Neighborhood Deprivation Predicts Infant Sleep Quality

Abstract

Objectives

The current study examined the relationship between neighborhood deprivation and infant sleep at 3 months of age.

Methods

Neighborhood and sleep data were collected from 80 African-American infants and their caregivers. A composite neighborhood deprivation score was created using census data. Infant sleep was measured via 7 nights of actigraphy monitoring when infants were 3 months of age. Current analyses considered the average number of infant night wakings as an index of sleep quality. Multilevel models were used, in which children (Level 1) were nested within census tracts (Level 2).

Results

Controlling for level 1 covariates, greater neighborhood deprivation (b = .07, p < .01) was associated with poorer infant sleep, as characterized by a greater number of wakings during the nighttime sleep period.

Conclusions

Findings suggest that infants who reside in communities marked by higher deprivation experience poorer quality sleep, even after controlling for family-level factors.

Sleep is an important biopsychosocial process, as it predicts cognitive, emotional, and behavioral functioning across the lifespan (1). Prior research has examined the effect of proximal influences on the development of sleep, including bedtime routines, parental psychopathology, and co-sleeping behaviors (e.g. 2–4). However, there is a need for work examining the broader contexts outside of the home environment that might play a role, such as the characteristics of the community within which a child resides (1). While researchers have begun to analyze the relations between community-level factors and sleep in older children and adolescents (e.g., 5–6), there have been no studies that focus on these associations early in the life course. Given that infancy is a critical period for the development of healthy sleep patterns, and these patterns may predict later sleep problems (7–8), it is important to examine the broader contextual factors that influence sleep during this period.

The first few months of life represent a period of rapid change for infant sleep (e.g., 9). Prior to 3 months of age, sleep is not yet consolidated, with infants sleeping in short bursts several times throughout the day (10). However, at 3 months of age, total sleep duration decreases from roughly 16 to 14 hours, with the majority of this occurring during the nighttime period (11). Notably, studies have found a decrease in number of nighttime wakings from birth through 3 months, and then a plateau for the remainder of the first year (e.g., 9). Since sleep becomes more stable from 3 months of age onward, it is crucial to understand the external factors, such as the neighborhood context, that influence sleep behaviors at this time period.

Studies that have expanded their analyses to broader contexts have found that poor neighborhood conditions are associated with negative sleep outcomes. In a study of school-aged children, higher neighborhood economic deprivation, defined using census data on income and public assistance variables, was related to fewer sleep minutes and poorer sleep efficiency (5). Similarly, in a sample of Latino adults, greater self-reported neighborhood disorder, as characterized by more unclean areas and crime, was associated with poorer sleep quality and more sleep disturbances (12). Additionally, in the only study to date to examine these associations in an entirely African-American adult sample, greater objective neighborhood disorder was related to poorer sleep efficiency and increased likelihood of short sleep duration (13). While the exact mechanisms through which the neighborhood impacts sleep are unclear, researchers have proposed links through environmental features such as noise levels and perceptions of safety (5–6).

In infancy, children indirectly interact with their neighborhood context through their relationships with caregivers and their exposure to the home environment. Prior studies have found links between disadvantaged neighborhoods and both poor parenting and low-quality features of the home such as an unsafe play area (14–15). Since sensitive parenting and a safe sleep environment are important for high-quality sleep in infants and children (3, 16), it is crucial to disentangle the links between the neighborhood context and sleep in the early life course. However, before more closely exploring these potential mechanisms, it is important to first establish whether, similar to adolescent and adults, there is an association between neighborhoods and infant sleep.

The current study aims to expand upon prior findings and explore the association between neighborhood deprivation and sleep in infancy, as measured by number of night wakings at 3 months of age.

Methods

The current study utilizes data from the Neonatal and Pediatric Sleep (NAPS) Study, a prospective, longitudinal study of 103 African-American infants and their caregivers. Women who resided within a 50 mile radius of a large public university in North Carolina and had an infant less than 3 months of age were recruited via public birth records, social media, and flyers. The current study includes only the participants (N = 80) for which at least 3 nights of actigraphy data were available at 3 months of age.

