Nonparametric "trim and fill" analysis of publication bias in meta-analysis ---------------------------------------------------------------------------
^metatrim^ { theta { se_theta | var_theta } | exp(theta) ll ul [cl] } [ ^if^ exp ] [ ^in^ range ] [^,^ { ^v^ar | ^ci^ } ^re^ffect ^pr^i > nt ^est^imator^(^{ ^r^un | ^l^inear | ^q^uadratic }^) ef^orm ^f^un > nel ^l^evel^(^#^) g^raph ^id^var^(^var^) sa^ve^(^filename [^, repla > ce^]^)^ ^flip^ ^mixed^ graph_options ]
where { a | b |...} means choose one and only one of {a, b, ...}.
Description -----------
^metatrim^ performs the Duval and Tweedie nonparametric "trim and fill" method of accounting for publication bias in meta-analysis. The method, a rank-based data augmentation technique, formalizes the use of funnel plots, estimates the number and outcomes of missing studies, and adjusts the meta-analysis to incorporate the theoretical missing studies. The authors claim that the method is effective and consistent with other adjustment techniques.
The user provides the effect estimate, ^theta^, to ^metatrim^ as a log risk ratio, log odds ratio, or other direct measure of effect. Along with theta, the user supplies a measure of theta's variability (i.e., its standard error, ^se_theta^, or its variance, ^var_theta^). Alternatively, the user may provide the exponentiated form, ^exp(theta)^, (i.e., a risk ratio or odds ratio) and it > s confidence interval, ^(ll, ul)^.
As an option, ^metatrim^ provides a funnel graph of the filled data. The funnel graph plots ^theta^ versus ^se_theta^. Guide lines to assist in visualizing th > e funnel are plotted at the variance-weighted (fixed effects) meta-analytic effect estimate and at pseudo confidence interval limits about that effect estimate (i.e., at ^theta +/- z * se_theta^, where ^z^ is the standard Normal variate for the confidence level specified by option ^level()^. Each filled (i.e., augmented) data point is indicated by the addition of a square placed around the data symbol.
Note: ^metatrim^ requires that program ^meta^ be installed on your system. Use ^findit meta^ to find and install ^meta^ from STB-43.
Options -------
^var^ indicates that ^var_theta^ was supplied on the command line instead of ^se_theta^. Option ^ci^ should not be specified when option ^var^ is specified > .
^ci^ indicates that ^exp(theta)^ and its confidence interval, ^(ll, ul)^, were supplied on the command line instead of ^theta^ and ^se_theta^. Option ^var^ should not be specified when option ^ci^ is specified.
^reffect^ requests that the analysis be based on random-effects meta-analytic point estimates. The default is to base the calculations on fixed-effects meta-analytic point estimates.
^mixed^ requests that the meta-analyses during iterations be different from those of the initial and final meta-analyses. That is, if ^reffect^ is specified, a random-effects meta-analysis is used for the initial point estimate and the final filled analysis but fixed-effects meta-analysis is used for point estimates during iteration. Conversely, if the default analysis is chosen (i.e. ^reffect^ not specified) then a fixed-effects meta-analysis is used for the initial point estimate and the final filled analysis but random- effects meta-analysis is used for point estimates during iteration.
^print^ requests that the weights used in the fixed and random effects estimation are listed for each study, together with the individual study estimates and confidence intervals. The studies are labelled by name if the ^idvar^ option is specified, or by number otherwise.
^eform^ requests that the results in the final meta-analysis and in the ^print^ option be reported in exponentiated form. This is useful when the data represent odds ratios or relative risks.
^estimator(^{ ^run^ | ^linear^ | ^quadratic^ }^)^ chooses the estimator to be u > sed to determine the number of points to trim in each iteration. The ^run^ estimator is non-robust to an isolated negative point and the ^quadratic^ estimator may not be defined when the number of points in the data set is small. The ^linear^ estimator is stable in most situations and is the default.
