Impact of Diaper Change Frequency on Preterm Infants’ Vital Sign Stability and Skin Health: A RCT

Abstract

Background:

Bundling nurse caregiving interventions are promoted to minimize infant stress.

Purpose:

To evaluate impact of bundled nursing care and diaper change frequency on vital sign stability and skin health of preterm infants born ≤ 32 weeks gestation.

Method:

Stable preterm infants on a 3-hour feeding schedule were randomly assigned to 3- vs. 6-hour diaper changes. Diapers were changed prior to 6 hours if stool was present. Direct observation of bundled care events (BCE) identify caregiving activities during each BCE. Skin pH, transepidermal water loss (TEWL), and neonatal skin condition scores (NSCS) were obtained. Vital sign data (HR, RR, O₂ saturation) was downloaded from bedside monitors.

Results:

Forty-six infants contributed to 605 BCEs. BCEs lasted on average 28 minutes and included nine different activities (e.g., vital signs, feeding). Significant increases in heart rate during BCEs occurred in approximately half of the observations. Among observations with a diaper change increases in heart rate during diapering occurred in over 74% of observations Infants who were awake at the beginning of BCEs had 48% lower odds of having a change in heart rate than infants who were sleeping (p=0.02). There were no group differences (3- vs. 6-hour diaper change) in skin health outcomes (TEWL, pH, NSCS).

Conclusion:

Reducing diaper change frequency without stool present should be considered to minimize caregiving stress in preterm infants. Additional research should evaluate the intrusiveness and clusters of activities that significantly impact physiologic stability to better individualize the timing of routine yet intrusive activities.

Introduction

Advances in neonatal care have resulted in an increased survival rate among very preterm infants. However, morbidities associated with prematurity including poor neurodevelopmental outcomes remain high, especially among the most immature infants.[1–3] The etiology of neurodevelopment impairments in preterm infants is multifactorial and includes prenatal, antenatal, and postnatal factors. One postnatal factor that may hinder infants’ health and development is the environment of the neonatal intensive care unit (NICU).[4–6]

The varying amounts of tactile, auditory, and visual stimulation in the NICU environment are just a few of the major differences from the intrauterine environment in which fetal development occurs. Following preterm delivery, both life-sustaining and routine caregiving interventions may be experienced by the infant as uncomfortable, painful, and/or disruptive to sleep and often result in excessive non-contingent sensory stimulation. A systematic review of painful procedures in the NICU found that infants experienced 7.5–17.3 painful procedures per day, with more severely ill and immature infants experiencing the most painful procedures.[7] Due to the atypical nature of the NICU environment compared to the intrauterine environment and the requirements of caregiving, the NICU has been characterized as stress provoking.

A body of literature has emerged highlighting the short- and long-term negative effects of ongoing NICU stressors on infant health and neurodevelopment. Preterm infants with higher exposure to acute (e.g. heelstick) and chronic (e.g. indwelling nasogastric tube) stressors demonstrated more stress signs and a poorer habituation response at 36–37 weeks postmenstrual age.[8] Also, the cumulative stress of skin breaking procedures of former preterm infants predicted lower cortisol responses and diurnal cortisol levels at seven years of age.[9] In addition, a recent systematic review illuminates the relationship between the stress experienced by preterm infants and epigenetic alterations of genes that influence infant health, disease, and neurodevelopmental outcomes. In this review, infants exposed to NICU environmental stressors exhibited alterations in genes that affect the hypothalamic-pituitary adrenal axis (HPA) (e.g. cortisol regulation) and serotonin function and regulation (e.g. negative emotionality).[10]

To minimize the negative impact the NICU environment may have on developmental outcomes, family centered developmental care programs have been broadly adopted worldwide. NICU developmental care interventions focus on environmental modifications that minimize excessive touch, noise and light, and structure infant caregiving in partnership with families that promotes sleep, proper positioning and handling, skin integrity, optimal nutrition, and minimizes stress and pain.[11] These interventions are all aimed to improve infant outcomes. A recent systematic review of the developmental care interventions delivered in the NICU found some short-term improvements in outcomes, but long-term positive impacts on preterm infant development have not been observed.[12]

One neuroprotective strategy used as part of developmental care includes the grouping of care activities around a single caregiving event described as “clustering” or “bundling care”.[13–15] Individualized developmental care dictates that a bundled care event is halted if an infant displays significant behavioral stress cues, but knowledge about the inclusion and exclusion of care activities that elicit distress have not been systematically studied. Certain care activities during bundled care may be dictated by the infant’s treatment plan or needs, but other activities like diapering may be optional. While painful and intrusive interventions like heel sticks are often described as a source of infant stress, routine interventions like diaper changes can also be a significant stressor.[16–18]

