Social isolation is a well-known risk factor for incident mental and physical health problems, including depression [1], distressing somatic symptoms (eg, pain and fatigue) [2], substance use [3], cognitive impairment [4], heart disease [5], and mortality [6]. In fact, meta-analytic work has shown that having social relationships is associated with reduced mortality risk comparable with that of well-known health factors such as not smoking (vs smoking), abstaining from alcohol (vs binge drinking), engaging in rigorous physical activity (vs low-to-no physical activity), and being lean (vs obese) [7]. Notably, however, the relationship between social activity and health is also reciprocal, such that chronic illnesses and diminished health can result in physical (eg, disability), emotional (eg, stigma and stress), and financial barriers to engagement in social activities, leading to social isolation [8].

Older persons living with HIV (PLWH) are a rapidly growing population of individuals who are at a significantly increased risk for social isolation compared with older persons without HIV [9]. In addition to social barriers that older HIV-uninfected adults often face (eg, retirement, hearing loss, and mobility impairment) [10], older PLWH face barriers to social support that are specifically related to HIV infection, including HIV-related stigma and discrimination [11], nondisclosure of HIV status [12], and loss of their social support network because of AIDS-related deaths [13]. Furthermore, these individuals also often have several medical comorbidities (eg, chronic diarrhea, lipodystrophy, cardiovascular disease, cancers, kidney and liver diseases, and osteoporosis) that can cause significant emotional distress, fatigue, and pain, leading to further social withdrawal [14-16]. Fatigue and pain are common somatic symptoms of HIV infection and can be critical barriers to engagement in social activity [17,18]. Results from a recent study conducted by Moore et al support findings of decreased social activity among older PLWH such that participants reported spending a majority of their sampled time at home, alone, and engaged in passive leisure activities (eg, watching television) [19]. Furthermore, older PLWH are more likely to live alone compared with both younger PLWH and their seronegative counterparts [9,12], greatly increasing the risk for social isolation.

In addition to the mental and physical health consequences of social isolation in the general population, older PLWH are also particularly vulnerable to worsening HIV disease progression without adequate social support [20]. Possible mechanisms for this include increased depression and reduced psychological well-being in older PLWH who are socially isolated [21], which in turn can contribute to decreased immune function and nonadherence to antiretroviral therapy medication [22,23]. Despite the increased risk for, and long-term detrimental consequences of, social isolation among older PLWH, there is a lack of research on the real-time temporal associations of social activity and mood- and health-related factors in this population.

Ecological momentary assessment (EMA) is a data collection method that has potential to help fill a gap in research on social activity among older PLWH. EMA is a highly feasible, valid, and reliable way to obtain real-world, real-time evaluations of various behaviors (eg, social activity) and experiences (eg, mood, fatigue, and pain) in a range of clinical and nonclinical samples, including older adults [19,24,25]. Thus, EMA can allow us to examine the temporal dynamics of social activity, in particular whether social activity predicts mood- and health-related factors (ie, fatigue and pain) or vice versa. Previous studies suggest that these relationships are often reciprocal and cyclical—ie, levels of social activity influence mental and physical health, and mental and physical health simultaneously influence levels of social activity, creating a potential downward spiral of decreasing social activity and worsening health outcomes [26]; however, such real-time relationships have yet to be examined among older PLWH.

Therefore, this study aims to use EMA to examine real-time relationships between social activity and mood, fatigue, and pain among older PLWH. Of note, the construct of fatigue in this study is represented by the reported current level of tiredness, consistent with geriatric literature [27]. First, we explored relationships between HIV disease characteristics and social activity (ie, proportion of time alone and number of social interactions per day). Next, we examined concurrent (ie, using responses within the same EMA survey) and lagged (ie, using responses from one survey to predict responses on the next survey) relationships between social activity (ie, alone vs not alone) and sad mood, happy mood, stress, fatigue, and pain level. Regarding concurrent relationships, we hypothesized that being alone would be related to lower ratings of happy mood and higher ratings of sad mood, stress, fatigue, and pain level. Regarding lagged relationships, we hypothesized that being alone on one survey would predict lower ratings of happy mood and higher ratings of sad mood, stress, fatigue, and pain level on the next survey; similarly, we also hypothesized that lower ratings of happy mood and higher ratings of sad mood, stress, fatigue, and pain level on one survey would predict greater likelihood of being alone on the next survey.

