Documente Academic
Documente Profesional
Documente Cultură
Introduction
Beck and Katcher have noted that the use of pets as
therapy to improve the morale of the elderly and psychiatrically ill is based in part upon an assumption that pet owners
have better health than non-owners.' They stress the need
for more controlled trials, since the evidence is not consistent.23 In the course of a study in Washington County,
Maryland, the opportunity arose to investigate the relation
between pet ownership and suicide. Although there is no
literature suggesting that pet ownership protects against
suicide, it seems reasonable to believe that psychological
well-being might have a bearing on suicide.
Methods
In July of 1975, a non-official census of Washington
County, conducted by the Johns Hopkins Training Center
for Public Health Research, obtained information on an
estimated 90 per cent of the total county population. For
each adult, the information included race, sex, date of birth,
marital status, and education. For each household, the
following question was asked: "On the property on which
you live, are there any of the following animals or pets?"
Cattle, horses, pigs, sheep or goats, dogs, cats, chickens or
pet birds, and "others" were listed, to be checked if present.
A list of all deaths of individuals identified as residents
of Washington County at time of death was available for the
period July 16, 1975 to July 15, 1983. The 1975 census
number was determined for each decedent who participated
in that census.
In the eight-year period, there were 85 suicides by
Washington County residents. Of these, 17 had moved to the
county since the 1975 census, 17 had not participated in the
census, and three had not answered the questions about
animals on premises. For this study, the 48 suicides who
participated in the 1975 census and answered the question
about animals were listed as cases. Two sets of two controls
per case were then selected.
Set #1 comprised two living controls for each case,
selected from the birth date listing of the 1975 census as
being the next younger and next older individual of the same
race and sex as the case. Both controls selected were
checked to be sure they were still alive in Washington
From the Department of Epidemiology, School of Hygiene and Public
Health, Johns Hopkins University. Address reprint requests to Dr. Knud J.
Helsing, Johns Hopkins Training Center for Public Health Research, Washington County Health Department, PO Box 2067, Hagerstown, MD
21742-2067. This paper, submitted to the Journal October 16, 1984, was
revised and accepted for publication March 27, 1985.
AND
TABLE 1-Number of Suicides 1975-83 by Washington County, Maryland Residents Who Had Participated in a Private Health
Census, by Sex and Age
Age (years)
15-24
25-44
45-64
65+
TOTALS
Men
Women
Total
4
10
14
5
33
0
5
8
2
15
4
15
22
7
48
Both Controls
Had a Pet
One Control
Had a Pet
Neither Control
Had a Pet
Had a Pet
Did Not Have a Pet
8
9
13
7
6
5
1 223
% of Suicides
Having a Pet
% of Controls
Having a Pet
Odds ratio
(95% confidence
interval)
48
48
48
56.2
41.7
29.2
56.2
42.7
25.0
1.00 (0.50-2.01)
0.95 (0.20-5.93)
1.22 (0.37-4.00)
33
15
38
19
29
57.6
53.3
47.4
52.6
58.6
56.1
56.7
43.4
60.5
53.4
1.06 (0.46-2.46)
0.88 (0.06-13.7)
1.19 (0.24-5.93)
0.73 (0.08-6.42)
1.21 (0.24-6.05)
TOTAL
Dogs
Cats
Subgroups
Men
Women
Non-farm residences
Died 1975-78
Died 1978-83
a) for total; b) for dog ownership and cat ownership separately; c) for men and women separately; d) for persons in
non-farm residences, where the dog or cat is more likely to
be a pet; and e) for suicides between 1975 and 1978, closer in
time to the 1975 census, and for suicides after 1978. All of the
odds ratio estimates are close to 1.0.
Data from the second set of controls yielded 27 cases
with two controls for each identified in the 1975 census, 16
cases with only one of the controls so identified, and five
cases with neither control identified in the 1975 census. The
controls in set #2 were Washington County residents who
had died of causes other than suicide, in the same year as the
suicide. As with the first set of controls, there seems to be
virtually no association of pet ownership with suicide (Table
4). The OR estimate is 0.91 and the 95 per cent CI is
0.49-1.69. When the dead controls were compared with
suicides in the same sub-groups as shown in Table 3, again
there were no noteworthy associations found between suicide and pet ownership.
Finally, the estimated OR and CI was calculated for the
48 suicides and all controls, both living and dead, with the
result OR = 0.98 (0.66-1.45).
TABLE 4-Pet Ownership among Suicides and Matched Controls Who
Died of other Causes, Washington County, Maryland, 1975-83
Discussion
These data have several important limitations. The
study has not measured the possible effects of introducing
pets into people's homes. The fact that there is dog or cat on
one's premises does not always mean that the person considers that animal his or her own pet, and there is nothing
that measures the degree of attachment that existed. Also
the presence of a cat or dog in 1975 does not necessarily
mean that one was present in 1976, 1983, or whenever the
suicide took place.
Nevertheless, the ownership or a pet or lack thereof in
1975 should distinguish two groups that also differ substantially in this regard in subsequent years. Whatever may be
the positive health effects of pets, they seemed not large
enough to affect suicide substantially in this population. The
small number of suicides in this study does, of course, limit
the precision of our measurement so that our data would still
be reasonably compatible with as much as a 50 per cent
decrease in risk for suicide.
One strength of the study is that the suicides and the
controls all came from the same population and provided
information about the cats and dogs prior to the event. The
use of two sets of controls also adds strength to the findings.
Finally, the consistency of results regardless of how the data
are broken down lends confidence to the conclusion that cat
or dog ownership is not an important factor in suicide by the
owner.
ACKNOWLEDGMENTS
This research was supported in part by the National Institute of Mental
Health, NIH, 1 ROI MH 38317. The authors thank Mary Elgin for her help in
selecting controls for this study.
Pet Ownership
among Suicides
Both Controls
Had a Pet
One Control
Had a Pet
Neither Control
Had a Pet
HadaPet
Did Not Have a Pet
7
3
4
3
4
7
1224
REFERENCES
Control Did Not
Have a Pet
3
2