KR20220102737A - Method And Computer Program for Predicting Recurrence Risk of Acute Coronary Syndrome - Google Patents

Abstract

The present disclosure provides a method and a computer program for predicting the recurrence risk of acute coronary syndrome. According to an aspect of the present disclosure, a method for predicting the recurrence risk of acute coronary syndrome and a computer program thereof are provided. The method comprises the steps of: collecting sample data on a patient group and a control group with recurrence of the acute coronary syndrome; performing logistic regression analysis to predict the risk of recurrence of the acute coronary syndrome based on the sample data; generating a machine learning-based logistic regression model which predicts the recurrence risk of the acute coronary syndrome of the patient based on the results of the logistic regression analysis, and performing learning based on the sample data. According to the present invention, rehabilitation is performed adaptively according to the degree of risk of recurrence of the patient.

Description

Method And Computer Program for Predicting Recurrence Risk of Acute Coronary Syndrome

The present invention relates to a method and computer program for predicting the risk of recurrence of acute coronary syndrome.

The content described in this section merely provides background information for the present embodiment and does not constitute the prior art.

Cardiac rehabilitation is a holistic treatment and follow-up management for heart disease patients, to restore cardiopulmonary function and motor function weakened by heart disease, and to prevent recurrence of the disease by managing risk factors for heart disease. Studies have shown that cardiac rehabilitation restores the patient's exercise capacity, reduces the need for heart disease recurrence, readmission, and reoperation, and reduces cardiac cause and all-cause mortality. In addition, it has the effect of reducing the recurrence rate of cardiovascular disease by 15-20%, as well as reducing the cardiovascular-related morbidity and mortality by 25% to 40%.

However, the lack of awareness and awareness of the need for cardiac rehabilitation, the lack of access to hospitals capable of cardiac rehabilitation and limited insurance coverage, the lack of time to participate in cardiac rehabilitation, the need for a guardian to take care of the patient, Participation in cardiac rehabilitation is low due to difficulties in going to hospital due to physical discomfort. For example, the participation rate of cardiac rehabilitation of patients with heart disease in Korea remains at 5% of all heart disease patients, and this phenomenon is not limited to patients with heart disease.

In order to make the patient aware of the need for rehabilitation and to improve the patient's awareness of rehabilitation, it is necessary to predict the risk of disease recurrence, furthermore, the probability of the disease recurrence risk, and use it to establish the patient's rehabilitation plan. Patients and medical staff, recognizing the risk of disease recurrence as a specific numerical value, can more actively work to prevent disease recurrence and respond to the degree of disease recurrence probability.

Accordingly, a method of analyzing the occurrence of disease by performing logistic regression analysis of various disease-related factors has been considered.

Logistic regression analysis is a statistical technique that estimates the causal relationship between the dependent and independent variables using a logistic function. Unlike linear regression analysis, logistic regression analysis analyzes the relationship between the dependent variable and the independent variable based on the probability of the value, not based on the value of the factor itself. Accordingly, when performing logistic regression analysis, there is an effect of preventing distortion due to outliers, but only performing simple logistic regression analysis can predict only the odds ratio for each factor or the recurrence of disease. There is a problem in that it is not possible to provide even specific numerical information about the probability of recurrence of the disease to the patients in need.

According to an aspect of the present disclosure, sample data about a patient group and a control group in which a disease has recurred is collected, and logistic regression analysis for predicting the risk of disease recurrence based on the sample data is performed, and logistic regression analysis is performed. To provide a method for predicting disease recurrence risk probability based on the result of has the main purpose of

According to one aspect of the present disclosure, with respect to a preset factor (factor), data about the factor in the patient group with recurrent acute coronary syndrome and sample data that is data about the factor of the control group are collected. process; performing logistic regression analysis to predict the recurrence risk of the acute coronary syndrome based on the sample data; Creating a machine learning-based logistic regression model for predicting the recurrence risk probability of the acute coronary syndrome based on the result of the logistic regression analysis, and learning the logistic regression model based on the sample data It provides a method for predicting the risk of recurrence of acute coronary syndrome, characterized in that it includes a process.

According to another aspect of the present disclosure, in the method for predicting the risk of recurrence of acute coronary syndrome, the control group is acute coronary syndrome, characterized in that it is composed of a group of people who have experienced the acute coronary syndrome once. provides a method for predicting the risk of recurrence of

According to another aspect of the present disclosure, in the aforementioned method for predicting the risk of recurrence of acute coronary syndrome, the verification of the logistic regression model is performed by determining whether an Area Under Curve (AUC) value is greater than or equal to a preset test level. It provides a method for predicting the risk of recurrence of acute coronary syndrome characterized by

According to another aspect of the present disclosure, the recurrence of acute coronary syndrome, characterized in that it further comprises the process of predicting the probability of recurrence of the disease from individual data of the patient regarding the factor using the learned logistic regression model. It provides a risk prediction method.

