Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2008 Sep 1.
Published in final edited form as: Womens Health Issues. 2007 Aug 17;17(5):316–324. doi: 10.1016/j.whi.2007.07.001

Factors associated with major depression among mothers in Los Angeles

Sandraluz Lara-Cinisomo 1, Beth Ann Griffin 2
PMCID: PMC2108528  NIHMSID: NIHMS30389  PMID: 17707124

Abstract

Purpose

The aim of this study was to identify factors associated with major depression among a sample of diverse mothers in Los Angeles while paying special attention to racial and ethnic as well as immigration status differences.

Methods

Using logistic regression models, we examined the association between major depression and race and ethnicity, immigration status, and other key covariates. Major depression was measured using the Comprehensive International Diagnostic Interview Short Form (CIDI-SF). This study was based on 1,856 racially and ethnically diverse mothers who participated in Wave 1 of the Los Angeles Family and Neighborhood Survey (L.A.FANS), which was fielded in 65 census tracts.

Main Findings

After controlling for key covariates, we found that non-Hispanic white mothers had more than 1.67 times the odds of having major depression than Hispanic mothers (95% CI =0.99, 2.80). In addition, single mothers had elevated rates of major depression compared to married mothers (OR = 1.54, 95% CI = 1.00, 2.37). Mothers with a college degree or higher had significantly lower odds of being depressed compared to mothers without a college degree (OR = .50, 95% CI = 0.29, 0.86) while mothers with only adolescents in the home had significantly higher odds of major depression than mothers with at least one preadolescent child in the home (OR = 1.73, 95% CI = 1.11, 2.70).

Conclusion

Given the links between depressed mothers and child outcomes, our results have important implication for mothers with adolescent children, particular those who are white, single or less educated.

Introduction and Background

Depression among mothers is an important indicator of maternal well-being, particularly because it can have debilitating effects on how well she functions as a whole and in the home. Beyond its effects on the mother, depression has also been shown to have important effects on children’s mental health (Luoma et al., 2001; Oyserman, Bybee & Mowbray, 2002; Petterson & Albers, 2001; Spence, Najman & Bor, 2002). For example, studies have found that mothers experiencing depression tend to be less interactive and less nurturing with their children (Cummings & Davies, 1994). Poor mother child interaction has also been shown to affect children’s behavior problems (Luoma et al., 2001). The issue that is unclear is what the risk factors for depression, particularly major depression, are for mothers from diverse backgrounds. Much of the research on depression and mothers has typically focused on a specific group of mothers with the primary interest being in detecting significant differences in rates of depression between mothers who differ on only one main factor such as race and ethnicity or income.

For example, numerous studies have examined racial and ethnic disparities in depressive symptoms (Cochran, Brown & McGregor, 1999; Myers et al., 2002) and major depression (Dunlop, Song, Lyons, Manheim & Chang, 2003; Somervell, Leaf, Weissman, Blazer & Bruce, 1989) among women, and more specifically, depressive symptoms among mothers (Heneghan, Silver, Bauman, Westbrook, & Stein, 1998; Herrick, 2002; Howell, Mora, Horowitz, & Leventhal, 2005). However, the results are mixed with most large studies showing Hispanics with higher rates of major depression compared to whites and African Americans and Asians demonstrating lower rates (Blazer, Kessler, McGonagle, & Swartz, 1994; Zhang & Snowden, 1999). Conversely, one recent large study found that whites had significantly higher prevalence of major depression compared to blacks and Hispanics (Riolo, Nguyen, Greden, & King, 2005). Other smaller studies have shown that ethnic minorities, as a group, have higher rates of depressive symptoms compared to whites (see Myers & Hwang, 2004 for a review).

Immigration status complicates these mixed results. Recent studies on immigration suggest that some immigrant groups, such as Hispanic immigrants, may have characteristics that protect them against adverse mental health outcomes. For example, Hispanics have been shown to have lower rates of chronic disease and illness despite limited economic resources (Morales, Lara, Kington, Valdez & Escarce, 2002), a phenomenon known as the “Hispanic Paradox.” Researchers suggest that family support systems among Hispanic immigrants may be the reason for this finding and may serve as buffers from poor mental health as well (Morales et al., 2002; Rogler, Cortes & Malgady, 1991). An alternative argument is selection, which suggests that Hispanics with pre-existing depressive symptoms do not migrate into the United States (see for example Goldman, Smith, & Sood, 2006). Results from another study suggest that African and Caribbean immigrant women, like Mexican immigrant women in other studies, have better mental health when they first immigrate into the U.S, but as their time in the U.S. increases, so does the probability for major depression among these women (Miranda, Siddique, Belin & Kohn-Wood, 2005). These researchers suggest that discrimination, poverty, and other social factors are to blame for this increased likelihood of depression among these immigrant women.

