CN103970973B - A kind of simulation RCT analysis methods based on True Data - Google Patents
A kind of simulation RCT analysis methods based on True Data Download PDFInfo
- Publication number
- CN103970973B CN103970973B CN201310038845.5A CN201310038845A CN103970973B CN 103970973 B CN103970973 B CN 103970973B CN 201310038845 A CN201310038845 A CN 201310038845A CN 103970973 B CN103970973 B CN 103970973B
- Authority
- CN
- China
- Prior art keywords
- odds
- power
- database
- intervention
- analysis
- 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.)
- Expired - Fee Related
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to randomization check experiment research field, it is related to a kind of simulation RCT analytic approach based on True Data;The analytic approach is a kind of for the data analysing method for intervening comparative studies CER exploration inventions, is analyzed by the outcome data for having received different dry pre-generated;It includes creating database, randomized grouping, statistical analysis, the step such as obtain a result.Analysis method of the present invention is to carry out intervention effect comparative analysis to the clinical True Data for producing, and can effectively control Confounding Factor, and analysis result has preferable epitaxy, and sample size requirements is low, moreover it is possible to while analyzing multiple final result indexs.The present invention can be used for the comparing of the clinically intervention effect and security of different interference methods or strategy, can provide a kind of new evidence-based medical for clinical decision.
Description
Technical field
The invention belongs to test data analyzer method field, it is related to a kind of simulation RCT analysis methods based on True Data;Should
Method is directed to randomized controlled trial comparative studies(Comparative Effectiveness Research,CER)Final result number
According to(Actual effect data)It is analyzed, for clinical decision provides circumstantial evidence.
Background technology
Evidence-based medicine EBM is the medical science based on evidence, and it is intended to help clinical practice using existing best evidence,
Solve clinical problem.Classification according to GRADE in 2011 to evidence quality, based on high-quality randomization check experiment research
(Randomized controlled trials,RCTs)Meta analyses and the RCT that implements of system evaluation and strict design
The high-quality evidence of comparitive study is considered to be, is the optimal path for obtaining evidence;And be in recent years using the CER of actual effect data
The new field being analyzed using the clinical actual effect data for producing of generation, but the equal base of the data analysing method of the studies above
In classical epidemiology and statistical basic skills.Due to RCT expenses, ethics, research object include it is more strict etc. because
The limitation of element, few, the effective limited, extrapolation of research quantity of RCT is limited, and this kind of evidence quantity is also limited.Additionally, design is tight
The careful and strict observational study of implementation is that CER obtains one of evidence approach, but because observational study does not carry out randomization
Packet so that different therapeutic modality therapeutic effects relatively become sufficiently complex.In order to exclude the influence of Confounding Factor, tradition side
Method is that the effect of Confounding Factor is controlled using complex multiplicity method, but observation by limited quantity is ground
Study carefully, accurately judge that the difference of the curative effect of different therapeutic modalities is still highly difficult.Additionally, Rosenbaum and Rubin are carried in nineteen eighty-three
The method of the Propensity Score (propensity score, PS) for going out also to a certain extent solve observational study in mix because
Unbalanced problem between plain group;However, due to the limitation of observational study itself, its result of study is defined as low-quality by GRADE
Amount evidence.Therefore, still there is the decision problem of very many diagnosis and treatment to be badly in need of evidence-based medicine EBM in the practice of current clinical medicine to demonstrate,prove
According to particularly recently as a large amount of novel medicines, the appearance of new Clinics so that at present carry out RCTs quantity far from
Meet clinical demand.Generally before RCTs implementations, clinical practice has generated the outcome data of substantial amounts of diagnosis and treatment,
Its patient be it is non-selective, how it is scientific and effective utilize above-mentioned actual effect data, generation be available for clinical decision refer to it is new
Evidence-based medical is problem to be resolved at present.
The content of the invention
The purpose of the present invention is the defect and deficiency for overcoming prior art, there is provided a kind of simulation RCT based on True Data
Analytic approach;The analytic approach is directed to intervention effect comparative studies(Comparative Effectiveness Research,CER)'s
Data are analyzed, especially to having received the outcome data that different intervention strateges are produced(Actual effect data)It is analyzed, can uses
In the comparing of clinically different intervention strateges or the tactful effect & Safety of implementation, for clinical decision provides new evidence-based doctor
Learn evidence.
