CN104091062A - Diagnostic test Meta analysis method based on power of test - Google Patents
Diagnostic test Meta analysis method based on power of test Download PDFInfo
- Publication number
- CN104091062A CN104091062A CN201410311640.4A CN201410311640A CN104091062A CN 104091062 A CN104091062 A CN 104091062A CN 201410311640 A CN201410311640 A CN 201410311640A CN 104091062 A CN104091062 A CN 104091062A
- Authority
- CN
- China
- Prior art keywords
- test
- meta
- power
- research
- 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.)
- Pending
Links
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a diagnostic test Meta analysis method based on power of test. The method is characterized in that the power of test serves as a main clue which is throughout systematic review evaluation with diagnostic test Meta analysis serving as the main body, and integration of the power of test, quality evaluation, total effect size merging, sensitivity analysis and Meta analysis aftereffect evaluation is achieved, wherein the quality evaluation, the total effect size merging, the sensitivity analysis and the Meta analysis aftereffect evaluation are introduced into traditional Meta analysis for study. According to the method, the conversion of the quality evaluation introduced for study from quality to quantity is achieved, and the shortage caused by effect size weighing and merging based on sample size is made up for. Accordingly, the power of test is expanded and extended in the field of evidence-based medicine, and meanwhile the method is beneficial to more science standardization, more accuracy materialization and more totalization of the diagnostic test Meta analysis.
Description
Technical field
The present invention relates to evidence-based medicine EBM field, relate in particular to a kind of method that diagnostic test Meta based on power of test analyzes.
Background technology
Diagnostic test refers to that the means such as application experiment, instrument and equipment check clients, thereby the detection method diagnosing the illness, mainly comprise various laboratory examinations (biological chemistry, immunology, microbiology, parasitology and pathology etc.), iconography (ultrasound wave, CT, x-ray and MRI etc.), inspect by instrument (cardiogram, electroencephalogram, radioisotope scanning and scope etc.), and inquiry medical history, physical examination etc., its fundamental purpose is by patient and suspicious ill but actual anosis people difference, to the patient who makes a definite diagnosis is given to corresponding clinical intervention.But due to hygiene and health resource-constrained, the often due to illness restriction such as kind, patient compliance and ethics clinically in addition, therefore the quantity of the enforcement that many diagnostic means can not be one by one and research object can not unlimited increase.Given this, the Meta of diagnostic test analyzes and arises at the historic moment, to the evidence by means of evidence-based medicine EBM, based on clinical practice, select simple and easy to do, relatively cheap, the noninvasive test mode of safety, in the assessment of suitably design, shown on the diagnostic mode of effective hygiene and health service thereby limited medical resource is reasonably used in to those.
But in current diagnostic test Meta analysis and system evaluation process, mainly to carry out qualitative evaluation by means of set convention to including research in based on researcher both at home and abroad to including the assessment of quality of research in, in the process of assessment because of researcher get sth into one's head and usually cause evaluation bias, and in the time of quality evaluation, for want of between quantizating index researcher, be also difficult to unified suggestion, so greatly reduced the statistical power that Meta analyzes.Existing Meta analyzes in the time merging gross effect amount, often taking include the sample size of research in or based on the derivative standard deviation of sample size etc. as weight object, although weighted sample amount can make the research of large sample obtain larger weight to a certain extent, but ignore the weight to quality of research, thereby cause weight bias, thereby affected science and reliability that diagnostic test Meta analyzes.In recent years, the quantity cumulative year after year that Meta analyzes, but because not yet there being the quantitative criteria for the quality evaluation of Meta analysis own specially, Meta quality is uneven, cause the significant wastage of Evidence-based medicine resource, Meta analytical probation intensity and public credibility are reduced, the serious development that is restricting evidence-based medicine EBM.
Summary of the invention
A kind of method that provides diagnostic test Meta based on power of test to analyze is provided fundamental purpose of the present invention, it is characterized in that serving as theme through analyzing in the systematic review of main body is evaluated taking diagnostic test Meta with power of test, realized the integrated integration of quality evaluation, effect quantity merging, sensitivity analysis and the aftereffect evaluation of including research in power of test and traditional Meta analysis in.Analyze including in the quality evaluation process of research at Meta, reference frame using power of test as evaluation study quality, based on power of test, the each research of including in is quantized to scoring, overcome because of tradition based on sample size as judge quality of research defect just; In Meta analyzes the effect quantity merging process of included in research, using power of test as the weight foundation that merges effect quantity, reduce the systematic error causing according to merging effect quantity as weight because of simple dependence sample size; In the sensitivity analysis of analyzing at Meta, using different powers of test as the normative reference that reconsolidates effect quantity, the intensity of variation based on gross effect and heterogeneity is judged the robustness of result of study, has improved the recall rate of bias in analyzing with Meta; After traditional Meta has analyzed, using power of test as Meta is analyzed to the criterion that quality aftereffect is evaluated, strengthen science that diagnostic test systematic review is evaluated and comprehensive.
