{smcl}
{* 03Sep2008}{...}
{hline}
help for {hi:metacum}
{hline}
{title:Cumulative meta-analysis, with graphics}
{p 8 12 2}
{cmd:metacum}
{it:varlist}
{ifin}
[{cmd:,}
[{it:{help metacum##binary:binary_data_options}} {c |}
{it:{help metacum##continuous:continuous_data_options}} {c |}
{it:{help metacum##precalculated:precalculated_effect_estimates_options}}]
{it:{help metacum##measure:measure_and_model_option}}
{it:{help metacum##output:output_options}}
{it:{help metacum##forest:forest_plot_options}}]
{phang2}{marker binary}
{it:binary_data_options}
{p 12 12 2}
{cmd:or}
{cmd:rr}
{cmd:rd}
{cmd:fixed}
{cmd:random}
{cmd:fixedi}
{cmd:randomi}
{cmd:peto}
{cmdab:noint:eger}
{cmd:cc(}{it:#}{cmd:)}
{phang2}{marker continuous}
{it:continuous_data_options}
{p 12 12 2}
{cmd:cohen}
{cmd:hedges}
{cmd:glass}
{cmd:nostandard}
{cmd:fixed}
{cmd:random}
{cmdab:noint:eger}
{phang2}{marker precalculated}
{it:precalculated_effect_estimates_options}
{p 12 12 2}
{cmd:fixed}
{cmd:random}
{phang2}{marker measure}
{it:measure_and_model_option}
{p 12 12 2}
{cmd:wgt(}{it:wgtvar}{cmd:)}
{phang2}{marker output}
{it:output_options}
{p 12 12 2}
{cmd:by(}{it:byvar}{cmd:)}
{cmd:log}
{cmd:eform}
{cmdab:il:evel(}{it:#}{cmd:)}
{cmd:sortby(}{it:varlist}{cmd:)}
{bind:{cmd:label(}[{cmd:namevar=}{it:namevar}]{cmd:,} [{cmd:yearvar=}{it:yearvar}]{cmd:)}}
{cmd:notable}
{cmd:nograph}
{phang2}{marker forest}
{it:forest_plot_options}
{p 12 12 2}
{cmdab:xla:bel(}{it:#}{cmd:,}...{cmd:)}
{cmdab:xt:ick(}{it:#}{cmd:,}...{cmd:)}
{opt texts:ize(#)}
{cmd:nowt}
{cmd:nostats}
{cmd:counts}
{cmd:group1(}{it:string}{cmd:)}
{cmd:group2(}{it:string}{cmd:)}
{cmd:effect(}{it:string}{cmd:)}
{cmd:force}
{opt lcols(varlist)}
{opt rcols(varlist)}
{opt astext(#)}
{opt double}
{cmd:summaryonly}
{cmd:null(}{it:#}{cmd:)}
{cmd:nulloff}
{cmd:favours(}{it:string} # {it:string}{cmd:)}
{cmd:pointopt(}{it:{help marker_options}} {c |} {it:{help marker_label_options}}{cmd:)}
{cmd:ciopt(}{it:{help line_options}}{cmd:)}
{cmd:olineopt(}{it:{help line_options}}{cmd:)}
{cmd:classic}
{cmd:nowarning}
{it:graph_options}
{title:Description}
{p 4 4 2}
{cmd:metacum} provides cumulative pooled estimates and confidence limits
obtained from fixed or random effects
meta-analysis and plots the cumulative pooled estimates
in the style of Lau et al. (1992).
{p 4 4 2}
This updated version requires that {cmd:metan} is installed, for which
{cmd:metacum} now acts as a wrapper. As such the syntax is very similar,
allowing the user to supply data in a variety of formats. Version 9 graphics
are displayed and most of the options for {cmd:metan}, and many general
graphics options, are permitted. This help file is very similar to that of
{cmd:metan}, although with the omission of some options.
{p 4 4 2}
{cmd:metacum} requires either two, three, four or six variables to be declared.
When four variables are specified these correspond to the number of events and
non-events in the experimental group followed by those of the control group,
and analysis of binary data is performed on the 2x2 table. With six variables,
the data are assumed continuous and to be the sample size, mean and standard
deviation of the experimental group followed by those of the control group. If
three variables are specified these are assumed to be the effect estimate and
its lower and upper confidence interval, and it is suggested that these are log
transformed for odds ratios or risk ratios and the {cmd: eform} option used.
If two variables are specified these are assumed to be the effect estimate and
standard error; again, it is recommended that odds ratios or risk ratios are
log transformed.
