Continuous meta
CONTINUOUS DATA
Learning outcomes:
You will have knowledge about and be able:
- Conduct fixed and random effects meta-analysis
- Measure effect size
- Explore heterogeneity using sub-group and meta-regression analyses
- Explore and address small study effects using funnel plot and trim and fill
STEPS in conducting Continuous Data
- STEP 1 - LOAD DATA
- STEP 2 - DECLARE, UPDATE & DESCRIBE meta data
- STEP 3 - SUMMARIZE meta data by using a TABLE or a FOREST PLOT
- STEP 4- EXPLORE HETEROGENEITY - SUB-GROUP and META-REGRESSION analysis
- STEP 5- EXPLORE and ADDRESS SMALL-STUDY EFFECTS
Materials and setup
Laptop users: you will need a copy of Stata installed on your machine.
- You can install a licensed version from https://warwick.ac.uk/services/its/servicessupport/software/list/stata
- Find class materials at Click to Download
- Download and extract to your desktop or any folder of your choice!
Link to YouTube Video Lecture
Click the image below:
STEP 1 - LOAD DATA
The meta command can also handle continuous outcomes (six variables).
n1 mean1 sd1 n2 mean2 sd2
Data for the continuous ‘outcome’:
- year: Year
- tsample: Number treated (exposed)
- tmean: Mean in treated (exposed)
- tsd: Standard deviation in treated (exposed)
- csample: Number not treated (unexposed)
- cmean: Mean in untreated (unexposed)
- csd: Standard deviation in untreated (unexposed)
Option 1 Copy and paste from Excel (see attached cont_data.xlsx below)
edit
Option 2
Input raw directly into Stata (using ‘do file editor’ type doedit)
Option 3 Load Stata data file directly from Excel file
import excel using data/cont_data.xlsx, firstrow clear
describe
(8 vars, 20 obs)
Contains data
obs: 20
vars: 8
--------------------------------------------------------------------------------
storage display value
variable name type format label variable label
--------------------------------------------------------------------------------
id str14 %14s id
year int %10.0g year
tsample int %10.0g tsample
tmean double %10.0g tmean
tsd double %10.0g tsd
csample int %10.0g csample
cmean double %10.0g cmean
csd double %10.0g csd
--------------------------------------------------------------------------------
Sorted by:
Note: Dataset has changed since last saved.
Option 4 Load Stata data file directly (if already saved)
use data/cont_data.dta, clear
describe
(Example dataset for meta-analyses, Ross Harris 2006)
Contains data from data/cont_data.dta
obs: 20 Example dataset for
meta-analyses, Ross Harris 2006
vars: 8 11 Oct 2015 21:15
--------------------------------------------------------------------------------
storage display value
variable name type format label variable label
--------------------------------------------------------------------------------
id str14 %14s Study identifier
year int %8.0g Year
tsample int %9.0g Number treated (exposed)
tmean float %9.0g Mean in treated (exposed)
tsd float %9.0g Standard deviation in treated
(exposed)
csample int %9.0g Number not treated (unexposed)
cmean float %9.0g Mean in untreated (unexposed)
csd float %9.0g Standard deviation in untreated
(unexposed)
--------------------------------------------------------------------------------
Sorted by: year
STEP 2- DECLARE, UPDATE & DESCRIBE meta data
| Continuous Type | Description |
|---|---|
| hedgesg | Hedges’s g standardized mean difference; the default |
| cohend | Cohen’s d standardized mean difference |
| glassdelta2 | Glass’s ∆ mean difference standardized by group 2 (control) standard deviation; more common than glassdelta1 |
| glassdelta1 | Glass’s ∆ mean difference standardized by group 1 (treatment) standard deviation |
| mdiff | (unstandardized) mean difference |
meta esize tsample tmean tsd csample cmean csd, studylabel(id) esize(hedgesg)
>
Meta-analysis setting information
Study information
No. of studies: 20
Study label: id
Study size: _meta_studysize
Summary data: tsample tmean tsd csample cmean csd
Effect size
Type: hedgesg
Label: Hedges's g
Variable: _meta_es
Bias correction: Approximate
Precision
Std. Err.: _meta_se
Std. Err. adj.: None
CI: [_meta_cil, _meta_ciu]
CI level: 95%
Model and method
Model: Random-effects
Method: REML
STEP 3- SUMMARIZE meta data by using a TABLE or a FOREST PLOT
meta forestplot, nullrefline
graph display
Effect-size label: Hedges's g
Effect size: _meta_es
Std. Err.: _meta_se
Study label: id
meta update, esize(mdiff)
-> meta esize tsample tmean tsd csample cmean csd , esize(mdiff) studylabel(id)
Meta-analysis setting information from meta esize
Study information
No. of studies: 20
Study label: id
Study size: _meta_studysize
Summary data: tsample tmean tsd csample cmean csd
Effect size
Type: mdiff
Label: Mean Diff.
