Heart & Lung xxx (2014) 1e7
Contents lists available at ScienceDirect
Heart & Lung journal homepage: www.heartandlung.org
Types of social support and their relationships to physical and depressive symptoms and health-related quality of life in patients with heart failure Seongkum Heo, PhD, RN a, *, Terry A. Lennie, PhD, RN b, Debra K. Moser, DNSc, RN b, Robert L. Kennedy, PhD c a b c
University of Arkansas for Medical Sciences College of Nursing, 4301 W. Markham Street #529, Little Rock, AR 72205, USA University of Kentucky, Lexington, USA Office of Educational Development, University of Arkansas for Medical Sciences, Little Rock, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 October 2013 Received in revised form 14 April 2014 Accepted 18 April 2014 Available online xxx
Objectives: To examine the various types of social support associated with physical and depressive symptoms and health-related quality of life (HRQOL) in patients with heart failure (HF) and the mediating effects of symptoms on the relationship between social support and HRQOL. Background: Patients with HF have a high burden of physical and depressive symptoms, along with poor HRQOL. Social support may improve symptoms and HRQOL. Methods: Data on social support (marital status, family relationships, relationships with health care providers, social networks, emotional support, and instrumental support), symptoms, and HRQOL were collected from 71 patients. Hierarchical regression was used to analyze the data. Results: Emotional support was related to all physical and depressive symptoms and HRQOL. Physical and depressive symptoms mediated the relationship between emotional support and HRQOL. Conclusions: Further studies are needed to identify ways to improve emotional support and determine whether the improvement leads to improvements in symptoms and HRQOL. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Heart failure Social support Emotional support Symptoms Quality of life
Introduction Health-related quality of life (HRQOL) is poorer in patients with heart failure (HF) than in healthy populations and in patients with other chronic diseases.1e3 Heart failure symptoms, including dyspnea and fatigue, and depressive symptoms are prevalent in this population and are strongly associated with poor HRQOL.3e7 Physical symptoms can prevent patients with HF from performing their daily activities that lead to poor HRQOL.8 For instance, many patients with HF have New York Heart Association (NYHA) functional class IIeIV,9 indicating that they experience HF symptoms when they perform daily activities.10 In addition, depressive symptoms also can cause functional impairment that leads to poor HRQOL in this population. For example, depressed patients with HF had reduced daily activities and walking distance compared to non-depressed patients.11 Thus, to improve
Abbreviations: HF, Heart failure; HRQOL, Health-related quality of life; NYHA, New York Heart Association. * Corresponding author. Tel.: þ1 501 686 5375; fax: þ1 501 296-1765. E-mail address: [email protected] (S. Heo). 0147-9563/$ e see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.hrtlng.2014.04.015
outcomes, it is important to identify modifiable factors affecting symptoms and HRQOL. The revised Wilson and Clearly model suggests that social support may be related to both symptoms and HRQOL.12 Social support may affect physical symptoms through effects on self-care. For instance, social support is associated with adherence to medication treatment and following a low sodium diet,13e15 and lack of adherence to medication treatment and a low sodium diet is associated with more severe symptom burden and higher hospitalization rates.16,17 Heart failure symptoms are important antecedents of hospitalizations in this population.4 Social support has also been associated with depressive symptoms in patients with HF.18 However, the findings on relationships between social support and HRQOL have been inconsistent. Patients with HF have reported that social support was a factor affecting their HRQOL,19 and perceived quality of support has been associated with HRQOL.20 However, the majority of HF studies that have examined the relationships between social support and HRQOL found that these relationships were not significant.21e24 These contradictory findings, in part, may be due to differing effects of different types of social support on HRQOL.
2
S. Heo et al. / Heart & Lung xxx (2014) 1e7
Social support in HF studies has been conceptualized in many ways, varying from simple structural support (e.g., marital status) to comprehensive combinations of structural and functional support.20e22,25,26 Structural support refers to the existence of a social network and the features of contacts with the network (e.g., marital status and social network).27,28 Martial status refers to a very simple social network with spouse or cohabitant (Fig. 1). Social networks represent a wider range of social networks that extend beyond marital status, including more extended family members, friends, and society. Functional support refers to individuals’ perceptions of the resources that social networks provide (e.g., emotional support, instrumental support, and relationships to health-care providers).27,28 Emotional support refers to intangible support from others except health care providers. Instrumental support refers to tangible support from others. Relationships to health care providers represent support from health care providers. Family relationships may combine structural and functional support because they include not only the existence of a social network but also individuals’ perceptions of the resources that family relationships provide. These different types of structural and functional support may affect symptoms and, in turn, HRQOL differently. The relationships between different types of social support and physical symptoms in patients with HF have been rarely examined. However, a meta-analysis that examined the relationship between different types of support and adherence to medical treatment found that instrumental support had the strongest association with adherence.25 In addition, adherence to medical treatment was greater in patients from cohesive families than in patients from families in conflict.25 In another study,29 patients with HF perceived health care providers as one source of support for medication adherence. Thus, instrumental support, family relationship, and relationships with health care providers may affect self-care, and, in turn, physical symptoms. Relationships between social support and depressive symptoms have been examined in this population. Among different types of structural and functional support, living with families and greater emotional support were the only variables significantly associated with less severe depressive symptoms.30 Finally, relationships between social support and HRQOL have been examined in this population, and the findings were inconsistent. In one study,20 one type of functional support (perceived quality of support), but not another type of functional support (emotional support), was associated with HRQOL. In two other studies,21,22 a combination of different types of structural and functional support was not associated with HRQOL.