Beneficial effect of the present invention is: the method that the Meta based on power of test analyzes, from including the evaluation of quality of research, the merging of gross effect amount, the aftereffect evaluation that sensitivity analysis is analyzed to Meta in, realize from traditional qualitative leap to the individuation assessment taking qualitative and quantitative combination as feature, made up the deficiency causing based on sample size balance merging effect quantity; Strengthen the feasibility of sensitivity analysis introducing power of test; Also start the precedent of the assessment to the quality of Meta analysis own.The integrated integration that power of test and Meta analyze, expansion and the innovation in evidence-based medicine EBM field to power of test, also be the development and improvement that existing Meta is analyzed, be conducive to make diagnostic test Meta to analyze more scientific and normalization, accurately specific and comprehensive generalization.
Brief description of the drawings
Accompanying drawing 1 is implementing procedure figure of the present invention.
Embodiment
Concrete grammar of the present invention is provided by following examples and accompanying drawing 1 thereof.
Accompanying drawing 1 is the process flow diagram of the method for the analysis of the Meta based on power of test of the proposition according to the present invention.Describe the concrete grammar proposing according to the present invention in detail below in conjunction with accompanying drawing:
(1) before enforcement diagnostic test Meta analyzes, should first determine the power of test formula that is applicable to this diagnostic test, the present invention adopts internationally recognized diagnostic test usefulness
p=(1-β), wherein β=(1-NORMSDIST (Z
β)) * 2,
,
,
,
,
, in formula, the clinical meaning of a, b, c and d sees table, and different I type mistake (α) value of the convention of the international statistics of foundation and clinical practice is also different conventionally.
(2) analyze including in the quality evaluation process of research at Meta, first include on the basis of quality of research at traditional qualitative evaluation, increase taking power of test as normative reference to each research individuation evaluation one by one of including in, by means of quantizing accurately score to judge the height of including quality of research in.If include the power of test of research in
pbe 1.0, illustrate that this research sample size has enough usefulness to reach the difference of screening between diagnostic test and goldstandard; If power of test
pbe between 0.9～1.0, can think that this research sample size can still can effectively screen the difference of the two; If power of test
pbe between 0.75～0.9, can think that this research sample size can still can screen the two difference to a certain extent; If power of test
pbe between 0～0.75, can think that this research sample size not yet can reach the usefulness of screening the two difference.Quantification by power of test to each research, can effectively control the evaluation bias causing because of researcher's subjective factor.
(3) in Meta analyzes the gross effect amount merging process of included in research, using the power of test based on including research in substitute merely based on sample size as the weight that merges gross effect amount according to and amount of analysis effect relationship.In the Meta of diagnostic test analyzes, the main statistics effect size merging is susceptibility Sen=a/ (a+c), specificity Spe=d/ (b+d) and diagnosis odds ratio Dor=(a/c)/(b/d).The combining step of its effect quantity:
If total number of the research of including in is during Meta analyzes
k(
k>=2),
iitem result of study is: susceptibility Sen
i=a
i/ (a
i+ c
i), specificity Spe
i=d
i/ (b
i+ d
i), diagnosis odds ratio Dor
i=(a
i/ c
i)/(b
i/ d
i), its natural logarithm is turned to ysen
i=In (Sen
i), yspe
i=In (Spe
i), ydor
i=In (Spe
i).Make μ
iand e
irepresent respectively
iindividual research ysen
i, yspe
iand ydor
ipopulation effect and stochastic effects, random-effect model is: ysen
i=μ sen
i+ esen
i, yspe
i=μ spe
i+ espe
i, ydor
i=μ sdor
i+ edor
i; Weighted mean ysen
w, yspe
w, ydor
wand variance S
ysen 2, S
yspe 2and S
ydor 2fitting formula be: ysen
w=
, yspe
w=
and ydor
w=
; S
2 ysen=
, S
2 yspe=
and S
2 ydor=
in formula, W
ibe
ithe weight coefficient of individual research, is calculated by the power of test of each research, and formula is
,
p i for each power of test of including research in;
For fixed-effect model,
kitem research merges effect quantity and 95% fiducial interval thereof estimated and is respectively:
Sen
f=exp (ysen
w) and 95% fiducial interval be exp (ysen
w± 1.96S
ysen);
Spe
f=exp (yspe
w) and 95% fiducial interval be exp (yspe
w± 1.96S
yspe);
Dor
f=exp (ydor
w) and 95% fiducial interval be exp (ydor
w± 1.96S
ydor).