Other factors have also been examined that function as explanatory variables for depressive symptoms and major depression among mothers, such as income (Scarinci, Beech, Naumann, Kovach, Pugh & Fapohunda, 2002), marital status (Davies, Avison & McAlpine, 1997; Earle, Smith, Harris & Longino, 1998; Gazmararian, James & Lepkowski, 1995; Scarinci, Beech, Naumann, Kovach, Pugh & Fapohunda, 2002), and education (Dunlop et al., 2003; Heneghan, Silver, Bauman, Westbrook & Stein, 1998; Scarinci et al., 2002). These studies have shown that lower levels of income and education, and single marital status are linked with higher levels of depressive symptoms as well as major depression. An important explanatory variable often left out of the equation is body weight. Previous studies have shown a link between body weight and depressive symptoms (Istvan, Zavela & Weidner, 1992; Jorm et al., 2003; Carpenter, Hasin, Allison & Faith, 2000; Roberts, Deleger, Strawbridge & Kaplan, 2003). Using longitudinal data, Roberts and his colleagues showed that obesity was predictive of major depression.

Limited research has been completed on the age composition of children and its association with depression among mothers. Previous studies on mothers from specific child age groups, such as infants, have shown high prevalence of depression (Cummings and Davies, 1994; Field, 1998; Heneghan et al., 1998; Luoma et al., 2001; Petterson & Albers, 2001). However, it is unclear whether having children in one age group versus another age group places mothers at higher risk for depression. This study provides unique insight into the associations that may exist between major depression in mothers and age composition of children in the home.

Despite the growing body of research on major depression among mothers, there are still gaps in our knowledge about its overall level of prevalence, key predictors, and about disparities by characteristics such as race and ethnicity and immigration status. The primary reason for these gaps is that large population-based social surveys rarely include valid and accurate depression screeners which allow for rigorous study of risk factors for depression. Even when these screeners have been included, these surveys often do not include sufficiently large samples of certain groups of interest, such as Hispanics and/or immigrants

This study sought to address some of the gaps described above. Specifically, we aimed to answer two primary questions: 1) Given the racial and ethnic diversity of mothers in Los Angeles, do we find racial and ethnic disparities and/or immigration status disparities in major depression? and 2) When controlling for race and ethnicity and immigration status, do we find differences in major depression by other key covariates (e.g., economic resources, marital status, education, body weight, number of children, and child age composition)? To answer these questions, we use data from the Los Angeles Family and Neighborhood Survey (L.A.FANS).

Method

Study Design

This study was based on data from Wave 1 of L.A.FANS, which was fielded in a sample of 65 census tracts throughout Los Angeles County in 2000–2001. L.A.FANS is based on a stratified multistage, clustered sampling design (see Sastry, Ghosh-Dastidar, Adams, & Pebley, 2006 for more detail). In the first step of the sample design, census tracts in Los Angeles County were divided into three strata based on the percent of the tract’s population in poverty in 1997: very poor (those in the top 10 percent of the tract poverty distribution), poor (tracts in the 60-89th percentiles of the distribution), and non-poor (tracts in the bottom 60 percent of the distribution). Tracts within each strata were sampled with probabilities proportional to the population size of the tract, and to achieve an oversampling of poor families, 20 tracts were sampled in both the poor and very poor strata and 25 tracts were sampled in the non-poor stratum, yielding a total of 65 tracts. Next, census blocks were sampled within each tract with probabilities proportional to the block population size, and all dwelling units in sampled blocks were listed. Fifty households were sampled within each block and screened, and approximately 40–50 households were interviewed in each tract, for a total sample size of 3,090 households. Households with children were oversampled; as a result, they comprise 70 percent of the sample, compared to 35 percent in Los Angeles County as a whole.

L.A.FANS collected extensive information on a variety of topics including household socioeconomic status and mental health status of mothers in the households. The response rate among mothers selected for the survey was 89 percent.a Of the mothers who completed the survey, just over 94 percent had all of the necessary data for the analysis conducted in this study (N = 1,873). Given the small number of Native American mothers in the sample (n=17), they were dropped from the analysis. Therefore, the results shown in the subsequent sections are based on 1,856 mothers.

Data Collection Procedures

L.A. FANS included both in-person interviews and a survey of neighborhoods (see Sastry et al., 2006). In-person interviews were conducted by trained data collectors. Computer Assisted Personal Interviewers (CAPI) was used to conduct these interviews. Data collectors were trained to use the CAPI and on each of the measures included in the interview protocol (see Peterson et al., 2003). Interviews were conducted Spanish or English depending on the participant’s preference. Only Spanish speaking data collectors were permitted to administer the Spanish questionnaire and assessments.

Sample

Four race and ethnic groups were represented in the data. Unweighted descriptive statistics for the analysis sample are reported in Table 1. The largest group was Hispanics (62 percent), followed by non-Hispanic whitesb (21 percent); Eight percent of mothers were non-Hispanic blacksc and eight percent were Asian and Pacific Islanders.

Table 1.

