Construction method of infectious disease patient sequelae risk prediction model

Abstract

The invention relates to a construction method of an infectious disease patient sequelae risk prediction model. The method comprises the following steps: determining all associated indexes of infectious disease patient sequelae risk prediction; determining weight values of all the associated indexes, and selecting the associated indexes of which the ranking of the weight values is higher than a ranking threshold value from all the associated indexes as prediction index data; carrying out infectious disease patient sampling based on the prediction index data to obtain a sampling data set, wherein the infectious disease patient sampling refers to sampling of patients with sequelae and patients without sequelae in infectious disease patients; obtaining prediction index data of a to-be-predicted infectious disease patient; and performing sequelae risk prediction on the prediction index data of the to-be-predicted infectious disease patient based on the sampling data set to obtain a sequelae risk prediction result. According to the mode provided by the invention, on one hand, a patient can be prevented from frequently going to a hospital to see a doctor so as to save time cost of a user, and on the other hand, the calculated sequelae risk prediction result is more comprehensive and accurate.

Description

Technical Field

[0001] This invention relates to the technical field of risk prediction models, and in particular to a method for constructing a risk prediction model for sequelae of infectious disease patients. Background Technology

[0002] Sequelae of infectious diseases refer to the situation where the function of certain organs fails to return to normal for a long period of time after the recovery period of an infectious disease patient.

[0003] Currently, methods for predicting the risk of sequelae in patients with infectious diseases often require long-term follow-up and monitoring to promptly identify and address sequelae issues. This approach necessitates frequent hospital visits for consultations with doctors, which is essential for high-risk patients but unnecessary for low-risk patients.

[0004] Therefore, there are some methods on the market for predicting the risk of sequelae of infectious disease patients. However, most of these methods monitor some key indicators of patients and then issue warnings based on whether preset indicator thresholds are exceeded. This method has poor monitoring effect and cannot provide users with comprehensive and accurate risk prediction. Summary of the Invention

[0005] The purpose of this invention is to at least address one of the shortcomings of the prior art and to provide a method for constructing a risk prediction model for sequelae of infectious disease patients.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Specifically, a method for constructing a risk prediction model for sequelae of infectious disease patients is proposed, including the following: Identify all relevant indicators for predicting the risk of sequelae in patients with infectious diseases; Determine the weight values ​​of all related indicators, and select the related indicators whose weight values ​​rank higher than the ranking threshold from all related indicators as the predictive indicator data; The sampling data set is obtained by sampling infectious disease patients based on predictive index data. The sampling of infectious disease patients refers to sampling infectious disease patients with sequelae and those without sequelae. The sampling data set refers to the data set composed of the predictive index data of each sampled infectious disease patient. Obtain predictive indicator data for patients with infectious diseases to be predicted; Based on the sampled data set, the risk of sequelae is predicted for the predictive indicator data of the patients with the infectious disease to be predicted, and the sequelae risk prediction results are obtained.

[0007] Furthermore, specifically, determining the weight values ​​for all related indicators includes, If there are a total of m related indicators; Standardize the m types of correlation indicators to obtain the standardized m types of correlation indicators; The information entropy of each correlation indicator is calculated using the following formula. , in, The information entropy of the j-th correlation index is represented. Let k be the proportion of the standardized i-th correlation index on the j-th correlation index, where k is a preset constant. The weight of each correlation indicator is calculated based on its information entropy, as shown in the following formula. , in, This represents the weight of the j-th correlation index.

[0008] Furthermore, specifically, from all related indicators, the related indicators with a weight value ranking equal to or higher than the 4th place are selected as predictive indicator data, and are respectively denoted as the first predictive indicator, the second predictive indicator, the third predictive indicator, and the fourth predictive indicator according to their ranking.

