Loneliness does (not) have cardiometabolic effects: A longitudinal study of older adults in two countries

Mass media increasingly report a "loneliness epidemic." A growing academic literature claims downstream effects of this experience on surrogate markers of cardiometabolic risk. Evidence on such influences is based on flawed samples and methodologies, rendering inferences questionable. The current study tested these claims.

Analysis was based on three-wave data on older adults from two national probability samples-the U.S. Health and Retirement Study (HRS) and the English Longitudinal Study of Ageing (ELSA). Models were gender-differentiated. Cardiovascular states were indexed by systolic and diastolic blood pressure, and metabolic condition by hemoglobin A1c. Fixed effects models were used for initial investigation, and subsequent triangulation was through a first-differencing approach with instrumental variables.

Loneliness had no linkage with any of the three outcomes. Nor were prevalences indicative of an epidemic of this affective state. Both gender and cross-national variations emerged: women were lonelier than men in each sample, while ELSA participants of both genders were less so than their HRS counterparts.

Contra previous literature, loneliness may not have cardiometabolic implications. Such nonreplications are increasingly common in the emerging "biosocial science" literature. Potential sources are discussed. More rigorous methods are available and urgently need incorporation to root out flawed inferences and conceptual models.