In this article, I describe and illustrate Stata programs that facilitate i) the fitting of smooth age–period–cohort models to event data and ii) the plotting of observed and fitted rates. The programs include postestimation functionality and flexibility to fit models not possible using Stata's glm command. What distinguishes this article from a recent Stata Journal article on age–period–cohort models by Rutherford, Lambert, and Thompson (2010, Stata Journal 10: 606–627) is that the emphasis here is on extrapolating the model fit to make projections into the future.
MøllerB., FekjærH., HakulinenT., SigvaldasonH., StormH. H., TalbäckM., and HaldorsenT.2003. Prediction of cancer incidence in the Nordic countries: Empirical comparison of different approaches. Statistics in Medicine22: 2751–2766.
5.
RutherfordM. J., LambertP. C., and ThompsonJ. R.2010. Age–period–cohort modeling. Stata Journal10: 606–627.
6.
SasieniP. D.1994. snp7: Natural cubic splines. Stata Technical Bulletin22: 19–22. Reprinted in Stata Technical Bulletin Reprints, vol. 4, pp. 171–174. College Station, TX: Stata Press.
7.
SasieniP. D., and AdamsJ.1999. Effect of screening on cervical cancer mortality in England and Wales: Analysis of trends with an age period cohort model. British Medical Journal318: 1244–1245.
8.
SasieniP. D., and AdamsJ.2000. Analysis of cervical cancer mortality and incidence data from England and Wales: Evidence of a beneficial effect of screening. Journal of the Royal Statistical Society, Series A163: 191–209.
