About 12.9 million preterm infants are born worldwide each year, with about 11 million (85%) in Africa and Asia (Beck et al., 2010). Preterm infants' survival requires staying in neonatal intensive care units (NICU) to allow continued development of the central nervous system and other body systems. However, infants in the NICU commonly undergo invasive procedures, creating stimuli and stress at odds with the requirements of their neurosensory development (Als, 1999). These multiple stressors can contribute to long lasting effects on their development. Indeed, preterm infants have higher rates of neurodevelopmental impairment, such as cerebral palsy, sensory deficits, and learning disabilities than children born at term (Beck et al., 2010; Sun, Mohay, & O'Callaghan, 2009).
In this prospective, descriptive study, we used a repeated-measures design to explore the 24-hour effects of caregiving and positioning on preterm infants' states and the factors associated with state changes. Thirty preterm infants (gestational age 27.6–36.1 weeks) were observed for 3 days in the neonatal intensive care unit to record six states: quiet sleep (QS), active sleep, transition, active awake, quiet awake, and fussy or crying. The occurrences of QS increased when infants received no caregiving, social interaction, non-nutritive sucking (NNS), and were laterally positioned. However, QS significantly decreased and fussy or crying state increased when infants received routine and intrusive caregiving. These results suggest that caregiving, NNS, and positioning should be appropriately provided to facilitate infants' sleep, and reduce fussiness or crying. © 2011 Wiley Periodicals, Inc. Res Nurs Health 35:132–145, 2012
Sleep is Important for Infants' Neurosensory Development
Preterm infants' neurological maturation and the intactness of their central nervous system have been closely related to the organization of sleep/wake states (Arditi-Babchuk, Feldman, & Eidelman, 2009; O'Callaghan et al., 2010). Sleep and sleep cycles were shown to be essential not only for sensory and motor system development, but also for preserving brain plasticity, creating long-term memory and learning, and cellular repair (Graven & Browne, 2008). Disrupting sleep and sleep cycles profoundly affects the creation of permanent neural circuits, interfering with the early process of sensory and brain development (Graven & Browne, 2008), which has sometimes led to developing attention problems in later life and adverse neurodevelopmental outcomes (O'Callaghan et al., 2010). Progressive lengthening of sleep periods was related to better mental development in later life, with long periods of uninterrupted sleep predicting better mental development at 12 months (Anders, Keener, & Kraemer, 1985). Therefore, sleep is an important determinant of infants' developmental outcomes (Arditi-Babchuk et al., 2009; A. Scher, 2005). Whether infants in the NICU have enough sleep for growth and neurosensory development is an important issue for clinicians to explore. Enhancing preterm infants' neurosensory development requires understanding the social-contextual and environmental factors that influence their sleep and using strategies to preserve infants' sleep in the NICU (Arditi-Babchuk et al., 2009; Lopez et al., 2008).
Sleep is a time of rest with decreased awareness, body movement, and metabolic activity (Anders, Sadeh, & Appareddy, 1995). A majority of preterm infants between 25 and 30 weeks postconceptional age demonstrate sleep/wake state cyclicity (M. S. Scher, Johnson, & Holditch-Davis, 2005). Infant sleep/wake states, defined by eye opening, body and limb movement, muscle tone, eye movement, and respiration, are categorized into quiet and active sleep (AS), quiet and active awake (AW), transition, and crying states (Holditch-Davis, Scher, Schwartz, & Hudson-Barr, 2004). A complete sleep cycle includes wakefulness, AS, and quiet sleep (QS; Graven & Browne, 2008; M. S. Scher et al., 2005). QS has been associated with energy maintenance, increased protein synthesis, and release of growth hormone, and is crucial to creating long-term memories and learning (Euston, Tatsuno, & McNaughton, 2007). AS is essential to the process of endogenous stimulation that is critical for neurosensory development (Euston et al., 2007).
The mean sleep cycle duration has been observed to be about 40 minutes at 27–30 weeks gestational age (GA), 45 minutes at 31–34 weeks GA, and 50–70 minutes at 35–41 weeks GA (Bertelle, Sevestre, Laou-Hap, Nagahapitiye, & Sizun, 2007). Increasing GA also was found to be correlated with increased QS and decreased AS (Holditch-Davis et al., 2004; Ingersoll & Thoman, 1999).
