Very less number of cases

[Akif Alig]

I want to do an analysis on cancer
my outcome variable is a binary variable (1 = cancer, 0 = not having cancer)

The total sample size is N = 72,250
and cancer patients are only 450

1 = 450
0 = 68,000


I have 8 independent variables (age, sex etc.)

When I am running the logistic regression, all the independent variables are coming insignificant.

I think it is because, in my dependent variable, there are very few cases.

I have tried bootstrapping but found similar results.
Is there any solution of this problem


Any help would be much appreciated.
Warmest regards

[Carlo Lazzaro]

Akif:
try to collect more data, especially on level 1, otherwise the -logistic- machine will return insignificant coefficients again.
As an aside, please note that here insignificant coefficients are in fact informative, as they warn you about increasing your sample size.

[Rich Goldstein]

first, it appears that you have a lot of missing data (450+68,000 is considerably smaller than 72,250) so that needs looking at

second, you may find the user-written -relogit- command to be helpful (use -search- to find and download)

third, depending on a lot of info you don't provide, you might find it useful to take a sample of the control group and/or to do matching and limit your analysis to successful matches

[Akif Alig]

Thank you so much Carlo and Rich for your expert advice