Aging & Mental Health
ISSN: 1360-7863 (Print) 1364-6915 (Online) Journal homepage: http://www.tandfonline.com/loi/camh20
Health-related social control among older men with depressive symptomatology Shahrzad Mavandadi, Natacha Jacques, Steven L. Sayers & David W. Oslin To cite this article: Shahrzad Mavandadi, Natacha Jacques, Steven L. Sayers & David W. Oslin (2015) Health-related social control among older men with depressive symptomatology, Aging & Mental Health, 19:11, 997-1004, DOI: 10.1080/13607863.2014.986646 To link to this article: http://dx.doi.org/10.1080/13607863.2014.986646
Published online: 15 Dec 2014.
Submit your article to this journal
Article views: 64
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=camh20 Download by: [University of Wisconsin Oshkosh]
Date: 05 November 2015, At: 13:50
Aging & Mental Health, 2015 Vol. 19, No. 11, 997 1004, http://dx.doi.org/10.1080/13607863.2014.986646
Health-related social control among older men with depressive symptomatology Shahrzad Mavandadia,b*, Natacha Jacquesa, Steven L. Sayersa,b and David W. Oslina,b a
Mental Illness Research, Education and Clinical Center, Philadelphia VA Medical Center, Philadelphia, PA, USA; bDepartment of Psychiatry, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
(Received 21 January 2014; accepted 28 October 2014) Objectives: Social control attempts, or attempts by social network members to influence a person’s behavior, significantly predict men’s health behaviors and psychological well-being. Despite the fact that depression is associated with compromised interpersonal functioning and poor health behaviors, the association between social control processes and depression has not been studied. Thus, this pilot study explored differential vulnerability to spouses’ social control attempts among older, male primary care patients with varying levels of depression symptom severity and the degree to which these attempts predicted patients’ behavioral and affective responses. Method: Participants included 88 older men referred by their primary care providers for a behavioral health assessment at a Veterans Affairs Medical Center. Data on sociodemographics, depressive symptomatology, health behaviors, spouses’ positive and negative social control attempts, and patients’ behavioral and affective responses to attempts were collected by telephone. Results: The sample was primarily Caucasian (mean age D 65.3 (SD D 8.1) years). Patients’ higher depressive symptoms were significantly associated with positive and negative affective responses to their spouses’ social control attempts. The frequency of control attempts and patients’ behavioral responses, however, were unrelated to patients’ depressive symptoms. Multiple regression models revealed that while spouses’ control attempts were unrelated to patients’ positive behavioral responses, more frequent negative attempts predicted greater negative behavioral responses (e.g., ignoring spouses’ attempts). Moreover, negative control attempts predicted greater negative affective responses (e.g., resentment, sadness). Conclusion: The findings highlight the value of identifying effective social control strategies that maximize positive behavioral change, emotional responses, and health outcomes among older men with depressive symptoms. Keywords: social support; depression; age; health behavior
Introduction Epidemiological work suggests that there is a robust association between depression and greater likelihood of comorbid medical conditions (Ormel et al., 1999; Penninx et al., 1998; Sobel & Markov, 2005; Tunks, Crook, & Weir, 2008), higher rates of medical care utilization and inpatient and outpatient costs (Rowan, Davidson, Campbell, Dobrez, & MacLean, 2002), and increased morbidity and mortality (Katon & Sullivan, 1990). One biobehavioral factor that has been shown to impact the association between depression and medical illness is the extent to which depression influences one’s engagement in health-promoting (e.g., exercise, proper diet, medication adherence) and healthcompromising (smoking, alcohol use) behaviors. In light of the fact that depression is related to compromised affective, cognitive, motivational, and coping processes, it follows that individuals with depression may be more likely to evidence unhealthy lifestyle choices and poor adherence to medical regimens (Sobel & Markov, 2005). Thus, it would be beneficial to determine what factors influence the adoption of and engagement in healthy and unhealthy practices among older patients with depression. The quality of one’s social relationships is an important psychosocial factor that may impact participation in
healthy and unhealthy behaviors. For example, social control, or attempts to influence and regulate another’s behavior, has been shown to significantly impact health outcomes (Rook & Pietromonaco, 1987). Indirect social control occurs when the target internalizes a sense of responsibility to significant others and therefore avoids health-compromising behaviors that might jeopardize certain role obligations (e.g., spouse, friend) (Rook, Thuras, & Lewis, 1990). On the other hand, direct social control occurs when network members prompt or persuade a target person to engage in certain behaviors. Existing literature suggests that direct social control attempts by others predict engagement in fewer harmful (e.g., smoking, poor diet) and more healthy (e.g., exercise, adherence to prescription drugs) behaviors by the target. However, social control attempts also appear to lead to considerable psychological distress (i.e., the ‘dual effects hypothesis’; Helgeson, Novak, Lepore, & Eton, 2004; Hughes & Gove, 1981; Lewis & Rook, 1999). Furthermore, more recent work indicates that additional factors, such as the nature of the social control attempt, might influence the relationship between social control attempts and psychological and behavioral outcomes (Lewis & Butterfield, 2005). For instance, positive attempts at social
*Corresponding author. Email: [email protected] This work was authored as part of the Contributor’s official duties as an Employee of the United States Government and is therefore a work of the United States Government. In accordance with 17 U.S.C. 105, no copyright protection is available for such works under U.S. Law.