Understanding the impact of when to include optional, yet stress provoking interventions like diapering is essential to minimize NICU environmental stress in very low birth weight (VLBW) infants. However, there is limited data on how changing the timing of these optional interventions impact infant health. For example, fewer diaper changes may negatively impact infant skin health in the diapered area. The skin has several important functions including being a barrier, acid-mantle formation, temperature regulation, water and electrolyte regulation, and tactile sensation. While skin adaptation is accelerated immediately following birth and stabilizes by one month of age, infant skin development continues for up to 12 months of life.[19] Diapered skin health is impacted most by prolonged exposure to urine and digestive stool enzymes, enteral feeding substrates, and the microbiome.[20–22] In addition, functional properties of diapered preterm infant skin, like moisture and pH, differ from elsewhere on the body. Transepidermal water loss (TEWL) on the buttocks is significantly higher than elsewhere on the body in preterm infants, and for these infants skin pH decreases on other body sites, but increases on the buttocks over time.[23] These differences across body sites likely contribute to the high incidence of diaper dermatitis in the neonatal intensive care unit (NICU) (21% to 45.5%).[20, 24, 25]. Therefore, implementation of potential stress reducing interventions such as decreased frequency of diaper changes could increase diaper dermatitis, and thereby impact infant health even if less stressful to the infant.

Given the lack of data around what specific activities to include in bundled care, we explored the impact of changes to the diapering routines (in a bundled care event) on the overall health of VLBW preterm infants exposed to every 3-hour bundled care events. We included a focus on skin health because it was important to understand the potential benefits of decreased infant stress and any potential skin health trade-offs associated with longer versus shorter time between diaper changes. Specific research questions addressed were: 1) Does infant vital sign status (heart rate, respiratory rate, oxygen saturation) during caregiving differ over time between infants randomized to 3- vs. 6-hour diaper changes, and 2) Does infant skin health [skin pH, transepidermal water loss (TEWL), neonatal skin condition score (NSCS)] of the buttocks differ over time between 3- vs. 6-hour diaper changes?

Methods

A randomized controlled parallel design was used to evaluate the effect of bundled diaper care on infant vital sign status and skin health.

Setting and Participants

Infants were recruited from the Level IV NICU at a Southeastern U.S. academic medical center between December of 2018 and July of 2019. The study was approved by the institutional review board. The ICN is a Level IV, 67-bed nursery and one of the area’s leading referral centers for infants born prematurely. The nursery practices individualized developmental care with standard bundled care intervals that vary based upon an infant’s birth weight and feeding status. Smaller, more immature and sicker infants receive “hands on” bundled care every 4 to 6 hours while more stable infants receive bundled care every 3 hours, typically corresponding to their feeding schedule. Infants who were transferred to one of the three stepdown units were also recruited and continued to receive the intervention when transferred to the stepdown unit, including the Level III nursery at an affiliate hospital in the same city. Inclusion criteria were infants who were born at ≤ 32 weeks gestation, received standard 3-hour bundled care, and were expected to remain in a study nursery for at least 4 weeks. Exclusion criteria were diagnosis of neonatal abstinence syndrome, humidified incubator at the time of random assignment, pre-existing or genetic skin condition, presence of a diaper rash, or severe illness requiring minimal stimulation. GEE modeling was used to examine the relationship between diaper change groups and heart rate instability. GEE models can require different sample sizes than linear mixed models. Following the power calculations outlined by Li and McKeague, the achieved power using GEE to model heart instability as a dichotomized variable was 94% to detect a 10% main effect.[26] As such, the study design was adequately powered.

Intervention

Infants were randomly assigned and stratified by birth weight (≤ 800 grams, >800 grams to 1150 grams, > 1150 grams) to receive either 3- or 6- hour diaper changes once they were physiologically ready for every 3-hour enteral feedings and thereby 3-hour bundled care. Other care activities included in the standard 3-hour bundle varied based upon infant treatment needs, but usually included a diaper change. While 3-hour bundled care is standard of care, the NICU is a clinical environment in which caregiving cannot be expected to occur exactly every three hours. The only difference in the 3- hour and 6-hour diaper change group was the timing of diaper changes; infants received a diaper change every 3 or 6 hours. Infants in the 6-hour bundled diaper change group continued to receive 3-hour bundled care for all other care activities (e.g., feeding, temperature check) as dictated by infant care needs. Infants assigned to have their diaper change every six hours would have their diaper checked by the During the 3-hour caregiving interval if an infant in the 6-hour diaper change group was not due for a diaper change, the nurse peeked into the leg of the diaper to ensure there was no stool present. If stool was present, the diaper was changed and the timing of the next diaper change was scheduled for 6 hours from that bundled care event.

Procedure

Parents were approached for pre-consent after eligibility screening was completed (born at ≤ 32 weeks gestation, without neonatal abstinence syndrome, or pre-existing or genetic skin condition) (see Figure 1). Following consent, infants were monitored until all inclusion criteria were met including the infant’s transition to every 3-hour interval enteral feedings which dictated the transition to the standard 3-hour bundled care. With all inclusion criteria met, infants were randomized to an intervention group by a predetermined statistical computation within REDCap™ to ensure the research staff was unaware of the group assignment ahead of intervention initiation. Due to the nature of research, the study staff and nurses who implemented the intervention could not be blinded to group assignment. Study staff had to be aware of group assignment to schedule and collect observational data. The intervention and data collection began after group assignment and data collection continued for six weeks after which bundled diapered care resumed according to unit protocol for the 6-hour group.

