Dear All:
I should be grateful if one could suggest a way of estimating standard errors using boostrap resampling method and predicted probabilities from a logistic regression model to calculate attributable risk percentage (AR%) and population attributable risk perentage (PAR%) 95% confidence intervals, for overweight & obesity on hypertension and diabetes.
I have used this formulae to calculate AR% and PAR%:
AR% = (sum(Pi) - sum(P*i))/sum(Pi),
where Pi are predicted probabilities for overweight or obesity on hypertenson/diabetes, P*i is the predicted probability when for non hypertensives/diabetics.
I will post the summary to the list.
Many thanks in advance for your kind help.
With best wishes,
Etienne Magloire Minkoulou
Research Associate
Endocrine & Diabetes Unit
Faculty of Medicine & Biomedical Sciences
B.P. 13033 YAOUNDE
CAMEROON - AFRICA
Fax: (237) 31 52 35
E-mail: [log in to unmask]
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|