{smcl} {* *! version 1.1 27may2015}{...} {vieweralsosee "" "--"}{...} {vieweralsosee "Main network help page" "network"}{...} {viewerjumpto "Syntax" "network_rank##syntax"}{...} {viewerjumpto "Description" "network_rank##description"}{...} {viewerjumpto "Examples" "network_rank##examples"}{...} {title:Title} {phang} {bf:network rank} {hline 2} Rank treatments after network meta-analysis {marker syntax}{...} {title:Syntax} {p 8 17 2} {cmdab:network rank} {cmd:min|max} {ifin} [, {it:mvmeta_pbest_options}] {pstd}{cmd:network rank} makes sensible choices for {cmd:if} and {cmd:in} which it would be unwise to change. {pstd}{it:mvmeta_pbest_options} are any {help mvmeta##pbest:suboptions for the pbest option of mvmeta}. {cmd:network rank} makes sensible choices for {cmd:zero} and {cmd:id()} which it would be unwise to change. The options listed below are likely to be useful. {synoptset 20 tabbed}{...} {synopthdr} {synoptline} {syntab:mvmeta_pbest_options} {synopt:{opt all}}Reports probabilities for all ranks. The default is to report only the probabilities of being the best treatment.{p_end} {synopt:{opt rep:s(#)}}Set the number of replicates - larger numbers reduce Monte Carlo error{p_end} {synopt:{opt seed(#)}}Set the random number seed for reproducibility{p_end} {synopt:{opt bar}}Draw a bar graph of ranks{p_end} {synopt:{opt line}}Draw a line graph of ranks{p_end} {synopt:{opt cum:ulative}}Make the bar or line graph show cumulative ranks{p_end} {synopt:{opt pred:ict}}Ranks the true effects in a future study with the same covariates, thus allowing for heterogeneity as well as parameter uncertainty, as in the calculation of prediction intervals {help mvmeta##Higgins++09:(Higgins et al, 2009)}. The default behaviour is instead to rank linear predictors and does not allow for heterogeneity.{p_end} {synopt:{opt mean:rank}}Tabulate the mean rank and the SUCRA {help mvmeta##Salanti++11:(Salanti et al, 2011)}. The SUCRA is the rescaled mean rank: it is 1 when a treatment is certain to be the best and 0 when a treatment is certain to be the worst.{p_end} {synopt:{opt saving(filename)}}Writes the draws from the posterior distribution (indexed by the identifier and the replication number) to {it:filename}.{p_end} {synopt:{opt replace}}Allows {it:filename} in {cmd:saving(}{it:filename}{cmd:)} to be overwritten.{p_end} {synopt:{opt clear}}Loads the rank data into memory and gives the commands needed to reproduce the table and graph.{p_end} {syntab:Other options} {synopt:{opt trtc:odes}}makes the display use the treatment codes rather than the treatment names.{p_end} {synoptline} {p2colreset}{...} {marker description}{...} {title:Description} {pstd} {cmd:network rank} ranks treatments after a network meta-analysis has been fitted. This currently only works after running {cmd:network meta consistency} or {cmd:network meta inconsistency} in the {cmd:augmented} format. {pstd} Use {cmdab:network rank min} if the best treatment is that with the lowest (most negative) treatment effect and {cmdab:network rank max} if the best treatment is that with the highest (most positive) treatment effect. {pstd} After fitting the inconsistency model, and after fitting a meta-regression, the treatment effect - and hence the ranking - differs across studies. {cmd:network rank} therefore computes and displays ranks for all studies unless {cmd:if} or {cmd:in} is specified. After fitting the consistency model without covariates, on the other hand, the ranks are the same for all studies and {cmd:network rank} displays ranks for just the first study. {marker examples}{...} {title:Examples} {pstd}Assume the thrombolytics data have been loaded and the consistency model has been fitted. {pstd}Find the probabilities that each treatment is the best (i.e. lowest log hazard ratio) under the consistency model: {pin}. {stata network rank min} {pstd}The same, setting the seed (for reproducibility); considering all ranks, not just the best; drawing a rankogram; and increasing the number of replicates (to reduce Monte Carlo error): {pin}. {stata network rank min, seed(48106) all line cumul reps(10000) meanrank} {p}{helpb network: Return to main help page for network}