Hi All
I am trying to calculate incidence of disease 'x' using stset and strate in Stata 12. I have split the data by age groups, calendar year and smoking(considering it as a time-varying covariate). Using strate I then calculated incidence rate of 'x' per 1000 person-years by smoking and calendar time. However, the incidence also increases by age and slightly different by sex. But stratifying the rates by so many variables will just complicate the results. So I wish to calculate age and sex adjusted/standardised incidence rates of 'x' by smoking and calendar year.
The options I have considered are:
1. using stdize command. However, the output doesn't look like it takes person years into account and that beats the purpose of survival analysis
The other option that I can think of is to use poisson regression and then use 'predict ir' to calculate adjusted incidence rates. But not sure what the syntax will be or if that's the correct way to do it.
Any help or suggestions will be appreciated.
Thanks,
Nafeesa
I am trying to calculate incidence of disease 'x' using stset and strate in Stata 12. I have split the data by age groups, calendar year and smoking(considering it as a time-varying covariate). Using strate I then calculated incidence rate of 'x' per 1000 person-years by smoking and calendar time. However, the incidence also increases by age and slightly different by sex. But stratifying the rates by so many variables will just complicate the results. So I wish to calculate age and sex adjusted/standardised incidence rates of 'x' by smoking and calendar year.
The options I have considered are:
1. using stdize command. However, the output doesn't look like it takes person years into account and that beats the purpose of survival analysis
Code:
. dstdize x pop ageband sex, by(smoking calendaryear) print
-----------Standard Population-----------
Stratum Pop. Dist.
----------------------------------------
50 Female 8580 0.204
50 Male 7402 0.176
60 Female 8309 0.197
60 Male 7867 0.187
70 Female 3579 0.085
70 Male 3979 0.095
80 Female 1082 0.026
80 Male 969 0.023
90 Female 182 0.004
90 Male 142 0.003
----------------------------------------
Total: 42091
-------------------------------------------------------------------
-> smoking calendaryear= 0 102
-----Unadjusted----- Std.
Pop. Stratum Pop.
Stratum Pop. Cases Dist. Rate[s] Dst[P] s*P
-------------------------------------------------------------------
50 Female 37 11 0.125 0.2973 0.204 0.0606
50 Male 40 9 0.135 0.2250 0.176 0.0396
60 Female 34 9 0.115 0.2647 0.197 0.0523
60 Male 38 9 0.128 0.2368 0.187 0.0443
70 Female 38 11 0.128 0.2895 0.085 0.0246
70 Male 28 11 0.095 0.3929 0.095 0.0371
80 Female 41 10 0.139 0.2439 0.026 0.0063
80 Male 27 3 0.091 0.1111 0.023 0.0026
90 Female 11 2 0.037 0.1818 0.004 0.0008
90 Male 2 0 0.007 0.0000 0.003 0.0000
-------------------------------------------------------------------
Totals: 296 75 Adjusted Cases: 79.3
Crude Rate: 0.2534
Adjusted Rate: 0.2681
95% Conf. Interval: [0.2099, 0.3262]
-------------------------------------------------------------------
-> smoking calendaryear= 0 104
-----Unadjusted----- Std.
Pop. Stratum Pop.
Stratum Pop. Cases Dist. Rate[s] Dst[P] s*P
-------------------------------------------------------------------
50 Female 35 12 0.127 0.3429 0.204 0.0699
50 Male 36 11 0.130 0.3056 0.176 0.0537
60 Female 33 11 0.120 0.3333 0.197 0.0658
60 Male 44 15 0.159 0.3409 0.187 0.0637
70 Female 33 14 0.120 0.4242 0.085 0.0361
70 Male 33 18 0.120 0.5455 0.095 0.0516
80 Female 29 17 0.105 0.5862 0.026 0.0151
80 Male 20 9 0.072 0.4500 0.023 0.0104
90 Female 7 2 0.025 0.2857 0.004 0.0012
90 Male 6 1 0.022 0.1667 0.003 0.0006
-------------------------------------------------------------------
Totals: 276 110 Adjusted Cases: 101.6
Crude Rate: 0.3986
Adjusted Rate: 0.3680
95% Conf. Interval: [0.3053, 0.4307]
-------------------------------------------------------------------
Summary of Study Populations:
smoking
calend~r N Crude Adj_Rate Confidence Interval
--------------------------------------------------------------------------
0
102 296 0.253378 0.268057 [ 0.209945, 0.326170]
0
104 276 0.398551 0.368006 [ 0.305284, 0.430729]
The other option that I can think of is to use poisson regression and then use 'predict ir' to calculate adjusted incidence rates. But not sure what the syntax will be or if that's the correct way to do it.
Any help or suggestions will be appreciated.
Thanks,
Nafeesa
