CN107516010B - Construction method of thrombolytic dose model - Google Patents
Construction method of thrombolytic dose model Download PDFInfo
- Publication number
- CN107516010B CN107516010B CN201710714752.8A CN201710714752A CN107516010B CN 107516010 B CN107516010 B CN 107516010B CN 201710714752 A CN201710714752 A CN 201710714752A CN 107516010 B CN107516010 B CN 107516010B
- Authority
- CN
- China
- Prior art keywords
- dose
- subgroups
- group
- subgroup
- 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.)
- Active
Links
- 238000010276 construction Methods 0.000 title claims abstract description 12
- 239000003146 anticoagulant agent Substances 0.000 title claims abstract description 10
- 230000002537 thrombolytic effect Effects 0.000 title claims abstract description 10
- 208000032843 Hemorrhage Diseases 0.000 claims abstract description 25
- 208000034158 bleeding Diseases 0.000 claims abstract description 25
- 230000000740 bleeding effect Effects 0.000 claims abstract description 25
- 238000011282 treatment Methods 0.000 claims abstract description 23
- 238000004458 analytical method Methods 0.000 claims abstract description 22
- 230000000694 effects Effects 0.000 claims abstract description 15
- 230000003993 interaction Effects 0.000 claims abstract description 7
- 230000011218 segmentation Effects 0.000 claims abstract description 3
- 208000006011 Stroke Diseases 0.000 claims description 29
- 238000003066 decision tree Methods 0.000 claims description 9
- 238000010801 machine learning Methods 0.000 claims description 5
- 238000000638 solvent extraction Methods 0.000 claims description 4
- 208000023516 stroke disease Diseases 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 7
- 208000032382 Ischaemic stroke Diseases 0.000 abstract description 5
- 230000002452 interceptive effect Effects 0.000 abstract description 4
- 238000007405 data analysis Methods 0.000 abstract description 3
- 238000011221 initial treatment Methods 0.000 abstract description 3
- 102100033571 Tissue-type plasminogen activator Human genes 0.000 description 14
- 108050006955 Tissue-type plasminogen activator Proteins 0.000 description 14
- 238000011160 research Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 239000003814 drug Substances 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000001225 therapeutic effect Effects 0.000 description 2
- 238000002560 therapeutic procedure Methods 0.000 description 2
- 206010018985 Haemorrhage intracranial Diseases 0.000 description 1
- 206010020772 Hypertension Diseases 0.000 description 1
- 208000008574 Intracranial Hemorrhages Diseases 0.000 description 1
- 101100328463 Mus musculus Cmya5 gene Proteins 0.000 description 1
- 230000001154 acute effect Effects 0.000 description 1
- 238000003915 air pollution Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 235000005911 diet Nutrition 0.000 description 1
- 235000021045 dietary change Nutrition 0.000 description 1
- 230000000378 dietary effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000003527 fibrinolytic agent Substances 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000000176 thermal ionisation mass spectrometry Methods 0.000 description 1
- 238000013055 trapped ion mobility spectrometry Methods 0.000 description 1
- 208000001072 type 2 diabetes mellitus Diseases 0.000 description 1
Images
Landscapes
- Medicines Containing Plant Substances (AREA)
- Medicines That Contain Protein Lipid Enzymes And Other Medicines (AREA)
Abstract
The invention provides a construction method of a thrombolytic dose model, which comprises the following steps: step (1): incorporating existing medical databases; step (2): performing subgroup analysis on the medical database; and (3): creating a segmentation model based on the subgroup analysis; and (4): based on NIHSS assigned patients, three subgroups were determined accordingly with the effect of the interaction between t-PA dose and subgroup on bleeding risk; and (5): the present invention analyzes the three subgroups to determine the efficacy between the high dose and the low dose, and analyzes the subgroups of patients with ischemic stroke by using the data analysis method of interactive tree to see if the patients with high dose are more likely to bleed (e.g., bleeding risk) (using "high dose" and "low dose" t-PA at the time of initial treatment) to find that the low dose of t-PA is similar to the high dose of t-PA (e.g., residence time, final efficacy of treatment).
Description
Technical Field
The invention relates to the field of medical treatment, in particular to a construction method of a thrombolytic dose model.
Background
Most stroke studies have been conducted in north america (where stroke is the fifth largest killer) or europe with relatively small sample sizes. In contrast, our background is in china. Stroke is the biggest killer in china due to its unique genetic, dietary, sports and environmental factors. Internal and external air pollution is an increasing significant factor in stroke events in china. With the tremendous increase in wealth, Chinese people also undergo dramatic dietary changes, resulting in dramatic increases in hypertension and type II diabetes.
