New methodological solutions need to be straightforwardly demonstrated and implemented

We are keenly aware that new methodological solutions need to be straightforwardly demonstrated and implemented, by ourselves and by others, in order to be taken up by the wider research community. Therefore, we often develop user-written code for use with commonly-used statistical software to accompany our work.

nstage

Design of multi-arm,multi-stage (MAMS) trials for time-to-event outcomes

Specify the design (sample size, duration, overall operating characteristics) of a multi-arm, multi-stage (MAMS) trial utilizing an intermediate outcome (I-outcome) at the intermediate stages and a definitive or primary outcome (D-outcome) at the final stage, nstagebin for binary responses

From within stata "ssc install nstage"

Main Publication

Tutorial

Stats Software: Stata

Author(s) / Maintainer(s): Royston, Barthel, Oskooei, Blenkinsop, Bratton, b.choodari-oskooei@ucl.ac.uk

Category: Clinical Trials METHODOLOGY

Where available: SSC (Boston College Statistical Software Components archive)

 

art (Assessment of Resources for Trials)

Complex sample size calculation in randomised trials

art is a menu- and command-driven set of programs to compute sample size or power for randomized controlled trials with a time-to-event (artsurv)or binary (artbin) outcome measure. ART accommodates complex features including non-proportional hazards, cross-over between treatments, loss to follow-up, staggered entry, flexible patient accrual patterns, and several different 'flavors' of the logrank test. Non-inferiority designs are supported. Multiple treatment groups with joint tests are allowed. Trend tests over dose levels of a covariate are supported. Projections of power and events (ARTPEP) are provided.

From within Stata "ssc install art" ,

Main Publication

Tutorial

Stats Software: Stata

Author(s) / Maintainer(s): Babiker, Bartel, Royston, Marley-Zagar e.marley-zagar@ucl.ac.uk

Category: Clinical Trials METHODOLOGY

Where available: SSC

 

network

Suite of programs for network meta-analysis

running a contrast-based network meta-analysis using mvmeta or metareg, assessing inconsistency, and graphing the data and results. The data in each arm of each study are assumed to be available either as binomial counts (successes/total) or as mean, standard deviation and number of individuals for a quantitative variable.

From within Stata "ssc install network" or "net from http://www.homepages.ucl.ac.uk/~rmjwiww/stata/network"

Main Publication

Tutorial

Stats Software: Stata

Author(s) / Maintainer(s): White, ian.white@ucl.ac.uk

Category: META-ANALYSiS

Where available: SSC and http://www.homepages.ucl.ac.uk/~rmjwiww

 

mvmeta

multivariate random-effects meta-analysis

running a contrast-based network meta-analysis using mvmeta or metareg, assessing inconsistency, and graphing the data and results. The data in each arm of each study are assumed to be available either as binomial counts (successes/total) or as mean, standard deviation and number of individuals for a quantitative variable.

From within Stata "ssc install mvmeta"

Main Publication

Stats Software: Stata

Author(s) / Maintainer(s): White, ian.white@ucl.ac.uk

Category: META-ANALYIS

Where available: SSC

 

rctmiss

Analyse a randomised controlled trial allowing for informatively missing outcome data

Multivariate meta-analysis combines estimates of several related parameters over several studies, mvmeta performs maximum likelihood, restricted maximum likelihood or method of moments estimation of random-effects multivariate meta-analysis models.

From within Stata "ssc install rctmiss"

Main Publication

Stats Software: Stata

Author(s) / Maintainer(s): White, ian.white@ucl.ac.uk

Category: Clinical Trials METHODOLOGY

Where available: SSC and https://github.com/UCL/rctmiss

 

admetan

Comprehensive meta-analyses routines including ipdmetan

rctmiss analyses a randomised control trial with missing outcome data under range of assumptions about the missing data. The data and missingness are modelled jointly using either a pattern-mixture model or a selection model. Assumptions about the missing data are expressed via a sensitivity parameter delta which measures the degree of departure from missing at random.

From within Stata "ssc install admetan"

Main Publication

Stats Software: Stata

Author(s) / Maintainer(s): Fisher, d.fisher@ucl.ac.uk

Category: META-ANALYSIS

Where available: SSC and https://github.com/UCL/admetan

 

mimix

Reference-based multiple imputation, sensitivity analysis of longitudinal trials with protocol deviation

admetan performs meta-analysis of aggregate (summary) data, in contrast to ipdmetan which performs meta-analysis of individual participant data (IPD). It has all the functionality of the popular metan package, and includes many additional features. As such, it may be seen as a direct update of metan. In particular, the syntax of admetan has been deliberately kept similar to that of metan; any differences are noted in the help file.

From within Stata "ssc install mimix"

Main Publication

Stats Software: Stata

Author(s) / Maintainer(s): Cro, s.cro@imperial.ac.uk Carpenter, Kenward

Category: MISSING VALUES

Where available: SSC

 

jomo

Multilevel joint modelling multiple imputation

mimix imputes missing numerical outcomes for a longitudinal trial with protocol deviation under distinct treatment arm-based assumptions for the unobserved data, following the general algorithm of Carpenter, Roger, and Kenward (2013).

From within R "install.packages("jomo")"

Main Publication

Stats Software: R

Author(s) / Maintainer(s): Quartagno, <m.quartagno@ucl.ac.uk> Carpenter

Category: MISSING VALUES

Where available: CRAN

 

dani

Tools to help design and analysis of non-inferiority and durations trials

Similarly to Schafer's package 'pan', 'jomo' is a package for multilevel joint modelling multiple imputation Novel aspects of 'jomo' are the possibility of handling binary and categorical data through latent normal variables, the option to use cluster-specific covariance matrices and to impute com

From within R "install.packages("dani")"

Main Publication

Stats Software: R

Author(s) / Maintainer(s): Quartagno, <m.quartagno@ucl.ac.uk>

Category: Clinical Trials METHODOLOGY

Where available: CRAN

 

bcss

create graphs to show how baseline data(prospective or retrospective) affects sample size for cluster randomised trial

Provides tools to help the design and analysis of resilient non-inferiority trials. These include functions for sample size calculations and analyses of trials, with either a risk difference, risk ratio or arc-sine difference margin, and a function to run simulations to design a trial with the methods described in Quartagno et al. (2019) .

From within stata "ssc install bcss", or " github install UCL/bcss"

Main Publication

Stats Software: Stata (Static charts) R (interactive)

Author(s) / Maintainer(s): Copas, Marley-Zagar (Stata), McGrath (R) e.marley-zagar@ucl.ac.uk

Category: Clinical Trials METHODOLOGY

Where available: SSC and https://github.com/UCL/bcss