Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Mar 16.
Published in final edited form as: Int J Contemp Sociol. 2014 Apr;51(1):7–25.

Neighborhood Disadvantage, Preconception Health Behaviors and Infant Birthweight: A Preliminary Study

Lee A Flagg a,✉, Belinda L Needham b, Julie L Locher c
PMCID: PMC4360973  NIHMSID: NIHMS609397  PMID: 25788757

Abstract

Prior studies have shown that neighborhood disadvantage and disorder are associated with birth outcomes. This study examined preconception diet and physical activity level, as well as body mass index, as mediators of the association between neighborhood conditions and birthweight. Secondary data analyses were conducted using data from the National Longitudinal Study of Adolescent Health (Add Health). The final analytic sample consisted of 523 adolescent and young adult mothers giving singleton live births between 1997 and 1998. In contrast to previous research, we found that neighborhood characteristics were unrelated to birthweight. Consistent with prior studies, compared to those who were White, on average, Blacks had birthweights that were 163.25 grams lighter. In addition, compared to mothers who were married or living with a partner, mothers who did not have a partner at the time of birth, on average, had offspring that were 127.20 grams lighter. No evidence was found for the mediation hypotheses as there were no associations between neighborhood characteristics and preconception diet or physical activity or between these behavioral variables and birthweight. To the authors’ knowledge, this is the first study examining diet and physical activity as possible behavioral pathways between the neighborhood context and birthweight.

Keywords: US, Add Health, pregnancy, neighborhood, exercise, diet

Introduction

Infant mortality rates are a key measure of the health of a community or society (Merrill and Timmreck 2006). Infant mortality is considered an indicator of the general health status of a population because many of the environmental factors that affect the health of childbearing women and infants affect the health of the rest of the population as well (Brown, Isaacs, Beate Krinke, Murtaugh, Sharbaugh, Stang, and Wooldridge 2008). However, it is becoming increasingly apparent that infant mortality is not equally distributed throughout cities or neighborhoods. For example, the infant mortality rate is 25 percent higher in the largest 100 U.S. cities compared to the national average (Fitzpatrick and LaGory 2000). A large risk factor for infant mortality is low birthweight (LBW), which can be caused by intrauterine growth restriction (IUGR), preterm birth (PTB), or a combination of the two (Abu-Saad and Fraser 2010; Masi, Hawkley, Piotrowski, and Pickett 2007; Mathews, Menacker, MacDorman, and Centers for Disease Control and Prevention 2004). There is also a growing body of research that examines the relationship between the neighborhood context and birth outcomes and the mediating mechanisms in these relationships, yet no previous studies have taken into consideration maternal diet and physical activity level (Ahern, Pickett, Selvin, and Abrams 2003; Masi, Hawkley, Piotrowski, and Pickett 2007; Morenoff 2003; Schempf, Strobino, and O'Campo 2009).

The purpose of this study is to test the relationship between neighborhood disadvantage and disorder, diet and physical activity, and birthweight using data from the National Longitudinal Study of Adolescent Health (Add Health). This study can have considerable implications for interventions directed towards adolescents and childbearing women to improve their health behaviors and birth outcomes and, in turn, decrease the infant mortality rate in the U.S.

In general, neighborhood disadvantage and disorder are inversely associated with size at birth. For example, neighborhood disorder, measured as block group violent crime counts and rates, increases the odds of LBW among both White and Black mothers.(Messer, Kaufman, Dole, Herring, and Laraia 2006) However, as previously mentioned, LBW can be caused by some form of IUGR or PTB or both and thus is not a particularly informative outcome (Masi, Hawkley, Piotrowski, and Pickett 2007). IUGR is difficult to quantify, but many researchers use small for gestational age (SGA) as an outcome indicating IUGR (Brown et al. 2008). Infants are classified as SGA if they are below the 10th percentile of sex- and parity-specific birthweight distributions for a given gestational age. Measures of neighborhood disadvantage are associated with increased odds of SGA in offspring, and neighborhood disorder, measured as violent crime rate, is associated with SGA among all races/ethnicities (Elo, Culhane, Kohler, O'Campo, Burke, Messer, Kaufman, Laraia, Eyster, and Holzman 2009; Masi, Hawkley, Piotrowski, and Pickett 2007). For Blacks and Hispanics, neighborhood violent crime rate also fully or partially mediates the association between neighborhood disadvantage and SGA (Masi, Hawkley, Piotrowski, and Pickett 2007). In addition, median household income of the neighborhood is positively related to the percentile of birthweight for gestational age (Farley, Mason, Rice, Habel, Scribner, and Cohen 2006).

