Socioeconomic differentials in the mortality of petsBMJ 1998; 317 doi: https://doi.org/10.1136/bmj.317.7174.1671 (Published 19 December 1998) Cite this as: BMJ 1998;317:1671
Probably reflect the same differences in material circumstances as in their owners
- George Davey Smith, Professor of clinical epidemiology,
- Brenda Bonnett, Professor
- Department of Social Medicine, University of Bristol, Bristol BS8 2PR
- Department of Ruminant Medicine and Epidemiology, Swedish University of Agricultural Sciences, Box 7019, 750 07 Uppsala, Sweden
Papers p 1686
Factors associated with mortality among dogs have not been extensively studied, though recent work has examined age, breed, and sex distributions of mortality.1 There has, however, been little investigation of socioeconomic differentials in pet mortality. The study by Moloo et al in this week's issue (p 1686) is therefore welcome, although the results should be interpreted with caution.2 They show that the probability of reporting the death of a pet in the past year is greater for young people from less favourable socioeconomic backgrounds than for those from more favourable backgrounds.
Pet ownership may have some influence on psychological health,3 but it does not itself seem to influence mortality4 and therefore the dynamics of mortality among pet owners are likely to reflect those of the total population. In this regard the socioeconomic inequalities in mortality—shown in a wide range of populations—will be a key feature. There has been considerable research into explanations of socioeconomic differentials in mortality among humans,5 and similar categories of explanation may account for such differentials in pet mortality.
The “artefact” explanation suggests that socioeconomic differentials in mortality are more apparent than real, resulting from the way data are collected and analysed. In Moloo et al's study a potential artefact is created by the fact that the experience of a pet's death is analysed without information on whether the study household actually possessed a pet. Thus the results may simply reflect socioeconomic differences in pet ownership. In the United States6 and in Sweden (A Egenvall, personal communication) ownership of a pet is positively associated with household income, though this plateaus at high incomes.7 This pattern might, therefore, create higher probabilities of pet death in the past 12 months for affluent than for poor households and this will lead to the current data underestimating the true socioeconomic gradient in pet mortality.
A second artefact could relate to the type of pet. If poorer households were more likely to have pets with shorter lifespans—gerbils, mice, or birds—than richer ones then a higher probability of experiencing a death of a pet in the past 12 months would be expected. Evidence from the US suggests that ownership of most types of pets is higher among better off families. Finally, pets often die as the result of euthanasia, and poorer families might be more likely to put down their pets because of an inability to support the animal. Thus several reasons exist why the socioeconomic gradient in pet mortality might be artefactual.
Health related selection has been extensively investigated as an underlying cause of socioeconomic differentials in health among humans.8 The basic proposition is that poor health leads to unfavourable social circumstances rather than vice versa. With respect to pet mortality possible health related selection would occur if poorer people took on pets who were less healthy or older at the time of acquisition. If this were so then higher mortality among the pets of less affluent families would be expected. This suggestion has some plausibility since one determinant of the cost of pets may be their health potential (though aesthetics seems more important). In humans evidence suggests that health related selection makes only a small contribution to overall health inequalities,8 but the situation in pets is unclear.
A third category of explanation suggests that health related behaviour is responsible for socioeconomic differentials in health. Clearly the usual suspects—smoking and drinking—are unlikely to apply to pets. Differences in diet and exercise levels may exist between pets belonging to more or less affluent families. In humans it is important to recognise that health related behaviours do not simply reflect individual choice but relate to structural constraints on the lives of people in straitened circumstances.9 This is even clearer with respect to the health related behaviours of pets.
Socioeconomic differentials in medical care may contribute to human health inequalities but are not considered key determinants.10 There is some evidence of socioeconomic inequalities in veterinary care, although the degree to which this contributes to differentials in pet mortality is not known.
Psychological and psychosocial factors have been advanced to explain inequalities in health among humans. These include social support11 and work stress.12 More affluent families may have more pets than less affluent households, providing their pets with more elaborate social networks, but greater social support is not necessarily health inducing 13; indeed, cats in multicat households may be more stressed. Work stress is clearly not a candidate factor. Indeed, pets serve as an additional largely non-working group (along with unemployed and retired people, and, earlier this century, many women) among whom socioeconomic differentials in health are seen which are at least as great as those in working populations.14
A final category of explanation relates to material and structural factors, the importance of which has probably been underestimated with respect to socioeconomic differentials in human health.5 Advantage clusters cross sectionally and longitudinally across the course of people's lives, with those born in adverse circumstances having a higher risk of embodied inequality (in the form of low birth weight and short stature) and a lower probability of succeeding educationally. This leads to entry into less privileged sections of the labour market, exposure to low pay and hazardous work, and reliance on limited welfare payments in old age. Within this framework the way in which apparent “lifestyle” factors are outcomes of social processes which generate disadvantage in a wide variety of spheres becomes clear. For pets, the material circumstances they encounter will largely depend on the socioeconomic wellbeing of their owners.
The study by Moloo et al has many limitations and does not firmly establish the existence of socioeconomic differentials in pet mortality. However, if extrinsic—particularly violent—causes of death are put aside, then humans and other animals show some similarities in mortality dynamics.15 A parsimonious explanation of health inequalities should perhaps account for differentials seen in different species, as well as for the fact that inequalities in mortality among humans are seen in different epochs and in different countries. In this light explanations solely in terms of artefact, selection, health related behaviours, or particular stresses seem parochial. The structuring of advantage and disadvantage in the material and social environment across the course of a life could reflect a more stable process of social differentiation. Further research into socioeconomic differentials in the mortality of animals other than humans may play a role in advancing our understanding of this basic process.
Log in using your username and password
Log in through your institution
Sign up for a free trial