Method and system for predicting dialysis timing in patients with chronic kidney disease stage 5 combined with type 2 diabetes mellitus

The dialysis timing prediction model is established through a multi-factor logistic regression method based on traditional Chinese medicine symptoms, and the problem of failure to effectively consider the clinical symptoms of patients in the existing technology is solved, and the accurate prediction of the dialysis timing of patients with stage 5 and type 2 diabetes is achieved. , reduces the uncertainty of doctors' subjective experience and provides efficient and non-invasive treatment suggestions.

CN117727459BActive Publication Date: 2024-11-01THE FIRST MEDICAL CENT CHINESE PLA GENERAL HOSPITAL
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410179365.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-02-18
Publication Date
2024-11-01
Estimated Expiration
2044-02-18

AI Technical Summary

Technical Problem

When predicting the dialysis timing of patients with stage 5 combined with type 2 diabetes in the prior art, the lack of considering the subjective feelings and clinical symptoms of the patients, resulting in greater uncertainty in the selection of the optimal dialysis timing.

Method used

Based on the patient's Chinese medicine symptoms, we have collected modeling indicators such as gender, disease location heart, Yin deficiency syndrome, water stop syndrome, kidney and phlegm syndrome, and established a dialysis timing prediction model, and used multi-factor logistic regression method for training to obtain Dialysis probability and determine the dialysis timing based on the preset threshold.

Benefits of technology

It provides more accurate and specific prediction of dialysis timing, reduces the uncertainty of doctors' subjective experience, is non-invasive, non-invasive, intuitive, simple and low operating cost, and can provide effective suggestions for dialysis treatment timing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117727459B_ABST
    Figure CN117727459B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for predicting the timing of dialysis for chronic kidney disease stage 5 combined with type 2 diabetes, belonging to the field of medical care information technology. The method comprises the following steps: according to modeling indicators, collecting characteristics of patients who have not yet undergone dialysis, the modeling indicators include: gender, diseased heart, Yin deficiency syndrome, water retention syndrome, diseased kidney and phlegm turbidity syndrome; obtaining a dialysis timing prediction model; predicting the patient characteristics through the dialysis timing prediction model to obtain a dialysis probability; obtaining the dialysis timing according to the dialysis probability and a preset first threshold. According to the gender, diseased heart, Yin deficiency syndrome, water retention syndrome, diseased kidney and phlegm turbidity syndrome of patients who have not yet started dialysis, and the dialysis timing prediction model, the dialysis timing is obtained, providing a basis for the timing of dialysis treatment, and providing effective suggestions for the doctor's treatment prescription; reducing the uncertainty caused by the doctor's subjective experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of medical care information technology, and in particular to a method and system for predicting the dialysis timing of chronic kidney disease stage 5 combined with type 2 diabetes. Background Art

[0002] Chronic kidney disease stage 5 combined with type 2 diabetes refers to chronic kidney disease stage 5 (CKD5) combined with type 2 diabetes, which is equivalent to end-stage kidney disease. Compared with renal failure, it has developed to the end stage and the condition is more serious. The conventional treatment methods for CKD5 are hemodialysis, peritoneal dialysis, etc.

[0003] The current research directions for dialysis are mainly: predicting the future physiological characteristics and risk of complications of patients after dialysis (CN117316457A); predicting the mortality rate of hemodialysis patients (CN115547502A). These research directions do not take into account the subjective feelings of patients, etc.; before patients start dialysis, they do not distinguish which CKD5 patients need dialysis and which CKD5 patients do not need dialysis for the time being. After CKD5 patients are diagnosed, some patients do not need dialysis immediately.

[0004] At present, whether to perform dialysis on a patient depends mainly on clinical indicators such as the patient's glomerular filtration rate, some of the patient's symptoms, and the doctor's subjective experience. No clinical studies have yet proven that relying solely on clinical indicators such as glomerular filtration rate to determine the timing of dialysis is beneficial to the patient's prognosis. Therefore, there is great uncertainty in the selection of the best time for dialysis. A large number of clinical studies have shown that referring to clinical indicators while referring to the patient's clinical symptoms before deciding when to start dialysis is more beneficial to the patient's prognosis. Therefore, it is necessary to design a method, system, and device for predicting the timing of dialysis based on the clinical symptoms of patients with CKD stage 5 and type 2 diabetes. Summary of the invention

[0005] In view of the above-mentioned technical problems existing in the prior art, the present invention provides a method and system for predicting the timing of dialysis for chronic kidney disease stage 5 combined with type 2 diabetes, which predicts the timing of dialysis for the patient based on the patient's traditional Chinese medicine symptoms.