Researchers have also examined the relationship between neighborhood contextual factors and birth outcomes by considering the entire life course, beginning with the neighborhood context in childhood. Overall, LBW and SGA rates are elevated with increasingly disadvantaged lifetime neighborhood economic context (Love, David, Rankin, and Collins 2010). Additionally, Black women, but not White women, experience weathering in terms of LBW and SGA (Love, David, Rankin, and Collins 2010). Weathering is a phenomenon commonly found among Black mothers, in which the rates of adverse infant outcomes increase with advancing maternal age. It has been hypothesized that weathering is a result of the cumulative effects of socioeconomic disadvantage (Geronimus 1992). These patterns are reversed for Black women who live in advantaged neighborhoods throughout the course of their lives. Overall, it appears that the lifetime neighborhood context can alter the relationship between age and birth outcomes (particularly LBW and SGA) among Black women. Researchers have attributed these patterns to the weathering effect and the decrease in health status with increasing age, as well as an increase in the incidence of pregnancy risk factors (smoking, hypertension, etc.) (Love, David, Rankin, and Collins 2010).

The results of one study indicated that an increase in measures of neighborhood disadvantage and neighborhood disorder and decay are associated with a decrease in birthweight in grams (Schempf, Strobino, and O'Campo 2009). Regardless of race, neighborhood disorder assessed by violent crime rate is negatively associated with birthweight and accounts for much of the association between neighborhood economic disadvantage and birthweight (Masi, Hawkley, Piotrowski, and Pickett 2007). Another study corroborated these results, demonstrating that neighborhood disadvantage is negatively related to birthweight and that this association is fully mediated by violent crime rate and neighborhood exchange/voluntarism. In other words, violent crime, reciprocal exchange, and participation in local volunteer associations explain the relationship between neighborhood disadvantage and birthweight. More specifically, the violent crime rate within a neighborhood is negatively associated with birthweight while neighborhood social relations and engagement are positively associated with birthweight (Morenoff 2003).

Other research suggests that the association between the neighborhood context and birthweight may be explained or partially explained by stress, hypertensive disorders, drug use, and inadequate/delayed prenatal care (Ahern, Pickett, Selvin, and Abrams 2003; Brown et al. 2008; Elo et al. 2009; Farley et al. 2006; Finch, Vega, and Kolody 2001; Morenoff 2003; Ross and Mirowsky 2001; Schempf, Strobino, and O'Campo 2009). However, to the authors’ knowledge, previous research has not considered maternal diet and physical activity level, as well as prepregnancy body mass index (BMI), as other explanatory factors of this relationship.

There appears to be little research on the relationship between the neighborhood context, nutrition, and birth outcomes. The poorest neighborhoods have a decreased prevalence of supermarkets, as well as an increased prevalence of smaller grocery stores and bars/taverns. Predominantly Black neighborhoods, an indicator of neighborhood disadvantage, have a decreased prevalence of supermarkets, full-service restaurants, and carryout specialty places. They also have a higher prevalence of convenience stores (Morland, Wing, Diez Roux, and Poole 2002). Neighborhood disadvantage is characterized by decreased access to amenities such as grocery stores that supply fresh and nutritious foods, and access to supermarkets is associated with an increased consumption of fruits and vegetables and a better dietary intake (Schempf, Strobino, and O'Campo 2009). The nutritional status of the mother in the pre- and periconceptional periods is an important factor in birth outcomes (Abu-Saad and Fraser 2010; Brown et al. 2008).

While there is a considerable amount of research relating the neighborhood context and physical activity, there does not appear to be research specifically related to physical activity prior to or during pregnancy. In one study, results indicated that the neighborhood context was an important predictor of physical activity level in the general population (van Lenthe, Brug, and Mackenbach 2005). Neighborhood disadvantage was measured as the residents' reported education and occupation levels, as well as employment status. Neighborhoods were also characterized according to their proximity to food shops, general physical design of the neighborhood, quality of green facilities, noise pollution, and required police attention. Residents of the most disadvantaged and disordered neighborhoods reported the least amount of sports participation and physical activity during leisure time while being more likely to walk or cycle to shops or work. The positive association between neighborhood disadvantage and disorder and odds of almost never participating in physical activity during leisure time was partially mediated by poorer general physical design of the neighborhood. In addition, the positive association between neighborhood disadvantage and odds of almost never participating in sports was partially explained by increased amounts of required police attention (van Lenthe, Brug, and Mackenbach 2005). Poor general physical design of a neighborhood and increased levels of required police attention appear to discourage physical activity among its residents. The physical environment of the neighborhood and possibly the fear of victimization may prevent residents from partaking in health-promoting leisure time physical activity.

Disadvantaged neighborhoods can be characterized by a lack of parks or recreation facilities and thus may fail to encourage exercise (Fitzpatrick and LaGory 2011). There is some evidence that neighborhood disorder assessed in terms of fear or victimization and lack of perceived neighborhood safety can reduce physical activity in residents (Schempf, Strobino, and O'Campo 2009). For example, parental perceptions of neighborhood safety are positively associated with youth physical activity levels, and parental perceptions of neighborhood resources (e.g., sidewalks, parks, recreation centers, etc.) are positively associated with a healthy weight status among youth (Duke, Borowsky, and Pettingell 2012). Exercise preconception and during pregnancy are associated with a decreased risk for adverse maternal and infant outcomes (Artal and O'Toole 2003; Dempsey, Butler, and Williams 2005; Weissgerber, Wolfe, Davies, and Mottola 2006).

