Derivation and validation of a predictive model for chronic stress in patients with cardiovascular disease

Background: Chronic stress in patients with cardiovascular disease (CVD), including peripheral artery disease (PAD), is independently associated worse outcomes. A model that can reliably identify factors associated with risk of chronic stress in patients with CVD is needed. Methods: In a prospective myocardial infarction (MI) registry (TRIUMPH), we constructed a logistic regression model using 27 patient demographic, socioeconomic, and clinical factors, adjusting for site, to identify predictors of chronic stress over 1 year. Stress at baseline and at 1-, 6- and 12-month follow-up was measured using the 4-item Perceived Stress Scale (PSS-4) [range 0–16, scores ≥6 depicting high stress]. Chronic stress was defined as at least 2 follow-up PSS-4 scores ≥6. We identified and validated this final model in another prospective registry of patients with symptomatic PAD, the PORTRAIT study. Results: Our derivation cohort consisted of 4,340 patients with MI (mean age 59.1 ± 12.3 years, 33% females, 30% non-white), of whom 30% had chronic stress at follow-up. Of the 27 factors examined, female sex, current smoking, socioeconomic status, and economic burden due to medical care were positively associated with chronic stress, and ENRICHD Social Support Instrument (ESSI) score and age were inversely related to chronic stress. In the validation cohort of 797 PAD patients (mean age 68.6±9.7 years, 42% females, 28% non-white, 18% chronic stress) the c-statistic for the model was 0.77 and calibration was excellent. Conclusions: We can reliably identify factors that are independently associated with risk of chronic stress in patients with CVD. As chronic stress is associated with worse outcomes in this population, our work identifies potential targets for interventions to as well as the patients that could benefit from these.