Descriptive Statistics for Analysis Sample (N=1856)

Unweighted
Percent or
Mean		 Unweighted
Standard
Deviation
Race and ethnicity
 White non-Hispanic 21%
 Hispanic 62%
 Black non-Hispanic 8%
 Asian/Pacific Islander 8%
Immigration Status
 Undocumented Immigrant 22%
 Documented Immigrant 40%
 Non-Immigrant 38%
Economic Resources
 Income (dollars) 52,846 96,240
 Assets (dollars) 130,789 528,313
Marital Status
 Married 64%
 Cohabitating 12%
 Single 24%
Education
 Less than College 83%
 College and Beyond 17%
Body weight
 Normal (BMI <25) 38%
 Overweight and Obese (BMI ≥ 25) 62%
Number of Children 2.03 1.08
Child Composition
At least one infant [1] 10%
At least one preadolescent child [2] 82%
Only adolescents [3] 8%
Open in a new tab

Notes:

[1]

At least one infant (child under the age of one year) in the household—though household could include one or more children between one and 17 years of age.

[2]

This variable identifies mothers with at least one child between 1 and 14 years of age but with no infants, and can include an adolescent between 15 and 17 years of age.

[3]

This variable identifies mothers with one or more child between 15 and 17 years of age but no children from any other age groups.

Measures

Major Depression

The dependent variable of interest was an indicator of whether a mother had major depression. L.A.FANS screened for major depression using the Comprehensive International Diagnostic Interview short form (CIDI-SF), an international protocol adopted by the World Health Organization (Kessler, Andrews, Mroczek, Ustrun & Wittchen, 1998). This instrument screens for a major depressive episode during the 12-month period preceding the interview and estimates the probability that a respondent had major depression based on the Diagnostic and Statistical Manual of Mental Disorders - Fourth Edition (DSM-IV) criteria for a major depressive episode (Wang & Patten, 2002). The CIDI-SF is a valid and reliable diagnostic interview and has been shown to have 93 percent classification accuracy for major depressive disorders (Kessler et al., 1998).

Individuals could meet the criteria for major depression by either responding “yes” to all the stem questions about dysphoric mood (i.e., sadness or anxiety) or responding “yes” to questions about anhedonia (i.e., an inability to experience joy).d Dysphoric or anhedonic symptoms should have lasted for two weeks for most of the day and should have occurred almost every day during the period to meet requirements for classification (Kessler et al., 1998).

The CIDI-SF screener only identifies individuals who have a high probability of being classified with major depression (see Kessler et al., 1998). Neither the severity nor the duration of major depression was assessed in this study. Probability rates were calculated based on participant responses and criteria described above. The CIDI-SF scoring algorithm indicates that probability scores equal to or greater than 0.55 classifies a respondent as being likely to have major depression.e Thus, probability scores were dichotomized to indicate whether a woman has major depression (e.g., has a probability score greater than or equal to 0.55) or not.

Immigration Status

We examined three immigration status categories for mothers in the sample: undocumented immigrant, documented or legal immigrant (including naturalized citizens), and non-immigrant (native born). Immigration status was coded as a categorical variable with documented immigrant mothers as the reference group and undocumented and non-immigrant mothers were modeled using two separate dummy variables.

Economic Resources

Total family income and total family assets were used as covariates. Total family income includes income for all members of the nuclear family from all sources (including earnings, public assistance, and income from assets). Total family assets include savings accounts, property, and business investments, as well as ownership of stocks and bonds. By including a measure of assets, we account for total resources available to mothers, which is an important and often ignored component of economic status. Missing information for income and assets was imputed as part of a separate study (Bitler & Peterson, 2005) that drew on the extremely detailed information for household income and assets in L.A.FANS and included unfolding brackets to identify a successively narrower range for each variable when a respondent was unable or unwilling to provide a specific dollar value.f Income and assets were included as log-transformed continuous variables in our models to account for skewness in these measures.

Marital Status

Mothers were classified as single, cohabitating, or as married. Marital status was dummy-coded with married as the reference group and cohabitating and single coded as dummy variables.

Education

Education was dichotomized with mother with less than a college degreeg = 0 and those with a college degree or higher = 1. Preliminary analyses explored the use of three education groups (e.g., mothers with less than a high school degree, mothers with a high school degree and some college, and mothers with a college degree or higher). Since mothers with less than a high school degree and mothers with a high school degree and some college had similar rates of depression, these categories were collapsed for the final analysis.

Body weight

The relationship between mother’s body weight and major depression was assessed using the body mass index (BMI) to classify women as overweight or obese (BMI > 25) or not, h according to standards established by the National Heart, Lung and Blood Institute (1998). We did not distinguish underweight mothers from those of normal weight after preliminary analyses showed similar rates of depression in these two groups. Body weight was dichotomized with normal and underweight mothers as the reference group.

Child Age Composition

Detailed information was collected in L.A.FANS on the number of children that were born to the mother, as well as the children’s individual ages. We examined a number of different ways to represent the age composition of children in a household. First, we included a continuous variable for the number of children in our regression models. We also examined the age composition of the children based on developmental age groups that distinguished among infants (less than one year of age), preschoolers (ages 1–5 years), elementary school-age children (ages 6–8 years), middle school-age children (ages 9–12 years), pre-adolescents (ages 13–14 years), and adolescents (ages 15–17 years). We conducted a preliminary analysis to simplify the large number of combinations that were possible given these age groupings, as well as to provide information on the number of children in each group and their effects on the mother’s probability of major depression.