Based on the review of the literature, the authors propose four hypotheses:

  • H1: Neighborhood disadvantage is negatively associated with birthweight.

  • H2: Perceived neighborhood disorder mediates the association between neighborhood disadvantage and birthweight.

  • H3: Diet, physical activity level, and prepregnancy BMI mediate the association between neighborhood disadvantage and birthweight.

  • H4: Diet, physical activity level, and prepregnancy BMI also mediate the association between perceived neighborhood disorder and birthweight.

Methods

The data for this study are from Add Health. Add Health is a nationally representative, longitudinal study of adolescents that collected data from 1994 to 2008. This dataset provides information on social, economic, psychological, and physical well-being, as well as contextual information on the family, neighborhood, community, school, friends, peer groups, and romantic relationships. Overall, Add Health imparts data on the social, behavioral, and biological influences on health across the life course (CPC]).

The target population for the Add Health study was adolescents in grades 7 through 12 across the U.S. starting in the 1994–1995 school year. This cohort has been followed through 2008 into young adulthood (aged 24 to 32 years) (CPC]). The unit of analysis for the current study is individual births given by the adolescents initially enrolled in the Add Health study (adolescence is defined as anywhere from the age of 10 to 24 years old) (Office of Population Affairs n.d.). The analytic sample was restricted to one singleton live birth of at least 510 grams and 20 weeks gestation during 1997 and 1998 to females giving birth at least 3 years after the onset of menses. Only one birth per mother was included to simplify the analysis (two levels instead of three levels). Births of multiples were excluded due to their increased risk for adverse birth outcomes. Birthweights that were reported to be less than 510 grams and gestational ages less than 20 weeks were also excluded because these are generally considered coding errors (Masi, Hawkley, Piotrowski, and Pickett 2007). The sample was restricted to pregnancies that ended during 1997 and 1998 and, therefore, to offspring who were conceived after the first wave of data was collected. Additional years were not included in the analysis because the diet and physical activity data for the analytic sample were collected in 1995 and may not be relevant exposures to births occurring later. Females who reported that the birth occurred prior to 3 years after the onset of menses were excluded due to possible coding errors (i.e., potentially indicating mothers of a very young age at the pregnancy outcome) and the heightened risk for adverse birth outcomes (Brown et al. 2008). Data are from Waves 1 and 3 of the Add Health Study. Wave 1 contains the neighborhood variables and the diet and physical activity variables, while Wave 3 contains the pregnancy and birth files. For the analytic sample, Wave 1 data was collected in 1995 and Wave 3 data was collected from 2001 to 2002. About 73 percent of Wave 1 participants were interviewed in Wave 3 (CPC]). The main advantages of the Add Health dataset are that it is nationally representative and data from Wave 1 is temporally prior to the data in Wave 3. Thus, the dependent variables may be considered endogenous, and some issues of selectivity are addressed because the study sampled adolescents, who cannot self-select into neighborhoods. Since Add Health also surveyed the grandparents (parents of the adolescent mothers), some confounding factors and selection processes can be controlled for in the analyses.

Measurement

Birthweight of the offspring was self-reported in pounds and ounces by the adolescent mother. Birthweight was converted to grams and left as a continuous variable to preserve statistical power, consistent with previous research (Schempf, Strobino, and O'Campo 2009). Births that were less than 510 grams were excluded because they usually indicate coding errors (Masi, Hawkley, Piotrowski, and Pickett 2007). This variable comes from Wave 3 of Add Health.

Gestational age at birth was also reported by the mother in weeks. Again, gestational age was restricted to gestations of at least 20 weeks because reports of less than 20 weeks may be misclassified stillbirths or errors in reporting gestational age. In addition, births that occurred prior to 37 weeks of gestation were considered preterm and those births that occurred at 37 weeks or more of gestation were considered term births (Masi, Hawkley, Piotrowski, and Pickett 2007). These variables also come from Wave 3.

Neighborhood disadvantage was measured by an index of census tract-level measures that were standardized and then summed. These measures included the proportions of families that are below the poverty line, Black residents, female-headed households with dependents, unemployed, and housing that is not owner occupied in the respondent's census tract (α = 0.87), similar to previous research using Add Health data (Burdette and Needham 2012) and other neighborhood research (Schempf, Strobino, and O'Campo 2009). According to Wilson, the census tract is considered the measureable unit most appropriate to represent neighborhoods (Wilson 1996). Principal components analysis indicated that all of the index measures loaded on the same factor (families below the poverty line: 0.88, Black residents: 0.79, female-headed households with dependents: 0.92, unemployed: 0.86, and housing not owner occupied: 0.61) with an Eigen value of 3.37. In addition, 67.36 percent of the variance in this factor was explained by these five indicators. These variables are from Wave 1 of the study.

