1
Racial/ethnic disparities in sleep in mothers and infants during the Covid-19 pandemic
Authors: Maristella Lucchini
1
, Margaret Kyle
1
, Nicolò Pini
1
, Ayesha Sania
1
, Vanessa
Babineau
1
, Morgan R. Firestein
1
, Cristina R. Fernández
2,3
, Lauren C. Shuffrey
1
, Jennifer R
Barbosa
1,4
, Cynthia Rodriguez
4
, William P. Fifer
1,4
, Carmela Alcántara
6
, Catherine Monk
1,4,5
,
Dani Dumitriu
3,1,7
1
Division of Developmental Neuroscience, Psychiatry Department, Columbia University Irving
Medical Center, New York, NY USA
2
New York-Presbyterian Hospital, New York, NY USA
3
Division of Child and Adolescent Health, Department of Pediatrics, Columbia University Irving
Medical Center, New York, NY USA
4
New York State Psychiatric Institute, New York, NY
5
Department of Obstetrics and Gynecology, Columbia University Irving Medical Center, New
York, NY USA
6
School of Social Work, Columbia University, New York, NY USA
7
Sackler Institute, Zuckerman Institute, and the Columbia Population Research Center, Columbia
University, New York, NY USA
Corresponding author:
Dani Dumitriu, MD, PhD
Pardes Rm 4932, 1051 Riverside Drive, New York, NY 10032
dani.dumitriu@columbia.edu
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted March 26, 2021. ; https://doi.org/10.1101/2021.03.22.21254093doi: medRxiv preprint
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
2
ABSTRACT
Study Objectives: To quantify the association between race/ethnicity and maternal and infant
self-reported sleep health at 4 months, exploring the role of maternal depression, stress and
symptoms of trauma related to the COVID-19 pandemic as potential mediators. Methods:
Participants were recruited as part of the COVID-19 Mother Baby Outcomes (COMBO) cohort at
Columbia University (N=71 non-Hispanic White, N=14 African American (AA), N=113 Hispanic,
N=40 other/declined). Data on infant sleep were collected at 4 months postpartum. A subset of
149 women also completed questionnaires assessing maternal mental health and sleep.
Multivariable regressions were used to separately estimate associations of race/ethnicity and
mental health with multiple sleep domains for infants and mothers adjusting for individual-level
covariates. Results: Compared to non-Hispanic White, Hispanic infants slept less at night (β=-
101.7±17.6, p<0.0001) and AA and Hispanic infants went to bed later (respectively β =1.9±0.6,
p<0.0001, β=1.7±0.3, p<0.0001). Hispanic mothers were less likely to perceive infant sleep as a
problem (β=1.0±0.3, p=0.006). Compared to non-Hispanic White mothers, Hispanic mothers
reported worse maternal sleep latency (β=1.2±0.4, p=0.002), and efficiency (β=0.8±0.4, p=0.03),
but better subjective sleep quality (β=-0.7±0.4, p=0.05), and less daytime dysfunction (β=-0.8±0.4,
p=0.04). Maternal mental health scores were statistically significant predictors of multiple domains
of maternal sleep but did not mediate the association between race/ethnicity and sleep.
Conclusions: Racial/ethnic disparities in maternal and infant sleep are observable at 4 months
post-partum. Maternal stress, depression and symptoms of trauma related to the COVID-19
pandemic did not mediate these associations.
KEYWORDS: sleep, race, ethnicity, disparities, infant, mother, post-partum, COVID-19
INTRODUCTION:
Sleep health is an essential component of general health and it is associated with numerous
mental and physical health outcomes in the general population.
1,2
The postpartum period poses
unique challenges to mothers with respect to maintaining good sleep health.
3
In the perinatal
period, poor sleep has been associated with numerous maternal physical and mental outcomes,
such as changes in interpersonal relationships, risk of postpartum weight retention, and perinatal
mood disorders.
4–6
For the infant, the early postnatal period is characterized by rapid development
of sleep and wake patterns, marking a major neurobiological milestone. Infants’ sleep problems
have been reported in up to 15–25% of infants
7
and are a source for concerns for both parents
and pediatricians. Such sleep problems have been reported to negatively affect physical,
cognitive, and socioemotional development.
