Article Text

Download PDFPDF

Social networks and health: it’s time for an intervention trial
  1. Anthony F Jorm
  1. Correspondence to:
 Professor A F Jorm
 Centre for Mental Health Research, Australian National University, Canberra, ACT 0200, Australia;

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

A randomised controlled trial is necessary to find out if intervention on social networks has a health benefit for older people and if this could be extended to the whole population.

Back in 1988, James House and colleagues reviewed a number of studies on social relationships and mortality. They concluded that “…the theory and evidence on social relationships and health increasingly approximate that available at the time of the U.S. Surgeon General’s 1964 report and smoking and health…with similar implications for future research and public policy.”1 Since that time, the evidence has continued to accumulate. While not all studies have found an association, there is sufficient confirming evidence to take it very seriously. These findings raise three issues: Is the association causal? If it is causal, what is the mechanism? And can we use this finding to improve the health of older people?

The evidence base consists of prospective studies of social networks as a predictor of mortality. Such studies indicate either a causal relationship or some unknown confounder that is the true cause. It is clear that the association persists after adjusting for obvious confounders, such as baseline health status and health habits, but we can never fully rule out some unknown third factor by using a prospective methodology.

Nevertheless, prospective studies can help us to refine the possible causal factors. If social networks are protective, what components contribute to the effect? A study by Giles et al in this issue reports that networks of friends and confidants (which includes spouse confidants) predict mortality in older people, but networks of children and other relatives do not.2 The authors suggest that networks of friends might have a protective effect by influencing health behaviours and help seeking, as well as the psychological benefits they confer for depression, self efficacy, self esteem, coping, and morale. These findings suggest what sort of interventions might be useful in improving the health of older people.

Despite the many studies on this topic, this line of research has yet to be translated into health benefits for older people. The only sure way to find out if intervention has a health benefit, and whether the association is causal, is to carry out a randomised controlled trial. Fortunately, friendship networks are likely to be more modifiable than family relationships. We are now arguably at the point where a large scale trial is warranted. Befriending schemes have been developed and trialled for people suffering from depression and could provide a model for a potential intervention.3 The assumption of a befriending intervention is that the resulting friendships are qualitatively similar enough to serve the same function as lengthier, naturally occurring friendships. The challenge would be to carry out a trial on a sufficiently large sample, and over a sufficiently long enough time, to see an impact on mortality. Such a trial should examine effects on other health indicators besides mortality, such as disability, and a combined indicator like disability adjusted life years would be warranted. Factors that might mediate the association, such as health habits, help seeking, psychological state, and specific disease processes, should also be measured. If such a trial successfully showed that social networks could be modified with benefits to health, this would provide the basis for more widespread health promotion action, with the aim of increasing friendship networks in the whole population. Only then would we be ready for a US surgeon general’s report on social relationships and health.

A randomised controlled trial is necessary to find out if intervention on social networks has a health benefit for older people and if this could be extended to the whole population.



  • Funding: the author’s salary is paid by a National Health and Medical Research Council Research Fellowship.

  • Conflicts of interest: none declared.

Linked Articles