[0006] The first aspect of the present invention discloses a method for predicting the timing of dialysis for chronic kidney disease stage 5 combined with type 2 diabetes, comprising the following steps: according to modeling indicators, collecting characteristics of patients who have not yet undergone dialysis, the modeling indicators include: gender, disease location heart, yin deficiency syndrome, water retention syndrome, disease location kidney and phlegm turbidity syndrome; obtaining a dialysis timing prediction model; predicting the patient characteristics through the dialysis timing prediction model to obtain a dialysis probability; and obtaining the dialysis timing based on the dialysis probability and a preset first threshold.

[0007] Preferably, the diseased heart includes any one of the following symptoms: palpitations, insomnia, dreaminess and forgetfulness;

[0008] The Yin deficiency syndrome includes any two of the following symptoms: dry throat and mouth, fever in the five parts of the body, fear of heat and sweating or night sweats, yellow urine and dry stool, red tongue with little coating, and weak pulse;

[0009] The water retention syndrome includes any two of the following symptoms: edema, pleural effusion, ascites, and heaviness of limbs;

[0010] The kidney disease includes any one of the following symptoms: soreness of waist and knees, tinnitus and deafness, hair loss and loose teeth, and frequent urination at night;

[0011] The phlegm and turbidity syndrome includes any two of the following symptoms: nausea and vomiting, spitting or coughing up phlegm, sticky mouth, greasy and turbid tongue coating, and slippery pulse.

[0012] Preferably, the training method of the dialysis timing prediction model includes:

[0013] A non-dialysis group and a dialysis group are established, wherein the non-dialysis group includes data of patients with chronic kidney disease stage 5 and type 2 diabetes who have not started dialysis treatment, and the dialysis group includes data of patients with type 2 diabetic nephropathy stage 5 who have received dialysis treatment for a time interval less than a second threshold;

[0014] The data of the non-dialysis group and the dialysis group were quantified;

[0015] Screening candidate indicators from the non-dialysis group and the dialysis group based on binary logistic regression method;

[0016] Determine a modeling indicator based on the candidate indicators;

[0017] Based on the multivariate logistic regression method, the data of the modeling indicators are trained to obtain a dialysis timing prediction model.

[0018] Preferably, after training the data of the modeling indicators, a dialysis timing prediction model is obtained in combination with collinearity diagnosis.

[0019] Preferably, the indicators of the non-dialysis group and the dialysis group include: gender, age, physical health indicators, systolic blood pressure, diastolic blood pressure, total cholesterol, triglycerides, low-density lipoprotein, history of diabetes, glycosylated hemoglobin, glomerular filtration rate, hemoglobin, serum albumin, 24-hour urine protein quantification, 24-hour urine volume, 24-hour urine protein concentration, blood creatinine, blood urea nitrogen, blood uric acid, hypertension, hyperuricemia, hyperlipidemia, coronary heart disease, atherosclerosis, fatty liver, qi deficiency syndrome, blood deficiency syndrome, yin deficiency syndrome, yang deficiency syndrome, heat syndrome, qi stagnation syndrome, water stagnation syndrome, dampness obstruction syndrome, blood stasis syndrome, phlegm turbidity syndrome, diseased heart, diseased liver, diseased spleen, diseased lung, and diseased kidney.

[0020] Preferably, the dialysis timing prediction model is expressed as:

[0021] (3);

[0022] P = e Logit(P) / (1 + e Logit(P) ) (2);

[0023] in,

[0024] ;

[0025] P represents the dialysis probability, are coefficients respectively, and A is a constant.

[0026] The second aspect of the present invention discloses a prediction system for implementing the above method, comprising an acquisition module, a prediction module and an analysis module, wherein the acquisition module is used to acquire the characteristics of patients who have not yet started dialysis according to modeling indicators; the prediction module is used to predict the patient characteristics through the dialysis timing prediction model to obtain the dialysis probability; the analysis module is used to calculate the dialysis timing according to the dialysis probability and a preset first threshold.