The results of this analysis revealed several interesting findings. First, the only group that had statistically significant higher probability of being classified with major depression was mothers who only had adolescent-age children. These women had a substantially higher likelihood of having major depression when compared to all other mothers, with an odds ratio of 1.72 (95% CI: 1.19, 2.48). There were no significant differences in probability of being classified as depressed between mothers who had an infant or did not only have adolescents, based on the ages of their children. These results led to the coding of this variable into three groups: mothers with at least one infant (<1), mothers with only adolescents (ages 15 – 17) and mothers with at least one pre-adolescent child and no infants. These groups are mutually exclusive and exhaustive.

Thus, three exclusive child age composition categories were created: at least one infant; at least one pre-adolescent; and adolescent-only, and child age composition was coded using dummy variables with mothers with at least one pre-adolescent as the reference group.

Analysis

Our analysis examined major depression among mothers using a binary indicator of major depression as the dependent variable of interest. The primary goal of our analysis was to identify predictors of major depression among mothers in the L.A.FANS data.

In our modeling approach, we used logistic regression models which appropriately controlled for the sampling design used in the L.A.FANS study. Specifically, three aspects of the L.A. FANS design must be accounted for when analyzing the data: stratification of tracts by poverty level, clustering of women within tracts, and use of sampling weights equal to the inverse probability that a woman was sampled for the study. Controlling for the clustering of women within neighborhoods ensured that our regression models had unbiased standard error estimates. Controlling for the stratification and use of the sampling weights in the regression models ensured that our inferences would be generalizable to the population from which these women were sampled (i.e., all neighborhoods and households in Los Angeles County). The SVY command in STATA was used to control for these three factors in our logistic regression models.

Our analysis proceeded in two steps. First, we examined differences in the probability of major depression by race and ethnicity, immigration status, income and asset categories, marital status, education, weight, and child age composition in bivariate models. Rao-Scott chi-squared statistics were calculated to test for statistically significant differences across all groups while adjusting for the stratification and clustering of the sample design and incorporating sample weights (Lee & Forthofer, 2006).

Next, we estimated multivariable logistic regression models of the probability of major depression to determine which demographic variables were significantly associated with the probability of depression among mothers in the L.A.FANS data. We report the results from both the unadjusted and adjusted models for race and ethnicity and immigration status in order to examine the effects of key demographic variables on the probability of depression after controlling for race and ethnicity and immigration status. For each set of models, all two-way interaction terms were fitted and significance was assessed at the 0.05 level. All logistic regression models appropriately controlled for stratification, clustering, and sampling weights.

Results

The results of our analysis are presented in Tables 2 and 3. In Table 2, we show the weighted distribution of major depression in the sample by key demographic variables. Rao-Scott chi-squared tests are also reported to identify significant differences in major depression across groups. In Table 3, shows results from the logistic regression models.

Table 2.

Distribution of Depression Rates by Key Covariates (N=1856)

Weighted Percent
Depressed
Total 15
Race and ethnicity
 White non-Hispanic 18
 Hispanic 14
 Black non-Hispanic 20
 Asian/Pacific Islander 7
 Chi-Squared (3) 14.28*
Immigration Status
 Documented Immigrant 14
 Undocumented Immigrant 12
 Non Immigrant 16
 Chi-Squared (2) 3.09
Log Income
 Lower 50% 15
 Top 50% 14
 Chi-Squared (1) 0.98
Log Assets
 Lower 50% 17
 Top 50% 12
 Chi-Squared (1) 4.40*
Marital Status
 Married 12
 Cohabitating 16
 Single 19
 Chi-Squared (2) 6.95*
Education
 Less than College Degree 16
 College Degree 8
 Chi-Squared (1) 13.32***
Weight
 Normal or Underweight 13
 Overweight or Obese 16
 Chi-Squared (1) 5.07*
Child Age Composition
 At least one infant 17
 At least one preadolescent child/no 13
 infants
 Only Adolescents 22
 Chi-Squared (2) 7.03*
Open in a new tab

Notes:

*

p < .05;

**

p < .01;

***

p<.001

Table 3.

Results from the Logistic Regression Models for Major depression

Model 1
OR (95% CI)		 Model 2
OR (95% CI)
Race and ethnicity
 Non-Hispanic White 1.29 (0.80,2.09) 1.67 (0.99,2.80)*
 Black 1.55 (0.93,2.58) ± 1.43 (0.84,2.43)
 Asian/Pacific Islander .42 (0.24,0.75)** 0.68 (0.37,1.23)
 Hispanic 1 1
Immigration Status
 Legal 1 1
 Undocumented .75 (0.50,1.13) .70 (0.45,1.06)±
 Non-Immigrant .93 (0.54,1.59) .90 (0.52,1.56)
Economic Resources
 Log Income 1.03 (0.95, 1.11)
 Log Assets .98 (0.94,1.03)
Marital Status
 Married 1
 Cohabitating 1.29 (0.79,2.10)
 Single 1.54 (1.00,2.37)*
Education
 Less than College Degree 1
 College Degree .50 (0.29,0.86)**
Body Weight
 Normal and Underweight 1
 Overweight or Obese 1.31 (0.95,1.79)±
Children
 At least one infant 1.59 (0.93,2.70)±
 At least one preadolescent 1
 child/no infants
 Only Adolescents 1.73 (1.11,2.70)*
Number of Children 1.02 (0.88,1.18)
Number of Observations 1856
Open in a new tab

Notes: Robust standard errors are reported that adjust for clustering of observations by tract and that the regressions are weighted using the sampling weights provided by L.A.FANS.

