Social Support and Social Networks in COPD: A Scoping Review

Abstract

A scoping review was conducted to determine the size and nature of the evidence describing associations between social support and networks on health, management and clinical outcomes amongst patients with COPD. Searches of PubMed, PsychInfo and CINAHL were undertaken for the period 1966–December 2013. A descriptive synthesis of the main findings was undertaken to demonstrate where there is current evidence for associations between social support, networks and health outcomes, and where further research is needed. The search yielded 318 papers of which 287 were excluded after applying selection criteria. Two areas emerged in which there was consistent evidence of benefit of social support; namely mental health and self-efficacy. There was inconsistent evidence for a relationship between perceived social support and quality of life, physical functioning and self-rated health. Hospital readmission was not associated with level of perceived social support. Only a small number of studies (3 articles) have reported on the social network of individuals with COPD. There remains a need to identify the factors that promote and enable social support. In particular, there is a need to further understand the characteristics of social networks within the broader social structural conditions in which COPD patients live and manage their illness.

Keywords:

• COPD
• psychosocial factors
• review
• social support
• social networks

Introduction

Chronic Obstructive Pulmonary Disease (COPD) can result in significant and disabling psychological distress and social isolation, for both males and females, and younger as well as older persons (1). The main symptoms of COPD are dyspnoea, fatigue and cough, which can have significant impacts on activities of daily living and quality of life (2). Significant advances have been made in the past decade toward understanding the impact of COPD on psychological distress and mental health and the role of these factors in management and outcomes of COPD (3, 4); but no reviews have been written to consolidate our understanding of the impact of COPD on social relationships, and how social relationships, including social support and social networks, impact management and outcomes in COPD.

Over the past 30 years there has been increasing empirical evidence supporting the important role of social relationships in mediating health outcomes. For example a recent meta-analysis of all-cause mortality found that individuals with adequate social relationships have a 50% greater likelihood of survival compared to those with poor or insufficient social relationships (5). This effect is comparable with other well-established risk factors for mortality such as smoking and alcohol consumption and exceeds the influence of other risk factors such as physical inactivity and obesity (5). Studies investigating disease outcomes other than mortality have shown a beneficial effect of perceived social support for conditions such as coronary heart disease in terms of both development and prognosis (6) and that larger social networks are associated with positive health outcomes and decreased incidence of cardiovascular disease (7).

Several theoretical and conceptual models have been proposed to explain processes through which social factors may influence health. The first of these emerged in the 1980s following social research published in the 1970s demonstrating associations between social support and disease aetiology and mortality. One influential model known as the “stress buffering and main effects models” proposed by Cohen and colleagues envisions social support ‘buffering’ individual well-being from the potentially pathogenic influence of stressful events (8, 9). A second theory known as the main-effects model, proposes that social resources have beneficial effects irrespective of whether a person is under stress or not. This theory underlies explanations of results from recent meta-analyses that determined an overall effect size corresponding with a 50% increase in odds of survival as a function of social relationships (5).

However, Cohen's model (and other models that emphasise the role of social support) doesn't take into consideration the importance of characteristics of social networks and the macro social environment that social support occurs within. These “upstream” social factors form part of a broader conceptualisation of how social relationships impact upon health proposed by Berkman and Glass (10). In their model, Berkman and Glass envision a cascading causal process whereby the macro-social context conditions the structure of social networks. The social network structure and function influence social and interpersonal behaviours including social support, which in turn influence more proximate pathways to health status. These proximate pathways included 1) direct physiological stress responses; 2) psychological states and traits; and 3) health damaging behaviours such as tobacco consumption (11).

COPD has long been known to impact negatively upon social roles and place individuals at risk of social isolation, which may diminish confidence in their ability to self-manage their illness (12). Disler et al.’s (12) review highlights the social influences impacting self-management of COPD including social isolation, loss of social role and social support and how these interact with physical, psychological, existential and other social determinants to influence self-care behaviour (12). However, despite an increasing body of evidence describing associations between social relationships and important outcomes in COPD there have not been any published reviews assimilating this literature.

We began by asking: what do we know about the social relationships of patients with COPD and what are the health impacts of these relationships? An initial search of the literature did not identify any review articles addressing the issue of social relationships and health outcomes amongst patients with COPD. Consequently a scoping review was conducted in order to identify gaps in the literature and make recommendations for future primary research. Scoping studies differ from standard systematic reviews in that they do not attempt to synthesise the evidence but are used to determine the size and nature of the evidence base for a particular topic area. This approach is particularly useful when the aim is to identify gaps in the literature and make recommendations for future primary research (13).

The aim of this review was to determine the size and nature of the evidence describing associations between perceived social support and social networks on health, management and clinical outcomes amongst patients with COPD.

