Understanding the Complexity of Analyzing Veteran Suicide Rates

Andrew Morral, PhD

Credit: RAND Corporation

Recent estimates suggest that between 2017 and 2020, veteran suicide rates were 1.57 to 1.66 times higher than those of nonveterans, after adjusting for age and sex differences.¹

The issue of veteran suicide rates in the US has been a matter of concern for the US Department of Veterans Affairs (VA) for some time.²

However, it is important to understand the nuances of these findings and the methodologies employed to arrive at these conclusions. In a research letter authored by Andrew R. Morral, PhD, and investigators from RAND Corporation, they addressed the 2 standard procedures used to compare suicide rates among veterans and nonveterans: direct standardization and indirect standardization.¹

The analysis consisted of 4 years' worth of data (2017-2020) published by the VA on suicide counts and population sizes for both veterans and nonveterans. A couple of standardized approaches were employed: direct standardization, and indirect standardization.

Direct standardization involved weighting mortality rates for different demographic strata of veteran and nonveteran populations to match the age and sex distribution of the US population. The second method adjusted nonveteran suicide rates to align with the age and sex distribution of the veteran population.

The analysis of the data revealed significant differences in age and sex distribution between the veteran and nonveteran populations. Veterans tended to be older and predominantly male.

These demographic disparities, coupled with variations in suicide risk within each stratum, led to notable differences in the estimates produced by the 2 standardization methods. While both approaches indicated an elevated suicide risk among veterans compared with nonveterans, the direct standardization method yielded an adjusted mortality ratio of 1.59, which was approximately 6 times higher than the ratio derived from indirect standardization (1.10).

Investigators noted both direct and indirect standardization are valid methods for comparing mortality rates. However, direct standardization is particularly useful when assessing disease burdens across multiple groups.

The weighted rates generated by direct standardization lack intrinsic meaning and do not represent the actual mortality risk of the populations being compared, according to the letter.

Furthermore, when a group is demographically dissimilar from the reference population, as is the case with veterans compared to the US population, direct estimates heavily favor the relative mortality risk of groups that are underrepresented in the reference population used for standardization.

However, the indirect standardization approach treats all veteran suicides equally when determining relative risks. This contrasts the direct standardization method, which would require a significantly higher number of suicides among elderly male veterans to impact the resulting adjusted mortality ratio as much as a single additional suicide by a young female veteran.

The letter stated that given the substantial variations in suicide risk across different demographic strata, age- and sex-specific mortality ratios provide the most meaningful statistics for most purposes.

"When an aggregate statistic is required to compare suicide risk among veterans and nonveterans, the correct interpretation of indirect standardization results might be more easily conveyed to general audiences because it compares actual veteran suicide rates with those of similar nonveterans," investigators wrote.

References:

1. Morral AR, Schell TL, Smart R. Comparison of Suicide Rates Among US Veteran and Nonveteran Populations. JAMA Netw Open. 2023;6(7):e2324191. doi:10.1001/jamanetworkopen.2023.24191
2. Walter K. US Government Updates Guidelines to Reduce Veteran Suicide. HCPLive. August 28, 2019. https://www.hcplive.com/view/us-government-updates-guidelines-reduce-veteran-suicide