Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
A study investigating the moderating effect of social self-efficacy on the association between perceived social support and anxiety in young adults (18-29 years old) who had a relative admitted to the ICU in the past 18 months. The study also examines the relationship between time since discharge, anxiety, social support, and social self-efficacy.
What you will learn
Typology: Study Guides, Projects, Research
1 / 46
This page cannot be seen from the preview
Don't miss anything!
Running head: MODERATION OF SOCAL SELF-EFFICACY ON ANXIETY AND SOCIAL SUPPORT 1
Bachelor Thesis 10 - 03 - 2021 Lena Fitzian s 1st Supervisor: Jorinde E. Spook, PhD 2nd Supervisor: Britt E. Bente, MSc B.Sc. Psychology Faculty of Behavioural, Management and Social Sciences (BMS) University of Twente, Enschede, Netherlands
Abstract Background: The Intensive Care Unit is one of the places where family members suffer the most. Many of them develop symptoms of the Post Intensive Care Syndrome – Family (PICS-F) lasting after the release from the ICU. Anxiety is one of the most prevalent symptoms of PICS-F and one of the main health related threats to young adults. Therefore, this study focused on anxiety in people between the ages of 18 and 29 filling the research gap on this target group and extending the current insights to 18 months following the release of the relative from the ICU. Furthermore, social support has been named one of the main coping resources with social self-efficacy influencing someone’s perceived social support. It was expected that symptoms of anxiety decrease with time progressing. Further, social self-efficacy was expected to moderate the association between social support and anxiety. Method: A convenience sample of 42 participants was included in the analysis with 69% being female and the majority from Germany ( M age = 22.76). A survey was administered using the anxiety subscale of the Hospital Anxiety and Depression Scale (α=.71), the Multidimensional Scale for Perceived Social Support (α=.90), and the social subscale of the Self-Efficacy Scale (α=.59). Afterwards, the association was tested using a linear regression model whereas the moderation analysis was assessed using a multiple regression model. Results: No significant association between time and anxiety could be noted ( p =.44). Additionally, social self-efficacy was not found to be a significant moderator of the association between social support and anxiety ( p =. 51 ). Conclusion: The study helped contribute to the little body of information available on the topic for this specific target group. It indicated that young adults may differ in their utilization of social support and could benefit more from receiving additional support that is not part of their usual social surrounding. Therefore, suggestions for future studies, such as focusing on specific personalities and different types of social support, may help develop more tailored interventions to prevent PICS-F using social support. Keywords: Post Intensive Care Syndrome – Family, Anxiety, Social Support, Social Self- efficacy, Social Cognitive Theory, young Adults
Introduction Background Due to the COVID-19 pandemic starting in March of 2020, a stark increase in the number of intensive care unit (ICU) patients has been noted throughout the Netherlands. Even though the numbers are generally decreasing again, consequences might linger on for longer (Stewart, 2020). Oftentimes, the ICU environment comes with increased stress levels, not only having an impact on the patients but also on the family members (Davidson, Jones, & Bienvenu, 2012). Harvey (1998) stated that the ICU is one of the places where the family members of patients suffer the most. Two thirds of all family members experience symptoms of posttraumatic stress, depression, or anxiety (Harlan et al., 2018). However, the impact does not vanish when the patient is released from the hospital or even dies. Symptoms can linger on or arise afterwards. About half of the family members noted symptoms for months after the stay, known as the post-intensive care syndrome – family (PICS-F), having a great influence on their daily lives (Harlan et al., 2018). As Pochard et al. (2005) have found symptoms of anxiety in around 75% of family members of ICU patients, decreasing their quality of life (Fridriksdottir, et al., 2010), it will be the focal point of this study. Many coping variables associated with anxiety have already been researched, such as social support (Davidson et al., 2012). Based on a finding by Carmeli, Peng, Schaubroeck, & Amir (2020), the already established association of social support and anxiety will be extended by including social self-efficacy. Therefore, this study fills the research gap and contributes to the understanding of the PICS-F. PICS-F and Anxiety Davidson et al. (2012) describe the PICS-F as a cluster of psychological outcomes due to exposure to critical care. This cluster consists of symptoms of anxiety, acute stress disorder, posttraumatic stress, depression, and complicated grief. Symptoms can remain for up to four years after the discharge of the patient from the hospital and affect the ability to execute care-giving functions that might be necessary. Additionally, many of the disorders are comorbid. For example, severe PTSD is associated with an increased prevalence of anxiety as well as depression (Pochard et al., 2005). Even though the previous findings were based on family members of patients at the day of the release from the hospital, previous studies have already shown that symptoms can remain.
