I am an honorary professor of statistics in the Department of Statistical Science at University College London. My "day job" is as a senior scientist (statistician) at the MRC Clinical Trials Unit in London, where I have worked since 2000 on cancer trials and statistical methodology.
My interests centre around statistical modelling and its medical applications, modelling continuous predictors, survival analysis, methodology of clinical trial design, and imputation of missing covariate data. I have a strong interest in developing and disseminating software implementation of new research methods in the form of Stata programs.
- prognostic/diagnostic modelling
- modelling in epidemiology - dose-response relationships, building confounder models
- treatment/covariate interactions in clinical trials (tailoring treatment to the patient)
- modelling interactions in epidemiological studies (effect modification)
- modelling continuous predictors using fractional polynomials or splines
- multiple imputation of missing covariate values
- selection of variables and of functions of variables
- flexible parametric alternatives to the Cox model
- presenting survival data from trials more informatively than Kaplan-Meier curves
- adjusted survival curves
- application of flexible parametric survival models to relative survival
- modelling time-dependent effects
Clinical trials methodology
- design of multi-arm, multi-stage (MAMS) trials
- operating characteristics of MAMS trials
- accommodating different types of intermediate outcome measure
- Stata programs for MAMS designs
- complex sample size calculations for time-to-event trials (ART - Assessment of Resources for Trials)
You can obtain the latest versions of my Stata software, provided your Stata has Internet access.
In stata, type net from http://www.homepages.ucl.ac.uk/~ucakjpr/stata and follow the instructions on the screen to select, download and install the program(s) of your choice.
To check and if necessary update one of the 'packages' (programs) use Stata's adoupdate command, for example adoupdate stpm or adoupdate mfpboot, update.
Some of the programs are the work of several people. Some may be less well tested, and some may even have bugs. If you have a problem with a program, please report it to me at email@example.com. If it is a complex problem or program, it is always helpful to receive a .do file, a .log file and if needed a .dta data file so I can try to reproduce the issue.
You can view here a (reasonably) complete list of my publications from 1979 to date.
Our Wiley book on Multivariable Model-Building, written jointly with Willi Sauerbrei of Freiburg University, Germany, may be viewed/purchased at Amazon or Wiley. Further information on the book, together with datasets, Stata software, potential teaching materials and other useful stuff, may be obtained at the website of our book. Alternatively, you may download the Stata programs, datasets and Stata do-files and a readme file for the book from here. The datasets are also available in SAS, Excel and ASCII format.
In 1996, I wrote a paper on modelling running speeds in athletic track events. The paper was never published, but there has been some subsequent interest in the claim I made then that the fastest possible human speeds would, on average, be achieved in sprint races of around 133 metres. With updated data, I now believe that the overall speed in a 141 m. race would exceed that in 100 or 200 metre races. The original paper, not updated with subsequent world-record data, is still available. If anyone has any comments or interest, or would like to use the material for some purpose, please contact me by email.
I am passionately interested in classical music (excluding opera), particularly instrumental and chamber music. Duplicate Bridge consumes about one of my evenings a week - decent Club level. Usual other stuff (fitness, swimming, walking, eating, reading, concerts, etc - you know the sort of thing). I spend a lot of time in Germany and love France even more.
Patrick Royston, March 2013