Gender Differences in Depressive Symptoms in U.S. Chinese Older Adults

Background: This study aim s to explore gender dif ferences in depressive sym ptoms in U.S. Chinese older adu lts. Methods: Data were from the PINE study , a population-based study of U.S. Chinese older adults aged 60 years and a bove. The PHQ-9 was used to m easure depressive symptoms. Results: Depressive symptoms were m ore prevalent in U.S. Chinese older wo men (59.2%) than in older m en (48.5%). Older women were more likely to present som atic depressive symptoms and to develop moderate to severe depressive symptoms. Older age (r = 0.09, P < 0.001), lower income (r = 0.07, P < 0.01), poorer health status (r = 0.34, P < 0.001), inferior quality of life (r = 0.17, P < 0.001) and worsening health ch anges over the past year ( r = 0.23, P < 0.001) were positively correlated with any depr essive symptom in older wom en. Conclusions: This study emphasizes the need for developing tailored interv entions for depressive symptoms in the subgroup of U.S. Chinese older adults. Further longitudinal studies shoul d be conducted to better understand gender differences in risk factors and outcom es associated with depressiv e symptoms in U.S. Chinese older adults.


Introduction
Depression is a widespread m ental health issue and is predicted to be the second leading cause of disability worldwide by 2020 [1]. Older adults are disproportionally af fected by depressive symptoms [2]. It is estim ated that 15%-20% of older adults suf fer from significant depressive symptoms [3]. The presence of depressive symp toms may result in adverse health outcomes, including physical and cognitive impairment [4] increased levels of social isolation, the occurrence of other m ental health issues such as anxiety and hopelessness [5 ], and elder m istreatment [6,7]. More importantly, depressive symptoms are linked to increased mortality and suicide risks [8,9]. Yet the study of depressive sym ptoms in older adults is fraught with many challenges, such as older adults tending to assume that depression is a part of the normal aging process and thus are less likely to seek treatment in a timely manner. In addition, detecting depressive symptoms in older adults may be very co mplicated, as depressive sym ptoms may be associated with m edical commodities, functional impairments and comorbid dementia disorders [10].
Gender may be one of t he most salient factors th at influence the occurrence, m anifestation and recognition of depressive symptoms. There is a large body of literature suggesting that wom en are more susceptible to depressiv e symptoms than men [11,12]. A variety of psychological factors, socio-cultural factors, and biologi cal factors s uch as hormonal ch anges may contribute to gender differences in rates of depressive sym ptoms. However, prior studies on gender dif ferences in depressive symptoms in late lif e have yielded inconsistent result s, with som e studies showing no gender difference in rates of depressive sym ptoms [13] and others sugge sting that older wom en experienced higher rates of depressive sym ptoms than older m en [14,15]. The variations m ay be attributed to dif ferences in setting s, sample procedures and assessm ent tools. T hus far, gender differences in depressive symptoms among older adults remain less clear.
The study on gender dif ferences in depressive symptoms cannot be separated from the familial and cultural context in which they oc cur. Traditional gender norms may play roles in shaping gender differences in depress ive symptoms. Historically, guided by patriarchal cultural values, Chinese women are expected to be subordi nate to m en and therefore of lo wer social status. The gender inequality in Chinese cultur e has resulted in wo men facing educational, econom ic and health care disadvantages [16]. Thus, social expectations of gender roles in Chinese culture m ay lead to variations in depressive sym ptoms among Chinese older m en and wom en. Prior studies on depressive symptoms consistently demonstrate gender differences in depressive symptoms varied by cultural context [17,18]. Y et existing knowledge of gender dif ferences in depressive sym ptoms among Chinese older adults, especially among U.S. Chinese older adults is scarce.
The Chinese population is the lar gest Asian American subgroup population in the U.S., numbering 4 m illion in 2010 [19] . Acculturation stress brought about by cultural and linguis tic barriers may predispose Chinese older adults to higher risk fo r depressive sym ptoms [20]. Additionally, disparities in health care access along with the influence of traditional values in emotional restraint may prevent Chinese o lder adults from seeking tim ely professional help. Depressive symptoms may be detrimental to individuals and families of U.S. Chinese older adults [10]. Not withstanding the scope and se verity of the issue, there is a dearth of investigations examining gender differences in depressive symptoms in U.S. community-dwelling Chinese older adults. The objectiv es of this study are to: 1) ex amine the gender dif ferences in depressive symptoms in U.S. Chinese older adults ; and 2) explore the corr elates of depressive symptoms in U.S. Chinese older men and older women.

