Correlates of Maternal Depressive Symptoms in a National Head Start Program Sample FREE

DEPRESSIVE symptomatology among mothers of young children is a very common and important problem that often goes untreated and undiagnosed.^{1 - 7} The prevalence of a significant number of maternal depressive symptoms in samples of several hundred indigent mothers of young children from primary care clinics or community-based samples has been consistently reported to be in the range of 40% to 50%.^{3 ,8} Maternal depressive symptoms have been shown to adversely affect parents' attitudes and competence.^{9 - 11} Further, children of mothers with depressive symptoms have been shown to exhibit significant physical,¹² mental health,^{9 ,13} and school/behavioral problems.^{14 - 18} Maternal depression in the form of poor responsivity and inappropriate child rearing contribute to children's poor outcomes.^{6 ,12} Emotional regulation, separation, individuation, and consistent discipline are all difficult for mothers with depression.^{6 ,9 ,19} Mothers with depression rarely encourage academic or social achievement in their children or provide stimulating activities that promote achievement, often as a result of their feelings of helplessness, preoccupation with their own situations, and decreased memory.^{6 ,18} For very young children, depressed maternal affect seems to be a major factor; long-term exposure to abnormal parental behavior as the child gets older is even more damaging.⁶

Despite the negative effect maternal depression can have on children, earlier studies in this area have focused on smaller samples typically collected within 1 site or community.^{3 ,8 ,20} The purpose of this multisite, cross-sectional study of 5820 children who attended a Head Start program and their families was to describe key aspects of the family environment and demographic factors related to maternal depressive symptoms in a diverse national sample of low-income families whose children attended Head Start as preschoolers and were attending kindergarten at the time of the interview.

The lifetime prevalence for major depression in women is 21.3%, compared with 12.7% for men. Most episodes of depression are treated in the primary care setting, not by mental health professionals.²¹ A major depressive disorder is diagnosed according to a specific set of criteria²² including at least 2 weeks of a depressed mood, loss of pleasure, or both occurring on most days and for most of each day. In addition, 4 or more symptoms from a list of 8 symptoms (eg, insomnia, decrease in ability to concentrate) must be present during the same 2-week period. Of course, sadness is a common response to disappointment, failure, or loss. In 1996, the Diagnostic and Statistical Manual for Primary Care, Child and Adolescent Version (DSM-PC)²³ was published. The DSM-PC describes a wide range of feelings of sadness, including sadness as a normal response, sadness that is a problem but is too mild to meet Diagnostic and Statistical Manual of Mental Disorders, Revised Fourth Edition (DSM-IV) diagnostic criteria, and major depressive disorder as defined by the DSM-IV.

Depressive disorder and depressive symptoms in the absence of disorder are associated with limitations to several areas of patient well-being and functioning when compared with patients with no chronic conditions. In addition, the only chronic condition with an effect on well being and functioning that is as strong as depression is a current heart condition.²¹ In short, patients with symptoms not severe enough to qualify for the diagnosis of depression have impaired working and social lives as well as many unexplained physical symptoms that lead to greater use of medical services.²²

Primary care physicians are often the professionals who work with young mothers on a regular basis. Using a validated screen, primary care physicians can play a key role in early screening and identification of mothers with depression.^{6 ,13 ,24 - 26} Several screening questionnaires and interview protocols have been developed to aid in the screening for and diagnosis of depression. These methods have been primarily of 2 types: self-report measures of symptoms (eg, Center for Epidemiologic Studies Depression scale²⁷ ) and structured interview protocols that identify specific disorders (eg, the National Institute of Mental Health Diagnostic Interview Schedule²⁸ ), which uses the criteria set forth in the DSM-IV.²⁹ By far, the self-report measures are less time consuming and cheaper to administer; however, they do not provide a specific diagnosis. While structured interview protocols do provide a specific diagnosis, the diagnostic process is very costly and often inconvenient, which may hinder their use in studies with large numbers of patients and in clinical settings.

