To the Editor: We read with interest a report based on SEER data, in the CCG Quarterly, Newsletter, Summer 1999, which suggested that the decline in cancer mortality in young people in North America would stop by 2010 if the current trend continues .
However, we consider that there may be an alternative interpretation.
Data on childhood cancer mortality from England and Wales [2] clearly show a linear fall from around 8 to 3 deaths per 100,000 children age 0-14 per year from 1958 to 1995 (Fig. ).
Extrapolation of this curve would suggest the possibility of zero deaths from malignancy in this age group with the next 20 years. Perhaps a little optimistic?
The data from 1975-1995 from England and Wales (the same time period as presented for the SEER data) are equally well described by a straight line (r β«Χ‘β¬ 0.97) which extrapolates to zero deaths in 2018, or a second order polynomial (r β«Χ‘β¬ 0.97) which, in agreement with the presentation of SEER data, suggests no further reduction in mortality after around 2018 (Fig. ).
It is probable that both scenarios are equally unlikely. Deaths will not be eliminated completely within the next two decades and, conversely, progress in curing these children of their disease will not cease. Since these data support both the worst case and best case scenario equally well, it is clearly an important reminder of the dangers of extrapolation into the unknown. The inspection and interpretation of such complex things as mortality rates are probably best done in retrospect.