anxiety and depression or a family history of a mental illness are risk factors (Hettema, Prescott, Myers, Neale, & Kendler, 2005). Furthermore, the preferred decision-making role has an influence on the prevalence of anxiety (Anderson et al., 2009; Davidson et al., 2012). Next, symptoms are more common when felt that information on the patient’s condition is incomplete (Azoulay, 2005, as cited in Davidson et al., 2012). For example, people were less likely to show symptoms if they felt that the staff delivered all important information (Anderson et al., 2009). Finally, many studies have mentioned perceived social support as an important influence in the development of anxiety. Perceived Social Support as a Coping Resource Social support can be a coping resource people use when confronted with stressors. It is defined by Cobb (1976, p. 300) as “information leading the subject to believe that he is cared for and loved, esteemed, and a member of a network of mutual obligations” (as cited in Carmeli et al., 2020). It has been shown that relationships can give people social support and contribute to their mental and physical health and help reduce the likelihood of developing depression and anxiety after negative events (Thoits, 1995). Further, Anxiety levels of family members of former ICU patients have been shown to negatively correlate with perceived social support (Davidson et al., 2012). Family members of former ICU patients perceive different amounts of social support. Those who experienced more social support had a lower-state anxiety (Davidson et al., 2012). This goes along with a finding by Harlan et al. (2018) specifying six main coping strategies family members used, one of which is seeking support. Relatives seem to have the need to receive the help or emotional support of others. This can be in the form of spiritual support, help from medical personnel but also the comfort of a friend or just general contact with others. Additionally, Carmeli et al. (2020) note that a large body of research can be found emphasizing the role of perceived social support in mitigating potential outcomes of stressors such as symptoms of anxiety. They conducted a study among older students which showed that social support influences vitality and mental health. However, they also note an interaction of social self-efficacy and social support. Pursuing this finding, the social cognitive theory by Bandura can be employed to explain the interaction of perceived social support and social self-efficacy.
Social Self-efficacy through the Social Cognitive Theory Following Bandura’s social cognitive theory of 1977, the interaction of social support and self-efficacy can explain symptoms of anxiety. The social cognitive theory explains that behaviour is learned through someone’s social surrounding, cognition, and previous experiences. (Bandura, 1977, as cited in Kassin, Fein, & Markus, 2017). It emphasizes the role of cognition in someone’s ability to assess a situation, perform a behaviour and construct the reality including concepts like social self-efficacy (Bandura, 2002, as cited in Riaz Ahmad, Yasien, & Ahmad, 2014). Self- efficacy is a person’s belief that they have the capability to conduct a certain task or action (Bandura, 1995). Therefore, social self-efficacy can be defined as the “confidence of the individual to participate and engage in social interactions” (Carmeli et al., 2020, p. 352). Someone’s self- efficacy can strongly affect their overall functioning. It is needed to develop new subskills and adapt to a changing environment or new situations that may be stressful, or unpredictable, such as the admission of a relative to the ICU (Bandura, 2002, as cited in Riaz et al., 2014). Someone with high social self-efficacy is more likely to display behaviour that would lead to more perceived social support. First, someone with high social self-efficacy can approach others leading to a greater support system that can help them deal with their environment. For example, Riaz Ahmad et al. (2014) state that people with a higher perceived social self-efficacy are more likely to have a strong support system. Second, people with higher social self-efficacy are more likely to approach their social surrounding when they need support, therefore perceiving more support. Carmeli et al. (2020) have shown that high social self-efficacy is positively correlated to confidence asking for social support (Carmeli et al., 2020). Thus, the theory can explain how someone’s social self-efficacy and social behaviour interact in dealing with environmental factors. Based on the social cognitive theory, it