monitoring according to claim 5, characterized in that: The method for obtaining the fluctuation period and fluctuation range includes: Under each monitoring indicator, all stable monitoring curves are screened out from all the corresponding monitoring curves in each patient group according to the instability degree; In each of the patient groups, under each monitoring index, the average of the periods of all the stable monitoring curves is used to obtain the fluctuation period of the respiratory monitoring data; In each of the patient groups, under each monitoring index, the maximum monitoring data in all the stable monitoring curves is taken as the upper limit of the fluctuation range, and the minimum monitoring data in all the stable monitoring curves is taken as the lower limit of the fluctuation range to obtain the fluctuation range.

8. An intelligent medical system for clinical nursing respiratory monitoring according to claim 7, characterized in that: The method for obtaining the stable monitoring curve includes: In each of the patient groups, under each monitoring indicator, all the monitoring curves whose instability is less than a preset threshold are regarded as stable monitoring curves.

9. The intelligent medical system for clinical nursing respiratory monitoring according to claim 3, characterized in that: The method for obtaining the reference group includes: The physical sign similarity coefficient between the patient to be monitored for respiratory failure and the corresponding cluster center of each patient group is used as a matching parameter; the patient group corresponding to the minimum matching parameter is used as a reference group of the patient to be monitored for respiratory failure.

10. The intelligent medical system for clinical nursing respiratory monitoring according to claim 1, characterized in that: The method for performing respiratory monitoring early warning on a patient undergoing respiratory monitoring comprises: Under each monitoring index, the change cycle of the real-time monitoring data of the patient to be monitored for respiratory function is obtained based on the autocorrelation function; If the real-time monitoring data of the patient to be monitored for respiratory failure under any monitoring indicator exceeds the fluctuation range of the respiratory monitoring data under the corresponding monitoring indicator in the reference group, or the period difference between the change period and the fluctuation period of the real-time monitoring data of the patient to be monitored for respiratory failure under any monitoring indicator is greater than a preset period threshold, a respiratory monitoring warning will be issued to the patient to be monitored for respiratory failure.

Images