Methods

A preliminary search of the literature was first conducted using broad terms such as “social” AND “COPD.” From the articles identified it was apparent that most studies assessed aspects of social support and to a lesser extent, social networks. The search also revealed that both qualitative and quantitative methods were used. Guided by both the frameworks we previously identified that linked social relationships to health outcomes and this first pass search of the literature, it was decided that a more focussed search was the most useful way forward, focussing on quantitative assessment of social support and social networks.

Focussed searches of PubMed, PsychInfo and CINAHL, and inspection of reference lists of relevant papers for the period 1966–December 2013 was then undertaken by one author (CB). The databases were searched for all articles using the terms (“social* support*” or “social* isolation*” or “social* network*”) AND COPD. No reviews specifically addressing the impact of social relationships on relevant outcomes amongst patients with COPD were identified, but relevant related review papers were sourced and checked for additional papers meeting our selection criteria.

Peer-reviewed studies written in English, German or Dutch languages were considered for inclusion. Articles were excluded when outcomes were not reported independently for social support or social network (for example if only a full regression model was reported), or if results for COPD patients were not reported separately to those of participants with other chronic illnesses (including asthma). Further, if outcomes were only focussed on smoking cessation or caregivers the study was excluded. Studies that reported social function or social isolation only as part of a quality-of-life instruments (such as the SF 36 or Nottingham Health Profile) were also excluded.

Our focus was on studies that measured and reported outcomes for social support or social networks using instruments or questions designed to assess these variables. These included both functional measures of social support (6) (or perceived social support) and structural support in the form of social networks. Social support measures assess resources provided by others (8, 9) and can include dimensions of emotional, instrumental, appraisal and informational supports (10, 14). The term social network refers to the web of relationships that surround an individual (10) and in particular the structural features such as the type and strength of each social relationship (15).

We excluded studies that only used marital status or living arrangements as a measure of structural support. We decided to focus on the measurement of the perceived availability of functional support because of our belief that a person's perceptions about available support are most important (16) and the difficulty in determining how marital status was assessed and recorded in individual studies. We have only reported the findings from quantitative studies here; results from qualitative studies will be reported elsewhere as part of a meta-synthesis, mostly due to difficulty in integrating the findings from these different study types in a coherent manner with findings from the quantitative studies.

Data extraction tables were developed by the authors and then information from papers was initially extracted by one author (CB). The accuracy of this information was then independently verified by the other 2 co-authors (TE and PC). Any disputes were resolved by discussion amongst the 3 authors. As this was a scoping study we did not attempt meta-analysis but we provide a descriptive synthesis of the main findings to demonstrate where there is current evidence for associations between social relationships and health outcomes. Conversely, we are also able to highlight where further research is needed to determine an association between social relationships and health outcomes (13).

Results

The search yielded 253 articles from PubMed; 82 from CINAHL and 45 from PsychInfo. After removing duplicates a total of 318 papers were considered for inclusion in the review, of which 287 were excluded following a reading of the title and/or abstract for relevance and applying selection criteria, leaving a final number of 31 papers that were included (See Figure 1).

Figure 1. Flow chart depicting selection of papers for inclusion in the review and reasons why papers were excluded.

The included articles described a range of outcomes against which the impact of social support and social networks were assessed. These were grouped for data extraction into studies reporting on either the effect of social support (Tables 1–4) or social networks (Table 5).

