Age heaping is an important source of demographic bias in many
countries. There is little scholarship on the effect of age heaping on
cancer rate estimation. We use Nigerian demographic and cancer registry
data to (1) quantify age heaping in states and within cancer
registries using the Myers Blended Index and (2) examine the effect
of residual age heaping bias on age standardized cancer rates (ASRs).
We find severe age heaping at the level of the state and within cancer
registries which is more pronounced in the north than the south. Further,
we find less age heaping among women in registries compared to
both men in registries and to women in the general population. Lastly,
1
we find little evidence of residual age heaping bias comparing ASRs
estimated using the population data as given (Base) to Arriaga and
Strong smoothed imputed populations. The geographic results likely
reflect economic and educational disparities between the north and
south. The gender difference may signify that higher educated women
do more screening and seek care more often. Lastly, the absence of
ASR differences between the Base and smoothed populations provides
evidence that age heaping may not seriously bias ASRs in countries
with severe age heaping.