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
998
S. Mavandadi et al.
control, such as persuasion and encouragement, are related to increased engagement in healthy behaviors, while negative attempts, such as nagging, are related to greater psychological distress and little behavioral change (Fekete, Stephens, Druley, & Greene, 2006; Lewis & Butterfield, 2005; Lewis & Rook, 1999; Stephens, Rook, Franks, Khan, & Lida, 2010). Social control has not been studied among older adults with significant psychiatric and physical comorbidity, and only a few studies have specifically examined gender differences in social control receipt and response. This is unfortunate given that depression may help determine the frequency and type of social control directed toward an individual, as well as the impact of social control experienced by that person. While there is a robust literature supporting the notion that poor marital quality is associated with greater psychological distress and other adverse health outcomes (Kiecolt-Glaser & Newton, 2001), there also is evidence that depression may have a negative impact on the social interactions of married couples (Coyne, Wortman, & Lehman, 1988; Druley, Stephens, Martire, Ennis, & Wojno, 2003). Accordingly, one might expect that individuals experiencing greater depressive symptomatology are more likely to receive negative forms of social control. Individuals with greater depressive symptoms also might be more likely to respond negatively to both positive and negative social control attempts due to the negative attributional style often observed in individuals with depression (Bradbury, Beach, Fincham, Frank, & Nelson, 1996). It also may be particularly valuable to explore these processes specifically among married older men with depression. Compared to women, men appear to derive greater health benefits from marriage (Kiecolt-Glaser & Newton, 2001). One potential explanation for this disparity may be that women are more likely to exert health-related social control than men. Married men, for example, have been found to receive more health-related social control from their wives than vice versa (August & Sorkin, 2010; Umberson, 1992). Furthermore, spousal social control has been found to be more effective in reducing engagement in risky health behaviors for men relative to women (Westmaas, Wild, & Ferrence, 2002). Thus, the current pilot study sought to understand the role that spousal social control plays in influencing adherence and self-care behaviors among older, male primary care patients referred for a behavioral health assessment following a positive screen. First, we explored differential vulnerability to spousal social control attempts among older men with varying levels of depression symptom severity (i.e., none or mild to severe). Second, we examined the degree to which depression severity moderates the association between the frequency of positive and negative social control attempts and patients’ behavioral and affective responses to those attempts. Given the negative association between depression on marital quality and interaction, we hypothesized that patients with higher depressive symptoms would report more negative and fewer positive social control attempts from their spouses. We also anticipated that relative to patients with milder symptoms, patients with greater depression symptom
severity and those who met criteria for probable depressive disorder would experience significantly greater negative affect and less positive affect in response to both positive and negative social control attempts, and would be less likely to engage in the targeted health behavior(s).
Method Participants and procedures The sample included 88 older, male Veterans receiving primary care at the Philadelphia VA Medical Center and affiliated community-based outpatient clinics who completed a behavioral health assessment by the Behavioral Health Laboratory (BHL) from November 2009 through January 2011. The BHL is an evidence-based, clinical management program that uses a telephone-based structured interview for the identification, screening, assessment, and triage of primary care patients who may be in need of care for behavioral health issues such as depression, anxiety, alcohol misuse, and post-traumatic stress disorder (PTSD) (Oslin et al., 2006). A total of 152 patients completing the BHL assessment who met initial inclusion criteria were referred to the study team and contacted by research staff. Patients were recruited for participation in the current study if they were 55 years of age or older and reported living with a spouse or significant other. Individuals with the following characteristics were excluded: (1) meeting criteria for psychosis, bipolar disorder, post-traumatic stress disorder, mania, panic disorder, drug use, or alcohol dependence, (2) hearing, visual, or other physical impairments leading to difficulty with assessment, and/or (3) poor cognitive functioning, defined by a Blessed Orientation-Memory-Concentration (BOMC) test score of >15 (Kawas, Karagiozis, Resau, Corrada, & Brookmeyer, 1995). Of the patients contacted, 40 (26.3%) refused, 15 (9.9%) were found to be ineligible following further evaluation of inclusion/exclusion criteria, and 97 (63.8%) consented to participate. Of the 97 patients who consented, an additional 9 were administratively withdrawn, yielding a final analytic sample of 88 participants in the research study. Upon giving informed consent, participants were called by the research assistant by phone to complete the study assessment. This assessment battery included measures of psychosocial factors, health behaviors, and social control processes. Each baseline call took approximately 45 minutes to complete. All study procedures were approved by the Institutional Review Board of the Philadelphia VA Medical Center.