Table 1. Impact of social support on mental health

Author/Date	Sample/Setting	Study Type and Intervention if Applicable	Social Support Measure	Outcome and Outcome Measure	Main Finding for Social Support (Include n analysed if different from n in sample)	Social support of benefit
Toshima et al. (17) 1992	N=119 patients with COPD from a PR program at University of California, San Diego, USA.	RCT – comprehensive PR program vs educational control	Short version of The Social Support Questionnaire	Depression assessed using the Center for Epidemiological Studies Depression scale (CES-D)	At baseline, depression score was significantly correlated with social support quantity and social support satisfaction.	+
Keele-Card et al. (18) 1993	N=30 clients with COPD and their spouses recruited from a private pulmonary clinic in a small midwestern city, USA.	Cross-sectional	The Social Support questionnaire	Loneliness assessed by UCLA loneliness scale and Depression symptoms assessed by CES-D	For patients with COPD, social support satisfaction correlated negatively with loneliness and depression, but no correlation found for spouses.	+
^Grodner et al. (19) 1996	N=110 patients with COPD who participated in an RCT of PR in San Diego, USA.	RCT of 8-week intervention involving either comprehensive PR or education control with follow up at 2, 6 and 12 months, then 6 years.	Short form of The Social Support Questionnaire including number of persons (SSQ-N) AND satisfaction with support (SSQ-S)	Depression symptoms assessed by CES-D	Negative correlation for depression symptoms and satisfaction with social support.	+
^McCathie et al. (20) 2002	N=92 males including both discharged in-patients and outpatients, Sydney Australia.	Cross-sectional	Illness specific (rheumatoid arthritis) social support scale	Depression assessed using BDI and anxiety symptoms assessed by State Trait Anxiety Inventory.	Higher levels of positive social support were linked to lower levels of depression and anxiety.	+
Kara et al. (21) 2004	N=30 patients with COPD and their spouses recruited from an outpatient clinic of a university hospital in eastern Turkey	Cross-Sectional	Perceived Social Support from Friends and Family Scale	Loneliness assessed by UCLA loneliness scale and Depression assessed by Beck Depression Inventory (BDI)	A negative relationship found between loneliness and social support for both patients and their spouse.
Similarly, negative association found between depression and social support for both patients and spouses	+
Alexopoulos et al. (22) 2006	N=63 consecutively admitted PR patients with COPD and major depression, New York, USA	Before and after study to assess the effectiveness of a comprehensive intervention incorporating tailored treatment regimen and PR.	Duke Social Support Index	Depression symptoms assessed by the Hamilton Depression Rating Scale (HAM-D)	Social support and satisfaction with treatment were predictors of improvement in depression following PR	+
Koenig (23) 2006	N=711 consecutively hospitalised Chronic Pulmonary Disease Patients at Duke University medical centre, North Carolina, USA.	Prospective. Patients with minor depression assessed at 6 and 12 weeks. Those with major depression at 6, 12, 18 and 24 weeks.	Duke Social Support Index	Course of depression assessed using Longitudinal Interval Follow-up Evaluation (LIFE-II).	Patients with minor depression remitted more quickly during 12 weeks of follow up with greater social support, but no difference for those with major depression.	+
Coultas et al. (24) 2007	N=207 COPD patients from 17 primary care clinics part of an urban academic health centre, Jacksonville, Florida, USA.	Cross-sectional analysis as part of a RCT of nurse assisted home care for patients COPD.	Medical Outcome Study (MOS) Social Support questionnaire	Depression symptoms assessed using the CES-D	Social support inversely related to the level of depressive symptoms	+
Holm et al. (27) 2010	N=542 COPD living in the community who had been assessed or treated for COPD at one of 2 medical centres in Denver Colorado, USA	Cross-sectional	General Family Functioning sub-scale of the Family Assessment Device	Psychological distress assessed by summing depression and anxiety domains of the Hospital Anxiety and Depression Scale (HADS)	Unsupportive family relationships were associated with psychological distress. Unsupportive family relationships were indirectly associated with smoking status via psychological distress	+
Medinas-Amoros et al. (28) 2012	N=80 consecutively hospitalised COPD patients, assessed after 10 days of hospitalisation, Majorca, Spain.	Cross-sectional	MOS Social Support Survey	Hospital Stress rating scale	Linear regression analysis showed the predictive value of social support on stress level during hospitalisation.	+
DiNicola et al. (26) 2013	N=452 individuals from a population based cohort of US adults living with obstructive airways disease.	Cross-sectional	Positive and Negative Social Exchanges scale	Anxiety Symptoms assessed using the HADS.	Multiple regression analysis revealed that the receipt of instrumental support, feeling let down by others and unsympathetic support predicted higher patient anxiety.	+
Lee et al. (25) 2013	N=209 patients with COPD recruited from 3 university hospitals and 2 clinics in the southern area of South Korea	Cross-sectional	Personal Resource Questionnaire (PRQ)	Depression and Anxiety assessed using the Korean version of the HADS.	Patients with more perceived social support experienced fewer depressive symptoms through the use of problem-oriented coping strategies.	+

+ indicates social support of statistically significant benefit; 0 indicates social support of no statistically significant benefit.

PR = Pulmonary Rehabilitation.

^ Paper reports results for multiple outcomes.

Social support

Studies describing relationships between perceived social support and health outcomes are summarised in Tables 1–4. There were 2 areas in which there was consistent evidence of benefit of social support for patients with COPD; namely mental health (Table 1) and self-efficacy (Table 2). Twelve studies were identified that reported benefits of social support for COPD patients with depression (9 studies) (17–25), anxiety (2 studies) (25, 26), psychological distress or stress (2 studies) (27, 28) in community, outpatient and in-patient hospital settings (Table 1).

