Combining psychosocial data to improve prediction of cardiovascular disease risk factors and events: The National Heart, Lung, and Blood Institute-sponsored women's ischemia syndrome evaluation study

BACKGROUND: There is overlap among psychosocial predictors of cardiovascular disease (CVD). The usefulness of combining psychosocial variables as risk markers for CVD needs investigation. METHODS: Participants were 493 women in the NHLBI WISE study. Multivariate combination of Beck Depression Inventory (BDI), State-Trait Anxiety Inventory (STAI), Social Network Index (SNI), and Cook-Medley hostility subscales was evaluated, and principal components analysis also conducted. Relationships of composite psychosocial risk markers to CVD events and risk factors were assessed. RESULTS: The multivariate block of SNI, Cook-Medley Hostile Affect subscale, STAI, and BDI predicted CVD events (χ = 27.8, df = 6, p < .001). Scalewise factor analysis revealed 2 factors: negative affectivity (NA) and hostility (explained variance, 45.6% and 17.1%, respectively). NA was associated with BMI (β [SE] = 0.18 [0.09], p = .04), hostility with metabolic syndrome (exp(β) = 0.60 [0.28], p = .04). Both factors were associated with blood pressure (BP): NA with SBP (β = 2.53 [1.04], p = .02) and DBP (β = 1.66 [0.60], p = .02); hostility with SBP (β = 2.72 [1.13], p = .02) and DBP (β = 1.83 [0.65], p = .005). Neither factor predicted CVD events. Original scales predicted CVD events: lower SNI (HR = 0.74, CI = 0.57-0.96), lower Hostile Affect (HR = 0.80, CI = 0.56-1.03), and higher BDI (HR = 1.33, CI = 1.08-1.74). CONCLUSIONS: In women with suspected ischemia, multivariate combination of psychosocial risk markers predicts CVD events; derived psychosocial factors were associated with CVD risk factors but not events. Measuring common variance among psychosocial variables may be a useful research strategy.