Posttraumatic Stress Disorder, Health-Related Social Needs, and Cognitive Outcomes in US Veterans.

Abstract

Posttraumatic stress disorder (PTSD) and health-related social needs (HRSNs) have been associated with cognitive decline, but their joint associations with mild cognitive impairment (MCI), Alzheimer disease (AD), and AD-related dementias (ADRD) are unclear. To evaluate the independent and joint associations of PTSD and HRSNs with incident MCI, AD, and ADRD and to assess additive and multiplicative interaction. Retrospective cohort study using 2015 to 2024 National Veterans Health Administration (VHA) electronic health record data. Participants were veterans aged 55 to 100 years with a qualifying VHA encounter in 2015 and no preindex MCI, AD, or ADRD, followed up through December 31, 2024. Time-varying PTSD, defined by at least 2 coded encounters on different dates, and time-varying HRSNs across 9 domains identified from structured data and clinical notes using natural language processing. The primary outcome was incident composite MCI, AD, or ADRD; secondary outcomes were each component separately. Time-varying Cox models estimated adjusted hazard ratios (aHRs) and 95% CIs. Multiplicative interaction was assessed using a PTSD × HRSN product term, and additive interaction using the relative excess risk due to interaction (RERI) and attributable proportion (AP). The cohort included 3 817 970 veterans (mean [SD] age, 70.3 [9.1] years; 3 664 349 [96.0%] male), of whom 270 282 (7.1%) met the composite outcome with median (IQR) follow-up of 8.84 (5.02-9.03) years for the composite outcome. PTSD (aHR, 1.33; 95% CI, 1.32-1.35) and HRSNs (aHR, 2.17; 95% CI, 2.13-2.20) were independently associated with the composite outcome. Compared with neither exposure, aHRs were 2.14 (95% CI, 2.11-2.17) for HRSNs only, 1.13 (95% CI, 1.07-1.19) for PTSD only, and 2.87 (95% CI, 2.82-2.92) for both exposures. Multiplicative interaction was observed for the composite outcome (aHR, 1.19; 95% CI, 1.13-1.26), with positive additive interaction (RERI, 0.61; 95% CI, 0.54-0.67; AP, 0.21; 95% CI, 0.19-0.23). Additive interaction was observed for MCI and ADRD but not AD; multiplicative interaction was observed for ADRD but not MCI or AD. In this cohort study, PTSD and HRSNs were independently associated with incident cognitive outcomes, and co-occurring PTSD and HRSNs may identify veterans with higher observed hazards of cognitive impairment and dementia.