View {title}

Search
!app sctoplatforms theme default menu

sts_graph_landmark: Landmark analysis graphs for avoiding survival bias

Survival bias, or immortal time bias, arises from the incorrect analysis of time-dependent events. Incorrect modelling ignores the time dependence of the failure event, thus creating artificial scenarios in which statistical units can be immune to failure for a period of time. Landmark analysis is an approach that adequately circumvents this issue by resetting the analysis time to one or more specified time points, after which only surviving statistical units are included. As a result, only statistical units with comparable survivorship status at the landmark time are compared within each landmark epoch. sts_graph_landmark is available as a Stata package. 

Landmark analysis for avoiding survival bias

What it is:

sts_graph_landmark is a Stata package for performing landmark analysis and producing landmark graphs in order to avoid survival bias.

What it can do:

sts_graph_landmark can draw Kaplan-Meier curves associated with landmark analysis and display corresponding population at risk tables underneath the curves. The programme automates the boilerplate code as much as possible to produce these graphs and tables using sensible default parameter values while providing options to fully customise all aspects of each graph and table.

Who should use it:

This tool is for statisticians and principal investigators who perform survival analysis in their clinical research project.

Who is behind this resource:

The code was developed by statisticians at the CTU Bern on behalf of the SCTO’s Statistics & Methodology Platform. It is one of the projects that won the platform’s statistical programming grant in 2020.

Download

The sts_graph_landmark Stata package is available on GitHub. For detailed instructions on how to use this tool, please read an example with comments by Arnaud Künzi (2020). 

Before you download this document, we would appreciate you letting us know who you are.

Our templates and tools are free. As a publicly funded organisation, we strive to inform the public about the impact of our work and to continually improve our services. We therefore kindly ask you to leave us your email address so we can contact you with a short one-time survey. You may download the document without providing your personal data.

Cancel Close Skip confirmation and download directly

Related tools and resources