Grandparental perceived neighborhood disorder was an index of two questions that were asked of the grandparents in Wave 1. Parents of the adolescent mothers were asked to rate the significance of physical and social problems in their neighborhood, which included trash on the sidewalks and drug dealers/drug use. Responses included no problem at all (0), a small problem (1), and a big problem (2). These two items were summed to create an index of perceived neighborhood disorder (physical and social) that ranged from zero to four, also consistent with previous research using Add Health data (Burdette and Needham 2012). These data come from Wave 1 of the study.

Maternal diet was assessed by a series of questions that asked the respondents the number of servings of milk or dairy products, fruit or fruit juice, vegetables, grains, and pastries or sweets consumed on the previous day. Responses to each question included zero servings (0), one serving (1), and two or more servings (2). Each of these five variables comes from Wave 1.

Physical activity level was measured ordinally using responses from a series of questions from Wave 1 of Add Health. Respondents were asked how many times they skated or cycled, played sports, and participated in aerobic exercise during the past week. Responses included not at all (0), one or two times (1), three or four times (2), and five or more times (3).

Prepregnancy BMI at Wave 1 was calculated from self-reported weight and height. BMI is the ratio of weight to height squared (kg/m2 or lb/in2 × 703) ([CDC] 2011).

Grandparent’s highest level of education completed was recoded into four categories: less than high school education (0); high school diploma, GED, or equivalent, including those who went to college, but did not graduate (1); bachelor’s degree (2); and professional training postbaccalaureate (3). Maternal racial/ethnic categories included were White (reference category), Black, Hispanic, and Other. Blacks, Hispanics, and Other were not collapsed into a single category because of their differential associations with birth outcomes, also known as the “Latina Paradox” (Barr 2008). Since the neighborhood variables came from Wave 1, only personal disadvantage from Wave 1 was controlled for in the models (grandparent's education, as well as mother’s race/ethnicity, were used as indicators of SES or personal disadvantage because the respondents were dependents at Wave 1). The adolescent mother’s birthweight was also included in the models (self-reported by the grandparent), which for 82.08 percent of the adolescents was their biological mother. This variable was included to partially address issues of selection and any biological predisposition for adverse birth outcomes. Life-cycle and intergenerational factors, such as the intrauterine environment the mother experienced herself, can influence her reproductive outcomes later in life (Abu-Saad and Fraser 2010). All of these variables come from Wave 1 of Add Health.

Age at the pregnancy outcome was calculated by subtracting the respondent's birth month and year (reported at Wave 1) from the month and year that the pregnancy ended (reported at Wave 3). Marital status was measured nominally with those who were married and/or those who were living with their partners at the pregnancy outcome as the reference category. Prenatal care was also measured nominally with those who accessed prenatal care during the pregnancy as the reference category. In addition, drug use during pregnancy was included in the analyses with those who abstained from alcohol, tobacco, and other drugs (ATOD) as the reference category and those who reported any ATOD use as the comparison category. These variables come from Wave 3 of the Add Health study.

While there was no missing data for the dependent or mediating variables (birthweight, diet, and physical activity), as well as gestational age and age at the pregnancy outcome, means were imputed for missing data on other continuous variables to minimize the loss of cases due to missing data. The variables that were imputed with means for cases with missing values included grandparent’s perceived neighborhood disorder, mother’s BMI, and mother’s birthweight. In addition, those grandparents with missing education data were coded as having a high school education or equivalent. Means were not imputed on missing data for neighborhood disadvantage, but only eight cases were lost due to these missing values.

Analytical Procedure

Prior to calculating any inferential statistics, descriptive statistics were computed for the analytic sample. Because of the multilevel and potentially clustered nature of the data, hierarchal linear modeling (HLM) was initially considered the most appropriate method for the multivariable analyses. HLM estimates regression equations that explain variation in an outcome for individuals that are a part of groups as both a function of individual characteristics and as a function of group characteristics (Arnold 1992). Regular Ordinary Least Squares (OLS) regression assumes that each observation is independent. If individuals are clustered within census tracts, the variance between participants is not constant, but likely varies between schools and neighborhoods as well. Currently, HLM is an appropriate method of addressing this issue, as HLM can model both within and between group variance at the same time and produce more accurate estimates (Arnold 1992). However, the respondents in the analytic sample were not sufficiently nested, meaning there were too few cases per census tract to allow for multilevel modeling. Previous research using data from Wave 1 of Add Health has also found that adolescents were not sufficiently nested within census tracts (median of 2 adolescents per tract) to conduct a multilevel analysis (Cubbin, Santelli, Brindis, and Braveman 2005).

Regardless, multilevel modeling does not account for the effects of multistage clustered sampling designs (Cubbin, Santelli, Brindis, and Braveman 2005). These effects were not taken into account because of the unique characteristics of the analytic sample, which will be discussed further in the following section. Therefore, the results of these analyses cannot be generalized to the entire adolescent population of the U.S.