[0027] Preferably, the prediction system further comprises a training module, and the training module is used to train the dialysis timing prediction model.

[0028] The third aspect of the present invention discloses a prediction device for implementing the above method, comprising an input unit and a prediction unit, wherein the input unit is used to input the characteristics of a patient who has not yet started dialysis according to a modeling indicator, and send the patient characteristics to the prediction unit; the prediction unit is used to predict the patient characteristics through the dialysis timing prediction model to obtain a dialysis probability; and obtain the dialysis timing according to the dialysis probability and a preset first threshold.

[0029] Preferably, the prediction device further comprises an output unit or a display unit, wherein the output unit is used to output the dialysis probability or dialysis timing; and the display unit is used to display the dialysis probability or dialysis timing.

[0030] Compared with the prior art, the beneficial effects of the present invention are as follows: according to the gender, disease location of the heart, Yin deficiency syndrome, water retention syndrome, disease location of the kidney and phlegm turbidity syndrome of patients who have not yet started dialysis, as well as the dialysis timing prediction model, analysis is performed to obtain the dialysis timing of patients with chronic kidney disease stage 5 and type 2 diabetes, provide a basis for the selection of dialysis treatment timing, and provide effective suggestions for doctors' treatment prescriptions; reduce the uncertainty caused by doctors' subjective experience; have good specificity and accuracy; are non-invasive, intuitive, simple, easy to operate and have low operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1This is a flow chart of a method for predicting the timing of dialysis for chronic kidney disease stage 5 combined with type 2 diabetes mellitus of the present invention;

[0032] Figure 2 It is the ROC curve graph;

[0033] Figure 3 It is a system logic block diagram of the present invention;

[0034] Figure 4 Schematic diagram of the prediction device of the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0037] The first aspect of the present invention provides a method for predicting the timing of dialysis for patients with chronic kidney disease stage 5 combined with type 2 diabetes, such as Figure 1 As shown, the following steps are included:

[0038] Step 101: According to the modeling indicators, the characteristics of patients who have not started dialysis are collected, wherein the modeling indicators include: gender, disease location heart, Yin deficiency syndrome, water retention syndrome, disease location kidney and phlegm turbidity syndrome. Not yet started dialysis means that patients with type 2 diabetes have not undergone dialysis after being diagnosed with chronic kidney disease stage 5.

[0039] Step 102: Obtain a dialysis timing prediction model.

[0040] Step 103: predicting the features by using the dialysis timing prediction model to obtain the dialysis probability.

[0041] Step 104: According to the dialysis probability and the preset first threshold, the current dialysis opportunity is obtained. More specifically, according to the dialysis probability and the preset first threshold, it is determined whether dialysis is currently being performed. For example, when the dialysis probability is greater than 80% or 60%, it is determined that dialysis is currently available; otherwise, it is determined that dialysis is currently not available.

[0042] According to the gender, disease location of the heart, Yin deficiency syndrome, water retention syndrome, disease location of the kidney and phlegm syndrome of patients who have not yet started dialysis, as well as the dialysis timing prediction model, analysis is performed to obtain the dialysis timing, provide a basis for the timing of dialysis treatment, and provide effective suggestions for the doctor's treatment prescription; reduce the uncertainty caused by the doctor's subjective experience; have good specificity and accuracy; be non-invasive, intuitive, simple, easy to operate and have low operating costs.

[0043] Among them, when any one of the following symptoms appears, it is determined that the disease is in the heart: palpitations, insomnia and dreaminess, and forgetfulness. When any two of the following symptoms appear, it is determined to be Yin deficiency syndrome: dry throat and dry mouth, fever in the five hearts, fear of heat and sweating or night sweats, yellow urine and dry stool, red tongue with little coating, and thin pulse. When any two of the following symptoms appear, it is determined to be water retention syndrome: edema, pleural effusion, ascites, and heaviness of limbs. When any one of the following symptoms appears, it is determined to be kidney disease: sore waist and knees, tinnitus and deafness, hair loss and loose teeth, and frequent urination at night. When any two of the following symptoms appear, it is determined to be phlegm turbidity syndrome: nausea and vomiting, spitting or coughing up phlegm, sticky mouth, greasy and turbid coating, and slippery pulse.