[0009] Furthermore, specifically, based on the predicted indicator data of the patients with the infectious disease to be predicted from the sampled data set, the risk of sequelae is predicted, including: Let the set of sampled data be set Sp_set, and let mm be the number of elements in set Sp_set, which is the number of sampled infectious disease patients. In set Sp_set, Sp_set(mk) represents the relevant data of the sampled infectious disease patient with index mk, where mk∈[1, mm]. A three-dimensional spatial coordinate system is constructed using the second, third, and fourth prediction indicators as the X-axis, Y-axis, and Z-axis, respectively. Then the position coordinates of Sp_set(mk) in the three-dimensional coordinate system are x(mk), y(mk) and z(mk); Where x(mk) represents the coordinate of the element with index mk on the X-axis, y(mk) represents the coordinate of the element with index mk on the Y-axis, and z(mk) represents the coordinate of the element with index mk on the Z-axis, which is also the height of the element with index mk from the x-axis plane. In other words, z(mk) represents the plane formed by the target prediction points that are equidistant from the x-axis plane. Add the predictive index data of patients with infectious diseases to be predicted to the set Sp_set, and denote it as Sp_set(mk); At this point, we obtain the two elements with the smallest Euclidean distance to Sp_set(mk) in the three-dimensional coordinate system, and denote one of these elements as Sp_set(mk1) and the other as Sp_set(mk2). The coordinates of element Sp_set(mk1) are x(mk1), y(mk1), and z(mk1), and the coordinates of element Sp_set(mk2) are x(mk2), y(mk2), and z(mk2). Get the first prediction index corresponding to element Sp_set(mk1) as W(mk1), and get the first prediction index corresponding to element Sp_set(mk2) as W(mk2); The plane corresponding to z(mk1) is denoted as Pla(mk1), and the plane corresponding to z(mk2) is denoted as Pla(mk2). Obtain the set of elements falling on Pla(mk1) in the three-dimensional spatial coordinate system as Set(mk1), the number of elements in Set(mk1) is r1, the index of the elements in Set(mk1) is v1, v1∈[1 , r1], the element with index v1 in Set(mk1) is Set(mk1 ,v1), and the position coordinates corresponding to Set(mk1 ,v1) are x(mk1 ,v1), y(mk1 ,v1) and z(mk1 ,v1); Obtain the set of elements falling on Pla(mk2) in the three-dimensional spatial coordinate system as Set(mk2), the number of elements in Set(mk2) is r2, the index of the elements in Set(mk2) is v2, v2∈[1 , r2], the element with index v2 in Set(mk2) is Set(mk2 ,v2), and the position coordinates corresponding to Set(mk2 ,v2) are x(mk2 ,v2), y(mk2 ,v2) and z(mk2 ,v2); The sequelae association index seq_rel(mk) of Sp_set(mk) is calculated as follows: , in, (O,P) represents calculating the Euclidean distance between elements O and P; Get the elements in the sampled data set that are on the Pla(mk1) plane and calculate the mean of the first prediction index corresponding to each element as W(Pla(mk1)). Get the elements in the sampled data set that are on the Pla(mk2) plane and calculate the mean of the first prediction index corresponding to each element as W(Pla(mk2)). The formula for calculating the sequelae risk index W(mk) of Sp_set(mk) is as follows: W(mk)= [W(Pla(mk1))+ W(Pla(mk2))]* seq_rel(mk)*β / 2, Where β is the compensation coefficient, a value set manually; Next, it is determined whether W(mk) is greater than the preset sequelae risk index threshold. If so, the patient with the infectious disease to be predicted is judged to have a high risk of sequelae; otherwise, the patient with the infectious disease to be predicted is judged to have a low risk of sequelae.

[0010] Furthermore, the method also includes, For patients with infectious diseases at high risk of sequelae, they will be reminded to visit the hospital for a follow-up examination via SMS in the first pre-set cycle; for patients with infectious diseases at low risk of sequelae, they will be reminded via SMS in the second pre-set cycle to provide predictive indicator data again for a second verification.

[0011] The beneficial effects of this invention are as follows: This invention proposes a method for constructing a risk prediction model for sequelae of infectious disease patients. First, by weighting all relevant indicators for predicting the risk of sequelae of infectious disease patients, the most influential indicators are identified as prediction indicators. Then, using these prediction indicators, the risk of sequelae for the patients to be predicted is calculated through historical big data correlation analysis, yielding the predicted risk result. This method avoids frequent hospital visits, saving users' time, and provides a more comprehensive and accurate prediction of sequelae risk. Attached Figure Description

[0012] The above and other features of this disclosure will become more apparent from the detailed description of the embodiments illustrated in conjunction with the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same or similar elements. Obviously, the drawings described below are merely some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort. In the drawings: Figure 1 The diagram shows a flowchart of the method for constructing the risk prediction model for sequelae of infectious disease patients according to the present invention. Detailed Implementation

[0013] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The same reference numerals used throughout the accompanying drawings indicate the same or similar parts.