±

p < .10;

*

p < .05;

**

p < .01;

***

p<.001.

The results in Table 2 indicate that 15 percent of mothers in our sample met the classification for major depression. White and black mothers had the highest levels of major depression (18 percent and 20 percent, respectively), while 14 percent of Hispanic mothers were depressed and only 7 percent of Asian and Pacific Islander mothers had major depression. A Rao-Scott chi-squared test indicated these racial and ethnic differences were statistically significant with p< .05.

Our results also showed that 16 percent of non-immigrants mothers had major depression compared to 14 percent and 12 percent of documented and undocumented immigrant mothers, respectively. According to the Rao-Scott chi-squared test, these differences were not statistically significant.

There were no significant differences in rates of major depression in the upper and lower 50th percentile brackets for log income. In contrast, there was a significant association between log assets and major depression (X2 =4.40, p< .05) when we compared mothers in the upper 50th percentile versus mothers in the lower 50th percentile. Seventeen percent of mothers in the lower 50th percentile were depressed compared to 12 percent in the upper 50th percentile.i

Results from the Rao-Scott chi-squared test indicated that differences in major depression among married mothers, cohabitating mothers, and single mothers was statistically significant (X2 = 6.95, p< .05), with 12 percent of married mothers classified as depressed compared with 16 and 19 percent of cohabitating and single mothers.

Sixteen percent of mothers with less than a college degree were depressed and eight percent of mothers with at least a college education had major depression. These differences were also statistically significant (X2 = 13.32, p< .001).

Finally, based on the Rao-Scott chi-squared test, there was a statistically significant association between body weight and major depression (X2 =5.07, p< .05) and between children’s age composition and major depression (X2 =7.03, p< .05). Women who were overweight or obese had higher levels of depression than women of normal or under weight, 16 and 13 percent respectively. Twenty-two percent of mothers with only adolescents in the household were depressed, compared to 13 percent of mothers with at least one pre-adolescent child and 17 percent of mothers with at least one infant.

Table 3 provides results from the multivariate logistic regression models. Model 1 shows that when comparing mothers in our sample, Asian/Pacific Islanders had significantly lower odds (OR = 0.42, 95% CI = 0.24, 0.75) of being depressed compared to Hispanic mothers. Additionally, black mothers had marginally significant higher odds of having major depression than Hispanic mothers (OR = 1.55, 95% CI = 0.93, 2.58). Results from the joint test indicated that all three race and ethnicity parameter estimates were jointly significant [F (3, 60) = 4.49, p<.001]. Furthermore, results from the pairwise statistical tests of covariate estimates indicated that there were statistically significant differences in major depression between Asian/Pacific Islanders, on one hand, and whites and blacks, on the other hand (test results not shown). Immigration status was not significantly associated with major depression in this model.

Model 2 shows the results from the fully adjusted model for major depression, which controlled for all the covariates of interest simultaneously. The results indicate that after controlling for the other key covariates, race remained strongly associated with major depression. Specifically, after adjusting for other covariates, white mothers were now significantly more likely to have major depression compared to Hispanic mothers (OR = 1.67, 95% CI = 0.99, 2.80). As in Model 1, results from the joint test of race indicate that all three race and ethnicity parameter estimates are jointly significant [F (3, 60) = 3.24, p<.05]. Results from the pairwise statistical tests indicated that race and ethnic group differences in major depression were statistically significant between white and Asian/Pacific Islander mothers as well (test results not shown for pairwise tests).

Additionally, model 2 shows that single mothers had significantly higher odds of having major depression compared to married mothers (OR = 1.54, 95% CI = 1.00, 2.37). Mothers who had a college degree or greater were significantly less likely to be depressed than mothers without a college degree (OR = 0.50, 95% CI = 0.29, 0.86). Finally, mothers with only adolescents were significantly more likely to have major depression (OR = 1.73, 95% CI = 1.11, 2.70) than mothers with at least one pre-adolescent child and no infants.j

Discussion

Understanding predictors of disparities in mental health outcomes among mothers is important because a mother’s poor emotional well-being can have negative effects on her functioning and can have lasting negative effects on her children (Cummings & Davies, 1994; Field, 1998; Heneghan et al., 1998; Luoma et al., 2001; Petterson & Albers, 2001). Understanding which women, and ultimately which children, are at risk for major depression is important. Thus, our study aimed to understand mental health disparities as they exist in a sample of Los Angeles mothers who participated in the L.A.FANS. To achieve this goal, we used logistic regression models which appropriately controlled for the complex survey sample design of L.A.FANS. Our analysis focused on drawing inferences about racial and ethnic and immigration status disparities in depression rates among mothers in Los Angeles County as well as on drawing inferences about other key predictors of major depression.

After controlling for race and ethnicity, immigration status, family economic resources, marital status, maternal education, body weight, number of children and child age composition, results from our study indicated that race, marriage, education, and child-age composition were the strongest predictors of major depression in the L.A. FANS sample.

