, Volume 52, Issue 1, pp 155-156
Date: 23 Sep 2012

Comments on “Background stratified Poisson regression analysis of cohort data” by Richardson and Langholz, Radiat Environ Biophys 51(1): 15–22

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access
This is an excerpt from the content

We were interested to read the paper by Drs Richardson and Langholz (2012), concerning the fitting of Poisson regression models to grouped survival data. Their approach provides a useful alternative to the fitting of possibly many stratum-specific parameters; for example, via the AMFIT program within the Epicure package (Preston et al. 1993). This program has been used in various analyses of radiation-exposed populations. However, we would like to correct the suggestion that Muirhead et al. (2009) used AMFIT in the 3rd analysis of the UK National Registry for Radiation Workers (NRRW). Rather, both in that analysis and in earlier NRRW analyses (Kendall et al. 1992; Muirhead et al. 1999), bespoke software was used to fit linear relative rate models based on the score function for a multinomial likelihood.

The finding that stratum-specific parameters can be eliminated from inference based on multiplicative models for Poisson counts and so used to test for trends in proportions goes back a