[0014] Example 1, referring to Figure 1 This invention proposes a method for constructing a risk prediction model for sequelae of infectious disease patients, comprising the following: Step 110: Identify all relevant indicators for predicting the risk of sequelae in patients with infectious diseases; Step 120: Determine the weight values ​​of all related indicators, and select the related indicators whose weight values ​​rank higher than the ranking threshold from all related indicators as the predictive indicator data. Step 130: Sampling of infectious disease patients based on predictive index data to obtain a sampling data set. The sampling of infectious disease patients refers to sampling of infectious disease patients with and without sequelae. The sampling data set refers to the data set composed of the predictive index data of each sampled infectious disease patient. Step 140: Obtain predictive indicator data for patients with infectious diseases to be predicted; Step 150: Based on the sampled data set, predict the risk of sequelae for the patients with infectious diseases to be predicted, and obtain the sequelae risk prediction results.

[0015] In this embodiment 1, firstly, by weighting all relevant indicators for predicting the risk of sequelae in patients with infectious diseases, the most influential indicators are identified as predictive indicators. Then, using these predictive indicators, the risk of sequelae in the patients to be predicted is calculated through historical big data correlation analysis, yielding the predicted risk result. The method proposed in this invention, on the one hand, avoids patients frequently visiting hospitals, saving users' time; on the other hand, the calculated risk prediction result is more comprehensive and accurate.

[0016] In a preferred embodiment of the present invention, specifically, determining the weight values ​​of all related indicators includes, If there are a total of m related indicators; Standardize the m types of correlation indicators to obtain the standardized m types of correlation indicators; The information entropy of each correlation indicator is calculated using the following formula. , in, The information entropy of the j-th correlation index is represented. Let k be the proportion of the standardized i-th correlation index on the j-th correlation index, where k is a preset constant. The weight of each correlation indicator is calculated based on its information entropy, as shown in the following formula. , in, This represents the weight of the j-th correlation index.

[0017] In this preferred embodiment, the weight values ​​of all related indicators can be accurately determined by the above method.

[0018] In a preferred embodiment of the present invention, specifically, the related indicators whose weight value ranks equal to or higher than the 4th place are selected from all related indicators as the prediction indicator data, and are respectively recorded as the first prediction indicator, the second prediction indicator, the third prediction indicator and the fourth prediction indicator according to their ranking.

[0019] In a preferred embodiment of the present invention, specifically, the prediction of sequelae risk is performed based on the predictive indicator data of the patients with the infectious disease to be predicted, according to the sampled data set, including: Let the set of sampled data be set Sp_set, and let mm be the number of elements in set Sp_set, which is the number of sampled infectious disease patients. In set Sp_set, Sp_set(mk) represents the relevant data of the sampled infectious disease patient with index mk, where mk∈[1, mm]. A three-dimensional spatial coordinate system is constructed using the second, third, and fourth prediction indicators as the X-axis, Y-axis, and Z-axis, respectively. Then the position coordinates of Sp_set(mk) in the three-dimensional coordinate system are x(mk), y(mk) and z(mk); Where x(mk) represents the coordinate of the element with index mk on the X-axis, y(mk) represents the coordinate of the element with index mk on the Y-axis, and z(mk) represents the coordinate of the element with index mk on the Z-axis, which is also the height of the element with index mk from the x-axis plane. In other words, z(mk) represents the plane formed by the target prediction points that are equidistant from the x-axis plane. Add the predictive index data of patients with infectious diseases to be predicted to the set Sp_set, and denote it as Sp_set(mk); At this point, we obtain the two elements with the smallest Euclidean distance to Sp_set(mk) in the three-dimensional coordinate system, and denote one of these elements as Sp_set(mk1) and the other as Sp_set(mk2). The coordinates of element Sp_set(mk1) are x(mk1), y(mk1), and z(mk1), and the coordinates of element Sp_set(mk2) are x(mk2), y(mk2), and z(mk2). Get the first prediction index corresponding to element Sp_set(mk1) as W(mk1), and get the first prediction index corresponding to element Sp_set(mk2) as W(mk2); The plane corresponding to z(mk1) is denoted as Pla(mk1), and the plane corresponding to z(mk2) is denoted as Pla(mk2). Obtain the set of elements falling on Pla(mk1) in the three-dimensional spatial coordinate system as Set(mk1), the number of elements in Set(mk1) is r1, the index of the elements in Set(mk1) is v1, v1∈[1 , r1], the element with index v1 in Set(mk1) is Set(mk1 ,v1), and the position coordinates corresponding to Set(mk1 ,v1) are x(mk1 ,v1), y(mk1 ,v1) and z(mk1 ,v1); Obtain the set of elements falling on Pla(mk2) in the three-dimensional spatial coordinate system as Set(mk2), the number of elements in Set(mk2) is r2, the index of the elements in Set(mk2) is v2, v2∈[1 , r2], the element with index v2 in Set(mk2) is Set(mk2 ,v2), and the position coordinates corresponding to Set(mk2 ,v2) are x(mk2 ,v2), y(mk2 ,v2) and z(mk2 ,v2); The sequelae association index seq_rel(mk) of Sp_set(mk) is calculated as follows: , in, (O,P) represents calculating the Euclidean distance between elements O and P; Get the elements in the sampled data set that are on the Pla(mk1) plane and calculate the mean of the first prediction index corresponding to each element as W(Pla(mk1)). Get the elements in the sampled data set that are on the Pla(mk2) plane and calculate the mean of the first prediction index corresponding to each element as W(Pla(mk2)). The formula for calculating the sequelae risk index W(mk) of Sp_set(mk) is as follows: W(mk)= [W(Pla(mk1))+ W(Pla(mk2))]* seq_rel(mk)*β / 2, Where β is the compensation coefficient, a value set manually; Next, it is determined whether W(mk) is greater than the preset sequelae risk index threshold. If so, the patient with the infectious disease to be predicted is judged to have a high risk of sequelae; otherwise, the patient with the infectious disease to be predicted is judged to have a low risk of sequelae.

