Age adjustment in ecological studies: using a study on arsenic ingestion and bladder cancer as an example

BMC Public Health

Table 3 Estimates of incremental relative risks (IRR) for women obtained by three different methods

Predictors	Direct Method	Indirect Method	Variable Method
	IRR [SE]	IRR [SE]	IRR [SE]
Arsenic Exposure^a
0.05-0.08 mg/L	-0.125 [0.045]	-0.077 [0.030]	-0.095 [0.046]
0.09-0.16 mg/L	0.259 [0.062]	0.162 [0.041]	0.214 [0.063]
0.17-0.32 mg/L	-0.222 [0.082]	-0.136 [0.055]	-0.143 [0.082]
0.33-0.64 mg/L	0.103 [0.094]	0.060 [0.062]	0.032 [0.093]
> 0.64 mg/L	0.489 [0.075]	0.349 [0.050]	0.354 [0.076]
Urbanization Index^b	0.324 [0.097]	0.241 [0.064]	0.299 [0.212]
Cigarette Sale^c	0.011 [0.219]	0.009 [0.146]	-0.004 [0.219]
Age^d
20-29 years	NA^e	NA	0.122 [0.153]
30-39 years	NA	NA	-0.249 [0.146]
40-49 years	NA	NA	0.346 [0.217]
50-59 years	NA	NA	-0.170 [0.287]
60-69 years	NA	NA	0.065 [0.373]
> 69 years	NA	NA	-0.417 [0.304]
p Value for the Model^f	< 0.001	< 0.001	< 0.001

^aincremental relative risk for each 1% increase in residents exposed to arsenic levels in each exposure category.
^bincremental relative risk for each one-unit increase in urbanization index.
^cincremental relative risk for each 100 cigarettes sold per year.
^dincremental relative risk for each 1% increase in residents in each age group.
^enot included in the analyses.
^fp value for F test of the significance of the model.

ISSN: 1471-2458