{title:Options}
{title:{it:binary_data_options}}
{p 4 8 2}
{cmd:or} pools ORs.
{p 4 8 2}
{cmd:rr} pools RRs; this is the default.
{p 4 8 2}
{cmd:rd} pools risk differences.
{p 4 8 2}
{cmd:fixed} specifies a fixed-effect model using the Mantel-Haenszel
method; this is the default.
{p 4 8 2}
{cmd:random} specifies a random-effects model using the DerSimonian and Laird
method, with the estimate of heterogeneity being taken.
{p 4 8 2}
{cmd:fixedi} specifies a fixed-effect model using the inverse-variance method.
{p 4 8 2}
{cmd:randomi} specifies a random-effects model using the
DerSimonian and Laird method, with the estimate of heterogeneity being
taken from the inverse-variance fixed-effect model.
{p 4 8 2}
{cmd:peto} specifies that the Peto method is used to pool ORs.
{p 4 8 2}{marker binary_nointeger}
{cmd:nointeger} allows the cell counts to be nonintegers. This option may be
useful when a variable continuity correction is sought for studies containing
zero cells but also may be used in other circumstances, such as where a
cluster-randomized trial is to be incorporated and the "effective sample size"
is less than the total number of observations.
{p 4 8 2}
{cmd:cc(}{it:#}{cmd:)} defines a fixed-continuity correction to add
where a study contains a zero cell. By default, {cmd:metan8} adds 0.5 to
each cell of a trial where a zero is encountered when using
inverse-variance, DerSimonian and Laird, or Mantel-Haenszel weighting to
enable finite variance estimators to be derived. However, the {cmd:cc()}
option allows the use of other constants (including none). See also the
{cmd:nointeger} option.
{title:{it:continuous_data_options}}
{p 4 8 2}
{cmd:cohen} pools standardized mean differences by the Cohen method;
this is the default.
{p 4 8 2}
{cmd:hedges} pools standardized mean differences by the Hedges method.
{p 4 8 2}
{cmd:glass} pools standardized mean differences by the Glass method.
{p 4 8 2}
{cmd:nostandard} pools unstandardized mean differences.
{p 4 8 2}
{cmd:fixed} specifies a fixed-effect model using the Mantel-Haenszel
method; this is the default.
{p 4 8 2}
{cmd:random} specifies a random-effects model using the DerSimonian and Laird
method, with the estimate of heterogeneity being taken.
{p 4 8 2}
{cmd:nointeger} denotes that the number of observations in each arm does not
need to be an integer. By default, the first and fourth variables specified
(containing N_intervention and N_control, respectively) may occasionally be
noninteger (see {helpb metacum##binary_nointeger:nointeger} under binary
data).
{title:{it:precalculated_effect_estimates_options}}
{p 4 8 2}
{cmd:fixed} specifies a fixed-effect model using the
Mantel-Haenszel method; this is the default.
{p 4 8 2}
{cmd:random} specifies a random-effects model using the DerSimonian and Laird
method, with the estimate of heterogeneity being taken.
{title:{it:measure_and_model_option}}
{p 4 8 2}
{cmd:wgt(}{it:wgtvar}{cmd:)} specifies alternative weighting
for any data type. The effect size is to be computed by assigning
a weight of {it:wgtvar} to the studies. When RRs or ORs are
declared, their logarithms are weighted. This option should be used only if
you are satisfied that the weights are meaningful.
{title:{it:output_options}}
{p 4 8 2}
{opt by(byvar)} specifies that the meta-analysis is to be stratified
according to the variable declared.
{p 4 8 2}
{cmd:log} reports the results on the log scale
(valid only for ORs and RRs analyses from raw data counts).
{p 4 8 2}
{cmd:eform} exponentiates all effect sizes and confidence intervals
(valid only when the input variables are log-ORs or log-hazard
ratios with standard error or confidence intervals).
{p 4 8 2}
{cmd:ilevel(}{it:#}{cmd:)} specifies the coverage (e.g., 90%, 95%, 99%) for the
individual trial confidence intervals; the default is {cmd:$S_level}.
See {helpb set level}.
{p 4 8 2}
{cmd:sortby(}{it:varlist}{cmd:)} sorts by variable(s) in {it:varlist}.
{p 4 8 2}
{cmd:label(}[{cmd:namevar=}{it:namevar}]{cmd:,} [{cmd:yearvar=}{it:yearvar}]{cmd:)}
labels the data by its name, year, or both. Either or both
variable lists may be left blank. For the table display, the overall
length of the label is restricted to 20 characters. If the {cmd:lcols()}
option is also specified, it will override the {cmd:label()} option.
