CN113744825A - Model for prognosis of gram-negative bacillus blood stream infection of tumor patient and construction method - Google Patents
Model for prognosis of gram-negative bacillus blood stream infection of tumor patient and construction method Download PDFInfo
- Publication number
- CN113744825A CN113744825A CN202111007806.XA CN202111007806A CN113744825A CN 113744825 A CN113744825 A CN 113744825A CN 202111007806 A CN202111007806 A CN 202111007806A CN 113744825 A CN113744825 A CN 113744825A
- Authority
- CN
- China
- Prior art keywords
- gram
- death
- infection
- blood flow
- patients
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a prognosis model of gram-negative bacillus blood flow infection of a tumor patient, which comprises a plurality of optimal prognosis factors, corresponding assessment scores of the optimal prognosis factors in different states and survival probability corresponding to a total score obtained by adding the assessment scores; wherein the prognostic factors consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection, and shock; the different status, and optimal prognostic factors include the different counts of the platelets prior to death in gram-negative bacilli blood flow infected tumor patients, the different counts of the lymphocytes prior to death in gram-negative bacilli blood flow infected tumor patients, whether gram-negative bacilli blood flow infected tumor patients entered the ICU prior to death, whether gram-negative bacilli blood flow infected tumor patients experienced the shock prior to death, and whether tumor patients prior to gram-negative bacilli blood flow infection had the pulmonary infection. The risk of death of a patient is predicted quickly, timely, and accurately.
Description
Technical Field
The invention relates to the technical field of medical treatment, in particular to a model and a construction method for prognosis of gram-negative bacillus blood flow infection of a tumor patient.
Background
Blood Stream Infection (BSI) is a serious, systemic infectious disease that can trigger systemic infection, poisoning, and systemic inflammatory responses. Further leading to systemic Multiple Organ Dysfunction Syndrome (MODS) and even death. The immunity of tumor patients is low due to the tumor and the treatment mode, and the patients with the tumor are more prone to blood stream infection compared with patients with other diseases. The main pathogenic bacteria of BSI of tumor patients are gram-negative bacteria, wherein the gram-negative bacteria mainly comprise Escherichia coli and Klebsiella pneumoniae. The disease of a tumor patient is fast after the patient is infected by blood flow, and the disease course is easy to further progress to sepsis, severe sepsis, septic shock and even death. Therefore, a timely and accurate biological index or scoring tool is needed to predict the death risk of tumor patients after blood stream infection.
Currently, scoring tools commonly used clinically to predict the risk of death include SIRS criteria (Systemic inflammation Response Syndrome, SIRS), Sequential Organ failure Assessment score (SOFA score), qSOFA score, pitter bacteremia (qPitt score), and APACHE ii score. Among the criteria for SIRS are: the body temperature is higher than 38 ℃ or lower than 36 ℃; heart rate greater than 90 beats/minute; breathing for more than 20 times/min or carbon dioxide partial pressure less than 32 mmHg; the white blood cells are more than 12000/ul or less than 4000/ul, or the immature cells are more than 10 percent, and the SIRS can be diagnosed by 2 or more than 2. The SOFA score requires multi-system data for scoring, mainly including respiratory system (PaO2/FiO2 oxygenation index) (mmHg), platelet count (x109/L), bilirubin (umol/L), circulatory system function, GCS score, renal function. The qsfa score was evaluated based on systolic blood pressure, respiratory rate and changes in consciousness. The qPitt score was evaluated based on systolic blood pressure, respiratory rate, changes in consciousness, body temperature and whether the heart was beating or not. The APACHE II score is assessed based on the patient's history of organ dysfunction, presence or absence of acute renal failure, age, body temperature, Mean Arterial Pressure (MAP), arterial blood pH, heart rate, respiratory rate, serum sodium (mmol/L), serum potassium (mmol/L), serum creatinine (μmol/L), hematocrit, white blood cell count, Glasgow coma score, alveolar-arterial oxygen differential pressure.
The research on the scoring tool shows that after careful analysis, the SIRS, SOFA, qPitt and APACHE II scores are complex to calculate and cannot be scored timely and quickly, and the research shows that the sensitivity of the qSOFA scores is 46.7 percent and the specificity is 81.3 percent, so that the qSOFA scores cannot predict the death risk of the patient timely; second, recent studies have found that the scoring tool has limitations for predicting the risk of death due to bloodstream infections, resulting in poor adaptability and accuracy.
Disclosure of Invention
Aiming at the problems, the invention provides a model and a construction method for the prognosis of the gram-negative bacillus blood flow infection of a tumor patient, which can quickly and timely predict the death risk of the patient and have high accuracy.
The technical scheme of the invention is as follows: a prognosis model of gram-negative bacillus blood flow infection of a tumor patient comprises a plurality of survival probabilities corresponding to total scores obtained by adding evaluation scores and evaluation scores corresponding to the evaluation scores when the evaluation scores are in different states, wherein the evaluation scores are selected from all risk factors and have the best prognosis ability; wherein the prognostic factors consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection, and shock; the different status, and optimal prognostic factors include the different counts of the platelets prior to death in gram-negative bacilli blood flow infected tumor patients, the different counts of the lymphocytes prior to death in gram-negative bacilli blood flow infected tumor patients, whether gram-negative bacilli blood flow infected tumor patients entered the ICU prior to death, whether gram-negative bacilli blood flow infected tumor patients experienced the shock prior to death, and whether tumor patients prior to gram-negative bacilli blood flow infection had the pulmonary infection.
The working principle of the technical scheme is as follows:
the survival probability corresponding to the total score obtained by adding the plurality of optimal prognostic factors, the corresponding assessment scores of the optimal prognostic factors in different states and the assessment scores is obtained; wherein the prognostic factors are designed to consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection and shock, forming a prognostic model capable of rapidly and timely predicting the risk of death of a patient, and through various different states of the prognostic factors including different counts of platelets before death of a gram-negative bacilli bloodstream-infected tumor patient, different counts of lymphocytes before death of a gram-negative bacilli bloodstream-infected tumor patient, whether a gram-negative bacilli bloodstream-infected tumor patient entered the ICU before death, whether a gram-negative bacilli bloodstream-infected tumor patient experienced the shock before death, and whether a gram-negative bacilli bloodstream-infected tumor patient experienced the pulmonary infection before death, not only forming a design capable of rapidly, and rapidly predicting the risk of death of a patient, A prognostic model that predicts the risk of death of a patient in a timely manner, and achieves accuracy of the prognostic model in predicting the risk of death of a patient. On one hand, the embodying of the accuracy is objective and accurate due to the risk factors of platelet counting, lymphocyte counting, ICU, pulmonary infection and shock; on the other hand, the determination of the different states corresponding to each risk factor has reasonable limitations (different counts of platelets, different counts of lymphocytes, whether to enter ICU and whether to shock are all limitations of tumor patients infected by gram-negative bacilli bloodstream infection before death, and whether to have the pulmonary infection before infection by gram-negative bacilli bloodstream).