Factors That Influence Infant Sleep
Infants' sleep/wake patterns have been conceptualized in the transactional model (Thomas, Tsai, & Brown, 2008) as resulting from interactions between infants and their social-contextual environment. Indeed, sleep organization is shaped by early interactions between preterm infants and the caregiving environment (Sameroff, 2009; Thomas et al., 2008). Development of infant sleep was also influenced by infant maturation, infant traits, and nursing caregiving (Brandon, Holditch-Davis, & Beylea, 1999; Ingersoll & Thoman, 1999; Thomas et al., 2008). However, infant states in these studies were recorded for only 3–4 hours (Brandon et al., 1999; Thomas et al., 2008) via video camera (Ingersoll & Thoman, 1999; Thomas et al., 2008). Furthermore, caregiving was not clearly defined, nor were relevant confounding factors analyzed. Such factors include positioning and use of non-nutritive sucking (NNS), which might have simultaneously influenced infants' states. Thus, these studies offer an incomplete analysis of how caregiving influences infants' states.
Nurse caregiving has been categorized into contact, routine, and procedural care, with intrusiveness of care correlated to infant sleep–wake development (Brandon et al., 1999). Common caregiving events in the NICU have been rated by physicians and nurses as stressful to preterm infants at different levels (Newnham, Inder, & Milgrom, 2009).
Infants in the prone position sometimes sleep more efficiently (Ariagno et al., 2003; Bhat et al., 2006; Sahni, Schulze, Kashyap, Ohira-Kist, & Myers, 2005). Furthermore, preterm infants in the NICU commonly receive NNS, an intervention considered benign, to modulate infant behavioral states (Bingham, Ashikaga, & Abbasi, 2011; Pinelli & Symington, 2010). However, little is known about the effect of NICU caregiving, position, or NNS on preterm infants' states over time, nor about the links among preterm infant sleep quality, caregiving, and positioning.
To fill these gaps in knowledge, we conducted this study to (a) explore the effects of 24-hour caregiving, positioning, and NNS use in the NICU on preterm infants' sleep/wake states and (b) identify the factors associated with infants' state changes. Specifically, we asked three questions:
- 1.What are the effects of different patterns of NICU caregiving on the occurrence rates of QS, AS, wake, transition, and fussy or crying states in preterm infants after adjusting for the effects of time, infants' GA, postmenstrual age (PMA), illness severity, positioning, and NNS use?
- 2.What are the effects of position on the occurrence rates of the preterm infants' sleep–wake states after adjusting for the effects of time and other factors?
- 3.What are the effects of NNS use on the occurrence rates of preterm infants' sleep–wake states after adjusting for the effects of time and other factors?
A prospective, descriptive repeated-measures design was used. Sleep/wake states were continuously observed and recorded for three consecutive 24-hour periods. At the same time, caregiving patterns, position, and NNS use were measured and recorded.
Setting and Sample
Infants were recruited by convenience sampling from a level III NICU at a medical center in northern Taiwan from October 2008 to October 2010. This study was approved by the study site's institutional review board. The principal investigator (PI) selected infants and collected their background data after reviewing medical and nursing charts. Infants were included if they met these criteria: (a) GA 27–37 weeks and PMA 27–37.5 weeks, (b) postbirth age 3–28 days, and (c) disease condition acceptable for observation (illness severity indicated by the Neonatal Therapeutic Intervention Scoring System [NTISS] score ≤ 21). In this study, GA was defined as the time from the first day of the last menstrual period to birth (in weeks). PMA was defined as GA plus the time since birth (in weeks). Infants were excluded by these criteria: (a) congenital anomalies, (b) neurologic impairment including convulsion, intraventricular hemorrhage >grade II, or periventricular leukomalacia, (c) documented congenital or nosocomial sepsis (infection acquired at hospital after birth), (d) surgery, (e) severe growth restriction at birth, (f) substance-abusing mother, (g) severe medical conditions requiring treatments such as sedatives, muscle relaxants, antiepileptic, aminophylline, or analgesic drugs. Nurse caregivers were included by these criteria: (a) employed in the unit for ≥18 months, (b) familiar with preterm infant care based on standardized NICU procedures, and (c) agreed to be observed while caring for preterm infants.
Of 110 infants screened from 2008 to 2010, 54 met our criteria (Fig. 1). Participation was refused by 16 parents who did not want their infants observed in the NICU due to their small size and unstable condition. Another eight infants did not complete the 3-day observation period due to change in illness severity (NTISS > 21) or use of medications such as phenobarbital and aminophylline. Thus, 30 infants were enrolled and completed three consecutive 24-hour periods of observation. Infants who did and did not participate did not differ significantly in sex, GA, PMA, days since birth, or birth weight.
