I wish to make comparisons between injury death rates and rates of child
maltreatment. I have computed age-specific injury death rates per 100,000,
Rate Ratios (RR) and 95% confidence intervals using the Stata command "ir".
Similarly, I have computed maltreatment rates per 1,000, RR, and confidence
intervals for these same age and race groups (and during the same period).
For example (white is the reference group):
ir deaths black pop, by(agecat)
ir subs black pop, by(agecat)
Now, I would like to compute a ratio of the maltreatment rate to the rate of
death for a given age/race group (I am exploring the all-cause injury death
rate as a proxy for child welfare service needs). In other words, I would
like to examine the association between maltreatment rates and death rates,
within race. To accomplish this, I have used the "iri" command, entering the
population count twice (since the same age/race population was exposed to
both maltreatment and death in a given year).
For example, having computed black to white RR for deaths and maltreatment
above, I now compute a RR of maltreatments to deaths for each race/age group
combination:
black infants:
iri subs deaths pop pop
Are there any obvious problems with this method from a statistical
standpoint? Can I consider the confidence intervals to be valid? Any other
suggestions for how I might compare these two rates? I've explored some
epidemiology text books, but haven't identified anything relevant. Any
thoughts/suggestions would be very welcome.
Thank you.
Emily
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