Table 2. Impact of social support on self-efficacy and adherence

Author/Date	Sample/Setting	Study Type and Intervention if Applicable	Social Support Measure	Outcome and Outcome Measure	Main Finding for Social Support (Include n Analysed if Different From n in Sample)	Social Support of Benefit
^Grodner et al. (19) 1996	N=110 patients with COPD who participated in an RCT of PR in San Diego, USA.	RCT of 8 week intervention of either comprehensive PR or education control with follow up at 2, 6 and 12 months, then 6 years.	The Social Support Questionnaire including number of persons (SSQ-N) AND satisfaction with support (SSQ-S)	Self-Efficacy assessed from instrument adapted from Kaplan et al. (1984) using walking sub-scale	Positive correlation at baseline for satisfaction with social support and self-efficacy for walking	+
Young et al. (33) 1999	N=88 outpatients enrolled in a 4 week PR program, Auckland, New Zealand.	Cross sectional	Modified O'Reilly and Thomas (1989) General and Social Support Questions specific to risk reduction efforts	Adherence to PR program	52 participants completed the program and 36 did not. Those who did not complete the program were more likely to lack social support for COPD related problems.	+
Xiaolian et al. (32) 2002	N=120 COPD outpatients in Chengdu, China	Cross sectional	Modified Perceived Social Support from ‘Family Scale’	Self-care behaviour scale for COPD patients	91% of patients reported high levels of family support. A slightly statistically significant positive relationship between family support and total self-care behaviour score was found.	+
^Hesselink et al. (39) 2004	N=53 patients with COPD and N=220 with asthma from primary care in the Netherlands.	Cross-Sectional	Social Support List - Interactions	Self-Efficacy, mastery and self-esteem assessed by the Self-Efficacy Scale, 5-item Pearlin Mastery Scale, and the question “on the whole, I'm satisfied with myself?”	Social support was not correlated with self-efficacy, mastery or esteem for COPD patients.	0
Kara-Kasikci and Alberto (30) 2007	N=200 outpatients with COPD in eastern Turkey	Cross-Sectional	Perceived Social Support from ‘Family scale’	Self-efficacy and self-care behaviour score from the COPD Self-Efficacy Scale	Significant positive relationship between family support and self-care behaviour, and family support and self-efficacy.	+
Bonsaksen et al. (31) 2012	N=86 (39.1%) COPD and N=134 (60.9%) morbid obesity attending a patient education course, Oslo, Norway.	Baseline data (cross-sectional) from a longitudinal study	Single question “I think I have enough support from people with whom I have a close relationship?”	Self-efficacy assessed using the General Perceived Self-Efficacy Scale.	COPD participants had higher levels of social support than obese participants. In the COPD group, higher social support predicted higher self-efficacy after controlling for physical activity and other illness perception variables.	+
Lee et al. (29) 2013	N=249 outpatients with COPD from 5 hospitals in South Korea	Cross sectional	Personal Resource Questionnaire 2000	Self-efficacy assessed using COPD Self-Efficacy Scale	In uni-variate analyses, social support was significantly different between patients with high vs low self-efficacy, but was not a significant predictor in a binary logistic regression model.	0

+ indicates social support of statistically significant benefit; 0 indicates social support of no statistically significant benefit.

PR = Pulmonary Rehabilitation.

^ Paper reports results for multiple outcomes.

Self-efficacy and self-care behaviour (including adherence) was also consistently found to be better when study participants reported better social support (Table 2). Five of seven studies reported positive associations [and one of the negative studies found benefits in univariate analyses but the relationship failed to reach statistical significance in multi-variate analyses (29)]. Three studies reported positive relationships between social support and self-efficacy (19, 30, 31) and studies from Xiaolian et al. (32) and Kara-Kasickci et al. (30) reported positive associations with self-care behaviour. One further study [Young et al. (33)] reported adherence to a pulmonary rehabilitation program was higher amongst patients with better social support.

There was inconsistent evidence for a relationship between social support and the variables of functional status, quality of life, and self-rated health (Table 3). The study from Arne et al. (34) comprising a community sample of 1,475 patients with COPD found that social support and the absence of economic problems almost doubled the odds for better self-rated health and quality of life. However, the occurrence of COPD in this study was self-reported rather than objectively assessed and the assessment of social support was based only on a single question (Do you have any persons in your vicinity that can provide you with personal support in case of personal problems or life crises?). One other study [Marino et al. (35)] found that amongst 156 depressed COPD patients admitted to hospital, overall functioning as assessed by the WHO disability assessment schedule II was predicted by subjective social support assessed by the Duke Social Support Inventory.

Table 3. Impact of social support on functional status, quality of life and self-rated health

