Pet Attachment and the Interpersonal Theory of Suicide

The media and public often claim that pets can work wonders for physiological or psychological problems, ranging from hypertension to loneliness (Allen, 2003). People may feel compelled by these claims because they fit with their personal experiences of enjoying their pets (Herzog, 2011). Often, individuals assume pet ownership confers a “blanket” positive effect for everyone and for whatever health concern is at hand, especially psychiatric concerns. This so-called “pet effect” does have support from some studies (Allen, 2003), but it does not have as much empirical backing as assumed. In the general pet ownership literature, some studies support positive benefits for different populations. Having pets has been associated with decreased loneliness in rural adolescents (Black, 2012), unmarried women living alone (Zasloff & Kidd, 1994), older adults living alone (Stanley et al., 2014), and older women (Krause-Parello, 2012). However, a systematic review of this literature reported mixed results (Brooks et al., 2018) and casts doubt on the idea that having pets is always helpful. In addition, minimal research has investigated the link between suicide and pet ownership, with an epidemiological letter to the editor finding no link between ownership and suicide deaths (Batty & Bell, 2018). To understand in what circumstances pet ownership is detrimental or helpful for suicide risk, investigation of related variables is imperative.

One such variable is attachment to one's pet. In the broadest definition of the term, pet attachment is the perceived strength and quality of the bond to one's pet (Johnson et al., 1992). When examining the strength of pet attachment, it has been found that the relationship is much more complicated. Some studies have found that pet attachment does not exert an effect: such as when examining loneliness in unmarried women living alone (Zasloff & Kidd, 1994) and loneliness and depression in adults living alone (Antonacopoulos & Pychyl, 2010). Other studies have found that increased pet attachment is detrimental. Higher pet attachment is associated with increased depressed mood in older women (Krause-Parello, 2012), higher levels of depression and loneliness in adults living alone (Antonacopoulos & Pychyl, 2010), and increased depression in rural older adults (Miltiades & Shearer, 2011).

It may be the type of attachment, rather than the strength of attachment, that determines whether pets are helpful or harmful to mental health (Zilcha-Mano et al., 2011). Using a two-dimensional model of attachment based on Bowlby's (1969) attachment theory, Zilcha-Mano et al. (2011) proposed that pet attachment anxiety or pet attachment avoidance orientation levels are better indicators of pet ownership effects on mental health. An anxious attachment style is associated with worries that the attachment partner will not be supportive or available when needed, and thus, the individual tries to maximize proximity to the attachment partner (in this case, an animal). An avoidant attachment style is associated with the individual striving to be self-reliant because the individual is distrustful of the attachment partner's intentions (again, a pet in this case; Zilcha-Mano et al., 2011).

Multiple studies have found that people who were low in pet attachment avoidance or anxiety consistently had better outcomes for psychological well-being and psychological distress (e.g., Zilcha-Mano et al., 2011, 2012). Specifically, pet attachment anxiety was linked to lower psychological well-being and greater psychological distress (Zilcha-Mano et al., 2011). The researchers also conducted an experimental study to see how pet attachment orientation would influence stress (Zilcha-Mano et al., 2012). They found that, in general, participants who had their pet in the room or who were instructed to think of their pet were able to generate more goals they wanted to pursue, were more confident they could achieve these goals, and had a lower increase in blood pressure when completing a stressful task, as compared to a control group who thought of an acquaintance. However, when specifically looking at those high in attachment avoidance, the opposite of these results was found. Compared to those with elevated avoidance scores in the control condition, those with elevated avoidance in the pet condition generated fewer goals, had less confidence in achieving these goals, and had elevated blood pressure while completing a stressful task. This study suggests that, for individuals with increased pet attachment avoidance, the presence of their pets can increase stress instead of alleviating it (Zilcha-Mano et al., 2012).

As stated before, there is minimal research investigating the effect of pet ownership on suicide risk, let alone investigating variables such as pet attachment on the relationship. To begin to parse apart these relationships, we utilized the interpersonal theory (IPT) of suicide as our framework (Chu et al., 2017; Van Orden et al., 2010). The IPT posits that multiple factors must converge before a person attempts or dies by suicide. To begin, an individual must feel disconnected from others and that they have no reciprocal, caring relationships (thwarted belongingness [TB]). The individual must simultaneously experience self-hatred to the point that they erroneously believe they are a burden on others and that others would be better off without them (perceived burdensomeness [PB]). When an individual experiences hopelessness about these states ever changing, it results in a desire for suicide. Although PB, TB, and hopelessness are necessary, they are not sufficient for someone to die by suicide or to make a lethal/near-lethal attempt. The individual must also display capability for suicide such as a low fear of death and a high pain tolerance (Chu et al., 2017; Van Orden et al., 2010), qualities that are believed to emerge from genetic factors and painful, provocative life experiences (Chu et al., 2017).

