BackgroundPrognostic models in peripheral T cell lymphoma (PTCL) have identified biological factors including age, performance status, LDH, and bone marrow involvement as prognostic for survival. The association of nonbiological factors, termed social determinants of health (SDH), on PTCL outcomes remains unexplored. MethodsTo evaluate the impact of actionable SDH on PTCL mortality across race groups, we conducted a retrospective cohort study that included all White, Hispanic, Asian/Pacific Islander (PI) and Black adult patients with PTCL (AITL, ALCL, NKTCL and PTCL, NOS), diagnosed from 2000 to 2020, in California. We utilized Chi2 and Wilcoxon rank-sum tests for descriptive metrics and Kaplan-Meier statistics for mortality estimation. Regression models included patient- (age, sex, race, stage, Charlson Comorbidity Index, histology, treatment type, treatment at an academic center, payer), and neighborhood-level factors (socioeconomic (SES) quintile, proportion without a high school diploma, and rural/urban). Risk factors significant in univariate regression at a P-value < .10 were incorporated into the multivariable model. FindingsOur analysis included 6158 patients: 51.8% White, 25.8% Hispanic, 14.7% Asians/PI, and 7.6% Black. Hispanics exhibited the longest median survival (33 months) followed by Whites, Blacks, and Asian/PI (25, 20, and 14 months, respectively; P = .011). Risk factors independently associated with inferior lymphoma-specific survival (LSS) included Asian/PI compared with NH White race (HR, 1.23; 95% CI, 1.10-1.34; P = .0002), AITL and ALCL compared with PTCL, NOS histology (AITL HR, 1.14; 95% CI, 1.02-1.25; P = .011; ALCL HR, 1.15; 95% CI, 1.04-1.26; P = .004), academic compared to nonacademic facility-type (HR 0.71; 95% CI, 0.64-0.77; P < .01), Medicare compared with uninsured status (HR 1.48, 95% CI, 1.25-1.73; P < .01), and the lowest 3 education quartiles compared to the highest education quartile (Q2 HR 1.13; 95% CI, 1.01-1.25; P = .021; Q3 HR 1.14; 95% CI, 1.02-1.26; P = .018; Q4 HR 1.22; 95% CI, 1.08-1.36; P < .001). In the least resourced patients, histology, treatment, treatment facility-type, payer and education were independently prognostic for LSS. Treatment at an academic institution was associated with a striking improvement in LSS (academic institution: yes = 101 months, no = 17 months; P < .01). InterpretationThese data identify treatment facility-type, payer and education, as an independent actionable SDH for PTCL mortality. Treatment center-type had the strongest prognostic association with LSS in our PTCL cohort, conferring a risk reduction of PTCL mortality by nearly 30%. Interventions targeted at improving access to academic centers may improve outcomes for all PTCL patients.