The method of the present invention, for the dependency number for having received intervention to be measured and the object of other interventions for clinically producing
According to(Including essential characteristic data and outcome data), through simulating randomized grouping, scheme collection is met according to RCT after packet every time
Analysis strategy (Per-protocol Analysis strategies/PP strategies) carries out intervention effect comparing, obtains discrepant experiment frequency
The ratio between experiment frequency percentage of several percentage and indifferences(Odds values)And its 95%CI carries out different intervention sides as judgement
Whether effect and security have discrepant foundation between case or mode;Work as odds<When the upper limit of 1 or 95%CI is less than 1, it is believed that
The fruit indifference of two kinds of intervention strateges or mode;Work as odds>When the lower limit of 1 or 95%CI is more than 1, it is believed that two kinds of intervention strateges
Or the effect of mode is variant;Odds=1 or 95%CI includes 1, wouldn't provide clear and definite conclusion.In embodiments of the invention, enter
100 simulation randomized groupings of row, carry out intervention effect comparative analysis after packet every time.
In the present invention, the principle that illustrates this analytic approach by 4 steps, the feasibility for proving this analytic approach, stability with
And the uniformity of result of determination and truth;Because the final result variable and Confounding Factor of True Data reflection intervention effect are in group
Between difference have no way of finding out about it, it is impossible to judge this analytic approach calculate odds and 95%CI distribution and real group between effect quantity
(EffectSize, E/S)And power of a test(power)The relation of size, cannot also judge to judge Different therapeutical effect using this analytic approach
Result and truth uniformity, therefore, the present invention is verified using the method for analogue data;
Specifically, the simulation RCT analysis methods based on True Data of the invention, it is characterised in that it includes step:
(1)First, two kinds of means of intervention and final result variable are contained using what computer simulation clinical intervention was produced(It is continuous to become
As a example by amount)The variant database of effect:With body-mass index(BMI, kg/m2)After two kinds of different means of intervention are received
Final result variable carry out Different therapeutical effect comparing;Experimental group in database(n1)And control group(n2)Sample size is 50, experimental group
Mean is 27.25kg/m2, and standard deviation is 0.85kg/m2, and control group mean is 30.26kg/m2, and standard deviation is 0.85kg/m2;
Set up and assume:H0There is no difference between two groups for final result variable mean;H1, the mean of final result variable is variant between two groups(Bilateral
Inspection);It is again that random sequence carries out randomized grouping with the numbering of research object, then to according to PP strategy error-free point of entrance point
The research object BMI of analysis compares between carrying out two groups;Repeat the comparing of 100 randomized groupings and BMI;Finally by calculating 100
Refuse null hypothesis experiment frequency in secondary BMI comparative analysis result and do not refuse the ratio (odds) of null hypothesis experiment frequency, repeat
The process obtains odds=2.03 for 100 times, and 95%CI is (1.20,3.09), odds>The lower limit of 1 and 95%CI be more than 1, show this two
The effect for planting means of intervention is variant;Result shows that the odds values and its 95%CI for obtaining can reflect the true poor of intervention effect
It is different;
(2)Whether the distribution for observing odds and 95%CI is consistent with E/S sizes and Power distributions;According to power from 0.5 to
0.85, the different simulations that E/S is corresponded to therewith create the simulation number of the variant database of intervention effect 8 and intervention effect indifference
It is analyzed according to 8, storehouse and respectively, is calculated odds and its 95%CI;Result shows changes of the odds and its 95%CI with power
Change and good linear relationship is presented, i.e., when being analyzed to the discrepant database of intervention effect, power is bigger, is calculated
Odds and 95%CI it is bigger;When database to intervention effect indifference is analyzed, power is bigger, the odds being calculated
It is smaller with 95%CI;No matter intervention effect whether there is difference, and when power is 0.5, close to 1, it 95% for the odds being calculated
CI includes 1;
(3)The distribution of observation Odds and its 95%CI is from the relation of real difference between group and power in different sample sizes(n=
50,100,500 and 1000)Under whether stablize;According under different sample sizes, the different simulations of E/S and power create multiple simulations
Database is simultaneously analyzed respectively, and 1 is all higher than to the odds that the variant database analysis of intervention effect under different sample sizes is obtained,
Odds values and its 95%CI increase with the increase of power;Intervention effect indifference database analysis under different sample sizes is obtained
The odds values for arriving are respectively less than 1, and odds values and its 95%CI reduce with the increase of power;Result shows, using this analysis