Population and settings
The Population Study of Chines e Elderly in Chicago (PINE) is a community-engaged, population-based epidemiological study of U.S. Chinese older adults aged 60 and over conducted in the greater Chicago area. Briefly, the purpose of the PINE study is to collect community-level data of U.S. Chinese older adults to exam ine the key cu ltural determinants of health and well-being. The project was initiated by a syner gistic community-academic collaboration among the Rush Institute for Healthy Aging, Northwestern University, and m any community-based social services agencies and organizations throughout the greater Chicago area [22].
In order to ensure study relevance to the we ll-being of the Chinese community and increas e community participation, the PINE study implem ented extensive cultura lly and linguistically appropriate community recru itment strategies strictly guided by a community-based participatory research (CBPR) approach. The form ation of this comm unity-academic partnership allowed u s to develop appropriate research m ethodology in accordance with the local Chinese cultural context, in which a community advisory board (CAB) pl ays a pivotal role in pro viding insights and strategies for conducting research. Board m embers were comm unity stakeholders an d residents enlisted through over twenty civic, health, social and advocacy groups, community centers and clinics in the city and suburbs of Chicago. The board works exte nsively with investigative team to develop and examine study instrument to ensure cultural sensitivity and appropriateness.
Over twenty social services agencies, community centers, health advocacy agencies, faith-based organizations, senior apartm ents and social clubs served as the basis of study recruitm ent sites. Community-dwelling older adults who aged 60 years and over and self-identified as Chinese were eligible to participant in the study. Out of 3,542 eligible ol der adults approached, 3,159 agreed to participate in the study, yielding a response rate of 91.9 %.
Our bilingual research team translated the scales into Chinese and back translated it into English. The translations were further sc rutinized by investigator s to ensure content and face validity . The Participant signed a consent form approved by the Institutional Review Board of the Rush University Medical Center prior to the interview. Trained multicultural and multilingual interviewers conducted face-to-face home interviews with participants in their p referred language and dialects, su ch as English, Cantonese, T aishanese, Mandarin, or T eochew dialect. Data were collected using state-of-science innovative web-based software whic h recorded simultaneously in English, Chinese traditional and sim plified characters. This tran sformative technological platform minimized any information that m ay have been "lost in translation", thus provi ding deeper m eaning to the data collected.
Based on the available census da ta drawn from U.S. Census 2010 and a random block census project conducted in th e Chinese comm unity in Chicago, the PINE study is representative of the Chinese aging population in the great er Chicago area with respect to key dem ographic attributes, including age, sex, incom e, education, num ber of children, and country of origin. The study was approved by the Institutional Review Board of the Rush University Medical Center.

2.2.1.
Socio-demographics Basic demographic information included age (in years), sex (female and male), education (years of education com pleted), personal income (0-$4,999 per year; $5,000-$9,999 per year; $10,000-$14,999 per year; $15,000-$19,999 per year; or m ore than $20,000 per year), marital status (married, separated, divorced, or widowed), number of children, number of grandchildren and living arrangement (living alone, living with 1 person, living with 2-3 persons, or living with 4 or more persons). Number of years in the community and years in the U.S. were also assessed in all participants.

2.2.2.
Overall health status, quality of life and health changes over the last year In general, how would you rate your health?" on a four-point scale (1 = poor, 2 = fair, 3 = good, 4 = very good). Quality of life was assessed by asking "In general, how would you rate you r quality of life?" also on a four -point scale ranging from 1 = poor to 4 = very good. Health changes over the last year was measured by "Compared to one year ago, how would you rate your health now?" on a three-point scale (1 = worsened, 2 = same, 3 = improved).