8–11
In addition to the direct adverse impact on infants
and young children, several studies have indicated that poor sleep in infancy and childhood also
affects parental levels of stress, depression, sense of competence, and overall quality of life,
12–14
which could further compound the negative impact on infant development.
Despite these known risks, increasing evidence indicates that we are in the middle of a sleep
crisis, with 70 million American adults estimated to suffer from chronic sleep loss and sleep
disorders.
15
This crisis also applies to children such that the average child sleep duration has
decreased over the last century, with children sleeping 0.75 min less per year, on average, since
1905,
16
and almost one third of children in the United States getting a sub-optimal amount of sleep
at least once a week.
17
However, the burden of poor sleep health is not experienced equally
across the U.S. population. Individuals from racial/ethnic minority backgrounds have been shown
to experience worse sleep health,
18–20
but less is known about racial/ethnic disparities in sleep in
the post-partum period.
21,22
Similar racial/ethnic disparities in sleep have also been reported for
children, such that racial/ethnic minority children compared to their White counterparts are
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted March 26, 2021. ; https://doi.org/10.1101/2021.03.22.21254093doi: medRxiv preprint
3
reporting to have the highest rates of insufficient sleep.
23
A few studies have focused on the infant
period and results suggest emergence of sleep disparities early on in life.
24,25
While these studies
suggest that racial/ethnic disparities in sleep exist across the life span, the potential pathways for
the relationship between race/ethnicity and poor sleep have not been fully understood. One
hypothesized pathway is that of psychosocial stress, which has been shown to be associated with
poor sleep
26,27
and has been reported to disproportionately affect racial/ethnic minorities.
28–30
In
addition, most recently, racial and ethnic minorities have been disproportionately affected by the
COVID-19 pandemic,
31–33
which itself represents a major psychosocial stressor.
34
To our knowledge, no published study exists investigating racial/ethnic disparities in infant and
maternal sleep during the COVID-19 pandemic. Here, we investigate racial/ethnic disparities in
sleep both in infants and mothers at 4 months post-partum, leveraging data collected as part of
the COVID-19 Mother-Baby Outcomes (COMBO) initiative at Columbia University Irving Medical
Center. The sample of this cohort is racially/ethnically diverse and since many participants gave
birth at the height of the pandemic, it represents a unique opportunity to investigate relationships
among race/ethnicity, psychosocial risk factors during the COVID-19 pandemic, and mother-
infant sleep.
METHODS
Participants and study design
Participants were recruited as part of the COMBO cohort at Columbia University Irving Medical
Center (www.ps.columbia.edu/COMBO), which aims to comprehensively describe the health of
mother-infant dyads during the COVID-19 pandemic. All procedures were approved by the
Columbia University Institutional Review Board. Eligible mothers delivered newborns at one of
two Columbia University-affiliated New York-Presbyterian Hospitals, the Morgan Stanley
Children’s Hospital or the Allen Hospital in New York City (NYC). On March 22, 2020,
universal SARS-CoV-2 PCR testing via nasal swab was initiated for all women admitted to labor
and delivery units. Starting in July 2020, all laboring mothers were also tested for COVID-19
antibodies. All mothers who gave birth at Morgan Stanley Children’s Hospital or Allen Hospital
since March 22, 2020 and who had a confirmed SARS-CoV-2 PCR positive test result during
pregnancy or a SARS-CoV-2 positive antibody test result with a confirmed or suspected
infection onset during pregnancy were approached for inclusion in the COVID-19 positive group
of COMBO. For each enrolled COVID-19 positive mother-baby dyad, a minimum of one case-
matched dyad was approached for enrollment into the control group. Control mothers must have
had negative SARS-CoV-2 PCR and antibody testing at admission to labor and delivery, no
history of COVID-like symptoms, and no history of a positive COVID-19 test at any point during
pregnancy. Control dyads were matched based on infant sex, gestational age in two-week