[0044] The training method of the dialysis timing prediction model includes the following steps:

[0045] Step 201: Establish a non-dialysis group and a dialysis group, wherein the non-dialysis group includes data of patients with chronic kidney disease stage 5 and type 2 diabetes who have not started dialysis treatment, and the dialysis group includes data of patients with type 2 diabetic nephropathy stage 5 who have received dialysis treatment for a time interval less than a second threshold. For example, patients with type 2 diabetes in CKD stage 5 who have received dialysis treatment for less than 2 or 3 months are used as the dialysis group. The data of the non-dialysis group and the dialysis group are collected with the consent of the patients and meet the approval of the Medical Ethics Committee.

[0046] The data of the non-dialysis group and the dialysis group include general information, Western medical clinical characteristics, TCM symptoms, signs, such as tongue and pulse information, etc. The relevant indicators include: gender (male, %), age (years), body health index BMI (kg / m 2 ), systolic pressure (mmHg), diastolic pressure (mmHg), total cholesterol (mmol / L), triglycerides (mmol / L), low-density lipoprotein (mmol / L), history of diabetes (months), glycosylated hemoglobin (%), estimated glomerular filtration rate eGFR (mL / min / 1.73 m 2)、Hemoglobin (g / L), serum albumin (g / L), 24-hour urine protein (g), 24-hour urine volume (L), 24-hour urine protein concentration (g / L), serum creatinine (umol / L), blood urea nitrogen (mmol / L), serum uric acid (umol / L), hypertension (no=0, yes=1), hyperuricemia (no=0, yes=1), hyperlipidemia (no=0, yes=1), coronary heart disease (no=0, yes=1), atherosclerosis (no=0, yes=1), fatty liver (no=0, yes=1), qi deficiency syndrome (no= =0, yes=1), blood deficiency syndrome (no=0, yes=1), yin deficiency syndrome (no=0, yes=1), yang deficiency syndrome (no=0, yes=1), heat syndrome (no=0, yes=1), qi stagnation syndrome (no=0, yes=1), water retention syndrome (no=0, yes=1), dampness obstruction syndrome (no=0, yes=1), blood stasis syndrome (no=0, yes=1), phlegm turbidity syndrome (no=0, yes=1), disease location heart (no=0, yes=1), disease location liver (no=0, yes=1), disease location spleen (no=0, yes=1), disease location lung (no=0, yes=1), and disease location kidney (no=0, yes=1). The data of the non-dialysis group and the dialysis group were summarized / sorted to obtain the adoptable TCM syndrome elements.

[0047] Step 202: quantify the data of the non-dialysis group and the dialysis group. For some data, quantify by using the method of 0 for none and 1 for yes; for age, quantize by using the method of interval assignment.

[0048] Step 203: Based on the binary logistic regression method, candidate indicators are screened from the non-dialysis group and the dialysis group. The candidate indicators are indicators with greater differences between the two groups of data, and indicators with p < 0.2 can be selected, specifically: Yin deficiency syndrome, water retention syndrome, phlegm turbidity syndrome, diseased heart, and diseased kidney.

[0049] Step 204: Determine a modeling index based on the candidate index.

[0050] Confounding factors such as gender and age were also included in the modeling indicators.

[0051] Step 205: Based on the multi-factor logistic regression method, the data of the modeling indicators are trained to obtain a dialysis timing prediction model. After the data of the modeling indicators are trained, the indicators of the dialysis timing prediction model can be selected in combination with collinearity diagnosis.

[0052] More specifically, multivariate logistic regression analysis was performed using the forward stepwise method, and the results showed that male sex (OR=1.780), Yin deficiency syndrome (OR=2.262), water retention syndrome (OR=1.984), and heart disease (OR=3.714) were related factors for type 2 diabetes combined with CKD stage 5 dialysis, as shown in Table 1. Among them, the OR value (odds ratio) is expressed as the exponential power of the regression coefficient, which is an indicator to measure the degree of influence of the independent variable on the dependent variable.

[0053] Table 1

[0054]

[0055] The trained dialysis timing prediction model is expressed as:

[0056] Logit(P)= -1.320+0.577*S 1 +1.312*S 2 +0.816*S 3 +0.685*S 4 -1.589*S 5 -1.125*S 6 (1);

[0057] P = e Logit(P) / (1 + e Logit(P) ) (2);

[0058] Among them, P represents the dialysis probability, S 1 -S 6 They are the quantitative values ​​of gender, disease location heart, Yin deficiency syndrome, water retention syndrome, disease location kidney and phlegm turbidity syndrome,

[0059] .