For random-effect model, weight coefficient W
iproofread and correct as W
i *, W
i *=(W
i -1+ S
μ 2)
-1in formula, S
μ 2for μ
i(i=1,2,3 ...,
k) estimation variance, S
μ 2=
; Weighted mean y
w *=
, S
2 y *=
;
In like manner,
kitem research merges effect quantity and 95% fiducial interval thereof estimated and is respectively:
Sen
r=exp (ysen
w *) and 95% fiducial interval be exp (ysen
w *± 1.96S
ysen *);
Spe
r=exp (yspe
w *) and 95% fiducial interval be exp (yspe
w *± 1.96S
yspe *);
Dor
r=exp (ydor
w *) and 95% fiducial interval be exp (ydor
w *± 1.96S
ydor *).
Merge gross effect based on power of test, effectively evaded the risk being excluded because of the false negative of clinical test results, can provide more scientific effect quantity to estimate and increase to add up usefulness.
(4) in the sensitivity analysis process of analyzing at Meta, using different powers of test as the reference that reconsolidates effect quantity, get rid of the research of abnormal test usefulness, reappraise and effect quantity inquire into this result of study and be combined effect quantity and heterogeneous influence degree.Get rid of successively power of test
pafter 1.0,0.9 and 0.75 research, reappraise merging effect quantity, and compare with the Meta analysis result before not getting rid of, robustness and the reliability of judgement research.If after getting rid of there is not large variation in result, illustrate that susceptibility is low, result is comparatively steadily and surely credible; On the contrary, if obtain the difference conclusion of more even completely contradicting after getting rid of, illustrate that susceptibility is higher, the robustness of result is lower, explanation results and draw a conclusion in should be very prudent, there is relevant to diagnosis effect, important, potential bias factor in prompting, needs further the clearly source of dispute.
(5) after completing Meta analysis, the power of test that tradition is only applied to single clinical diagnostic test is incorporated in the evaluation of systematic review, evaluates Meta analyze quality important references standard just as aftereffect, and its power of test formula is
p meta=(1-β), wherein β=(1-NORMSDIST (Z
β)) * 2,
,
,
,
,
, but in formula, a, b, c and d become the summation of including the corresponding entry of research in Meta analysis in, the same table of its clinical meaning, and wherein I type mistake (α) value can be set as 0.01, therefore the assessment of Meta analysis quality is transferred to quantitative evaluation by qualitative evaluation.
Claims (2)
1. the method that the diagnostic test Meta based on power of test analyzes, it is characterized in that, serve as theme through analyzing in the systematic review of main body is evaluated taking diagnostic test Meta with power of test, realize the integrated integration of quality evaluation, the merging of gross effect amount, sensitivity analysis and the evaluation of Meta analysis aftereffect of including research in power of test and traditional Meta analysis in.
