University Libraries

0 downloads 0 Views 492KB Size Report
Jun 10, 2017 - 765-285-1324 between 8:00 am – 5:00 pm during the academic year. .... Caribeña de Seguridad Alimentaria [ELCSA]; Melgar et al. 2012) that ...
Ball State University Libraries A destination for research, learning, and friends

Notice: Warning Concerning Copyright Restrictions The copyright Law of the United States (Title 17, United States Code) governs the making of photocopies or other reproductions of copyrighted material. Any electronic copy or copies, photocopies or any other type of Reproduction of this article or other distribution of this copyrighted Material may be an infringement of the Copyright Law. This copy is not to be “used for any purpose other than private study, scholarship, or research [section 107]. “If a user makes or later uses any form of reproduction of this copyrighted work for purposes in excess of section 107, Fair use, that user may be liable for copyright infringement.

This material is provided by the Ball State University Libraries. If you have questions concerning this material or are unable to access an electronic document, contact Interlibrary Loan Services via email at [email protected] or by telephone at 765-285-1324 between 8:00 am – 5:00 pm during the academic year.

University Libraries

Arch Womens Ment Health (2017) 20:561–568 DOI 10.1007/s00737-017-0736-7

ORIGINAL ARTICLE

Poor early childhood outcomes attributable to maternal depression in Mexican women Filipa de Castro 1 & Jean Marie Place 2 & Aremis Villalobos 1 & Rosalba Rojas 1 & Tonatiuh Barrientos 1 & Edward A. Frongillo 3

Received: 19 July 2016 / Accepted: 24 May 2017 / Published online: 10 June 2017 # Springer-Verlag Wien 2017

Abstract We aimed to estimate the population fraction of poor early child health and developmental outcomes attributable to maternal depressive symptoms (DS) contrasting it between low- and middle/high-income households. We used a nationally representative probabilistic sample of 4240 children younger than 5 years old and their mothers, derived from the Mexican National Health and Nutrition Survey Data (ENSANUT 2012). Complex survey design, sampling, and analytic weights were taken into account in analyses. DS was measured by CESD-7. Child outcomes were as follows: breastfeeding, attending well-child check-ups, respiratory disease, diarrhea and general health problems, immunization, accidents, growth, obesity, and food insecurity. Prevalence of DS among mothers was 21.36%. In low-SES households, DS was associated with higher risk of never being breastfed (RR = 1.77; p < .05), health problems (RR = 1.37; p < .05), acute respiratory disease (RR = 1.51; p < .05), accidents requiring child hospitalization (RR = 2.16; p < .01), and moderate or severe food insecurity (RR = 1.58; p < .001). In medium- or high-SES households, DS was associated with higher risk of never attending a developmental check-up (RR = 2.14; p < .05) and moderate or severe food insecurity (RR = 1.75; p < .01). Population risks attributable to DS ranged from 2.30 to 17.45%. Prevention of DS could lead to

* Jean Marie Place [email protected]

1

Reproductive Health Division, Center for Population Health Research, National Institute of Public Health, Cuernavaca, Mexico

2

Department of Nutrition and Health Science, Ball State University, Muncie, IN, USA

3

Department of Health Promotion, Education, and Behavior, University of South Carolina, Columbia, SC, USA

reduction of problematic early childhood outcomes in both low and medium/high SES. Keywords Maternal depression . Child health . Mexican National Health and Nutrition Survey Data (ENSANUT)