The method Odds that obtains of analysis and its 95%CI carry out intervention effect difference judges between two groups result and truth uniformity compared with
It is high and stability is preferable;
(4)Prove in the case of small sample(n=100), either two groups of numbers of sample size equal (n1=50, n2=50)
According to storehouse or sample size database not etc.(N1=30, n2=70), harmony of the Confounding Factor between two groups be preferable;It is right first
Database carries out 100 randomized groupings, then to according to wrong point of sample of PP strategy removals, then being ground into analysis to error-free point
Study carefully whether object Confounding Factor carries out hypothesis testing between two groups with difference, as a result show, Confounding Factor indifference between group
Possibility more than 95%(The possibility of Confounding Factor Homogeneity between groups is 97%, sample size when being respectively equal sample size
When Confounding Factor in a balanced way possibility be 99%).
Simulation RCT analysis methods based on True Data of the invention CER researchs high-level with the offer of prior art are followed
The analysis method for demonstrate,proving medical evidence is compared, with advantages below:
1)This method carries out intervention effect comparative analysis to the True Data that clinical intervention is produced, rather than to data or
It is that the data delivered are analyzed;
2)Effective control Confounding Factor:Although observational study can be by with the Confounding Factor between PS methods equilibrium group
The data of observational study are made to reach the effect of " close to randomization data ", but PS methods are only used for observable variable, it is right
The bias that potential unknown Confounding Factor causes is helpless;Only when all Confounding Factors are all Observables, PS methods can
Obtain the unbiased esti-mator of Treatment Effects;And this method utilizes RCT principles, clinically existing non-study data are simulated with
Machineization is grouped, and using unknown and known Confounding Factor between the balanced group of the method for randomized grouping, excludes each randomized grouping
The influence of Confounding Factor that may be present when different interventions and its effect compare afterwards;The calculating of statistic Odds and its 95%CI is
What two groups of intervention effects after based on randomized grouping were relatively carried out so that the good control that the influence of Confounding Factor is obtained;
In relative to single RCT research and designs and implementation process, due to being randomly assigned hiding and the factor such as blind imperfection for scheme, lead
Causing result of study can not accurately reflect the intervention effect of means of intervention to be studied, the original that this method is analyzed using approximate Meta
Reason, 100 results of intervention effect comparison in difference hypothesis testing of integrated use are different as assessment using odds values and its 95%CI
Foundation of the intervention effect with the presence or absence of difference between means of intervention;
3)Result has preferable epitaxy:During RCT research and designs, there is strict including to the selection of research object
Exclusion standard;Therefore its representativeness is relatively poor, it is impossible to reflects the overall picture of disease, causes the epitaxy of result of study poor;Base
In RCT Meta analyze, its analysis result also by include research multiple independent RCT research epitaxy influenceed;And this method
It is to be analyzed based on clinically existing non-study data, research object is non-selective, and result of study has preferable
Epitaxy;
4)Sample size requirements is low:Because RCT researchs are limited by two class statistics mistakes, it is desirable to have enough sample size drops
The possibility of I classes mistake and II class mistakes is made in low hypothesis testing;As shown in Figure 1 and Figure 2, in the method under different sample sizes
Odds values and its 95%CI are consistent with the trend that power changes, i.e., the odds for being obtained to the discrepant database analysis of intervention effect
Value is all higher than 1, odds values and its 95%CI and is continuously increased as power is continuously increased;To intervention effect indifference database
The odds values that analysis is obtained are respectively less than 1, odds values and its 95%CI is continuously increased with power and constantly reduces;Result shows,
In the case of sample size is less(Minimum n=50)Also the intervention effect of different means of intervention can be whether there is with this analytic approach
Difference is analyzed;
5)Multiple final result indexs can simultaneously be analyzed:Need to explicitly define the main and secondary of research in the RCT research and design stages
Want the calculating of final result index, sample size is carried out also based on major fate's index, therefore, research conclusion can only show intervention to be studied
Intervention effect of the mode to certain specific final result index;And this method can refer to the final result that multiple may be influenceed by intervention to be studied
Mark is studied, so as to synthetically evaluate the intervention effect of means of intervention to be studied.