Author/Date	Sample/Setting	Study Type and Intervention if Applicable	Social Support Measure	Outcome and Outcome Measure	Main Finding for Social Support (include n analysed if different from n in sample)	Social Support of Benefit
Blake et al. (40) 1990	N=32 intervention and N=36 control subjects assessed at 12 month follow up recruited from 2 outpatient speciality clinics, Missouri, USA.	RCT of a psychosocial intervention that emphasized the role of social supports and the use of stress reduction techniques. Results from 12 month follow up reported here.	McKay 13-item social support index.	Functional status measured using the Sickness Impact Profile (SIP)	The experimental group had better physical function and overall function, but there were no significant differences between groups for social support after controlling for baseline scores.	0
Lee et al. (41) 1991	N=30 COPD patients with FEV1 less than 40% pred. and on oxygen therapy from pt. list of 2 metropolitan hospitals, Toronto, Canada.	Cross-sectional	Personal Resource Questionnaire (PRQ)	Functional status measured using the Sickness Impact Profile (SIP)	PRQ score (social support) was moderately negatively correlated with functioning (SIP score) although was not a significant predictor in multiple regression analysis.	0
Anderson (42) 1995	N=126 COPD patients from a private pulmonary specialty practice in the Pacific North west region of the U.S.A.	Cross-sectional	Personal Resource questionnaire (PRQ)	QoL measured by the 16 item Quality of Life Scale (QOLS).	Social support had a direct positive effect on life quality and was moderately positively correlated with QOLS score.	+
Graydon et al. (43) 1995	N=143 entered the study but only N=71 participants completed the study. Community sample with COPD living in or adjacent to metropolitan Toronto, Canada.	Prospective. Six visits to participants homes at six month intervals.	Personal Resource Questionnaire (PRQ)	Functional status assessed by Sickness Impact Profile (SIP)	PRQ score (social support) was (weakly) negatively correlated with functioning (SIP score) although was not a significant predictor in multiple regression analysis.	0
^Grodner (19) 1996	N=110 patients with COPD who participated in an RCT of PR in San Diego, USA.	RCT of 8 week intervention of either comprehensive PR or education control with follow up at 2, 6 and 12 months, then 6 years.	Short form of The Social Support Questionnaire including number of persons (SSQ-N) AND satisfaction with support (SSQ-S)	QoL assessed using the Quality of Well Being (QWB) Scale	Higher satisfaction with support at baseline was not related to QWB scale score.	0
Scharloo et al. (44) 1998	N=244 consecutive adult patients, (80 COPD; 80 rheumatoid arthritis; 80 psoriasis) recruited from Leiden University Medical Centre hospital outpatient clinics, Netherlands.	Cross sectional	Utreschtse Coping List – seeking social support sub-scale	Functional status assessed by the Daily Activities of Life scale and generic function using the Dutch version of the Medical Outcome Study Short Form General Health Survey	Coping strategy characterised by seeking social support explained small but significant amounts of variance for patients with psoriasis, but not COPD.	0
^McCathie et al. (20) 2002	N=92 males including both discharged in-patients and outpatients, Sydney Australia.	Cross-sectional	Illness specific (rheumatoid arthritis) social support scale	QoL assessed using St George's Respiratory Questionnaire (SCRG)	Higher levels of positive social support were not associated with quality of life.	0
^Hesselink et al. (39) 2004	N=53 patients with COPD and N=220 with asthma from primary care in the Netherlands.	Cross-Sectional	Social Support List - Interactions	QoL assessed using the QoL in Respiratory Illness Questionnaire.	Social support was not associated with QoL for COPD or asthma patients	0
Trappenburg et al. (45) 2005	N=81 consecutive patients enrolled in an outpatient PR program at a University hospital, Leuven, Belgium	Before and after study involving a 12 week PR program administered as a multidisciplinary outpatient program.	The Social Support List	Outcomes of PR including exercise capacity, QOL, functional status assessed using the Chronic Respiratory Disease questionnaire and Modified Pulmonary Functional Status and Dyspnoea Questionnaire	At baseline, score for positive social interaction indicated an average level of social support. All outcomes improved following PR but no relationship between baseline social support scores and PR outcomes including 6 minute walk test, QoL, or functional status.	0
						(Continued)

Table 3. (Continued)

Author/Date	Sample/Setting	Study Type and Intervention if Applicable	Social Support Measure	Outcome and Outcome Measure	Main Finding for Social Support (include n analysed if different from n in sample)	Social Support of Benefit
Marino et al. (35) 2008	N=156 depressed COPD patients following hospital admission to the Burke Rehabilitation Hospital, New York, USA.	Cross-Sectional	Duke Social Support Inventory	Overall functioning assessed using the WHO Disability Assessment Schedule II	Instrumental support, and subjective social support predicted overall functioning in bivariate analysis, but social interaction frequency was not related to overall functioning. In multivariate analyses, only greater subjective social support was independently associated with better functioning	+
Arne et al. (34) 2011	N=1,475 Community sample with COPD selected from stratified random sample in five counties in central Sweden.	Repeated Cross-sectional	Single question “Do you have any persons in your vicinity who can provide you with personal support in case of personal problems or life crises?”	Self-rated health assessed from the question “how do you rate your general health status? and QoL assessed using the EuroQoL	Social support and absence of economic problems almost doubled the odds for better self-rated health and QoL	+
Mertens et al. (46) 2011	N=361 community dwelling Dutch men and women, (176 COPD; 185 TII Diabetes) with symptoms of depression.	Cross-Sectional	Social Support List of Interactions (SSL-I-12).	QoL - Physical, Mental and Social function domains of the SF36	Amongst the group with COPD, no associations were found between social support and physical, mental or social function.	0
Bonsaksen et al. (47) 2014	N=60 participants in a patient education course with moderate to severe COPD, at 6 different sites in Norway.	Prospective	Single question “I think I have enough support from people with whom I have a close relationship”	QoL measured using the SF12	Higher social support was related to higher Mental Component Summary (MCS) score at baseline controlling for other factors, but not physical component summary score (PCS). At follow up, social support was not directly related to MCS or PCS when the other factors were controlled for.	0

+ indicates social support of statistically significant benefit; 0 indicates social support of no statistically significant benefit.