Multiple mediation hypotheses were tested simultaneously using PROCESS, a SAS macro developed by Hayes in 2012. PROCESS can accommodate up to ten mediators operating in parallel (Hayes 2012). Bias-corrected bootstrap confidence intervals were used to test the significance of the indirect effects in the mediation models (Bollen and Stine 1990; Hayes 2012). If the 95% confidence interval does not contain zero, then the indirect effect is statistically significant at the p < 0.05 level.

Results

Descriptive statistics for the sample are reported in Table 1. It should be noted that this sample had a relatively high percentage of PTB (30 percent), a positively skewed distribution on the neighborhood disadvantage index, an age range of 14 years to 22 years, and a very small percentage of mothers who did not access prenatal care during their pregnancy (1 percent). Therefore, caution should be used when interpreting the results of this analysis, as the population to which the findings apply to is rather specific: adolescent and young adult mothers having singleton births at least 3 years after the onset of menses with presumably good access to prenatal care.

Table 1.

Descriptive Statistics for Grandparental, Maternal, and Infant Characteristics for the Analytic Sample

N Mean/
Proportion		 Standard
Deviation		 Minimum Maximum
Birthweight (g) 535 3264.96 545.66 765.44 4847.77
Gestational age (weeks) 535 37.81 2.96 24.00 40.00
Preterm birth 535 0.30 - - -
Neighborhood disadvantage 527 0.00 4.08 −5.34 14.84
Grandparent’s perceived neighborhood disorder 535 1.25 1.01 0.00 4.00
Diet
  Dairy 535 1.14 0.79 0.00 2.00
  Fruits 535 1.10 0.82 0.00 2.00
  Vegetables 535 0.78 0.78 0.00 2.00
  Grains 535 1.39 0.68 0.00 2.00
  Sweets 535 0.69 0.78 0.00 2.00
Exercise
  Cycling 535 0.31 0.64 0.00 3.00
  Sports 535 0.88 0.99 0.00 3.00
  Aerobics 535 1.64 1.05 0.00 3.00
Mother’s BMI 535 22.85 4.27 15.82 48.46
Grandparent’s education 535 1.23 0.96 0.00 4.00
Mother’s race/ethnicity 535
  White 0.48 - - -
  Black 0.28 - - -
  Hispanic 0.18 - - -
  Other 0.06 - - -
Mother’s birthweight (g) 535 3145.73 535.85 1360.78 4649.32
Mother’s age 535 18.82 1.50 14.00 22.00
Marital status 533
  Married or living with partner 0.55 - - -
  Single 0.45 - - -
No prenatal care 533 0.01 - - -
Drug use 535 0.18 - - -
Open in a new tab

Results from the multivariable analysis predicting birthweight are displayed in Table 2. Neighborhood disadvantage had no direct or indirect effects on birthweight. Perceived neighborhood disorder, as well as the other hypothesized mediators, was not related to birthweight. Compared to Whites, Blacks, on average, have newborns that are 163.25 grams lighter. In addition, a one gram increase in the mother’s birthweight is associated with a 0.12 gram increase in the offspring’s birthweight. Mothers who were not married or living with their partner at the pregnancy outcome had babies that were 127.20 grams lighter. As one would expect, those infants who were born preterm had considerably lighter birthweights (194.16 grams). According to the R-square, this model only explains 11 percent of the variance in birthweight.

Table 2.

Process Model Coefficients of Birthweight (g) Regressed on Neighborhood Disadvantage, Perceived Neighborhood Disorder, Mediators, and Control Variables

Model
(n = 523)		 Direct Effect Indirect Effect
Intercept 3206.06***
Neighborhood characteristics
Neighborhood disadvantage −1.23 −1.23 −1.62
Grandparent’s perceived neighborhood disorder −26.04 −2.18
Diet, Exercise, and BMI
Dairy 42.65 −1.13
Fruit 5.53 0.04
Vegetables 0.19 0.00
Grains −13.13 0.11
Sweets −15.03 0.11
Bicycling 29.59 0.03
Sports 5.58 0.02
Aerobics −24.16 0.34
Mother’s BMI 9.87 1.04
Individual controls
Grandparent’s education 21.09
Blacka −163.25*
Hispanica −85.37
Othera −29.86
Mother’s birthweight (g) 0.12**
Mother’s age −0.86
No partnerb −127.20*
No prenatal carec −110.41
Drug used −90.42
Preterm birthe −194.16***
F 2.87***
R2 0.11
Open in a new tab
*

p < .05,

**

p < .01,

***

p < .001.

a

Compared to White.

b

Compared to married or living with partner.

c

Compared to accessing prenatal care.

d

Compared to no drug use during pregnancy.

e

Compared to full term.

Discussion

Based on a review of the previous literature, it was hypothesized that neighborhood disadvantage and disorder would be negatively associated with birthweight and that these associations would be at least partially explained by prepregnancy diet, physical activity level, and BMI. These hypotheses were tested using data from Add Health. Despite what previous research suggests, no support was found for the hypotheses. There are several limitations of this study to consider, but there are several strengths to consider as well. Combined with the logical connection between neighborhood contextual factors and birthweight, future research should continue to consider the neighborhood environment when studying birth outcomes, but may also need to consider other factors that may explain the variation in these outcomes.