Brief description of the drawings
Fig. 1 shows what the odds values that are obtained to intervention effect indifference database analysis and its 95%CI changed with power
Trend, as a result shows, the odds values and its 95%CI of intervention effect indifference database are as power increases in different sample sizes
Plus and reduce.
Fig. 2 shows what the odds values that are obtained to the variant database analysis of intervention effect and its 95%CI changed with power
Trend, wherein, n represents sample size size, and power represents power of a test, and Lower represents the lower limit of odds values 95%CI, Upper tables
Show the upper limit of odds values 95%CI;
Result shows that the odds values and its 95%CI of the variant database of intervention effect are with power in different sample sizes
Increase and increase.
Specific embodiment
Embodiment 1
It is embodied using the programming tool of Stata11.0 statistical analysis softwares, using Stata11.0 software programmings
Complete program is formed, is analyzed accordingly using analogue data.Not comprising the mistake for creating simulated database during practical application
Journey.This analytic approach can directly be carried out repeating randomized grouping, Odds and 95%CI calculating, mixed for different true clinical datas
What the analyses such as the balance between the two groups analysis of miscellaneous factor, final offer Odds and 95%CI, and Confounding Factor balance between the two groups compared
As a result(Unbalanced probability between Confounding Factor group)As the foundation for judging different intervention effects;Comprise the following steps that:
1st, database is created
Using PS(powerandsamplesizeprogram)Software calculates different sample sizes(N=50,100,500, and
1000)Under, the size of the E/S corresponding to Power(Table 1).According to the difference of power, E/S under different sample sizes, using calculating
Machine simulation produces the database comprising Confounding Factor variable, final result variable and different means of intervention, contains under different sample sizes
16 databases, and database of intervention effect indifference each 8 variant comprising intervention effect;With height(cm)As experiment
During as a example by Confounding Factor that may be present, evaluate Confounding Factor whether balanced comparable between group;With body-mass index
(BMI, kg/m2)Intervention effect comparison in difference is carried out as the final result variable received after two kinds of different means of intervention;Below with sample
When this amount is 100, power is 0.5, presentation database creates process as a example by the establishment of the variant database of intervention effect, other
The establishment of database can be created according to the process, only need to change corresponding relevant parameter during establishment;Actually should
During, the step is skipped, since randomized grouping.
The feature of the simulated database of table 1:Effect value and hypothesis testing power of a test between sample size, group
Tab1Characteristicsofsimulationdatasets:samplesize、effectsizeandpower
NotesSignificant:Theeffectivenessoftwodifferenttreatmentshavesignific
antdifferences;No significant:Theeffectivenessoftwodifferenttreatmentshavenos
ignificantdifferences(Indicate conspicuousness:Effect value between two different disposal groups has significant difference;Without aobvious
Work property:Effect value no difference of science of statistics between two different disposal groups).