In the current study, we investigate TB and PB in particular due to attachments people can feel toward their pets. For example, pet owners often indicate that their pet provides them with companionship or a creature to love (Anderson, 2014). Therefore, having a pet could increase one's sense of belongingness and thereby buffer suicidal desire. Taking care of a pet could also affect one's perception of being a burden, as pet owners often feel responsibility, a sense of purpose, and reciprocity in caring for their pets (Langfield & James, 2009). This sense of responsibility or purpose could potentially alleviate feelings of being a burden and therefore also buffer against suicidal desire. Pet owners have been found to “humanize” their pets and thus view the animals on par with fellow humans (Uccheddu et al., 2019), which suggests that pet ownership may buffer the effects for individuals experiencing TB and PB, depending on the attachment relationship to the animal.

The current study investigates how pet ownership and attachment to one's pet relates to suicide risk. Many of the studies discussed above examined these variables in noncollege aged adults, and thus, it is important to expand this line of research into this population, as we do in this study. First, it was hypothesized that overall pet attachment would be negatively, indirectly associated with suicide risk through TB and PB (Hypothesis 1). This was hypothesized due to the earlier theorized positive associations of pet ownership on belongingness and burdensomeness. Second, it was hypothesized that the relationship between overall pet attachment and suicide risk would be conditioned on pet attachment avoidance and anxiety, with the relation of overall pet attachment and suicide risk weakening at higher levels of avoidance or anxiety and thus reducing the protective association of overall pet attachment (Hypothesis 2). This was based on findings that pet attachment anxiety is inversely related to psychological well-being and positively correlated to psychological distress and that those with an avoidant pet attachment style tend to have a lower attachment bond to their pets (Zilcha-Mano et al., 2011). Similarly, Hypothesis 3 predicted that pet attachment anxiety or avoidance, respectively, would be indirectly associated with suicide risk to increase risk via PB and TB.

Method

Results

Descriptive and bivariate correlations are shown in Table 1. Point estimates (midpoints) and 95% CI for paths are presented in the figures. Concerning Hypothesis 1, overall attachment was indirectly associated with suicide risk via PB (see Figure 1a; point estimate = −0.45, 95% CI = −0.91, −0.04), but not via TB (see Figure 1b; point estimate = −0.23, 95% CI = −0.65, 0.07). For Hypothesis 2, neither attachment anxiety (see Figure 2a; point estimate = −0.04, 95% CI = −0.11, 0.03) nor avoidance (see Figure 2b; point estimate = 0.06, 95% CI = −0.04, 0.15) conditioned the relationship between overall attachment and suicide risk. As a follow-up analysis, we investigated the link between overall attachment and suicide risk without any of the other variables via a linear regression (all assumptions met) and found that overall attachment was negatively associated with suicide risk, F(1, 152) = 5.00, β = −.70 (95% CI = −1.32, −0.08), p = .03. For Hypothesis 3, attachment avoidance was positively, indirectly associated with suicide risk via TB (see Figure 3a; point estimate = 0.03, 95% CI = 0.01, 0.06), but not via PB (see Figure 3b; point estimate = 0.02, 95% CI = −0.01, 0.05). PB was directly, positively associated with suicide risk. Attachment anxiety affected suicide risk via TB and directly (see Figure 4a; point estimate = 0.03, 95% CI = 0.01, 0.05). Attachment anxiety was indirectly, positively associated with suicide risk via PB (see Figure 4b; point estimate = 0.04, 95% CI = 0.02, 0.06).

Table 1 Descriptive statistics and bivariate correlations

Variables	1	2	3	4	5	6
Note. *p < .05; **p < .01.
1. Overall attachment	—
2. Anxious attachment	−0.36**	—
3. Avoidant attachment	0.61**	0.05	—
4. Perceived burdensomeness	−0.20*	0.34**	0.08	—
5. Thwarted belongingness	−0.13	0.34**	0.19*	0.71**	—
6. Suicide risk	−0.20*	0.29**	0.07	0.63**	0.46**	—
M	1.86	36.11	22.54	10.77	19.05	4.50
SD	0.57	12.27	7.06	5.51	8.79	2.34
Observed range	1–3.87	13–70	13–43	7–33	8–44	3–13

Table 1 Descriptive statistics and bivariate correlations

View as image HTML

Discussion

The effects of pet ownership on mental health and suicide risk are understudied. The current study examined the relationships between attachment to one's pet and suicide risk via TB and PB from the IPT (Van Orden et al., 2010). We hypothesized that overall pet attachment would negatively, indirectly affect suicide risk via TB and PB. Overall attachment was found to have no relation to suicide risk when examining the indirect associations via TB; however, TB was positively, directly associated with suicide risk. Increased overall attachment was associated with lower PB, whereas PB had a positive relationship to suicide risk. This suggests that when looking at overall attachment to one's pets, there may be a beneficial association on suicide risk via the PB pathway.