Burnam et al³⁰ developed an 8-item questionnaire that combines self-report symptom items with traditional diagnostic assessment items to serve as the first of a 2-stage screening procedure to identify depressive disorders in a primary care population. Compared with the lengthy Diagnostic Interview Schedule, the instrument had a sensitivity of 96% and specificity of 95% for detecting a current depressive disorder in a primary care population. Its clinical utility is somewhat limited owing to the scoring requirement of using a logistic regression equation. Therefore, to increase its clinical utility, Kemper and Babonis⁴ assessed the psychometric properties of a reduced 3-item set of the 8-item instrument that is easily scored. Compared with the 8-item instrument, the 3-item version had a sensitivity of 100%, specificity of 88%, and a positive predictive value of 66%. The current national study used this 3-item validated screen for depression.

All low-income families are needy, but this group is heterogeneous.³¹ It is critically important that data on family needs are collected from a diverse national sample of families to inform primary care physicians and lawmakers. This study describes family demographics, resources, and difficulties (eg, stress, social isolation) for a national sample of 5820 children who attended a Head Start program and their families during their child's kindergarten year.

Data were collected as part of the evaluation of the National Head Start/Public School Early Childhood Transition Demonstration Project, which was funded by the Administration on Children, Youth, and Families. This project, legislated in 1990 by Congress with reauthorization in 1994, was developed to foster collaboration between Head Start agencies and local education agencies to help low-income children and their families make successful transitions from Head Start programs to the primary grades of school. The program was designed to provide supportive social and health services for the child and family, activities to encourage parental involvement, and use of developmentally appropriate educational practices in schools. Through a competitive application process,³¹ independent local sites (in 30 different states and in the Navajo Nation) were selected to participate in the project.

This article includes data from the 30 local sites, including more than 5000 children from more than 430 schools in 81 school districts. Data included in these analyses were collected in the fall of 1992 or 1993 when the child was in kindergarten via in-person interviews with families and children and written questionnaires completed by teachers.

Details of the sample used in this study are described elsewhere.³² The study is based on 2 cohorts of mothers of children who attended a Head Start program, for a combined total of 5820 mothers. The first cohort of children and families were assessed in the fall of 1992 and the second cohort was assessed in the fall of 1993 during the children's kindergarten year. Families were randomly selected by each site for participation.

Information about family demographics, structure, resources, and life circumstances was drawn from the Family Background Interview³³ administered to the mother during the fall of the kindergarten year.

Information about maternal depression was obtained via the Parent Health and Depression Questionnaire,⁴ a 3-item screening tool abstracted from the Mental Health Battery of the RAND Health Insurance Survey. These 3 items are discussed in the data analysis section. The screening tool compares favorably with the RAND instrument, with a sensitivity of 100%, a specificity of 88%, and a positive predictive value of 66%.⁴

Interviews were administered in the native language of the respondent for those speaking English, Spanish, Vietnamese, or Hmong. All instruments were pretested with a Head Start sample and/or had been used in previous research involving low-income, culturally diverse children and families. Analyses were conducted using a commercial software program.³⁴

The following data analysis strategy was conducted:

(1) To assess the rate of positive depression screens among mothers whose children attended Head Start programs, a validated depression screen⁴ was employed. The depression screen was computed from the following 3 items:

(a) "Would you say that you have ever felt depressed? If yes, how many times in the last week has this statement been true for you? I have felt depressed . . . No days, 1 to 2 days, 3 to 4 days, 5 to 7 days."

(b) "In the past year, have you had 2 weeks or more during which you felt sad, blue, or depressed, or lost pleasure in things that you usually cared about or enjoyed?"

(c) "Have you had 2 or more years in your life when you felt depressed or sad most days, even if you felt okay sometimes?"

If mothers responded yes to 2 or 3 of these statements, the screen was scored "positive." Item a was scored as positive if the mother reported feeling depressed 1 or more days during the past week. Therefore, mothers with a positive depression screen represent a subgroup of mothers who had at least 2 depressive symptoms.

(2) To determine the most salient individual and family ecological factors associated with a positive maternal depression screen, the authors selected independent variables to include in the analysis based on their own clinical experience and a review of the literature on maternal depression. Several studies^{35 - 38} have found an association between maternal depression and social isolation as well as stress. Therefore, in addition to basic demographic variables (eg, maternal education, age) the investigators included variables from the national data set that they believed to be "proxies" for stress (eg, number of moves in past 2 years, number of children in the home, receiving public assistance) and social isolation (eg, number of adults in the home). Marital status was not included in the national data set.

Characteristics of the sample are given in Table 1. Most family incomes fell below $18,000 per year, with almost half earning less than $9600 per year. Most mothers were unemployed, receiving some type of public assistance, and had not moved in the past year. Approximately equal percentages (one third) of mothers did not complete high school, completed high school or had earned a high school equivalency diploma, or attended college.

Approximately 1 of 3 mothers reported that they were the only adult living in the home. Fifty-five percent of the mothers stated they had at least 3 children in the home. About one third of the mothers reported that the child's father helped with parenting duties on a regular basis. Three percent of the mothers reported they had a chronic health problem sufficient to interfere with their ability to care for their child.

The first question addressed the rate of depression among mothers of children who attended Head Start programs. Two (41%) of 5 mothers screened positive for depression. A majority of the mothers (80%) reported that they had ever felt depressed. More than half of the mothers (52%) reported that in the past year they felt depressed or lost pleasure in things that they usually cared about or enjoyed for 2 weeks or more, and more than one third of the mothers (35%) reported that they had felt depressed or sad most days, even if they felt okay sometimes, for 2 years or more.

Table 2 presents the crude odds ratio (OR) and 95% confidence interval (CI) from the logistic regression for each of the independent predictors. These results report on the prediction of a positive maternal depression screen for each variable. The presence of each variable—chronic maternal health problems, homelessness, and low income—elevated the risk of maternal depression at least 2-fold, with the largest effects seen for maternal chronic health problems (OR, 2.66; 95% CI, 1.98-3.59). The number of residential relocations made in the past year also related significantly to maternal depression screens. Mothers who had moved 2 or more times in the past year were more likely to have a positive depression screen than mothers who had not moved in the past year (OR, 1.64; 95% CI, 1.35-2.00). There were no significant differences in depression screens between mothers who had moved once in the past year vs those who had not moved in the past year.

Each of the 3 types of public assistance—living in public housing, receiving Aid to Families With Dependent Children (AFDC), and receiving Supplemental Security Income (SSI)—significantly increased the risk of maternal depression. The ORs were the highest for families that received SSI (OR, 1.61; 95% CI, 1.37-1.89), followed by families that received AFDC (OR, 1.46; 95% CI, 1.31-1.63), then families that lived in public housing (OR, 1.32; 95% CI, 1.16-1.49). Of the total sample, 14% were receiving AFDC and living in public housing, 6% were receiving AFDC and SSI, 3% were receiving SSI and living in public housing, and 2% were receiving AFDC and SSI and living in public housing. Furthermore, mothers who received SSI and AFDC were more likely to be depressed than those who did not receive SSI or AFDC (OR, 1.62; CI, 1.31-2.02).

Mothers who did not have a high school diploma were 1.45 time as likely (95% CI, 1.29-1.62) to have a positive screening for depression as those who did have a high school diploma or pass the General Educational Development test. Mothers who were unemployed were 1.37 times as likely (95% CI, 1.23-1.52) to have a positive screening for depression as those employed. Furthermore, mothers living in the South, Midwest, and West were more likely to be depressed than mothers living in the Northeast.

The number of adults and children in the home showed statistically significant effects on maternal depression screens, although the magnitude of these effects was relatively small. Women who received no parenting assistance from the father had slightly higher rates of maternal depression. Specifically, families with 2 adults in the home were less likely to have mothers who screened positive for maternal depression than those with only 1 adult (OR, 1.40; 95% CI, 1.26-1.57). Homes with 4 or more children showed slightly higher odds of positive maternal depression screens than those with 1 child (OR, 1.32; 95% CI, 1.11-1.57). The OR for maternal depression screens in households with fathers providing no parenting assistance compared with those households with fathers providing parenting assistance was 1.43 (95% CI, 1.28-1.60). Thus, it seems that the active involvement of the father in the child's upbringing may decrease positive maternal depression screens somewhat.

To determine the most salient individual and family ecological factors associated with positive maternal depression screens, a multivariate logistic regression analysis was performed. Each variable entered into the mulivariate logistic regression was independently significantly associated with maternal depression screens (95% CI). As indicated in Table 3, the strongest effects were shown for mothers who had a chronic health problem, were homeless, and/or had extremely low-income levels. These results parallel those found in the univariate analyses. Supplemental Security Income and moving 2 or more times in the past year continued to be strongly associated with rates of maternal depression screens; namely, those families who receive SSI were 1.43 times more likely (95% CI, 1.20-1.70) to have positive maternal depression screens and families who moved 2 or more times in the past year were 1.37 times more likely (95% CI, 1.10-1.71) to have positive maternal depression screens than those who did not move in the year.

An interesting finding was that rates of positive maternal depression screens varied as a function of region of the United States. Compared with mothers in the Northeast, mothers in the Midwest were 1.41 times as likely (95% CI, 1.19-1.67) to have positive maternal depression screens. Similar results were found for mothers in the South (OR, 1.37; 95% CI, 1.6-1.62) and the West (OR, 1.30; 95% CI, 1.08-1.56).

Additionally, the number of people living in the home greatly affected the rate of positive maternal depression screens. Having a large number of children in the home (ie, ≥4) vs having only 1 child in the home significantly elevated the risk of a positive maternal depression screen (OR, 1.29; 95% CI, 1.07-1.57). Single mothers were also 1.17 times as likely (95% CI, 1.02-1.34) to have positive screens for depression than mothers who had at least 1 other adult in the home. Furthermore, mothers who did not receive help from the child's father with parenting duties were 1.19 times as likely (95% CI, 1.05-1.35) to screen positive for maternal depression than mothers who did receive help from the child's father.

Rates for positive maternal depression screen were also elevated for mothers who had less than a high school education (OR, 1.28; 95% CI, 1.13-1.45). The effects of AFDC, public housing, and maternal employment, however, did not continue to be significant predictors in the multivariate model.

In a recent report, the Institute of Medicine identified 5 risk factors associated with the onset of depression: having a parent or other close biological relative with a mood disorder; experiencing a severely stressful event; having low self-esteem, a sense of low self-efficacy, and a sense of helplessness or hopelessness; being female; and living in poverty.³⁹ This national study of depressive symptoms in mothers of children in kindergarten who attended a Head Start program supports the predictive validity of several of the risk factors published by the Institute of Medicine and corroborates findings from several earlier studies that examine depression in mothers of young children. The prevalence of maternal depressive symptoms reported by screening this large national sample of indigent mothers interviewed between 1992 and 1993 is similar to the prevalence reported for low-income mothers of young children at a Baltimore pediatric primary care clinic in 1984 (41% vs 35%, respectively).⁸ In addition, the extent of family poverty in this study has a "dose-response" association with maternal depressive symptoms that is similar to that reported in another (smaller) national sample from the 1990s.¹⁵ In both studies, as well as this study, mothers with lower incomes reported higher levels of depressive symptoms. Maternal depressive symptoms remained inversely associated with income in this study, even after controlling for several factors. In addition, the indigent mothers in this sample who were homeless had a high rate of depressive symptoms, compared with the indigent mothers in this sample who lived in homes (adjusted OR, 2.00; 95% CI, 1.45-2.77).

MOTHERS' REPORT of their health status was another strong independent predictor of maternal depressive symptoms. Although only 158 (3%) of mothers reported that they suffered from a chronic illness sufficient to interfere with regular parenting duties, they were more almost 3 times more likely (adjusted OR, 2.77; 95% CI, 1.98-3.87) to report depressive symptoms than were mothers who were not chronically ill. Maternal chronic illness is one of several chronic stressors that have been associated with mothers' depressive symptoms.^{1 ,3}

Similarly, families that receive SSI were also more likely to have a mother screen positive for depression than families who did not receive SSI. Depressive symptoms and chronic medical conditions have unique, additive effects on adults' ability to function adequately at home and in the community.⁴⁰

In addition to poverty and health-related factors, the number of children and adults living in the home as well as family mobility are additional chronic stressors independently associated with maternal depressive symptoms. These chronic stressors have been shown to affect significantly the mothers' ability to care for themselves and their children.^{5 ,13 ,41}

Furthermore, low income is strongly associated with poor parental mental and physical health.^{40 ,42} Parental irritability and depressive symptoms have been associated with fewer interactions and more conflictual interactions with older children, leading to less satisfactory emotional, social, and cognitive development.⁴³ Specifically, the parents' emotional state and parenting has been shown to greatly affect their children's social adjustment, self-esteem, social competence, and externalizing as well as internalizing behaviors.^{10 ,13} As noted by the Institute of Medicine, there is an intergenerational transmission of depressive symptoms.¹⁷ Whether this relationship is due to poverty, home environment, family structure, family resources, social support, or other factors warrants further research.