windows, mode of delivery, and date of birth within approximately 1 week. A subset of
participants was initially approached to enroll in the prenatal arm of the COMBO study. Prior to
delivery, they were asked to complete the COVID-19 Perinatal Experiences (COPE) survey to
collect information on maternal mental and physical health during pregnancy. These participants
were subsequently enrolled in the postnatal arm of the COMBO study after giving birth and were
enrolled regardless of COVID-19 status and without case-matching. An additional group of
mothers who gave birth prior to the onset of the COVID-19 pandemic in NYC, during the month
of February 2020, were also approached for enrollment as a control group for pandemic-related
stress during pregnancy. All surveys were administered through the Columbia University Irving
Medical Center REDCap system (version 10.6.2) and participants were offered the option to
complete them in English or Spanish. After delivery, mothers were invited to complete the online
surveys at 1, 2, 4, 6, and 9 months of age.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted March 26, 2021. ; https://doi.org/10.1101/2021.03.22.21254093doi: medRxiv preprint
4
Race/ethnicity:
The main exposure variable was maternal racial/ethnic background. As part of hospital intake
records and through the online surveys, mothers self-reported their ethnic background as ‘non-
Hispanic’, ‘Hispanic’, or ‘decline to answer’ and were also asked to specify their racial background
choosing from categories recommended by the National Institutes of Health (NIH) as ‘White’,
‘Black or African American’, ‘Asian’, ‘American Indian or Alaskan Native’, ‘Native Hawaiian or
other Pacific Islander’, and/or ‘Other’. We combined these 2 variables to obtain one single
race/ethnicity variable as follows: non-Hispanic White (W), African American (AA), Hispanic (H),
and other (O).
Sleep and mental health measures:
The primary outcomes were mothers' report of their infant's sleep health as measured by the Brief
Infant Sleep Questionnaire (BISQ),
35
and of their own sleep as measured by the Pittsburgh Sleep
Quality Index (PSQI) questionnaire (Cronbach's alpha 0.83).
36
The entire cohort completed the BISQ infant sleep questionnaire. From the BISQ we extracted
the following measures: 1) nighttime sleep duration, 2) daytime sleep duration, 3) total sleep
duration, 4) number of night awakenings, 5) amount of wakefulness during the night, 6) sleep
latency, 7) bedtime, and 8) infant’s sleep perceived as a problem by the mother.
One-hundred and forty-nine women out of the total sample of 238 women also completed the
PSQI. From the PSQI we extracted the following measures: 1) subjective sleep quality, 2) sleep
latency, 3) sleep duration at night, 4) sleep efficiency, 5) sleep disturbances, 6) use of sleep
medication, 7) daytime dysfunction, and 8) overall sleep health. This same subset of mothers also
completed the Patient Health Questionnaire-9 (PHQ-9, Cronbach's alpha 0.851)
37
to assess
severity of depressive symptoms, the Perceived Stress Scale (PSS, Cronbach's alpha 0.82)
38
to
assess levels of perceived stress, and the post-traumatic stress PTSD Checklist for DSM-5 (PCL-
5) adapted for COVID-19 (see Supplementary Material for this adapted scale).
Other measures:
We collected information on infant sex, gestational age at birth, maternal age at delivery, and
health insurance status (commercial versus Medicaid) from the electronic medical records. In
addition, we defined maternal COVID-19 status, as 1) control pre-pandemic, 2) negative, 3)
positive. Positive indicated that the participant had tested positive by PCR or serology at any point
during pregnancy. Negative indicated that participant had not had a known COVID-19 infection at
any point in pregnancy and had tested negative for SARS-CoV-2 by PCR and/or antibody testing
at the time of delivery. The subset of mothers who completed the PSQI, PHQ-9, PSS, and PCL
do not include any pre-pandemic control participants.
Statistical analysis:
We tested for normal distribution of continuous variables obtained from the BISQ. In case of not
normally distributed variables (amount of wakefulness during the night, sleep latency, bedtime)
we discretized them and transformed from continuous to ordinal. Then, we used multiple
linear/ordinal/logistic regression models to assess the independent associations of maternal
race/ethnicity on maternal and infant sleep variables with non-Hispanic White group as reference.