These problems are particularly important for stroke care in china, as people going away from home, rounding off to a more distant hospital, may result in a too short stroke treatment window. In addition, over 60% of healthcare is self-funding, and about 30% of average household revenue will be used for the ongoing cost of stroke survival and hospitalization costs.
Another problem is that stroke treatment follows largely the tradition and guidelines of western research. These are not necessarily the best way to treat chinese patients and there is generally some clinical and research debate regarding asians. For example, two studies based on Japanese clinical trials have shown that lower doses of t-PA are as safe and effective as the recommended higher doses (studies from European and American populations) (Yamaguchi et al, 1995; Yamaguchi et al, 2006), but they are not directly comparable to higher doses. However, subsequent Japanese studies have reached almost the same conclusion (Mori et al, 2010; Nakagawara et al, 2010). One study in vietnam more strongly suggested that the results at low doses were better than at higher doses (Nguyen, 2010). However, mixed ethnic asian patient studies conducted in singapore showed higher dose efficacy (Sharma et al, 2010). However, recent korean studies found almost the same results as (Yamaguchi et al, 2006). Another study claimed that higher doses increased the risk of intracranial hemorrhage in asian patients (Menon et al, 2012), but they did not directly control the dose. However, one study in china concluded that high doses were more effective than low doses for chinese (Liao et al, 2014). However, a study involving most asian patients in the united states failed to conclude that the lower dose was more inferior than the standard dose (Anderson et al, 2016). Finally, researchers reviewed 23 trials involving asian patients, including the Anderson et al study-and concluded that the actual effect of low doses was not clear, although lower doses did result in less risk of bleeding (Dong, 2016). To solve the key problem of which guidelines for acute ischemic stroke treatment should be followed by the chinese medical center? The bet is obviously high. Excessive treatment is costly and can also lead to fatal bleeding; lack of treatment can perpetuate the effects of stroke, causing long-term and costly complications. Most importantly, a recent society of New England journal of medicine has called for more research to find therapeutic levels of t-PA suitable for Asians suffering from acute ischemic stroke. On the one hand, they indicate that asia now has 60% of world stroke victims, and that the average cost of a dose of t-PA increases by 111% over a decade, from $ 3,050 to $ 6,430, making acute phase costs a major problem in asia (Sila, 2016). Surprisingly, according to a study published in the same magazine, 63% of patients were asians and did not conclude a poor low dose, but the magazine supported a better high dose standing (Anderson et al, 2016). This "asian" stroke treatment debate is an ideal situation where evidence-based medicine is crucial, meaning that chinese may not be suitable for the rough classification of "asian" used in western studies in non-asian settings.
Disclosure of Invention
To address the above deficiencies, the present invention provides a method for constructing a thrombolytic dose model, and to solve this problem, a subgroup analysis is performed on patients with ischemic stroke using a data analysis method of cross-trees to see whether patients with high dose are more likely to bleed (e.g., bleeding risk) (using "high dose" and "low dose" t-PA at initial treatment) to find that low dose t-PA is similarly effective to high dose t-PA (e.g., residence time, final effect of treatment). Using bleeding index as the outcome of interest, dose as therapy (high and low dose based on threshold split), relevant patient characteristics as covariates; an interaction tree is then generated to identify a subset of patients that are characterized by certain interactive effects that are intense in relation to treatment (dose level). Finally, each subset obtained was examined to determine the efficacy of the low dose was not inferior to the high dose (e.g., improved NIHSS score) while the high dose had a higher risk of bleeding.
The invention provides a construction method of a thrombolytic dose model, which comprises the following steps:
step (1): incorporating existing medical databases;
step (2): performing subgroup analysis on the medical database;
and (3): creating a segmentation model based on the subgroup analysis;
and (4): based on NIHSS assigned patients, three subgroups were determined accordingly with the effect of the interaction between t-PA dose and subgroup on bleeding risk;
and (5): analysis of the above three subgroups resulted in efficacy between the high and low dose.
In the above construction method, in the step (1), the existing medical database includes: TIMS-china database containing 1440 stroke patients and SSSS database containing 11 hospitals and 2000 stroke patients, with scales to assess stroke severity, time/date stamp, time to TPA, final score of treatment effect, mortality, residence time and bleeding risk included in the existing medical database.
The above construction method, wherein the step (2) further comprises: subgroup analysis machine learning is used by recursive partitioning and decision trees.