Infants' illness severity was measured by the PI using the NTISS, which examines the therapeutic intensity and complexity of NICU therapies (Gray, Richardson, McCormick, Workman-Daniels, & Goldmann, 1992). The 63-item NTISS includes eight clinical subscales, which are used to compute the total score, ranging from 0 to 47. The internal consistency of the NTISS was .84 (α-coefficient), and its convergent validity was examined by correlations between its scores and mortality risk estimates (r = .70, p < .001; Gray et al., 1992). In this study, the NTISS interrater reliability between the PI and neonatologist was .98.
Outcome variables were occurrence of preterm infants' sleep/wake states: QS, AS, transition state (T), AW, quiet awake (QW), and fussing or crying (F). Mutually exclusive sleep/wake states were measured at 1-minute intervals by time-triggered observing and recording. In this way, the research nurses coded the infant states that occurred in every 1-minute interval during 24 hours of NICU caregiving.
The occurrence of infant states was observed and recorded using a coding scheme modified from previous studies (Brandon & Holditch-Davis, 2005; Holditch-Davis et al., 2004; Ingersoll & Thoman, 1999). Briefly, infants' states were coded based on eyes closed or open, regulation of respiration, muscle tone, and motor activity. QS was defined as eyes closed, respiration regular and abdominal, motor tone maintained at a tonic level, and limited motor movement (such as occasional subtle startles or sighs). AS was defined as eyes closed, respiration irregular and primarily costal, and sporadic motor movement with low muscle tone, or intermittent rapid eye movements. T was defined as eyes rapidly opening and closing, irregular breathing, some body movements, or fussy vocalization. QW was defined as eyes open, focused, respiration regular, very few, or no body movements. AW was defined as eyes open, respiration irregular, motor activity varies. F was defined as upset, fussing, highly aroused, and crying. When an infant's eyes opened and closed during an observation interval (1 minute), T was recorded. When an infant fussed or cried at any time during the 1-minute observation, F was recorded. The study rules for coding states were based on levels of infant wakefulness (F > AW > QW > T > AS > QS). For multiple occurrences during the same interval, only the most aroused state was recorded. A less aroused state was recorded only when it lasted the entire interval. If a more wakeful state than the previous state was observed during the interval, the more aroused state was coded. For example, QS had to last the whole interval to be coded QS. If the infant first showed AW, then fussed or cried at the end of the same interval, we coded F.
NICU caregiving, positioning, and NNS
The caregiving, positioning, and NNS offered to preterm infants were measured by a caregiving coding scheme, which measured four patterns of caregiving (no caregiving, social interaction, routine caregiving, and intrusive caregiving), four positions (supine, lateral, prone, and other), and NNS. We developed a list of 49 caregiving activities offered to preterm infants in the NICU based on chart review, direct observation, our work experiences, and relevant research (Brandon et al., 1999; Newnham et al., 2009). These 49 caregiving activities were categorized into social interactions, routine caregiving, or intrusive caregiving by 10 nurses with more than 2 years work experience in the NICUs from four medical centers. To categorize activities, nurses completed a brief questionnaire with the 49 activities and three caregiving categories. For each activity, they indicated the degree (as a percentage) to which they perceived the activity belonged to social interaction, routine caregiving, or intrusive caregiving. Thus, patting to burp might have been categorized as 7.5, 88.0, and 4.5. The nurses' consensus (mean percentage) on categorizing each caregiving activity is shown in Table 1. Their mean percentage agreement with the category of each caregiving activity ranged from 85.3% to 100%. Each caregiving category was operationally defined as follows: (a) no caregiving was no interaction with the preterm infant or no caregiving stimuli offered during the interval observed, (b) social interaction was gentle, supportive, comfort care activities, and interactions such as talking, gentle touch, or other comforting behavior, (c) routine caregiving was usual nursing or medical activities perceived as a little stressful but not intrusive or painful to the infant, and (d) intrusive caregiving was a nursing or medical activity perceived as moderately to severely stressful or intrusive and painful to the infant.