[0020] In this preferred embodiment, the contribution of indicator data is considered first. The related indicators with weight values ​​higher than the top four are selected from all related indicators as the predictive indicator data. Then, the coefficients of the second, third, and fourth predictive indicators are transformed using the Euclidean distance method. Finally, the product of these coefficients with the first predictive indicator is obtained to obtain the sequelae risk index. This not only utilizes the high-weight predictive indicators, but also uses Euclidean distance to calculate the sequelae risk of infectious disease patients using a more relevant historical data set, making the obtained sequelae risk index more accurate.

[0021] In a preferred embodiment of the present invention, the method further includes, For patients with infectious diseases at high risk of sequelae, they will be reminded to visit the hospital for a follow-up examination via SMS in the first pre-set cycle; for patients with infectious diseases at low risk of sequelae, they will be reminded via SMS in the second pre-set cycle to provide predictive indicator data again for a second verification.

[0022] In this preferred embodiment, by checking the risk of sequelae of patients with infectious diseases in the above manner, it is possible to conduct follow-up examinations and early warnings for users with corresponding sequelae risks in an appropriate manner, wherein the first cycle is shorter than the second cycle.

[0023] Although the description of the invention has been quite detailed and particularly of several described embodiments, it is not intended to limit it to any of these details or embodiments or any particular embodiment, but should be considered as providing a broad possible interpretation of the claims by referring to the appended claims and taking into account the prior art, thereby effectively covering the intended scope of the invention. Furthermore, the invention has been described above with respect to embodiments foreseeable by the inventors in order to provide a useful description, and non-substantial modifications to the invention that have not yet been foreseen may still represent equivalent modifications.

[0024] The above description is merely a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. Any embodiment that achieves the technical effects of the present invention using the same means should fall within the protection scope of the present invention. Within the protection scope of the present invention, various modifications and variations can be made to the technical solutions and / or implementation methods.

Claims

1. A method for constructing a risk prediction model for sequelae of infectious disease patients, characterized in that, Including the following: Identify all relevant indicators for predicting the risk of sequelae in patients with infectious diseases; Determine the weight values ​​of all related indicators, and select the related indicators whose weight values ​​rank higher than the ranking threshold from all related indicators as the predictive indicator data; The sampling data set is obtained by sampling infectious disease patients based on predictive index data. The sampling of infectious disease patients refers to sampling infectious disease patients with sequelae and those without sequelae. The sampling data set refers to the data set composed of the predictive index data of each sampled infectious disease patient. Obtain predictive indicator data for patients with infectious diseases to be predicted; Based on the sampled data set, the risk of sequelae is predicted for the predictive indicator data of the patients with the infectious disease to be predicted, and the sequelae risk prediction results are obtained.

2. The method for constructing a risk prediction model for sequelae of infectious disease patients according to claim 1, characterized in that, Specifically, determining the weight values ​​for all related indicators includes, If there are a total of m related indicators; Standardize the m types of correlation indicators to obtain the standardized m types of correlation indicators; The information entropy of each correlation indicator is calculated using the following formula. ， in, The information entropy of the j-th correlation index is represented. Let k be the proportion of the standardized i-th correlation index on the j-th correlation index, where k is a preset constant. The weight of each correlation indicator is calculated based on its information entropy, as shown in the following formula. ， in, This represents the weight of the j-th correlation index.