In a further technical solution, the survival probability refers to a survival probability of 30 days. On the premise of ensuring that the death risk of the patient can be predicted quickly and timely, the method has good capability of predicting the death risk of the patient in a short term.
In a further technical solution, the survival probability model for 30days is:
h (30days) ═ h0(30days) × exp (1.218 × ICU +0.8573 × pulmony. infection +1.3448 × shock +0.3759 × L-0.0062 × PLT — 1); wherein h0(30days) is a constant; the ICU is an intensive care unit; infection is a pulmonary infection; shock is shock; l is lymphocyte count; PLT is platelet count; whether a tumor patient infected by gram-negative bacillus blood flow enters an intensive care unit before death or not is judged, if yes, the ICU is 1, and if not, the ICU is 0; whether the shock occurs to a tumor patient infected by gram-negative bacillus blood flow before death, if so, taking 1 from the shock, and if not, taking 0 from the shock; whether a tumor patient before the bloodstream infection of gram-negative bacilli has the lung infection or not is judged, if yes, the lung infection is judged to be 1, and if not, the lung infection is judged to be 0. A prognostic model is formed which not only has the capability of predicting the death risk of a patient quickly and timely, but also has a good capability of predicting the death risk of a patient in a short period of time.
In order to solve the technical problem, a method for constructing a prognosis model of gram-negative bacillus bloodstream infection of a tumor patient comprises the following steps:
a. constructing a gram-negative bacillus blood flow infection tumor patient database;
b. dividing patients in a gram-negative bacillus patient database into a death group and a non-death group according to the survival condition of the gram-negative bacillus blood flow infected tumor patients within 30 days;
c. comparing laboratory and clinical data of the death group and the non-death group, and screening out risk factors between the death group and the non-death group from the laboratory and clinical data;
d. bringing the screened risk factors in the death group and the non-death group into a Cox regression model, obtaining contribution, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model and the p values are smaller than or equal to 0.05, screening the prognosis factors with the best prognosis capability from all the risk factors, and carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a prognosis model; wherein, in step d, the optimal prognostic factors include platelets, lymphocytes, ICU, pulmonary infection and shock in the prognostic model of gram-negative bacilli bloodstream infection in the above-mentioned tumor patients; different status, and optimal prognostic factors include different counts of the platelets in a gram-negative bacilli blood flow infected tumor patient prior to death, different counts of the lymphocytes in a gram-negative bacilli blood flow infected tumor patient prior to death, whether a gram-negative bacilli blood flow infected tumor patient entered the ICU prior to death, whether a gram-negative bacilli blood flow infected tumor patient experienced the shock prior to death, and whether a gram-negative bacilli blood flow infected tumor patient experienced the pulmonary infection prior to death in the prognostic model of such tumor patients;
e. and (3) verifying a prognosis model, diagnosing a cut-off value of death risk according to a Cox regression model, dividing the patients in the gram-negative bacillus patient database into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patients is low in 30days and the survival rate of the low-risk patients is high in 30 days.
The working principle of the technical scheme is as follows:
dividing patients into death groups and non-death groups according to the survival condition of 30days, bringing laboratory and clinical data (culture etiology result, blood routine, biochemistry, cell factor, PCT and other results) and clinical data (patient general information, current medical history, past history, physical sign, body temperature and the like) of the patients into a Cox regression model, obtaining contributions, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model, and then screening out the prognosis factors with the best prognosis among all the risk factors, carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a model, and finally combining platelet counting and lymphocyte counting whether the patients enter an ICU or not, Whether the patient has the lung infection or not and whether the shock occurs or not establishes a 30-day death risk prognosis model, and the diagnosis efficiency is good in a verification group. And (3) diagnosing the cut-off value of the death risk according to the Cox model, dividing all the patients into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patient is low in 30days and the survival rate of the low-risk patient is high in 30 days. According to the plotted survival curve, the difference between the two groups (high risk group and low risk group) is statistically significant (p is less than 0.0001), so that the model has a good prediction effect on the short-term death risk.
In a further technical scheme, the step a further comprises the following steps: a1, screening patients with positive bloodstream infection, counting the death situations and death times of the patients, and constructing a database of the patients with positive bloodstream infection; a2, screening tumor patients in a blood flow infected patient database according to death situations, receiving laboratory and clinical data of the tumor patients, and constructing the blood flow infected tumor patient database; a3, screening patients with gram-negative bacilli in a blood flow infection tumor patient database according to laboratory and clinical data, collecting the survival situation of the gram-negative bacilli patients, and constructing the gram-negative bacilli blood flow infection tumor patient database. On one hand, different databases are obtained to lay the foundation for other follow-up researches, and on the other hand, screening is carried out in a layer-by-layer indentation mode, so that the screening efficiency is higher and more accurate.
In a further technical scheme, the step e further comprises the following steps: e1, bringing patient indexes corresponding to risk factors of gram-negative bacillus blood flow infected tumor patients into a prognosis model, and calculating a predicted value; e2, drawing a roc curve according to the predicted value obtained by calculation, and obtaining the cut-off value of the death risk diagnosed by the Cox regression model according to the roc curve. By drawing the roc curve, the data embodiment is more visual, and the screening efficiency is higher and more accurate.
In a further aspect, the step d further comprises a step d1 of creating a nomogram based on the prognosis model; d2, converting the nomogram through the R language and presenting the nomogram on the webpage of the server terminal. And (4) making a nomogram based on the prognosis model and placing the nomogram on a webpage, wherein each risk factor in the nomogram is endowed with a corresponding proportional score. Taking the values of the risk factors in different states as a base point to make a vertical line segment, wherein the value of the vertical line segment corresponding to the score line at the top end (the transverse line at the Points) is the corresponding evaluation score of the value of the risk factor in different states, then adding the evaluation scores of all the risk factors to obtain a Total score (the transverse line at the Total Points), and finally obtaining the corresponding survival probability on the prediction line at the bottom of the graph (the transverse lines at 7_ days, 15_ days and 30_ days). The higher the survival rate, which is ultimately calculated from the bar graph on the web page, the higher the survival probability and the lower the risk of death for the patient. By inputting the indexes corresponding to the corresponding risk factors into the simple webpage tool, the death risk prediction of the patient with the gram-negative bacillus blood flow infected tumor can be obtained, and the method is simple, visual, timely and quick.