Introduction Globally, depression is the leading mental health problem in mothers, producing serious consequences that extend to their children (Instituto Nacional de las Mujeres [INMUJERES] 2006). The negative effects of maternal depression in children are stronger in resource-deprived environments, where links with preterm birth, low birth weight (Grote et al. 2010), undernutrition (Surkan et al. 2011), diarrheal diseases, and early cessation of breastfeeding (Patel et al. 2002; Falceto et al. 2004; Galler et al. 2006) have been observed. Other consequences include reduced likelihood of children receiving preventive healthcare (Minkovitz et al. 2005) and daily vitamin supplementation (Lieberman 2002) and an increased likelihood of children starting school without being adequately prepared due to developmental delays (Salt et al. 1988; Galler et al. 2004; Black et al. 2007). Several mechanisms have been hypothesized to explain how maternal depression affects children (Goodman and Gotlib 1999), including the following: (1) a genetic propensity for depression; (2) innate dysfunctional neurobiological mechanisms; (3) exposure to negative maternal cognitions, behaviors, and affect; and (d) exposure to chronic stress. The effect of maternal depression on children likely results from a combination of these mechanisms (Bouvette-Turcot et al. 2017). Socioeconomic status (SES) has been proposed as a moderator of maternal depression and negative children outcomes; children are less negatively affected by maternal

562

depression if mothers have higher SES (Lovejoy et al. 2000; Stein et al. 2008; Goodman et al. 2011; Pearson et al. 2013). Recent calls have been made to increase the knowledge about maternal depression, its impact on specific child outcomes, and the potential role of moderating conditions such as SES (Wachs et al. 2009). SES is particularly important given the influence of maternal depressive symptoms on child development in rural versus urban settings, where availability of resources and access to social services differ greatly. Little is known about the extent and impact of maternal depression in children from low- and middle-income countries (LMIC; Rahman et al. 2002). The self-reported prevalence of common postpartum mental disorders like depression in LMIC is 20.8% (Fisher et al. 2012), exceeding that of developed countries where a meta-analysis of 13 high-income countries reported a rate of postpartum depression at 13% (O’Hara 1996). In Mexico, the national prevalence of maternal depressive symptomatology in women with children younger than age five was 19.9% in 2012—consistent with Fisher and colleague’s estimate, meaning that at least 4.6 million children live with mothers who may experience moderate to severe depression (deCastro et al. 2015a). Among Mexican women, depression is known to be associated with lack of maternal self-efficacy and increased negative attributions of the baby, compromising the ability of the mother to interpret and respond to the baby’s needs (deCastro 2010; Navarro et al. 2011). To our knowledge, no study has examined the association between maternal depression and poor developmental outcomes in children and the potential moderating role of SES in Mexico. Identifying factors that put children at risk for poor outcomes is fundamental to inform policy oriented at promoting early child health and development in LMIC. We aimed to examine the association of maternal depressive symptoms with a set of early child health and developmental outcomes and to estimate the fraction of poor outcomes in children attributable to maternal depressive symptoms in both low- and high-SES mothers.

Methods We used a sample of 4240 children younger than 5 years old and their mothers derived from the Mexican National Health and Nutrition Survey Data (ENSANUT 2012). ENSANUT was a nationally representative cross-sectional complex survey with a probabilistic, multistage, stratified, clustersampling design. Further information regarding the survey design is available elsewhere (Romero-Martinez et al. 2013). Ensanut 2012 is representative at the national, regional, and state levels, as well as of the rural and urban strata. Data were collected between October 2011 and May 2012 in 50,528 households with a response rate of 87%. The households

de Castro F. et al.