Procedure

In-home data collection visits took place when infants were 3 months old. During the daytime portion of the home visit, infant-caregiver dyads took part in parent-child interaction tasks and caregivers completed questionnaires. Beginning on the night after the home visit, families completed a 1-week sleep assessment, including 7 days and nights of actigraphy monitoring. At the conclusion of each home visit, infants were given a small gift, and mothers received compensation of up to $130 in the form of a gift card. All procedures were approved by an institutional review board, and participants provided written consent prior to data collection.

Measures

Neighborhood deprivation

Census tracts were identified using participant addresses. Participants resided in a total of 67 different census tracts. The 2015 American Community Survey was used to collect data on four tract-specific measures: percent below poverty, percent receiving food stamps or SNAP, percent with less than a high school diploma, and percent of families with a female head-of-household and children under the age of 18. These four indicators of deprivation were selected for their consistency with prior literature examining the association between neighborhood characteristics and health behaviors (5–6, 17). The American Community Survey is a nationally representative, yearly survey conducted by the U.S. Census Bureau (18). The 2015 version was utilized here because it provides information regarding the conditions the majority (64%) of children were born into (birth range = October 14 – March 2016). Census measures were highly correlated (rs = .70 to .88, p < .001) and showed high internal consistency (α = .94). A composite neighborhood deprivation score was created by standardizing and summing these four measures.

Infant sleep

Infant sleep was measured via actigraphy throughout the nighttime period (7PM – 7AM) for 7 nights. A lightweight actogram (Actiwatch-2) was placed on the infant’s left ankle at the beginning of the home visit. Caregivers were told to keep the actogram on the infant for the entire study week, except during baths. Motion was measured in 15-s epochs, and data from the actogram were edited using Phillips Actiware software (version 6.0). The activity threshold for scoring the infant as awake was set to the Automatic setting (.888 × average activity count), which is consistent with prior studies (e.g., 19). A complete description of the actigraphy data collection has been published elsewhere (20). Actigraphy data were averaged across the 7-night period. Within our sample, 62 participants (77.5%) had 7 nights, 13 participants (16.25%) had 6 nights, 3 participants (3.75%) had 5 nights, and 2 participants (2.5%) had 4 nights of actigraphy data. For our analyses, we examined the average number of night wakings (>5 minutes) at 3 months of age.

Controls

Child gender and family sociodemographic risk were treated as Level 1 controls. A composite score for family sociodemographic risk was calculated from five dichotomized variables that denote increased risk (single parenthood, less than a high school diploma/GED, 3 or more children in household, income-to-needs ratio < 1.3, age at first childbirth < 19 years) (21). Detailed information regarding each individual risk factor is included in Table 1. The score indicates the proportion of risk factors relative to number of factors for which data is provided; for example, if data is provided for only three factors, but there is risk present in all three items, the score would be 1. Several other covariates were tested (i.e., gestational age, breastfeeding) that were not significant and were therefore not included in final models.

Table 1.

Descriptive Statistics of Individual Risk Factors

Neighborhood Deprivation Risk Factor	Mean (SD)	Min	Max
Percent Below Poverty	18.7 (13.8)	2.4	49.8
Percent Receiving Food Stamps/SNAP	15.5 (13.9)	0	59.8
Percent with Less than HS Diploma	14.4 (10.6)	0.4	38.8
Percent Female HOH with Children Under 18	12.5 (9.1)	0	36.1
Family Socioeconomic Risk Factor	N	%
Marital Status
Risk – Single	39	48.8
No Risk – Married	31	38.8
Missing	10	12.5
Education
Risk – No Degree	1	1.3
No Risk – ≥ HS Diploma/GED	78	97.5
Missing	1	1.3
Children in Household
Risk - ≥ 3 Children	10	12.5
No Risk - ≤ 2 Children	68	85.0
Missing	2	2.5
Income-to-Needs Ratio
Risk - ≤ 1.30	15	18.8
No Risk - ≥ 1.30	35	43.8
Missing	30	3.8
Age at 1st Childbirth
Risk - ≤ 19 Years Old	8	10.0
No Risk - ≥ 20 Years Old	54	67.5
Missing	18	22.5

Statistical Analyses

To account for the nesting of children within neighborhoods, a multilevel modeling approach was used, where infants (Level 1) were nested within census tracts (Level 2).