[0060] The modeling indicators are mainly TCM indicators, and there is no conclusion yet on whether TCM indicators are causes or effects. The present invention targets TCM indicators that have not been concluded, and verifies the association between these TCM indicators and the dependent variable (dialysis). It should be pointed out that when the training data is different and after the model is retrained, the coefficients of the model indicators are slightly different.

[0061] The dialysis timing prediction model can be expressed as:

[0062] (3)

[0063] in, are coefficients, A is a constant, and Logit (P) is the predicted score.

[0064] In addition, formula 1 can also be written as:

[0065] Logit (P) = -1.320+0.577*male gender (female=0, male=1)+1.312*disease location heart (no=0, yes=1)+0.816*yin deficiency syndrome (no=0, yes=1)+0.685*water retention syndrome (no=0, yes=1)-1.589*disease location kidney (no=0, yes=1)-1.125*phlegm turbidity syndrome (no=0, yes=1) (4).

[0066] The present invention provides a prediction reference for the timing of dialysis for patients with CKD stage 5 and type 2 diabetes who have not yet started dialysis treatment. It is specific and accurate for a single kidney disease: type 2 diabetic kidney disease (T2DKD), rather than a wide range of end-stage renal disease. The present invention is non-invasive, non-invasive, intuitive, simple, easy to operate, and has low operating costs. Each modeling indicator pays attention to the patient's subjective feelings, reflecting humanistic care. It provides a therapeutic theory and basis for clinical relief of the discomfort symptoms of patients with type 2 diabetes in CKD stage 5.

[0067] like Figure 2 In a specific prediction, the area under the curve (AUC) of the probability was 73.8% (95% CI: 0.680-0.797, p<0.001), and the optimal critical point value of the prediction model was 0.28, corresponding to a sensitivity of 70.5% and a specificity of 65.3%. Figure 2 In the figure, the horizontal axis is 1-specificity and the vertical axis is sensitivity.

[0068] It should be noted that the method of the present invention is used to predict and evaluate the timing of dialysis for patients with CKD stage 5 and type 2 diabetes, and is not used for diagnosis and treatment of the disease. For example, when a doctor prescribes dialysis, the method provides a data basis for the doctor to avoid uncertainty caused by the doctor's personal experience.

[0069] A second aspect of the present invention provides a prediction system for implementing the above method, such as Figure 3 As shown, it includes an acquisition module 1, a prediction module 2 and an analysis module 3. The acquisition module 1 is used to collect the characteristics of patients who have not yet started dialysis according to modeling indicators; the prediction module 2 is used to predict the patient characteristics through the dialysis timing prediction model to obtain the dialysis probability; the analysis module 3 is used to calculate the dialysis timing according to the dialysis probability and a preset first threshold.

[0070] The prediction system further includes a training module 4, and the training module 4 is used to train the dialysis timing prediction model.

[0071] The third aspect of the present invention provides a prediction device for implementing the above method, including an input unit 11, a prediction unit 12, and an output unit 13 or a display unit. The input unit is used to input the characteristics of patients who have not yet started dialysis according to the modeling indicators, and send the patient characteristics to the prediction unit. The prediction unit is used to predict the patient characteristics through the dialysis timing prediction model to obtain the dialysis probability; and obtain the dialysis timing according to the dialysis probability and a preset first threshold. The output unit is used to output the dialysis probability or dialysis timing; the display unit is used to display the dialysis probability or dialysis timing. The prediction device can be designed as a tablet, a PC or other forms of devices.