|Item||Caregiving activity||Social interaction||Routine caregiving||Intrusive caregiving|
|1||Pacifier use (non-nutritive sucking)||94.3||4.6||1.1|
|2||Gentle touch or talking||92.3||7.7||0|
|7||Adjust position; use towel rolls to support||87.0||13.0||0|
|8||Measure vital signs||0||94.6||5.4|
|9||Prepare baby for phototherapy||0||92.9||7.1|
|10||Diaper or clothes change||0||100||0|
|11||Bath or bed bath care||4.3||90.3||5.4|
|13||Remove oral gastric tube||0||88.0||12|
|14||Umbilical cord care||0||86.7||13.3|
|15||Change IV bottle or check IV tube or site||2.7||88.7||8.6|
|18||Brain, heart, or kidney sonogram||0||88.7||11.3|
|20||Measure abdomen girth||2.4||94.0||3.6|
|21||Change EKG leads||0.3||86.7||13.0|
|22||Use cotton swab to clean mouth||0.3||86.0||13.7|
|23||Use adhesive tape||0||86.7||13.3|
|24||Attach monitor sensor||2.1||90.6||7.3|
|25||Auscultate with stethoscope||1.5||93.3||5.2|
|27||Remove syringe feeding tube||0||100||0|
|28||Patting to burp||7.5||88.0||4.5|
|29||Change infant position||5.2||93.4||1.4|
|30||Apply cream to body||0||92.7||7.3|
|31||Use urine bag and collect urine||0||91.6||8.4|
|32||Remove infant from incubator||1.2||95.1||3.7|
|33||Suction endotracheal tube, nose, mouth||0||0||100|
|34||Insert IV, IA, UAC/UVC||0||0||100|
|35||Remove IV, IA, UAC/UVC||0||3.3||96.7|
|36||Hepatitis B IM injection||0||2.7||97.3|
|37||Glycerine ball enema||0||5.6||94.4|
|38||Flip at fingers or toes||0||6.7||93.3|
|42||Remove tape, change dressing, fix IV site, or fix oral gastric tube||0||13.3||86.7|
|44||Central venous puncture||0||0||100|
|45||Venous puncture for blood sampling||0||1.7||98.3|
|46||Insert oral gastric tube||0||6.7||93.3|
|47||Insert nasal CPAP tube||0||1.3||98.7|
|48||Clean milk from nose using cotton swab||0||13.3||86.7|
|49||Insert pneumothorax chest tube||0||0||100|
Infants' positions were defined as follows: (a) supine: chest positioned toward the top of the incubator with trunk flatly laid on the bed, and arms and legs spontaneously placed; (b) prone: chest positioned toward the incubator bed, head turned laterally, and small towel rolls used to support the flatly laid trunk and maintain limb and joint flexion; (c) lateral: right or left side of the trunk positioned toward the incubator bed, with back and abdomen supported using towel rolls; and (d) other: being breast or bottle fed, bathed, or out of the bed.
NNS was defined as the infant sucking on or holding a pacifier in his/her mouth without receiving breast milk or formula for nutrition. These definitions were used to categorically code occurrences of the four caregiving patterns, four types of positions, and use of NNS.
Parents of infants who met the study criteria received a pamphlet introducing the study. Interested parents returned a response sheet to a research nurse, who explained the study in detail, including the study procedures, their right to withdraw from the study at any time, protection of their infant's privacy, and right to receive caregiving, and the benefits of participating in this study. If parents agreed to the infant's participation, they signed parental permission. Consent was obtained from the nurse who took care of the infant.
Data were then collected at 1-minute intervals by six trained research nurses who directly observed preterm infants' states, caregiving patterns, positioning, and NNS use. These nurses were trained for 4 weeks by the PI to observe and record occurrences of infants' sleep/wake states, caregiving patterns, positioning, and use of NNS by watching videotapes of preterm infants in the NICU. The PI had 2 years of training on observing preterm infant behaviors and states from Dr. Evelyn Thoman, an expert at observing preterm infant sleep. The PI also had >10 years experience taking care of preterm infants, had conducted nine studies on infants' behavior and states, and was competent in measuring infants' states.
Each research nurse achieved internal consistency in coding infant states, caregiving patterns, positioning, and NNS use by watching videotapes of preterm infants and coding until intrarater reliability >.90 (range = .93–.97). Each research nurse then observed and recorded infants' sleep/wake states, caregiving patterns, position, and NNS use directly at the bedside with the PI until the interrater reliability between each nurse and the PI was ≥.85.