Measures Patient background characteristics Sociodemographic variables examined as covariates included age, ethnicity (non-Hispanic White vs. other), and financial status (report adequate finances (have at least ‘enough to get by’) vs. limited finances (‘can’t make ends meet’)). The physical component score (PCS) of the Medical Outcomes Study 12-Item Short Form Health
Aging & Mental Health Survey (SF-12) assessed general functional ability (Ware, Kosinski, & Keller, 1996). Relationship quality was assessed with the Quality of Marriage Index (QMI) (five items), summed, and included in analyses as a continuous variable (Norton, 1983).
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
Depressive symptomatology and diagnosis Severity of depressive symptoms was measured by the 20item Center for Epidemiological Studies Depression Scale (CES-D) (Radloff, 1977). The CES-D has demonstrated good psychometric properties among older adult samples. Responses were assessed using a Likert scale (0 D ‘none/ rarely’, 3 D ‘most of the time’). Item responses were summed and treated as a continuous variable in the analyses (possible range D 0 60; a D 0.93). The Patient Health Questionnaire-9 (PHQ-9), which is used by the BHL clinical program for screening and diagnostic purposes, was used to assess whether patients met criteria for probable minor or major depressive disorder (Kroenke, Spitzer, & Williams, 2001), and was entered as a categorical variable in models testing for moderation effects. Health behaviors Participants were asked to report whether their spouse had ever tried to get them to change one of nine health behaviors (e.g., stop or reduce alcohol use, exercise more, take medications as directed; Tucker & Mueller, 2000). Given the high proportion of patients who reported that their spouses did not try to get them to change any behaviors, for statistical analysis, participants were categorized into two groups based on whether their spouse had targeted any of the nine health behaviors (0 D ‘no’, 1 D ‘yes’). Social control processes The frequency of spousal social control attempts was assessed using the Social Control Tactic Use Scale (Butterfield & Lewis, 2002). Participants who reported that their spouse tried to get them to change their health behavior(s) were asked to think about situations where their spouse was currently trying or had recently tried to influence them to change a health-related behavior (examples such as drink/smoke less, take medication as prescribed, exercise more, and improve diet were provided based on specific patient responses to the health behavior questionnaire). They were then asked to indicate how frequently their spouse had engaged in a list of 28 social control tactics over the course of the last month (1 D ‘never’, 7 D ‘at least once a day’) (Butterfield & Lewis, 2002). Both positive (e.g., encouraging, modeling the behavior; a D 0.86) and negative (coercing, making the patient feel guilty; a D 0.81) tactics were included in the scale. Participants were then asked to rate how often (1 D ‘never’, 5 D ‘very often’) they responded to efforts from their spouses to regulate their behavior (i.e., behavioral response) in the following ways: attempt to engage in the behavior, successfully engage in the behavior, ignore the person, do nothing, do the opposite of what the person
999
wanted them to do, or hide the behavior (Lewis & Rook, 1999). Participants were also asked to rate their affective responses to social control attempts, including both positive (e.g., joy, affection, contentment; a D 0.87) and negative (e.g., sad, irritated, resentful; a D 0.91) affect, using 26 adjectives extracted from commonly used affective scales (e.g., Profile of Mood States, Affect Balance Scale; Curran, Andrykowski, & Studts, 1995; Derogatis, 1975). Analytic plan In addition to descriptive, univariate analyses, zero-order correlations among study variables were computed in order to identify potential covariates for inclusion in subsequent regression models. To address the study’s objectives, we first examined rates of meeting criteria for probable depressive disorder and mean group differences in depressive symptoms among patients who reported that their spouses had recently tried to get them to change at least one of nine health behaviors versus those who did not. Among the subsample of patients who endorsed that their spouses had tried to get them to change their health behavior(s), we examined correlations among depressive symptoms and the frequency of positive and negative social control attempts and behavioral and affective responses to those attempts. We also used one-way analysis of variance (ANOVA) to examine whether social control variables differed across patients meeting criteria for probable minor or major depressive disorder versus those who did not. Given the general lack of association between depressive disorder and other key variables, with the exception of models analyzing interaction effects, subsequent analyses examined only depressive symptoms (i.e., CES-D symptom score). Next, multiple regression models were run to explore the relationship between positive and negative social control attempts and patients’ behavioral and affective responses. To test for moderation, we initially explored whether the association between social control attempt type (i.e., positive vs. negative) and behavioral and affective responses would vary as a function of depressive symptoms and depression diagnosis by including both main effects and depressive symptoms/depression diagnosis £ social control attempt type interaction terms in the models. Due to the lack of statistical significance for interaction terms in any of the models, we next ran models including main effects only. All models controlled for age, relationship quality (i.e., the QMI), and physical functioning (i.e., the SF-12). In order to address the potential confounding effect of depressive symptomatology, multiple regression models also controlled for CES-D symptom scores. Coefficients were tested at the a D 0.05 level. Analyses were conducted using PASW SPSS version 17.0 software. Results Table 1 includes a summary of the samples’ sociodemographic characteristics and means on key study variables. The majority of the sample were non-Hispanic Whites (62.5%). Roughly half (55.7%) were retired and 20.5%