According to an aspect of the present disclosure, sample data about a patient group and a control group in which a disease has recurred is collected, and logistic regression analysis for predicting the risk of disease recurrence based on the sample data is performed, and logistic regression analysis is performed. By creating a machine learning-based logistic regression model that predicts the probability of disease recurrence based on the results of .

According to another aspect of the present disclosure, as the disease recurrence risk probability is predicted from individual patient data using the learned logistic regression model, the predicted disease recurrence risk probability is reflected in the patient's exercise prescription and / or disease management. It has the effect of adaptively performing rehabilitation according to the degree of the patient's risk of recurrence.

Therefore, when the method for predicting the risk of recurrence of a disease according to various aspects of the present disclosure is applied, there is an effect that allows the patient to increase the understanding of the risk of recurrence of the disease and to improve the compliance with the exercise prescription.

1 is a flowchart illustrating a method for predicting a recurrence risk of a disease according to an embodiment of the present disclosure.
2 is a statistical test result for predicting the probability of recurrence risk of acute coronary syndrome according to an embodiment of the present disclosure.

Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. It should be noted that in adding the reading reference numerals to the components of each drawing, the same components are to have the same reference numerals as much as possible even though they are displayed on different drawings. In addition, in describing the present disclosure, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present disclosure, the detailed description thereof will be omitted.

In addition, in describing the components of the present disclosure, terms such as second, first, etc. may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, or order of the elements are not limited by the terms. Throughout the specification, when a part 'includes' or 'includes' a certain component, this means that other components may be further included, rather than excluding other components, unless otherwise stated. . In addition, the '... Terms such as 'unit' and 'module' mean a unit that processes at least one function or operation, which may be implemented as hardware or software or a combination of hardware and software.

DETAILED DESCRIPTION The detailed description set forth below in conjunction with the appended drawings is intended to describe exemplary embodiments of the present disclosure and is not intended to represent the only embodiments in which the present disclosure may be practiced.

In the present disclosure, as a method for predicting the risk of recurrence of a disease, an example of predicting the risk of recurrence of acute coronary syndrome is described, but the method for predicting the risk of recurrence of the present disclosure is not necessarily limited to this use. In the present disclosure, in order to present the performance of the logistic regression model according to the present disclosure in predicting the risk of disease recurrence, the risk of disease recurrence is predicted based on the data of the acute coronary syndrome recurrence patient group and the control group. However, the recurrence risk prediction method of the present disclosure can be applied to any field requiring disease recurrence risk prediction.

1 is a flowchart illustrating a method for predicting a recurrence risk of a disease according to an embodiment of the present disclosure.

Sample data is generated or collected from the patient group and the control group in which the disease recurs (S100). Here, the sample data is data about each group, and is a factor value or category data that can be used as an independent variable in logistic regression analysis. Factors include, for example, age, gender, procedure code, prescription medicine code, diagnosis code, and treated disease code, as shown in Table 1, for example. etc. may be included, and in addition, various variables that can be used to predict the risk of disease recurrence may be further included.

In order to more accurately predict the risk of recurrence of the disease, the control group from which data will be collected is preferably from patients who have had the disease once.

variable Explanation Age age Gender gender PRO_CODE Procedure code (e.g., Percutaneous coronary intervention, etc.) MED_CODE prescription drug code
(eg Clopidogrel, etc.) CON_CODE diagnostic code
(eg, acute myocardial infarction, etc.) PRO_REASONCODE disease code treated
(eg Angina pectoris, etc.)

The sample data is pre-processed (S102). Preprocessing includes, for example, data scaling, generation of specific variable values by data processing, unification of data format or type, categorization of data, data merging, processing of missing values, analysis tools (analysis tool) may include data processing into a form that can be processed (eg, converting a string into a number, etc.), data shuffling, and the like.

A logistic regression analysis for predicting whether a disease will recur using the sample data is performed (S104). The logistic regression analysis may be performed using all or part of the sample data, and is preferably performed using training data to train the logistic regression model in step S106. This is because a logistic regression model is created based on the logistic regression analysis result.

Logistic regression analysis may be performed on all factors of the sample data, but may be performed on only some factors, and the composition of factors for which logistic regression analysis is performed may vary.