2. the method that the diagnostic test Meta based on power of test according to claim 1 analyzes, be characterised in that, analyze including in the quality evaluation process of research at Meta, increase taking power of test as normative reference and to quantize one by one scoring and judge the height of including quality of research in including research in; In Meta analyzes the gross effect amount merging process of included in research, using the power of test based on including research in substitute merely based on sample size as the weight that merges gross effect amount according to and statistics gross effect yardstick and inquire into dose-effect relationship; In the sensitivity analysis of analyzing at Meta, using different powers of test as the normative reference that reconsolidates effect quantity, judge the robustness of result of study by gross effect and heterogeneous intensity of variation; After traditional Meta has analyzed, evaluate the quality of Meta analysis itself taking calibrated power of test as criterion aftereffect.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410311640.4A CN104091062A (en) | 2014-07-03 | 2014-07-03 | Diagnostic test Meta analysis method based on power of test |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410311640.4A CN104091062A (en) | 2014-07-03 | 2014-07-03 | Diagnostic test Meta analysis method based on power of test |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104091062A true CN104091062A (en) | 2014-10-08 |
Family
ID=51638778
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410311640.4A Pending CN104091062A (en) | 2014-07-03 | 2014-07-03 | Diagnostic test Meta analysis method based on power of test |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104091062A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010051036A1 (en) * | 2008-11-03 | 2010-05-06 | Bruce Reiner | Automated method for medical quality assurance |
| CN101827522A (en) * | 2008-09-10 | 2010-09-08 | 阿索尔达治疗股份有限公司 | Continue to discharge the using method of aminopyridine compositions |
| CN103310105A (en) * | 2013-06-13 | 2013-09-18 | 浙江加州国际纳米技术研究院绍兴分院 | Method for screening non-small-cell lung cancer curative effect biomarker |
| CN103310129A (en) * | 2013-06-13 | 2013-09-18 | 浙江大学 | Evidence-based method for screening gastric cancer prognosis molecular markers |
-
2014
- 2014-07-03 CN CN201410311640.4A patent/CN104091062A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101827522A (en) * | 2008-09-10 | 2010-09-08 | 阿索尔达治疗股份有限公司 | Continue to discharge the using method of aminopyridine compositions |
| WO2010051036A1 (en) * | 2008-11-03 | 2010-05-06 | Bruce Reiner | Automated method for medical quality assurance |
| CN103310105A (en) * | 2013-06-13 | 2013-09-18 | 浙江加州国际纳米技术研究院绍兴分院 | Method for screening non-small-cell lung cancer curative effect biomarker |
| CN103310129A (en) * | 2013-06-13 | 2013-09-18 | 浙江大学 | Evidence-based method for screening gastric cancer prognosis molecular markers |
Non-Patent Citations (2)
| Title |
|---|
| 康德英 等: "临床试验Meta分析中的异质性评价研究", 《华西医学》 * |
| 曹明芹 等: "统计推断中检验效能的估计及应用", 《数理医药学杂志》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gambaro et al. | The complex associations between late life depression, fear of falling and risk of falls. A systematic review and meta-analysis | |
| Jácome et al. | Predicting anxiety and depression among family carers of people with chronic obstructive pulmonary disease | |
| Ewer et al. | Pulse oximetry as a screening test for congenital heart defects in newborn infants: a test accuracy study with evaluation of acceptability and cost-effectiveness | |
| Mawani et al. | Validation of self-rated mental health | |
| Ross et al. | Mood changes during pregnancy and the postpartum period: development of a biopsychosocial model | |
| Goldbecker et al. | Comparison of the most favoured methods for the diagnosis of hepatic encephalopathy in liver transplantation candidates | |
| Tan et al. | Freezing of gait and activity limitations in people with Parkinson's disease | |
| Chu et al. | The effects of a support group on dementia caregivers’ burden and depression | |
| Kelly et al. | Schizophrenia clinical symptom differences in women vs. men with and without a history of childhood physical abuse | |
| Dorros et al. | Interdependence in women with breast cancer and their partners: an interindividual model of distress. | |
| Dozier et al. | Validation of the clutter image rating in older adults with hoarding disorder | |
| Houston et al. | Type I Chiari malformation, RBANS performance, and brain morphology: Connecting the dots on cognition and macrolevel brain structure. | |
| Gangell et al. | Respiratory impedance in children with cystic fibrosis using forced oscillations in clinic | |
| Chen et al. | Greater functional connectivity between sensory networks is related to symptom severity in toddlers with autism spectrum disorder | |
| Eiden et al. | A conceptual model predicting internalizing problems in middle childhood among children of alcoholic and nonalcoholic fathers: The role of marital aggression | |
| Padden et al. | Psychophysiological measures of stress in caregivers of individuals with autism spectrum disorder: a systematic review | |
| van der Stouwe et al. | How to assess negative affective reactivity to daily life stress in depressed and nondepressed individuals? | |
| Hodge et al. | Predictors of adaptive functioning in preschool aged children with autism spectrum disorder | |
| De Donno et al. | Age assessment using the Greulich and Pyle method on a heterogeneous sample of 300 Italian healthy and pathologic subjects | |
| Kim | Factors influencing cognitive impairment of the elderly residents | |
| Sunderland et al. | Factor mixture analysis of DSM-IV symptoms of major depression in a treatment seeking clinical population | |
| Cheung et al. | Wrist accelerometry for monitoring dementia agitation behaviour in clinical settings: A scoping review | |
| Rougemont-Buecking et al. | Inter-informant agreement on diagnoses and prevalence estimates of anxiety disorders: direct interview versus family history method | |
| Yu et al. | Trajectories of health‐related quality of life during the natural history of dementia: a six‐wave longitudinal study | |
| Federici et al. | WHODAS II: Disability self-evaluation in the ICF conceptual frame |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20141008 |