11 10 9 7
0.49 ¡0.03 0.50 ¡0.41 0.12 0.46 ¡0.13
¡0.33 0.32 ¡0.35
6 5 4 3 2
reported limited finances. Moreover, 46.6% of the sample met criteria for minor or major depression (as per the PHQ-9), and patients had a mean CES-D score of 18.5 (SD D 14.5). Twelve (13.6%) patients reported that their spouse did not make any attempts to get them to change their health behavior(s). Subsequent analyses revealed that although patients whose partners did not attempt to change, their behavior did not differ with respect to sociodemographics (e.g., age, ethnicity), depression diagnosis, marital quality, or overall physical functioning; they did have significantly lower CES-D symptom scores (n D 12; M D 6.1, SD D 10.2) than patients who reported that their spouses tried to get them to change at least one health behavior (n D 76; M D 20.4, SD D 14.1; p D 0.001). With regard to the first main objective, analyses of correlations among key variables for patients who reported their spouses had tried to get them to change at least one health behavior (n D 76) suggest that depressive symptoms were unrelated to the frequency of positive or negative social control attempts reported by patients (Table 2). However, as discussed below, adjusted regression models that included age, relationship quality, physical functioning, and positive and negative social control attempts revealed that higher depressive symptoms were related to more negative affect (b D 0.02, SE D 0.01, p D 0.002) in response to social control attempts. Similarly, mean differences in social control variables across patients who did not meet criteria for any depressive disorder vs. those who met criteria for minor or major depressive disorder revealed that patients who met criteria for either minor or major depressive disorder had significantly greater negative affective responses to control attempts (mean D 2.6 (SD D 0.9) vs. 2.0 (SD D 0.8); p D 0.006; Cohen’s d D 0.7, indicating a medium to large effect). Finally, although depressive symptoms were not
8
Note. n D 76. Averages computed for positive (e.g., encouraging, modeling the behavior) and negative (e.g., coercing, making the patient feel guilty) social control attempts. Higher scores denote higher average frequency of positive and negative control attempts.
p < 0.05, p < 0.01, p < 0.001.
1.1 6.6 1.0 6.6
4.0 (1.2) 3.0 (1.4)
0.17 ¡0.01 0.47 ¡0.36 0.42 0.04 0.05 0.02 0.70 0.09 0.45
0 55
0.28 ¡0.06 0.19 ¡0.26 0.25 ¡0.33 ¡0.29 ¡0.32 0.71 0.11
47 (53.4%) 41 (46.6%) 18.5 (14.5)
0.01 0.36 0.34 ¡0.09 0.16 ¡0.03 ¡0.01 ¡0.03 0.56
7 35
0.19 0.12 0.51 ¡0.37 0.40 ¡0.16 ¡0.14 ¡0.15
31.1 (6.4)
12.2 61.9
¡0.21 0.21 0.00 0.29 ¡0.15 0.89 0.62
55 90
¡0.05 0.27 0.15 0.22 ¡0.09 0.91
65.4 (8.1) 88 (100%) 55 (62.5%) 18 (20.5%) 49 (55.7%) 39.6 (12.1)
1. Depressive symptoms 2. Positive social control 3. Negative social control 4. Positive affect 5. Negative affect 6. Positive behavioral response 7. Attempt to change behavior 8. Successfully change behavior 9. Negative behavioral response 10. Hide the behavior 11. Do nothing 12. Ignore spouse/partner 13. Do the opposite
Age (years) Married/partnered Non-Hispanic White Limited finances Retired General physical functioning (SF12 PCS score) Marital quality PHQ diagnosis Non-depressed Minor or major depression Depressive symptom severity (CES-D total score) Social control attempts Positive Negative
12
Mean (SD)/N (%)
¡0.14 0.27 0.08 0.28 ¡0.13
13 Observed range
Variable
Table 2. Correlations among key study variables (n D 76).
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
Table 1. Sample characteristics (n D 88).
0.06 0.04 0.36 ¡0.28 0.25 ¡0.12 ¡0.14 ¡0.08 0.75 0.32 0.37 0.35
S. Mavandadi et al.
1
1000
Aging & Mental Health
1001
Table 3. Behavioral and affective responses to social control attempts (n D 76). Behavioral response Positive Variables
b (SE)
Affective response
Negative p
b (SE)
Positive p
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
Constant 2.19 (1.45) 0.14 1.41 (1.11) 0.21 Age ¡0.01 (0.02) 0.74 ¡0.003 (0.01) 0.83 Marital quality 0.02 (0.02) 0.30 0.01 (0.02) 0.44 General physical functioning (SF12 PCS score) 0.001 (0.01) 0.96 ¡0.01 (0.01) 0.51 Depressive symptom severity (CES-D score) ¡0.01 (0.01) 0.59 0.01 (0.01) 0.29 Negative social control 0.04 (0.11) 0.74 0.37 (0.09)
ISSN: 1360-7863 (Print) 1364-6915 (Online) Journal homepage: http://www.tandfonline.com/loi/camh20
Health-related social control among older men with depressive symptomatology Shahrzad Mavandadi, Natacha Jacques, Steven L. Sayers & David W. Oslin To cite this article: Shahrzad Mavandadi, Natacha Jacques, Steven L. Sayers & David W. Oslin (2015) Health-related social control among older men with depressive symptomatology, Aging & Mental Health, 19:11, 997-1004, DOI: 10.1080/13607863.2014.986646 To link to this article: http://dx.doi.org/10.1080/13607863.2014.986646
Published online: 15 Dec 2014.
