Are volunteering and caregiving associated with suicide risk? A Census-based longitudinal study

There is increasing societal interest in health outcomes, and especially the mental health, associated with prosocial activities such as volunteering and caregiving. However, although helping others is the common defining feature of these activities, there are significant differences between them both in terms of the characteristics of the participants and the attributes of the activities, such that they might be expected to have very different effects on mental health and suicide risk.

On one hand the ageing population, and a growing tendency to care for incapacitated people in their own homes, has placed an increasing reliance on ‘invisible health systems’ - the informal and unpaid help and assistance provided by family, friends and neighbours [1, 2]. A consensus is forming that caregiving is associated with poorer mental health: caregivers tend to demonstrate higher levels of stress [3‐6] and a higher prevalence and incidence of depression, associations that may be more marked amongst people caring for those with dementia [7‐12]. A dose-response relationship is evident with more intensive levels of caregiving associated with a higher likelihood of scoring above the clinical threshold for the Clinical Interview Schedule [13]. Caregivers may also be exposed to a wider range of stressors including exposure to domestic violence, financial difficulties or stressful life events and in some UK-based studies, were more than twice as likely to report suicidal thoughts and wishing they were dead in the previous week compared to non-carers [14]. Similar findings have been reported from Australia where O’Dwyer et al found that 7.1% of middle-aged female carers felt that life was not worth living and that this was very strongly related to the presence of depression [15]. Furthermore, 26% of caregivers looking after family members with dementia had contemplated suicide more than once in the previous year, and almost 30% said they were likely to attempt suicide in the future [16].

Volunteering, on the other hand, is increasingly advocated as a win-win activity that brings tangible benefit to the volunteer as well as the wider society. A recent meta-analysis [17] and narrative review [18] concluded that there is now considerable evidence that volunteering is associated with less depression, more positive affect and happiness, and greater life satisfaction, though it is also acknowledged that most of the included studies were cross-sectional. There remains however, some uncertainty as to whether the benefits of volunteering are limited to older people, as most of the subjects were older, or diminish at higher levels of volunteering commitment, possibly due to role exhaustion [19]. The likelihood of a causative protective relationship has been increased by a strong theoretical underpinning, initially encapsulated by Wilson and Musick [20] and later echoed by Li and Ferraro [21], which lists some of the bio-psycho-social mechanisms by which volunteering should lead to better mental health: it requires participation which provides a sense of purpose and value [22, 23]; it increases social networks and interactions, strengthening existing friendships and establishing new ones [24‐27]; and finally, providing help to others can be a self-validating experience that enhances personal efficacy [28]. Another factor that connects volunteering and good mental health is the close correlation between volunteering and religiosity [29, 30] and one that differentiates it from caregiving is that it is more socially valued and publicly recognised [31]. These mechanisms may explain why volunteering is seen to act as a buffer against stress, work loss or bereavement [32‐35].

Collectively these studies suggest that caregivers and volunteers should have different levels of mental ill-health and attendant suicide risk. However, many of the findings related to caregiving were based on relatively small studies or of select sub-populations that may not be representative of the wider community of caregivers. Furthermore, an increasing number of studies emphasise the many positive psychosocial aspects to caregiving [36, 37] and recent studies suggest that, despite the recognised stresses and other difficulties of caregiving, it may be associated with reduced rather than raised risks of all-cause mortality [38‐43]. It is therefore possible that this relationship also holds for caregiving and suicide mortality.

This study aims to (i) compare the prevalence of poor mental health amongst volunteers and caregivers after adjustment for demographic and socio-economic factors; (ii) measure the risk of suicide amongst caregivers and volunteers, controlling for baseline health status and possible health selection effects - for example the degree of physical capability required to initiate and maintain a significant caregiving role [44]; and (iii) determine if these prosocial activities reduce suicide risk for those with poor mental health.