After this point, the six research nurses, blinded to the study purpose and infants' clinical information, observed and recorded preterm infant sleep/wake states and caregiving once per minute when signaled through an ear phone from a small electronic timer. The 24-hour observation period was divided into six 4-hour sessions, with one research nurse recording data in each session. These observers did not take breaks or talk with care providers. To overcome tiredness, observers could sit down or move around the incubator and drink tea as needed to remain vigilant. They also chose the sessions during which they thought they could work best, and wore comfortable clothes and shoes. Interrater reliabilities between observers and the PI were checked monthly for 30 minutes. Reliabilities for coding the six sleep/wake states were .89–.95, for the four caregiving patterns were .93–.98, for infant positioning were .96–.98, and for NNS use was .99. The occurrences of infants' states, caregiving patterns, positioning, and NNS use were recorded on observation sheets developed for the study. The sheet consisted of all codes for infant states and caregiving (144 rows for each 24-hour record, 10 minutes per row).
All nurse observers followed the NICU's standard procedures for taking care of preterm infants. Study fidelity was established by monthly meetings between investigators and research nurses. The PI also met the nurses caring for infants every time a new infant was enrolled to ensure consistency in providing caregiving activities and to discuss any problems.
Data were analyzed using SPSS for Windows, version 18.0.0. Continuous variables were described by means and SDs, and categorical variables were described by frequencies. Occurrences rates for all directly observed variables were summed over each hour for three continuous 24-hour periods. Because occurrence rates were low for active and quiet wake states, they were summed and considered as occurrences of the “wake” state. The occurrence rates of infant states, caregiving patterns, positioning, and NNS use, as well as time effects (hour and day) were analyzed after adjusting for the effects of other factors (independent variables other than the factor itself) using the generalized estimating equation (GEE) method's Poisson regression models (Liang & Zeger, 1986). We also accounted for within-subjects dependence due to repeated measures by specifying the structure of working correlation matrix as AR(1) (that is, the repeated measurements have a first-order autoregressive relationship). Based on these models, factors that potentially influenced occurrence rates of each infant's sleep/wake states were analyzed after adjusting for the other factors and time effects. Statistical significance was defined as p < .05.
The study sample of 30 preterm infants had a mean GA of 31.53 ± 2.69. Infants were equally divided between males and females; 13.33% were born by normal spontaneous delivery and 86.67% by Caesarean delivery. The infants' mean PMA was 32.54 weeks (range = 28.6–36.7), mean age at observation was 7.57 days (range = 3–24), mean birth weight was 1624.8 g (range = 1,005–2,630), and mean weight at observation was 1556.8 g (range = 880–2,450). Infants' mean NTISS score (illness severity) was 9.24 (range = 3–21), and mean Apgar scores at the first and fifth minutes after birth were 5.43 and 7.50, respectively. Of the 20 nurses employed in the unit during the study, 16 (80%) met the study criteria, cared for infant participants, and agreed to be observed. These nurses had on average 3.64 ± 2.90 years of neonatal experience.
Infant sleep was observed for a total of 2,160 hours. The most common state was QS (mean = 31.48 ± 16.65). The most commonly occurring caregiving pattern was no caregiving (mean = 50.96 ± 10.27), followed by routine care (mean = 5.14 ± 6.30). The most commonly occurring infant position was lateral (mean = 30.97 ± 26.82), followed by prone (14.72 ± 23.74). The mean occurrence of NNS use was 15.71 ± 21.67 per hour.
Infant Quiet and Active Sleep States
All infant states were analyzed after adjusting for effects of time, infant GA, and PMA, illness severity, positioning, NNS use, and patterns of caregiving. As the occurrence of no caregiving, social interaction, lateral positioning, and NNS use per unit increased, the occurrence rate of QS increased significantly by 3.6%, 3.4%, 3.4%, and 0.2%, respectively. However, as the occurrence of routine care and intrusive caregiving per unit increased, the occurrence rate of QS decreased by 2.9% and 12.7%, respectively. The occurrence rates of supine and prone infant positioning was not associated with the occurrence of QS. Infants' illness severity did not significantly increase the occurrence rate of QS. The occurrence of QS was significantly moderated by .1% due to the interaction effects between no caregiving and infant lateral positioning, between social interaction and supine positioning, and between social interaction and lateral positioning. For details, see Table 2.