The invention has the beneficial effects that:
1. the survival probability corresponding to the total score obtained by adding the plurality of optimal prognostic factors, the corresponding assessment scores of the optimal prognostic factors in different states and the assessment scores is obtained; wherein the prognostic factors are designed to consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection and shock, forming a prognostic model capable of rapidly and timely predicting the risk of death of a patient, and through various different states of the prognostic factors including different counts of platelets before death of a gram-negative bacilli bloodstream-infected tumor patient, different counts of lymphocytes before death of a gram-negative bacilli bloodstream-infected tumor patient, whether a gram-negative bacilli bloodstream-infected tumor patient entered the ICU before death, whether a gram-negative bacilli bloodstream-infected tumor patient experienced the shock before death, and whether a gram-negative bacilli bloodstream-infected tumor patient experienced the pulmonary infection before death, not only forming a design capable of rapidly, and rapidly predicting the risk of death of a patient, A prognostic model that predicts the risk of death of a patient in a timely manner, and achieves accuracy of the prognostic model in predicting the risk of death of a patient. On one hand, the embodying of the accuracy is objective and accurate due to the risk factors of platelet counting, lymphocyte counting, ICU, pulmonary infection and shock; on the other hand, the determination of the different states corresponding to each risk factor has reasonable limitations (different counts of platelets, different counts of lymphocytes, whether to enter ICU and whether to shock are all limitations of tumor patients infected by gram-negative bacilli bloodstream infection before death, and whether to have the pulmonary infection before infection by gram-negative bacilli bloodstream).
2. By designing the survival probability of 30days, the method has good capability of predicting the death risk of the patient in a short term on the premise of ensuring that the death risk of the patient is predicted quickly and timely.
3. The model of survival probability by 30days was:
h (30days) ═ h0(30days) × exp (1.218 × ICU +0.8573 × pulmony. infection +1.3448 × shock +0.3759 × L-0.0062 × PLT — 1). A prognostic model is formed which not only has the capability of predicting the death risk of a patient quickly and timely, but also has a good capability of predicting the death risk of a patient in a short period of time.
4. Dividing patients into death groups and non-death groups according to the survival condition of 30days, bringing laboratory and clinical data (culture etiology result, blood routine, biochemistry, cell factor, PCT and other results) and clinical data (patient general information, current medical history, past history, physical sign, body temperature and the like) of the patients into a Cox regression model, obtaining contributions, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model, and then screening out the prognosis factors with the best prognosis among all the risk factors, carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a model, and finally combining platelet counting and lymphocyte counting whether the patients enter an ICU or not, Whether the patient has the lung infection or not and whether the shock occurs or not establishes a 30-day death risk prognosis model, and the diagnosis efficiency is good in a verification group. And (3) diagnosing the cut-off value of the death risk according to the Cox model, dividing all the patients into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patient is low in 30days and the survival rate of the low-risk patient is high in 30 days. According to the plotted survival curve, the difference between the two groups (high-risk group and low-risk group) has statistical significance (p is less than 0.0001), the survival rate of the patients with high risk value is low in 30days, and the survival rate of the patients with low risk value in 30days is high, so that the model has a good prediction effect on the short-term death risk.
5. The method comprises the following steps: a1, screening patients with positive bloodstream infection, counting the death situations and death times of the patients, and constructing a database of the patients with positive bloodstream infection; a2, screening tumor patients in a blood flow infected patient database according to death situations, receiving laboratory and clinical data of the tumor patients, and constructing the blood flow infected tumor patient database; a3, screening gram-negative bacilli in a blood flow infected tumor patient database according to laboratory and clinical data, collecting survival conditions of gram-negative bacilli patients, and constructing a design of the gram-negative bacilli blood flow infected tumor patient database.
6. Through drawing the design of roc curve, the data embodiment is more directly perceived, and screening efficiency is higher, more accurate.
7. And (4) making a nomogram based on the prognosis model and placing the nomogram on a webpage, wherein each risk factor in the nomogram is endowed with a corresponding proportional score. Taking the values of the risk factors in different states as a base point to make a vertical line segment, wherein the value of the vertical line segment corresponding to the score line at the top end (the transverse line at the Points) is the corresponding evaluation score of the value of the risk factor in different states, then adding the evaluation scores of all the risk factors to obtain a Total score (the transverse line at the Total Points), and finally obtaining the corresponding survival probability on the prediction line at the bottom of the graph (the transverse lines at 7_ days, 15_ days and 30_ days). The higher the survival rate, which is ultimately calculated from the bar graph on the web page, the higher the survival probability and the lower the risk of death for the patient. By inputting the indexes corresponding to the corresponding risk factors into the simple webpage tool, the death risk prediction of the patient with the gram-negative bacillus blood flow infected tumor can be obtained, and the method is simple, visual, timely and quick.
Drawings
FIG. 1 is a time ROC for diagnosing the death of a gram-negative bacilli bloodstream infection in a tumor patient by the Cox model in example 2 of the present invention;
FIG. 2 is a nomogram of a Cox mortality risk prediction model in example 2 of the present invention;
FIG. 3 is a 30day survival curve for high risk and low risk patients in example 2 of the present invention;
fig. 4 is a schematic diagram of a cox model web page tool in embodiment 2 of the present invention.
Detailed Description
The embodiments of the present invention will be further described with reference to the accompanying drawings.
Example 1:
a prognosis model of gram-negative bacillus blood flow infection of a tumor patient comprises a plurality of survival probabilities corresponding to total scores obtained by adding evaluation scores and evaluation scores corresponding to the evaluation scores when the evaluation scores are in different states, wherein the evaluation scores are selected from all risk factors and have the best prognosis ability; wherein the prognostic factors consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection, and shock; the different status, and optimal prognostic factors include the different counts of the platelets prior to death in gram-negative bacilli blood flow infected tumor patients, the different counts of the lymphocytes prior to death in gram-negative bacilli blood flow infected tumor patients, whether gram-negative bacilli blood flow infected tumor patients entered the ICU prior to death, whether gram-negative bacilli blood flow infected tumor patients experienced the shock prior to death, and whether tumor patients prior to gram-negative bacilli blood flow infection had the pulmonary infection.