One reason for the lack of relationship of social support to HRQOL in HF studies may be that social support affects HRQOL mainly through its effects on other variables, including physical and depressive symptoms.12 If social support is associated with HRQOL indirectly through its effects on physical symptoms and depressive symptoms, this may explain the lack of direct, independent associations between social support and HRQOL in patients with HF.21e23 Thus, examination of direct and indirect associations of social support with HRQOL will provide valuable information on the theoretical framework of HRQOL and the associating factors. Therefore, we examined the relationships of several types of social support (marital status, social networks, relationships with health care providers, emotional support, instrumental support, and family relationships) to physical symptoms (dyspnea, fatigue, chest pain, edema, sleeping difficulty, and dizziness), depressive symptoms, and HRQOL. We also explored the mediating effects of physical and depressive symptoms on the relationship between social support and HRQOL in patients with HF. Methods Design, setting, and sample A cross-sectional correlational design was used to examine the relationships of social support to physical symptoms, depressive symptoms, and HRQOL in a convenience sample of patients with HF. Institutional Review Board approval was obtained for the current study. Eligible patients were referred to research associates by physicians or nurse practitioners in the HF clinic and then approached by the research associates who were trained in questionnaire administration and interviewing. Patients were recruited from an HF clinic at a hospital in a Mid-Atlantic city in the US between 2008 and 2009. Inclusion criteria were 1) a confirmed diagnosis of HF, 2) NYHA functional class IIeIV (symptomatic patients), 3) ability to read and write English, 4) no dementia, and 5) age 18 years or older. The diagnosis of HF was confirmed through medical record review using established criteria.31 Patients were carefully questioned by research associates to determine NYHA classification. We included only symptomatic patients because they need to be involved in self-management, including restricted sodium intake and adjustment of diuretics,32,33 and they may need more social support.34 We also included only those patients with no history of dementia identified on medical chart for collaboration during data collection. Written informed consent was obtained from all participants after explanation of the study purpose and procedures.
Fig. 1. Theoretical framework.
S. Heo et al. / Heart & Lung xxx (2014) 1e7
Sample size was calculated using G*Power 3.1.6 based on information from the literature.6,35,36 We assumed that social support was associated with HRQOL mainly through its effects on physical and depressive symptoms. Based on two studies,6,35 given an a of .05, power ¼ .95, 7 predictors, and a total expected R-square ¼ 27% (10% explained by physical symptoms and 17% by depressive symptoms), the estimated sample size was 67.36 Eighty patients were enrolled in the study, but 9 were excluded from the final analyses because 5 did not return the questionnaires, and 4 had missing data. Thus, a total of 71 patients were included in the analyses. Measures Health-related quality of life was defined as patients’ perceptions of the impact of HF on various aspects of their daily life,37 and was assessed using the Minnesota Living with Heart Failure Questionnaire.38,39 This instrument consists of 21 items rated on a scale from 0 (no impact on HRQOL) to 5 (considerable impact on HRQOL). Scores may range from 0 to 105; higher scores indicate poorer HRQOL. Reliability and validity were supported in several studies.38e41 Cronbach’s alpha in the current study was .94. Physical symptoms were defined as common HF-related symptoms and were measured using the Symptom Status Questionnaire-Heart Failure.42 The 7-item instrument assesses 7 common HF symptoms during the past 4 weeks: dyspnea during daytime, dyspnea when lying down, fatigue, chest pain, edema, sleeping difficulty, and dizziness. The items have 4 sub-items: presence (0 is given for no symptom), frequency, severity, and distress. Frequency sub-items have 4 response options, from 1 (less than once per week) to 4 (nearly daily). Severity sub-items have 4 response options, from 1 (slight) to 4 (very much). Distress subitems have 5 response options, from 0 (not at all) to 4 (very much). Possible scores range from 0 to 84, with higher scores indicating more severe physical symptoms. Reliability and validity were supported.42 Cronbach’s alpha in the current study was .92. Depressive symptoms were measured by the Patient Health Questionnaire (PHQ-9),43,44 which consists of 9 items corresponding to the major depressive disorder criteria of the Diagnostic and Statistical Manual of Mental Disorders-IV.43 Items assess the frequency of symptoms over the past 2 weeks using a 4-point Likert scale from 0 (not at all) to 3 (nearly every day). Possible scores range from 0 to 27, with higher scores indicating more severe depressive symptoms.43 Reliability and validity have been supported in patients with HF.45 Cronbach’s alpha was .78 in the current study. Social support was operationalized as marital status, social networks, relationships with health care providers, emotional support, instrumental support, and family relationships. Based on a standard sociodemographic questionnaire, patients were categorized as currently married/cohabitating or not married/ no cohabitating (single, divorced/separated, or widowed). Social networks were defined as the number of significant others who were contacted regularly46 and was also assessed using the standard sociodemographic questionnaire. Patients were asked to list persons whom they contacted regularly, including family, friends, neighbors, religious group members, and club or other group members. Possible scores range from 0 to 25, with higher scores indicating bigger social networks. Relationships with health-care providers were defined as individuals’ perceptions of whether their health-care providers did their best to provide good care and were measured using the Wake Forest Physician Trust Scale.47 This measure consists of 10 items using a 5-point Likert Scale ranging from 1 (strongly disagree) to 5 (strongly agree). Possible total scores range from 10 to 50, with higher scores indicating more trust.47 Reliability and validity were supported.47,48 Cronbach’s alpha in the current study was .84.
3
Emotional support was defined as individuals’ perceptions of affective support from family, friends, and important others and was measured using the Multidimensional Scale of Perceived Social Support.49e51 This measure consists of 12 items and uses a 7-point Likert scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). Possible total scores range from 12 to 84, with higher scores indicating greater emotional support. Reliability and validity were supported.51,52 Cronbach’s alpha was .93 in the current study. Instrumental support was defined as tangible support from others and was measured by the Social Support Scale-Instrumental (Heart Failure) developed by the first author. Content validity was supported by two experts in HF research. The measure consists of 5 items on a 5-point Likert Scale ranging from 0 (never) to 4 (always). The 5-items represent 5 types of instrumental support: shopping, preparation of low sodium foods and encouragement to eat low sodium foods, symptom management, monetary support, and readiness to help when instrumental support is needed. Possible total scores range from 0 to 20, with higher scores indicating greater instrumental support. Cronbach’s alpha was .88 in the current study. Family relationships were measured using two subscales of the FES-Family Relationship Index (Cohesion and Conflict).53 Cohesion is defined in the index as the extent of family members’ helpfulness and supportiveness, and conflict as the extent of family members’ expressions of anger and conflicteladen interactions.53 We did not include the Expression subscale in the analysis because of poor reliability in this sample (Cronbach’s alpha ¼ .41). Each subscale has 9 items with dichotomous response options (true or false). Possible total scores for the Cohesion and Conflict subscales range from 4 to 65 and 33 to 80, respectively. Higher scores indicate more cohesive relationships or more conflicted relationships. Reliability and construct validity have been supported.54 The Kuder-Richardson reliability coefficients for the Cohesion and Conflict subscales in the current study were .75 and .74, respectively. Comorbidities,55 age,56 and gender57 were selected as covariates for symptoms. Comorbidities,20 age,20,35 and NYHA functional class58 were selected as covariates for the relationships among social support, symptoms, and HRQOL. Data on covariates and other sociodemographic and clinical characteristics (education, left ventricular ejection fraction, ethnicity, and etiology of HF) were collected using the standard sociodemographic questionnaire and a clinical questionnaire. Data on comorbidities were collected using the Charlson Comorbidity Index, which was included in the clinical questionnaire.59 This index uses the number and seriousness of comorbidities to assess the risk of mortality. The total score is calculated by adding the weighted scores; higher scores indicate more severe comorbidities.59 Data analysis All analyses were done using SPSS (version 21). Bivariate relationships between sociodemographic and clinical characteristics and HRQOL were examined using Pearson’s r for continuous variables and the independent t-test for dichotomous variables. Hierarchical regression analyses were used to determine the types of social support that were significantly related to each of the physical and depressive symptoms, controlling for comorbidities, age, and gender. Hierarchical regression analysis was used to determine the types of social support that were significantly related to HRQOL, controlling for comorbidities, age, and NYHA functional class. A mediated effect implies that the independent variable leads to the mediators, which, in turn leads to the outcome variable (HRQOL).60 General linear regression analysis and simple and hierarchical multiple regression analyses were used to determine whether
4
S. Heo et al. / Heart & Lung xxx (2014) 1e7
Table 1 Characteristics of the study sample and their bivariate relationships to healthrelated quality of life (N ¼ 71). Characteristics
Mean (SD)
Age (Years) Education (Years) Left ventricular ejection fraction (%) Charlson Comorbidity Index
72 12 39 5
Pearson’s r
Number (%)
t statistics
p value
Gender (male) Marital status (married/co-habitant) Ethnicity (Caucasian) Heart failure etiology (Ischemic) NYHA functional class Class II Class III/IV
47 51 65 44
1.438 2.138 .414 1.458 2.488
.155 .036 .680 .149 .015
.200 .061 .079 .279
(11) (2) (14) (3)
(66) (72) (92) (62)
p value .097 .618 .510 .018
37 (52) 34 (48)
NYHA ¼ New York Heart Association. SD ¼ standard deviation.
