st: New commands -stjm- and -stjmgraph- available on SSC

["Crowther, Michael J." <mjc76@leicester.ac.uk>]

Many thanks to Kit Baum, 2 new packages -stjm- and -stjmgraph- are available from SSC. Stata version 11.2 is required and please use -ssc- to install.

-stjm- fits shared parameter joint models of longitudinal and survival data. A single continuous longitudinal response and a single survival outcome are allowed. A linear mixed effects model is used for the longitudinal submodel, which allows time to be modelled using fixed and/or random fractional polynomials. Two choices are currently allowed for the survival submodel; the first being the flexible parametric survival model (-stpm2- available on SSC), modelled on the log cumulative hazard scale, and the second the Weibull proportional hazards model. The association between the two processes can be induced via the default current value parameterisation, the first derivative of the longitudinal submodel, and/or a random coefficient such as the intercept. Gauss-Hermite quadrature, coded in Mata, is used to evaluate the joint likelihood. Under a Weibull submodel, Gauss-Kronrod quadrature is used to evaluate the cumulative hazard. The dataset must be stset correctly into enter a!
 nd exit times, using the enter option. stjm uses _t0 to denote measurement times. Delayed entry models are allowed.

-stjmgraph- creates a longitudinal trajectory plot, whereby the timescale is adjusted by taking away each patient's event/censoring time. This form of graph can be useful to display joint longitudinal and survival data, giving an indication of any association between the two processes. A separate plot is created for panels who were censored and for panels who experienced the event of interest.

More details on -stjm- can be found in a recent talk I gave at the UK Users' Group Meeting:
http://repec.org/usug2011/UK11_crowther.pdf

-stjm- is very much still in development and will be updated frequently; future updates will include further choices for the survival submodel, and extended prediction options. Any feedback from users would be gratefully received!


Thanks,
Michael


Michael Crowther
Research Assistant in Medical Statistics
Centre for Biostatistics and Genetic Epidemiology
Department of Health Sciences
University of Leicester
Room 212, Adrian Building
University Road
Leicester LE1 7RH
Tel: 0116 229 7278
E-mail: mjc76@le.ac.uk

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/