Submit your article to this journal
Article views: 64
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=camh20 Download by: [University of Wisconsin Oshkosh]
Date: 05 November 2015, At: 13:50
Aging & Mental Health, 2015 Vol. 19, No. 11, 997 1004, http://dx.doi.org/10.1080/13607863.2014.986646
Health-related social control among older men with depressive symptomatology Shahrzad Mavandadia,b*, Natacha Jacquesa, Steven L. Sayersa,b and David W. Oslina,b a
Mental Illness Research, Education and Clinical Center, Philadelphia VA Medical Center, Philadelphia, PA, USA; bDepartment of Psychiatry, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
(Received 21 January 2014; accepted 28 October 2014) Objectives: Social control attempts, or attempts by social network members to influence a person’s behavior, significantly predict men’s health behaviors and psychological well-being. Despite the fact that depression is associated with compromised interpersonal functioning and poor health behaviors, the association between social control processes and depression has not been studied. Thus, this pilot study explored differential vulnerability to spouses’ social control attempts among older, male primary care patients with varying levels of depression symptom severity and the degree to which these attempts predicted patients’ behavioral and affective responses. Method: Participants included 88 older men referred by their primary care providers for a behavioral health assessment at a Veterans Affairs Medical Center. Data on sociodemographics, depressive symptomatology, health behaviors, spouses’ positive and negative social control attempts, and patients’ behavioral and affective responses to attempts were collected by telephone. Results: The sample was primarily Caucasian (mean age D 65.3 (SD D 8.1) years). Patients’ higher depressive symptoms were significantly associated with positive and negative affective responses to their spouses’ social control attempts. The frequency of control attempts and patients’ behavioral responses, however, were unrelated to patients’ depressive symptoms. Multiple regression models revealed that while spouses’ control attempts were unrelated to patients’ positive behavioral responses, more frequent negative attempts predicted greater negative behavioral responses (e.g., ignoring spouses’ attempts). Moreover, negative control attempts predicted greater negative affective responses (e.g., resentment, sadness). Conclusion: The findings highlight the value of identifying effective social control strategies that maximize positive behavioral change, emotional responses, and health outcomes among older men with depressive symptoms. Keywords: social support; depression; age; health behavior
Introduction Epidemiological work suggests that there is a robust association between depression and greater likelihood of comorbid medical conditions (Ormel et al., 1999; Penninx et al., 1998; Sobel & Markov, 2005; Tunks, Crook, & Weir, 2008), higher rates of medical care utilization and inpatient and outpatient costs (Rowan, Davidson, Campbell, Dobrez, & MacLean, 2002), and increased morbidity and mortality (Katon & Sullivan, 1990). One biobehavioral factor that has been shown to impact the association between depression and medical illness is the extent to which depression influences one’s engagement in health-promoting (e.g., exercise, proper diet, medication adherence) and healthcompromising (smoking, alcohol use) behaviors. In light of the fact that depression is related to compromised affective, cognitive, motivational, and coping processes, it follows that individuals with depression may be more likely to evidence unhealthy lifestyle choices and poor adherence to medical regimens (Sobel & Markov, 2005). Thus, it would be beneficial to determine what factors influence the adoption of and engagement in healthy and unhealthy practices among older patients with depression. The quality of one’s social relationships is an important psychosocial factor that may impact participation in
healthy and unhealthy behaviors. For example, social control, or attempts to influence and regulate another’s behavior, has been shown to significantly impact health outcomes (Rook & Pietromonaco, 1987). Indirect social control occurs when the target internalizes a sense of responsibility to significant others and therefore avoids health-compromising behaviors that might jeopardize certain role obligations (e.g., spouse, friend) (Rook, Thuras, & Lewis, 1990). On the other hand, direct social control occurs when network members prompt or persuade a target person to engage in certain behaviors. Existing literature suggests that direct social control attempts by others predict engagement in fewer harmful (e.g., smoking, poor diet) and more healthy (e.g., exercise, adherence to prescription drugs) behaviors by the target. However, social control attempts also appear to lead to considerable psychological distress (i.e., the ‘dual effects hypothesis’; Helgeson, Novak, Lepore, & Eton, 2004; Hughes & Gove, 1981; Lewis & Rook, 1999). Furthermore, more recent work indicates that additional factors, such as the nature of the social control attempt, might influence the relationship between social control attempts and psychological and behavioral outcomes (Lewis & Butterfield, 2005). For instance, positive attempts at social
*Corresponding author. Email: [email protected] This work was authored as part of the Contributor’s official duties as an Employee of the United States Government and is therefore a work of the United States Government. In accordance with 17 U.S.C. 105, no copyright protection is available for such works under U.S. Law.