Next, we hypothesized that overall pet attachment would directly affect suicide risk to increase risk as conditional upon pet attachment anxiety and avoidance, by reducing the protective association of overall pet attachment. Attachment anxiety did not moderate the relationship between overall attachment and suicide risk but had a positive association with suicide risk. This suggests that attachment anxiety may be a risk factor despite not affecting the main relationship at hand. Attachment avoidance did not moderate the positive association between overall attachment and suicide risk. A follow-up analysis found an overall negative effect of overall attachment on suicide risk, with higher overall attachment being associated with lower risk.

We then examined the hypothesized indirect associations of attachment anxiety and avoidance on suicide risk via PB and TB. Attachment avoidance was associated with increased TB, which was associated with increased suicide risk, whereas no direct or indirect associations were found when PB was included. In the context of attachment theory, those with an avoidant attachment style are distrustful of an attachment partner's intentions and strive to be self-reliant (Bowlby, 1969), so it would be consistent that avoidance would predict one not feeling like they belong, even with their animal. Attachment anxiety's relationship with suicide risk was mediated by both TB and PB. Also, even when accounting for TB, attachment anxiety directly affected suicide risk. These results are in congruence with previous research that found attachment anxiety was associated with lower psychological well-being and greater psychological distress (Zilcha-Mano et al., 2011).

Overall, these preliminary results suggest that the IPT should be investigated within the context of human and animal relationships as overall attachment and attachment style to pets may affect a person and their suicide risk in different ways. For example, someone who overall is attached to their pet and has lower levels of attachment anxiety or avoidance may have neutral or positive associations. However, those who display higher levels of attachment avoidance or anxiety may not see the same neutral or beneficial associations and may, in some contexts, see a detrimental association. These preliminary results will need to be replicated and extended to assess any temporality or causality in these relationships as the current study is limited by its cross-sectional nature.

Future directions will need to expand into the predictive feasibility of pet attachment and suicide risk as this could tangibly assist clinicians. Psychologists in the United States have increasingly been pressured and engaged in practices such as “prescribing” emotional support animals and comfort animals to individuals who have mental health problems (Younggren et al., 2020). For clinicians who are trying to assess if a pet may be helpful or harmful for a client, they will want to know potential ramifications before a pet is brought on board. One limitation of this model is that we are measuring pet attachment to already existing relationships. Future research will need to examine if there are any feasible proxies to examine (e.g., would human relationship attachment styles be indicators of pet attachment styles?) or if the PAQ is useful when a participant is instructed to imagine how they may feel regarding a future pet. However, for clinicians who are attempting to determine whether an emotional support animal has a demonstrable beneficial effect on the owner and their specific psychiatric disability (Younggren et al., 2020), examining the attachment style to the pet may be useful.

The current study has limitations that should be considered when interpreting the findings. The INQ and SBQ-R have different time frames (recent vs. lifetime) that may affect results. The sample was recruited from a university population, and the majority of participants indicated they lived on campus (which is normally pet-free), the majority indicated they did not live with their pets all year, and the majority indicated that they were not the primary caregivers of their pets. This suggests that the participants may have limited exposure to their pets, and thus, their relationship with their pet may not be that prominent in their day-to-day life, limiting the potential effect of owning a pet on their risk. A community sample, who would be more likely to live with and be exposed to their pet daily, may have a more prominent relationship with their pets and may see stronger or different associations. If this is true, however, the presence of any associations for a sample who is not normally exposed to their pet suggests that the relationship one has with their pet may be an important one.

Overall, when examining the results of this study, it suggests that pet ownership may have a complex relationship to suicide risk, at least when viewing it through the lens of the IPT. As this is the first study to directly examine the relationships between pet ownership and associations with suicide risk via the TB and PB, these results need to be replicated to assess for robustness and fully explore the relationships between these constructs. Clinicians may believe the “pet effect” to be true, which may influence them to be agreeable to signing emotional support animal letters or encouraging their client to obtain a pet for companionship. It has already been argued that clinicians need to exhibit care and consideration is when deciding to encourage a client to obtain a pet for their mental health (Younggren et al., 2020), and our results suggest that the broad encouragement of clients to obtain pets may be too soon. These preliminary results also suggest that if a clinician is hoping to reduce a client's TB or PB via owning a pet, it may be best to seek other interventions (Joiner et al., 2009).

Electronic Supplementary Material