Mother-Infant Room-Sharing and Sleep Outcomes in the INSIGHT Study

As previously described,^26,–28 mothers and their newborns were recruited after delivery from one maternity ward (Penn State Milton S. Hershey Medical Center, Hershey, PA) for participation in the INSIGHT study from January 2012 to March 2014. The study design and inclusion and exclusion criteria have been previously published.²⁶ Major eligibility criteria specified that the subjects be term singletons weighing ≥2500 g delivered to English-speaking, primiparous mothers ≥20 years old. Dyads were randomly assigned 2 weeks after delivery to a responsive parenting (RP) intervention or a safety control group. 279 mother-infant dyads completed the first home visit and are considered the study cohort for outcomes as specified before study initiation in the protocol. For the current analyses, those parents who reported bed-sharing with their newborn were excluded because the goal was to compare those who were sleeping independently in a separate room from parents with those still room-sharing but not bed-sharing (as recommended by the AAP).

Research nurses conducted home visits at infant ages of 3 to 4 weeks and 4, 6, and 9 months. Research center visits occurred at 1 and 2 years of age. Intervention components have been previously described.²⁷ Among other sleep-related topics discussed at the 3- to 4-week home visit, the RP group received guidance encouraging parents to consider moving their child’s sleep location at the age of 3 months to the room they would prefer the infant to sleep in at 1 year of age, with the advice that such a move would be more difficult if the family waited much beyond that point.²⁵ This point was reinforced at the 4-month home visit. The safety control group received education regarding prevention of SIDS, which included direct observation of the sleep environment by the visiting nurses with individual feedback to improve safety at the first 3 home visits. This study was approved by the Penn State College of Medicine’s Human Subjects Protection Office and registered at www.clinicaltrials.gov.

Data were collected and managed by using REDCap (Vanderbilt University, Nashville, TN).²⁹ Mothers completed the online surveys with paper surveys mailed to those lacking internet connectivity (n = 20). A detailed list of measures was published.²⁶ 

Demographic information collected at enrollment included parent and child race/ethnicity, maternal education, annual income, and marital status. Data extracted from medical charts included maternal age, prepregnancy weight, gestational weight gain, infant gestational age, and birth anthropometrics.

The validated Brief Infant Sleep Questionnaire was used to assess sleep at 4 and 9 months of age with an abbreviated version administered at 12 and 30 months.³⁰ This survey assesses infant sleep location, before-bedtime activities, and sleep patterns. Sleep duration is divided into nighttime (7 pm–7 am) and daytime (7 am–7 pm) and is reported in hours and minutes. Additional questions regarding sleep location, night waking, night feedings, bedtime routine components and duration, infant sleep behaviors and environment, and parenting responses to night wakings were asked at 4 and 9 months. Instructions for the infant sleep survey guided mothers to think about her child’s sleep during the past week when answering the questions and base responses on a “typical night.” Maternal sleep quality was measured at 4 months of infant age by using the PROMIS Sleep Disturbance instrument (PROMIS Network, American Institutes for Research, Chapel Hill, NC).³¹ 

Feeding mode was assessed by using a food frequency questionnaire. Infants were categorized as predominantly breastfed if ≥80% of milk feedings were breastmilk.³² 

Data were analyzed by using SAS version 9.4 (SAS Institute, Cary, NC). Descriptive statistics were generated for sleep duration and behaviors on the basis of the timing of our 2 sleep assessments during infancy at 4 and 9 months of age, which created 3 groups: early independent sleepers (own room <4 months), later independent sleepers (own room between 4 and 9 months), and room-sharers at 9 months. For continuous outcomes, differences by sleep location were assessed by using analysis of variance (ANOVA) at each of the 4 times, adjusting for study group assignment. Categorical outcomes were assessed by using logistic regression, adjusting for study group. Analyses including socioeconomic covariates (eg, income, living in a single-family home) did not result in major differences in results, and thus results from models that included only study group are presented. Statistical significance was defined as P ≤ .05.

Among 279 participants,^27,28 259 (93%) remained in the study through 9 months of age. Among this group, 20 were bed-sharing at 4 and/or 9 months and 9 had missing data on sleep. These participants were excluded from further analyses, leaving a sample of 230 mother-infant pairs as the sample for this manuscript. The majority of infants, 142 (62%), were “early independent sleepers,” sleeping independently without room-sharing by 4 months (Table 1). Another 62 (27%) were “later independent sleepers” who began sleeping independently between 4 and 9 months of age, whereas 26 (11%) were room-sharing at 9 months. There were no differences between intervention groups in terms of frequency of bed-sharing or room-sharing at 4 or 9 months. Room-sharing was associated with demographic variables including race/ethnicity, lower income and education, unmarried and/or not cohabiting with partner, fewer bedrooms in the home, and extended family and/or other people living in the home.