PR = Pulmonary Rehabilitation.

QoL = Quality of Life.

WHO = World Health Organisation.

^ Paper reports results for multiple outcomes.

Readmission following hospital discharge for treatment of COPD (3 studies) (Table 4) was not associated with level of perceived social support. One study reported mixed benefits of social support in terms of reduced risk of mortality amongst patients following a program of pulmonary rehabilitation (benefits for females but not males) (Table 4).

Table 4. Impact of social support on hospital re-admission and mortality

Author/Date	Sample/Setting	Study Type and Intervention if Applicable	Social Support Measure	Outcome and Outcome Measure	Main Finding for Social Support (include n analysed if different from n in sample)	Social Support of Benefit
Almagro et al. (48) 2006	N=129 consecutive hospitalised COPD patients in Barcelona, Spain	Prospective	Social Resources Scale	Hospital readmission in 12 months	Social variables not significantly related to readmission	0
Chen and Narsavage (49) 2006	N=145 consecutive patients with COPD scheduled for hospital discharge in 6 regional hospitals in Pingtung area, Taiwan.	Prospective	Personal Resource Questionnaire	Hospital readmission at 14 days and 90 days	Perceived family support was not related to readmission.	0
Coventry et al. (37) 2011	N=79 COPD patients cared for by nurse led early discharge service in Manchester, UK.	Prospective	ENRICHD Social Support Inventory	Hospital readmission in 12 months	Perceived social support not associated with readmission.	0
^Grodner et al. (19) 1996	N=110 patients with COPD who participated in an RCT of PR in San Diego, USA.	RCT of 8 week intervention of either comprehensive PR or education control with follow up at 2, 6 and 12 months, then 6 years.	Short form of The Social Support Questionnaire including number of persons (SSQ-N) AND satisfaction with support (SSQ-S)	Mortality	Higher satisfaction with social support at baseline related to improved survival at 6 years for females, but not males.	+

+ indicates social support of statistically significant benefit; 0 indicates social support of no statistically significant benefit.

^ Paper reports results for multiple outcomes.

Social networks

Only three studies met our selection criteria that have reported on the social network of individuals with COPD (Table 5) and only one study [Marino et al. (35)] was published in the past decade. The study by Grodner et al. (19) found positive relationships between the number of persons in the COPD patient's social network and self-efficacy for walking, disease severity, exercise tolerance and self-reported shortness of breath and a negative correlation with depression symptoms. In this study, following a program of pulmonary rehabilitation, the number of persons in the patients social network was associated with improved breathlessness after 2, 6 and 12 months of follow up. Marino et al. (35) found social network size measured using the Duke Social Support Inventory to be associated with overall functioning in bivariate analyses, but this relationship just failed to reach statistical significance in multi-variate analyses. Finally, Keele-Card and Barron (36) found no relationship between the number of people in the COPD patient's social network and reports of loneliness or depression (Table 5).

Table 5. Social networks and COPD

Author/Date	Sample/Setting	Study Type	Network Measure	Outcomes	Main Finding for Social Network	Larger Social Network of Benefit
Keele-Card and Barron (36) 1993	N=30 clients with COPD and their spouses recruited from a private pulmonary clinic in a small midwestern city, USA.	Cross-sectional	6-item, The Social Support questionnaire	Network size for COPD clients and their spouse, Loneliness and Depression	Number of people reported as sources of social support was similar for patients and spouses (n=2.38 and 2.07 respectively, p=NS). COPD patients relied more on spouse for support, cw. spouse who relied more on non-family.	0
					Number of people in network was not correlated with loneliness or depression.
Grodner et al. (19) 1996	N=110 patients with COPD who participated in an RCT of PR in San Diego, USA.	RCT of 8 week intervention of either comprehensive PR or education control with follow up at 2, 6 and 12 months, then 6 years.	Short form of The Social Support Questionnaire including number of persons (SSQ-N) AND satisfaction with support (SSQ-S)	Self-Efficacy, Well-Being, Depression, Mortality	Positive relationships at baseline between number of persons in network and self-efficacy for walking, disease severity, exercise tolerance and self-reported shortness of breath. Negative correlation found between number of persons and depression symptoms.	+
					After PR, number of persons in network was associated with improved breathlessness at 2, 6 and 12 months.
Marino et al. (35) 2008	N=156 depressed COPD patients following hospital admission to the Burke Rehabilitation Hospital, New York, USA.	Cross-Sectional	Duke Social Support Inventory	Overall functioning assessed by the WHO Disability Assessment Schedule II	Participants had a mean social network size of 4.7 (SD=4.2).	0
					Social network size was associated with overall functioning in bivariate analysis, but just failed to reach statistical significance in multivariate analyses.