{p 4 8 2}
{cmd:notable} prevents the display of a results table.
{p 4 8 2}
{cmd:nograph} prevents the display of a graph.
{title:{it:forest_plot_options}}
{p 4 8 2}
{cmd:xlabel(}{it:#}{cmd:,}...{cmd:)} defines x-axis labels. This option has been
modified so that any number of points may be defined. Also, checks are no
longer made as to whether these points are sensible, so the
user may define anything if the {cmd:force} option is used. Points must
be comma separated.
{p 4 8 2}
{cmd:xtick(}{it:#}{cmd:,}...{cmd:)} adds tick marks to the x-axis. Points must
be comma separated.
{p 4 8 2}
{opt textsize(#)} specifies the font size for the text display on the graph.
This option has been modified so that the default is
{cmd:textsize(100)} (as in 100%) and the percentage may be increased or
decreased (e.g., 80 or 120 for 20% smaller or larger, respectively).
{p 4 8 2}
{cmd:nowt} prevents the display of study weight on the graph.
{p 4 8 2}
{cmd:nostats} prevents the display of study statistics on the graph.
{p 4 8 2}
{cmd:counts} displays data counts (n/N) for each group when using
binary data or the sample size, mean, and standard deviation for each
group if mean differences are used (the latter is a new feature).
{p 4 8 2}
{cmd:group1(}{it:string}{cmd:)} and {cmd:group2(}{it:string}{cmd:)} may be
used with the {cmd:counts} option, and the text should contain the
names of the two groups.
{p 4 8 2}
{opt effect(string)} allows the graph to name the summary statistic used
when the effect size and its standard error are declared.
{p 4 8 2}
{cmd:force} forces the x-axis scale to be in the range specified
by {cmd:xlabel()}.
{p 4 8 2}
{cmd:lcols(}{it:varlist}{cmd:)} and {cmd:rcols(}{it:varlist}{cmd:)}
define columns of additional data to the left or right of the graph. The
first two columns on the right are automatically set to effect size and
weight, unless suppressed by using the options {cmd:nostats} and
{cmd:nowt}. If {cmd:counts} is used, this will be set as the third column.
{cmd:textsize()} can be used to fine-tune the size of the text to achieve a
satisfactory appearance. The columns are labeled with the variable label or
the variable name if this is not defined. The first variable specified in
{cmd:lcols()} is assumed to be the study identifier and this is used in the
table output.
{p 4 8 2}
{cmd:astext(}{it:#}{cmd:)}
specifies the percentage of the graph to be taken up by text.
The default is 50%, and the percentage must be in the range 10-90.
{p 4 8 2}
{cmd:double}
allows variables specified in {cmd:lcols()} and {cmd:rcols()} to run over
two lines in the plot. This option may be of use if long strings are used.
{p 4 8 2}
{cmd:summaryonly}
shows only summary estimates in the graph. This option may be of use for
multiple subgroup analyses.
{p 4 8 2}
{cmd:null(}{it:#}{cmd:)}
displays the null line at a user-defined value rather than at 0 or 1.
{p 4 8 2}
{cmd:nulloff}
removes the null hypothesis line from the graph.
{p 4 8 2}
{cmd:favours(}{it:string} {cmd:#} {it:string}{cmd:)}
applies a label saying something about the treatment effect to either
side of the graph (strings are separated by the {cmd:#} symbol). This
option replaces the feature available in {cmd:b1title} in the previous
version of {cmd:metan}.
{p 4 8 2}
{opt pointopt(marker_options)}, {opt ciopt(line_options)}, and
{opt olineopt(line_options)}
specify options for the graph routines within the program, allowing the
user to alter the appearance of the graph. Any options associated with a
particular graph command may be used, except some that would cause incorrect
graph appearance. For example, diamonds are plotted using the
{helpb twoway pcspike} command, so options for line styles are available
(see {it:{help line options}}); however, altering the x-y orientation with
the option {cmd:horizontal} or {cmd:vertical} is not allowed. So,
{cmd:ciopt(lcolor(green) lwidth(thick))} feeds into a command such as
{cmd:pcspike(y1 x1 y2 x2, lcolor(green) lwidth(thick))}
{p 8 8 2}
{opt pointopt(marker_options)} controls the point estimate by using marker
options. See {it:{help marker_options}} and
{it:{help marker_label_options}}.
{p 8 8 2}
{opt ciopt(line_options)} controls the confidence intervals for studies by using
options for {cmd:twoway pcspike} (not horizontal/vertical). See
{it:{help line_options}}.
{p 8 8 2}
{opt olineopt(line_options)} controls the overall effect line with options for
another line (not position). See {it:{help line_options}}.