[0072] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Prediction method of dialysis timing for chronic kidney disease stage 5 combined with type 2 diabetes mellitus, It is characterized in that The following steps are involved: According to the modeling indicators, the characteristics of patients who have not yet undergone dialysis are collected, and the modeling indicators include: gender, disease location in the heart, Yin deficiency syndrome, water retention syndrome, disease location in the kidney, and phlegm turbidity syndrome; Obtain a dialysis timing prediction model; Predicting the patient characteristics by using the dialysis timing prediction model to obtain a dialysis probability; Obtaining a dialysis opportunity according to the dialysis probability and a first threshold; The heart disease includes any one of the following symptoms: palpitations, insomnia, dreaminess, and forgetfulness; the kidney disease includes any one of the following symptoms: soreness of waist and knees, tinnitus, deafness, hair loss, loose teeth, and frequent urination at night; Among them, the dialysis timing prediction model is expressed as: Logit(P)=A+λ 1 ×S 1 +λ 2 ×S 2 +λ 3 ×S 3 +λ 4 ×S 4 +λ 5 ×S 5 +λ 6 ×S 6 (3) P=e Logit(P) / (1+e Logit(P) ) (2) in, P represents the dialysis probability, λ 1 , 2 , 3 , 4 , 5 and λ 6 are coefficients respectively, and A is a constant; Among them, the training method of the dialysis timing prediction model includes: A non-dialysis group and a dialysis group are established, wherein the non-dialysis group includes data of patients with chronic kidney disease stage 5 and type 2 diabetes who have not started dialysis treatment, and the dialysis group includes data of patients with chronic kidney disease stage 5 and type 2 diabetes who have received dialysis treatment for a time interval less than a second threshold; The data of the non-dialysis group and the dialysis group were quantified; Screening candidate indicators from the non-dialysis group and the dialysis group based on binary logistic regression method; Determine a modeling indicator based on the candidate indicators; Based on the multivariate logistic regression method, the data of the modeling indicators are trained to obtain a dialysis timing prediction model.

2. The method for predicting dialysis timing according to claim 1, It is characterized in that The Yin deficiency syndrome includes any two of the following symptoms: dry throat and mouth, fever in the five parts of the body, fear of heat and sweating or night sweats, yellow urine and dry stool, red tongue with little coating, and weak pulse; The water retention syndrome includes any two of the following symptoms: edema, pleural effusion, ascites, and heaviness of limbs; The phlegm and turbidity syndrome includes any two of the following symptoms: nausea and vomiting, spitting or coughing up phlegm, sticky mouth, greasy and turbid tongue coating, and slippery pulse.

3. The method for predicting dialysis timing according to claim 1, It is characterized in that After training the data of the modeling indicators, combined with collinearity diagnosis, a dialysis timing prediction model is obtained.

4. The method for predicting dialysis timing according to claim 1, It is characterized in that The indicators of the non-dialysis group and the dialysis group include: gender, age, physical health indicators, systolic blood pressure, diastolic blood pressure, total cholesterol, triglycerides, low-density lipoprotein, history of diabetes, glycosylated hemoglobin, glomerular filtration rate, hemoglobin, serum albumin, 24-hour urine protein quantification, 24-hour urine volume, 24-hour urine protein concentration, blood creatinine, blood urea nitrogen, blood uric acid, hypertension, hyperuricemia, hyperlipidemia, coronary heart disease, atherosclerosis, fatty liver, qi deficiency syndrome, blood deficiency syndrome, yin deficiency syndrome, yang deficiency syndrome, heat syndrome, qi stagnation syndrome, water stagnation syndrome, dampness obstruction syndrome, blood stasis syndrome, phlegm turbidity syndrome, disease location of the heart, disease location of the liver, disease location of the spleen, disease location of the lung, and disease location of the kidney.

5. A prediction system, It is characterized in that Used to implement the dialysis timing prediction method according to any one of claims 1 to 4, the prediction system comprises: a collection module, a prediction module and an analysis module, The acquisition module is used to acquire characteristics of patients who have not yet started dialysis based on modeling indicators; The prediction module is used to predict the patient characteristics through the dialysis timing prediction model to obtain a dialysis profile; The analysis module is used to calculate the dialysis opportunity according to the dialysis probability and the first threshold.

6. The prediction system according to claim 5, It is characterized in that It also includes a training module, which is used to train the dialysis timing prediction model.

7. A prediction device, It is characterized in that Used to implement the dialysis timing prediction method as described in any one of claims 1 to 4, comprising an input unit and a prediction unit, The input unit is used to input the patient characteristics that have not yet started dialysis according to the modeling index, and send the patient characteristics to the prediction unit; The prediction unit is used to predict the patient characteristics through the dialysis timing prediction model to obtain the dialysis probability; and to obtain the dialysis timing according to the dialysis probability and the first threshold.

8. The prediction device according to claim 7, It is characterized in that Also includes an output unit or a display unit, The output unit is used to output the dialysis probability or dialysis opportunity; The display unit is used to display the dialysis probability or dialysis opportunity.