can be expected that social self-efficacy has a role in the association of perceived social support and anxiety. Someone with high social self-efficacy can find and perceive the support they need to deal with the admission of a relative to the ICU. Therefore, symptoms of anxiety may be reduced. For example, Carmeli et al. (2020) have shown that the association of perceived social support and vitality is moderated by social self-efficacy, especially for people scoring low on social self-efficacy. They have more trouble utilizing their social surrounding and therefore benefit more from receiving additional social support, i.e., the relationship between social support and vitality is stronger (Carmeli et al., 2020). As an association between social support and anxiety has already been established, it can be expected that social self-
Methods This study was part of a larger cooperation of six bachelor students gathering data together for the aim of writing their bachelor thesis. While all students have the same overall topic of mental or physical health after the stay of a relative in the ICU, the different aims are examined independently. The variables in the full survey study include symptoms of anxiety and depression, the quality of life, stress, sleep disturbances, eating patterns, social support, flourishing, social self- efficacy and coping strategies. For this study, the focus is only on symptoms of anxiety, social support, and social self-efficacy. Design A survey design was used to conduct a between-subjects, cross-sectional study to meet the aim of assessing whether social self-efficacy moderates the relationship between social support and anxiety. Before data collection began, the questionnaire was approved by the ethics committee of the BMS faculty at the University of Twente (approval number: 210239). The data collection took place over the course of one month starting in April 2021. The final dataset was downloaded on the 18th of May 2021. Participants To maximize the number of participants, people were recruited using convenience sampling with the advantage of having a possible snowballing effect. The demographics of the participants are listed in Table 1. In total, 42 participants filled out the relevant questionnaires meeting the inclusion criteria of having had a relative in the ICU within the last 18 months, being within the age group and finishing all necessary questionnaires. Of the final participants, the majority was female and from Germany ( M age = 22.76, SD = 2.46). Furthermore, the majority indicated that their relative in the ICU was their grandparent.
Table 1 Sample Characteristics (n= 42 ) Characteristics N (%) M SD Min Max Total sample 42 (100) Gender Female Male non-binary/ third gender Prefer not to say
Nationality Dutch German Other
Age 22.76 2.46 18 29 Time since the release (in months)
The relative is the participant’s Child Parent Grandparent Sibling Aunt/ Uncle Cousin Other
Length of the ICU stay (in days) <2 4 (9.5)
often the relative stayed in the ICU and whether they are still in need of receiving care (for the form see Appendix A). The Hospital Anxiety and Depression Scale (HADS). Second, the hospital anxiety and depression scale (HADS) by Zigmond and Snaith (1983) was used to assess the state of anxiety of family members who previously had a relative in the ICU. The questionnaire consists of 14 items, seven items relating to depression (α=.83) and seven relating to anxiety (α=.84), scored on 4-point Likert scales (0 = not at all to 3 = most of the time ) (Dagnan, Chadwick, & Trower, 2000, as cited in Bjelland, Dahl, Haug, & Neckelmann, 2002). Both subscales are assessed separately. Therefore, this study can make use of only the anxiety subscale. An example item is “I get sudden feelings of panic”. Bjelland et al. (2002) reviewed several studies concluding a good to very good validity. This is supported by factor analyses indicating two underlying factors. Additionally, the Cronbach’s alpha was at least .60 in all studies, showing very good internal consistency in most cases (Tavakol and Dennick, 2011). The score of a participant was assessed by summing up all outcome variables on the scale. This score can be categorized as normal (between 0 and 7), as borderline abnormal (between 8 and 10) or as an abnormal case (between 11 and 21). The Multidimensional Scale of Perceived Social Support (MSPSS). Third, the multidimensional scale of perceived social support was used (MSPSS; Zimet, Dahlem, Zimet, & Farley (1988)). The MSPSS is a three-factor model consisting of 12 items rated on a 7-point Likert scale (1 = very strongly disagree to 7 = very strongly agree ). Each factor is scored on four items respectively that measure the perceived social support provided by family (α=.91), friends (α=.89) and a significant other (α=.91) (Canty-Mitchell and Zimet, 2000). Example items for each subscale are “my family really tries to help me”, “I can talk about my problems with my friends” and “there is a special person who is around when I am in need”. The sources providing social support used in the questionnaire are designed to allow the participants interpretations that are most relevant to them. For example, a “significant other” can be interpreted as a romantic relationship but can also be a teacher, a priest or any other person that is important to the participant (Zimet et al., 1988). Overall, the scale showed a coefficient α of .93. The factor structure, reliability and validity have been demonstrated several times across many populations and age ranges including university students which are approximately the same age as the target group which are also included in this study. Additionally, the questionnaire has been shown to have an excellent internal validity across
different subgroups (Canty-Mitchell and Zimet, 2000). Thus, it represents a great fit for this study. The sum of all items represents the final score of an individual. The Self-efficacy Scale (SES). Lastly, the self-efficacy scale (SES) by Sherer et al. (1982) was used. However, as it consists of two subscales, general self-efficacy (α=.86) and social self- efficacy (α=.71), only the latter is assessed in this study. It has been later adapted to consist of six items rated on a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree ) (Sherer and Adams, 1983). Three of the six items are negatively formulated and must be recoded. An example item for a positively framed item is “when I am trying to make friends with someone who seems uninterested at first, I do not give up easily”. An example item for a negatively frame items is “I do not handle myself well in social gatherings”. After the recoding, a total score can be calculated by summing up all the elements, with a high score indicating more social self-efficacy expectations. Here, 30 represents the highest possible score and 5 the lowest. In a follow-up study, Sherer and Adams (1983) assessed the subscale among 101 subjects measuring a mean score of 21. ( SD =3.63). They documented the criterion validity by showing that the social self-efficacy subscale has predicted past vocational success and construct validity by comparing it to already established measures (Sherer et al., 1982). All questionnaires can be found in their entirety in Appendix B. Procedure Participants were recruited either by the researchers providing them with a link or via Sona- System and SurveyCircle redirecting them to the questionnaire. After giving informed consent (see Appendix A), they answered a few questions regarding demographics. Afterwards, the described scales were administered. First, participants filled in questions regarding their mental Health and later regarding their physical health. At the end of the survey, they received further information on the specific theses and were thanked for their participation. For people coming from the SurveyCircle website, a code was displayed to redeem points. Overall, filling in all questionnaires took about 25 minutes. Data Analysis SPSS version 27 was used to conduct the following analyses. To begin with, participants that failed to meet the inclusion criteria were removed from the final dataset. Participants were removed from the dataset when they indicated to not have had a relative in the ICU, if they did not
anxiety as the dependent variable. The hypothesis would be accepted when the interaction effect is significant ( p < .05). Results Stripping of Data Before the actual testing of the hypotheses, the data was stripped and explored. Table 2 shows the participants that had to be removed due to the exclusion criteria. It can be seen that more than half of the participants had to be excluded from the original dataset. Afterwards, the data of the remaining participants ( n =53) was examined further. Table 2 Initial stripping of data (N=129) Exclusion Criteria N Did not finish demographics/ consent 48 Indicated to have no relative in the ICU 24 Number of stays: zero 2 Stay more than 18 months ago 2 Of the remaining participants, four had to be removed after an exploration of the data identified them as outliers (see Appendix D, for boxplots). Additionally, it must be noted that of the remaining participants, twelve did not finish the entire survey. To explore possible reasons, a post-hoc analysis using an independent t - test was conducted comparing the participants who did not finish the questionnaire ( N = 12 ) to those who did ( N= 39 ). However, the scales were tested for normality for both groups first. The Shapiro-Wilk test indicated that the mean scores of the HADS subscale were normally distributed for the group that finished the survey [ W (37) = .98, p = .57] and for the group that did not finish [ W (12) = .95, p = .60]. Similar results were found for the mean scores of the social self-