At 4 months, no significant differences in nighttime or total sleep duration were observed between those who were early independent sleepers and those who room-shared at 4 and/or 9 months (Fig 1). However, the early independent sleepers had better sleep consolidation at 4 months of age with a longest (mean ± SD) sleep stretch averaging 469 ± 139 minutes compared with 423 ± 158 minutes for those who were room-sharing (Table 2) (P = .02). While the mean number of night wakings was similar between early independent and nonindependent sleeping infants at 4 months of age, those sleeping independently had fewer night feedings (1.1 vs 1.4; P = .02). Nonindependent sleepers also had twice the odds of being fed back to sleep overnight compared with independent sleepers (adjusted odds ratio [aOR] 1.95, 95% confidence interval [CI] 1.09–3.50).

By 9 months of age, significant differences in nighttime sleep duration emerged between groups (Table 3). The early independent sleepers slept on average 627 ± 67 minutes per night compared with 601 ± 73 minutes for later independent sleepers and 587 ± 83 minutes for those who were room-sharing at 9 months of age (P = .008). Those who were early independent sleepers had the best sleep consolidation with the longest sleep stretches, averaging 542 ± 162 minutes versus 497 ± 166 minutes for later independent sleepers and 442 ± 158 minutes for room-sharers (P = .01). Although night wakings, night feedings, and odds of being fed back to sleep overnight did not differ among groups at 9 months of age, total daily sleep duration was longest among early independent sleepers (796 ± 86 vs 760 ± 94 vs 773 ± 96 minutes, P = .04).

At 12 months of age, no differences existed between groups for nighttime or total daily sleep duration. However, the 30-month assessment revealed that both early and later independent sleepers slept on average more than 45 minutes longer at night than those who were room-sharing at 9 months (614 ± 51 vs 617 ± 70 vs 569 ± 79; P = .004). Total daily sleep duration was similar among groups at 30 months of age.

Significant differences existed for bedtime routines at 4 months of age. Early independent sleepers had increased odds of having a consistent bedtime routine (aOR 1.93, 95% CI 1.05–3.53) and being put to bed by 8 pm (aOR 1.93, 95% CI 1.06–3.53) relative to room-sharers. At 9 months of age, room-sharing was associated with greater odds of having a bedtime after 8 pm; room-sharers also had lower odds than independent sleepers of being put to sleep drowsy but awake (P ≤ .05 for both).

At 4 months of age, room-sharing infants had greater odds of having an unapproved object (blankets, pillows, positioners, etc) on their sleep surface than those who were sleeping independently (aOR 2.04, 95% CI 1.17–3.57). Additionally, room-sharing parents with infants at 4 months of age had over 4 times the odds of bringing their infant into their bed overnight than those parenting early independent sleepers (aOR 4.24, 95% CI 1.64–10.95). Similarly, at 9 months of age, room-sharing parents had 4 times the odds than those who were sleeping independently of bringing their infant into their bed after waking overnight (aOR 3.97, 95% CI 1.59–9.90).

These secondary analyses from the INSIGHT study demonstrate that infant-parent room-sharing at 4 and 9 months of age is associated with less reported nighttime sleep in both the short- and long-term, worse sleep consolidation, and unsafe sleep practices previously associated with sleep-related death. The findings raise questions about the well-intended AAP recommendation that room-sharing should ideally occur for all infants until their first birthday.²¹ 

The recommendation to room-share until the age of 1 year is not evidence-based for the prevention of SIDS, given that 90% of these tragedies occur before the age of 6 months.²⁰ The data presented by the AAP Task Force on Sudden Infant Death Syndrome to support the recommendation²¹ came from 3 European case-control reports.^33,–35 In the first by Blair et al,³³ infants who died of SIDS at over 4 months of age were more likely than controls to be room-sharing without bed-sharing (46% vs 41%) and less likely to be sleeping independently in their own rooms (43% vs 53%; statistical comparison not performed), yet the age stratification presented in this paper was not acknowledged in a recently published letter by the Task Force.³⁶ The second report involved mostly infants <6 months old and used data from the early- to mid-1990s, when prone sleeping was still common.³⁴ The third study involved only 123 cases of SIDS with a mean age of 13.6 ± 9.5 weeks, indicating that most cases included infants <6 months old.³⁵ This small study included only 15 infants who slept in their own room and died of SIDS. Therefore, it is unlikely that any conclusions can be made about those who were 6 to 12 months old on the basis of this small sample size and the epidemiology of the condition. Importantly, the biggest risk factor for SIDS in this third report was “sharing the same sleep surface as the last sleep location,” something that more commonly occurred among infants who initially shared their mother’s room in our study.