Figure 1.

Consort Flow Diagram

Measures

Study data included maternal and infant demographics and medical history extracted from the electronic health record (EHR) to describe the sample and use as covariates. Data collected around each bundled care event included physiologic data (e.g., heart rate, respiratory rate, oxygen saturation), caregiving activities included in the events, and biomarkers of skin health (pH, transepidermal water loss, Neonatal Skin Condition Score).

Electronic Health Record Data: EHR data included maternal age and history of skin disease and infant demographic, birth history data, pertinent illness characteristics (e.g., diaper dermatitis), and care needs with the potential to influence study outcomes, such as the use of products to protect the diapered skin area. The actual number of diaper changes during the study period was also collected.

Physiologic instability: Heart rate (HR), respiratory rate (RR), and oxygen saturation (O₂ sats) were measured continuously with the unit’s standard GE Cardiopulmonary monitor from the first day of study initiation through the end of the study period. Extraction of vital sign data occurred every 1-minute as a point in time variable and was stored on a secure server. The 1-minute values were extracted to match the time of each bundled care observation beginning with the 90 minutes prior to the beginning of the bundled care event and ending with the conclusion of bundled care. Exception to the following timeframe occurred with a few observations due to removal of monitor leads for bathing or repositioning/replacing of monitor leads.

Physiological instability was defined as having a mean vital sign value during a bundled care event that was ≥1 standard deviation above or below the mean physiologic value during the 90 minutes before the event started. In addition, clinically meaningful parameters were added to indicate exceptions to this definition. For HR, mean values ≥ one standard deviation below the mean was categorized as physiologically unstable only if the mean value was < 100. For respiratory rate, mean values that were ≥ one standard deviation above or below the mean at baseline were categorized as physiologically unstable only if the mean value during bundled care was < 20 or ≥ 60. Finally, for oxygen saturation, mean oxygen saturation values that were ≥ one standard deviation below were categorized as physiologically unstable if the mean value during bundled care was < 90.

Bundled Care Event: The presence of caregiving activities in a bundled care event was recorded through direct observation of all activities at the bedside during weekly observations conducted by study staff. Nursing care was observed at the bedside 3 times a week based on their bundled care schedule. For the 6-hour diaper change group, a 4^th observation occurred each week in order to capture a bundled care event that did not include a diaper change or skin health measures. Parents could be present and participate in the bundled caregiving event. The length of time for each bundled care event was operationally defined as beginning when the nurse or parent touched the infant for the first time (beginning of caregiving). The bundled care event ended when the last care activity was completed (typically the feeding) whether by the nurse or parent. All care activities were recorded in the sequential order they were observed and included interventions such as obtaining vital signs, physical assessment, feeding, medications, procedures, and diapering. Diapering was defined as the time from when the diaper was unfastened until it was fastened again with the new diaper in place. During diapering biophysical measurements of skin health status were collected as well as other caregiving activities such as abdominal girth measurements and skin cleansing. In addition to recording all care activities, the infant’s sleep-wake state at the beginning of the bundled care event was recorded. Infant state was defined as either sleeping (eyes closed, few or no body movements), or as quietly alert (eyes open, few or no body movements), actively alert (eyes open and frequent body movements), or fussy or irritable (occasional to full crying). During the observation, study staff also documented any developmental interventions that occurred, including offering a pacifier, providing flexed support or containment, swaddling with a blanket, or a two-person assist with a care activity. The entire bundled care event was also categorized as developmentally sensitive (yes/no) by the study staff based upon the nurse’s responsiveness to infant cues and use of appropriate developmental care interventions throughout bundled care. Interrater reliability established between three study staff raters was > 90%.

Skin Health Assessment: Biophysical measurements of skin health measures were obtained three times per week during a diaper change at the time of direct observation of a bundled care event. Both pH and TEWL measures were obtained from the buttocks and chest prior to standard cleansing during the diaper change. Chest pH and TEWL measures were included as a control site outside of the diapered area to determine whether any changes in skin pH or TEWL associated with diaper changing frequency were specific to the diapered area or reflected generalized skin health. If there was stool present in the diaper, the diaper was used to remove as much stool as possible from the buttocks before proceeding with the pH and TEWL data collection.

The ExStik™ PH100 by Extech® instruments was used to measure skin pH in the range of 0.00 to 14.00 during a diaper change. The small flat surface electrode was placed at the center of the test site (either chest or buttocks). Holding the probe flush over the skin at the test site, the measurements were taken first at the chest followed by the buttocks.