Examining real-time, immediate predictors of social activity is an important and novel approach to further our understanding of social experiences of older PLWH. Specifically, our findings provide a framework for understanding social activity in the context of mood, fatigue, and pain among older PLWH. This study found being alone was associated with lower concurrent happiness and later-day fatigue, and higher pain levels predicted greater likelihood of being alone earlier in the day and lower likelihood of being alone later in the day (compared with lower pain levels). Simply stated, although being with others was related to increases in fatigue and pain throughout the day, people were happier when they were with others than when they were alone.

The importance of the association between more social activity and greater happiness warrants a thorough discussion. This finding is critical for our understanding of psychological well-being of older PLWH. Although seemingly contradictory, our finding that participants were happier when they were with others (compared with when they were alone) even though pain and fatigue increased is supported by neurobiological and behavioral medicine literature. Neurobiologically, positive social interactions induce activity in the mesocorticolimbic (ie, reward) system and are experienced as pleasurable [34]; thus, social activity is likely reinforcing beyond potentially limiting factors such as pain or fatigue. Our finding is also consistent with the “stress buffering” theoretical model such that social relationships may “buffer” the adverse effects of a stressor (eg, pain) on health outcomes (eg, mood) [35]. Furthermore, happiness has been associated with a variety of positive health and lifestyle factors among older adults, including physical activity, competence, functional capacity, disease management, quality of social network, and general life satisfaction [36,37]. These positive health and lifestyle factors are critical for maintaining optimal quality of life among older adults, especially those living with a chronic disease such as HIV [38]. Thus, our findings support further examination of the benefits of social activity on real-time happiness and psychological well-being, and whether increased social activity may have longer-term effects on happiness. On the basis of the results of this study and existing literature on associations of overall health and happiness, the immediate positive influence of increased social activity on happiness has potential to improve many aspects of an older adult’s quality of life.

Although there were no significant associations between social activity and concurrent sadness or stress, this is likely because our participants reported overall low levels of sadness and stress with little variability compared with reported levels of happiness. Possible interpretations for this finding include the following: (1) our sample may not be very representative of older PLWH among whom depressive symptoms and stress are common, (2) our measures of sadness and stress may not be sensitive enough for this population, or (3) our measure of positive affect (happiness) may be a more sensitive indicator of mood and well-being in older PLWH with minimal-to-no depressive symptoms (as indicated by baseline BDI-II scores). Future research is needed to understand the clinical relevance of fluctuations in EMA-measured happiness among older PLWH without clinically elevated depressive symptoms or mood disorders, and whether consistently low happiness ratings in the context of decreased social activity or social isolation may predict feelings of sadness, loneliness, or onset of a depressive episode.

In contrast to our finding regarding increased happiness with social activity, results also showed that spending time with others predicted being more fatigued on the next survey. The relationship between social activity and fatigue is important to consider in the context of the association between social activity and happiness, and provides preliminary evidence to support potential clinical interventions promoting social activity [39-41] while simultaneously improving our understanding of strategies that help older PLWH manage fatigue in an effort to improve happiness and psychological well-being in this population. Notably, lagged analyses also elucidated an interaction effect between pain and time of day on later social activity. Participants reporting higher pain levels were more likely to be alone in the morning and less likely to be alone in the evening and night compared with those reporting lower pain levels. Although we are unable to determine reasons for this relationship with our available data, results show that participants in our sample of older PLWH were engaging in evening social activities despite experiencing high levels of pain. Further examination of this pattern is needed to determine whether this may be evidence of the ability to utilize effective coping techniques (eg, social support) when needed among older PLWH.