PR = pulmonary rehabilitation.

+ Indicates larger social network of statistically significant benefit; 0 indicates larger social network of no statistically significant benefit.

Discussion

This scoping review started with a broad question: what do we know about the health impacts of social relationships for patients with COPD? The search criteria and inclusion criteria were deliberately kept broad in order to facilitate the inclusion of the largest number of studies possible. This approach enabled the determination of the size and nature of the evidence which in turn allowed us to identify gaps in the literature and make recommendations for future primary research. Following an initial inspection of the literature we focussed on studies reporting associations with perceived social support and social networks. Most of the identified literature was concerned with ‘perceived’ social support. Clear trends emerged in terms of the outcomes that benefit from social support. There were benefits for mental health and sense of self-efficacy, but inconsistent support was found in terms of quality of life, functional status, and hospital readmission. From the literature identified, little can be determined about where this perceived social support is derived and the characteristics of the social network of COPD patients.

Much of the research investigating the role of social relationships amongst patients with COPD has focussed on perception of social support and so we also focussed our review on this variable. However, conceptual models of how social networks impact health incorporate other “upstream” factors, such as the size and structure of the social network and the characteristics of network ties such as frequency of contact, reciprocity, duration and intimacy (10). Revealing these upstream factors could enable a more comprehensive understanding of the pathways by which health outcomes are impacted and opportunities for intervention. These upstream factors such as network size, range, density, boundedness, reachability and the characteristics of network ties are conceptualised as providing the opportunities through which downstream psychosocial (e.g., social support) and behavioural pathways influence health and wellbeing outcomes (10). However, our understanding of the social network of individuals with COPD is limited and those studies included in this review only provided minimal information about the number of individuals in the network (its size) and outside of this no information was found about social network structures or characteristics of network ties.

A large number of studies report associations between social support and important health outcomes for COPD patients and some clear trends emerged as to where there is good consistent evidence of benefits, and where further research is needed to clarify any benefits of perceived social support. A number of validated questionnaires were used by these studies which provides some confidence in the observed outcomes, although a small number of studies relied on single item assessments of support and some used instruments developed initially for use with different patient groups (e.g., (20, 37)). It is unclear if the same results would have been achieved given more thorough assessment and/or using generic social support instruments or instruments providing COPD specific social support assessment (Table 6).

Table 6. Instruments used to assess social support and social networks

Instrument Name (abbreviation) and Original Reference	What is Assessed	Length	Used by
McKay 13-item index [McKay et al. (50)]	Six dimensions of support (personal networks, community networks, intimacy, satisfaction-loneliness, appreciation-understanding, tangible assistance.)	13 items	Blake et al. (1990) (40)
The Social Support Questionnaire (SSQ-N and SSQ-S) [Sarason et al. (51)]	Number of persons in network providing social support (SSQ-N) and Satisfaction with support (SSQ-S)	6 (2 part) items	Toshima et al. (1992) (17); Keele-Card and Barron 1993 (18); Grodner et al. 1996 (19)
Illness specific (rheumatoid arthritis) social support scale (ISSS) [Revenson and Schiaffino (52)]	The scale rates how often the social network responds in different ways when individuals are not feeling well due to illness.	37 items	^McCathie et al. (20) 2002
Perceived Social Support from Friends and Family Scale [Procidano and Heller (53)]	Provides a score for perceived support from family and from friends	40 items	Xiaolian et al. (32) 2002; Kara and Mirici 2004 (21); Kara-Kasikci and Alberto (30) 2007
Duke Social Support Index (DSSI) [abbreviated versions Koening et al. (55)]	DSSI measures multiple dimensions of social support including network size, social interaction frequency, instrumental support and subjective support.	11 items	Alexopoulos et al. 2006 (22); Koenig 2006 (23); Marino et al. 2008 (35)
Medical Outcome Study (MOS) Social Support questionnaire [Sherbourne and Stewart (16)]	Assesses perceptions of availability of different functional aspects of social support	19 items	Coultas et al. 2007 (24); Medinas-Amoros et al. 2012 (28)
General family functioning subscale of the Family Assessment Device (FAD) [Byles et al. (57)]	Provides a global assessment of general family functioning	12 items	Holm et al. (27) 2010
Positive and Negative Social Exchanges scale (PANSE) [Newsom et al. (58)]	Developed for use with older adults and addresses several aspects of both positive and negative domains of social exchanges.		DiNicola et al. (26) 2013
Personal Resource Questionnaire (PRQ) [Brandt and Weinert (60)]	Assesses the extent to which the individual feels valued, is an integral part of a group, provides attachment/intimacy and has available informational, emotional and material help.	25 items	Lee et al. (41) 1991; Anderson (42) 1995; Graydon et al. (43) 1995; Chen and Narsavage 2006 (49); Lee et al. (29) 2013
Utrechtse Coping List – seeking social support sub-scale [Schreurs et al. (61)]	Subscale that asks about seeking social support	5 items	Scharloo et al. (44) 1998
Social Support List – Interactions (SSL12-I) [Kempen and van Eijk (62)]	Measures received social support in terms of emotional, informative and instrumental support.	12 items	Hesselink et al. 2004 (39); Trappenburg et al. 2005 (45); Mertens et al. 2011 (46)
Social Resources Scale of the Older Americans Resources and Services (OARS) Program	The Duke OARS (Older Americans Resources and Services) Program, was developed as a means of determining the impact of services and alternative service programs on the functional status of older persons. The first section of this structured interview seeks information on five dimensions of functioning including Social Resources	15 items	Almagro et al. (48) 2006
ENRICHD Social Support Inventory [Berkman et al. (63)]	Initially developed for use with coronary heart disease patients. Measures level of perceived emotional support, instrumental support and informational support.	7 items	Coventry et al. (37) 2011
Single question “Do you have any persons in your vicinity who can provide you with personal support in case of personal problems or life crises?”	Perceived availability of social support	1 question	Arne et al. (34) 2011
Single question “I think I have enough support from people with whom I have a close relationship?”	Satisfaction with social support	1 question	Bonsaksen et al. (31) 2012; Bonsaksen et al. (47) 2014