TEWL values were obtained using the DermaLab® TEWL probe (Cortex Technology, Hadsund, Dermark), which consists of an open probe with paired sensors placed at different distances from the skin. Humidity and temperature are measured in each sensor to calculate vapor pressure gradients. The difference between two vapor pressure gradient measures is representative of TEWL at that point on the skin.[27]

The NSCS is routinely documented by the care nurse at least twice daily in the infants EHR flowsheet. The NSCS is a reliable and valid measure of preterm infant skin condition[28]. The NSCS assesses three areas: dryness, erythema, and skin breakdown that are scored 1–3 with scores of 1 being normal.

Statistical Analysis

Descriptive statistics and bivariate tests were conducted to summarize infant demographic and illness characteristics and maternal characteristics for the entire sample and each diaper change group. Descriptive statistics and bivariate tests also summarized the characteristics of the intervention and bundled care events such as postmenstrual age (PMA) at intervention initiation, number of diaper changes in the 24 hours prior to observation of bundled care and skin health data collection, number of diaper changes across the intervention, and diagnoses at discharge. In addition, the frequency of physiologic instability (HR, RR, O2 saturation) from baseline (90 minutes prior to the bundled care event) compared to the bundled care event was summarized.

To examine the relationship between bundled care event predictors and heart rate instability, generalized estimating equations (GEE) with a logit link were used. Due to the low occurrence of instability (see Table 2), models were not conducted for respiratory rate and oxygen saturation. In this study, diaper change group (3- vs. 6-hour diaper change) and presence of a diaper change were confounded because only infants in the 6-hour group had bundled care events without a diaper change. To control for this, diaper change group and presence of a diaper change were tested with separate samples: an analysis of events within the 6-hour group tested the impact of a diaper change, while an analysis of events with a diaper change, which included events from both diaper change groups, tested the impact of diaper change group. In all models, days since intervention initiation was entered as a fixed effect to assess the effect of time, as well as fixed effects for infant demographics and bundled care observation characteristics, including PMA in weeks at study initiation, chronological age in weeks at study initiation, infant sleep-wake state at the time of the observation, and presence of developmentally appropriate caregiving, which include pacifier use and swaddling. Because the effect of the presence of a diaper change is confounded by observation duration and number of observation events, the within 6-hour group model also included fixed effects for these covariates. Models of heart rate instability during diapering included a fixed effect for the duration of diapering. Continuous predictors were centered, by subtracting the mean from each score, so that the exponentiated intercept represented the odds of heart rate instability at the mean of continuous predictors. For infant sleep-wake, categories representing states of being awake (quietly alert, actively alert, fully or irritable) were collapsed due to small cell sizes, resulting in two categories: sleep vs. awake. A random effect for the intercept was also entered. Estimates were exponentiated to produce odds ratios. While GEE is a robust modeling technique and can handle large imbalances (less than 90%:10%) in sample sizes between groups, an imbalance in diaper changes to no diaper change could threaten our non-significant result. We had 257 observations in the 6-hour diaper change group, 77.8% that had a diaper change and 22.2% (n=73) that did not have a diaper change. When we used other techniques (cross-validation, weighted logistic regression, etc.) to evaluate our results, the analyses suggested the results hold consistent.

Table 2.

Intervention and Bundled Care Observation Characteristics

	6-Hour n=26			3-Hour n=20			Total Group n=46
	Mean	SD	Median	Mean	SD	Median	Mean	SD	Median	t	p-value
PMAa at intervention initiation	34.80	3.95	34.07	35.43	3.59	34.86	35.07	3.77	34.64	0.56	0.58
PMA at hospital discharge	40.24	5.76	37	41.50	4.80	41	40.80	5.33	40	−0.78	0.44
Intervention length (days)	25.31	14.10	23.50	31.80	9.24	33.0	28.13	12.54	29.50	−1.78	.08
Total # study diaper changes	147.81	83.60	138.0	240.45	77.74	247.50	188.09	92.69	204.0	3.84	<.001
# diaper changes in last 24 hours prior to observation	5.68	1.38	5	7.87	0.64	8	6.68	1.55	7	25.63	<.001
	6-Hour n=330			3-Hour n=275			Total Group n=605
	Mean	SD	Median	Mean	SD	Median	Mean	SD	Median	t	p-value
Duration in minutes: BC event	26.98	13.39	24.83	29.41	13.74	28.23	28.09	13.60	26.63	2.19	.03
Duration in minutes: diapering	5.85	2.28	5.67	5.97	1.84	5.65	5.91	2.07	5.67	0.69	.49
Caregiving activities	9.33	3.14	10	9.55	2.59	9	9.43	2.90	9	0.94	.35
Pacifier	1.61	1.72	1.00	1.89	1.77	2.00	1.74	1.75	1.00	−1.96	.051
Flexed support	0.82	1.44	0.00	0.73	1.39	0.00	0.78	1.42	0.00	0.81	.42
Swaddling	1.48	1.27	1.00	1.72	1.38	2.00	1.59	1.33	1.00	−2.17	.03
Two-person assist	0.38	0.93	0.00	0.43	0.75	0.00	0.40	0.85	0.00	−0.71	.48
	n	%		n	%		n	%		χ2	p-value
Developmentally sensitive	278	92.36		243	96.05		521	94.04		3.34	.07
Infant state: awake	79	38.92		74	37.0		153	37.97		0.16	.69
HR change during observation	148	48.84		129	47.96		277	48.43		.05	.83
RR change during observation	2	0.66		0	0		2	0.35			.50
O2 sat change during observation	0	0		0	0		0	0
HR change during diaper change	185	73.41		200	75.19		385	74.32		0.21	.64
RR change during diaper change	5	1.98		6	2.25		11	2.12		0.05	.83
O2 sat change during diaper change	5	2.08		2	0.78		7	1.41			.27

