Table 2.

Robustness of carcinogenesis parameter estimation for prostate cancer.

Year of observation	k	b	A/1000	R2
	Number of carcinogenic events ± s.e.m.	Average time between events, years ± s.e.m.	Maximal populational susceptibility, % ± s.e.m.	Goodness of fit
1999	40.72 ± 1.28	1.876 ± 0.063	31.79 ± 0.48	0.9992
2000	39.56 ± 1.28	1.931 ± 0.067	32.23 ± 0.50	0.9992
2001	40.59 ± 1.16	1.873 ± 0.057	32.00 ± 0.43	0.9993
2002	38.82 ± 0.99	1.955 ± 0.053	31.57 ± 0.38	0.9994
2003	38.37 ± 1.25	1.981 ± 0.069	28.82 ± 0.45	0.9991
2004	38.10 ± 1.41	1.992 ± 0.079	27.94 ± 0.49	0.9988
2005	38.67 ± 1.29	1.959 ± 0.070	27.33 ± 0.43	0.9990
2006	39.85 ± 1.21	1.886 ± 0.061	28.30 ± 0.39	0.9991
2007	40.14 ± 1.46	1.863 ± 0.072	28.67 ± 0.47	0.9987
2008	41.56 ± 1.58	1.784 ± 0.072	25.49 ± 0.43	0.9984
2009	42.91 ± 1.79	1.711 ± 0.075	23.35 ± 0.42	0.9979
2010	44.39 ± 2.16	1.651 ± 0.084	21.62 ± 0.45	0.9971
2011	44.97 ± 2.48	1.623 ± 0.094	21.14 ± 0. 50	0.9962
2012	44.19 ± 2.32	1.648 ± 0.090	16.84 ± 0.38	0.9964

The parameters are determined for the gamma distribution fitted to actual cancer incidence data (see Fig. 1). The gamma distribution was selected instead of the Erlang distribution to show precise estimates for the number of carcinogenic events. Prostate cancer was selected due to the highest incidence, the highly efficient screening procedure, the highest estimated number of carcinogenic events and the dramatic variation in incidence between the years of observation.