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
998
S. Mavandadi et al.
control, such as persuasion and encouragement, are related to increased engagement in healthy behaviors, while negative attempts, such as nagging, are related to greater psychological distress and little behavioral change (Fekete, Stephens, Druley, & Greene, 2006; Lewis & Butterfield, 2005; Lewis & Rook, 1999; Stephens, Rook, Franks, Khan, & Lida, 2010). Social control has not been studied among older adults with significant psychiatric and physical comorbidity, and only a few studies have specifically examined gender differences in social control receipt and response. This is unfortunate given that depression may help determine the frequency and type of social control directed toward an individual, as well as the impact of social control experienced by that person. While there is a robust literature supporting the notion that poor marital quality is associated with greater psychological distress and other adverse health outcomes (Kiecolt-Glaser & Newton, 2001), there also is evidence that depression may have a negative impact on the social interactions of married couples (Coyne, Wortman, & Lehman, 1988; Druley, Stephens, Martire, Ennis, & Wojno, 2003). Accordingly, one might expect that individuals experiencing greater depressive symptomatology are more likely to receive negative forms of social control. Individuals with greater depressive symptoms also might be more likely to respond negatively to both positive and negative social control attempts due to the negative attributional style often observed in individuals with depression (Bradbury, Beach, Fincham, Frank, & Nelson, 1996). It also may be particularly valuable to explore these processes specifically among married older men with depression. Compared to women, men appear to derive greater health benefits from marriage (Kiecolt-Glaser & Newton, 2001). One potential explanation for this disparity may be that women are more likely to exert health-related social control than men. Married men, for example, have been found to receive more health-related social control from their wives than vice versa (August & Sorkin, 2010; Umberson, 1992). Furthermore, spousal social control has been found to be more effective in reducing engagement in risky health behaviors for men relative to women (Westmaas, Wild, & Ferrence, 2002). Thus, the current pilot study sought to understand the role that spousal social control plays in influencing adherence and self-care behaviors among older, male primary care patients referred for a behavioral health assessment following a positive screen. First, we explored differential vulnerability to spousal social control attempts among older men with varying levels of depression symptom severity (i.e., none or mild to severe). Second, we examined the degree to which depression severity moderates the association between the frequency of positive and negative social control attempts and patients’ behavioral and affective responses to those attempts. Given the negative association between depression on marital quality and interaction, we hypothesized that patients with higher depressive symptoms would report more negative and fewer positive social control attempts from their spouses. We also anticipated that relative to patients with milder symptoms, patients with greater depression symptom
severity and those who met criteria for probable depressive disorder would experience significantly greater negative affect and less positive affect in response to both positive and negative social control attempts, and would be less likely to engage in the targeted health behavior(s).
Method Participants and procedures The sample included 88 older, male Veterans receiving primary care at the Philadelphia VA Medical Center and affiliated community-based outpatient clinics who completed a behavioral health assessment by the Behavioral Health Laboratory (BHL) from November 2009 through January 2011. The BHL is an evidence-based, clinical management program that uses a telephone-based structured interview for the identification, screening, assessment, and triage of primary care patients who may be in need of care for behavioral health issues such as depression, anxiety, alcohol misuse, and post-traumatic stress disorder (PTSD) (Oslin et al., 2006). A total of 152 patients completing the BHL assessment who met initial inclusion criteria were referred to the study team and contacted by research staff. Patients were recruited for participation in the current study if they were 55 years of age or older and reported living with a spouse or significant other. Individuals with the following characteristics were excluded: (1) meeting criteria for psychosis, bipolar disorder, post-traumatic stress disorder, mania, panic disorder, drug use, or alcohol dependence, (2) hearing, visual, or other physical impairments leading to difficulty with assessment, and/or (3) poor cognitive functioning, defined by a Blessed Orientation-Memory-Concentration (BOMC) test score of >15 (Kawas, Karagiozis, Resau, Corrada, & Brookmeyer, 1995). Of the patients contacted, 40 (26.3%) refused, 15 (9.9%) were found to be ineligible following further evaluation of inclusion/exclusion criteria, and 97 (63.8%) consented to participate. Of the 97 patients who consented, an additional 9 were administratively withdrawn, yielding a final analytic sample of 88 participants in the research study. Upon giving informed consent, participants were called by the research assistant by phone to complete the study assessment. This assessment battery included measures of psychosocial factors, health behaviors, and social control processes. Each baseline call took approximately 45 minutes to complete. All study procedures were approved by the Institutional Review Board of the Philadelphia VA Medical Center.