By performing step S104, regression coefficients, odds ratios, confidence intervals, and the like corresponding to each factor may be calculated as a result of logistic regression analysis. Here, the odds ratio is a value representing the probability that an event will occur versus the probability that the event does not occur, and the relationship between the factor and the predicted value may be analyzed based on the odds ratio or the log-odds ratio obtained by taking the logarithm of the odds ratio.

A logistic regression model is generated using all or a part of regression coefficients corresponding to each factor of the sample data (S106). The regression coefficient may be an estimated value generated as a result of performing the regression analysis in step S104. Each factor serving as an independent variable of the logistic regression model may be a factor selected in consideration of a ratio between odds before and after a change in a value of a specific factor.

In another embodiment, in the logistic regression analysis, when 1 is included in the confidence interval or the odds ratio is less than or equal to a preset reference value, the value of each regression coefficient or constant calculated by the logistic regression analysis is adjusted, or the sample data used for the analysis is sampled again. Therefore, it is possible to improve the predictive performance of the conglomerate risk of disease by re-performing the logistic regression analysis.

The logistic regression model in the present disclosure is a machine learning-based model. The machine learning methodology used in the logistic regression model may be a decision tree, a random forest, a K-Nearest Neighbor (K-NN), a support vector machine (SVM), etc., but is not limited thereto, and is usually Any machine learning methodology that can be easily employed by those skilled in the art to implement a logistic regression model for predicting the risk of recurrence of a disease may be used for the implementation of the logistic regression model of the present disclosure.

Each of training data and/or test data is extracted from the sample data, and a logistic regression model for predicting the probability of recurrence of a disease is trained using the training data (S108). As described above, the process of extracting training data from the sample data may be performed before step S104.

In order to test the logistic regression model trained in step S108, the prediction quality of the logistic regression model is evaluated by using an area under curve (AUC) value to test ( S110 ). AUC is the area under the ROC curve (Receiver Operating Characteristic curve) with the true positive rate as the vertical axis and the false positive rate as the horizontal axis. The AUC value is a value representing the ratio of the area under the ROC curve to the total area. AUC is an invariant value in the evaluation scale, and is an invariant value in which classification threshold or cutoff is selected for the ROC curve. The test in step S108 may be performed, for example, by determining whether the AUC value is equal to or greater than a preset test level.

Another way to evaluate the prediction quality of a logistic regression model is to evaluate the goodness-of-fit of the model, and whether it is a model that minimizes the log loss obtained from the loss function. can be considered.

In order to predict the probability of disease recurrence risk from individual data of patients who have suffered the disease, the probability of disease recurrence risk is predicted using a logistic regression model in any one of steps S106 to S108 ( S112 ).

Although it is described that each process is sequentially executed in FIG. 1 , this is merely illustrative of the technical idea of an embodiment of the present disclosure. In other words, those of ordinary skill in the art to which an embodiment of the present disclosure pertain may change the order described in FIG. Since it may be applied by various modifications and variations to be executed in parallel, it is not limited to the time-series order of FIG. 1 . For example, the data preprocessing process of step S102 or the model testing process of step S110 may be omitted, and steps S104 and S106 may be performed as one step to generate a fit logistic regression model at the same time as logistic regression analysis. have.

Hereinafter, the probability of the recurrence risk of acute coronary syndrome is predicted using the disease recurrence risk prediction method according to an embodiment of the present disclosure, and a statistical test result of the predicted probability will be described. The statistical test result of the predicted probability is shown in FIG. 2 .

2 is a statistical test result for predicting the probability of recurrence risk of acute coronary syndrome according to an embodiment of the present disclosure.

In order to predict the risk of recurrence of acute coronary syndrome, age, sex, procedure code, prescription drug code were obtained from the patient group with recurrence of acute coronary syndrome and the control group of subjects who had one occurrence of acute coronary syndrome (eg, acute myocardial infarction syndrome). , diagnostic codes and treated disease codes were collected, respectively. Each data was obtained from the patient's hospital medical record and an external database (inpatient dataset of Health Insurance Review and Assessment Service and Synthea's Synthetic patient records), and the number of data obtained is shown in Table 2.

data type recurrent patient group control total number of people hospital medical records 0 4 4 Health Insurance Review and Assessment Service 703 5828 6531 Synthea 0 778 778 gun 703 6610 7313

In predicting the recurrence risk of acute coronary syndrome, 70% of the total collected data was used as training data and 30% was used as test data.

Referring to the ROC curve of FIG. 2 , it can be seen that the predicted true positive rate is higher than the false positive rate, and the AUC value is 0.821. Since the AUC value is 0.821, it was verified that the predictive performance of the logistic regression model for predicting the probability of the recurrence risk of acute coronary syndrome learned by the recurrence risk prediction method of an embodiment of the present disclosure is excellent.