Contrary to previous studies, we found that white mothers had significantly higher odds of being classified with major depression when compared to Hispanic mothers. With the exception of one study (Riolo, Nguyen, Greden & King, 2005), previous studies have shown the opposite, that white women have better mental health outcomes when compared to ethnic minority women (Cochran, Brown & McGregor, 1999; Dunlop, Song, Lyons, Manheim & Chang, 2003; Myers et al., 2002; Somervell, Leaf, Weissman, Blazer & Bruce, 1989; Vega, Kolody, Valle & Hough, 1986). One possible explanation might be that ethnic minority women, such as Hispanic and Asian/Pacific Islanders, have protective factors that white women do not such as large family units, which bring with them high levels of social support.

On the other hand, results from our analysis support previous studies which have shown that single mothers (Davies, Avison & McAlpine, 1997) as well as those who are less educated (Heneghan, Silver, Bauman, Westbrook & Stein, 1998) are at risk for depression compared to their married and highly educated counterparts.

Unlike other studies, our study showed that having children in a specific age group put mothers at higher risk for major depression. Specifically, we found that mothers with only adolescents in the home had significantly higher odds of being depressed compared to mothers with a mix of children other than infants. These results highlight the importance of screening mothers with older children. Numerous studies emphasize the need to screen mothers both during and after pregnancy (for example Georgiopoulos, Bryan, Wollan & Yawn, 2001; Yonkers et al., 2001). However, our results indicate that there may also be a need to screen mothers with adolescents in the home.

Limitations

Our study had a number of limitations. First, our measure of major depression does not rely on a diagnostic interview with a clinician nor does it provide a second rater to confirm the measure on an individual mother. Second, with only 276 depressed mothers in our sample, our analysis lacked sufficient power to detect significant two-way interactions between our key predictors of major depression. Of particular interest would be to detect interactions which may exist between race and ethnicity and immigration status as well as the interactions these variables may have with the other covariates studied in this paper. An analysis which examines these possible interactions is needed and would provide an important contribution to the field. Unfortunately, the L.A.FANS data were not designed with sufficient power to address such an analysis completely. All two way interactions were fit in our analysis and significance was assessed. Only one was found to be significant at the 0.05 level and this interaction was most likely due to small sample sizes and thus was not reported or over interpreted.

When examining the amount of power the study had to detect significant interactions between race and ethnicity and immigration status, we found that we had sufficient power to detect small effect size differences between legal Hispanic immigrant mothers and undocumented or non-immigrant Hispanics and white non-immigrant mothers. However, for the other race and ethnic groups, small sample sizes lead to poor power for detecting significant differences in the effects of immigration status among these groups. If these interactions do exist, caution must be taken when drawing conclusions about the joint effects of ethnicity, immigration status, and the other covariates.

Acknowledgments

We would like to acknowledge Narayan Sastry, Co-Director of the L.A. FANS for his guidance, editing and support.

The authors are grateful to the National Institute of Child Health and Human Development (R01HD41486) for its support of this research.

Footnotes

a

See Sastry & Pebley (2003) for details about rates and patterns of non-response in L.A.FANS.

b

We refer to non-Hispanic whites as “whites.”

c

We refer to non-Hispanic blacks as “blacks.”

d

While the CIDI-SF was designed to classify individuals by probabilities for major depression, it does not distinguish between major depressive episodes that are a result of bipolar or psychotic disorders.

e

Due to the nature of the scoring algorithm, only seven distinct scores appeared in the L.A.FANS data. Specifically, 83% of mothers had a score of 0, <1% a score of 0.06, 2% a score of 0.24, 2% a score of 0.55, 3% a score of 0.81, 8% a score of 0.89, and 2% a score of 0.91.

f

Eight-nine cases of family income and 382 cases of assets were imputed.

g

Less than a college degree includes mothers with some college education, but not completion of a degree.

h

The body mass index is defined as weight in kilograms divided by the squared value of height in meters.

i

In a bivariate logistic regression containing both log assets and log income, the same pattern arose. Higher log assets was significantly associated with lower rates of depression (OR = 0.96, 95% CI = 0.92, 1.00) while there was no statistically significant relationship between log income and depression.

j

Marginally significant results (p < .10) were found for immigration status, body weight, and mothers with at least one infant (see Table 3).

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Contributor Information

Sandraluz Lara-Cinisomo, RAND Corporation

Beth Ann Griffin, RAND Corporation

References

  1. Bitler M, Peterson C. Working Paper. Santa Monica: RAND Corporation; 2005. L.A.FANS income and assets imputations. [Google Scholar]
  2. Blazer DG, Kessler RC, McGonagle KA, Swartz MS. The prevalence and distribution of major depression in a national community sample: the National Comorbidity Survey. American Journal of Psychiatry. 1994;151:979–986. doi: 10.1176/ajp.151.7.979. [DOI] [PubMed] [Google Scholar]
  3. Carpenter KM, Hasin DS, Allison DB, Faith MS. Relationship between obesity and DSM-IV major depressive disorder, suicide ideation, and suicide attempts: Results from a general population study. American Journal of Public Health. 2000;90(2):251–257. doi: 10.2105/ajph.90.2.251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cochran DL, Brown DR, McGregor KC. Racial differences in the multiple social roles of older women: implications for depressive symptoms. Gerontologist. 1999;39(4):465–72. doi: 10.1093/geront/39.4.465. [DOI] [PubMed] [Google Scholar]
  5. Cummings EM, Davies PT. Maternal depression and child development. Journal of Child Psychology and Psychiatry. 1994;35:73–112. doi: 10.1111/j.1469-7610.1994.tb01133.x. [DOI] [PubMed] [Google Scholar]
  6. Davies L, Avison WR, McAlpine DD. Significant life experiences and depression among single and married mothers. Journal of Marriage and Family. 1997;59(2):294–308. [Google Scholar]
  7. Dunlop DD, Song J, Lyons JS, Manheim LM, Chang RW. Racial/Ethnic differences in rates of depression among preretirement adults. American Journal of Public Health. 2003;93(11) doi: 10.2105/ajph.93.11.1945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Earle JR, Smith MH, Harris CT, Longino CF., Jr Women, marital status, and symptoms of depression in a midlife national sample. Journal of Women and Aging. 1998;10(1):41–57. doi: 10.1300/j074v10n01_04. [DOI] [PubMed] [Google Scholar]
  9. Field T. Maternal depression effects on infants and early interventions. Preventive Medicine. 1998;27:200–203. doi: 10.1006/pmed.1998.0293. [DOI] [PubMed] [Google Scholar]
  10. Gazmararian JA, James SA, Lepkowski JM. Depression in black and white women. The role of marriage and socioeconomic status. Annals of Epidemiology. 1995;5(6):455–463. doi: 10.1016/1047-2797(95)00061-5. [DOI] [PubMed] [Google Scholar]
  11. Georgiopoulos A, Bryan T, Wollan P, Yawn B. Routine screening for postpartum depression. Journal of Family Practice. 2001;50:117–22. [PubMed] [Google Scholar]
  12. Goldman DP, Smith JP, Sood N. Immigrants and the cost of medical care. Health Affairs. 2006;25(6):1700–1711. doi: 10.1377/hlthaff.25.6.1700. [DOI] [PubMed] [Google Scholar]
  13. Heneghan AM, Silver EJ, Bauman LJ, Westbrook LS, Stein REK. Depressive symptoms in inner-city mothers of young children: Who is at risk? Pediatrics. 1998;102:1394–1400. doi: 10.1542/peds.102.6.1394. [DOI] [PubMed] [Google Scholar]
  14. Herrick H. Postpartum depression: who gets help? Statistical Brief 24. Raleigh, NC: Department of Health and Human Services State Center for Health Statistics; 2002. [Google Scholar]
  15. Howell EA, Mora PA, Horowitz CR, Leventhal H. Racial and ethnic differences in factors associated with early postpartum depressive symptoms. Obstetrics & Gynecology. 2005;105:1442–1450. doi: 10.1097/01.AOG.0000164050.34126.37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Istvan J, Zavela K, Weider G. Body weight and psychological distress in NHANES I. International Journal of Obesity and Related Metabolic Disorders. 1992;12:999–1003. [PubMed] [Google Scholar]
  17. Jorm AF, Korten AE, Christensen H, Jacomb PA, Rodgers B, Parslow RA. Association of obesity with anxiety, depression and emotional well-being: a community survey. Australian and New Zealand Journal of Public Health. 2003;27(4):434–440. doi: 10.1111/j.1467-842x.2003.tb00423.x. [DOI] [PubMed] [Google Scholar]
  18. Kessler RC, Andrews G, Mroczek D, Ustrun B, Wittchen HU. The World Health Organization Composite International Diagnostic Interview short-form (CIDI-SF) International Journal of Methods in Psychiatric Research. 1998;7:172–185. [Google Scholar]
  19. Lee ES, Forthofer RN. Analyzing complex survey data. 2. Thousand Oaks: Sage Publications; 2006. [Google Scholar]
  20. Luoma I, Tamminen T, Kaukonen P, Laippala P, Puura K, Salmenlin R, Almqvist F. Longitudinal study of maternal depressive symptoms and child well-being. Journal of the American Academy of Child and Adolescent Psychiatry. 2001;40:1367–1374. doi: 10.1097/00004583-200112000-00006. [DOI] [PubMed] [Google Scholar]
  21. Miranda J, Siddique J, Belin TR, Kohn-Wood LP. Depression prevalence in disadvantaged young black women: African and Caribbean immigrants compared to US-born African Americans. Social Psychiatry and Psychiatric Epidemiology. 2005;40(4):252–258. doi: 10.1007/s00127-005-0879-0. [DOI] [PubMed] [Google Scholar]
