Reply to comments by Dr. Frisch and Dr. Van Howe
We appreciate the opportunity to respond to the thoughtful comments by Dr. Frisch and Dr. Van Howe.
They suggest the need for a trend test to determine whether the changes in age-standardized rates are statistically significant. We used variance-weighted least squares linear regression to estimate the average year-on-year change in the incidence rate per 100,000 men per year between 1979 and 2009. The average annual increase in the incidence rate is numerically small at 0.007 (95 % CI 0.005–0.010) per 100,000 per year, but it is statistically significant (p < 0.001). This modeled trend represents an absolute increase in the age-standardized incidence rate of 0.22 per 100,000, or about a 20 % increase over the 31-year period 1979–2009.