(1)Create BMI data
It is that 100, mean and standard deviation are respectively 27.5kg/m to set sample size2And 6kg/m2, using invnorm
(uniform ()) * σ+μ order establishments meet the control group BMI data of normal distribution and calculate standard deviation;Obtaining control group
On the basis of BMI data and its standard deviation (sd), according to sample size in table 1 be 100, power be 0.5 when, the variant number of curative effect
According to the corresponding E/S sizes in storehouse, establishment meets the experimental group BMI data of normal distribution;
(2)Produce height data
Sample size is set as 100, mean and standard deviation are respectively 174.5cm and 4cm, using invnorm (uniform
()) * σ+μ orders are respectively created the height data of the control group and experimental group that meet normal distribution;Control group and experimental group height
Though data are created using identical parameter, due to unassigned species subnumber, therefore the experimental group that produces and control group
Height data and unequal;
(3)Generation database
Ordered using stack, merge etc., merge height and BMI data form database;
2nd, randomized grouping
Use simple randomizing packet mode to simulation produce database, using recording mechanism as random sequence number carry out with
Machineization is grouped;
3rd, statistical analysis
Data analysis use Stata11.0 statistical softwares, with PP strategy retain randomized grouping after with it in former data
Implement the consistent sample of intervention group at place, remove wrong point of sample, continuous data is with mean ± standard deviationRepresent;
Compare between two groups using two independent samples t tests, with α=0.05 as inspection level, P<0.05 is that difference is statistically significant;
(1)Compare intervention effect difference
To compare two kinds of means of intervention its intervention effects with the presence or absence of difference, simulated database is carried out first 100 times with
Machineization is grouped, and carries out hypothesis testing to error-free point after each randomized grouping BMI entered in the sample of analysis, then distinguishes
Calculate and refuse null hypothesis in 100 BMI hypothesis testing results and do not refuse the frequency of null hypothesis;
(2)Odds values and 95%CI
According to 100 results of BMI hypothesis testings, it is calculated refusal null hypothesis frequency and does not refuse null hypothesis frequency
Ratio(Odds values), and by repeating the calculating process 100 times of odds values, obtain distribution and its 95%CI of odds values;
(3)Confounding Factor is harmonious between comparative group
For compare two groups of sample sizes it is equal when randomized grouping after Confounding Factor it is whether balanced comparable between group, first to two
Group sample size is equal and is 50 simulated database and carries out 100 randomized groupings, and to error-free after each randomized grouping
The height divided in the sample for entering analysis carries out hypothesis testing;Then calculate respectively and refuse in 100 height hypothesis testing results
Exhausted null hypothesis and the frequency of null hypothesis is not refused;
At the same time, it is equilibrium of the Confounding Factor between group after randomized grouping when the above-mentioned two groups of sample sizes of checking are unequal
Property, be first according to above-mentioned database creation process simulate an experimental group sample size be 30, the data that control group sample size is 70
Storehouse, then carries out 100 randomized groupings, and enter analysis to error-free point after each randomized grouping to the simulated database
Height in sample carries out hypothesis testing;Finally calculate respectively and refuse in 100 height hypothesis testing results null hypothesis and not
Refuse the frequency of null hypothesis;
4th, result
Relevant parameter in table 1 creates database, and it is 50,100,500 that table 2,3,4,5 describes sample size respectively,
The essential characteristic of database when 1000;In all database, height equal indifference between group(P>0.05);Have in intervention effect
In discrepancy database, BMI has differences (P between group<0.05);In intervention effect indifference database, BMI indifferences between group
It is different(P>0.05).
The odds values and its 95%CI obtained to the analysis of whole simulated databases according to this analytic approach are as shown in table 6, to intervening
The odds values that the variant database analysis of effect is obtained are all higher than 1, and the database analysis of intervention effect indifference is obtained
Odds values are respectively less than 1.
16 simulated database final result variables and Confounding Factor variable essential characteristic of the sample size 50 of table 2
Tab2Characteristicsofoutcomevariablesandconfoundervariablesinsixteensimulatio
ndatasets of50samples
16 simulated database final result variables and Confounding Factor variable essential characteristic of the sample size 100 of table 3
Tab3Characteristicsofoutcomevariablesandconfoundervariablesinsixteensimulatio
ndatasetsof100samples
16 simulated database final result variables and Confounding Factor variable essential characteristic of the sample size 500 of table 4
Tab4Characteristicsofoutcomevariablesandconfoundervariablesinsixteensimulatio
ndatasets of500samples
16 simulated database final result variables and Confounding Factor variable essential characteristic of the sample size 1000 of table 5
Tab5Characteristicsofoutcomevariablesandconfoundervariablesinsixteensimulatio
ndatasets of1000samples
Odds values and 95%CI under the different sample sizes of table 6
Tab6Oddsand95%CIindifferentsamplesgroups
Notes Significant differences:The effectiveness of two different
treatments have significant differences or not;Y:The effectiveness of two
different treatments have significant differences;N:The effectiveness of two
different treatments have no significant differences。
(Illustrate significant difference:Whether the effect value between two different disposal groups has significant difference;Y:Two differences
Effect value between treatment group has notable difference;N:Effect value no significant difference between two different disposal groups.)