Regression models were adjusted for infant sex, gestational age at birth, and age in weeks at
time of assessment in Model 1. No adjustments were made for regression models predicting
maternal sleep in Model 1. Given that maternal socio-economic status (SES) has the potential to
be a confounder of the relationship between race/ethnicity and maternal and infant sleep, Model
2 also adjusted for health insurance status as an indicator of SES, with commercial insurance as
reference group. Two additional models tested the potential effects of SARS-CoV-2 prenatal
exposure versus pandemic-related stress, respectively. In Model 3, we adjusted for COVID-19
status, with control pre-pandemic mothers as reference group. In Model 4, we adjusted for
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted March 26, 2021. ; https://doi.org/10.1101/2021.03.22.21254093doi: medRxiv preprint
5
delivery timing with respect to the pandemic in New York City, which had a sharp peak in March-
April 2020, to understand the different impact of various phases on the pandemic on maternal
and infant sleep. Time periods of the pandemic at delivery were grouped in 2 months bin based
on infant date of birth: 1) pre pandemic controls, 2) March-April 2020, 3) May-June 2020, 4) July-
August 2020, and 5) September-October 2020.
After Bonferroni adjustments for multiple comparison analyses, the adjusted p-values for analyses
on infant and maternal sleep measures was p=0.006.
We ran successive multiple linear/ordinal/logistic regression models to assess the independent
associations of PHQ-9, PSS, and PCL-5-COVID scores with maternal and infant sleep variables.
We performed mediation analysis via the CMAverse R package
39
to assess whether PHQ-9, PSS
and PCL scores were mediators in the relationship between race/ethnicity and maternal and infant
sleep variables. The mediation was tested only for non-Hispanic White and Hispanic mothers.
RESULTS
Infants included in this analysis were born between 2/1/20 and 9/30/20. The racial/ethnic
composition of the entire sample was 71 non-Hispanic White, 14 African American, 113 Hispanic
and 40 other. Additional demographics are in Table 1. The racial/ethnic composition of the subset
of mothers who also completed the PSQI and the mental health questionnaires was 44 White
non-Hispanic, 8 African American, 70 Hispanic and 27 other. The subset was not significantly
different from the entire subset for any socio-demographic variable, except for infant’s age at the
assessment, with infants from the subset being younger (p<0.001) and COVID-19 status, since
in the subset there were no pre-pandemic control mothers. Additional socio-demographics for the
subset are presented in Table S1.
Infants sleep by race/ethnicity
Table 2 displays infant sleep variables across racial/ethnic groups.
The results of the multivariable regression models for Model 1, with maternal race/ethnicity
predicting infant sleep variables and infant’s sex and gestational age at birth as covariates,
indicated that compared to non-Hispanic White infants, Hispanic infants and infants from other
race/ethnicities slept less at night, and African American infants showed the same directional
trend (H: β=-101.7±17.6, p<0.0001, O: β=-64.7±23.0, p=0.005, AA: β=-63.6±33.2, p=0.06).
Table 1 Demographic for the overall cohort and by race
Entire cohort
Overall
W
(N=71)
AA
(N=14)
H
(N=113)
O
(N=40)
p-val
Gestational age (weeks)
38.68 (1.7)
38.34 (1.9)
38.43 (2.3)
38.95 (1.6)
38.83 (1.3)
p=.020
Age at assessment
(weeks)
18.99 (2.40)
19.15 (2.42)
19.04
(2.71)
18.79 (2.34)
19.26
(2.48)
p=0.68
9
Sex (F/M)
119/119
31/40
6/8
58/55
24/16
p=.376
Covid status p=.035
Pre-pandemic Control
66
21
5
28
12
Negative
92
34
6
34
18
Positive
80
16
3
51
10
Maternal age (years)
31.65 (5.4)
33.0 (4.6)
33.0 (6.6)
30.1 (5.4)
33.1 (5.4)
p=.001
Medical Insurance p>.001
Commercial
130
64
8
30
28
Medicaid
108
7
6
83
12
Mode of delivery p=.375
Vaginal
144
45
7
64
28
Ceserean
94
26
7
49
12
Parity
1.8 (1.0)
1.7 (1.2)
2.1 (1.0)
1.95 (0.9)
1.5 (0.7)
p=.089
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)
The copyright holder for this preprint this version posted March 26, 2021. ; https://doi.org/10.1101/2021.03.22.21254093doi: medRxiv preprint