3. The method for constructing a risk prediction model for sequelae of infectious disease patients according to claim 2, characterized in that, Specifically, from all related indicators, the related indicators with a weight value ranking equal to or higher than the 4th place are selected as the predictive indicator data, and are respectively recorded as the first predictive indicator, the second predictive indicator, the third predictive indicator, and the fourth predictive indicator according to their ranking.

4. The method for constructing a risk prediction model for sequelae of infectious disease patients according to claim 3, characterized in that, Specifically, based on the sampled data set, the risk of sequelae is predicted for the predictive indicator data of patients with the infectious disease to be predicted, including: Let the set of sampled data be set Sp_set, and let mm be the number of elements in set Sp_set, which is the number of sampled infectious disease patients. In set Sp_set, Sp_set(mk) represents the relevant data of the sampled infectious disease patient with index mk, where mk∈[1, mm]. A three-dimensional spatial coordinate system is constructed using the second, third, and fourth prediction indicators as the X-axis, Y-axis, and Z-axis, respectively. Then the position coordinates of Sp_set(mk) in the three-dimensional coordinate system are x(mk), y(mk) and z(mk); Where x(mk) represents the coordinate of the element with index mk on the X-axis, y(mk) represents the coordinate of the element with index mk on the Y-axis, and z(mk) represents the coordinate of the element with index mk on the Z-axis, which is also the height of the element with index mk from the x-axis plane. In other words, z(mk) represents the plane formed by the target prediction points that are equidistant from the x-axis plane. Add the predictive index data of patients with infectious diseases to be predicted to the set Sp_set, and denote it as Sp_set(mk); At this point, we obtain the two elements with the smallest Euclidean distance to Sp_set(mk) in the three-dimensional coordinate system, and denote one of these elements as Sp_set(mk1) and the other as Sp_set(mk2). The coordinates of element Sp_set(mk1) are x(mk1), y(mk1), and z(mk1), and the coordinates of element Sp_set(mk2) are x(mk2), y(mk2), and z(mk2). Get the first prediction index corresponding to element Sp_set(mk1) as W(mk1), and get the first prediction index corresponding to element Sp_set(mk2) as W(mk2); The plane corresponding to z(mk1) is denoted as Pla(mk1), and the plane corresponding to z(mk2) is denoted as Pla(mk2). Obtain the set of elements falling on Pla(mk1) in the three-dimensional spatial coordinate system as Set(mk1), the number of elements in Set(mk1) is r1, the index of the elements in Set(mk1) is v1, v1∈[1 , r1], the element with index v1 in Set(mk1) is Set(mk1 ,v1), and the position coordinates corresponding to Set(mk1 ,v1) are x(mk1 ,v1), y(mk1 ,v1) and z(mk1 ,v1); Obtain the set of elements falling on Pla(mk2) in the three-dimensional spatial coordinate system as Set(mk2), the number of elements in Set(mk2) is r2, the index of the elements in Set(mk2) is v2, v2∈[1 , r2], the element with index v2 in Set(mk2) is Set(mk2 ,v2), and the position coordinates corresponding to Set(mk2 ,v2) are x(mk2 ,v2), y(mk2 ,v2) and z(mk2 ,v2); The sequelae association index seq_rel(mk) of Sp_set(mk) is calculated as follows: ， in, (O,P) represents calculating the Euclidean distance between elements O and P; Get the elements in the sampled data set that are on the Pla(mk1) plane and calculate the mean of the first prediction index corresponding to each element as W(Pla(mk1)). Get the elements in the sampled data set that are on the Pla(mk2) plane and calculate the mean of the first prediction index corresponding to each element as W(Pla(mk2)). The formula for calculating the sequelae risk index W(mk) of Sp_set(mk) is as follows: W(mk)= [W(Pla(mk1))+ W(Pla(mk2))]* seq_rel(mk)*β / 2, Where β is the compensation coefficient, a value set manually; Then, it is determined whether W(mk) is greater than the preset sequelae risk index threshold; if so, it is determined that the patient with the infectious disease to be predicted has a high risk of sequelae; if not, it is determined that the patient with the infectious disease to be predicted has a low risk of sequelae.

5. The method for constructing a risk prediction model for sequelae of infectious disease patients according to claim 4, characterized in that, The method also includes, For patients with infectious diseases who have a high risk of sequelae, they will be reminded by SMS to visit the hospital for follow-up examinations according to the pre-set first cycle; For patients with infectious diseases who have a low risk of sequelae, they will be reminded via SMS to provide predictive indicator data again for a second verification according to the pre-set second cycle.

Images