The steps to calculate SMR is given below:

1) The first step in calculating the SMR is to calculate the age-specific death rates in the reference population.

2) Next, enter the population size estimate for each age group in the study population of interest

3) Calculate the number of expected deaths in each age group

4) Sum across all the age-specific expected deaths to get the expected number of deaths for the study population.

5) The observed number of deaths, that is, the actual number of deaths in the study population during the same time period was then compared with the expected number in step 4 to get the SMR.

My question is usually the age-specific death rates uses annual rates and the study population is the mid-year population. Is it possible to use the death rates calculated from a different period, say 6 months (Jan-Jun). Then the expected deaths is compared with the observed deaths during the same period of Jan-Jun? For the reference and study population, mid-year population estimates are used.

I am interested in comparing deaths for a specified time period. Is the above logical? Am I violating any "person-years at risk" issue?

Any comments are welcomed. Thanks.