seqtte

Stata package for sequential target trial emulation.

Installation

net install seqtte, from("https://raw.githubusercontent.com/remlapmot/seqtte/main/") replace

If installation fails with a permission error

On managed machines Stata’s default installation directory (PLUS) is sometimes not writable, and net install fails with an error such as cannot write in directory ... (r(603)). In that case redirect the installation to a directory you can write to using net set ado.

Option 1 — install into your current working directory. In Stata, cd to a folder you know you can write to (for example your Documents folder or your project folder), then run:

net set ado "`c(pwd)'"
net install seqtte, from("https://raw.githubusercontent.com/remlapmot/seqtte/main/") replace

Stata always searches the current working directory for programs, so seqtte will work whenever Stata’s working directory is that folder. To use it from other folders as well, run adopath + "<that folder>" (per session, or add that line to your profile.do).

Option 2 — install into your PERSONAL ado directory. Run sysdir to see where PERSONAL points on your machine and create that folder if it does not exist (including any missing parent folders — note Stata’s mkdir only creates one level at a time, so it may be easier in File Explorer/Finder). Then run:

// Find PERSONAL location with either
sysdir
// or
adopath
// or
di c(sysdir_personal)
// create the directory if it does not exist

net set ado PERSONAL
net install seqtte, from("https://raw.githubusercontent.com/remlapmot/seqtte/main/") replace

PERSONAL is on Stata’s search path in every session, so the package will then be found automatically.

Note that net set ado only lasts for the current Stata session; subsequent net install/adoupdate runs in a new session will need it set again.

After installation has succeeded launch the helpfile with

help seqtte

To check for an update run

adoupdate

To update the package run

adoupdate seqtte, update

To uninstall the package run

ado uninstall seqtte

Commands

Command	Description
`seqtte`	Intent-to-treat (ITT) and per-protocol (PP) estimators via sequential trial emulation

Usage

seqtte outcomevar, id(varname) time(varname) treatment(varname) [covariates(varlist)] [estimator(itt|pp)] [wdenominator(varlist)] [wnumerator(varlist)] [truncation(#)]

The input data should be in long (person-period) format with: - one row per individual per time period - outcomevar equal to 1 only in the period the event first occurs - consecutive integer values for time - binary treatment (0/1)

For developers

Building the website

The Quarto source for the package website is in site/ and renders into docs/, which is served by GitHub Pages. From the top level of the repo run

just render

See site/README.md for the R and Stata dependencies.

Running certification scripts

From the cscripts/ directory in Stata:

do master

Or from top level of repo

just test

Authors

Tom Palmer, University of Bristol, Bristol, UK.
Paul Madley-Dowd, University of Bristol, Bristol, UK.
Michalis Katsoulis, UCL, London, UK.

References

Hernán MA, Robins JM (2016). Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available. American Journal of Epidemiology 183(8): 758–764. https://doi.org/10.1093/aje/kwv254