The working principle of the technical scheme is as follows:
the survival probability corresponding to the total score obtained by adding the plurality of optimal prognostic factors, the corresponding assessment scores of the optimal prognostic factors in different states and the assessment scores is obtained; wherein the prognostic factors are designed to consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection and shock, forming a prognostic model capable of rapidly and timely predicting the risk of death of a patient, and through various different states of the prognostic factors including different counts of platelets before death of a gram-negative bacilli bloodstream-infected tumor patient, different counts of lymphocytes before death of a gram-negative bacilli bloodstream-infected tumor patient, whether a gram-negative bacilli bloodstream-infected tumor patient entered the ICU before death, whether a gram-negative bacilli bloodstream-infected tumor patient experienced the shock before death, and whether a gram-negative bacilli bloodstream-infected tumor patient experienced the pulmonary infection before death, not only forming a design capable of rapidly, and rapidly predicting the risk of death of a patient, A prognostic model that predicts the risk of death of a patient in a timely manner, and achieves accuracy of the prognostic model in predicting the risk of death of a patient. On one hand, the embodying of the accuracy is objective and accurate due to the risk factors of platelet counting, lymphocyte counting, ICU, pulmonary infection and shock; on the other hand, the determination of the different states corresponding to each risk factor has reasonable limitations (different counts of platelets, different counts of lymphocytes, whether to enter ICU and whether to shock are all limitations of tumor patients infected by gram-negative bacilli bloodstream infection before death, and whether to have the pulmonary infection before infection by gram-negative bacilli bloodstream).
In another embodiment, the survival probability refers to a 30day survival probability. On the premise of ensuring that the death risk of the patient can be predicted quickly and timely, the method has good capability of predicting the death risk of the patient in a short term.
In another embodiment, the 30-day survival probability model is:
h(30days)＝h0(30days)*exp(1.218*ICU+0.8573*pulmonary.infection+1.3448*shock+0.3759*L-0.0062*PLT_1)；
wherein h0(30days) is a constant; the ICU is an intensive care unit; infection is a pulmonary infection; shock is shock; l is lymphocyte count; PLT is platelet count;
whether a tumor patient infected by gram-negative bacillus blood flow enters an intensive care unit before death or not is judged, if yes, the ICU is 1, and if not, the ICU is 0;
whether the shock occurs to a tumor patient infected by gram-negative bacillus blood flow before death, if so, taking 1 from the shock, and if not, taking 0 from the shock;
whether a tumor patient before the bloodstream infection of gram-negative bacilli has the lung infection or not is judged, if yes, the lung infection is judged to be 1, and if not, the lung infection is judged to be 0. A prognostic model is formed which not only has the capability of predicting the death risk of a patient quickly and timely, but also has a good capability of predicting the death risk of a patient in a short period of time. As the disease of a tumor patient is fast after the patient is infected by blood flow, the disease course is easy to further progress to sepsis, severe sepsis, septic shock, even death and the like, only the prognosis model with strong short-term prognosis capability can be better adapted to the condition.
In another embodiment, a prognostic model for gram-negative bacilli bloodstream infection of a tumor patient comprises a plurality of screening out the best prognostic factors with prognostic capability from all risk factors, evaluation scores corresponding to different states of each best prognostic factor, and survival probabilities corresponding to total scores obtained by adding the evaluation scores; wherein the prognostic factors consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection, and shock; the different status, and optimal prognostic factors include the different counts of the platelets prior to death in gram-negative bacilli blood flow infected tumor patients, the different counts of the lymphocytes prior to death in gram-negative bacilli blood flow infected tumor patients, whether gram-negative bacilli blood flow infected tumor patients entered the ICU prior to death, whether gram-negative bacilli blood flow infected tumor patients experienced the shock prior to death, and whether tumor patients prior to gram-negative bacilli blood flow infection had the pulmonary infection. And the survival probability refers to a 30-day survival probability. On the premise of ensuring that the death risk of the patient can be predicted quickly and timely, the method has good capability of predicting the death risk of the patient in a short term. And the 30-day model of survival probability is:
h (30days) ═ h0(30days) × exp (1.218 × ICU +0.8573 × pulmony. infection +1.3448 × shock +0.3759 × L-0.0062 × PLT — 1); wherein h0(30days) is a constant; the ICU is an intensive care unit; infection is a pulmonary infection; shock is shock; l is lymphocyte count; PLT is platelet count; whether a tumor patient infected by gram-negative bacillus blood flow enters an intensive care unit before death or not is judged, if yes, the ICU is 1, and if not, the ICU is 0; whether the shock occurs to a tumor patient infected by gram-negative bacillus blood flow before death, if so, taking 1 from the shock, and if not, taking 0 from the shock; whether a tumor patient before the bloodstream infection of gram-negative bacilli has the lung infection or not is judged, if yes, the lung infection is judged to be 1, and if not, the lung infection is judged to be 0. A prognostic model is formed which not only has the capability of predicting the death risk of a patient quickly and timely, but also has a good capability of predicting the death risk of a patient in a short period of time. Survival probability corresponding to a total score obtained by adding the risk factors, the evaluation scores corresponding to the optimal prognostic factors in different states and the evaluation scores; wherein the prognostic factors are designed to consist essentially of platelet count, lymphocyte count, ICU, pulmonary infection and shock, forming a prognostic model capable of rapidly and timely predicting the risk of death of a patient, and through various different states of the prognostic factors including different counts of platelets before death of a gram-negative bacilli bloodstream-infected tumor patient, different counts of lymphocytes before death of a gram-negative bacilli bloodstream-infected tumor patient, whether a gram-negative bacilli bloodstream-infected tumor patient entered the ICU before death, whether a gram-negative bacilli bloodstream-infected tumor patient experienced the shock before death, and whether a gram-negative bacilli bloodstream-infected tumor patient experienced the pulmonary infection before death, not only forming a design capable of rapidly, and rapidly predicting the risk of death of a patient, A prognostic model that predicts the risk of death of a patient in a timely manner, and achieves accuracy of the prognostic model in predicting the risk of death of a patient. On one hand, the embodying of the accuracy is objective and accurate due to the risk factors of platelet counting, lymphocyte counting, ICU, pulmonary infection and shock; on the other hand, the determination of the different states corresponding to each risk factor has reasonable limitations (different counts of platelets, different counts of lymphocytes, whether to enter ICU and whether to shock are all limitations of tumor patients infected by gram-negative bacilli bloodstream infection before death, and whether to have the pulmonary infection before infection by gram-negative bacilli bloodstream). As the disease of a tumor patient is fast after the patient is infected by blood flow, the disease course is easy to further progress to sepsis, severe sepsis, septic shock, even death and the like, only the prognosis model with strong short-term prognosis capability can be better adapted to the condition. The use of Sequential Organ failure Assessment score (SOFA score), qSOFA score, pittesy (qPitt score) and APACHE ii score for predicting the risk of death due to bloodstream infection in Response to SIRS criteria (SIRS), Sequential Organ failure Assessment score (SOFA score), is highly limited and less accurate, and when used to further predict patients with gram negative bacilli-infected tumors, the scoring tools will be less timely and accurate in predicting the risk of death in patients with gram negative bacilli-infected tumors.
