Incident rate calculation and 95% CI

[Mayura Shinde]

Hi,

I have to calculate the 95% CI for incident rates. I have data on the number of events, person years and would like to estimate incident rates by a grouping variable comparing the incident rates by categories of group variable. STATA calculates the incident rate ratios, I used the ir command but it only gives incident rates estimates for each category of group variable. I wanted to know if I could calculate the 95% CI for incident rates as well. Appreciate your help!

. ir episode_ct	spirometry risk_years, level(95)
	spirometry
Exposed Unexposed Total
episode_ct	20917 53886 74803
risk_years	48018.5 169428.8 217447.3
Incidence rate	.4356029 .318045 .3440051
Point estimate [95% Conf. Interval]
------------------------+------------------------
Inc. rate diff.	.1175579 .1110726 .1240432
Inc. rate ratio	1.369626 1.347868 1.391697	(exact)
Attr. frac. ex.	.2698739 .2580875 .281453	(exact)
Attr. frac. pop	.0754643
+-------------------------------------------------
(midp) Pr(k>=20917) = 0.0000	(exact)
(midp) 2*Pr(k>=20917) = 0.0000	(exact)

Thanks,
Mayura

[Rich Goldstein]

your output is very hard to read and your question is not clear; please review the FAQ; meanwhile you might want to try poisson regression (I'm guessing somewhat here)

[Mayura Shinde]

Hi Rich,

I am sorry that the question was not clear. To give more details about the data file, I have counts of COPD exacerbation episodes "episode_ct", person years estimated as "risk_year" and spirometry is whether the patients had or did not have a spirometry test. I have to estimate the crude exacerbation rates overall and stratified by spirometry. I used the ir command which estimates the incident rate ratios and incident rates for patients with and without spirometry. However, I have to provide 95% CI for the incident rates. Could you please provide any guidance regarding whether STATA can estimate the confidence intervals for the incident rates.

ir episode_ct  spirometry risk_years, level(95)