interviewed were distributed across all 31 states and the Federal District and represented the estimated 29,429,252 households that conformed Mexico in 2012 according to the 2010 Census and population growth estimates. The ethics review board at the National Institute of Public Health approved this study, and informed consent was obtained from all individual participants. Maternal depressive symptomatology (DS), or the frequency of women reporting moderate to severe depressive symptoms in the past 7 days, was measured by CESD-7, validated for use in Mexico, with a cut-off point at 9. While the instrument is not a diagnostic tool, it can indicate probable cases with a sensitivity of 0.90 and specificity of 0.82 based on the gold standard of the Diagnostic Statistical Manual of Mental Disorders (DSM-V) (Salinas-Rodriguez et al. 2013). Early childhood outcomes included the following variables. Developmental check-ups was measured by the number of well-child visits that the mothers or primary caregiver reported that their child, age 0–12 months, had attended. Breastfeeding was measured by the number of weeks or months mothers or primary caregivers reported breastfeeding their last child. Exclusive breastfeeding among infants less than 6 months was measured by those who received only breast milk the day prior to survey administration. Respiratory disease was assessed by whether the child had episodes of flu, colds, sore throat, cough, bronchitis, or earaches, accompanied by fever, in the previous 2 weeks. Pneumonia was assessed by whether the mother or the primary caregiver reported a doctor diagnosing the child with pneumonia in the last 3 months. Diarrhea was measured by whether the mother or primary caretaker affirmed an episode of diarrhea among children younger than age 5 in the 2 weeks prior to survey administration. Measles vaccine was measured by whether or not the mother or primary caretaker affirmed the measles, mumps, and rubella vaccination among children under age 5, as determined through written records or verbal response. Correct knowledge about oral rehydration therapy (ORT) was measured by providing correct answers regarding which foods could be provided to the child in case of acute diarrhea. Accidents was measured by report of an accident occurring in the previous year that required hospital admission among infants younger than age 1 and children younger than age 5. Anthropometric variables were directly measured and pre-calculated in the ENSANUT database to reflect stunting (low height-for-age), wasting (low weight-for-height), underweight (low weight-for-age), and overweight. Overweight among children younger than age 5 was measured by calculating the z-score for body mass index (BMI) by age. Nutrition indicators were calculated according to the 2006 WHO standards; children were classified as overweight or obese if they had z-scores above two standard deviations and as low height-for-age if they had z-scores below two standard deviations. Household Moderate or Severe Food Insecurity was measured by the Latin American and

Poor early childhood outcomes

Caribbean Food Insecurity scale (Escala Latinoamericana y Caribeña de Seguridad Alimentaria [ELCSA]; Melgar et al. 2012) that included 15 items for households with children younger than 18 years old; the households were classified into four categories ranging from no food insecurity to severe food insecurity. For the multivariable models, we created a dichotomous variable for food insecurity according to whether or not the households experienced food insecurity (light, moderate and severe). The SES indicator was generated imputing deciles of income level to the households on the sample, using demographic and socioeconomic characteristics, and based on the National Income and Expenditure Survey 2010.1 The resulting SES indicator adequately described heterogeneity on standard socioeconomic variables, such as education level of the head of household, income, access to services, and household assets (Gutierrez 2013). For analytic purposes, SES was categorized as low vs medium or high based on the index measured by the reported income, housing characteristics, and possession of household goods. Multivariable models were adjusted for rural household, defined as households in areas with populations less than 2500 inhabitants, and having more than four children. For statistical analysis, we first calculated weighted proportions and 95% confidence intervals for socio-demographic characteristics and early childhood outcomes, stratified by depression status. Group differences between mothers with and without DS were assessed for socio-demographic and early childhood outcomes using Wald test for complex samples (Lohr 2000). Multivariable models to assess the association between each child outcome and maternal depression were built including all variables significant at 10% or less in the bi-variate models. Because binomial models for prevalence ratios generate confidence intervals that tend to be narrower than they should be (Zou 2004), we estimated a general linear model with log-link function, Poisson distribution, and robust standard errors (Zou 2004; Cummings 2009). We calculated the population fraction of each early childhood outcome (AFpop) attributable to exposure to DS using Miettinen’s formula to allow for adjusted risk ratios (Miettinen 1974): AF pop ¼ pc AF exp ¼ pc

RR−1 RR

where pc is the percentage of cases exposed to DS in the population and RR is the risk of early childhood outcome (e.g., risk of not being breastfed) associated with DS. Because the survey was cross-sectional, we substituted risk ratios with prevalence ratios. All analyses accounted for the stratified two-stage cluster sample design for both parameter

1 INEGI. Encuesta Nacional de Ingresos y Gastos de los Hogares 2010.Aguascalientes: Instituto Nacional de Estadística y Geografía, 2011

563

estimates and standard errors by using procedures for complex survey designs (svy suite, Stata 13.0).