In addition, platelet count, lymphocyte count, ICU, lung infection, and shock are risk factors, and the status of each risk factor is not a different count of platelets, a different count of lymphocytes, whether to enter the ICU, whether to present the shock, and whether to have the lung infection, but a different count of the platelets before death in a tumor patient infected with gram-negative bacillus blood flow, a different count of the lymphocytes before death in a tumor patient infected with gram-negative bacillus blood flow, whether to enter the ICU before death in a tumor patient infected with gram-negative bacillus blood flow, whether to present the shock before death in a tumor patient infected with gram-negative bacillus blood flow, and whether to have the lung infection in a tumor patient infected with gram-negative bacillus blood flow.
The traditional thinking ways focus on complex data (the first data is that the body temperature is more than 38 ℃ or less than 36 ℃, the heart rate is more than 90 times/minute, the respiration is more than 20 times/minute or the partial pressure of carbon dioxide is less than 32mmHg, the second data is that the white blood cells are more than 12000/ul or less than 4000/ul or the immature cells are more than 10%, 2 or more than 2 pieces of SIRS can be diagnosed, the third data is that the respiratory system (PaO2/FiO2 oxygenation index) (mmHg), the platelet count (x109/L), bilirubin (umol/L), the circulatory system function, the GCS score and the kidney function, the fourth data is that the systolic pressure, the respiratory frequency and the consciousness are changed, the fifth data is that the systolic pressure, the respiratory frequency, the consciousness is changed, the body temperature and whether the heart is stopped, and the sixth data is that the patient has the history of organ dysfunction, the acute renal failure and the premature cell failure, Age, body temperature, Mean Arterial Pressure (MAP), arterial blood pH, heart rate, respiratory rate, serum sodium (mmol/L), serum potassium (mmol/L), serum creatinine (μmol/L), hematocrit, white blood cell count, Glasgow coma score, alveolar-arterial oxygen differential pressure. ) Studies have completely omitted this data (platelet count, lymphocyte count, ICU, pulmonary infection and shock) study, which allows rapid acquisition of test results, but these data are used as a routine way. Thereby improving the ability of the prognostic model to predict the risk of death quickly, in time, and accurately, and the ability to predict the risk of death in the short term, as a whole. The optimal prognostic factors refer to screening out the risk factors with prognostic capability from all the risk factors, wherein the risk factors with prognostic capability meet the requirements that the risk factors are repeatedly brought into a Cox regression model, and the risk factors with the p value larger than 0.05 are removed until all the remaining risk factors are brought into the Cox regression model, and the p values are smaller than or equal to 0.05.
Example 2:
as shown in fig. 1 to fig. 2, a method for constructing a prognostic model of gram-negative bacillus bloodstream infection in a tumor patient includes the following steps:
a. constructing a gram-negative bacillus blood flow infection tumor patient database;
b. dividing patients in a gram-negative bacillus patient database into a death group and a non-death group according to the survival condition of the gram-negative bacillus blood flow infected tumor patients within 30 days;
c. comparing laboratory and clinical data of the death group and the non-death group, and screening out risk factors between the death group and the non-death group from the laboratory and clinical data;
d. bringing the screened risk factors in the death group and the non-death group into a Cox regression model, obtaining contribution, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until the p values of all the remaining risk factors are less than or equal to 0.05 after all the remaining risk factors are brought into the Cox regression model, screening the best prognosis factors with the prognosis capability (the prognosis capability satisfies that the risk factors are repeatedly brought into the Cox regression model, the risk factors with the p value larger than 0.05 are removed until all the remaining risk factors are brought into the Cox regression model, and carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a prognosis model; wherein, in step d, the optimal prognostic factors include the platelet count, the lymphocyte count, the ICU, the pulmonary infection, and the shock of example 1; different status, and optimal prognostic factors include different counts of the platelets before death in the gram-negative bacilli bloodstream-infected tumor patients, different counts of the lymphocytes before death in the gram-negative bacilli bloodstream-infected tumor patients, whether gram-negative bacilli bloodstream-infected tumor patients entered the ICU before death, whether gram-negative bacilli bloodstream-infected tumor patients experienced the shock before death, and whether gram-negative bacilli bloodstream-infected tumor patients experienced the pulmonary infection;
e. and (3) verifying a prognosis model, diagnosing a cut-off value of death risk according to a Cox regression model, dividing the patients in the gram-negative bacillus patient database into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patients is low in 30days and the survival rate of the low-risk patients is high in 30 days.
The working principle of the technical scheme is as follows:
dividing patients into death groups and non-death groups according to the survival condition of 30days, bringing laboratory and clinical data (culture etiology result, blood routine, biochemistry, cell factor, PCT and other results) and clinical data (patient general information, current medical history, past history, physical sign, body temperature and the like) of the patients into a Cox regression model, obtaining contributions, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model, and then screening out the prognosis factors with the best prognosis among all the risk factors, carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a model, and finally combining platelet counting and lymphocyte counting whether the patients enter an ICU or not, Whether lung infection occurs and whether shock occurs establishes a 30-day mortality risk prognostic model, and also achieves good diagnostic efficiency in the validation group (see fig. 1). And (3) diagnosing the cut-off value of the death risk according to the Cox model, dividing all the patients into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patient is low in 30days and the survival rate of the low-risk patient is high in 30 days. As shown in fig. 2, the survival curves plotted show that the difference between the two groups (high-risk group and low-risk group) is statistically significant (p <0.0001), and the survival rate of the high-risk patients is low in 30days, and the survival rate of the low-risk patients is high in 30days, so that the model has a good prediction effect on the short-term death risk. It can be seen that the model has a very good predictive role in short-term mortality risk. The use of Sequential Organ failure Assessment score (SOFA score), qSOFA score, pittesy (qPitt score) and APACHE ii score for predicting the risk of death due to bloodstream infection in Response to SIRS criteria (SIRS), Sequential Organ failure Assessment score (SOFA score), is highly limited and less accurate, and when used to further predict patients with gram negative bacilli-infected tumors, the scoring tools will be less timely and accurate in predicting the risk of death in patients with gram negative bacilli-infected tumors.