^aPostmenstrual Age (PMA) Note. Fisher’s exact test conducted for RR change during observation and O2 sat change during diaper change. Bivariate test not conducted for O2 sat change during observation due to no O2 sat changes during observation.

To examine the relationship between diaper change groups (3- vs. 6-hour) and continuous measures of neonatal skin health (skin pH, TEWL), linear mixed models (LMM) were used. Outcomes in these models consisted of within-subject change over time in skin pH and TEWL of the chest and buttock and the ratio of chest relative to buttock pH and TEWL. The ratio of chest relative to buttock pH and TEWL were examined to identify potential differences in the change in pH and TEWL over time by body area. In all models, outcomes were regressed on diaper change group and the number of days since study initiation. The interaction between these was also included to examine whether the rate of change across time depended on diaper change group. A random effect for the intercept was also included, to capture individual variation in baseline levels of skin health. Demographic and patient characteristics were also included as fixed effect covariates, including PMA and chronological age in weeks at study initiation, centered; presence of stool at the observation; and whether skin care products were used since the previous diaper change. Because the range of scores for the NSCS was limited (3–5), scores on this measure were dichotomized as 3 (normal) vs. 4 or 5 (abnormal). To examine the relationship between diaper change group and dichotomized NSCS, a generalized estimating equation (GEE) with a logit link was used. Fixed effect predictors in this model were the same as those in LMM models of skin pH and TEWL, with main effects and interaction for diaper change group and the number of days since study initiation, along with demographic and patient characteristic covariates. All analyses were conducted in SAS Version 9.4.

Results

Demographic statistics and bivariate tests summarizing infant and maternal characteristics are presented in Table 1. Infants had a mean gestational age at birth of 28 weeks and a mean birth weight of 1117 grams. The majority of infants were Non-Hispanic (89%). Gender differed by group (χ²=5.34, p=.02), in that female gender was more common in the 6-hour group (69%) than in the 3-hour group (35.00%). Feeding difficulty was common among infants (37%). C-section delivery occurred in 59% of infants.

Table 1.

Participant Demographics and Illness Characteristics

	6-Hour n=26		3-Hour n=20		Total Group n=46
Infant Variables	Mean	SD	Mean	SD	Mean	SD	t	p-value
Gestational age at birth (weeks)	28.4	2.8	27.6	2.8	28.0	2.8	0.95	.97
Birth weight	1190.1	484.5	1025.5	376.2	1117.0	442.7	1.25	.22
	n	%	n	%	n	%	χ2	p-value
Gender: Female	18	69.2	7	35.0	25	54.4	5.34	.02
Ethnicity: Non-Hispanic or Latino	24	92.3	17	85.0	41	89.1		.64
Problem/Diagnoses
Diaper dermatitis	0	0.0	1	5.0	1	2.2		.43
UTI	3	11.5	4	20.0	7	15.2		.68
Feeding difficulty	8	30.8	9	45.0	17	37.0	0.98	.32
Spontaneous perforation	0	0.0	1	5.0	1	2.2		.43
NEC: Surgical	1	3.9	0	0.0	1	2.2		>0.99
NEC: Medical	6	23.08	5	25.00	11	23.9		>0.99
BPD/CLD	1	3.9	1	5.0	2	4.4		>0.99
GERD	1	3.9	0	0.0	1	2.2		.43
C-section delivery	15	57.7	12	60.0	27	58.7	0.02	.87
Maternal variables	Mean	SD	Mean	SD	Mean	SD	t	p-value
Age	30.5	6.7	32.4	5.3	31.2	6.1	−0.98	.33
	n	%	n	%	n	%		p-value
History of skin disease	1	4.2	1	6.3	2	5.0		1.0

Note. UTI=urinary tract infection. NEC= necrotizing enterocolitis. BPD= bronchopulmonary dysplasia. GERD= gastroesophageal reflux disease. CLD=chronic lung disease. Fisher’s exact test conducted for ethnicity, diaper dermatitis, UTI, oliguria, and history of skin disease.

Descriptive statistics and bivariate tests for intervention and observation characteristics are presented in Table 2. Forty-six infants contributed to 605 observations (532 with a diaper change and 73 without a diaper change). The smaller number of observations without a diaper change corresponds to the number of weekly observations performed in each group. Only infants in the 6-hour group had bundled care events without a diaper change (a fourth weekly observation). Infants in both groups had three observations per week to collect skin health data.

