An exploration of the consequences of, and coping with loneliness in an ageing intellectual disability population

Introduction

Loneliness has been reported to affect nearly 50% of individuals with an intellectual disability (ID) (McCarron et al., 2011; Stancliffe et al., 2010). Loneliness in ageing populations has been associated with negative health (Leigh-Hunt et al., 2017), such as metabolic dysregulation (Shiovitz-Ezra & Parag, 2019), increased systolic blood pressure (Ong et al., 2012), and lifestyle consequences such as poor diets and smoking (Cacioppo & Patrick, 2008). To date, the consequences of loneliness and the coping strategies used by lonely older people with an ID have not been investigated.

Loneliness is the distressing feeling that accompanies discrepancies between one's desired and actual social relationships (Hawkley & Cacioppo, 2010). It is generally believed that the lonely tend not to seek help for their loneliness (Perlman & Peplau, 1998), withdrawing from others and using coping strategies that perpetuate their situation (Hawkley et al., 2008) such as self-blaming (Vanhalst et al., 2015) and lack of trust (Hensley et al., 2012).

Loneliness in older people has negative consequences for health and wellbeing, being associated with increased metabolic dysregulation (Shiovitz-Ezra & Parag, 2019) and increased systolic blood pressure (SBP) (Hawkley et al., 2010; Ong et al., 2012). Moreover, loneliness has been reported as altering a person's lifestyle and physiological reactions to stress (Hawkley & Cacioppo, 2007). In their social environment model of loneliness, Hawkley & Cacioppo (2007) argue that there are five pre-disease paths; health behaviours, exposure to stressful life events, coping style and support, physiology and recuperation, through which loneliness negatively influences a person's physiological resilience.

There is a growing body of evidence in older people to support each of the pre-disease pathways and their overall effect on physiological resilience. Hawkley & Cacioppo (2007) argue that unhealthy lifestyles contribute to poor health and early death in the first pathway health behaviours. Lonely people tend to be involved in more risky health behaviours (Shankar, 2017) and have been found to exercise less, smoke (Shankar et al., 2011), take in more fats and calories, and are more likely to have a higher body mass index (BMI) (Shiovitz-Ezra & Parag, 2019). It has been found that lonely people manage moods by eating, drinking and acting out sexually (Hawkley & Cacioppo, 2007), and they attend general practitioners surgeries and accident and emergency departments more frequently than those who are not lonely (Cleary, 2011).

In the second pathway, exposure to stressful life events, Hawkley & Cacioppo (2007) reported that the number of stressors experienced and the potency of those stressors are more prevalent in the lonely and diminish physiological resilience. Lonely people report being exposed to an increased number of stressful life events and (Hawkley et al., 2008) some argue that it is the accumulation of negative life events that lead to loneliness in older people (Jylhä, 2004). However, the association between life events and loneliness is not consistently reported, and others have found no association (Zebhauser et al., 2015). The effect of each stressor is particular to an individual's circumstances. For instance, work stress has been reported to have more of an effect on unmarried people's loneliness (Hawkley et al., 2008).

In the third path, coping style, lonely people are less likely to cope by seeking a confidant for support (Victor et al., 2008) and are more likely to regularly attend church (Hawkley & Cacioppo, 2007). It is notable that people with an ID lack support from friends (The Money, 2012), spouses or life partners.

Consistent with the fourth pathway, the physiology of chronic stress in older people signal the vulnerability to disease. Chronic loneliness leads to activation of the autonomic nervous system, leading to heart rate and blood pressure increases (Hawkley et al., 2010). However, the effects of loneliness on cardiovascular health have recently been questioned, with one study claiming loneliness does not affect SBP (Das, 2019). However, Das (2019) has been criticised for not considering the role of medications beyond baseline (Hawkley & Schumm, 2019). Other researchers have reported that lonely people record a different cardiovascular response than non-lonely people in specific conditions (Brown et al., 2019). There is no evidence about physiological responses to loneliness in people with an ID.

Finally, recuperation counteracts the forces that drain physiological reserves. Lonely people have less effective sleep (Coyle & Dugan, 2012), their sleep is more fragmented (Kurina et al., 2011) or altered (Leigh-Hunt et al., 2017), they take more time to go to sleep and have more night-time disturbances than non-lonely people (Cacioppo et al., 2000).

Perlman & Peplau (1998) claimed people had four mechanisms for coping with loneliness; sad passivity, active solitude, spending money and social contact. The evidence available tends to support the concept of sad passivity being the most common coping mechanism. People with an ID tend to be atypical of the general population in their health behaviours. It has already been reported that loneliness is not a predictor of healthcare utilisation in this population (McCallion et al., 2012). In general, people with an ID have healthier diets, smoke less, and drink less alcohol than the general population, but they complete very little vigorous physical exercise, and 66% are classified as overweight or obese (McCarron et al., 2014). Consideration of loneliness in people with ID must consider these different patterns of coping.