physical symptoms and depressive symptoms mediated the relationship between emotional support and HRQOL. The mediator effects were examined because only emotional support was associated with all these three variables in bivariate analyses. Simple and hierarchical multiple regression analyses were used to determine whether physical symptoms mediated the relationship between marital status and HRQOL. This mediator effect was examined because marital status was associated with only physical symptoms and HRQOL in bivariate analyses.
status and emotional support were associated with physical symptoms to similar degrees (bigger coefficients indicate greater effects on the outcome variable). When the covariates of comorbidities, age, and gender were entered into the second block of the model, the significant relationships of marital status and emotional support to physical symptoms remained, and none of the covariates were related to physical symptoms. Social networks, relationships with health care providers, instrumental support, and family relationships were not related to physical symptoms. When all types of social support were entered into the model at the same time, only emotional support was significantly related to depressive symptoms before controlling for comorbidities, age, and gender (R2 ¼ .186, p < .001). Patients who had greater emotional support had less severe depressive symptoms. When the covariates were entered into the second block of the model, the significant relationship of emotional support to depressive symptoms remained, and none of the covariates were related to depressive symptoms. Marital status, social networks, relationships with health care providers, instrumental support, and family relationships were not related to depressive symptoms. When all types of social support were entered into the model at the same time, only marital status was significantly related to HRQOL (R2 ¼ .062, p ¼ .036). Patients who remained in their marital relationship or had a cohabitant had better HRQOL. When covariates were entered into the second block, marital status was still associated with HRQOL. In addition, older age and lower NYHA functional class were associated with better HRQOL.
Results Sample characteristics and their relationships to health-related quality of life The mean age of the 71 patients was 72 years (Table 1). The majority were Caucasians, and they had mild to moderate functional impairment (NYHA functional class II or III). Slightly more than half were males, married or had a cohabitant, and had ischemic HF. Among the sociodemographic and clinical characteristics, marital status, comorbidities, and NYHA functional class were significantly associated with HRQOL in bivariate analyses. Relationships of social support to physical and depressive symptoms and health-related quality of life In the multiple regression analysis, all types of social support were entered into the model at the same time, and only marital status and emotional support were significantly related to physical symptoms (R2 ¼ .292, p < .001) (Table 2). Patients who remained in their marital relationship or had a cohabitant and had greater emotional support had less severe physical symptoms. Marital
Mediating relationships of physical symptoms and depressive symptoms on the relationship between social support and healthrelated quality of life Only emotional support was significantly related to physical symptoms, depressive symptoms, and HRQOL in bivariate analyses. Marital status was associated with physical symptoms and HRQOL, but not depressive symptoms in bivariate analyses. Thus, we explored whether physical symptoms and depressive symptoms mediated the relationship between emotional support and HRQOL, and whether physical symptoms mediated the relationship between marital status and HRQOL. First, we examined the relationship of emotional support to physical symptoms and depressive symptoms. Emotional support was significantly related to both physical symptoms and depressive symptoms in the general linear model analysis (R2 ¼ .568 and 540, respectively, p ¼ .003 and p ¼ .009, respectively, Fig. 2). In the multiple regression analysis, physical symptoms and depressive symptoms were significantly related to HRQOL (R2 ¼ .594, p < .001). Emotional support was significantly related to HRQOL in a simple regression model (R2 ¼ .058, p ¼ .042), but when physical symptoms and depressive
Table 2 Multivariate relationships of social support to physical symptoms, depressive symptoms, and health-related quality of life. Dependent variable
Independent variable
Standardized beta
95% Confidence interval
R2
F
p value
Physical symptoms
Model 1and 2
14.017
Contents lists available at ScienceDirect
Heart & Lung journal homepage: www.heartandlung.org
Types of social support and their relationships to physical and depressive symptoms and health-related quality of life in patients with heart failure Seongkum Heo, PhD, RN a, *, Terry A. Lennie, PhD, RN b, Debra K. Moser, DNSc, RN b, Robert L. Kennedy, PhD c a b c
University of Arkansas for Medical Sciences College of Nursing, 4301 W. Markham Street #529, Little Rock, AR 72205, USA University of Kentucky, Lexington, USA Office of Educational Development, University of Arkansas for Medical Sciences, Little Rock, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 October 2013 Received in revised form 14 April 2014 Accepted 18 April 2014 Available online xxx