When individual data of patients with actual acute coronary syndrome relapse are applied to the logistic regression model, as shown in Table 3, the probability of recurrence and non-recurrence of patients with actual acute coronary syndrome, respectively, is 0.999 (omitted below) and It can be confirmed that the prediction performance for individual data of the logistic regression model is also excellent by predicting each of 0.172 (hereinafter omitted).

Predict result : 1 with probability of : 0.9999999999996334
Predict result : 0 with probability of : 0.17207951877166552

Various implementations of the devices, units, processes, steps, etc., described herein may include digital electronic circuits, integrated circuits, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), computer hardware, firmware, software, and/or a combination thereof. These various implementations may include being implemented in one or more computer programs executable on a programmable system. The programmable system includes at least one programmable processor (which may be a special purpose processor) coupled to receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device. or may be a general-purpose processor). Computer programs (also known as programs, software, software applications or code) contain instructions for a programmable processor and are stored on a "computer-readable recording medium".

The computer-readable recording medium includes all types of recording devices in which data readable by a computer system is stored. These computer-readable recording media are non-volatile or non-transitory, such as ROM, CD-ROM, magnetic tape, floppy disk, memory card, hard disk, magneto-optical disk, and storage device. It may further include a medium or a transitory medium such as a data transmission medium. In addition, the computer-readable 　 recording medium may be distributed in a network-connected computer system, and the computer-readable code may be stored and executed in a distributed manner.

Various implementations of the systems and techniques described herein may be implemented by a programmable computer. Here, the computer includes a programmable processor, a data storage system (including volatile memory, non-volatile memory, or other types of storage systems or combinations thereof), and at least one communication interface. For example, the programmable computer may be one of a server, a network appliance, a set-top box, an embedded device, a computer expansion module, a personal computer, a laptop, a Personal Data Assistant (PDA), a cloud computing system, or a mobile device.

The above description is merely illustrative of the technical idea of this embodiment, and a person skilled in the art to which this embodiment belongs may make various modifications and variations without departing from the essential characteristics of the present embodiment. Accordingly, the present embodiments are intended to explain rather than limit the technical spirit of the present embodiment, and the scope of the technical spirit of the present embodiment is not limited by these embodiments. The protection scope of this embodiment should be interpreted by the claims below, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present embodiment.

Claims (9)

With respect to a preset factor, the process of collecting sample data, which is data on the factor in a patient group with recurrent acute coronary syndrome and data on the factor in a control group;
performing logistic regression analysis to predict the recurrence risk of the acute coronary syndrome based on the sample data;
Generating a machine learning-based logistic regression model for predicting the recurrence risk probability of the acute coronary syndrome based on the result of the logistic regression analysis, and learning the logistic regression model based on the sample data process
A method for predicting the risk of recurrence of acute coronary syndrome, characterized in that it comprises a. According to claim 1,
The control group, the acute coronary syndrome recurrence risk prediction method, characterized in that consisting of a group of patients who have had the acute coronary syndrome once. According to claim 1,
The factor, age, sex, procedure code, prescription drug code, diagnosis code, and a method for predicting recurrence risk of acute coronary syndrome, characterized in that it includes all or a part of the disease code treated. According to claim 1,
The result of the logistic regression analysis, each regression coefficient (regression coefficient) of the factor, an odds ratio (odds ratio) of all or part of the recurrence risk prediction method of acute coronary syndrome characterized in that it includes. 5. The method of claim 4,
The process of training the logistic regression model is,
A method for predicting recurrence risk of acute coronary syndrome, characterized in that the logistic regression analysis is re-performed so that the odds ratio is equal to or greater than a preset reference value, and the logistic regression model is generated based on the re-performed result. According to claim 1,
The learning process includes extracting training data and test data from the sample data, training the logistic regression model with the training data, and verifying the trained logistic regression model with test data.
A method for predicting the risk of recurrence of acute coronary syndrome, characterized by 7. The method of claim 6,
The verification is
A method for predicting the risk of recurrence of acute coronary syndrome, characterized in that it is performed by determining whether the Area Under Curve (AUC) value is greater than or equal to a preset test level. According to claim 1,
The process of predicting the recurrence risk probability of the acute coronary syndrome from individual data of the patient regarding the factor using a learned logistic regression model
A method for predicting the risk of recurrence of acute coronary syndrome, characterized in that it further comprises a. A computer program stored in a computer-readable recording medium to execute each process included in the method for predicting the recurrence risk of acute coronary syndrome according to any one of claims 1 to 8.

Images