{p 4 8 2}
{cmd:classic} specifies that solid black boxes without point estimate markers
are used, as in the previous version of {cmd:metan}.
{p 4 8 2}
{cmd:nowarning} switches off the default display of a note warning that studies
are weighted from random-effects analyses.
{p 4 8 2}
{it:graph_options}
are any of the options documented in {manhelpi twoway_options G}.
These allow the addition of titles, subtitles,
captions, etc.; control of margins, plot regions, graph size, aspect
ratio; and the use of schemes. Because titles may be added with
{it:graph_options}, previous options such as {cmd:b2title} are no longer
necessary.
{title:Remarks on metacum (calling metan)}
{p 4 4 2}
For two or three variables, a variance-weighted analysis is performed in
a similar fashion to the {helpb meta} command; the two variable syntax
is {it:theta} and {it:SE(theta)}. The 3 variable syntax is {it:theta},
{it:lower ci (theta)}, {it:upper ci (theta)}. Note that in this situation
"{it:theta}" is taken to be the logarithm of the effect size if the odds
ratio or risk ratio is used. {hi:This differs from the equivalent in the {cmd:meta} command}.
This program does not assume the three variables need log transformation:
if odds ratios or risk ratios are combined, it is up to the user to
log-transform them first. The {cmd:eform} option may be used to change
back to the original scale if needed. By default the confidence
intervals are assumed symmetric, and the studies are pooled by taking
the variance to be equal to (CI width)/2z.
{p 4 4 2}
Note that for graphs on the log scale (that is, ORs or RRs), values
outside the range [10e-8,10e8] are not displayed, and similarly graphs of
other measures (log ORs, RDs, SMDs) are restricted to the range [-10e8,10e8].
A confidence interval which extends beyond this, or the specified scale
if {cmd:force} is used, will have an arrow added at the end of the range.
{title:Examples}
{p 4 8 2}
All examples use a simulated example dataset (Ross Harris 2006)
{p 8 12 2}
{stata "use http://fmwww.bc.edu/repec/bocode/m/metan_example_data":. use http://fmwww.bc.edu/repec/bocode/m/metan_example_data}
{p 4 8 2}
Risk difference from raw cell counts, random effects model, "label" specification
{p 8 12 2}
{cmd:. metacum tdeath tnodeath cdeath cnodeath, }
{p_end}
{p 12 12 2}
{cmd:rd random label(namevar=id, yearid=year)}
{p_end}
{p 12 12 2}
{it:({stata "metacum_examples metacum_ex1":click to run})}
{p 4 8 2}
Generate log odds ratio and standard error. Graph has exponential form,
scale is forced within set limits and ticks added. Data columns syntax used and effect label specified.
{p 8 12 2}
{cmd:. gen logor = ln( (tdeath*cnodeath)/(tnodeath*cdeath) )}
{p 8 12 2}
{cmd:. gen selogor = sqrt( (1/tdeath) + (1/tnodeath) + (1/cdeath) + (1/cnodeath) )}
{p 8 12 2}
{cmd:. metacum logor selogor, eform xlabel(0.6, 0.8, 1, 1.2, 1.4, 1.6) }
{p_end}
{p 12 12 2}
{cmd:force xtick(0.7, 0.9, 1.1, 1.3, 1.5) lcols(id year country) effect(Odds ratio)}
{p_end}
{p 12 12 2}
{it:({stata "metacum_examples metacum_ex2":click to run})}
{title:Reference}
{phang} Lau, J., E. M. Antman, J. Jimenez-Silva, F. Mosteller, and T. C.
Chalmers. 1992. Cumulative meta-analysis of therapeutic trials for myocardial
infarction. {it:New England Journal of Medicine} 327: 248-254.
{title:Authors}
{title:First version}
{phang}
Jonathan A. C. Sterne{break}
Department of Social Medicine, University of Bristol,
Canynge Hall, Whiteladies Road, Bristol BS8 2PR, UK
{title:Version 9 update}
{phang}
Ross J. Harris{break}
Department of Social Medicine, University of Bristol,
Canynge Hall, Whiteladies Road, Bristol BS8 2PR, UK
{title:Also see}
{p 7 11 2}STB: STB-44 sbe24
{psee}Online: {helpb metan}, {helpb metannt} (if installed),
{helpb meta} (if installed), {helpb metareg} (if installed),
{helpb metabias} (if installed), {helpb metatrim} (if installed),
{helpb metainf} (if installed), {helpb galbr} (if installed),
{helpb metafunnel} (if installed)
{p_end}