Strengths and Limitations

The major strengths of this study are the use of a large, national dataset and the prospective cohort design of Add Health. To the authors’ knowledge, it is also the first study to consider diet and physical activity as possible behavioral pathways in the associations between neighborhood contextual variables and birthweight. In addition, there is a plausible biological link between the neighborhood context and birth outcomes, which is one of the criteria for increasing the strength of evidence for causation (HILL 1965). This argument will be outlined later in the following section.

As with any study, there are also limitations to consider. A major hurdle for neighborhood studies is determining whether associations are a result of a real effect or composition. However, individual-level disadvantage was included in the models and there was no evidence for an association between the neighborhood context and birthweight in this study. One of the main limitations of this study was the measurement of the mediators: diet and physical activity level. The questions used to assess these factors ask about these behaviors over the course of one day or one week and may not fully reflect a person's health behavior patterns over time. In addition, because these measures captured such a small exposure, the analysis was restricted to births occurring within 3 years of the assessment of these variables, restricting the sample size as well as the age range of mothers. While the sample size was still adequate, the study may not have been powered enough to detect very small, but statistically significant, associations. The relatively young ages (14 to 22 years) of the analytic sample generalizes to a very unique population that may be more likely to experience neighborhood and personal disadvantage. It is possible that there was not enough variance in these variables to detect associations. While it seemed that there was a large range in the neighborhood disadvantage index among the respondents, it also appears that this variable was positively skewed.

Linking the Neighborhood Context to Birth Outcomes

Despite these limitations, there are plausible biological pathways between the greater neighborhood context and birthweight. Neighborhoods may affect stress levels, prepregnancy BMI, diet, and physical activity (Burdette and Hill 2008; Duke, Borowsky, and Pettingell 2012; Grow, Cook, Arterburn, Saelens, Drewnowski, and Lozano 2010; Messer, Kaufman, Dole, Savitz, and Laraia 2006; Morland, Wing, Diez Roux, and Poole 2002; Ross and Mirowsky 2001; Schempf, Strobino, and O'Campo 2009; van Lenthe, Brug, and Mackenbach 2005; Wadhwa, Dunkel-Schetter, Chicz-DeMet, Porto, and Sandman 1996). Appropriate diet and physical activity may affect birth outcomes through the prevention of gestational diabetes mellitus (GDM) and hypertensive disorders of pregnancy (Artal and O'Toole 2003; Brown et al. 2008; Dempsey, Butler, and Williams 2005). More specifically, diet and physical activity may be used to improve maternal stress levels, adiposity, glucose tolerance, systolic and diastolic blood pressure, plasma lipid and lipoprotein concentrations, oxidative stress, concentrations of C-reactive proteins and proinflammatory cytokines in peripheral circulation, endothelial cell function, and blood flow to organs, which are underlying pathophysiologies of both GDM and hypertensive disorders of pregnancy (Abu-Saad and Fraser 2010; Artal and O'Toole 2003; Brown et al. 2008; Dempsey, Butler, and Williams 2005; Wadhwa et al. 1996; Weissgerber, Wolfe, Davies, and Mottola 2006; [ACSM] 2010). These conditions are associated with adverse birth outcomes (Brown et al. 2008; Dempsey, Butler, and Williams 2005; Masi, Hawkley, Piotrowski, and Pickett 2007; Mocarski and Savitz 2012; Wadhwa et al. 1996; Weissgerber, Wolfe, Davies, and Mottola 2006). While no support was found for the hypotheses, some of the limitations of the current study and the plausible link between the neighborhood and birth outcomes suggest that future research could still take into consideration these factors and explore these various pathways further.

Conclusion

Based on the review of the literature, it was expected that neighborhood characteristics would be associated with birthweight and that prepregnancy diet, physical activity, and other factors would mediate these associations. Results from the analyses did not support these hypotheses. Despite these findings, previous research and theory still suggest there may be a causal relationship and there are plausible biological pathways in which the neighborhood context may influence birth outcomes. Future research should use nationally representative longitudinal or experimental data and consider these possible contextual, behavioral, and biological factors in birth outcomes. Ideally, these studies should aim to include a wide range of maternal ages, as well as relevant exposures of health behaviors, such as diet and physical activity. Further investigation of these relationships, as well as consideration of other factors than can influence birth outcomes, may prove to be worthwhile in promoting the health of mothers and their children in the U.S.

Acknowledgements

This research uses data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from grant P01-HD31921 for this analysis.

The first author is supported by the UAB Predoctoral Training Program in Obesity-Related Research (T32 HL105349). The last author is supported by the UAB Nutrition Obesity Research Center (P30 DK056336). The contents of this report are solely the responsibility of the authors and do not necessarily represent the views of the NIH or any other organization with which the authors are affiliated.

