In Reply: Drs Maserejian and McKinlay argue that the differences in prescribing that we noted for white vs nonwhite patients could be attributable to differences in SES. This is a possibility, given that we had only limited information about each patient's SES and therefore could not adjust for it completely. The National Hospital Ambulatory Medical Care Survey did record the type of insurance each patient carried, which is a crude measure of SES. Adjustment for this measure led to only a minimal reduction in effect size (for example, the odds ratio for receiving an opioid medication for black vs white patients with pain was 0.635 before adjustment and 0.646 after adjustment). Full adjustment for SES might be expected to attenuate this effect a bit more, but probably not completely.
Pletcher MJ, Kohn MA, Gonzales R. Demographic Characteristics and Opioid Prescribing—Reply. JAMA. 2008;299(15):1773-1774. doi:10.1001/jama.299.15.1774-a