|Variable||B||SE||Wald Chi-squared||p-value||Rate ratios||95% Confidence interval|
|3rd versus 1st||0.031||0.054||0.339||.560||1.032||−0.074||0.137|
|2nd versus 1st||−0.006||0.036||0.031||.859||0.994||−0.077||0.064|
|Use of NNS||0.002||0.001||14.962||<.001||1.002||0.001||0.003|
|No caregiving × lateral||−0.0006||0.0002||5.586||.018||0.999||−0.001||−0.0001|
|Social interaction × supine||−0.0006||0.0003||4.484||.034||0.999||−0.001||−0.00005|
|Social interaction × lateral||−0.001||0.0003||13.407||<.001||0.999||−0.002||−0.0005|
As the occurrence rates of no caregiving, social interaction, or routine caregiving per unit increased, the occurrence of AS increased significantly by 5.0%, 2.5%, and 5.9%, respectively. However, NNS use was negatively associated with the occurrence of AS. As the frequency of infants' lateral and prone positioning per unit increased, the occurrence of AS increased significantly by 4.7% and 3.2%. The occurrence of AS was significantly moderated by 0.1% due to the interaction effects between no caregiving and infant lateral positioning, and between no caregiving and infant prone positioning. For details of AS, see Table 3.
|Variable||B||SE||Wald Chi-squared||p-value||Rate ratios||95% Confidence interval|
|3rd versus 1st||−0.102||0.091||1.257||.262||0.903||−0.280||0.076|
|2nd versus 1st||0.004||0.063||0.004||.947||1.004||−0.119||0.127|
|Use of NNS||−0.006||0.001||27.111||<.001||0.994||−0.008||−0.003|
|No caregiving × lateral||−0.0008||0.000||8.830||.003||0.999||−0.001||−0.0003|
|No caregiving × prone||−0.0006||0.000||4.825||.028||0.999||−0.001||−0.0001|
Infant Transition and Wake States
As the occurrence of social interaction or routine caregiving per unit increased, the occurrence rates of the transition state increased significantly by 6.4% and 7.0%, respectively. However, as the occurrence of infant supine positioning per unit increased, the occurrence of the transition state decreased significantly by 14.7%. As the infant's GA per unit increased, the occurrence of the transition state decreased by 19.1%. As infant illness severity per unit increased, the occurrence rate of the transition state decreased significantly by 15.3%. Occurrence of the transition state was significantly higher during the third observation day than during the first day. As the occurrences of no caregiving increased, the occurrence of the transition state increased by 4.8%, but intrusive caregiving did not. Occurrence of the transition state was significantly amplified by 0.3% due to the interaction effects between no caregiving and infant supine positioning, between social interaction and supine positioning, and between routine caregiving and infant supine positioning. For details of the transition state, see Table 4.
|Variable||B||SE||Wald Chi-squared||p-value||Rate ratios||95% Confidence interval|
|3rd versus 1st||0.286||0.124||5.363||.021||1.331||0.044||0.528|
|2nd versus 1st||0.164||0.128||1.639||.201||1.178||−0.087||0.415|
|No caregiving × supine||0.003||0.001||12.414||<.001||1.003||0.001||0.004|
|Social interaction × supine||0.003||0.001||9.969||.002||1.003||0.001||0.004|
|Routine caregiving × supine||0.003||0.001||12.137||<.001||1.003||0.001||0.005|
As the occurrence of intrusive caregiving per unit increased, the occurrence rates of wake states increased significantly by 90.4%. The occurrences of no caregiving, social interaction, or routine caregiving were not associated with the occurrence of wake states, nor did the occurrence of infant positioning and NNS use. As infants' GA per unit increased, the occurrence of wake states decreased significantly by 24.9%. As infant illness severity per unit increased, the occurrence rates of wake states decreased significantly by 13.9%. The occurrence of wake states was significantly moderated by 0.9% due to the interaction effects between intrusive caregiving and infant lateral positioning. For details, see Table 5.
|Variable||B||SE||Wald Chi-squared||p-value||Rate ratios||95% Confidence interval|
|3rd versus 1st||0.143||0.210||0.464||.496||1.154||−0.269||0.556|
|2nd versus 1st||0.179||0.196||0.829||.362||1.196||−0.206||0.564|
|Use of NNS||0.002||0.005||0.192||.661||1.002||−0.007||0.011|
|Intrusive caregiving × lateral||−0.009||0.004||4.308||.038||0.991||−0.018||−0.0005|
Infant Fussy or Crying State
As the occurrence of intrusive caregiving per unit increased, the occurrence rates of the fussy or crying state increased significantly by 100%. The occurrence of infant supine positioning was not associated with the occurrence of the fussy or crying state, nor was no caregiving and social interaction. As infants' illness severity per unit increased, the occurrence of fussy or crying state increased significantly by 9.7%. The occurrence of fussy or crying state was significantly moderated by 1.1%, 0.9%, and by 1.0% due to the interaction effects between intrusive caregiving and infant supine positioning, between intrusive caregiving and infant lateral positioning, and between intrusive caregiving and prone positioning, respectively. For details, see Table 6.
