Citation

Zandberg J, Waller R, Barzilay R. JAMA Psychiatry 2023; ePub(ePub): ePub.

Copyright

(Copyright © 2023, American Medical Association)

DOI

10.1001/jamapsychiatry.2023.1409

PMID

37256593

Abstract

In Reply We respond here to issues raised by Kelly about our recent article,1 focusing on our description of publicly available data, treatment of missing data, and approach to control variables. As described in our article, we used WONDER from the US Centers for Disease Control and Prevention (CDC)3 to download data using the CDC's predetermined age bins, 20 to 24 years and 25 to 34 years, which we combined to create our target group (20 to 34 years). From 1989 forward, in a policy implemented to ensure confidentiality, WONDER data were suppressed if there were fewer than 10 deaths per cell (in our case, state per year within the age bin). Our conservative approach was to only include data if rates were available for both age bins. We also conducted sensitivity analyses separately for each age bin, which were included in the supplementary materials. Alternative approaches to downloading the data could allow for calculation of suicide rates that are otherwise suppressed, thus breaching the confidentiality policy2 (eg, subtracting rates from the 20 to 34-year age bin could lead to a calculation of rates for the 20 to 24-year age bin that might be less than 10). In our article, we clearly described how we obtained data and provided a supplemental table presenting suicide rates and missing data for every year and state used in our analyses.

Language: en