The above construction method, wherein the step (3) further comprises: based on subgroup analysis, combining the characteristics of stroke disease species and the characteristics of the existing medical database, a segmented model is created, and the data is used for determining different treatment parameters corresponding to each patient subgroup, and a decision tree is associated with the prediction factors of the subgroups.
The above construction method, wherein the three subgroups are: (1) NIHSS is less than or equal to 4, (2)5 is less than or equal to NIHSS is less than or equal to 14, and (3) NIHSS is more than or equal to 15, and the estimated difference between the logarithmic probability of high-dose and low-dose bleeding of the three subgroups is respectively as follows: group (1) is-0.85, where p is 0.15, group (2) is 1.95, where p is 0.02, and group (3) is 0.39, where p is 0.56.
The above construction method wherein within the above three subgroups the estimated difference in log odds of efficacy between high and low dose is: group (1) was 0.16, where p is 0.49, group (2) was-0.16, where p is 0.57, and group (3) was-0.08, where p is 0.80, indicating that the high and low doses did not differ significantly in efficacy.
The invention has the following advantages: 1. the present invention finds that low dose t-PA is similarly effective to high dose t-PA (e.g., residence time, ultimate therapeutic effect of treatment) by performing a subgroup analysis of ischemic stroke patients using cross-tree data analysis to see if patients with high doses are more likely to bleed (e.g., risk of bleeding) (using "high dose" and "low dose" t-PA at initial treatment). Using bleeding index as the outcome of interest, dose as therapy (high and low dose based on threshold split), relevant patient characteristics as covariates; an interaction tree is then generated to identify a subset of patients that are characterized by certain interactive effects that are intense in relation to treatment (dose level). Finally, each subset obtained was examined to determine the efficacy of the low dose was not inferior to the high dose (e.g., improved NIHSS score) while the high dose had a higher risk of bleeding.
Drawings
The invention and its features, aspects and advantages will become more apparent from reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings. Like reference symbols in the various drawings indicate like elements. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.
FIG. 1 is a flow chart of the construction method of the thrombolytic drug dose model of the present invention.
Detailed Description
In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the present invention. It will be apparent, however, to one skilled in the art, that the present invention may be practiced without one or more of these specific details. In other instances, well-known features have not been described in order to avoid obscuring the invention.
In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The following detailed description of the preferred embodiments of the invention, however, the invention is capable of other embodiments in addition to those detailed.
Referring to fig. 1, the present invention provides a method for constructing a thrombolytic dose model, comprising the steps of:
step (1): incorporating an existing medical database, wherein the existing medical database comprises: TIMS-china database containing 1440 stroke patients and SSSS database containing 11 hospitals and 2000 stroke patients, wherein the existing medical databases include scales to assess stroke severity, time/date stamp, time to TPA, final score of treatment effect, mortality, residence time and bleeding risk; the method specifically comprises the following steps: the first one contained the TIMS-china database of 1440 stroke patients (67 major stroke centers involved) and the second one contained the SSSS database of 11 hospitals and 2000 stroke patients. The database includes a scale to assess stroke severity, time/date stamp, time to TPA (initial t-PA application), final score of treatment effect, mortality, residence time and bleeding risk. The invention concludes with longitudinal data in one of two things: death or release from hospital (TIMS still under evaluation with 90 days follow-up data). Despite careful tracking and building of these databases, several advanced steps of data cleansing are required in order to perform interactive tree analysis.
