Women's status and depressive symptoms: A multilevel analysis

Abstract

The effects of state-level women's status and autonomy on individual-level women's depressive symptoms were examined. We conducted a multi-level analysis of the 1991 longitudinal follow up of the 1988 National Maternal Infant Health Survey (NMIHS), with 7789 women nested within the fifty American states. State-level women's status was assessed by four composite indices measuring women's political participation, economic autonomy, employment & earnings, and reproductive rights. The main outcome measure was symptoms of depression (Center for Epidemiologic Studies Depression Scale, CES-D). The participants were a nationally representative stratified random sample of women in the USA aged between 17 and 40 years old who gave birth to live babies in 1988, were successfully contacted again in 1991 and provided complete information on depressive symptoms. Women who were younger, non-white, not currently married, less educated or had lower household income tended to report higher levels of depressive symptoms. Compared with states ranking low on the employment & earnings index, women residing in states that were high on the same index scored 0.85 points lower on the CES-D (p<0.01). Women who lived in states that were high on the economic autonomy index scored 0.83 points lower in depressive symptoms (p<0.01), compared with women who lived in states low on the same index. Finally, women who resided in states with high reproductive rights scored 0.62 points lower on the CES-D (p<0.05) compared with women who lived in states with lower reproductive rights. Gender inequality appears to contribute to depressive symptoms in women.

Introduction

The World Health Organization (WHO) ranks depression fourth among health problems that contribute to the global burden of disease (World Health Organization, 2001). Epidemiological and clinical studies have consistently shown that depression is two to three times more prevalent among women than men (Horwath & Weissman, 1995). In the Epidemiological Catchments Area Study (ECA), the female to male ratio of the lifetime prevalence of major depression was approximately 2.7:1 and the annual incidence rate ratio was approximately 1.8:1 in the United States (Horwath & Weissman, 1995). Furthermore, in the US, women are twelve times more likely to take antidepressant medication than men (Rousseau, 2000).

There are several contending theories that seek to explain women's over representation in the prevalence and incidence of depression (Piccinelli & Wilkinson, 2000). Starting from different assumptions about gender/sex differences in health and disease, researchers have employed a variety of “lenses” to examine women's excess of depression (Tesh, 1988; Walsh, Sorenson, & Leonard, 1995). At the most micro level, the “biomedical lens” focuses on genetic differences, sex differences in the dysregulation of neurotransmitters, hormone imbalances and other sex differences in biology (Ussher, 1992; Walsh et al., 1995). By contrast, the “psychosocial lens” attributes sex differences to the contribution of psychology, measured by such factors as differential help-seeking behaviors, coping style and self-efficacy.

Moving to a higher level of analysis, the “epidemiological lens” attributes gender differences to the population distribution of risk factors for depression, such as exposure to social isolation, intimate partner violence, and childhood physical and sexual abuse. At the most macro-level of analysis, Walsh et al. propose a “society and health lens” that explores the broader socio-cultural, political and economic processes that shape and determine the distribution of power and resources between men and women, and how these social processes in turn determine gender differences in health (Walsh et al., 1995). Specifically, the society and health lens traces the “social roots”,—such as, gendered social stratification, gendered division of labor and structures of power, that shape and constrain women's life experiences, behavior and social roles, and, ultimately, their health risks (Connell, 1987; Walsh et al., 1995). Resonating with the recent recognition of the importance of the social context in determining health and disease (Diez-Roux, 1998; Macintyre & Ellaway, 2000), the society and health lens directly points to the gendered social, economic and political arrangements within society in generating gender differences in physical and mental health and offers a useful framework for examining women's over-representation in depression.

Yllo pioneered research on the relationship of contextual gender inequality and women's health outcomes (Yllo, 1983). She demonstrated that US states with greater gender inequality had higher rates of violence against women perpetrated by their husbands. Using a similar ecological approach, Kawachi et al. (1999) found that US state-level women's status as measured by women's political participation, economic autonomy, employment & earnings, and reproductive rights, was associated with both male and female morbidity and mortality rates (Kawachi et al., 1999).

Turning to research on depression, more attention has been paid to the impact of economic context on women's depression, including socioeconomic status, income inequality, and neighborhood poverty (Belle, 1982; Dohrenwend et al., 1992; Kahn, Wise, Kennedy, & Kawachi, 2000; Lorant et al., 2003). For example, Belle demonstrated that poverty increases stress and leads to depression in women, especially for mothers with young children (Belle, 1982). Furthermore, a burgeoning literature has revealed the importance of contextual economic conditions in determining depression. For example, a 1999 study based in the United Kingdom, demonstrated the impact of material deprivation at the ward-level on depression (Wilson, Chen, Taylor, McCracken, & Copeland, 1999). In addition, a recent study by Kahn and colleagues using data from the United States showed that state-level income inequality was associated with higher odds of depressive symptoms in women, net of individual income (Kahn et al., 2000). Further evidence of the importance of neighborhood economic context in determining depression stems from a recent housing voucher experiment in the US (Moving to Opportunity), which documented decreased rates of depression in women and children who moved from high-poverty urban neighborhoods to low-poverty suburban neighborhoods (Katz, Kling, & Liebman, 2001).

However, while the research on the effect of disadvantaged economic conditions (at both the individual and contextual level) on women's depression is quite rich, the role of the “gendered context” (i.e. structural gender inequality) in shaping the gender discrepancies in depression remains unexplored. Although women's subordinate position in society is closely associated with their lack of access to economic resources, material disadvantage also cannot fully account for the pervasive system of gender-based oppression (Connell, 1987). Examples of other contributing factors include social controls over women's sexuality, restrictions on women's reproductive rights, violence against women, undervaluation of women's work, gender divisions in paid and unpaid work, and sexual segregation in waged work, to name a few. These issues cannot be adequately covered by using gender-neutral economic indicators.

With the rise in federalism in the past decades, the state has increasingly become the unit of legislation and policy implementation in the US. The power to legislate, fund and enforce policies and programs has been devolved to the individual states (Daniels, 1997). States that stipulate policies that enforce gender equality in reproductive, economic, political or other social domains can create a more women- and family-friendly environment. Conversely, states that neglect, tolerate or sanction women's unequal social status can perpetuate women's disadvantaged positions and consequently harm their mental and physical health. The Institute for Women's Policy Research (IWPR) assembled 4 composite indices in 1996 to assess the status of women in fifty US states in four separate domains—“political participation”, “employment & earnings”, “economic autonomy” and “reproductive rights” to inform, identify and measure the barriers to gender equality at the state level (Institute for Women's Policy Research, 1996). The women's status indicators developed by IWPR thus provide an opportunity to test the association between the “gendered context” of society and women's depression through the “society and health lens”.

We employed a multilevel analytical framework to examine the contribution of women's status at the state-level for women's depressive symptoms, taking into account other individual and contextual determinants of depression. Individual level factors, such as age, race/ethnicity, income, education and unemployment status, are important determinants of depression, and were considered in our analysis (Horwath & Weissman, 1995). We also examined state level income inequality as an independent contextual predictor of depressive symptoms (Kahn et al., 2000). We hypothesized that women who live in states with higher women's status and autonomy would report lower levels of depressive symptoms, and vice versa, after controlling for other individual and contextual determinants of depression. In addition, while women as a group have shared issues, they are by no means a homogenous group (Doyal, 1995). We therefore hypothesized that contextual women's status has differential effects on women from varying racial and socioeconomic backgrounds. We tested for cross-level interactions to examine whether women with lower household incomes or from racial minority groups are more vulnerable to depression as a result of residing in areas with lower women's status.