efficacy scale. The data of the group that answered all questions was also normally distributed [ W (37) = .97, p = .45] as well as for the group that had missing items [ W (8) = .93, p = .48]. However, results for the MSPSS scale indicated that for participants that finished the survey [ W (37) = .92, p = .01] and for participants that did not finish [ W (9) = .81, p = .03] the null hypothesis that the data is distributed normally had to be rejected. These variables had to be transformed using independent log transformation to account for the sensibility of the t - test to non-normality. Afterwards, the t - test was conducted. Table 3 shows that no significant difference in means was measured. As the sample sizes differ significantly, Levene’s test of homogeneity of variances was run as well. The results indicate that the null hypothesis of equal population variances can be accepted for the relevant variables of anxiety [ F (1,47) = 3.17, p = .08], social support [ F (1,44) = .02, p = .90] and self-efficacy [ F (1,43) = .55, p = .46]. Therefore, the assumption of homogeneity of variance is met for all variables. Based on these results indicating no significant difference between the two groups, participants who filled in all relevant questions for the hypotheses were included in the final dataset ( n =42). Table 3 Results of t-test for Equality of Means Finished Dropped Out M SD M SD t - test Anxiety 19.91 3.06 11.25 1.96. Logarithmic social support 1.85 .07 1.86 .07. Social self-efficacy 19.05 3.80 17.00 5.01 - 1.
Table 5 Correlation Matrix between main variables and potential covariates N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Validity and Reliability of Scales To conclude the exploration of data, the validity and reliability of the three scales was assessed. The Bartlett’s test of sphericity indicated that overall, the correlations within the correlation matrix were significant for anxiety [ χ² (91)=224.12, p<.001], social support [ χ²( 66)=487.80, p<.001] and social self-efficacy [ χ² (15)=39.84, p<.001]. The results of the Kaiser- Meyer-Olkin test for sampling adequacy indicated that the strength of the relationship among variables was mediocre for the variables anxiety ( KMO =.79), social support ( KMO =.76) and social self-efficacy ( KMO =.62). This indicated a sufficient fit of the factor model for this sample. Therefore, the factor analyses were run for all three scales. First, for the HADS, the scree plot and eigenvalue criterion ( ev ≥ 1) indicated four factors. Item four, “I can sit at ease and feel relaxed”, of the anxiety subscale loaded negatively on one of the factors despite being formulated positively and was therefore excluded from the final analyses. Second, scree plot and eigenvalue criterion indicated the expected three factors for the MSPSS. Together all three factors accounted for 84.38% of the variance with each one accounting for approximately the same proportion. Lastly, scree plot and eigenvalue criterion of the SES subscale indicated two factors instead of one. The factor analysis yielded that the first factor accounts for 35.38% of the variance while together both factors explain 57.78%. However, all items loaded correctly with the already reversed items loading positively. The results of the reliability scores differed among the scales. To begin with, the reliability test for the HADS was run twice. The first test was run with the entire scale, indicating no sufficient result (α=-.16). However, the second time it was run without the previously mentioned item four that falsely scored negative. Now, Cronbach’s alpha indicated better reliability (α=.71). Additionally, while the Cronbach’s alpha score was high for the MSPSS scale (α=.90), the score for the SES subscale was low and barely acceptable (α=.59). Testing of Hypothesis 1 Before running the moderation analysis, the first hypothesis was tested. Table 6 displays the ANOVA between time since discharge and symptoms of anxiety. Results show that the first hypothesis expecting a negative association between time and symptoms of anxiety in family