Methods

Results

Participants for this study had to self-report their loneliness on the loneliness scale in wave two of data collection (Figure 1). Table 1 represents the demographic breakdown of participants, comparing those who completed the scale to those who did not complete the scale. The average age of 56.16 (SD=8.578) was not significantly different for those who did not complete (mean=56.95, SD=9.875). This subpopulation had a higher percentage of females (59.3%) than the main population (56%), but there was no significant difference in the gender balance (χ²=2.671, p=0.102). There was an overrepresentation of those with mild and moderate disability in those who completed the loneliness scale compared to those who did not respond to the loneliness scale (χ²=179.190 p<0.001).

Figure 1. Flow diagram for participant inclusion in the loneliness research.

Cross-tabulations were calculated for each health variable against the dichotomised loneliness scale variable (Table 2). All cells had an expected value of more than five participants, allowing chi-squared analysis to be conducted. Analysis of the loneliness scale score found in path 3, the not lonely on the loneliness scale were more likely to confide in staff (χ²=6.625 p<0.05). In path 4, those who were lonely were more likely to have raised SBP (χ²=4.424, p<0.05). Analysis of path 5 revealed that those who were lonely were proportionally more likely to have difficulties falling asleep (χ²=9.176, p<0.01) and waking too early (χ²=6.539, p<0.05).

The mean loneliness scores were subject to ANCOVA, with gender and functional limitations being held constant (Table 3). Those who had difficulty falling asleep tended to be lonelier (mean = 6.000) than those who did not have difficulty falling asleep (mean = 5.150). In path 1, participants who did moderate activity had significantly higher loneliness scores (mean = 5.524) than those who did not (mean = 5.221, F=4.171, p<0.05). There were no significant results in Path 2. In path 3, coping, those who confided in staff (mean = 5.152) had significantly lower scores on the loneliness scale than those who did not (mean = 5.589, F=5.716, p<0.05).

Analysis of SBP in path 4 reveals that those who had high SBP scores had higher mean loneliness scores (mean=5.570) than those who were not lonely (mean=5.137), but differences were not significant (F=10.207, p>0.05). In path 5, those who had trouble falling asleep had significantly higher loneliness scores than those that did not have trouble falling asleep (F=13.907, p<0.01). People who reported having disturbed sleep (mean=5.456) had significantly higher loneliness scores than those who did not (mean=5.178, F=3.930, p<0.05). When the sleep scale individual items were combined, those who recorded higher scores had significantly higher scores on the loneliness scale (F=10.284, p<0.01).

The binary logistic regression results (Table 4) indicate the strength of the loneliness variable's influence, with gender and functional limitations, held constant. The table details the adjusted odds ratio (AOR), the B statistic (the original scale coefficient), the standard error of the B value (created with bootstrap analysis), the bootstrap created confidence intervals, and the Naglekerk R², for variables where loneliness was a significant predictor.

In path 3, those who were lonely were less than half as likely to confide in staff as the not lonely participants (AOR=0.464, B=-0.769, SE=0.292, 95% CI=-1.333, -0.237). Loneliness accounted for 2.8% of the confiding in staff variance (Naglekerke R²=.028).

Analysis of SBP in path 4 reveals that being categorised as lonely was a significant predictor of having raised SBP. The lonely were twice as likely to have raised blood pressure, with loneliness accounting for 2.6% of the SBP variance (AOR=2.051, B=0.718, SE=0.355, 95% CI =0.019, 1.512, Nagelkerke R²=0.026). In path 5, being lonely was a significant predictor of having trouble falling asleep (AOR=2.543, B=0.933, SE=0.342, 95% CI=0.262, 1.631). Loneliness accounted for 4.4% of the variance of the difficulty in falling asleep variable (Naglekerke R²=0.044). Loneliness was also a significant predictor of waking too early, accounting for 3.2% of the variance (AOR=2.225, B=0.800, SE=0.348, 95% CI =0.103, 1.480, Naglekerke R²=0.032). When the sleep scale individual items were combined, being lonely was a significant predictor of having sleep difficulties, accounting for 4.7% of the sleep scale variance (AOR=2.526, B=0.927, SE=0.315, 95% CI =0.307, 1.574, Nagelkerke R²=0.047).

Discussion

This research offers the first evidence of how older people with an ID react to and cope with loneliness. The results indicate that older people with an ID reacted to loneliness with sleeping difficulties, raised systolic blood pressure, and were less likely to confide in staff/caregivers. The results also found that people with an ID coped with loneliness through increased moderate physical activity.