|Variable||B||SE||Wald Chi-squared||p-value||Rate ratios||95% Confidence interval|
|3rd versus 1st||0.041||0.154||0.071||.791||1.042||−0.261||0.343|
|2nd versus 1st||0.074||0.113||0.433||.510||1.077||−0.147||0.296|
|Use of NNS||0.002||0.002||0.677||.411||1.002||−0.003||0.007|
|Intrusive caregiving × supine||−0.011||0.002||20.910||<.001||0.989||−0.015||−0.006|
|Intrusive caregiving × lateral||−0.009||0.002||13.537||<.001||0.991||−0.013||−0.004|
|Intrusive caregiving × prone||−0.010||0.003||11.965||<.001||0.990||−0.016||−0.004|
The unique contribution of this study was to relate the effects of different stress-inducing caregiving activities on preterm infants' states after adjusting for time effects and possible influencing factors across three 24-hour periods in the NICU. In contrast, previous studies did not thoroughly categorize caregiving activities by their stress-inducing level, nor did they adjust for effects of time and other factors influencing infants' states (Ingersoll & Thoman, 1999; Thomas et al., 2008).
Our study findings are similar to a report that QS increased when the infant was alone with no caregiving offered, and waking states increased when the infant was with caregivers (Brandon et al., 1999). Our results are also consistent with reports that with more intrusive caregiving, infants' fussy, or crying state increased (Brandon et al., 1999; Liaw, Yang, Chou, Yang, & Chao, 2010; Liaw, Yang, Yuh, & Yin, 2006). In the present study, the occurrence of infants' wake and fussy or crying states increased with increasing frequency of intrusive caregiving. The increased frequency of infants' AS and transition states with increasing frequency of no caregiving was also likely due to environmental stimuli such as noise and light or infants' younger GA (Holditch-Davis et al., 2004; Ingersoll & Thoman, 1999). The frequency of no caregiving can increase the occurrence of both QS and AS, which are essential for preterm infants' energy preservation, protein synthesis, recovery from illness, growth, and neurosensory development (Euston et al., 2007; Graven & Browne, 2008). No caregiving should be reinforced to increase the time of AS and QS.
Social interactions may calm infants to stop crying and go to sleep, serve as stressful stimuli for extremely ill preterm infants, or be ineffective (infants continue crying) if infants' needs are not met. In this study, social interactions increased the occurrences of QS, AS, and transition states. Social interactions are necessary for infant development (Snow & McGaha, 2003), but such interactions should be offered based on infants' needs and states to avoid the possibility of becoming stressful stimuli.
Routine caregiving was provided by caregivers at regular times, and intrusive caregiving was usually delivered based on clinical needs. Both routine and intrusive caregiving may stress infants and increase their arousal and fussiness (Brandon et al., 1999; Liaw et al., 2010; Liaw et al., 2006; Peters, 1998). As infants' arousal and fussiness increase, their motor movements increase, in turn increasing their metabolic rates as well as oxygen and energy consumption (Lehtonen & Martin, 2004). In addition, stimuli from caregiving interrupt infants' sleep, potentially influencing their brain development (Graven & Browne, 2008).
Routine and intrusive caregiving often present a dilemma to NICU caregivers because the numerous procedures necessary for infant survival are also stressful and potentially disruptive to infant sleep, growth, and development (Brandon et al., 1999; Liaw et al., 2006; Peters, 1998). Caregiving activities that are necessary to maintain an infant's survival should be provided with minimal stress and pain to protect infants' sleep. To avoid disturbing infants' sleep, caregiving activities should be clustered to allow time for sleep cycles and offered based on infant states. For example, when an infant has just gone into QS, it is not a good time to wake him/her to offer even benign caregiving, such as feeding or social interaction. This does not mean that nurses should wait for infants to awaken fully before giving care. Creative caregivers can note infants' sleep time, recognize changes in their states and sleep cycles, and time caregiving activities to fit infant sleep cycles. After several sleep cycles, infants start to increase body movements and open their eyes occasionally; then caregivers can wake them for caregiving.
Positioning is another important influence on infant states. Our findings are unique in reporting that lateral positioning facilitated infant QS and AS, and prone positioning increased the occurrence of AS. Moreover, when infants were supine, non-intrusive caregiving increased the occurrence of the infant transition state. However, the effects of supine positioning decreased the occurrence of the transition state in this GEE model. Most of our findings differ from previous reports that infants in the prone position sleep more efficiently (Bhat et al., 2006) and their arousal decreases (Sahni et al., 2005). Sleeping in the prone position, however, has been identified as a major risk factor for sudden infant death syndrome (SIDS; Richardson, Walker, & Horne, 2008). To reduce the risk of SIDS, both the American Academy of Pediatrics (AAP, 2011) and the International Society for the Study and Prevention of Perinatal and Infant Death (ISPID, 2011) recommend that parents place infants in a supine position and provide a pacifier.