Measures Patient background characteristics Sociodemographic variables examined as covariates included age, ethnicity (non-Hispanic White vs. other), and financial status (report adequate finances (have at least ‘enough to get by’) vs. limited finances (‘can’t make ends meet’)). The physical component score (PCS) of the Medical Outcomes Study 12-Item Short Form Health
Aging & Mental Health Survey (SF-12) assessed general functional ability (Ware, Kosinski, & Keller, 1996). Relationship quality was assessed with the Quality of Marriage Index (QMI) (five items), summed, and included in analyses as a continuous variable (Norton, 1983).
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
Depressive symptomatology and diagnosis Severity of depressive symptoms was measured by the 20item Center for Epidemiological Studies Depression Scale (CES-D) (Radloff, 1977). The CES-D has demonstrated good psychometric properties among older adult samples. Responses were assessed using a Likert scale (0 D ‘none/ rarely’, 3 D ‘most of the time’). Item responses were summed and treated as a continuous variable in the analyses (possible range D 0 60; a D 0.93). The Patient Health Questionnaire-9 (PHQ-9), which is used by the BHL clinical program for screening and diagnostic purposes, was used to assess whether patients met criteria for probable minor or major depressive disorder (Kroenke, Spitzer, & Williams, 2001), and was entered as a categorical variable in models testing for moderation effects. Health behaviors Participants were asked to report whether their spouse had ever tried to get them to change one of nine health behaviors (e.g., stop or reduce alcohol use, exercise more, take medications as directed; Tucker & Mueller, 2000). Given the high proportion of patients who reported that their spouses did not try to get them to change any behaviors, for statistical analysis, participants were categorized into two groups based on whether their spouse had targeted any of the nine health behaviors (0 D ‘no’, 1 D ‘yes’). Social control processes The frequency of spousal social control attempts was assessed using the Social Control Tactic Use Scale (Butterfield & Lewis, 2002). Participants who reported that their spouse tried to get them to change their health behavior(s) were asked to think about situations where their spouse was currently trying or had recently tried to influence them to change a health-related behavior (examples such as drink/smoke less, take medication as prescribed, exercise more, and improve diet were provided based on specific patient responses to the health behavior questionnaire). They were then asked to indicate how frequently their spouse had engaged in a list of 28 social control tactics over the course of the last month (1 D ‘never’, 7 D ‘at least once a day’) (Butterfield & Lewis, 2002). Both positive (e.g., encouraging, modeling the behavior; a D 0.86) and negative (coercing, making the patient feel guilty; a D 0.81) tactics were included in the scale. Participants were then asked to rate how often (1 D ‘never’, 5 D ‘very often’) they responded to efforts from their spouses to regulate their behavior (i.e., behavioral response) in the following ways: attempt to engage in the behavior, successfully engage in the behavior, ignore the person, do nothing, do the opposite of what the person
999
wanted them to do, or hide the behavior (Lewis & Rook, 1999). Participants were also asked to rate their affective responses to social control attempts, including both positive (e.g., joy, affection, contentment; a D 0.87) and negative (e.g., sad, irritated, resentful; a D 0.91) affect, using 26 adjectives extracted from commonly used affective scales (e.g., Profile of Mood States, Affect Balance Scale; Curran, Andrykowski, & Studts, 1995; Derogatis, 1975). Analytic plan In addition to descriptive, univariate analyses, zero-order correlations among study variables were computed in order to identify potential covariates for inclusion in subsequent regression models. To address the study’s objectives, we first examined rates of meeting criteria for probable depressive disorder and mean group differences in depressive symptoms among patients who reported that their spouses had recently tried to get them to change at least one of nine health behaviors versus those who did not. Among the subsample of patients who endorsed that their spouses had tried to get them to change their health behavior(s), we examined correlations among depressive symptoms and the frequency of positive and negative social control attempts and behavioral and affective responses to those attempts. We also used one-way analysis of variance (ANOVA) to examine whether social control variables differed across patients meeting criteria for probable minor or major depressive disorder versus those who did not. Given the general lack of association between depressive disorder and other key variables, with the exception of models analyzing interaction effects, subsequent analyses examined only depressive symptoms (i.e., CES-D symptom score). Next, multiple regression models were run to explore the relationship between positive and negative social control attempts and patients’ behavioral and affective responses. To test for moderation, we initially explored whether the association between social