Estimating relative risk of disease from outputs of logistyic regression when the disease is not rare See original record