Claims (7)
1. a kind of simulation RCT analysis methods based on True Data, it is characterised in that it includes step,
(1) first, the intervention effect containing two kinds of means of intervention and final result variable for being produced using computer simulation clinical intervention
Variant database:Intervention effect is carried out as the final result variable received after two kinds of different means of intervention using body-mass index BMI
Fruit comparison in difference;It is again that random sequence carries out randomized grouping with the numbering of research object, then to tactful error-free point according to PP
Compare between carrying out two groups into the research object body-mass index BMI for analyzing;Repeat randomized grouping and body-mass index
The comparing of BMI;Finally by refusal null hypothesis experiment frequency in calculating body-mass index BMI comparative analysis results and do not refuse
Null hypothesis tests the ratio calculation Odds of frequency;
Repeat above-mentioned Odds calculating process and obtain Odds values and 95% fiducial limit scope (95%CI);
(2) observation Odds and 95% fiducial limit scope (95%CI) distribution whether real difference E/S sizes and assurance between group
Degree Power distributions are consistent;Odds and 95% fiducial limit scope (95%CI) change with the change of power of a test power, and present
Linear relationship;
(3) real difference E/S and power of a test power between the distribution of observation Odds and its 95% fiducial limit scope (95%CI) and group
Relation whether stablize under different sample sizes;
(4) in the case of small sample, including the equal database of sample size and sample size database not etc., compare mix because
The plain harmony between two groups.
2. the method as described in claim 1, it is characterised in that in the step (1), experimental group n1 and control group in database
N2 sample sizes are 50, wherein, H0There is no difference, H between two groups for final result variable mean1, the mean of final result variable is between two groups
It is variant.
3. the method as described in claim 1, it is characterised in that in the step (1), described randomized grouping repeats 100
It is secondary.
4. the method as described in claim 1, it is characterised in that in the step (2), according to power of a test power values and therewith
The different simulations of real difference E/S create the mould of the variant database of intervention effect 8 and intervention effect indifference between correspondence group
Intend database 8 and be analyzed respectively, be calculated Odds and its 95% fiducial limit scope (95%CI).
5. the method as described in claim 1, it is characterised in that in the step (2), to the discrepant database of intervention effect
When being analyzed, power of a test power is bigger, and the Odds and 95% fiducial limit scope (95%CI) being calculated are bigger;To intervening
When the database of effect indifference is analyzed, power of a test power is bigger, the Odds being calculated and 95% fiducial limit scope
(95%CI) is smaller;No matter intervention effect whether there is difference, and when power of a test power is 0.5, the Odds being calculated is approached
1, its 95% fiducial limit scope (95%CI) includes 1.
6. the method as described in claim 1, it is characterised in that in the step (3) is true between group according under different sample sizes
The different simulations of real difference E/S and power of a test power create multiple simulated databases and are analyzed respectively, to different sample sizes
The Odds that the lower variant database analysis of intervention effect is obtained is all higher than 1, Odds values and its 95% fiducial limit scope (95%CI)
Increase with the increase of power of a test power;The Odds obtained to intervention effect indifference database analysis under different sample sizes
Value is respectively less than 1, and Odds values and its 95% fiducial limit scope (95%CI) power of a test reduce with the increase of power.