Objectives: To examine the various types of social support associated with physical and depressive symptoms and health-related quality of life (HRQOL) in patients with heart failure (HF) and the mediating effects of symptoms on the relationship between social support and HRQOL. Background: Patients with HF have a high burden of physical and depressive symptoms, along with poor HRQOL. Social support may improve symptoms and HRQOL. Methods: Data on social support (marital status, family relationships, relationships with health care providers, social networks, emotional support, and instrumental support), symptoms, and HRQOL were collected from 71 patients. Hierarchical regression was used to analyze the data. Results: Emotional support was related to all physical and depressive symptoms and HRQOL. Physical and depressive symptoms mediated the relationship between emotional support and HRQOL. Conclusions: Further studies are needed to identify ways to improve emotional support and determine whether the improvement leads to improvements in symptoms and HRQOL. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Heart failure Social support Emotional support Symptoms Quality of life
Introduction Health-related quality of life (HRQOL) is poorer in patients with heart failure (HF) than in healthy populations and in patients with other chronic diseases.1e3 Heart failure symptoms, including dyspnea and fatigue, and depressive symptoms are prevalent in this population and are strongly associated with poor HRQOL.3e7 Physical symptoms can prevent patients with HF from performing their daily activities that lead to poor HRQOL.8 For instance, many patients with HF have New York Heart Association (NYHA) functional class IIeIV,9 indicating that they experience HF symptoms when they perform daily activities.10 In addition, depressive symptoms also can cause functional impairment that leads to poor HRQOL in this population. For example, depressed patients with HF had reduced daily activities and walking distance compared to non-depressed patients.11 Thus, to improve
Abbreviations: HF, Heart failure; HRQOL, Health-related quality of life; NYHA, New York Heart Association. * Corresponding author. Tel.: þ1 501 686 5375; fax: þ1 501 296-1765. E-mail address: [email protected] (S. Heo). 0147-9563/$ e see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.hrtlng.2014.04.015
outcomes, it is important to identify modifiable factors affecting symptoms and HRQOL. The revised Wilson and Clearly model suggests that social support may be related to both symptoms and HRQOL.12 Social support may affect physical symptoms through effects on self-care. For instance, social support is associated with adherence to medication treatment and following a low sodium diet,13e15 and lack of adherence to medication treatment and a low sodium diet is associated with more severe symptom burden and higher hospitalization rates.16,17 Heart failure symptoms are important antecedents of hospitalizations in this population.4 Social support has also been associated with depressive symptoms in patients with HF.18 However, the findings on relationships between social support and HRQOL have been inconsistent. Patients with HF have reported that social support was a factor affecting their HRQOL,19 and perceived quality of support has been associated with HRQOL.20 However, the majority of HF studies that have examined the relationships between social support and HRQOL found that these relationships were not significant.21e24 These contradictory findings, in part, may be due to differing effects of different types of social support on HRQOL.
2
S. Heo et al. / Heart & Lung xxx (2014) 1e7
Social support in HF studies has been conceptualized in many ways, varying from simple structural support (e.g., marital status) to comprehensive combinations of structural and functional support.20e22,25,26 Structural support refers to the existence of a social network and the features of contacts with the network (e.g., marital status and social network).27,28 Martial status refers to a very simple social network with spouse or cohabitant (Fig. 1). Social networks represent a wider range of social networks that extend beyond marital status, including more extended family members, friends, and society. Functional support refers to individuals’ perceptions of the resources that social networks provide (e.g., emotional support, instrumental support, and relationships to health-care providers).27,28 Emotional support refers to intangible support from others except health care providers. Instrumental support refers to tangible support from others. Relationships to health care providers represent support from health care providers. Family relationships may combine structural and functional support because they include not only the existence of a social network but also individuals’ perceptions of the resources that family relationships provide. These different types of structural and functional support may affect symptoms and, in turn, HRQOL differently. The relationships between different types of social support and physical symptoms in patients with HF have been rarely examined. However, a meta-analysis that examined the relationship between different types of support and adherence to medical treatment found that instrumental support had the strongest association with adherence.25 In addition, adherence to medical treatment was greater in patients from cohesive families than in patients from families in conflict.25 In another study,29 patients with HF perceived health care providers as one source of support for medication adherence. Thus, instrumental support, family relationship, and relationships with health care providers may affect self-care, and, in turn, physical symptoms. Relationships between social support and depressive symptoms have been examined in this population. Among different types of structural and functional support, living with families and greater emotional support were the only variables significantly associated with less severe depressive symptoms.30 Finally, relationships between social support and HRQOL have been examined in this population, and the findings were inconsistent. In one study,20 one type of functional support (perceived quality of support), but not another type of functional support (emotional support), was associated with HRQOL. In two other studies,21,22 a combination of different types of structural and functional support was not associated with HRQOL.