In another embodiment, step a further includes the following steps: a1, screening patients with positive bloodstream infection, counting the death situations and death times of the patients, and constructing a database of the patients with positive bloodstream infection; a2, screening tumor patients in a blood flow infected patient database according to death situations, receiving laboratory and clinical data of the tumor patients, and constructing the blood flow infected tumor patient database; a3, screening patients with gram-negative bacilli in a blood flow infection tumor patient database according to laboratory and clinical data, collecting the survival situation of the gram-negative bacilli patients, and constructing the gram-negative bacilli blood flow infection tumor patient database. On one hand, different databases are obtained to lay the foundation for other follow-up researches, and on the other hand, screening is carried out in a layer-by-layer indentation mode, so that the screening efficiency is higher and more accurate.
In another embodiment, step e further includes the following steps: e1, bringing patient indexes corresponding to risk factors of gram-negative bacillus blood flow infected tumor patients into a prognosis model, and calculating a predicted value; e2, drawing a roc curve according to the predicted value obtained by calculation, and obtaining the cut-off value of the death risk diagnosed by the Cox regression model according to the roc curve. By drawing the roc curve, the data embodiment is more visual, and the screening efficiency is higher and more accurate.
In another embodiment, as shown in fig. 3-4, the step d further comprises a step d1 of creating a nomogram based on the prognostic model; d2, converting the nomogram through the R language and presenting the nomogram on the webpage of the server terminal. The method comprises the steps of converting a graph R language and presenting the graph R language on a webpage of a server terminal, wherein the graph R language is converted into the webpage of the server terminal and the webpage is presented on the webpage of the server terminal by the method of converting the nomogram R language. A nomogram is generated based on the prognostic model and placed on a web page, with each risk factor in the nomogram assigned a corresponding proportional score (as shown in fig. 3). Taking the values of the risk factors in different states as a base point to make a vertical line segment, wherein the value of the vertical line segment corresponding to the score line at the top end (the transverse line at the Points) is the corresponding evaluation score of the value of the risk factor in different states, then adding the evaluation scores of all the risk factors to obtain a Total score (the transverse line at the Total Points), and finally obtaining the corresponding survival probability on the prediction line at the bottom of the graph (the transverse lines at 7_ days, 15_ days and 30_ days). The higher the survival rate, which is ultimately calculated from the bar graph on the web page, the higher the survival probability and the lower the risk of death for the patient. By inputting the indexes corresponding to the corresponding risk factors into the simple webpage tool, the death risk prediction of the patient with the gram-negative bacillus blood flow infected tumor can be obtained (as shown in figure 4), and the method is simple, visual, timely and quick. The present study can be rapidly and accurately concluded based on the different counts of the platelets before death in gram-negative bacilli blood-flow infected tumor patients, the different counts of the lymphocytes before death in gram-negative bacilli blood-flow infected tumor patients, and whether gram-negative bacilli blood-flow infected tumor patients entered the ICU before death, whether gram-negative bacilli blood-flow infected tumor patients experienced the shock before death, and whether gram-negative bacilli blood-flow infected tumor patients suffered from the pulmonary infection before death. The method has the effects of high accuracy and timeliness when predicting the death risk of the tumor patient infected by the gram-negative bacillus blood flow. And the prognosis model only uses the existing information of the good subjects, does not generate additional examination cost and does not bring excessive economic pressure to patients. And the webpage tool can be convenient for patients and doctors to use.
In another embodiment, a method for constructing a prognostic model of gram-negative bacilli bloodstream infection in a patient with a tumor comprises the steps of: a. constructing a gram-negative bacillus blood flow infection tumor patient database; b. dividing patients in a gram-negative bacillus patient database into a death group and a non-death group according to the survival condition of the gram-negative bacillus blood flow infected tumor patients within 30 days; c. comparing laboratory and clinical data of the death group and the non-death group, and screening out risk factors between the death group and the non-death group from the laboratory and clinical data; d. bringing the screened risk factors in the death group and the non-death group into a Cox regression model, obtaining contribution, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model and the p values are smaller than or equal to 0.05, screening the prognosis factors with the best prognosis capability from all the risk factors, and carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a prognosis model; wherein in step d, the optimal prognostic factors include the platelet count, the lymphocyte count, the ICU, the pulmonary infection and the shock of example 1, the different status, and optimal prognostic factors include the different counts of the platelets before death in the gram-negative bacillus blood flow infected tumor patient, the different counts of the lymphocytes before death in the gram-negative bacillus blood flow infected tumor patient, whether the gram-negative bacillus blood flow infected tumor patient entered the ICU before death, whether the gram-negative bacillus blood flow infected tumor patient experienced the shock before death, and whether the tumor patient before infection with gram-negative bacillus blood flow had the pulmonary infection; e. and (3) verifying a prognosis model, diagnosing a cut-off value of death risk according to a Cox regression model, dividing the patients in the gram-negative bacillus patient database into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patients is low in 30days and the survival rate of the low-risk patients is high in 30 days. Further, the step a further includes the steps of: a1, screening patients with positive bloodstream infection, counting the death situations and death times of the patients, and constructing a database of the patients with positive bloodstream infection; a2, screening tumor patients in a blood flow infected patient database according to death situations, receiving laboratory and clinical data of the tumor patients, and constructing the blood flow infected tumor patient database; a3, screening patients with gram-negative bacilli in a blood flow infection tumor patient database according to laboratory and clinical data, collecting the survival situation of the gram-negative bacilli patients, and constructing the gram-negative bacilli blood flow infection tumor patient database. In step e, the method further includes the steps of: e1, bringing each patient index (prognostic factor in different states) corresponding to the risk factor of the tumor patient infected by gram-negative bacillus blood flow into a prognostic model, and calculating a predicted value (survival probability); e2, drawing a roc curve according to the predicted value obtained by calculation, and obtaining the cut-off value of the death risk diagnosed by the Cox regression model according to the roc curve. Dividing patients into death groups and non-death groups according to the survival condition of 30days, bringing laboratory and clinical data (culture etiology result, blood routine, biochemistry, cell factor, PCT and other results) and clinical data (patient general information, current medical history, past history, physical sign, body temperature and the like) of the patients into a Cox regression model, obtaining contributions, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model, and then screening out the prognosis factors with the best prognosis among all the risk factors, carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a model, and finally combining platelet counting and lymphocyte counting whether the patients enter an ICU or not, Whether lung infection occurs and whether shock occurs establishes a 30-day mortality risk prognostic model, and also achieves good diagnostic efficiency in the validation group (see fig. 1). And (3) diagnosing the cut-off value of the death risk according to the Cox model, dividing all the patients into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patient is low in 30days and the survival rate of the low-risk patient is high in 30 days. As shown in fig. 2, the survival curves plotted show that the difference between the two groups (high-risk group and low-risk group) is statistically significant (p <0.0001), and the survival rate of the high-risk patients is low in 30days, and the survival rate of the low-risk patients is high in 30days, so that the model has a good prediction effect on the short-term death risk. It can be seen that the model has a very good predictive role in short-term mortality risk. The use of Sequential Organ failure Assessment score (SOFA score), qSOFA score, pittesy (qPitt score) and APACHE ii score for predicting the risk of death due to bloodstream infection in Response to SIRS criteria (SIRS), Sequential Organ failure Assessment score (SOFA score), is highly limited and less accurate, and when used to further predict patients with gram negative bacilli-infected tumors, the scoring tools will be less timely and accurate in predicting the risk of death in patients with gram negative bacilli-infected tumors. On one hand, different databases are obtained to lay the foundation for other follow-up researches, and on the other hand, screening is carried out in a layer-by-layer indentation mode, so that the screening efficiency is higher and more accurate. By drawing the roc curve, the data embodiment is more visual, and the screening efficiency is higher and more accurate. Therefore, the efficiency of constructing the prognosis model is increased more efficiently and accurately on the whole.