References

  1. Abu-Saad K, Fraser D. Maternal nutrition and birth outcomes. Epidemiol Rev. 2010;32(1):5–25. doi: 10.1093/epirev/mxq001. [DOI] [PubMed] [Google Scholar]
  2. Ahern J, Pickett KE, Selvin S, Abrams B. Preterm birth among African American and white women: a multilevel analysis of socioeconomic characteristics and cigarette smoking. J Epidemiol Community Health. 2003;57(8):606–611. doi: 10.1136/jech.57.8.606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. ACSM’s guidelines for exercise testing and prescription. Philadelphia, PA: Lippincott Williams & Wilkins; 2010. American College of Sports Medicine [ACSM] [Google Scholar]
  4. Arnold Carolyn L. An introduction to hierarchical linear models. Measurement and Evaluation in Counseling and Development. 1992;25(2):58–90. [Google Scholar]
  5. Artal R, O’Toole M. Guidelines of the American College of Obstetricians and Gynecologists for exercise during pregnancy and the postpartum period. Br J Sports Med. 2003;37(1):6–12. doi: 10.1136/bjsm.37.1.6. discussion 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Barr Donald A. Health disparities in the United States: Social class, race, ethnicity, and health. Baltimore, MD: Johns Hopkins University Press; 2008. [Google Scholar]
  7. Bollen Kenneth A, Stine Robert. Direct and indirect effects: Classical and bootstrap estimates of variability. Sociological methodology. 1990;20(1):15–140. [Google Scholar]
  8. Brown Judith E, Isaacs Janet S, Krinke U Beate, Murtaugh Maureen A, Sharbaugh Carolyn, Stang Jamie, Wooldridge Nancy H. Nutrition through the life cycle. Belmont, CA: Thomson Wadsworth; 2008. [Google Scholar]
  9. Burdette AM, Hill TD. An examination of processes linking perceived neighborhood disorder and obesity. Soc Sci Med. 2008;67(1):38–46. doi: 10.1016/j.socscimed.2008.03.029. [DOI] [PubMed] [Google Scholar]
  10. Burdette AM, Needham BL. Neighborhood environment and body mass index trajectories from adolescence to adulthood. J Adolesc Health. 2012;50(1):30–37. doi: 10.1016/j.jadohealth.2011.03.009. [DOI] [PubMed] [Google Scholar]
  11. Centers for Disease Control and Prevention [CDC] Healthy Weight: Assessing Your Weight: BMI: About Adult BMI. 2011 Retrieved March 3, 2012 ( http://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html#Interpreted). [Google Scholar]
  12. Cubbin C, Santelli J, Brindis CD, Braveman P. Neighborhood context and sexual behaviors among adolescents: findings from the national longitudinal study of adolescent health. Perspect Sex Reprod Health. 2005;37(3):125–134. doi: 10.1363/psrh.37.125.05. [DOI] [PubMed] [Google Scholar]
  13. Dempsey JC, Butler CL, Williams MA. No need for a pregnant pause: physical activity may reduce the occurrence of gestational diabetes mellitus and preeclampsia. Exerc Sport Sci Rev. 2005;33(3):141–149. doi: 10.1097/00003677-200507000-00007. [DOI] [PubMed] [Google Scholar]
  14. Duke NN, Borowsky IW, Pettingell SL. Parent perceptions of neighborhood: relationships with US youth physical activity and weight status. Matern Child Health J. 2012;16(1):149–157. doi: 10.1007/s10995-010-0731-3. [DOI] [PubMed] [Google Scholar]
  15. Elo IT, Culhane JF, Kohler IV, O’Campo P, Burke JG, Messer LC, Kaufman JS, Laraia BA, Eyster J, Holzman C. Neighbourhood deprivation and small-for-gestational-age term births in the United States. Paediatr Perinat Epidemiol. 2009;23(1):87–96. doi: 10.1111/j.1365-3016.2008.00991.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Farley TA, Mason K, Rice J, Habel JD, Scribner R, Cohen DA. The relationship between the neighbourhood environment and adverse birth outcomes. Paediatr Perinat Epidemiol. 2006;20(3):188–200. doi: 10.1111/j.1365-3016.2006.00719.x. [DOI] [PubMed] [Google Scholar]
  17. Finch BK, Vega WA, Kolody B. Substance use during pregnancy in the state of California, USA. Soc Sci Med. 2001;52(4):571–583. doi: 10.1016/s0277-9536(00)00161-1. [DOI] [PubMed] [Google Scholar]
  18. Fitzpatrick Kevin, LaGory Mark. Unhealthy places: The ecology of risk in the urban landscape. New York: Routledge; 2000. [Google Scholar]
  19. Fitzpatrick Kevin, LaGory Mark. Unhealthy cities: Poverty, race, and place in America. New York: Routledge; 2011. [Google Scholar]
  20. Geronimus AT. The weathering hypothesis and the health of African-American women and infants: evidence and speculations. Ethn Dis. 1992;2(3):207–221. [PubMed] [Google Scholar]