The Northern Ireland Mortality Study (NIMS) is a record-linkage study comprising the Census returns for the whole enumerated population and subsequently registered deaths. Details of NIMS and how the linkage was conducted have been previously reported [45]. For this study, we defined the population-at-risk as those enumerated in the Northern Ireland Census (March 2011), aged 25–74 years and not living in institutional care, with mortality follow-up from the Census until December 2014 (45 months). All personal characteristics were drawn from the Census and selected on the basis of their known association with suicide risk, including age (in 10-year bands); gender; and marital status (married, never married, and – as a single group - those widowed, separated or divorced). Due to the ethnic homogeneity of Northern Ireland we summarised ethnicity as white and non-white. Regarding household variables, households were summarised as single person or not, on the basis of number of residents. Socioeconomic status was assessed using two parameters: (i) household car availability (two or more cars, one only, no household access), and (ii) a combination of housing tenure and capital value of the property. Data on housing tenure were drawn from the Census (grouped as owner occupiers, private renters or social renters), while capital value had been previously derived from a 2005 exercise by central government which aimed to determine the level of local payable tax of each household. These two variables were combined, producing an eight-fold classification of housing tenure/ property value: private renting; social renting; and, for owner-occupiers, five categories ranging from less than £75,000 to over £200,000 (Table 1), with owners of unvalued homes treated separately.

Of the 1,018,000 individuals included for analysis: 12.9% (130,816) were providing care and 10.9% (110,467) were volunteering. There was considerable overlap in these activities with 42,099 people engaged in both volunteering and caregiving, representing 32.2% of caregivers and 38.1% of volunteers.

Compared to those not engaged in any prosocial activity, non-volunteer caregivers were more likely to be older, female, married and slightly more affluent in terms of house value and car availability, though more likely to be in part-time employment and much more likely to be a homemaker. They were also less likely to report chronic mental health or mobility problems. Those involved in volunteering only were more similar to those not engaged in these prosocial activities in terms of age, sex, and household composition, though were more likely to belong to a more conservative Christian denomination. The socio-economic gradients are more evident amongst volunteers than caregivers and this is particularly marked for educational attainment. Volunteers had a lower prevalence of all three chronic health problems. Those engaged in both activities were more like caregivers in terms of age, sex, marital status and household composition but more similar to volunteers in terms of religious affiliation, socio-economic status and educational attainment.

This large representative study shows that prevalence of mental ill-health is related to the type of prosocial activity – while more intensive caregivers record worse mental health than non-caregivers, volunteers record better mental health. Additionally, when compared to those who don’t engage in either prosocial activity, both caregivers and (separately) volunteers have reduced suicide risks, though in fully-adjusted models these associations lose significance. However, volunteers with caregiving responsibility maintain a significantly reduced suicide risk even after full adjustment. In the analysis of suicide stratified by absence/presence of mental health problems at baseline, those caregivers and volunteers with no mental health problems both recorded lower likelihoods of suicide mortality than their non-prosocial peers. However, for those with mental health problems a similar analysis showed no suicide mortality differences when compared with their non-prosocial peers. Thus, while volunteering is associated with lower likelihood of poor mental health [17, 18], it does not appear to be associated with reduced suicide risk for those with poor mental health. This is perhaps not surprising given the strength of the relationship between poor mental health and suicide [50, 51].

To our knowledge this is the first study to examine suicide risk amongst volunteers and caregivers and these findings provide some reassurance that despite the higher prevalence and incidence of depression and common mental disorders amongst caregivers [7‐13] this does not translate into higher suicide risk. The majority of caregivers do not have poor mental health and their suicide risk is about one-third that of non-caregivers, and amongst the minority with poor mental health there is no evidence of an increased suicide risk. This may be surprising given reports suggesting increased suicide ideation [14‐16] though, as Gunnell et al [52] point out, the epidemiology of suicide ideation and suicide completion is different and, importantly, the relationship can be modified, most obviously by age and sex. It appears that while a disproportionate number of caregivers suffer from increased suicide ideation, this does not generally translate into higher suicide risk. It may be that the benefits of caregiving outweigh the more negative aspects of the role. Altruistic behaviour has also been suggested to improve wellbeing and reduce mortality [39], while the increased sense of purpose which has been associated with caregiving has been known to act as a buffer against mortality risk across different adult age groups [53, 54], irrespective of other markers of psychological or affective well-being [55]). Although the psychological burden may be high, this sense of purpose may protect against completed suicide due to a sense of responsibility on the part of the caregiver.

This study also found a significant overlap in prosocial activities with approximately one quarter of volunteers and caregivers engaged in both activities and we would agree with other researchers that future studies should consider them simultaneously [56, 57]. Indeed, as there is some indication that people combining caregiving and volunteering have the lowest risk this suggests a synergism of actions and mechanisms [58], perhaps the increased sense of bonding associated with caregiving complements the outward-looking social engagement associated with volunteering. However, we are also aware that too many roles can offset the benefits and may result in worse health outcomes [19, 59].