Variable: _meta_es
Precision
Std. Err.: _meta_se
Std. Err. adj.: None
CI: [_meta_cil, _meta_ciu]
CI level: 95%
Model and method
Model: Random-effects
Method: REML
meta forestplot, nullrefline
graph display
Effect-size label: Mean Diff.
Effect size: _meta_es
Std. Err.: _meta_se
Study label: id
STEP 4- EXPLORE HETEROGENEITY - SUB-GROUP and META-REGRESSION analysis
meta regress year
Effect-size label: Mean Diff.
Effect size: _meta_es
Std. Err.: _meta_se
Random-effects meta-regression Number of obs = 20
Method: REML Residual heterogeneity:
tau2 = .2539
I2 (%) = 58.12
H2 = 2.39
R-squared (%) = 0.00
Wald chi2(1) = 0.00
Prob > chi2 = 0.9729
------------------------------------------------------------------------------
_meta_es | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
year | -.0008941 .0262719 -0.03 0.973 -.0523862 .0505979
_cons | 1.871691 52.49334 0.04 0.972 -101.0134 104.7568
------------------------------------------------------------------------------
Test of residual homogeneity: Q_res = chi2(18) = 50.78 Prob > Q_res = 0.0001
STEP 5- EXPLORE and ADDRESS SMALL-STUDY EFFECTS
meta funnelplot, metric(invse)
graph di
Effect-size label: Mean Diff.
Effect size: _meta_es
Std. Err.: _meta_se
Model: Common-effect
Method: Inverse-variance
meta bias, egger
Effect-size label: Mean Diff.
Effect size: _meta_es
Std. Err.: _meta_se
Regression-based Egger test for small-study effects
Random-effects model
Method: REML
H0: beta1 = 0; no small-study effects
beta1 = 0.36
SE of beta1 = 0.894
z = 0.40
Prob > |z| = 0.6898
meta trimfill
Effect-size label: Mean Diff.
Effect size: _meta_es
Std. Err.: _meta_se
Nonparametric trim-and-fill analysis of publication bias
Linear estimator, imputing on the left
Iteration Number of studies = 20
Model: Random-effects observed = 20
Method: REML imputed = 0
Pooling
Model: Random-effects
Method: REML
---------------------------------------------------------------
Studies | Mean Diff. [95% Conf. Interval]
---------------------+-----------------------------------------
Observed | 0.084 -0.214 0.382
Observed + Imputed | 0.084 -0.214 0.382
---------------------------------------------------------------
CONTINUOUS DATA
STEPS in conducting Continuous Data
- STEP 1 - LOAD DATA
- STEP 2 - DECLARE, UPDATE & DESCRIBE meta data
- STEP 3 - SUMMARIZE meta data by using a TABLE or a FOREST PLOT
- STEP 4- EXPLORE HETEROGENEITY - SUB-GROUP and META-REGRESSION analysis
- STEP 5- EXPLORE and ADDRESS SMALL-STUDY EFFECTS