As previously reported by Moore et al [19], our sample of PLWH aged 50 years and older reported being alone for a majority of participants’ typical waking hours. This is considerably higher than results reported by the national American Time Use Survey in which adults aged 65 years and older reported spending 23% to 42% of waking, nonwork time alone [42]. Although our results may be influenced by the high rate of unemployment in our sample, our results are consistent with the previous research showing that older PLWH are at an increased risk for reduced social activity [8-10]. Future research on social activity must not only characterize levels of social activity but must also focus on understanding and disentangling different reasons for decreased social activity within different clinical populations, with the ultimate goal of developing targeted interventions.

Results also showed that HIV disease characteristics (ie, current and nadir CD4 counts, detectable plasma viral load, estimated duration of HIV, antiretroviral therapy status, and history of AIDS) were unrelated to the proportion of time spent alone and the average number of reported social interactions per day. Although previous research links social isolation to accelerated HIV disease progression [20], this is often in the context of perceived loneliness and social support. In contrast, the current EMA social activity data capture a more objective level of overall social activity. Findings from this study warrant further examination of the association between HIV disease progression and levels of social activity. Furthermore, participants who reported more social support at baseline (via the MOS Social Support survey) demonstrated a lower proportion of time alone and a higher number of daily social interactions over the 1-week EMA period. This finding also supports the convergent validity of our EMA social activity measures, and suggests that participants’ perceptions of support received from their social networks may be accurately related to their average levels of social activity. These relationships between social support and EMA-assessed social activity were expected, as social support often relates to the number of social contacts and frequency of social interactions across many clinical and nonclinical populations [43]. Finally, we found that females reported spending less time alone compared with males. Because we cannot make any definitive conclusions from a sample that only includes 3 women, further exploration of possible sex differences in social activity and health outcomes is needed among older PLWH in future studies.

This study is strengthened by its use of real-time data; however, there are several limitations. First, the relatively small sample size limits generalizability and statistical power to detect associations between social activity and baseline, laboratory-measured clinical characteristics. Furthermore, our sample was relatively healthy and had little-to-no HIV disease burden, which may not be widely representative of the general population of older PLWH in the United States. Also of note, the majority of participants in our sample were unemployed. Because employment is an activity that often engenders social activity (eg, interactions with coworkers and increases financial resources), our results may not be generalized to older PLWH who are employed; however, our sample also represents a real-world sample of older PLWH among whom unemployment is very common [44]. Future studies would benefit from further understanding associations between reasons for unemployment and levels of social activity. Next, because the data came from a small pilot study with the aim of assessing overall daily functioning in older PLWH, there were no data collected on other specific factors related to social isolation, such as perceived loneliness, reasons for being alone (eg, preference vs access to social activities), or structural social factors (eg, marital or relationship status). Relatedly, future research may benefit from understanding the real-time impact of pet ownership and time spent with pets on mood and physical health. In addition, the EMA item related to being alone or with others may not be the best representation of social isolation because it only allows us to capture participants’ social activity at the exact moment of the survey administration—that is, if participants were alone when completing the survey but were with others right up until the survey completion time, then they will report being alone although they may not be socially isolated. Future studies may consider using mobile sensing technologies and global positioning system data to better characterize social activity [45]. Assessing real-time associations between social activity and markers of mental and physical health, as we have done in this report, also lays the groundwork for developing potential real-time interventions for improving happiness, psychological well-being, and overall quality of life.

Overall, this study provides preliminary evidence for the real-time associations between social activity and mood, fatigue, and pain in older PLWH. Our findings showed that spending time with others was associated with increased happiness. The results of this study warrant additional research on the potential benefits of social activity-based interventions for improving well-being among older PLWH despite presence of fatigue or pain. Furthermore, with mounting evidence showing that older PLWH spend a great deal of time alone, future research must continue to characterize social activity as well as explore causes for and consequences of limited social activity among this population.