Psychological Health Benefits of Companion Animals Following a Social Loss

Dawn C. Carr, PhD; Miles G. Taylor, PhD; Nancy R. Gee, PhD; Natalie Sachs-Ericsson, PhD

Disclosures

Gerontologist. 2020;60(3):428-438. 

In This Article

Results

Descriptive Results

The characteristics for individuals in each pet–loss group are given in Table 1. As noted earlier, only 75% of the overall sample was included in the models used to assess changes in loneliness. In analyses not shown (available on request), we examined differences between those retained in both samples relative to those not included in the loneliness models. There were no differences in any of the control measures. As a result, we report descriptive information for the larger sample. The overall sample reported an average overall 4-year increase in depressive symptoms of approximately 0.30 symptoms, and a small overall decrease in loneliness score of 0.02. However, these changes varied by pet–loss group. Without accounting for selection into pet/loss group, the average change in depressive symptoms for the groups experiencing a social loss was +0.89 (loss/no-pet) and +1.33 (loss/pet). The groups that did not experience a social loss reported small increases in depressive symptoms +0.21 (no-loss/no-pet) and +0.09 (no-loss/pet). With respect to loneliness, the loss/no-pet group (+0.21 points) and the no-loss/pet group (0.06) experienced an overall increase in loneliness. The other groups showed a small overall decrease in loneliness.

There are several other notable characteristics. First, although the average age of the sample is 66, the average age of the loss/pet group is 5 years younger (67.2) than the loss/no-pet group (72.4). In addition, the groups with pets (loss/pet, no-loss/pet) have a higher proportion non-Hispanic White (88.0% and 80.0%, respectively) relative to the nonpet owner groups (no-loss/no-pet, loss/no-pet; 69.1% and 70.0%, respectively). The no-loss/pet group has the highest proportion engaged in paid work with about 47% working, and all other groups reporting between 26% and 38% working.

Inverse-probability Weighted Regression Analysis

Results from the inverse-probability weighted regression models are provided in Table 2 and Table 3 and shown in Figures 1 and 2. Regarding depressive symptoms, after accounting for selection and controlling for well-established factors influencing psychological health, the groups who experienced a social loss (i.e., no-pet/loss and pet/loss) both reported a 4-year increase in depressive symptoms as shown by the POMs reported in the first column, and both experienced greater increases in depressive symptoms (no-pet/loss 2.576, pet/loss 1.207, respectively) relative to both control conditions (no-pet/no-loss 0.004 and pet/no-loss groups 0.117). However, among those who experienced a loss, those with a pet accumulated significantly fewer depressive symptoms. This suggests that even though experiencing a loss was associated with more depressive symptoms, the effects were significantly lower for those with a pet. Both of these key findings support our first hypothesis.

Figure 1.

Change in depressive symptoms by pet/loss group. Statistical significance shown in bars relates to difference relative to the no-pet/no-loss group. All other differences indicated by brackets. Significance is based on the following values: ***p < .001; **p < .01; *p < .05; + p < .08. Results are based on models given in Table 2.

Figure 2.

Change in loneliness score by pet/loss group. Statistical significance shown in bars relates to difference relative to the no-pet/no-loss group. All other differences indicated by brackets. Significance is based on the following values: ***p < .001; **p < .01; *p < .05; + p < .08. Results are based on models given in Table 3.

Regarding loneliness, experiencing a loss was associated with an overall increase in loneliness (e.g., +1.669 no-pet/loss and +0.288, pet/loss). The groups who did not experience a social loss showed a decrease in loneliness (−0.328, no-pet/no-loss and −0.023, pet/no-loss). Moreover, the no-pet/loss group had a significantly greater increase in loneliness relative to both non-loss control conditions (e.g., no-pet/no-loss and pet/no-loss; p < .001) and the pet/loss group (p < .01). Also, consistent with expectations, the pet/loss group did not experience significantly greater negative consequences with respect to loneliness than either group that did not experience a loss. Thus, we conclude in relation to our second hypothesis that having a pet buffered the negative consequences of loneliness associated with a social loss.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....