Infants were approximately 35 weeks PMA at intervention initiation and 41 weeks PMA at hospital discharge. Observations lasted on average 28 minutes and included on average 9 different caregiving activities but these were not always the same 9 activities. Notably, there were 10 activities (diaper change, first and second repositioning, oral gastric/nasal gastric feeding, monitor probe adjustment, temperature taking, bottle feeding, auscultation assessment, oral gastric/nasal gastric residual check, and blood pressure) present in 40 to 90% of the observations in each group. Observation duration was significantly longer (t=2.19, p=.03) in the 3-hour diaper change group (29.41 vs. 26.98), likely due to the 73 observations in the 6-hour diaper change group that did not include a diaper change. Descriptive statistics of all activities observed are included in supplemental digital content 1. The total number of diaper changes during the study and the number of diaper changes in the 24 hours prior to the observation differed by group (t=3.84, p<.001; t=25.63, p<.001), in that those in the 3-hour group had more diaper changes. The diaper change activity lasted on average 6 minutes and did not differ by group. The frequency of specific developmental interventions throughout the observations were calculated. Flexed support and two-person assist did not differ by group, though total use of a pacifier differed marginally (t=−1.96, p=.051) and swaddling differed (t=−2.17, p=.03), in that the three-hour group used pacifier and swaddling more often. Categorization of the entire bundled care event as developmentally sensitive caregiving occurred in almost all study observations (94%). Infants were awake at the beginning of the observations 38% of the time. A significant change in heart rate during the bundled care event occurred in approximately half (48.4%) of the observations in both groups (100% were increases in HR). However, there were only two significant changes in respiratory rate and no significant change in oxygen saturation during observations of bundled care. Among observations with a diaper change and non-missing physiologic data, a change in heart rate during diapering occurred in over 74% of observations (>99% were increases in HR) with few changes in respiratory rate or oxygen saturations. Among observations without a diaper change and non-missing physiologic data, a change in heart rate during the entire BC event occurred in 37% of observations (100% were increases in HR), with one change in respiratory rate and no changes in oxygen saturation.

Results of GEE models regressing heart rate instability during the entire bundled care event on diapering variables are presented in Table 3. Results regressing heart rate instability on presence of a diaper change, among infants in the 6-hour diaper change group, indicated that this predictor was not significant (OR=1.71, p=.18). In this model, infant sleep-wake state was significant (OR=0.42, p=.02), in that infants who were awake had 58% lower odds of having a significant change in heart rate than infants who were sleeping. The remaining infant characteristic covariates were not significant. GEE model results regressing heart rate instability on diaper change group, in observations with a diaper change, are also presented in Table 3. These indicated that group was not significantly related to heart rate instability (OR=0.69, p=.30). Other infant characteristics covariates in this model were not significant.

Table 3.

Factors Influencing Heart Rate Instability

	6-hour Infant Observations (N=330) (n=257 with; 73 without a Diaper Change)		3- and 6-hour Infant Observations with a Diaper Change (N=532)
	OR	OR 95% CI; [LL,UL]	OR	OR 95% CI; [LL,UL]
Intercept	0.52	[0.21, 1.33]	4.17	[2.18, 8.00]
BC Group	-	-	0.69	[0.35, 1.36]
Diaper change	1.71	[0.78, 3.72]	-	-
Time	1.01	[0.99, 1.03]	1.00	[0.98, 1.02]
PMA	0.93	[0.80, 1.09]	1.01	[0.82, 1.26]
Age	1.05	[0.94, 1.16]	0.95	[0.82, 1.10]
BC duration	1.00	[0.97, 1.03]	-	-
# of activities	0.98	[0.86, 1.11]	-	-
Diapering duration	-	-	0.96	[0.81, 1.14]
Infant State: awake	0.42 *	[0.24, 0.73]	0.63	[0.35, 1.12]
Pacifier	0.91	[0.74, 1.11]	1.03	[0.90, 1.17]
Swaddling	0.93	[0.71, 1.22]	1.00	[0.87, 1.16]

Note. Unstandardized coefficients shown.

^*p<.05

Continuous predictors were centered, by subtracting the mean from each score.

During BC= during BC observation; ORs indicate odds of HR change during each BC observation

BC group=bundled care group

Reference for bundled care diaper group: 6-hour

Time=days since study initiation

PMA= postmenstrual age in weeks at study initiation.

Age=chronological age in weeks at study initiation.

BC duration= BC duration in minutes.

Diapering duration=Diapering duration in minutes.

Number of activities=number of activities in BC.

Reference category for infant state: sleeping.