control attempt type (i.e., positive vs. negative) and behavioral and affective responses would vary as a function of depressive symptoms and depression diagnosis by including both main effects and depressive symptoms/depression diagnosis £ social control attempt type interaction terms in the models. Due to the lack of statistical significance for interaction terms in any of the models, we next ran models including main effects only. All models controlled for age, relationship quality (i.e., the QMI), and physical functioning (i.e., the SF-12). In order to address the potential confounding effect of depressive symptomatology, multiple regression models also controlled for CES-D symptom scores. Coefficients were tested at the a D 0.05 level. Analyses were conducted using PASW SPSS version 17.0 software. Results Table 1 includes a summary of the samples’ sociodemographic characteristics and means on key study variables. The majority of the sample were non-Hispanic Whites (62.5%). Roughly half (55.7%) were retired and 20.5%
11 10 9 7
0.49 ¡0.03 0.50 ¡0.41 0.12 0.46 ¡0.13
¡0.33 0.32 ¡0.35
6 5 4 3 2
reported limited finances. Moreover, 46.6% of the sample met criteria for minor or major depression (as per the PHQ-9), and patients had a mean CES-D score of 18.5 (SD D 14.5). Twelve (13.6%) patients reported that their spouse did not make any attempts to get them to change their health behavior(s). Subsequent analyses revealed that although patients whose partners did not attempt to change, their behavior did not differ with respect to sociodemographics (e.g., age, ethnicity), depression diagnosis, marital quality, or overall physical functioning; they did have significantly lower CES-D symptom scores (n D 12; M D 6.1, SD D 10.2) than patients who reported that their spouses tried to get them to change at least one health behavior (n D 76; M D 20.4, SD D 14.1; p D 0.001). With regard to the first main objective, analyses of correlations among key variables for patients who reported their spouses had tried to get them to change at least one health behavior (n D 76) suggest that depressive symptoms were unrelated to the frequency of positive or negative social control attempts reported by patients (Table 2). However, as discussed below, adjusted regression models that included age, relationship quality, physical functioning, and positive and negative social control attempts revealed that higher depressive symptoms were related to more negative affect (b D 0.02, SE D 0.01, p D 0.002) in response to social control attempts. Similarly, mean differences in social control variables across patients who did not meet criteria for any depressive disorder vs. those who met criteria for minor or major depressive disorder revealed that patients who met criteria for either minor or major depressive disorder had significantly greater negative affective responses to control attempts (mean D 2.6 (SD D 0.9) vs. 2.0 (SD D 0.8); p D 0.006; Cohen’s d D 0.7, indicating a medium to large effect). Finally, although depressive symptoms were not
8
Note. n D 76. Averages computed for positive (e.g., encouraging, modeling the behavior) and negative (e.g., coercing, making the patient feel guilty) social control attempts. Higher scores denote higher average frequency of positive and negative control attempts.
p < 0.05, p < 0.01, p < 0.001.
1.1 6.6 1.0 6.6
4.0 (1.2) 3.0 (1.4)
0.17 ¡0.01 0.47 ¡0.36 0.42 0.04 0.05 0.02 0.70 0.09 0.45
0 55
0.28 ¡0.06 0.19 ¡0.26 0.25 ¡0.33 ¡0.29 ¡0.32 0.71 0.11
47 (53.4%) 41 (46.6%) 18.5 (14.5)
0.01 0.36 0.34 ¡0.09 0.16 ¡0.03 ¡0.01 ¡0.03 0.56
7 35
0.19 0.12 0.51 ¡0.37 0.40 ¡0.16 ¡0.14 ¡0.15
31.1 (6.4)
12.2 61.9
¡0.21 0.21 0.00 0.29 ¡0.15 0.89 0.62
55 90
¡0.05 0.27 0.15 0.22 ¡0.09 0.91
65.4 (8.1) 88 (100%) 55 (62.5%) 18 (20.5%) 49 (55.7%) 39.6 (12.1)
1. Depressive symptoms 2. Positive social control 3. Negative social control 4. Positive affect 5. Negative affect 6. Positive behavioral response 7. Attempt to change behavior 8. Successfully change behavior 9. Negative behavioral response 10. Hide the behavior 11. Do nothing 12. Ignore spouse/partner 13. Do the opposite
Age (years) Married/partnered Non-Hispanic White Limited finances Retired General physical functioning (SF12 PCS score) Marital quality PHQ diagnosis Non-depressed Minor or major depression Depressive symptom severity (CES-D total score) Social control attempts Positive Negative
12
Mean (SD)/N (%)
¡0.14 0.27 0.08 0.28 ¡0.13
13 Observed range
Variable
Table 2. Correlations among key study variables (n D 76).
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
Table 1. Sample characteristics (n D 88).
0.06 0.04 0.36 ¡0.28 0.25 ¡0.12 ¡0.14 ¡0.08 0.75 0.32 0.37 0.35
S. Mavandadi et al.
1
1000
Aging & Mental Health
1001
Table 3. Behavioral and affective responses to social control attempts (n D 76). Behavioral response Positive Variables
b (SE)
Affective response
Negative p
b (SE)
Positive p
Downloaded by [University of Wisconsin Oshkosh] at 13:50 05 November 2015
Constant 2.19 (1.45) 0.14 1.41 (1.11) 0.21 Age ¡0.01 (0.02) 0.74 ¡0.003 (0.01) 0.83 Marital quality 0.02 (0.02) 0.30 0.01 (0.02) 0.44 General physical functioning (SF12 PCS score) 0.001 (0.01) 0.96 ¡0.01 (0.01) 0.51 Depressive symptom severity (CES-D score) ¡0.01 (0.01) 0.59 0.01 (0.01) 0.29 Negative social control 0.04 (0.11) 0.74 0.37 (0.09)