There is a consensus that caring is associated with an increased risk of stress and poorer mental health. However, as researchers such as Brown [38] point out, the interpretation of this association is less clear and may be a consequence of witnessing the suffering or deterioration of a spouse or close family member with a chronic disabling illness, or of anticipatory bereavement rather than the effects of caring per se. Others, such as Roth [60], have pointed to results from national surveys showing high proportion of carers reporting satisfaction with their role and how this might be linked to forging stronger bonds with the care recipient [30‐32]. However, some caring could be associated with stress and poor mental health and this is most likely when demands exceed available psychological, social or financial resources [34]. Personality traits might also be linked to caregiver burden, with neuroticism being negatively associated with mental health-related quality of life while extraversion has an opposite effect [61]. Volunteering has also been strongly related with extraversion [62], suggesting that personality traits could affect the relationship between prosocial activities and suicide risk. It is of note that a relatively recent twin study [63] has suggested that the interplay between caring and distress might also be due to a vulnerability which is shaped, at least in part, by genetic and early life factors.

This study has significant strengths and limitations that should be acknowledged. To our knowledge it is one of the largest and most representative studies of suicide risk for caregivers and volunteers. Caregiving and volunteering status were defined at baseline and the linked administrative data ensured accurate assessment of cause of death extracted from Official Statistics. However, we acknowledge that the data relating to prosocial activity is limited: there is no indication of the relationship between carer and the cared-for person or of the type of care provided. Earlier mortality studies have emphasised the role of carer stress, but here only the number of hours spent caring and baseline mental and emotional health status were available as a proxy for stress. The classification of volunteer status included no indication of its duration, nature or intensity - limiting our understanding of its interrelationship with more intense levels of caregiving. Furthermore, it was not possible to differentiate between volunteering undertaken as part of an organisation or in a private capacity. However, this seemingly naive instrument does identify the same demographic, socio-economic features and health and mortality relationships that characterise volunteers in other studies. Although the protective effect of engaging in both caregiving and volunteering activities in relation to suicide risk was highlighted, future researchers might wish to explore the potential influence of competing mortality risk on these effects. Additionally, despite the fact that this observational study shows that caregiving is associated with a reduced risk of suicide, it cannot definitively prove that caregiving reduces suicide risk. While a concern (not easily dismissed) is that caregivers and volunteers are healthier than those who are neither [44], after adjustment for health status, there is very little change in hazard ratios, suggestive that health selection was not a major contributor to suicide risk, even among those with poor mental health. Finally, it is worth considering that our sample is of predominantly white ethnicity, with variation in the associations of interest to be expected in people from other cultures: for example, in Asian countries where government support for informal carers is limited [64, 65] one study found that informal caregiving does not appear to be linked with all-cause mortality [66]; though this may not pertain when focusing on suicide risk.

One major point raised by this study relates to the additional evidence it provides that current putative health consequences associated with informal caregiving are too pessimistic. Two recent overviews, the first by Brown and Brown [67] the second by Roth et al [60] have argued for a more balanced perspective and state that ‘Policy reports, media portrayals, and many research reports commonly present an overly dire picture of the health risks associated with caring and largely ignore alternative positive findings’. There are growing concerns that this negative view of caregiving, with an over-emphasis on burden and stress, may deter some from undertaking a caregiving role. Therefore, while it will be increasingly important to identify how best to assess the physical and mental effects of the caring role so as to provide the most appropriate support [13, 14], it is important to recognise both the benefits accruing from caregiving and reiterate the more general observation that caregivers as a general group have reduced mortality risk compared to their non-caregiving peers.

These analyses suggest benefits associated with engagement in pro-social activity (caregiving and/or volunteering): for caregivers the evidence proposes that, while there were increased risks of reporting mental ill-health, there was no evidence of increased suicide risk; for volunteers the risks of both reporting mental ill-health and suicide were reduced, though not significantly so in the more fully adjusted mortality models; and for persons engaging in both suicide risk was significantly reduced. However, while we stress these findings be approached with caution, it is important to further research the relationship between pro-social activity (especially caregiving) and health outcomes – if only because its societal role is predicted to grow.