summary.mvmeta: Summarizing mvmeta Models

In mvmeta: Multivariate and Univariate Meta-Analysis and Meta-Regression

Description

Print and summary method functions for fitted univariate or multivariate meta-analytical models represented in objects of class "mvmeta".

Usage

## S3 method for class 'mvmeta'
summary(object, ci.level=0.95, ...)

## S3 method for class 'summary.mvmeta'
print(x, digits=4, ...)

## S3 method for class 'mvmeta'
print(x, digits=4, ...)

Arguments

object	an object of class "mvmeta" produced by a call to mvmeta.
x	an object of class "mvmeta" or "summary.mvmeta", produced by calls to mvmeta or summary.mvmeta, respectively.
ci.level	a numerical value between 0 and 1, specifying the confidence level for the computation of confidence intervals.
digits	an integer specifying the number of digits to which printed results must be rounded.
...	further arguments passed to or from other methods.

Details

The print method function for class "mvmeta" only returns basic information on the fitted model, namely the call, estimated fixed-effects coefficients, dimensions and fit statistics (log-likelihood, AIC, BIC).

The summary method function computes additional statistics and tests, and produces a list object of class "summary.mvmeta". The print method function for this class shows additional information, such as tables reporting the estimates for the fixed and random-effects parts of the model, Cochran Q test for heterogeneity and I^2.

Value

The summary method function for mvmeta objects produces a list of class "summary.mvmeta". The components of the lists are some of those stored in the related mvmeta object, plus the following:

coefficients	a matrix reporting point estimates, standard errors, z statistics and related p-values of the test, and confidence intervals for the kp fixed-effects coefficients. Note this is different than the component with the same name stored in mvmeta objects, simply reporting the point estimates (see mvmetaObject).
AIC	the value of the Akaike information criterion for the fitted mvmeta model, obtained through a call to AIC.
BIC	the value of the Bayesian information criterion for the fitted mvmeta model, obtained through a call to BIC.
corFixed	the kp x kp correlation matrix of the fixed-effects coefficients, obtained from the (co)variance matrix vcov (see mvmetaObject and vcov).
corRandom	the k x k correlation matrix of the random effects, obtained from the between-study (co)variance matrix Psi (see see mvmetaObject).
qstat	results from the Cochran Q test for heterogeneity, namely a list corresponding to a qtest.mvmeta object without its class, obtained through qtest.
ci.level	the confidence level used for defining the confidence intervals for the estimates of the fixed-effects coefficients.

As usual, the print method functions for classes "mvmeta" and "summary.mvmeta" do not return any value.

Author(s)

Antonio Gasparrini, [e-mail]

References

Sera F, Armstrong B, Blangiardo M, Gasparrini A (2019). An extended mixed-effects framework for meta-analysis.Statistics in Medicine. 2019;38(29):5429-5444. [Freely available here].

Gasparrini A, Armstrong B, Kenward MG (2012). Multivariate meta-analysis for non-linear and other multi-parameter associations. Statistics in Medicine. 31(29):3821–3839. [Freely available here].

See Also

See mvmeta and mvmetaObject.

Examples

# RUN THE MODEL 
model <- mvmeta(cbind(PD,AL)~pubyear,S=berkey98[5:7],data=berkey98)

# SIMPLE PRINT
model
# DEFINE DIGITS
print(model,digit=2)
# SUMMARY WITH 80TH CONFIDENCE INTERVALS
summary(model,ci.level=0.80)

Example output

This is mvmeta 0.4.7. For an overview type: help('mvmeta-package').
Call:  mvmeta(formula = cbind(PD, AL) ~ pubyear, S = berkey98[5:7], 
    data = berkey98)

Fixed-effects coefficients:
                  PD       AL
(Intercept)  -9.2816  22.5414
pubyear       0.0049  -0.0115

5 studies, 10 observations, 4 fixed and 3 random-effects parameters
 logLik      AIC      BIC  
-3.5400  21.0799  19.6222  

Call:  mvmeta(formula = cbind(PD, AL) ~ pubyear, S = berkey98[5:7], 
    data = berkey98)

Fixed-effects coefficients:
                PD     AL
(Intercept)  -9.28  22.54
pubyear       0.00  -0.01

5 studies, 10 observations, 4 fixed and 3 random-effects parameters
logLik     AIC     BIC  
 -3.54   21.08   19.62  

Call:  mvmeta(formula = cbind(PD, AL) ~ pubyear, S = berkey98[5:7], 
    data = berkey98)

Multivariate random-effects meta-regression
Dimension: 2
Estimation method: REML

Fixed-effects coefficients
PD : 
             Estimate  Std. Error        z  Pr(>|z|)  80%ci.lb  80%ci.ub   
(Intercept)   -9.2816     43.3420  -0.2141    0.8304  -64.8266   46.2634   
pubyear        0.0049      0.0219   0.2225    0.8239   -0.0231    0.0329   
AL : 
             Estimate  Std. Error        z  Pr(>|z|)  80%ci.lb  80%ci.ub   
(Intercept)   22.5414     59.4309   0.3793    0.7045  -53.6223   98.7052   
pubyear       -0.0115      0.0300  -0.3850    0.7002   -0.0499    0.0269   
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Between-study random-effects (co)variance components
	Structure: General positive-definite
    Std. Dev    Corr
PD    0.1430      PD
AL    0.2021  0.5614

Multivariate Cochran Q-test for residual heterogeneity:
Q = 125.7557 (df = 6), p-value = 0.0000
I-square statistic = 95.2%

5 studies, 10 observations, 4 fixed and 3 random-effects parameters
 logLik      AIC      BIC  
-3.5400  21.0799  19.6222  

mvmeta documentation built on Dec. 10, 2019, 5:07 p.m.