Step (2): subgroup analysis is performed on the medical database described above, using subgroup analysis machine learning by recursive partitioning and decision trees, in particular using a subgroup analysis machine learning method by recursive partitioning and decision trees (Su et al, 2009). This is an advanced analytical method, well suited to the model of the invention, with data sets with natural treatment groups (e.g., low and high dose) and detailed longitudinal stroke treatment outcomes until death or hospital release. This approach has been used for several medical studies (e.g., Seibold et al, 2016; Su et al, 2011), but has not been applied to stroke studies. Based on subgroup analysis, combining the characteristics of the stroke disease species and the characteristics of the two databases;
and (3): creating a segmented model based on the subgroup analysis, the present invention creates a segmented model using the data to determine different treatment parameters for each patient subgroup, and a decision tree associated with the predictor of the subgroup;
and (4): three subgroups were determined accordingly based on NIHSS assigned patients, with the effect on bleeding risk of interaction between t-PA dose and subgroup, and specifically based on NIHSS assigned patients, with the guidance of optimizing the effect on bleeding risk of interaction between t-PA dose (high/low) and subgroup, three subgroups were determined accordingly: (1) NIHSS is less than or equal to 4, (2) NIHSS is less than or equal to 5 and less than or equal to 14, and (3) NIHSS is more than or equal to 15. For these three subgroups, the estimated difference between the log odds of high and low dose bleeding was: group 1 is-0.85 (p ═ 0.15), group 2 is 1.95(p ═ 0.02), and group 3 is 0.39(p ═ 0.56). We also performed post-effect adjustment of other factors and reduction of bias (Firth, 1993);
and (5): analysis of the above three subgroups resulted in efficacy between high and low doses, where the number of bleeding cases in some subgroups may be small. Within these three subgroups, the estimated difference in log odds of efficacy (improvement/non-improvement) between the high and low doses after post-effect adjustment for other factors is: group 1 was 0.16(p ═ 0.49), group 2 was-0.16 (p ═ 0.57), and group 3 was-0.08 (p ═ 0.80), indicating that the high and low doses did not differ significantly in efficacy. The results of the present study advocate the use of lower doses of t-PA for treatment of group 2 patients: obviously reduces the bleeding risk and simultaneously achieves the curative effect without bad effect. We provide a tool for clinical thrombolytic dose selection that balances bleeding risk and efficacy.
The invention aims to solve the problem that the traditional North American stroke research result cannot be applied to Chinese patient groups. In the scheme, two relatively similar national medical databases, namely a TIMS-China database and a SSSS database, are combined and applied, and the database is the database with the largest sample size in the current known stroke research in China. In the experiment, patients were grouped based on their NIHSS scores, independent segmented models were formed, and a decision tree was connected for each subgroup's predictor. In the analysis for subgroups we introduced a machine learning scheme and recursively partitioned with a decision tree. The test scheme is based on subjective treatment grouping, challenges are initiated to the traditional stroke research scheme according to data processing, a support tool for clinical thrombolytic dose decision is provided, and bleeding risk and curative effect are balanced.
The above description is of the preferred embodiment of the invention. It is to be understood that the invention is not limited to the particular embodiments described above, in that devices and structures not described in detail are understood to be implemented in a manner common in the art; those skilled in the art can make many possible variations and modifications to the disclosed embodiments, or modify equivalent embodiments to equivalent variations, without departing from the spirit of the invention, using the methods and techniques disclosed above. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention are still within the scope of the protection of the technical solution of the present invention, unless the contents of the technical solution of the present invention are departed.
Claims (1)
1. The construction method of the thrombolytic dose model is characterized by comprising the following steps:
step (1): incorporating existing medical databases;
step (2): performing subgroup analysis on the medical database;
and (3): creating a segmentation model based on the subgroup analysis;
and (4): based on NIHSS assigned patients, three subgroups were determined accordingly with the effect of the interaction between t-PA dose and subgroup on bleeding risk;
and (5): analyzing the three subgroups to obtain the efficacy between the high dose and the low dose;
in the step (1), the existing medical database includes: TIMS-china database containing 1440 stroke patients and SSSS database containing 11 hospitals and 2000 stroke patients, wherein the existing medical databases include scales to assess stroke severity, time/date stamp, time to TPA, final score of treatment effect, mortality, residence time and bleeding risk;
the step (2) further comprises: using subgroup analysis machine learning by recursive partitioning and decision trees;
the step (3) further comprises: based on subgroup analysis, combining the characteristics of the stroke disease species and the characteristics of the existing medical database, creating a segmented model, using data to determine different treatment parameters, corresponding to each patient subgroup, and associating a decision tree with the prediction factors of the subgroups;