Results

Descriptive Statistics

In the final sample of 80 infants, 43.8% were female. Neighborhood deprivation ranged from −4.85 to 9.44 (mean = 0, SD = 3.68), with negative scores indicating less deprivation. Family sociodemographic risk ranged from 0 to .80 (mean = .22, SD = .20). Proportion of individual risk factors are shown in Table 1. Throughout the study period, infants woke an average of 2.23 times per night (range = .71 – 5.50, SD = .85). Neighborhood deprivation was correlated with family sociodemographic risk (r = .21, p = .05) and nighttime wakings (r = .37, p < .01). Family sociodemographic risk was not correlated with nighttime wakings (r = .17, p = .14).

Multilevel Models

An initial random-effects ANOVA model yielded an intraclass correlation coefficient (ICC) of .38, which represents the amount of variance in night wakings that can be attributed to differences among neighborhoods. This high level of variance provided support for the use of a multilevel modeling approach, as it was necessary to account for the nesting of individual children within census tracts.

Controlling for Level 1 covariates, neighborhood deprivation was significantly associated with greater wakings throughout the nighttime period (b = .07, SE = .02, p < .01, Figure 1). Family sociodemographic risk (b = .51, SE = .46, p =.29) and child gender (b = −.18, SE = .17, p = .35) were not associated with night wakings. Each of the individual census measures that comprised the neighborhood deprivation score were independently associated with greater wakings during the night (bs = .021 - .028, p < .05), controlling for Level 1 covariates.

Figure 1.

Predicted Values of Infant Night Wakings Based On Neighborhood Deprivation Score

Discussion

This study assessed the association between neighborhood deprivation and sleep in infancy. We found that, after accounting for family-level sociodemographic characteristics, neighborhood factors significantly predict sleep behaviors at 3 months of age. The results of the current study are consistent with those from prior work examining the association between neighborhood deprivation and sleep in older populations (5, 12–13). While some of the results are modest in effect size, the current study is the first to produce similar findings in a sample of infants, providing support for the notion that these associations are present very early on in the life course.

The current study adds to the growing body of literature focused on examining the influence of risk on early sleep. While prior findings regarding the links between socioeconomic status and sleep in childhood are mixed (e.g., 22–23), this study advances our understanding of this association through objective measurement of the neighborhood via census information, while also taking into account individual family sociodemographic risk. While we treated this sociodemographic risk score as a control variable, the lack of significant associations between this score and infant sleep indicates that it is likely something beyond demographic risk alone that matters. Whether it is noise level, fear of violence, or the way in which a deprived neighborhood affects parenting and overall stress in the home, it seems that this context exerts an influence on infant sleep above and beyond familial demographic risk. These findings advance our understanding of the influence of risk on sleep by underscoring the need to examine contexts larger than demographic risk when considering the many ways in which risk can impact sleep in infancy.

A major strength of this study is its use of objective measurement, including census data and actigraphy, which has been lauded as a more accurate tool for measuring infant night wakings than traditionally relied-upon methods of parent report (24). In addition, the consideration of risk at multiple levels is a strength of this study. By controlling for individual-level risk factors, we demonstrate the independent contributions that neighborhood and individual contexts make to infant sleep. Furthermore, this study is strengthened by its use of an African American sample. Since prior research has provided evidence of race-related sleep disparities throughout the life course (e.g., 25), it is crucial to understand the early manifestations of these disparities. To our knowledge, this study provides the first descriptive evidence of sleep quality in 3-month-old African American infants in a naturalistic setting.

It is important to consider these findings in terms of the study’s limitations. Although census data provides objective information about an individual’s neighborhood, it does not tell us how the individual perceives his or her environment. Individual perceptions might drive the associations between the neighborhood context and sleep, such that a mother perceiving a neighborhood as more disadvantaged could influence her behaviors in ways that hinder sleep for her child. Furthermore, although the participants of this study resided in a number of census tracts, they were all residents of the same larger region. Thus, it may be difficult to generalize these findings to the entire population. Additionally, we were unable to test for mechanistic pathways that may link neighborhood deprivation to infant sleep, including, but not limited to, caregiving behaviors and parental stress.

This study is among the first to examine the role of neighborhoods in the development of infant sleep. While others have examined these important influences on sleep and health more broadly, the focal populations of such studies have been older children and adults. Future research should aim to explain how the neighborhood environment trickles down to impact infant sleep. Examining the more proximal mechanisms by which the environment might impact infant sleep, such as parent-child relations and the physical sleep environment, could help bring these processes to light.