This study supports previous research in the general population that indicated associations between loneliness and sleeping difficulties (Kurina et al., 2011) and loneliness and systolic blood pressure (Hawkley & Cacioppo, 2007). These findings also extend previous knowledge (Victor et al., 2008), indicating the importance of whom people confide in over merely confiding as an act. Finally, for people ageing with an ID, the results disagree with the general ageing population's findings that lonely people are more likely to have worse health behaviours than non-lonely (Hawkley & Cacioppo, 2007; Lauder et al. 2006). There was no association found between health behaviours and loneliness, and only those who did moderate activity were found more likely to score higher on the loneliness scale.

The relationship found in this study between sleep and loneliness was consistent with research from the wider population. Loneliness often affects sleep because of the unconscious scanning for social threats caused by hypervigilance (Hawkley & Cacioppo, 2010). When people have had their sleeping checked by the use of electronic devices such as nightcaps, researchers have found that lonelier participants have more disrupted sleep (Kurina et al., 2011), they take longer to fall asleep and have poorer sleep quality (Cacioppo et al., 2000). Sleep is an important precipitator of health problems as sleep counteracts the forces that drain the body, and lower quality sleep does not allow the restorative processes to operate. (Hawkley & Cacioppo, 2007)

Hypervigilance is also reported to reduce sleep quality and be associated with increased vascular resistance (Cacioppo & Cacioppo, 2014), which leads to increased SBP (Hawkley & Cacioppo, 2007). Disrupted sleep has also been postulated as a causal route for raised SBP (Bonnet & Arand, 2003).

Among people ageing with an ID, the lonely were twice as likely to have raised SBP. This evidence supports research from the wider population where it has been found that for every standard deviation rise in loneliness, SBP increased significantly (Ong et al., 2012). Although establishing a causal relationship remains for future research, a need is confirmed for attention to highlight blood pressure concerns among people with ID who report feelings of loneliness.

Hypervigilance has also been reported to cause people to be wary of others (Cacioppo & Cacioppo, 2014). Here, whom a person confided in influenced their chances of becoming lonely and additionally, being lonely influenced whom people confided in, creating a cycle of protection or harm. More specifically, those who confided in staff were less likely to be lonely, and those who were lonely were more than one and a half times more likely to confide in others. It is possible that people who confided in staff were confiding in someone who could make a difference to underlying issues, whereas others may not affect a person's circumstances directly.

There is no evidence that loneliness influenced the health behaviours of this population. This finding is divergent to findings in the general population, where the lonely have been described as utilising a coping state of sad passivity, where they eat worse diets, exercise less, and are generally more lethargic than the non-lonely (Hawkley et al., 2008). In this study, those who did moderate exercise had higher loneliness scores than those who did not.

Limitations

In analysing the data as reactions to and coping with loneliness, this research may imply causality. To show causality, three criteria need to be met; covariation, temporal ordering and elimination of competing theories (Hayes, 2013). This analysis cannot prove causality since it is not an experiment controlling the above conditions; however, steps have been taken in the analysis to approximate the three criteria. Covariation was dealt with through the type of analysis conducted that showed the variables did have covariation. Two competing theories were accounted for through the utilisation of the co-variates gender and functional limitations. There are more than two possibilities for competing causes of the variables health paths, and these need to be considered with further investigation of the data. This analysis is the first work to look at how older people with an intellectual disability react to and cope with loneliness, and some results confirmed findings from the general population, further suggesting validity. Additional investigation assuming findings hold will add to the validity of the findings.

The data used here were taken at a single time point and, therefore, do not have the support of longitudinal analysis. Unfortunately, the full loneliness scale was only available from wave two of data in the IDS-TILDA study. Further analysis will offer a more detailed insight as further waves of data also incorporating the scale become available.

The data collection techniques employed in this study meant that only people who could self-report their feelings are represented. This limitation excludes any understanding of loneliness in those who have difficulty communicating. In future research, alternative research methodologies must be employed to help those with communication difficulties express their feelings of loneliness.

Conclusion

This study was the first to explore reactions to and coping with loneliness in an ageing population of people with an ID. This study used a social environment model of loneliness described by Hawkley & Cacioppo (2007) that analysed the effects of loneliness on physiological resilience through five pre-disease paths. The results add support to path 4 (health effects) and path 5 (recuperation). They extend parts of path 3 (coping) but find little supporting evidence for paths 1 (health behaviours) and path 2 (exposure to stress). The analysis undertaken supports the concept of hypervigilance and suggests that it is experienced in this population, leading to sleep disruption, raised SBP and wariness of other people. The analysis does not support any hypothesised coping mechanisms (Perlman & Peplau, 1998), finding no differences in the health behaviours between the lonely and the non-lonely. Future research needs to investigate the longitudinal relationships of loneliness and health in this ageing ID population.

Data availability