Results Maternal DS was significantly associated with moderate or severe food insecurity (47.74 vs 28.31%) and was marginally (but not significantly) associated with increased frequency of no developmental check-ups (34.1 vs 21.46%) and never having been breastfed (13.59 vs 9.79%) (Table 1). Children of mothers reporting more DS had a significantly higher frequency of any health problem (31.99 vs 25.75%) and experiencing a respiratory disease with fever in the past 2 weeks (23.03 vs 17.7%) and marginally (but not significantly) higher having an accident resulting in injuries requiring hospital admission in children younger than 1 year (4.1 vs 0.79%), but not in children between 2 and 5 years old. From multivariable models, children in low SES households whose mother reported DS, compared with children of mothers with no DS, had significantly 1.77 times higher prevalence of no breastfeeding, 1.30 times higher prevalence of presenting health problems, 1.37 times higher prevalence of presenting acute respiratory disease, 2.17 times higher prevalence of accidents, and 1.58 times higher prevalence of living in a household with moderate or severe food insecurity (Table 2). In medium or high SES households, children of mothers with DS, compared with children of mothers with no DS, had significantly 2.14 times higher prevalence of never attending a developmental check-up and 1.75 times higher prevalence of living in a household with moderate or severe food insecurity. All multivariable models had overall p < 0.01. In low SES households, DS accounted for 10.84% of the new cases of children receiving zero developmental check-ups, 5.89% of babies never breastfed, 6.22% of children with at least one health problem, and 17.45% of cases with food insecurity (Table 3). In medium or high SES households, DS accounted for 11.46% of the new cases of children receiving zero developmental check-ups and 12.17% of the cases with food insecurity.

Discussion In this nationally representative study in Mexico, maternal DS was significantly associated with health problems, respiratory diseases, and food insecurity among all mothers. In addition, maternal DS was significantly associated with children never having been breastfed and accidents among low SES mothers and with zero check-ups among medium or high SES mothers. The PAR results provide compelling motivation to create efforts to prevent or reduce maternal DS, given that the lack of developmental check-ups and food insecurity would decrease in households regardless of SES if maternal DS was addressed

564

de Castro F. et al.

Table 1

Maternal depressive symptomatology (n = 4239, N* = 3,296,100) Depressed 21.36% [95% CI: 19.44–23.42]

Zero developmental check-up

34.1 [23.15,47.05]

Non-depressed 78.64% [95% CI: 76.58–80.56]

p

21.46 [17.26,26.35]

0.0743

Never breastfed

13.59 [10.2,17.89]

9.79 [8.417,11.35]

0.0801

Exclusive BF 4 months Exclusive BF 6 months

19.01 [7.96,38.92] 16.06 [6.69,33.8]

16.88 [11.72,23.71] 14.63 [10.23,20.5]

0.7936 0.8385

Health problems Respiratory disease

31.99 [27.34,37.02] 23.03 [18.89,27.76]

25.75 [23.51,28.12] 17.70 [15.84,19.74]

0.0200 0.0198

Pneumonia

2.02 [0.9,4.38]

0.70 [0.45,1.19]

0.1176

Diarrhea (2 weeks) No measles vaccine

12.3 [9.27,16.16] 12.77 [9.35,17.19]

10.24 [8.65,12.09] 10.35 [8.7,12.25]

0.2966 0.2616

Poor knowledge of ORT

16.53 [7.25,33.42]

6.49 [3.62,11.36]

0.1337

4.10 [1.68,9.64] 4.73 [2.91,7.60]

0.79 [.2367,2.60] 3.75 [2.88,4.86]

0.0819 0.4371

Wasting Stunting

2.65 [1.31,5.29] 15.78 [12.19,20.19]

2.10 [1.45,3.03] 14.69 [12.84,16.75]

0.5968 0.6335

Underweight Overweight Household moderate or severe food insecurity

4.09 [2.35,7.00] 10.5 [7.73,14.10] 47.74 [43.02,52.50]

3.18 [2.48,4.07] 9.72 [8.35,11.30] 28.31 [25.93,30.81]

0.4475 0.6633