  22. Morales L, Lara M, Kington RS, Valdez RO, Escarce JJ. Socioeconomic, cultural and behavioral factors affecting Hispanic health outcomes. Journal of Health Care for the Poor and Underserved. 2002;13(4):477–503. doi: 10.1177/104920802237532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Myers HF, Hwang W. Cumulative psychosocial risks and resilience: a conceptual perspective on ethnic health disparities in late life. In: Anderson NA, Bulatao RA, Cohen B, editors. National Research Council, critical perspectives on racial and ethnic disparities in health in later life. Washington, D.C.: The National Academies Press; 2004. Committee on Population, Division of Behavioral and Social Sciences and Education. [PubMed] [Google Scholar]
  24. Myers HF, Lesser R, Rodriguez N, Mira C, Hwang WC, Camp C, Anderson D, Lucy E, Wohl M. Ethnic differences in clinical presentation of depression in adult women. Cultural Diversity and Ethnic Minority Psychology. 2002;8:138–156. doi: 10.1037/1099-9809.8.2.138. [DOI] [PubMed] [Google Scholar]
  25. National Heart, Lung and Blood Institute. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. 1998 [PubMed] [Google Scholar]
  26. Oyserman D, Bybee D, Mowbray C. Influences of maternal mental illness on psychological outcomes for adolescent children. Journal of Adolescence. 2002;25:587–602. doi: 10.1006/jado.2002.0506. [DOI] [PubMed] [Google Scholar]
  27. Peterson CE, Sastry N, Pebley AR, Ghosh-Dastidar B, Williamson S, Lara-Cinisomo S. DRU-2400/2-LAFANS. Labor & Population. RAND Publication; 2003. The Los Angeles Family and Neighborhood Survey: Codebook. [Google Scholar]
  28. Petterson SM, Albers AB. Effects of poverty and maternal depression on early child development. Child Development. 2001;72:1794–1813. doi: 10.1111/1467-8624.00379. [DOI] [PubMed] [Google Scholar]
  29. Riolo SA, Nguyen TA, Greden JF, King CA. Prevalence of depression by race/ethnicity: Findings from the National Health and Nutrition Examination Study III. American Journal of Public Health. 2005;6:998–1000. doi: 10.2105/AJPH.2004.047225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Roberts RE, Deleger S, Strawbridge WJ, Kaplan GA. Prospective association between obesity and depression: evidence from that Alameda County Study. International Journal of Obesity and Related Metabolic Disorders. 2003;27:514–521. doi: 10.1038/sj.ijo.0802204. [DOI] [PubMed] [Google Scholar]
  31. Rogler LH, Cortes DE, Malgady RG. Acculturation and mental health status among Hispanics. American Psychologist. 1991;46:585–597. doi: 10.1037//0003-066x.46.6.585. [DOI] [PubMed] [Google Scholar]
  32. Sastry N, Pebley AR. Report DRU–2400/7–LAFANS, Labor and Population Program. Santa Monica: RAND; 2003. Non-Response in the Los Angeles Family and Neighborhood Survey. [Google Scholar]
  33. Sastry N, Ghosh-Dastidar B, Adams J, Pebley AR. The design of a multilevel longitudinal survey of children, families, and communities: The Los Angeles Family and Neighborhood Study. Social Science Research. 2006;35:1000–1024. [Google Scholar]
  34. Scarinci IC, Beech BM, Naumann W, Kovach KW, Pugh L, Fapohunda B. Depression, socioeconomic status, age, and marital status in black women: a national study. Ethnicity and Disease. 2002;12(3):421–8. [PubMed] [Google Scholar]
  35. Somervell PD, Leaf PJ, Weissman MM, Blazer DG, Bruce ML. The prevalence of major depression in black and white adults in five United States communities. American Journal of Epidemiology. 1989;130(4):725–735. doi: 10.1093/oxfordjournals.aje.a115394. [DOI] [PubMed] [Google Scholar]
  36. Spence SH, Najman JM, Bor W. Maternal anxiety and depression, poverty and marital relationship factors during early childhood as predictors of anxiety and depressive symptoms in adolescence. The Journal of Child Psychology and Psychiatry and Allied Disciplines. 2002;43:457–469. doi: 10.1111/1469-7610.00037. [DOI] [PubMed] [Google Scholar]
  37. Vega WA, Kolody B, Valle JR, Hough R. Depressive symptoms and their correlates among immigrant Mexican women in the United States. Social Science and Medicine. 1986;22(6):645–652. doi: 10.1016/0277-9536(86)90037-7. [DOI] [PubMed] [Google Scholar]
  38. Wang J, Patten SB. The moderating effects of coping strategies on major depression in the general population. The Canadian Journal of Psychiatry. 2002;47(2):167–173. doi: 10.1177/070674370204700207. [DOI] [PubMed] [Google Scholar]
  39. Yonkers KA, Ramin SM, Rush AJ, Navarrete CA, March D, Heartwell S, Leveno KJ. Onset and persistence of postpartum depression in an inner-city maternal health clinic system. American Journal of Psychiatry. 2001;158(11):1856–63. doi: 10.1176/appi.ajp.158.11.1856. [DOI] [PubMed] [Google Scholar]
  40. Zhang AY, Snowden LR. Ethnic characteristics of mental disorders in five U.S. communities. Cultural Diversity and Ethnic Minority Psychology. 1999;5(2):134–146. doi: 10.1037/1099-9809.5.2.134. [DOI] [PubMed] [Google Scholar]

RESOURCES