One reason for the lack of relationship of social support to HRQOL in HF studies may be that social support affects HRQOL mainly through its effects on other variables, including physical and depressive symptoms.12 If social support is associated with HRQOL indirectly through its effects on physical symptoms and depressive symptoms, this may explain the lack of direct, independent associations between social support and HRQOL in patients with HF.21e23 Thus, examination of direct and indirect associations of social support with HRQOL will provide valuable information on the theoretical framework of HRQOL and the associating factors. Therefore, we examined the relationships of several types of social support (marital status, social networks, relationships with health care providers, emotional support, instrumental support, and family relationships) to physical symptoms (dyspnea, fatigue, chest pain, edema, sleeping difficulty, and dizziness), depressive symptoms, and HRQOL. We also explored the mediating effects of physical and depressive symptoms on the relationship between social support and HRQOL in patients with HF. Methods Design, setting, and sample A cross-sectional correlational design was used to examine the relationships of social support to physical symptoms, depressive symptoms, and HRQOL in a convenience sample of patients with HF. Institutional Review Board approval was obtained for the current study. Eligible patients were referred to research associates by physicians or nurse practitioners in the HF clinic and then approached by the research associates who were trained in questionnaire administration and interviewing. Patients were recruited from an HF clinic at a hospital in a Mid-Atlantic city in the US between 2008 and 2009. Inclusion criteria were 1) a confirmed diagnosis of HF, 2) NYHA functional class IIeIV (symptomatic patients), 3) ability to read and write English, 4) no dementia, and 5) age 18 years or older. The diagnosis of HF was confirmed through medical record review using established criteria.31 Patients were carefully questioned by research associates to determine NYHA classification. We included only symptomatic patients because they need to be involved in self-management, including restricted sodium intake and adjustment of diuretics,32,33 and they may need more social support.34 We also included only those patients with no history of dementia identified on medical chart for collaboration during data collection. Written informed consent was obtained from all participants after explanation of the study purpose and procedures.
Fig. 1. Theoretical framework.
S. Heo et al. / Heart & Lung xxx (2014) 1e7
Sample size was calculated using G*Power 3.1.6 based on information from the literature.6,35,36 We assumed that social support was associated with HRQOL mainly through its effects on physical and depressive symptoms. Based on two studies,6,35 given an a of .05, power ¼ .95, 7 predictors, and a total expected R-square ¼ 27% (10% explained by physical symptoms and 17% by depressive symptoms), the estimated sample size was 67.36 Eighty patients were enrolled in the study, but 9 were excluded from the final analyses because 5 did not return the questionnaires, and 4 had missing data. Thus, a total of 71 patients were included in the analyses. Measures Health-related quality of life was defined as patients’ perceptions of the impact of HF on various aspects of their daily life,37 and was assessed using the Minnesota Living with Heart Failure Questionnaire.38,39 This instrument consists of 21 items rated on a scale from 0 (no impact on HRQOL) to 5 (considerable impact on HRQOL). Scores may range from 0 to 105; higher scores indicate poorer HRQOL. Reliability and validity were supported in several studies.38e41 Cronbach’s alpha in the current study was .94. Physical symptoms were defined as common HF-related symptoms and were measured using the Symptom Status Questionnaire-Heart Failure.42 The 7-item instrument assesses 7 common HF symptoms during the past 4 weeks: dyspnea during daytime, dyspnea when lying down, fatigue, chest pain, edema, sleeping difficulty, and dizziness. The items have 4 sub-items: presence (0 is given for no symptom), frequency, severity, and distress. Frequency sub-items have 4 response options, from 1 (less than once per week) to 4 (nearly daily). Severity sub-items have 4 response options, from 1 (slight) to 4 (very much). Distress subitems have 5 response options, from 0 (not at all) to 4 (very much). Possible scores range from 0 to 84, with higher scores indicating more severe physical symptoms. Reliability and validity were supported.42 Cronbach’s alpha in the current study was .92. Depressive symptoms were measured by the Patient Health Questionnaire (PHQ-9),43,44 which consists of 9 items corresponding to the major depressive disorder criteria of the Diagnostic and Statistical Manual of Mental Disorders-IV.43 Items assess the frequency of symptoms over the past 2 weeks using a 4-point Likert scale from 0 (not at all) to 3 (nearly every day). Possible scores range from 0 to 27, with higher scores indicating more severe depressive symptoms.43 Reliability and validity have been supported in patients with HF.45 Cronbach’s alpha was .78 in the current study. Social support was operationalized as marital status, social networks, relationships with health care providers, emotional support, instrumental support, and family relationships. Based on a standard sociodemographic questionnaire, patients were categorized as currently married/cohabitating or not married/ no cohabitating (single, divorced/separated, or widowed). Social networks were defined as the number of significant others who were contacted regularly46 and was also assessed using the standard sociodemographic questionnaire. Patients were asked to list persons whom they contacted regularly, including family, friends, neighbors, religious group members, and club or other group members. Possible scores range from 0 to 25, with higher scores indicating bigger social networks. Relationships with health-care providers were defined as individuals’ perceptions of whether their health-care providers did their best to provide good care and were measured using the Wake Forest Physician Trust Scale.47 This measure consists of 10 items using a 5-point Likert Scale ranging from 1 (strongly disagree) to 5 (strongly agree). Possible total scores range from 10 to 50, with higher scores indicating more trust.47 Reliability and validity were supported.47,48 Cronbach’s alpha in the current study was .84.
3
Emotional support was defined as individuals’ perceptions of affective support from family, friends, and important others and was measured using the Multidimensional Scale of Perceived Social Support.49e51 This measure consists of 12 items and uses a 7-point Likert scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). Possible total scores range from 12 to 84, with higher scores indicating greater emotional support. Reliability and validity were supported.51,52 Cronbach’s alpha was .93 in the current study. Instrumental support was defined as tangible support from others and was measured by the Social Support Scale-Instrumental (Heart Failure) developed by the first author. Content validity was supported by two experts in HF research. The measure consists of 5 items on a 5-point Likert Scale ranging from 0 (never) to 4 (always). The 5-items represent 5 types of instrumental support: shopping, preparation of low sodium foods and encouragement to eat low sodium foods, symptom management, monetary support, and readiness to help when instrumental support is needed. Possible total scores range from 0 to 20, with higher scores indicating greater instrumental support. Cronbach’s alpha was .88 in the current study. Family relationships were measured using two subscales of the FES-Family Relationship Index (Cohesion and Conflict).53 Cohesion is defined in the index as the extent of family members’ helpfulness and supportiveness, and conflict as the extent of family members’ expressions of anger and conflicteladen interactions.53 We did not include the Expression subscale in the analysis because of poor reliability in this sample (Cronbach’s alpha ¼ .41). Each subscale has 9 items with dichotomous response options (true or false). Possible total scores for the Cohesion and Conflict subscales range from 4 to 65 and 33 to 80, respectively. Higher scores indicate more cohesive relationships or more conflicted relationships. Reliability and construct validity have been supported.54 The Kuder-Richardson reliability coefficients for the Cohesion and Conflict subscales in the current study were .75 and .74, respectively. Comorbidities,55 age,56 and gender57 were selected as covariates for symptoms. Comorbidities,20 age,20,35 and NYHA functional class58 were selected as covariates for the relationships among social support, symptoms, and HRQOL. Data on covariates and other sociodemographic and clinical characteristics (education, left ventricular ejection fraction, ethnicity, and etiology of HF) were collected using the standard sociodemographic questionnaire and a clinical questionnaire. Data on comorbidities were collected using the Charlson Comorbidity Index, which was included in the clinical questionnaire.59 This index uses the number and seriousness of comorbidities to assess the risk of mortality. The total score is calculated by adding the weighted scores; higher scores indicate more severe comorbidities.59 Data analysis All analyses were done using SPSS (version 21). Bivariate relationships between sociodemographic and clinical characteristics and HRQOL were examined using Pearson’s r for continuous variables and the independent t-test for dichotomous variables. Hierarchical regression analyses were used to determine the types of social support that were significantly related to each of the physical and depressive symptoms, controlling for comorbidities, age, and gender. Hierarchical regression analysis was used to determine the types of social support that were significantly related to HRQOL, controlling for comorbidities, age, and NYHA functional class. A mediated effect implies that the independent variable leads to the mediators, which, in turn leads to the outcome variable (HRQOL).60 General linear regression analysis and simple and hierarchical multiple regression analyses were used to determine whether
4
S. Heo et al. / Heart & Lung xxx (2014) 1e7
Table 1 Characteristics of the study sample and their bivariate relationships to healthrelated quality of life (N ¼ 71). Characteristics
Mean (SD)
Age (Years) Education (Years) Left ventricular ejection fraction (%) Charlson Comorbidity Index
72 12 39 5
Pearson’s r
Number (%)
t statistics
p value
Gender (male) Marital status (married/co-habitant) Ethnicity (Caucasian) Heart failure etiology (Ischemic) NYHA functional class Class II Class III/IV
47 51 65 44
1.438 2.138 .414 1.458 2.488
.155 .036 .680 .149 .015
.200 .061 .079 .279
(11) (2) (14) (3)
(66) (72) (92) (62)
p value .097 .618 .510 .018
37 (52) 34 (48)
NYHA ¼ New York Heart Association. SD ¼ standard deviation.