Depressive symptoms
We used the Patient Health Questionnaire (P HQ-9) to assess depressive sym ptoms among Chinese older adults. T he PHQ-9 consists of nine item s assessing e ach of the nine symptoms of depression from the Diagnostic and St atistical Manual of Mental Disorders (DSM-IV) [23] P articipants were asked if they had the following symptoms in the last two weeks: (1) changes in sleep; (2) changes in appetite; (3) fatigue; (4) feelings of sadness o r irritability; (5) loss of interest in activities; (6) inability to experience pleasur e, feelings of guilt or worthlessness; (7) inability to concentrate or make decisions; (8) feeling restless or slowed down; and (9) suicide thoughts. Respondents indicated answers to each question on a 4-point scale ranging from 0 = "not at all" to 3 = "nearly every day." The total score could range from 0-27, with a score of 1-4 indicating m inimal level of depressive symptoms, 5-9 indicating mild level of depressive symptoms, 10-14 indicating moderate level of depressive symptoms and 15 and more indicating severe level of depressive sym ptoms [24]. The PHQ-9 has been validated am ong Chinese Americans and has good inter-ra ter reliability [25]. The standardized Cronbach's alpha of the PHQ-9 in the PINE study was 0.82 [26].

Data analysis
Descriptive statistics were used to s ummarize demographic information of the older m en and women with any depressive sym ptoms. Chi-square statistics and Fisher 's exact test were us ed to compare the prevalence of dif ferent levels of depressive symptoms in each health status group. Pearson Correlation coef ficients and Spearman's rank correlation were calcu lated to exam ine the correlations of socio-demographic and health related factors depressive sym ptoms in older wome n and older men. Analyses were carried out using SAS, Version 9.2 (SAS Institute Inc., Cary, NC).

Socio-demographic and health related characteristics of older women and older men with any depressive symptoms
This study consisted of 1,831 olde r women and 1,328 older m en. More than half of the older women (59.2%) reported having depre ssive symptoms, as compared to 48.5% of the older m en. The characteristics of older women and older men with any depressive symptom are presented in Table 1. Compared to the group of older m en with any depressive symptoms, the group of older women with any depressive symptom had a higher proportion of older adults who had no education at all (10.2% vs. 1.1%, P < 0.001), had more than four children (32.5% vs. 24.9%, P < 0.01), had more than three grandchildren (71.4% vs. 59.2%,P < 0.001), were widowed (38.4% vs. 7.2%, P < 0.001), and lived alone (28.7% vs. 13.0%, P < 0.001).

Gender differences in severity of depressive symptoms
The prevalence of depressive sym ptoms by different levels is presented in Table 3. Significant gender differences were found with respect to the severity of depressive symptoms. Among those who reported any depre ssive symptoms, 66.7% of older men had mini mal depressive symptoms, compared with 60.5% of olde r women. Among thos e who repor ted any de pressive symptoms, moderate to severe depressive symptoms were more prevalent in women (16.6%) than in men (12.0%).

Prevalence of depressive symptoms by health status
The gender specific distribution of overall he alth status, quality of life and health changes in the past year is presented in Table 4. Having any depressive symptom was most prevalent among older adults with poor overall health status. In addition, compared to their male counterparts, a higher percentage of older wom en with poor health status had depressive sym ptoms (85.9% vs. 77.2%, P < 0.01), and in particular, severe depressive symptoms (12.1% vs. 5.9%, P < 0.05).

Note: Percentage represents prevalence of depressive symptoms within each of the health status group.
Likewise, the prevalence of any depressive sym ptoms was highest among female older adults with th e lowest level of quality of l ife. The proportion of older women with severe depre ssive symptoms in those with fair quality of life was significa ntly higher than that of older me n with fair quality of life (4.6% vs. 1.7%, P < 0.01), but no significant gender dif ference of severe symptoms were found am ong those who reported poor quality of life.
Older adults with worsened health status had the highest proportion of people reporting depressive symptom(s). The prevalence of moderate and severe depressive symptoms in older women with worsened health status was all significantly higher than that of older m en with worsened health status.

Correlation of depressive symptoms in older men and older women
The socio-demographic and health related correlates of any depressive symptom among older women is presented in Table 5. Older age (r = 0.09, P < 0.001), lower income (r = 0.07, P < 0.01), poorer health status (r = 0.34, P < 0.001), inferior quality of life (r = 0.17, P < 0.001) and worsening health changes over the past year (r = 0.23, P < 0.001) were positively correlated with any depressive symptom in older women.