7. the method as described in claim 1, it is characterised in that in the step (4), database is carried out first 100 times with
Machineization is grouped, then to according to wrong point of sample of PP strategy removals, then existing to the error-free point of research object Confounding Factor for entering analysis
Whether hypothesis testing is carried out between two groups with difference.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310038845.5A CN103970973B (en) | 2013-01-31 | 2013-01-31 | A kind of simulation RCT analysis methods based on True Data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310038845.5A CN103970973B (en) | 2013-01-31 | 2013-01-31 | A kind of simulation RCT analysis methods based on True Data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103970973A CN103970973A (en) | 2014-08-06 |
CN103970973B true CN103970973B (en) | 2017-07-07 |
Family
ID=51240463
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310038845.5A Expired - Fee Related CN103970973B (en) | 2013-01-31 | 2013-01-31 | A kind of simulation RCT analysis methods based on True Data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103970973B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108242268B (en) * | 2017-07-18 | 2021-02-09 | 嘉兴太美医疗科技有限公司 | Error identification and correction method for randomized grouping and treatment method distribution of clinical study subjects |
US20220245219A1 (en) * | 2019-07-19 | 2022-08-04 | Ebay Inc. | Sample Delta Monitoring |
CN112907364A (en) * | 2021-04-01 | 2021-06-04 | 重庆度小满优扬科技有限公司 | Assessment method and device of wind control analysis model |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101334770A (en) * | 2007-06-28 | 2008-12-31 | 科凌力医学软件(深圳)有限公司 | Evidence-based medical data base establishment method and related evidence-based medical disease diagnostic method |
-
2013
- 2013-01-31 CN CN201310038845.5A patent/CN103970973B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101334770A (en) * | 2007-06-28 | 2008-12-31 | 科凌力医学软件(深圳)有限公司 | Evidence-based medical data base establishment method and related evidence-based medical disease diagnostic method |
Non-Patent Citations (3)
Title |
---|
代谢综合征药物治疗的系统评价;勾忠平;《中国博士学位论文全文数据库》;20080515(第5期);全文 * |
循证医学对中医药治疗糖尿病足的临床疗效系统评价研究;李树法;《中国博士学位论文全文数据库》;20111215(第12期);全文 * |
阿加曲班治疗急性缺血性卒中的系统评价;孔燕;《中国优秀硕士学位论文全文数据库》;20110815(第8期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN103970973A (en) | 2014-08-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Lycett et al. | Toward a “quantitative genetic” approach to lithic variation | |
CN107369151A (en) | System and method are supported in GISTs pathological diagnosis based on big data deep learning | |
Zhang | Big data and clinical research: perspective from a clinician | |
Li et al. | Identifying patent conflicts: TRIZ-led patent mapping | |
Chatziioannou et al. | Exploiting statistical methodologies and controlled vocabularies for prioritized functional analysis of genomic experiments: the StRAnGER web application | |
CN105843870B (en) | The analysis method and its application of repeatability and reproducibility | |
Kennedy et al. | Why data coherence and quality is critical for understanding interareal cortical networks | |
CN103970973B (en) | A kind of simulation RCT analysis methods based on True Data | |
Zheng et al. | Elevated CO2, warming, N addition, and increased precipitation affect different aspects of the arbuscular mycorrhizal fungal community | |
CN104866710B (en) | The method for predicting Cytochrome P450 1A2 inhibitor inhibition concentrations | |
Meng et al. | The application of machine learning techniques in clinical drug therapy | |
CN108931815A (en) | A kind of hierarchical identification method of lithology | |
Esteves et al. | The ATB framework: Quantifying and classifying epistemic strategies in tangible problem-solving tasks | |
Radinović et al. | Outlining the knapping techniques: Assessment of the shape and regularity of prismatic blades using elliptic Fourier analysis | |
Sobhan et al. | Explainable machine learning to identify patient-specific biomarkers for lung cancer | |
JP2007156721A (en) | Effective factor extraction system | |
Soberón et al. | Visualizing species richness and site similarity from presence-absence matrices | |
Pernice et al. | A new computational workflow to guide personalized drug therapy | |
Tsakalerou et al. | Systematic reviews and metastudies: A meta-analysis framework | |
Gupta et al. | A comprehensive review on systematic and Meta-analysis methods. | |
CN104978497B (en) | A kind of large sample high throughput biological data association analysis method | |
DE112018004558T5 (en) | MODELING GENETIC DISEASES | |
Sratthaphut et al. | Principal component analysis coupled with artificial neural networks for therapeutic indication prediction of Thai herbal formulae | |
CN107679366A (en) | A kind of computational methods of genome mutation data | |
CN110322945B (en) | Intelligent screening method for plant compounding |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170707 Termination date: 20210131 |
|
CF01 | Termination of patent right due to non-payment of annual fee |