Results of LMM and GEE models of skin health on diaper change group are presented in Tables 4 and 5. Buttocks pH decreased over time (b=−0.01, p=.01), with each one-day increase in time since study initiation being associated with a 0.01 decrease in buttock pH. Diaper change group, and the interaction between group and time, were not significant (b=−0.07, p=.60; b=0.01, p=.10). Other infant characteristic covariates were not significant. In the model of pH on the chest, group, and the interaction between group and time were not significant (b=−0.07, p=.53; b=0.01, p=.10), whereas time was significant (b=−0.01, p=.03), in that each one-day increase in time since study initiation was associated with a 0.01 decrease in chest pH. Other covariates in this model were not significant. In the model of pH ratio (pH buttocks/pH chest), group, time, and the interaction between these were not significant (b=0.005, p=.84; b=−0.0003, p=.72; b=−0.001, p=.89). In this model, stool was significant (b=0.04, p=.004), in that infants with stool present had 0.04 higher buttock pH relative to chest pH than infants without stool present.

Table 4.

pH Skin Health on Bundled Care Group and Time

	pH Buttocks		pH Chest		pH Ratio
	b	b 95% CI; [LL,UL]	b	b 95% CI; [LL,UL]	b	b 95% CI; [LL,UL]
Intercept	5.73	[5.53, 5.93]	5.22	[5.04, 5.39]	1.11	[1.07, 1.14]
Bundled Care Group	−0.07	[−0.32, 0.19]	−0.07	[−0.29, 0.15]	0.005	[−0.04, 0.05]
Time	−0.01*	[−0.02, −0.00]	−0.01*	[−0.01, −0.00]	−0.0003	[−0.002, 0.001]
Group by time	0.01	[−0.00, 0.02]	0.01	[−0.00, 0.02]	−0.0001	[−0.002, 0.002]
PMA	0.02	[−0.05, 0.08]	0.00	[−0.05, 0.06]	0.002	[−0.01, 0.01]
Age	−0.01	[−0.05, 0.03]	−0.01	[−0.05, 0.02]	0.001	[−0.01, 0.01]
Stool present	0.11	[−0.01, 0.23]	−0.04	[−0.15, 0.06]	0.04*	[0.01, 0.06]
Buttock skin care product use	−0.01	[−0.14, 0.11]	0.05	[−0.06, 0.16]	−0.02	[−0.04, 0.01]

Note. Unstandardized coefficients shown. For NSCS, odds (for intercept) and odds ratios shown.

^*p<.05

^**p<.01.

pH ratio=pH Buttocks/pH Chest. TEWL ratio= TEWL Buttocks/ TEWL Chest. NSCS=Neonatal Skin Condition Score. Time=days since study initiation. PMA=postmenstrual age in weeks at study initiation, centered. Age=chronological age in weeks at study initiation, centered.

Reference for bundled care diaper group: 6-hour.

Table 5.

TEWL and NSCS Skin Health on Bundled Care Group and Time

	TEWL Buttocks		TEWL Chest		TEWL Ratio		NSCS
	b	b 95% CI; [LL,UL]	b	b 95% CI; [LL,UL]	b	b 95% CI; [LL,UL]	OR	OR 95% CI; [LL,UL]
Intercept	36.03	[28.15, 43.90]	26.71	[21.47, 31.95]	1.77	[1.29, 2.25]	0.10	[0.03, 0.36]
Bundled Care Group	−5.26	[−15.33, 4.81]	−8.85*	[−15.54, −2.16]	0.59	[−0.00, 1.19]	0.79	[0.17, 3.68]
Time	0.03	[−0.17, 0.23]	−0.08	[−0.19, 0.04]	0.01	[−0.01, 0.03]	1.00	[0.96, 1.03]
Group by time	0.11	[−0.19, 0.41]	0.08	[−0.08, 0.25]	−0.01	[−0.04, 0.02]	1.04	[0.99, 1.09]
PMA	0.49	[−2.53, 3.51]	0.20	[−1.91, 2.30]	0.06	[−0.07, 0.20]	1.47	[0.92, 2.36]
Age	0.13	[−1.78, 2.04]	0.10	[−1.23, 1.43]	−0.03	[−0.11, 0.06]	0.83	[0.61, 1.12]
Stool present	6.50*	[2.93, 10.08]	−0.92	[−2.88, 1.04]	0.52*	[0.20, 0.84]	0.98	[0.55, 1.74]
Buttock skin care product use	−0.93	[−4.52, 2.66]	0.65	[−1.34, 2.63]	−0.36*	[−0.68, −0.03]	1.05	[0.59, 1.86]

In models of buttocks TEWL, group, time, and the interaction between these were not significant (b=−5.26, p=.31; b=0.03, p=.78; b=0.11, p=.48). For this outcome, stool was also significant, in that TEWL was 6.50 higher in infants with stool present. In models of TEWL on the chest, group was significant (b=−8.85, p=.01), in that TEWL was 8.85 lower in the 6-hour diaper change group than in the 3-hour group. Time, and the interaction between group and time were not significant (b=−0.08, p=.18; b=0.08, p=.32). In the model of TEWL buttocks/chest ratio, group, time, and the interaction between these were not significant (b=0.59, p=0.05; b=0.01, p=.32; b=−0.01, p=.43). In this model, stool was significant (b=0.52, p=.002), in that infants with stool present had 0.52 higher buttock TEWL relative to chest TEWL than infants without stool present. The application of skin care products to the diapered area was also significant (b=−0.36, p=.03). Infants who had a skin care product had 0.36 lower buttock TEWL relative to chest TEWL than infants who did not have a skin care product. In the GEE model of dichotomized neonatal skin condition score, group, time, and the interaction between these were not significant (OR=0.79, p=.76; OR=1.00, p=.83; OR=1.04, p=.13). Covariates in this model were not significant.

