Background
Depression is reportedly the most common mental disorder following stroke [
1‐
4], with an incidence ranging from 10 to 64% [
4‐
6]. Poststroke depression has an adverse effect on functional recovery and increases the mortality rate [
7]. Furthermore, the rehabilitation of depressed stroke patients is more difficult because poststroke depression is associated with disruption of daily activities, function, and quality of life [
8,
9].
Depressive symptoms may be a psychological consequence of low levels of physical functioning and different life experiences after a stroke [
10]. Indeed, one study found that there was no significant association between depression levels and the pathological development of stroke such as type of stroke, side of stroke, and comorbidity [
10]. Furthermore, a study on comorbidity and primary care costs found that depression was a stroke comorbidity associated with increased costs [
11].
Demographic characteristics are one of factors that affect depressive symptoms in stroke patients. A meta-analysis in stroke patients reported that education level, income, and age showed significant effects on depressive symptoms [
12]. Other studies have found that depressive symptoms are significantly related to presence of a spouse [
13] and higher levels of education [
14], while low economic status has also been associated with a higher prevalence of depression at any level of morbidity [
15]. Stroke patients were also shown to receive care and emotional and physical support from their family members [
16]. Support from the family was a protective factor affecting poststroke depression [
17]. Indeed, the severity of depressive symptoms has been found to be associated with lower levels of certain relational variables, such as social support and satisfaction thereof [
18]. The literature has shown that depressive symptoms can be reduced through interventions that aim to improve the self-concept and self-esteem, such as positive thinking, as well as social/family support [
19‐
21]. Functional independence was also found to be a crucial factor contributing to the level of depressive symptoms [
22]. Overall, the literature has found evidence for relationships between gender, age, education level, presence of a spouse, income status, and poststroke depressive symptoms.
Demographic characteristics including gender roles, family role, caregiving support, and socioeconomic status make up a shared environment and are interdependent and interaction dynamics [
23‐
25]. There is, however, some controversy regarding the relationships between demographic characteristics and depressive symptoms: namely, one study found no significant differences or correlations between depression score and these demographic characteristics [
17]. Given the chronic nature and long-term care after stroke, it seems necessary to precisely understand how depressive symptoms result from this pathology, which can be done through attending to the moderating effects of demographic variables on psychological status.
To our knowledge, little is known about the potential role of demographic factors in depression following a stroke. The aim of our study was to evaluate the interaction of demographic characteristics including gender, age, education level, the presence of a spouse, and income status on depressive symptoms after stroke and to identify groups of patients that may be at risk of depressive symptoms after stroke.
Discussion and conclusion
In this study, we investigated the relationships of various demographic characteristics—including gender, age, education level, income status, and the presence of a spouse—with depressive symptoms among stroke patients, as well as the interaction effects of these variables.
In this study, the depressive symptoms scores of women living with a spouse were lower than were those of women without a spouse. By contrast, the difference in depressive symptoms scores according to the presence of a spouse was trivial among male patients. Past studies have similarly shown gender-based differences in stroke outcomes, including daily activity independence and quality of life [
35]. Female stroke patients appear to be less likely to achieve independence in daily activities and have poorer physical, cognitive, and emotional functioning (manifesting as problems such as thinking difficulties, language difficulties, and less energy) after discharge [
36]. Considering that spouses are the primary caregivers for most patients are physically dependent on others [
37], it is possible that there are partner effects on emotional functioning, such as the development of depressive symptoms, among female stroke patients.
Family support has been reported as a protective factor for major depression in chronic diseases such as cancer [
38]. Research has similarly shown that depressive symptoms are negatively correlated with social support [
17]; thus, it is logical to suggest that stroke patients who do not have a spouse might be more depressed than those with a spouse. In terms of the social contextual model, which accounts for the transactional and interdependent nature of social relationships, a diagnosis of chronic illness might lead to changes that influence one’s partner as well [
23‐
25]. However, this model cannot explain why depressive symptoms were found among male stroke patients who had a spouse in our study. There are studies suggesting gender differences in partner effects [
13]. Thus, our results are also supported by a study reporting that there are no partner effects in male patients. In other words, husbands’ depressive symptoms do not appear to be influenced by the wives’ characteristics, whereas the husband’s characteristics can contribute to women’s somatic symptoms and depressive symptoms [
13].
In our results, the depressive symptoms of stroke patients with an education level above high school and insufficient income were higher than were those in the sufficient income group. Education has been found to be negatively associated with general depressive symptoms [
13,
14,
39,
40], likely because of its relationship with future income, socioeconomic status, and life satisfaction [
41]. However, education did not have a protective role in the insufficient income group; it was only protective in the sufficient group in our study. This might be because material deprivation is an important variable in stroke patients. Patients with stroke are long-term users of health services and typically have high health care costs [
42]. Low income has been found to be associated with less participation following stroke [
43]. Additionally, around 20% of stroke survivors are unemployed post-stroke and around half must change jobs [
37]. Thus, under conditions of material deprivation, higher levels of education do not guarantee higher income. Because there was a strong association between low income and depression among patients [
15], further studies should be conducted to examine economic status changes after a stroke.
Note that this study employed a subjective measure of income status. Subjective income status is an integral aspect of one’s economic well-being because it can improve one’s assessment of one’s capacity to meet financial needs, including maintaining independent community-based living [
44]. Perceived income adequacy has been reported in the literature in various ways and has been verified as a predictor of other outcome measures such as self-rated health [
45], life satisfaction [
46], and depressive symptoms [
47].
In the same line of reasoning, with insufficient income, the presence and support of a spouse did not have positive effects on depressive symptoms in stroke patients. In our study, the depressive symptoms scores of stroke patients living with a spouse in the insufficient income bracket were higher than were those living without a spouse. Overall decline in social function and burden for both the stroke survivors and their caregivers [
37] would be much greater among those with a low income status.
Stroke is a complex condition that can have multiple comorbidities [
48]. Because managing complex patients requires greater clinical effort, a better understanding of this complex patient population is needed [
49]. The complexity of these patients increases the need for healthcare resources and substantial family and community support; in particular, healthcare systems and services need to be redesigned to better meet the needs of these patients [
49]. Personal characteristics, social determinant factors, and social/family support have been reported as some of the elements that contribute to the complexity of stroke patients [
48]. Because depressive mood leads to a lowered likelihood of help-seeking intention [
50], identifying the risk factors of depressive symptoms after a stroke is necessary. This study showed that assessment of risk factors should consider the interactions between gender, economic status, education level, and presence/absence of a spouse. This would also be relevant for interventions aiming to reduce depressive symptoms during stroke rehabilitation.
The results of this study help in comprehensive understanding of the importance of screening for and treating depressive symptoms during rehabilitation after stroke. In particular, our results showed that the researcher took into account the interaction between general characteristics such as gender, socioeconomic status, and the presence/absence of a spouse. It might be noted, however, that researchers must be cautious in generalizing our findings, as the sample may not be fully representative of all stroke patients, especially those who are institutionalized or have cognitive deficits. Furthermore, our sample size was rather small, which increases the likelihood of spurious associations among a large number of interactions. Investigating such a large number of potential interactions indeed presents a statistical challenge for studies with relatively small sample sizes. Moreover, the probability of making a type II error could increase due to lack of adjustment for multiple comparisons. There was a need to be cautious when interpreting these results from the point of view that chance could increase with each subsequent test. Finally, the data were all collected through a self-report questionnaire, and some of the variables, in particular income status, were subjectively measured. These facts suggest that the validity of our results might have been influenced by social desirability bias and intrinsic self-reporting bias.