Background: Pet ownership is thought to make a positive contribution to health, health behaviours and the general well-being of older people. More specifically pet ownership is often proposed as a solution to the problem of loneliness in later life and specific 'pet based' interventions have been developed to combat loneliness. However the evidence to support this relationship is slim and it is assumed that pet ownership is a protection against loneliness rather than a response to loneliness. The aim of this paper is to examine the association between pet ownership and loneliness by exploring if pet ownership is a response to, or protection against, loneliness using Waves 0-5 from the English Longitudinal Study of Ageing (ELSA).
Methods: Using data from 5,210 men and women in the English Longitudinal Study of Ageing, cross-sectional and longitudinal regression analysis was used to assess the bi-directional relationship between loneliness and pet ownership among adults aged 50 + .
Results: In 2001 (wave 0) 41% of participants were pet owners compared with 30% in 2010 (Wave 5). The association between pet ownership and loneliness is stronger in women than men, and in both directions (i.e. pet ownership predicting loneliness and loneliness predicting pet ownership) and of the similar magnitude (OR 1.2-1.4). Age, social relationships, demographic factors and health behaviour variables have only a minimal influence upon the association between loneliness and pet ownership. The results of our longitudinal analysis showed that women who reported being lonely always in Waves 0 to 5 were more likely to have a pet in Wave 5.
Conclusion: Reported loneliness is dependent on socio-demographic characteristics such as gender, household income, household living arrangements and health status. Taking those factors into account, owning a pet significantly influences later reporting of loneliness in women in our longitudinal analysis. In the reverse direction, reported loneliness influences pet ownership in later waves. In both directions, the relatively strong gender interaction suggests the association is limited to women with effects for men minimal or non-existent.