LETTERS TO THE EDITOR

In Response

We thank Tumin and Tobias¹ for their thorough reading of our article, “Race/Ethnicity and Sex Both Affect Opioid Administration in the Emergency Room,” and for their astute comments. In response to their concern, we have carefully reconsidered our original analysis.

After reanalysis, we can confirm that there was, indeed, an error in the statistical coding of the plan of analysis. Specifically, the code we used did not properly consider the sampling design of the National Hospital Ambulatory Medical Care Survey. We specified the sampling weight (PTWT) as a frequency weight and inadvertently omitted the sampling strata (CSTRATM) from the survey design. This caused the precision estimates (eg, 95% confidence interval) from the models to be far too narrow in their coverage. We have now double-checked every line of code and can ensure the accuracy of the article revision.

We sincerely appreciate Tumin and Tobias’s¹ sighting this coding error and bringing it to our attention—it reflects a readership committed to responsible and reproducible science.