  21. Grow HM, Cook AJ, Arterburn DE, Saelens BE, Drewnowski A, Lozano P. Child obesity associated with social disadvantage of children’s neighborhoods. Soc Sci Med. 2010;71(3):584–591. doi: 10.1016/j.socscimed.2010.04.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hayes Andrew F. PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling. Manuscript submitted for publication. 2012 [Google Scholar]
  23. Hill AB. THE ENVIRONMENT AND DISEASE: ASSOCIATION OR CAUSATION? Proc R Soc Med. 1965;58:295–300. doi: 10.1177/003591576505800503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Love C, David RJ, Rankin KM, Collins JW. Exploring weathering: effects of lifelong economic environment and maternal age on low birth weight, small for gestational age, and preterm birth in African-American and white women. Am J Epidemiol. 2010;172(2):127–134. doi: 10.1093/aje/kwq109. [DOI] [PubMed] [Google Scholar]
  25. Masi CM, Hawkley LC, Piotrowski ZH, Pickett KE. Neighborhood economic disadvantage, violent crime, group density, and pregnancy outcomes in a diverse, urban population. Soc Sci Med. 2007;65(12):2440–2457. doi: 10.1016/j.socscimed.2007.07.014. [DOI] [PubMed] [Google Scholar]
  26. Mathews TJ, Menacker F, MacDorman MF N.tional Center for Health Statistics Centers for Disease Control and Prevention. Infant mortality statistics from the 2002 period: linked birth/infant death data set. Natl Vital Stat Rep. 2004;53(10):1–29. [PubMed] [Google Scholar]
  27. Merrill Ray M, Timmreck Thomas C. Introduction to epidemiology. Sudbury, MA: Jones and Bartlett Publishers; 2006. [Google Scholar]
  28. Messer LC, Kaufman JS, Dole N, Herring A, Laraia BA. Violent crime exposure classification and adverse birth outcomes: a geographically-defined cohort study. Int J Health Geogr. 2006;5:22. doi: 10.1186/1476-072X-5-22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Messer LC, Kaufman JS, Dole N, Savitz DA, Laraia BA. Neighborhood crime, deprivation, and preterm birth. Ann Epidemiol. 2006;16(6):455–462. doi: 10.1016/j.annepidem.2005.08.006. [DOI] [PubMed] [Google Scholar]
  30. Mocarski M, Savitz DA. Ethnic differences in the association between gestational diabetes and pregnancy outcome. Matern Child Health J. 2012;16(2):364–373. doi: 10.1007/s10995-011-0760-6. [DOI] [PubMed] [Google Scholar]
  31. Morenoff JD. Neighborhood mechanisms and the spatial dynamics of birth weight. AJS. 2003;108(5):976–1017. doi: 10.1086/374405. [DOI] [PubMed] [Google Scholar]
  32. Morland K, Wing S, Diez Roux A, Poole C. Neighborhood characteristics associated with the location of food stores and food service places. Am J Prev Med. 2002;22(1):23–29. doi: 10.1016/s0749-3797(01)00403-2. [DOI] [PubMed] [Google Scholar]
  33. Ross CE, Mirowsky J. Neighborhood disadvantage, disorder, and health. J Health Soc Behav. 2001;42(3):258–276. [PubMed] [Google Scholar]
  34. Schempf A, Strobino D, O’Campo P. Neighborhood effects on birthweight: an exploration of psychosocial and behavioral pathways in Baltimore, 1995--1996. Soc Sci Med. 2009;68(1):100–110. doi: 10.1016/j.socscimed.2008.10.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. The University of North Carolina at Chapel Hill Carolina Population Center [UNC CPC] Add health. Retrieved October 19, 2010 ( http://www.cpc.unc.edu/projects/addhealth)
  36. U.S. Department of Health and Human Services Office of Population Affairs [US DHHS OPA]. n.d. What is Adolescence? Retrieved August 14, 2013 ( http://www.hhs.gov/opa/familylife/tech_assistance/etraining/adolescent_brain/Overview/what_is_adolescence/#2).
  37. van Lenthe FJ, Brug J, Mackenbach JP. Neighbourhood inequalities in physical inactivity: the role of neighbourhood attractiveness, proximity to local facilities and safety in the Netherlands. Soc Sci Med. 2005;60(4):763–775. doi: 10.1016/j.socscimed.2004.06.013. [DOI] [PubMed] [Google Scholar]
  38. Wadhwa PD, Dunkel-Schetter C, Chicz-DeMet A, Porto M, Sandman CA. Prenatal psychosocial factors and the neuroendocrine axis in human pregnancy. Psychosom Med. 1996;58(5):432–446. doi: 10.1097/00006842-199609000-00006. [DOI] [PubMed] [Google Scholar]
  39. Weissgerber TL, Wolfe LA, Davies GA, Mottola MF. Exercise in the prevention and treatment of maternal-fetal disease: a review of the literature. Appl Physiol Nutr Metab. 2006;31(6):661–674. doi: 10.1139/h06-060. [DOI] [PubMed] [Google Scholar]
  40. Wilson William Julius. When work disappears: The world of the new urban poor. New York: Vintage Books; 1996. [Google Scholar]

RESOURCES