In this review, we limited inclusion of studies to those that assessed functional support which is characterised by the perception of aid and encouragement provided to the individual by their social network. It was uncommon however for the studies we identified to report which components of social support were most beneficial (emotional, instrumental, appraisal or ­informational), and usually just a single number was given as a total social support score. It would be useful for future studies to better delineate which aspects of social support are most beneficial as this information could support the development of interventions to enhance social support.

Social support was most clearly of benefit in the area of mental health, and in particular a consistent negative relationship is seen with depression. This is consistent with Cohen's theory that social support influences the well-being of individuals by buffering the adverse effects of stressful life events and that the perception of available social support can be a source of general positive affect as well as enhanced self-esteem and feelings of belonging and security (9, 38). Furthermore, social support was also beneficial to a sense of self-efficacy. This is consistent with the view that social support is health promoting, because it might facilitate healthier behaviours as well as greater adherence to medical regimens.

This review began with a broad question, but following an initial inspection of the available literature focussed on quantitative studies investigating two aspects of social relationships, namely perceived social support and the available social network. A number of qualitative studies detailing the lived experience of COPD and its impact on social relationships were caught by these search terms but we were unable to effectively integrate these with the emerging quantitative findings. A separate review using appropriate methodology would be useful to expand upon this initial identification of qualitative literature using dedicated search terms and meta-synthesis of qualitative findings.

Other aspects of social relationships not assessed in the studies included in this review might also be important mediators or moderators of health outcomes for COPD patients. Psychosocial mechanisms such as social influence, social engagement, person to person contact and access to resources (such as ability to work) may also influence pathways to health outcomes (10). Further upstream factors associated with the social network (mezzo level factors) still need to be characterised and understood amongst COPD patients including the characteristics of social network ties and the structure of social networks, which provide the opportunities for social support and those other micro psychosocial mechanisms.

Further, in social network theory, the characteristics of network ties are often delineated as strong ties (spouse and family) and weak ties (friends and acquaintances) but no studies were found that explicitly attempted to evaluate the impact of weak social network ties or how those ties relate to management or health outcomes for COPD patients. Finally, these micro and mezzo level factors occur within the context of broader macrolevel social structural conditions such as culture, socioeconomic status, politics and social change, social capital and social cohesion, and it remains undetermined how these factors impact upon the social networks and availability of social support for patients with COPD.

Conclusion

There is consistent evidence of a positive benefit to mental health and sense of self-efficacy for patients with COPD who have adequate perceived social support, and there is an increasing body of evidence suggesting benefits in terms of self-care and adherence in the presence of adequate social support. However, there remains a need to identify the factors that promote and maintain social support as understanding these will aid in designing optimal interventions for patients with COPD. In particular there is a need to further understand the characteristics of social networks amongst patients with COPD and how this influences health and well-being outcomes for patients with COPD; within the broader social structural conditions in which COPD patients live and manage their illness.

Declaration of Interest Statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.