physical symptoms and depressive symptoms mediated the relationship between emotional support and HRQOL. The mediator effects were examined because only emotional support was associated with all these three variables in bivariate analyses. Simple and hierarchical multiple regression analyses were used to determine whether physical symptoms mediated the relationship between marital status and HRQOL. This mediator effect was examined because marital status was associated with only physical symptoms and HRQOL in bivariate analyses.
status and emotional support were associated with physical symptoms to similar degrees (bigger coefficients indicate greater effects on the outcome variable). When the covariates of comorbidities, age, and gender were entered into the second block of the model, the significant relationships of marital status and emotional support to physical symptoms remained, and none of the covariates were related to physical symptoms. Social networks, relationships with health care providers, instrumental support, and family relationships were not related to physical symptoms. When all types of social support were entered into the model at the same time, only emotional support was significantly related to depressive symptoms before controlling for comorbidities, age, and gender (R2 ¼ .186, p < .001). Patients who had greater emotional support had less severe depressive symptoms. When the covariates were entered into the second block of the model, the significant relationship of emotional support to depressive symptoms remained, and none of the covariates were related to depressive symptoms. Marital status, social networks, relationships with health care providers, instrumental support, and family relationships were not related to depressive symptoms. When all types of social support were entered into the model at the same time, only marital status was significantly related to HRQOL (R2 ¼ .062, p ¼ .036). Patients who remained in their marital relationship or had a cohabitant had better HRQOL. When covariates were entered into the second block, marital status was still associated with HRQOL. In addition, older age and lower NYHA functional class were associated with better HRQOL.
Results Sample characteristics and their relationships to health-related quality of life The mean age of the 71 patients was 72 years (Table 1). The majority were Caucasians, and they had mild to moderate functional impairment (NYHA functional class II or III). Slightly more than half were males, married or had a cohabitant, and had ischemic HF. Among the sociodemographic and clinical characteristics, marital status, comorbidities, and NYHA functional class were significantly associated with HRQOL in bivariate analyses. Relationships of social support to physical and depressive symptoms and health-related quality of life In the multiple regression analysis, all types of social support were entered into the model at the same time, and only marital status and emotional support were significantly related to physical symptoms (R2 ¼ .292, p < .001) (Table 2). Patients who remained in their marital relationship or had a cohabitant and had greater emotional support had less severe physical symptoms. Marital
Mediating relationships of physical symptoms and depressive symptoms on the relationship between social support and healthrelated quality of life Only emotional support was significantly related to physical symptoms, depressive symptoms, and HRQOL in bivariate analyses. Marital status was associated with physical symptoms and HRQOL, but not depressive symptoms in bivariate analyses. Thus, we explored whether physical symptoms and depressive symptoms mediated the relationship between emotional support and HRQOL, and whether physical symptoms mediated the relationship between marital status and HRQOL. First, we examined the relationship of emotional support to physical symptoms and depressive symptoms. Emotional support was significantly related to both physical symptoms and depressive symptoms in the general linear model analysis (R2 ¼ .568 and 540, respectively, p ¼ .003 and p ¼ .009, respectively, Fig. 2). In the multiple regression analysis, physical symptoms and depressive symptoms were significantly related to HRQOL (R2 ¼ .594, p < .001). Emotional support was significantly related to HRQOL in a simple regression model (R2 ¼ .058, p ¼ .042), but when physical symptoms and depressive
Table 2 Multivariate relationships of social support to physical symptoms, depressive symptoms, and health-related quality of life. Dependent variable
Independent variable
Standardized beta
95% Confidence interval
R2
F
p value
Physical symptoms
Model 1and 2
14.017