In addition, platelet count, lymphocyte count, ICU, lung infection, and shock are risk factors, and the status of each risk factor is not a different count of platelets, a different count of lymphocytes, whether to enter the ICU, whether to present the shock, and whether to have the lung infection, but a different count of the platelets before death in a tumor patient infected with gram-negative bacillus blood flow, a different count of the lymphocytes before death in a tumor patient infected with gram-negative bacillus blood flow, whether to enter the ICU before death in a tumor patient infected with gram-negative bacillus blood flow, whether to present the shock before death in a tumor patient infected with gram-negative bacillus blood flow, and whether to have the lung infection in a tumor patient infected with gram-negative bacillus blood flow.
The traditional thinking ways focus on complex data (the first data is that the body temperature is more than 38 ℃ or less than 36 ℃, the heart rate is more than 90 times/minute, the respiration is more than 20 times/minute or the partial pressure of carbon dioxide is less than 32mmHg, the second data is that the white blood cells are more than 12000/ul or less than 4000/ul or the immature cells are more than 10%, 2 or more than 2 pieces of SIRS can be diagnosed, the third data is that the respiratory system (PaO2/FiO2 oxygenation index) (mmHg), the platelet count (x109/L), bilirubin (umol/L), the circulatory system function, the GCS score and the kidney function, the fourth data is that the systolic pressure, the respiratory frequency and the consciousness are changed, the fifth data is that the systolic pressure, the respiratory frequency, the consciousness is changed, the body temperature and whether the heart is stopped, and the sixth data is that the patient has the history of organ dysfunction, the acute renal failure and the premature cell failure, Age, body temperature, Mean Arterial Pressure (MAP), arterial blood pH, heart rate, respiratory rate, serum sodium (mmol/L), serum potassium (mmol/L), serum creatinine (μmol/L), hematocrit, white blood cell count, Glasgow coma score, alveolar-arterial oxygen differential pressure. ) Studies have completely omitted this data (platelet count, lymphocyte count, ICU, pulmonary infection and shock) study, which allows rapid acquisition of test results, but these data are used as a routine way.
The above-mentioned embodiments only express the specific embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention.
Claims (7)
1. A prognosis model of gram-negative bacilli blood stream infection of a tumor patient is characterized by comprising a plurality of survival probabilities corresponding to a total score obtained by screening the best prognosis factors with prognosis ability, evaluation scores corresponding to the best prognosis factors in different states and addition of the evaluation scores from all risk factors; wherein the optimal prognostic factors consist essentially of platelets, lymphocytes, ICU, lung infection and shock; the different status, and optimal prognostic factors include the different counts of the platelets prior to death in gram-negative bacilli blood flow infected tumor patients, the different counts of the lymphocytes prior to death in gram-negative bacilli blood flow infected tumor patients, whether gram-negative bacilli blood flow infected tumor patients entered the ICU prior to death, whether gram-negative bacilli blood flow infected tumor patients experienced the shock prior to death, and whether tumor patients prior to gram-negative bacilli blood flow infection had the pulmonary infection.
2. The prognostic model for gram-negative bacilli bloodstream infection in tumor patients according to claim 1, wherein the survival probability is 30 days.
3. The prognostic model of gram-negative bacilli bloodstream infection in tumor patients according to claim 1, wherein the model of survival probability for 30days is:
h(30days)＝h0(30days)*exp(1.218*ICU+0.8573*pulmonary.infection+1.3448*shock+0.3759*L-0.0062*PLT_1)；
wherein h0(30days) is a constant; the ICU is an intensive care unit; infection is a pulmonary infection; shock is shock; l is lymphocyte count; PLT is platelet count;
whether a tumor patient infected by gram-negative bacillus blood flow enters an intensive care unit before death or not is judged, if yes, the ICU is 1, and if not, the ICU is 0;
whether the shock occurs to a tumor patient infected by gram-negative bacillus blood flow before death, if so, taking 1 from the shock, and if not, taking 0 from the shock;
whether a tumor patient before the bloodstream infection of gram-negative bacilli has the lung infection or not is judged, if yes, the lung infection is judged to be 1, and if not, the lung infection is judged to be 0.