Our findings suggest that both lateral positioning and NNS use can facilitate infant sleep. Future researchers should explore whether the lateral sleep position can reduce the incidence of SIDS. Some researchers, however, have reported no significant differences in sleep/wake states between the supine and prone positions (Ariagno et al., 2003; Goto, Maeda, Mirmiran, & Ariango, 1999). The effects of positioning or the interaction of caregiving and positioning on infant sleep still warrant further study. In the mean time, infants should be positioned based on their individual sleep cycles and habits. When infants become more awake, the supine position is more suitable for interacting with caregivers, the environment, or receiving routine caregiving. If infants need rest, the lateral or prone position is better.
NNS use increased the occurrence of QS but decreased the occurrence of AS and was not associated with the occurrence of wake, transition, and fussy or crying states. This finding is comparable to a report that NNS use did not alter the frequency or duration of arousal in sleeping infants (Hanzer et al., 2009). NNS use has been shown to help modulate infants' sleep (Bingham et al., 2011; Pinelli & Symington, 2010). This effect was proposed to be due to the emergence of immature developmental patterns of NNS (burst organization and consistency of suck waves) associated with decreased arousal level. However, mature patterns of NNS activity have been correlated with increased arousal. Both the quality and quantity of NNS behaviors had complex relationships with preterm infants' behavioral states (Bingham et al., 2011). In this study, we measured the frequency of NNS but not its quality. More studies are needed to differentiate the effects of NNS patterns on infant states.
Finally, our results suggest that infants' states are influenced by their own characteristics. Infants with younger GA tended to have higher occurrences of the transition and wake states. However, older PMA infants were apt to have more occurrences of the wake state. Infants whose illness was more severe spent less time in wake and transition states, but more time in fussy or crying state. This finding is likely explained by severely ill infants often feeling uncomfortable and tending to be fussy or crying. They also may have been too ill to stay awake. These findings are consistent with a report that older infants spent more time in the wake state, and those with severe disease were less awake but fussier (Holditch-Davis et al., 2004). Overall, our findings demonstrate how the social-contextual environment influences infant states and are consistent with assumptions of the transactional model (Sameroff, 2009).
First, infant state changes in response to caregiving and the environment varied greatly at each observation. In this study we only considered the social contexts that influenced infant sleep/wake states. Future investigators should measure sleep cycles and how often infants are disturbed by caregiving. Second, other factors such as noise, light, and temperature in the NICU, incubator humidity, and emergency treatment may have influenced infant states. Such factors need to be controlled in future studies. Third, our sample included only 30 preterm infants, due to time, staffing, and financial limitations. A larger sample would allow certain infant characteristics, such as GA and illness severity, to be blocked in analysis to better control factors that might influence outcome variables. Fourth, in this study we measured only the frequency of NNS use. Future researchers may need to include measures of quality of NNS use (e.g., burst organization and consistency of suck waves). Fifth, the short-term design limited determination of long-term outcomes. Future researchers could use the results of this study to develop interventions that facilitate infant sleep and then follow infants' long-term outcomes such as neurodevelopmental or biological outcomes.
Sleep and the rest it provides are important for infants' neurosensory development, illness recovery, healing, and health. QS and AS are both essential for preterm infants' healing, growth, and neurosensory development. Our results suggest that no caregiving, social interaction, lateral position, and NNS use can be offered to facilitate the occurrence of QS, whereas, routine and intrusive caregiving should be minimized to protect the infants' QS. No caregiving, social interaction, lateral, and prone positioning can facilitate AS. Routine and intrusive caregiving can increase infants' fussy or crying state, which can increase their energy consumption. Clinicians need to ensure that these caregiving activities are provided only when truly necessary. Finally, caregiving, positioning, and NNS use should be provided based on preterm infants' sleep/wake states, needs and with minimal stress and pain to preserve their sleep. Wise and creative clinicians can recognize infants' states, and adjust caregiving, positioning, and NNS use to modulate infant sleep and facilitate complete sleep cycles. Our results provide evidence to guide clinicians in facilitating preterm infant sleep by reinforcing positive factors and avoiding negative ones.