in the step (4), the three subgroups are respectively: (1) NIHSS is less than or equal to 4, (2)5 is less than or equal to NIHSS is less than or equal to 14, and (3) NIHSS is more than or equal to 15, and the estimated difference between the logarithmic probability of high-dose and low-dose bleeding of the three subgroups is respectively as follows: group (1) is-0.85, where p is 0.15, group (2) is 1.95, where p is 0.02, group (3) is 0.39, where p is 0.56;
in said step (5), the estimated difference in log odds of efficacy between the high and low doses within the above three subgroups were: group (1) was 0.16, where p is 0.49, group (2) was-0.16, where p is 0.57, and group (3) was-0.08, where p is 0.80, indicating that the high and low doses did not differ significantly in efficacy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710714752.8A CN107516010B (en) | 2017-08-19 | 2017-08-19 | Construction method of thrombolytic dose model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710714752.8A CN107516010B (en) | 2017-08-19 | 2017-08-19 | Construction method of thrombolytic dose model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107516010A CN107516010A (en) | 2017-12-26 |
| CN107516010B true CN107516010B (en) | 2021-02-09 |
Family
ID=60723198
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710714752.8A Active CN107516010B (en) | 2017-08-19 | 2017-08-19 | Construction method of thrombolytic dose model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107516010B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105787252A (en) * | 2016-01-18 | 2016-07-20 | 贡京京 | Medical decision supporting method and system |
| CN107073292A (en) * | 2014-11-03 | 2017-08-18 | 溶栓科学有限责任公司 | Method and composition for safely effectively thrombolysis |
-
2017
- 2017-08-19 CN CN201710714752.8A patent/CN107516010B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107073292A (en) * | 2014-11-03 | 2017-08-18 | 溶栓科学有限责任公司 | Method and composition for safely effectively thrombolysis |
| CN105787252A (en) * | 2016-01-18 | 2016-07-20 | 贡京京 | Medical decision supporting method and system |
Non-Patent Citations (2)
| Title |
|---|
| 同严重程度急性缺血性卒中患者静脉溶栓预后的影响因素分析;严晓波 等;《浙江大学学报（医学版）》;20150131;第54-60页 * |
| 心源性脑栓塞出血转化的危险因素及其与预后的关系;吴章薇 等;《中国康复理论与实践》;20160531;第22卷(第5期);第559-562页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107516010A (en) | 2017-12-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zuraw | Hereditary angioedema with normal C1 inhibitor: four types and counting | |
| Lopez et al. | Genomic evidence for local adaptation of hunter-gatherers to the African rainforest | |
| Flaherty et al. | Location and outcome of anticoagulant-associated intracerebral hemorrhage | |
| Undas et al. | Altered fibrin clot structure in patients with advanced coronary artery disease: a role of C‐reactive protein, lipoprotein (a) and homocysteine | |
| Liu et al. | Proteomic profile of edible bird’s nest proteins | |
| González et al. | Evaluation of inter-day and inter-individual variability of tear peptide/protein profiles by MALDI-TOF MS analyses | |
| Hwang et al. | Effect of the skincare product on facial skin microbial structure and biophysical parameters: A pilot study | |
| Gervais et al. | Validation of a 30-year-old process for the manufacture of L-asparaginase from Erwinia chrysanthemi | |
| Nouraei et al. | Comparative analysis of virulence factors of homozygous and heterozygous strains of Candida albicans vaginal isolates | |
| Yu et al. | Vitreous proteomic analysis of idiopathic epiretinal membranes | |
| Wielsch et al. | Rapid validation of protein identifications with the borderline statistical confidence via de novo sequencing and MS BLAST searches | |
| Yang et al. | Interaction between cervical microbiota and host gene regulation in caesarean section scar diverticulum | |
| CN107516010B (en) | Construction method of thrombolytic dose model | |
| Gonçalves et al. | Revealing Candida glabrata biofilm matrix proteome: global characterization and pH response | |
| Sosa‐Acosta et al. | Proteomics of ZIKV infected amniotic fluids of microcephalic fetuses reveals extracellular matrix and immune system dysregulation | |
| Millea et al. | Comparative profiling of human saliva by intact protein LC/ESI-TOF mass spectrometry | |
| Wei et al. | Fungal microbiome related to mycotoxin contamination in medicinal and edible seed Semen Persicae | |
| US20100280759A1 (en) | Mass spectrometer output analysis tool for identification of proteins | |
| Cai et al. | Identification of immunogenic cell death-related gene classification patterns and immune infiltration characterization in ischemic stroke based on machine learning | |
| CN116895359A (en) | A method and application of traditional Chinese medicine combination analysis based on analytic hierarchy process | |
| Barsuren et al. | Differences in serum protein 2D gel electrophoresis patterns of P rzewalski's (M ongolian wild horse) and thoroughbred horses | |
| Du et al. | Nomogram model for predicting the prognosis of high-grade glioma in adults receiving standard treatment: a retrospective cohort study | |
| Xiao et al. | Association between basal platelet count and all-cause mortality in critically ill patients with acute respiratory failure: a secondary analysis from the eICU collaborative research database | |
| Huang et al. | Disentangling the potential roles of the human gut mycobiome and metabolites in asthma | |
| Pruinelli et al. | Clustering the whole-person health data to predict liver transplant survival |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |