 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
a Section for Child and Family Health Studies, Evanston Northwestern Healthcare, Evanston, Illinois
b Northwestern University, Feinberg School of Medicine, Chicago, Illinois
c Center for Health and Public Service Research, Robert F. Wagner Graduate School Of Public Service, New York University, New York, New York
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS. After a school-based ascertainment of asthma and asthma symptoms in 15 low-income, racially/ethnically diverse public elementary schools, 1149 eligible mothers agreed to participate in a longitudinal study. Mothers either had a child with previously diagnosed asthma or a child with symptoms consistent with possible asthma or were in the randomly selected comparison group in which no child in the household had asthma. During the first interview, mothers responded to questions about their own life stressors, supports and mental health, and their children's health.
RESULTS. In bivariate analyses of a community-based sample of children who share low-income neighborhoods, mothers of children with diagnosed or with possible undiagnosed asthma had more symptoms of depression than did mothers of children who have no asthma. Mothers of children with diagnosed or with possible undiagnosed asthma also experienced more life stressors than did mothers of children without asthma. Using nested linear regression, we estimated a model of maternal symptoms of depression. Most of the variation in Center for Epidemiologic Studies–Depression score was accounted for by life stressors and social support. There were no independent effects of either asthma status or asthma status–specific child health status on maternal symptoms of depression.
CONCLUSION. Children who are under care for chronic conditions such as asthma live and manage their illness outside the clinical setting. Their social context matters, and maternal mental health is related to their children's physical health. Although having a child with asthma may be "just" another stressor in the mother's social context, complex treatment plans must be followed despite the many other pressures of neighborhood and family lives.
Key Words: asthma • stress • social support • maternal depression
Abbreviations: CES-D— Center for Epidemiologic Studies–Depression • CRISYS-R—Crisis in Family Systems–Revised • BPAS+—Brief Pediatric Asthma Screen Plus
Recent national survey data indicate an overall asthma prevalence of 12.2% for children who are younger than 18 years.1 These data reveal dramatic differences in the prevalence of lifetime asthma by racial/ethnic group: Puerto Rican individuals have the highest lifetime asthma rate (19.6%), followed by non-Hispanic black individuals (13.8%), non-Hispanic white individuals (11.1%), and Mexican individuals (6.1%).2 Previous research in clinical samples of children with asthma suggest that their mothers are at greater risk for symptoms of depression,3–5 but this research generally lacks comparison groups of children with no asthma. Therefore, there is no available literature on how much influence the additional stress of childhood asthma has on community levels of maternal depression. The relationship between maternal symptoms of depression and having a child with asthma in a community-based sample is the subject of this article.
Accumulating evidence suggests that children's asthma morbidity may be greater when their mothers have symptoms of depression.6 The mechanism that contributes to the increased morbidity is not clear. If the mental health of the primary caregiver is compromised, then the demands of asthma management may be overwhelming.7 Anderson et al8 found that the family factor that was associated most with undertreatment of asthma symptoms was poor maternal mental health. In addition, symptoms of depression in mothers has been associated with increased use of the emergency department for her child's asthma3 and increased school absence, suggesting difficulty in self-management of asthma symptoms.9 Although the decisions to go to the emergency department and to keep a child home from school are based on the parent's perception of illness and therefore subject to bias, maternal symptoms of depression also have been associated with higher levels of asthma illness morbidity as rated by an asthma specialist.4
Caring for a child with asthma may have an impact on the mother's mental health, and mother's mental health may affect her child's asthma.5 Less often discussed is that mother and child also share a common social context of stressors and supports that may influence both maternal mental health and asthma prevalence. The association of symptoms of depression and life stress is widely known,10–12 and social support is thought to moderate the effects of life stress.10,13,14 In our conceptual model, having and caring for a child with asthma is yet another stressor in the caregiver's life. The question is whether, if one accounts for the social context, the number of maternal symptoms of depression remains significantly associated with having a child with asthma. In this article, we focus on the influence of social context on maternal mental health, taking into account the additional stress of having a child with asthma. Despite the cross-sectional design, we advance the literature on maternal mental health and childhood asthma in several important ways. First, we drew our sample from among participants in a nonclinical, community-based ascertainment of asthma in racially and ethnically diverse low-income Chicago neighborhoods. Second, although the participants shared a similar neighborhood context, we measured their individual social circumstances, that is, the sources of the mothers' life stress (other than childhood illness) and social support. Third, we included a natural comparison group of these children's schoolmates who had no previous diagnosis of asthma. Within this group, we identified children with no respiratory symptoms and children who had respiratory symptoms but no asthma diagnosis. This latter group is unique and interesting for this type of study because it may represent children who have asthma and whose caregivers have not been able to gain access to health care for their children.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1200, we surveyed 15 public elementary schools with a total enrollment of 14171. Students in all classrooms brought the survey home for completion and returned the questionnaire to their teachers. Caregivers participated by returning the survey, yielding 12699 children, or 90% of the school enrollments. Fifty-three percent of caregivers checked the box on the survey that indicated an interest in additional contact regarding the research project. This article reports the results of a subset of variables from the cross-sectional analyses of the first of 3 interviews in this longitudinal panel study of caregivers of children with previously professionally diagnosed asthma (diagnosed asthma) or at least 1 symptom consistent with asthma but no professional diagnosis (possible asthma). Caregivers who were approached for interview were drawn from the pool of 53% who expressed interest in being contacted. Families were selected further using the following criteria: (1) all children who had possible or previously diagnosed asthma and were aged 5 to 12 years and; 2) all black, all white, and a random sample of Latino children. A comparison group of schoolchildren without asthma were drawn randomly from among their schoolmates. This comparison sample was "purified of childhood asthma" further at the time of the interview by ensuring that no child in the household had asthma. The overall response rate was 64% (1245 completed interviews of 1947 attempted eligible respondents). One interview subsequently was dropped because of extensive missing data. The resulting sample of 1244 primary caregivers participated in a 45- to 60-minute interview in English or Spanish that was conducted by trained interviewers under contract to a national survey firm.
The subject of this report are the 1149 (92%) primary caregivers who were the children's mothers or female primary caregivers (included as "mothers"). Twelve percent of the children were white, 34% were black, and 52% were Latino. Just more than half (57%) of the mothers were US born, but 90.6% of the children were US born. Thirty percent of the interviews were conducted in Spanish. Thirty-nine percent of the children were commercially insured, and 12% were uninsured. Twenty-two percent of the white children, 42.1% of black children, and 37.5% of Latino children were on Medicaid. Thirteen percent of the Latino children were uninsured. Sixty-three percent of white children, 40% of black children, and 37% of Latino children were privately insured. Six percent of white children, 10% of black children, and 13% of Latino children had KidCare. The Institutional Review Boards of all participating institutions approved the study.
Measures
Dependent Variable
Maternal Symptoms of Depression
We measured symptoms of depression using the Center for Epidemiologic Studies–Depression (CES-D).15 The CES-D has been validated in similar samples in English and in Spanish. The CES-D is a 20-item self-report measure of depressive symptoms experienced in the past 7 days.16 Response options range on a 4-point scale from 0 (rarely) to 3 (most of the time). Scoring yields a total score with a cutoff of 15; a higher score has been shown to reflect clinically significant symptoms of depression. Schulberg et al17 provided cutoff criteria to categorize depression as mild versus severe. The CES-D is a widely used scale and is internally consistent with 
of .85 or higher among English speakers.15 In addition, this measure has been translated into Spanish and has been validated in previous studies.18,19 Perczek et al18 found that it had a high internal consistency ( = .89) for Spanish speakers. Likewise, Masten19 and others found very high coefficients (between .80 and .86). The coefficients by language group for this sample were .88 for the English version and .86 for the Spanish version.18,20
Independent Variables
Maternal Life Stressors
We measured life stressors with the Crisis in Family Systems–Revised (CRISYS-R)21,22 and the Spanish CRISYS-R.23 The CRISYS-R has been validated in English and in Spanish. We ascertained the occurrence of a series of life stressors during the previous 6 months using the CRISYS-R instrument. The CRISYS-R21,22 has 63 items and 3 dimensions (Valence, Difficulty, and Chronicity). Respondents indicated whether each of the 63 stressors or incidents had happened to them in the 6 months before the interview. When the event had occurred, the respondent rated the experience of the incident as positive, negative, or neutral (Valence); the outcome as resolved or ongoing (Chronicity); and the level of difficulty ascribed to having experienced the stressor (Difficulty). The respondent rated the level of difficulty on a 4-point scale (1 = not at all, 4 = a lot). The average level of difficulty was obtained by dividing the sum of the ratings by the number of stressors reported. Counts of ratings represent the other dimensions, for example, number of total life stressors and number of endorsed stressors rated as negative. For this article, we removed the 2 child health–related questions from the analyses to model child health separately from other sources of life stress.
Maternal Health Status
We measured maternal health status with the SF-12 Health Survey.24 The SF-12 Health Survey measures functional health status of adults using physical and mental component scales. Respondents answer 12 questions about general health, activity and social limitations, pain, and feelings. Responses may range from excellent to poor, very limited to not limited at all, all of the time to never, or not at all to extremely, depending on the item. The physical and mental health components are scored separately; for these analyses we used the physical health score. We used the Spanish translation that was provided by the authors.
Child General Health Status
We measured child health status using the parent form of the Child Health Questionnaire, a generic assessment tool that was developed specifically for children and adolescents who are 5 years and older.25 We used the general health item for the analyses in this report. Mothers rated their child's health in general on a 5-point scale from excellent to poor. We collapsed poor and fair (labeled "fair") because of a small number of respondents who reported poor child health. The questionnaire overall and the general health item in particular have been shown to have good internal consistency and item discriminant validity in a variety of populations. We used the Spanish translation that was provided by the authors.
Maternal Social Support
We measured perception of the level of social support with The Medical Outcomes Study Social Support Survey.26 This brief social support index was developed for patients with chronic conditions and measures perception of social support distinct from structural measures of support. Respondents state how often various kinds of support are available to them on a 5-point scale ranging from "none of the time" to "all of the time." Multitrait scaling analyses support the dimensionality of its 4 functional scales.27 We administered only the emotional/informational support subscale. A higher score indicates more support, and the subscale score is calculated by averaging the scores of its 8 items. This subscale has high internal consistency (Cronbach's = .96) and high 1-year stability (.72). We used the Spanish translation that was provided by the authors. The coefficients by language group for this sample were .94 for the English version and .92 for the Spanish version.
Maternal Language
To flag primary caregivers who might face particular challenges because they are non–English speaking, we divided the Latino caregivers into 2 groups by the language in which they chose to complete the interview. The study was conducted only in English and Spanish, so we do not have representation of other non–English-speaking immigrant groups.
Education
We derived a series of dichotomous variables (less than high school, high school diploma, some college, and college degree or more) from our question on primary caregiver's education.
Insurance
We derived a series of dichotomous variables (private insurance, Medicaid, KidCare [State Children's Health Insurance Program funds], and self-pay) from our question on the child's source of health insurance.
Partner in Home
This variable was dichotomized as yes or no. The question was intended to capture the potential source of support from a cohabiting partner, whether married or not.
Asthma Status
We determined asthma status with the Brief Pediatric Asthma Screen Plus (BPAS+) and the Spanish BPAS+.28,29 The BPAS+ respiratory symptom questions were validated by comparing the questionnaire results with medical history and physical examination by a pediatric asthma specialist. We surveyed students in the schools with the asthma section of the BPAS+ and the Spanish BPAS+. We repeated the BPAS+ at the time of the first interview and used this administration for the analyses in this report. These parent-report questionnaires have been found to be useful for identifying children with diagnosed asthma as well as children who are in need of additional evaluation for possible undiagnosed asthma. The English BPAS+ had 73% sensitivity and 74% specificity for black children28 and 61% sensitivity and 83% specificity for Latino children. The Spanish BPAS+ had 74% sensitivity and 86% specificity.29
Analysis
The full interview sample of 1244 caregivers was weighted to be representative of the larger, screened school population, based on race/ethnicity, language choice (English/Spanish), asthma status (no, possible, or diagnosed asthma), age, and gender. The resulting data set therefore represented the sample of children who were surveyed in 15 schools. All analyses were performed using SAS software, Version 9.1.3 of the SAS System for Windows.30 Results are reported for the 1149 mothers only. We first present descriptive statistics, then bivariate relationships between each of the key variables of interest and symptoms of depression, and finally multivariate models of symptoms of depression. Bivariate analyses included 
2 tests for 2 categorical variables (race/ethnicity–language group and asthma status), Pearson correlations for 2 continuous variables (life stressors and depression symptoms), and 1-way analyses of variance for 1 categorical and 1 continuous variable (asthma status and life stressors). The interpretations of these tests are noted in the text. Multivariate analysis proceeded using nested linear regressions. This technique permits the progressive addition of independent variable groups (entered together on the basis of theory or face validity) to specify further the regression model. This method quantifies the additional contribution of the added variable group to the amount of variation in the dependent variable (symptoms of depression) explained by the model. This additional contribution to the model is noted as the R2 change. In successive models, we regressed the CES-D score on (1) maternal demographics (race/ethnicity–language group and education), (2) sources of life stress other than the child's health (life stressors and caregiver physical health), (3) sources of support (perception of support, insurance, and partner in the home and insurance), (4) the child's overall general health and asthma status, and (5) and interaction terms (child general health*asthma status) to separate the child's asthma status–specific health from the child's overall general health.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Similar to the pattern seen for maternal symptoms of depression, CRISYS life stressor scores for diagnosed and possible undiagnosed asthma did not differ, and both were higher than for mothers of children with no asthma (P < .001). Black mothers experienced more life stressors in the previous 6 months than any other race/ethnicity–language group (P < .001). Having a partner in the home (P < .001) and having commercial insurance for the child (P < .001) each were associated with fewer life stressors in the previous 6 months. Spanish-speaking Latino mothers reported the lowest level of perceived social support (P < .001), and the most highly educated reported the highest level of perceived social support (P < .001).
Maternal Symptoms of Depression, Stressors, Supports, and Childhood Asthma: Nested Multivariate Models
We estimated a series of nested linear regression models of maternal symptoms of depression to identify the independent contributions of life stress, supports, and childhood asthma to maternal depression. In the nested models, we accounted for elements of the social context of the caregivers first, describing the relationships of stressors and supports to the mother's symptoms of depression before we incorporated the additional stress of having a child with asthma. The fully saturated model therefore was used to assess the effect of having a child with asthma on maternal mental health over and above the effect of the social context (Table 3).
|
Model 2: Addition of Life Stressors and the Mother's Physical Health
Sources of life stress for the mother (viz, life stressors and physical health) explained a considerable amount of additional variation in maternal depression, over and above demographics. The total variation of the CES-D score that was explained by this model was 26.7%. The change in R2 from model 1 to model 2 was substantial, 20.8%.
Model 3: Addition of Social Support
Sources of support contributed significantly to the explained variation in maternal symptoms of depression, over and above variables that already were in the model. The total variation of the CES-D score that was explained by adding mothers' life perception of social support, having a partner in the home, and children's source of insurance was 32.4%. The change in R2 from model 3 to model 4 was 6.3%.
Model 4: Addition of Children's Asthma Status and Children's General Health
Children's general health and their asthma status explained relatively little additional variation in the cumulative model of maternal symptoms of depression. There was no relationship between the children's asthma status and the mothers' symptoms of depression, controlling for the child's general health. The total variation in CES-D score that was explained when the children's asthma status and general health were added to the cumulative model was 33.1%. The change in R2 from model 3 to model 4 was 0.7%.
Because there was no systematic difference in depression across asthma groups, we combined the "no asthma" and "possible asthma" groups and reestimated the model. With this modification, the change in the "diagnosed asthma" coefficient was very small (.16), so we used the simplified model for subsequent estimations.
Model 5: Addition of the Interaction of Asthma Status and Child's General Health
We interacted asthma diagnosis with the child's general health to test whether asthma status affected the relationship between maternal depression and child's health. The independent variables in this final, fully saturated model were race/ethnicity–language group, education, life stressors, maternal physical health, perception of social support, presence of a partner, insurance status, child general health, child asthma status, and the interaction of asthma status and child's general health.
In this final model, the race/language group was borderline significant (F = 3.15, P = .056). Specifically, pairwise comparisons showed that controlling for maternal stressors and supports and the children's asthma status and general health, the black subgroup had lower CES-D scores than both the Spanish-speaking Latino (P < .05) and white (P < .05) subgroups (multiple comparison tests not shown). Likewise, the only significant comparison within the education indicator variable was that between having less than a high school education and being college educated (P < .05; multiple comparison tests not shown). Maternal life stressors (P < .0001) and worse physical health (P < .0001) each predicted increasing symptoms of depression. Greater perceived social support was associated with fewer symptoms of depression (P < .0001), as was having a partner in the home (P < .01). Having insurance was unrelated to symptoms of depression. Each decrement in child health was associated with an increment in the mother's symptoms of depression. Children who were in fair health had mothers with CES-D scores that were 2.3 points higher than that for mothers of children who were in excellent health (P < .05). Having diagnosed asthma did not make a significant additional contribution to the relationship between the general health of the child and symptoms of depression in the mother, as evidenced by the nonsignificance of the interaction terms. The change in R2 from model 4 to model 5 was 0.01%. The total variation of the CES-D score that was explained by the final model was 33.2%.
Summary
Using nested linear regression, we created a model of maternal symptoms of depression, with most of the variation in CES-D score accounted for by life stressors and social support. Although reporting that the child was in fair health was associated with more symptoms of depression, there was no additional association of maternal depression with either asthma status or the interaction of asthma status and child general health status. Therefore, it seems neither that a child's asthma status increased maternal depression on average, over and above the child's general health nor that the child's asthma further exacerbated maternal depression at different levels of the child's general health.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The child's actual asthma status did not have a relationship to the mother's symptoms of depression. When the mother reported that the child's general health was fair, she also was likely to report more symptoms of depression. The interaction terms tell us that the children's asthma status–specific health status was not associated with maternal mental health. In other words, in regard to maternal depression, the mothers did not distinguish their children's health as a result of asthma independent of their children's general health.
It should be noted that despite these conclusions and consistent with the predictions in the clinical literature, our bivariate analyses show more symptoms of depression and life stressors among mothers of children with diagnosed asthma compared with children who have no asthma. In addition, we identified another group of children who were at risk for asthma and were not in the clinical samples. The mothers of these children likewise demonstrated more symptoms of depression and life stressors compared with children who had no asthma. Therefore, in regard to maternal life stress and symptoms of depression, children with possible undiagnosed asthma shared a similar social context with children who were under clinical care for asthma that is different from those who do not have asthma.
To the best of our knowledge, this study is the first moderately large-scale, community-based study of maternal mental health and childhood asthma that incorporated a comparison group of children without asthma and accounted for children with possible undiagnosed asthma. It also is the first study of its type to control for neighborhood and the individual social context before examining the additional contribution of having a child with asthma. A study by Stein et al32 shares some similarities, using random-digit dialing to create 2 nonclinical samples: 1 of families who lived in the "inner city" (household N = 654) and the other a national sample (household N = 712). They showed an association between parental anxiety/depression and the presence of a childhood chronic condition for the inner-city sample only. Because of our school selection strategy, our study was urban and low income but not inner city, suggesting that the association between maternal symptoms of depression and chronic illness status may be a more generalized phenomenon.
Limitations
Our estimates of childhood asthma and possible undiagnosed asthma approximate prevalence among schoolchildren because of the high rate of return of the survey questionnaires (90%) and the strong validity of the respiratory questions in the survey instrument. Although we did not conduct community-based ascertainment of maternal symptoms of depression, the weighting of our sample gave us a good approximation of the high levels of depression among women with elementary school–aged schoolchildren in these communities. It clearly is possible that social context affected maternal mental health and childhood asthma directly. We cannot rule out the chance association of 2 high-prevalence conditions.
We note, as others have noted, that when the mother reports that the child is less healthy, she is more likely to have more symptoms of depression33 and worse physical health herself. These findings may reflect an interaction of circumstances and behavior, but because these data all are derived from maternal perception, they also may be confounded by respondent bias. It would require objective measures of health and depression to verify these relationships.
Because of the cross-sectional nature of our design, we cannot make statements about causality. For this, we would need longitudinal data to assess temporal precedence; that is, whether increasing symptoms of depression are preceded by or follow changes in life stressors, social supports, and maternal or child physical health or asthma status. Likewise, we cannot disentangle relationships that arise because of a possible biological connection (eg, stress and asthma morbidity, and stress and depression) or behavioral consequences (eg, depression leading to suboptimal asthma management and thereby worse physical health). Additional work will address these issues.
Implications
Children who are under care for chronic conditions such as asthma live and manage their illness outside the clinical setting. The clinician's awareness of the family context of children's health is important for guiding effective intervention.34 Providing support to caregivers improves their children's asthma status.13,35,36
Despite pediatric health providers' concerns, becoming comfortable eliciting information about maternal mental health, stress, and available supports is critically important. Health care providers may be the mother's single most frequent health professional contact and source for referrals.37,38 The social context matters, and caregivers' mental health is related to their children's physical health. Although having asthma per se may not be causally related to maternal depression, complex treatment plans must be followed despite the many other pressures of neighborhood and family lives. Rather than assuming that having a child with asthma is depressing, it is better to say that when a child with asthma has a caregiver who has symptoms of depression, then a broader case management approach is likely to offer extended benefits to the family that will accrue to better children's health. Improving families' access to needed resources (even if not directly health related) may allow the parent to focus with fewer distractions on asthma care. This approach is equally valid for children who fall "under the radar" with respiratory symptoms but no professional diagnosis of asthma, suggesting that this group would benefit from intervention in primary care or in the community, perhaps through their schools.
|
ACKNOWLEDGMENTS |
|---|
We acknowledge our terrific research team and our colleagues Lisa Sharp, PhD, Molly Martin, MD, and David Shalowitz, BA, for thoughtful comments on early versions of the manuscript.
|
FOOTNOTES |
|---|
Address correspondence to Madeleine U. Shalowitz, MD, MBA, Northwestern University Feinberg School of Medicine, Pediatrics, Evanston Northwestern Healthcare Research Institute, 1001 University Pl, Evanston, IL 60201. E-mail: m-shalowitz{at}northwestern.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. B. Weiss Eliminating Asthma Disparities: A National Workshop To Set a Working Agenda Chest, November 1, 2007; 132(5_suppl): 753S - 756S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