Discussion
This study dem onstrates that depressive sym ptoms are m ore prevalent in U.S. Chinese older women than in older men. The prevalence of all depressive symptoms of the PHQ-9, especially somatic symptoms, was higher in older women than in older m en. In addition, older wom en were more likely to report moderate to severe depressive sym ptoms. Older age, lower incom e, poorer health status, inferior quality of life and worsening health status were positively correlated with any depressive symptom in both ge nder groups. H owever, being unm arried and fewer years in the community were positively correlated with any depr essive symptoms in older men but not in older women.
This study represents the first lar ge scale inve stigation on gender dif ferences in depressive symptoms in U.S. Chinese comm unity-dwelling older adults. It lays the groundw ork for a more comprehensive understanding of how depressive sym ptoms in older adults m ay vary across gender and cultural groups. Ou r academic-community partnership and comm unity engagement facilitated the design of culturally and lingu istically appropriate research methods [27]. Due to our CBPR approach, participants may have been more comfortable conversing in their preferred dialects, more trusting of research assistants, and more willing to express emotions and acknowledge their feelings.
This study suggests that depressive symptoms are more prevalent in U.S. Chinese older wom en than in olde r men, consistent with the m ajority of studies of older adults in both comm unity and clinical settings. Using the PHQ-9, a recent study of 1,659 community-dwelling older adults aged 60 to 85 years in German showed that 62.4% of older women had depressive symptoms, compared with 51.1% of older m en [14]. In the U.S., a study of 2,732 older patients aged 60 years and over in the clinical setting reported that 17.8% of the olde r women had life tim e depression, com pared with 9.4% of older m en. The prevalence of current de pression in the older wom en (10.6%) was also higher than in the older men (5.7%) when assessed using criteria from the Diagnostic and Statistical Manual of Mental Disorders, (DSM-III-R) [28].
Our finding is also consistent with studies am ong Chinese older adults, which generally suggest that older women are more likely to present depr essive symptoms. In a study of 1,062 Chinese older adults aged 65 years and older in Singapore, the prevalence of depression in older wom en (6.1%) was slightly higher than that in older men (5.9%) [29]. In a community sample survey of older adults in southern Taiwan, older wom en also reported a h igher prevalence of depression than older m en (21.8% vs. 14.0%) [30]. Several reasons m ay account for gender disp arities in th e prevalence of depressive symptoms among Chinese older adults; one key factor is that older wom en tend to be exposed to more negative life events in later life than older men. For instance, the subordinate role of women in Chinese culture m ay make older wom en more vulnerable to elder abuse, espe cially domestic abuse. Prior studies ha ve consistently dem onstrated the association betw een depressive symptoms and elder abuse [20,31]. W e suspect that violence against women may in part be driving the high prevalence of depressive symptoms in Chinese wom en. In addition, compared with older men, U.S. Chinese older women may encounter greater cultural, language and transportation barriers as well as financial hardship, increasing their risk for experiencing depressive symptoms.
In addition to gender dif ferences in the preval ence of depressive sym ptoms, this study also suggests a tendency for older women to report m oderate to severe depressive symptom s. Similar findings are reported by a study of 50 community-dwe lling Chinese older immigrants in New York, in which older women reported having higher rates of moderate to severe depressive symptoms than older men (4.0% vs. 0%) [32]. Variations in stress coping st rategies may contribute to the observed gender differences in the severity of depressive symptoms. For example, men may be more likely to use problem-focused and rational coping strategies, while women tend to adopt emotion-focused and avoidance coping strategies [33]. Thus, older women's adaptation of rumination as the primary stress coping strategy m ay maintain or worsen symptom s of depression. We are aware that depressive symptoms are more severe in imm igrant women, particularly among those who are separated from their children or grandchildren due to imm igration. Obsessive wo rry about childre n in the home country and longing for fam ily reunification m ay exacerbate d epressive symptoms among U.S. Chinese older women.
This study also supports earlier studies that found that manifestations of depressive sym ptoms differed by gender. Somatic depressive symptoms such as sleeping disturbances or fatigue are m uch more prevalent in older wom en than in older m en. A study using data on m ajor depression from the National Comorbidity Survey found that women exhibited a higher prevalence of somatic depression than men [34]. One explanation may be that women are more likely to assume the primary caregiver role-a role that m ay require substantial physical and em otional demands. In addition to providing care to the spouse, in Chinese culture, older wom en often take on the task of raising grandchildren. Taking care of grandchildren may trigger positive emotional effects such as high levels of self-esteem, but may also exacerbate stress and b urden, resulting in sleep disturb ances or fatigue in older adults. There is also evidence to suggest th at somatic depression in women may be associated with anxiety disorders and chronic pain [35]. Futu re studies should explore the factors associated with somatic depressive symptoms in older adults, and among older women in particular.