In addition to the lack of data supporting the recommendation to room-share until 1 year of age, the Policy Statement authored by the Task Force did not describe a mechanism by which room-sharing would prevent SIDS or how parents should apply this recommendation to achieve better outcomes than those parents choosing to have their infants sleep independently. Although the Task Force’s recent letter cited more frequent arousals among room-sharers as the potential protective mechanism,³⁶ the 2 studies referenced do not support that conclusion. The first is a case-control study of 18 infants.³⁷ Among 9 room-sharers, 5 were bed-sharers, making conclusions about arousals among those who are room-sharing but not bed-sharing (versus independent sleepers) problematic. The second study, which also included bed-sharers in the room-sharing group, states that parental presence in the room at bedtime (eg, holding, rocking, feeding to sleep) explained much of the variance in night wakings and that night-long room-sharing had a “negligible (<1%) independent contribution to the explained variance.”²² 

The AAP recommendation to room-share until the age of 1 year conflicts with sleep expert guidance,²⁵ which recognizes developmental changes that occur over the first year. The suggestion that parents wait to move the infant out of their bedroom until the end of the first year, when separation anxiety is normative and increasing, is likely to result in frustrated parents and unhappy infants. It also conflicts with other data that room-sharing is associated with more sleep disruption for mothers.^24,38 

Our data showing shorter night sleep and shorter sleep bouts for infants who are room-sharing has potentially important consequences. Inadequate sleep affects nearly every health domain, including emotional regulation, mood, and behavior in infancy and childhood.^3,–6 As demonstrated in our analyses, room-sharing and short sleep duration during infancy are also associated with inadequate sleep later in childhood,¹⁵ which emphasizes the importance of developing healthy sleep-related behaviors during the first months of life. Infant sleep also affects parental sleep, which can affect other outcomes, including maternal sensitivity, relationship quality, parental emotional health, and parenting practices.^16,17 Fatigue is related to somatic complaints, anxiety, depression, paranoia, and crashes or unintentional injuries involving motor vehicles among healthy adults^39,40 and has been associated with postpartum depression,⁴¹ abusive head trauma, and child abuse.⁴² At a time when family-centered guidance is strongly encouraged, parent fatigue should be considered in any guideline related to infant sleep.

Perhaps our most troubling finding was that room-sharing was associated with overnight transitions to bed-sharing, which is strongly discouraged by the AAP.²¹ Bed-sharing overnight was more common in our sample among 4- and 9-month-olds who began the night on a separate surface in their parents’ room. This finding is consistent with those from another recently described cohort.⁴³ 

There are limitations to our analyses. The data come from a randomized controlled trial with 2 interventions. The control group received information aimed at reducing the risk of SIDS, although no recommendation was made regarding sleep location. The RP intervention included guidance to improve sleep duration and sleep-related behaviors, including recommending independent infant sleep by 4 months of age. However, sleep location did not vary significantly by intervention group (data not shown), and all analyses were adjusted for study group. Next, the findings we report are correlational, and causal effects cannot be inferred. It is possible that parents of infants with poorer sleep prefer that infants sleep more closely to them. The relatively homogeneous INSIGHT sample is another potential limitation, as are the self-reported data, although the Brief Infant Sleep Questionnaire has been validated against objective measures of sleep.³⁰ Some previous studies have shown discrepancies between subjective and objective measures of sleep,^24,38 and the lack of an objective measure may have resulted in unreported night wakings in which the infant self-soothed back to sleep. This possibility would particularly influence the sleep consolidation data, especially for independent sleepers. However, this possibility, if true, would only reinforce the concept that room-sharing may result in either unnecessary parental responses to infant night wakings or, alternatively, the infant’s expectation of caretaking behaviors from parents, both of which can be expected to lead to decreased sleep for infant and parent. In addition, parental perceptions of their infant’s sleep are the basis for whether they report sleep problems and request guidance from health care providers.

Our data showing that room-sharing families had lower education, income, and less likelihood of living in a single-family residence emphasize that not all parents have a choice regarding their infant’s sleep location and may not have the option to have the infant sleep independently, a factor we could not fully address although demographics and home characteristics were assessed in data analyses. Cultural norms may also lead to preferences for room-sharing in some patient populations, as described by Mileva-Seitz et al.⁴⁴ Lastly, we did not assess room-sharing beyond 9 months of age.

While substantial progress has been made over the past several decades to improve the safety of infant sleep, the AAP recommendation that parents room-share with their infants until the age of 1 year is not supported by data, is inconsistent with the epidemiology of SIDS, is incongruent with our understanding of socioemotional development in the second half of the first year, and has the potential for unintended consequences for infants and families. Our findings showing poorer sleep-related outcomes and more unsafe sleep practices among dyads who room-share beyond early infancy suggest that the AAP should reconsider and revise the recommendation pending evidence to support room-sharing through the age of 1 year.

 
• AAP

  American Academy of Pediatrics

 
• ANOVA

  analysis of variance

 
• aOR

  adjusted odds ratio

 
• CI

  Confidence Interval

 
• INSIGHT

  Intervention Nurses Start Infants Growing on Healthy Trajectories

 
• RP

  responsive parenting

 
• SIDS

  sudden infant death syndrome