^funnel^ requests the filled funnel graph showing the data, the fixed-effects (variance-weighted) meta-analytic effect, and the pseudo confidence interval limits about the meta-analytic effect.
^level(^#^)^ sets the confidence level % for the pseudo confidence intervals; the default is 95%.
^idvar(^varname^)^ indicates the character variable used to label the studies.
^graph^ requests that the point estimates and confidence intervals be plotted. The combined estimate in the graph is derived using fixed or random effects meta-analysis, as specified by option ^reffect^.
^save(^filename [^, replace^]^)^ saves the filled data in a separate Stata data file. The filename is assumed to have extension ^.dta^ (an extension should not be provided by the user). If filename does not exist, it is created. If filename exists, an error will occur unless ^replace^ is also specified. Only three variables are saved: a study id variable and two variables containing the filled theta and setheta values. The study id variable, named ^id^ in the saved file, is created by ^metatrim^ but, when option ^idvar()^ is specified, is based on that id variable. The filled theta and setheta variables are named ^filled^ and ^sefill^ in the saved file.
^flip^ is a switch that changes the direction in which studies are assumed to b > e suppressed by publication bias. By default, ^metatrim^ assumes that studies wit > h negative or small effect sizes are suppressed. Specifying ^flip^ causes ^metatr > im^ to assume that studies with positive or large effect sizes are suppressed. Use of this option would be appropriate, for example, where a beneficial treatment reduces the odds or risk of death. Studies not showing the desired effect (i.e., a decrease) may be more likely to be suppressed by publication bias.
Graph options are those allowed with ^graph, twoway^, except ^ylabel()^, ^symbo > l()^, ^xlog^, ^ytick^ and ^gap^ are not recognized by ^graph^. For ^funnel^, the def > ault graph_options include ^connect(lll..)^, ^symbol(iiioS)^, and ^pen(35522)^ for displaying the meta-analytic effect, the pseudo confidence interval limits (two lines), the data points, and the augmented data, respectively.
Required input variables ------------------------
^theta^ the effect estimate ^se_theta^ the corresponding standard error
or
^theta^ the effect estimate ^var_theta^ the corresponding variance
or
^exp(theta)^ the risk (or odds) ratio ^ll^ the lower limit of the risk ratio's confidence interval ^ul^ the upper limit of the risk ratio's confidence interval [^cl^] optional (see below)
Optional input variable -----------------------
^cl^ contains the confidence level of the confidence interval defined by ^ll^ and ^ul^. If ^cl^ is not provided, the procedure assumes that each confidence interval is at the 95% confidence level. ^cl^ allows the user to provide the confidence level, by study, when the confidence interval is not at the default level. ^cl^ can be specified with or without a decimal point. For example, 90 and .90 are equivalent and may be mixed (i.e., 90, .95, 80, .90 etc.).
Note ----
If your data are in raw count format, program ^metan^ can be used to facilitate conversion to effect format. ^metan^ automatically adds ^exp(theta)^ and ^se_t > heta^ variables to the dataset, calling them ^_ES^ and ^_seES^. You must manually generate ^theta^ as the natural log of ^_ES^ (for example, ^gen _lnES = ln(_ES) > ^) then input the effect-format variables, ^_lnES^ and ^_seES^, using ^metatrim^'s default input method.
Examples --------
. ^metatrim logrr selogrr, funnel print^ . ^metatrim logrr varlogrr if site==3, var graph^ . ^metatrim rr ll ul, ci eform reffect^ . ^metatrim logor selogor if region==4, funnel level(90)^
References ----------
Duval, S., Tweedie, R., 2000. A nonparametric "trim and fill" method of accounting for publication bias in meta-analysis. JASA 95(449): 89-98.
Author ------
Thomas J. Steichen, steichen@triad.rr.com
Also see --------
STB: STB-57 sbe39 (STB-58 sbe-39.1; STB-61 sbe-39.2) Manual: On-line: help for @meta@, @metabias@, and @metan@ (if installed)