This national study has focused on risk factors for a positive maternal depression screen, which is itself a risk factor for poor child health and development outcomes.^{9 - 18} This "parental pathway" through which low income influences children suggests some general recommendations for both primary care physicians and maternal and child health public policy advocates to improve maternal and child health status. Primary care physicians can identify parents in their practice struggling with depressive symptoms and depressive disorders by using the screening instrument described in this article and by keeping in mind the risk factors identified by the Institute of Medicine.³⁹ During the clinic visit, primary care physicians should observe the parent's posture, demeanor, voice, and tone; note any previous depression in the parent (from the depression screen) and inquire about a family history of depression; and ask about the parent's mood, feelings, and relationships.⁴⁴ It is extremely important that primary care physicians show that they care about the depressed parent and the child during the clinical encounter. The primary care physicians may be the important "ray of hope" for the depressed parent. It is also important that the primary care physicians tell the parent that depression is common, use the word "hope," and emphasize the real possibility of recovery from symptoms.⁴⁴ It has been shown that individuals with depressive symptoms are responsive to therapy and it is critically important that parents understand their positive prognosis.^{21 ,45}

From a community health and public policy perspective, about half of the effect of family income on children's cognitive ability is the home environment.⁴⁶ Thus, interventions to improve cognitive development in low-income children must focus on the parents' mental health and their ability to provide the children with adequate learning opportunities. Community interventions that should have a positive effect on the parents' depressive symptoms and the home environment include interventions that are started prenatally rather than postnatally and provide family support in the home at least weekly rather than less frequently.⁴⁷

Maternal depression is a common, important national problem for low-income mothers of preschool and school-aged children. It interferes more with social functioning than with chronic physical illnesses such as hypertension and diabetes.⁴⁸ Maternal depressive symptoms reverberate throughout the family and adversely affect children and their mothers.

The results of this large national study support earlier findings from smaller samples that found maternal depression to be associated strongly with factors amenable to interventions (eg, poverty and chronic illness). The early identification of maternal depressive symptoms by primary care physicians and the ongoing development of effective preventive strategies and treatments can have highly substantial public health implications for the prevention of major child health and development problems, service use, and health care costs.⁴⁹

The analyses reported here were based on the national core data sets collected for the National Evaluation of the Head Start/Public Schools Early Childhood Transition Demonstration Project directed by Sharon Landesman Ramey, PhD, and Craig Ramey, PhD, which was designed and implemented by the National Transition Demonstration Consortium. The Consortium comprises Principal Investigators and Project Directors from each of the 31 local sites, Principal Investigators and staff from the Civitan International Research Center, staff from the Administration on Children, Youth and Families, and a National Research Advisory Panel for the project. For a full description of the Consortium and its members, see Head Start Children's Entry into Public School: An Interim Report on the National Head Start-Public School Early Childhood Transition Demonstration Study.²⁷

Accepted for publication December 23, 1998.

The National Evaluation is supported by grants from the Head Start Bureau of the Administration on Children, Youth, and Families to each of the 31 local demonstration sites, and by a coordinating contract (105-91-1541) to the Civitan International Research Center at the University of Alabama at Birmingham.

We would especially like to thank Carl Brezausek, MS, for assistance in data preparation and coordination.

Corresponding author: Robin Gaines Lanzi, 1719 Sixth Ave South, Civitan International Research Center, University of Alabama at Birmingham, Birmingham, AL 35294-0021 (e-mail: rlanzi@uab.edu).

Editor's Note: When I contemplate the likelihood of the proposed interventions being enacted, I get depressed.—Catherine D. DeAngelis, MD