Hostility and stress are known risk factors for heart disease, but are costly to assess on a large scale. Therefore, researchers used language expressed on Twitter to characterize county-level psychological correlates of age-adjusted mortality from artherosclerotic heart disease. Language patterns reflecting negative social relationships, disengagement, and negative emotions emerged as risk factors; and positive emotions and engagement emerged as protective factors against heart disease. Thse correlations remained significant after controlling for income and education levels. A cross-sectional, regression model based only on Twitter language predicted heart disease better than a model that combined ten well-known risk factors such as smoking, diabetes, obesity, and hypertension.