Browsing by Author "Klein, Liviu"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Predicting Incident Heart Failure in Women With Machine Learning: The Women's Health Initiative Cohort
    (Elsevier, 2021) Tison, Geoffrey H.; Avram, Robert; Nah, Gregory; Klein, Liviu; Howard, Barbara V.; Allison, Matthew A.; Casanova, Ramon; Blair, Rachael H.; Breathett, Khadijah; Foraker, Randi E.; Olgin, Jeffrey E.; Parikh, Nisha I.; Medicine, School of Medicine
    Background: Heart failure (HF) is a leading cause of cardiac morbidity among women, whose risk factors differ from those in men. We used machine-learning approaches to develop risk- prediction models for incident HF in a cohort of postmenopausal women from the Women's Health Initiative (WHI). Methods: We used 2 machine-learning methods-Least Absolute Shrinkage and Selection Operator (LASSO) and Classification and Regression Trees (CART)-to perform variable selection on 1227 baseline WHI variables for the primary outcome of incident HF. These variables were then used to construct separate Cox proportional hazard models, and we compared these results, using receiver-operating characteristic (ROC) curve analysis, against a comparator model built using variables from the Atherosclerosis Risk in Communities (ARIC) HF prediction model. We analyzed 43,709 women who had 2222 incident HF events; median follow-up was 14.3 years. Results: LASSO selected 10 predictors, and CART selected 11 predictors. The highest correlation between selected variables was 0.46. In addition to selecting well-established predictors such as age, myocardial infarction, and smoking, novel predictors included physical function, number of pregnancies, number of previous live births and age at menopause. In ROC analysis, the CART-derived model had the highest C-statistic of 0.83 (95% confidence interval [CI], 0.81-0.85), followed by LASSO 0.82 (95% CI, 0.81-0.84) and ARIC 0.73 (95% CI, 0.70-0.76). Conclusions: Machine-learning approaches can be used to develop HF risk-prediction models that can have better discrimination compared with an established HF risk model and may provide a basis for investigating novel HF predictors.
  • Thumbnail Image
    Item
    When the At-Risk Do Not Develop Heart Failure: Understanding Positive Deviance Among Postmenopausal African-American and Hispanic Women
    (Elsevier, 2021) Breathett, Khadijah; Kohler, Lindsay N.; Eaton, Charles B.; Franceschini, Nora; Garcia, Lorena; Klein, Liviu; Martin, Lisa W.; Ochs-Balcom, Heather M.; Shadyab, Aladdin H.; Cené, Crystal W.; Medicine, School of Medicine
    Background: African American and Hispanic postmenopausal women have the highest risk for heart failure compared with other races, but heart failure prevalence is lower than expected in some national cohorts. It is unknown whether psychosocial factors are associated with lower risk of incident heart failure hospitalization among high-risk postmenopausal minority women. Methods and results: Using the Women's Health Initiative Study, African American and US Hispanic women were classified as high-risk for incident heart failure hospitalization with 1 or more traditional heart failure risk factors and the highest tertile heart failure genetic risk scores. Positive psychosocial factors (optimism, social support, religion) and negative psychosocial factors (living alone, social strain, depressive symptoms) were measured using validated survey instruments at baseline. Adjusted subdistribution hazard ratios of developing heart failure hospitalization were determined with death as a competing risk. Positive deviance indicated not developing incident heart failure hospitalization with 1 or more risk factors and the highest tertile for genetic risk. Among 7986 African American women (mean follow-up of 16 years), 27.0% demonstrated positive deviance. Among high-risk African American women, optimism was associated with modestly reduced risk of heart failure hospitalization (subdistribution hazard ratio 0.94, 95% confidence interval 0.91-0.99), and social strain was associated with modestly increased risk of heart failure hospitalization (subdistribution hazard ratio 1.07, 95% confidence interval 1.02-1.12) in the initial models; however, no psychosocial factors were associated with heart failure hospitalization in fully adjusted analyses. Among 3341 Hispanic women, 25.1% demonstrated positive deviance. Among high-risk Hispanic women, living alone was associated with increased risk of heart failure hospitalization (subdistribution hazard ratio 1.97, 95% confidence interval 1.06-3.63) in unadjusted analyses; however, no psychosocial factors were associated with heart failure hospitalization in fully adjusted analyses. Conclusions: Among postmenopausal African American and Hispanic women, a significant proportion remained free from heart failure hospitalization despite having the highest genetic risk profile and 1 or more traditional risk factors. No observed psychosocial factors were associated with incident heart failure hospitalization in high-risk African Americans and Hispanics. Additional investigation is needed to understand protective factors among high-risk African American and Hispanic women.