In line with prior stud ies, this study suggests that overall health status m ay be an im portant indicator for depressive sym ptoms in older adults. In particular , moderate to severe depressive symptoms was more prevalent in older women with poorer overall health status. This m ay be explained by higher levels of perceived burdens omeness among older wom en who are ill. Due to cultural and language barriers, old er Chinese wom en with illnes s may have to depend on adult children for scheduling doctor 's appointments and taking them to the doctor 's clinic. Such intensive demands on support from adult children m ay result in older wom en's higher levels of perceived burdensomeness and depressive symptom s. The inve rse association between health and depression indicates that depressive sym ptoms may contribute to poorer health outcomes in older adults. The presence of depressive sym ptoms may inhibit older adults' ability to perform physical activities, giving rise to unhealthy beha viors such as poor diet and insuf ficient sleep that may affect the health status of older adults.
Interestingly, marital status is significantly correlated with any depressi ve symptom in men but not with that in wom en, which corresponds to previous studies on gender dif ferences in the association between m arital status and depre ssion. In a study of 3,056 comm unity-dwelling older Dutch individuals, the risk of depressive symptoms for men who were not or were no longer m arried (OR = 2.51, CI = 1.69-3.91) was higher than that for wom en (OR = 1.57, CI = 1.07-2.29) [15]. This is perhaps due to d ifferences in th e nature of social engagem ent among older men and wom en. Women are more likely to develop and maintain close confiding relationships with persons other than the spouse. In contrast, m en may have smaller social networks in later li fe and may depend largely on their spouse's emotional and instrument support, which makes it more difficult for them to adjust to life after spousal death [36]. Si milarly, we found that years in the co mmunity is negatively correlated with any depressive sym ptom in older men but not with that in older wom en. We suspect that men may experience greater financial and social status losses due to relocation. As a consequence, residential transitions m ay have gr eater adverse psychological ef fects on older men than on older women.
The findings of the study should be interpreted with limitations in m ind. First, reporting bias may be present in this study , and m ay vary by ge nder. It is possible that older wom en are more willing to o penly express psychological distress, thereby r esulting in higher ra tes of depressive symptoms. Second, this study did not exam ine gender differences in the link between intergenerational relationships and depressive symptoms. Given the role of family relationship in the well-being of Chinese older adults, it may be of considerable im portance to take into account the contributions of intergenerational relationships when examining gender differences. Third, this study aimed to understand depressive symptom s and we did not delve into the co mplex issue of clinical depression. Future studies should continue to exam ine the optim al cut-off point for clin ical depression among U.S. Chinese older adults and im prove our understanding of the scope of clinical depression in this vulnerable population. Additionally, gender differences may also persist in patterns of health seeking behaviors am ong those with depr essive symptoms. Understanding dif ferences in health seeking behaviors may facilitate the diagnosis of depressive symptoms in U.S. Chinese old er adults. Lastly, this stud y utilizes a cross-sectio nal design, so we could not postulate on potential temporal relationships. Future longi tudinal studies are needed to e xplore risk factors and outcom es associated with depressive symptoms in U.S. Chinese older women and older men.
This study has im portant implications for researchers, mental health professionals, community service workers, and policy m akers. First, commun ity-based mental health services shou ld be tailored to m eet the specific needs of older wo men. In doing so, community or ganizations should: prioritize educational workshops that focus on the relevant knowledge regarding depressive symptoms among older wom en, address violence and abuse closely associat ed with depressive symptoms, and enhance stress m anagement and c oping skills among older adults, especially older women. Second, health care prof essionals should be educated on gender -specific depressive symptoms and the cultu rally effective approaches to tackle the symptoms. Specific attention should be given to those who are with older age, lower income, and poor health status and quality of life.

Conclusion
This study demonstrates that U.S. Chinese older women are more prone to depressive symptoms than older men. Moreover, significant gender differences in the prevalence, sym ptom manifestation, severity, and correlates of depressive sym ptoms exist among U.S. Chinese older adults. This study highlights the need to develop tailored interventions for depressive symptoms in this subgroup of U.S. Chinese older adults. Future longitudinal studies are needed to better un derstand gender differences in risk factors and outcomes associated with depressive symptoms in U.S. Chinese older adults.