4. A method for constructing a prognosis model of gram-negative bacillus blood flow infection of a tumor patient is characterized by comprising the following steps:
a. constructing a gram-negative bacillus blood flow infection tumor patient database;
b. dividing patients in a gram-negative bacillus patient database into a death group and a non-death group according to the survival condition of the gram-negative bacillus blood flow infected tumor patients within 30 days;
c. comparing laboratory and clinical data of the death group and the non-death group, and screening out the risk factors of the death group from the laboratory and clinical data;
d. bringing the screened risk factors in the death group and the non-death group into a Cox regression model, obtaining contribution, removing the risk factors with the p value larger than 0.05, bringing the remaining risk factors with the p value smaller than or equal to 0.05 into the Cox regression model again, removing the risk factors with the p value larger than 0.05 until all the remaining risk factors are brought into the Cox regression model and the p values are smaller than or equal to 0.05, screening the prognosis factors with the best prognosis capability from all the risk factors, and carrying out multi-factor regression analysis on the best prognosis factors in different states to obtain a prognosis model; wherein, in step d, the optimal prognostic factors include platelets, lymphocytes, ICU, pulmonary infection and shock according to any one of claims 1 to 3; different status, and optimal prognostic factors include different counts of the platelets before death in a gram-negative bacilli blood flow infected tumor patient according to one of claims 1 to 3, different counts of the lymphocytes before death in a gram-negative bacilli blood flow infected tumor patient, whether a gram-negative bacilli blood flow infected tumor patient entered the ICU before death, whether a gram-negative bacilli blood flow infected tumor patient experienced the shock before death, and whether a gram-negative bacilli blood flow infected tumor patient experienced the pulmonary infection before death;
e. and (3) verifying a prognosis model, diagnosing a cut-off value of death risk according to a Cox regression model, dividing the patients in the gram-negative bacillus patient database into a high-risk group and a low-risk group, drawing survival curves of the two groups, and verifying whether the survival rate of the high-risk patients is low in 30days and the survival rate of the low-risk patients is high in 30 days.
5. The method for constructing a prognostic model of gram-negative bacillus blood stream infection in a patient with tumor according to claim 4, wherein the step a further comprises the steps of:
a1, screening patients with positive bloodstream infection, counting the death situations and death times of the patients, and constructing a database of the patients with positive bloodstream infection;
a2, screening tumor patients in a blood flow infected patient database according to death situations, receiving laboratory and clinical data of the tumor patients, and constructing the blood flow infected tumor patient database;
a3, screening patients with gram-negative bacilli in a blood flow infection tumor patient database according to laboratory and clinical data, collecting the survival situation of the gram-negative bacilli patients, and constructing the gram-negative bacilli blood flow infection tumor patient database.
6. The method for constructing a prognostic model of gram-negative bacillus blood flow infection in tumor patients according to claim 4, wherein the step e further comprises the following steps:
e1, bringing patient indexes corresponding to risk factors of gram-negative bacillus blood flow infected tumor patients into a prognosis model, and calculating a predicted value;
e2, drawing a roc curve according to the predicted value obtained by calculation, and obtaining the cut-off value of the death risk diagnosed by the Cox regression model according to the roc curve.
7. The method for constructing a prognostic model of gram-negative bacilli bloodstream infection in a patient with tumor according to claim 4, wherein the step d further comprises the steps of,
d1, preparing a nomogram according to the prognosis model;
d2, converting the nomogram through the R language and presenting the nomogram on the webpage of the server terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111007806.XA CN113744825A (en) | 2021-08-30 | 2021-08-30 | Model for prognosis of gram-negative bacillus blood stream infection of tumor patient and construction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111007806.XA CN113744825A (en) | 2021-08-30 | 2021-08-30 | Model for prognosis of gram-negative bacillus blood stream infection of tumor patient and construction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113744825A true CN113744825A (en) | 2021-12-03 |
Family
ID=78733976
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111007806.XA Withdrawn CN113744825A (en) | 2021-08-30 | 2021-08-30 | Model for prognosis of gram-negative bacillus blood stream infection of tumor patient and construction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113744825A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117877649A (en) * | 2023-12-05 | 2024-04-12 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | Marker group and system for predicting bacteremia of tumor patient |
-
2021
- 2021-08-30 CN CN202111007806.XA patent/CN113744825A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117877649A (en) * | 2023-12-05 | 2024-04-12 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | Marker group and system for predicting bacteremia of tumor patient |
| CN117877649B (en) * | 2023-12-05 | 2024-05-31 | 中山大学肿瘤防治中心（中山大学附属肿瘤医院、中山大学肿瘤研究所） | Marker group and system for predicting bacteremia of tumor patient |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110827993A (en) | Early death risk assessment model establishing method and device based on ensemble learning | |
| US20100185399A1 (en) | Method for Analyzing Clinical Trial Results for Efficacy of a Therapy | |
| Schneider et al. | Predicting acute kidney injury among burn patients in the 21st century: a classification and regression tree analysis | |
| CN112117006A (en) | Type 2 diabetes kidney disease risk assessment system based on ensemble learning | |
| Links et al. | Minor physical anomalies in childhood autism. Part I. Their relationship to pre-and perinatal complications | |
| CN113327679A (en) | Pulmonary embolism clinical risk and prognosis scoring method and system | |
| CN114023441A (en) | Severe AKI early risk assessment model and device based on interpretable machine learning model and development method thereof | |
| Sikka et al. | A comparison of severity of illness scoring systems for elderly patients with severe pneumonia | |
| CN116013516A (en) | Death risk prediction system and method for sepsis-related acute kidney injury | |
| Seiler et al. | Lung Ultrasound for Identification of Patients Requiring Invasive Mechanical Ventilation in COVID‐19 | |
| CN113744825A (en) | Model for prognosis of gram-negative bacillus blood stream infection of tumor patient and construction method | |
| Castellanos-Ortega et al. | A new prognostic scoring system for meningococcal septic shock in children. Comparison with three other scoring systems | |
| CN114023433A (en) | Early prediction system for predicting severity of acute pancreatitis patient | |
| Zhong et al. | Combined detection of procalcitonin, heparin‐binding protein, and interleukin‐6 is a promising assay to diagnose and predict acute pancreatitis | |
| WO2021253484A1 (en) | System for predicting mortality risk of patient | |
| CN114023440A (en) | Model and device capable of explaining layered old people MODS early death risk assessment and establishing method thereof | |
| CN108508211A (en) | Starting non-medication schizophrenic patients blood serum designated object FGF9 and its application | |
| KR102310888B1 (en) | Methods for providing information of mortality risk and devices for providing information of mortality risk using the same | |
| CN116564529A (en) | Early sepsis liver failure prediction system based on SVM | |
| CN116646080A (en) | Risk prediction model construction method and system for ET patient to progress to post-ETMF | |
| Liu et al. | A machine learning method for predicting the probability of MODS using only non-invasive parameters | |
| CN112180093B (en) | Critical disease mortality diagnostic biomarker tenascin-c and application thereof | |
| CN113707295A (en) | Prediction method and system for senile postoperative systemic inflammatory response syndrome | |
| CN108417267B (en) | System for determining short-term prognosis of patient with liver cirrhosis hepatorenal syndrome | |
| CN110504030A (en) | A kind of traumatic coagulopathy prediction technique |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20211203 |
|
| WW01 | Invention patent application withdrawn after publication |