We do not encourage sub-national estimates of the MMR. In Nigeria, even though the MMR is high, the number of deaths per state will be insufficient to produce statistically stable estimates. Also we do not know where the sister lived when she died.

To calculate the MMR, you first need to calculate the maternal death rate (deaths per 1000 women 15-49) and then you divide it by the general fertility rate (births per 1000 women 15-49. I don't see the calculation of the GFR in your code.

The GFR would not be the one in the report, but is calculated for the past seven years, with the age distribution standardized on the age distribution of the household population at the time of the survey. (At least that's how we do it here.)

Later in 2016 we will post Stata programs for complex indicators, including the MMR, on the website. In the meantime, I can distribute my own Stata program for this purpose on a case-by-case basis.