Recent publications in the journal suggested that living alone was a significant predictor of mortality in subjects at risk of or with atherothrombosis1 and that loneliness was a significant predictor of functional decline and death in those 60 years or older.2 The authors and an accompanying commentary elegantly crystallized the great difficulty in defining and studying “social support.”1,3 In the Dubbo Study4 of 2805 Australian subjects 60 years or older with long-term follow-up, we have re-examined some of these issues using a social support scale derived from the 19-point Medical Outcomes Study Social Support Survey.5