Discussion

In this study, we sought to determine whether changes over time in vital signs or skin health are associated with diapering routines in a 3-hour bundled care event. Stress from the NICU environment and caregiving activities has the potential to impact preterm infant outcomes negatively. When the entire bundled care event was evaluated almost half of the episodes of caregiving resulted in a significant change in HR from the 90-minutes before the start of the event. Consistent with prior research [16–18, 29] a diaper change was associated with stress indicated by a significant increase in heart rate over 74% of the time for infants in both intervention groups. Given the population of infants studied was chronologically and developmentally older preterm infants (PMA >34 weeks) who were born at ≤ 32 weeks gestation, the impact on younger and less stable infants is unclear, but would likely be greater.

There were few significant changes in respiratory rate and none in oxygen saturation associated with the bundled care event as a whole or with a diaper change. Significant increases in infants’ mean heart rate from baseline for the entire bundled care event did not differ between intervention groups (3- vs. 6-hour with a diaper change), nor based upon the presence of a diaper change within the 6-hour group. One factor that may have contributed to the lack of group differences was the lack of a baseline stress measure. While infants served as their own control by determining heart rate instability based upon the 90 minutes prior to beginning bundled care, we did not have a baseline measure of stress. In addition, given each bundled care event included other activities, some of which are more intrusive than diapering (e.g. line placements, gastric tube placement, dressing changes), these activities may have contributed to the observed increase in heart rate during bundled care events without a diaper change. Changes in respiratory rate or oxygen saturation may be more prevalent in other caregiving activities like oral feeding. Additional research should explore the stress associated with caregiving activities based upon their perceived intrusiveness. Knowledge about which caregiving activities are most stressful could inform decisions about when to alter the timing of modifiable intrusive activities. Identification of modifiable activities associated with stress would also allow caregivers to minimize the inclusion of multiple stressful activities in the same bundled care event.

Longer bundled care events and greater numbers of caregiving activities within an event have been thought to be more stressful for preterm infants and could affect vital sign stability. However, in our study the duration of the bundled care event and the number of activities included in the event did not predict heart rate instability. The characteristics, or the level of intrusiveness, of the included activities are likely more important than the absolute number of activities and length of bundled care. For example, bundled caregiving with several intrusive activities (e.g., eye exams diaper changes, oral medications, heel sticks) would likely be more stressful than bundled care without intrusive activities. In addition, the intrusiveness of an activity is not always related to the length of time it contributes to the bundled care event. Consequently, you could have a bundled care event of shorter duration with more intrusive activities than a longer bundled care event, where the level of intrusiveness contributes more to infant stress than the duration of an activity.

Unexpectedly offering a pacifier and swaddling during bundled care did not predict heart rate instability. This is consistent with other research looking specifically at diapering,[29] where the lack of developmental interventions did not predict a change in heart rate during diapering or bundled care. Interventions like offering a pacifier, swaddling, flexed support or a two-person assist paired with a specific activity such as obtaining a blood pressure may have a more significant impact on stress reduction. Only infant sleep-wake state assessed at the beginning of bundled caregiving predicted significant changes in heart rate for bundled care events. Infants who were awake before the bundled care event began were less likely to have a change in heart rate than if they were asleep. While maturation of the autonomic nervous system is thought to be less developed in preterm infants,[30, 31] transitioning to wakefulness from sleeping is associated with increases in heart rates,[30, 32]. Consequently, an increase in heart rate would have already occurred during the 90 minutes before caregiving was initiated. Allowing an infant to self-wake or gently waking an infant prior to caregiving would enable the infant to prepare for engagement. In addition, allowing infants to self-awake before initiation of caregiving, especially for oral feeding is recommended for feeding effectiveness.[33, 34]

While infant sleep-wake state affected heart rate stability for the entire bundled care event, as expected, infant state at the initiation of bundled care did not predict heart rate stability during diapering alone. To understand the impact of sleep-wake state on specific activities like diapering it would be important to know the infant’s sleep-wake state immediately prior to the activity in addition to the state at the beginning of bundled care. Significant changes in heart rate during a diaper change were more likely to occur during longer bundled care events. The significance of this finding is unclear given the length of the bundled care event did not impact heart rate change and the diaper change typically occurred near the beginning of bundled care. The stress of an intrusive event like diaper changes may have greater or less impact depending upon their placement within the event. The inclusion of multiple intrusive activities in bundled care may impact the infants response to individual caregiving activities.[14]

Importantly, less frequent diaper changes in the 6-hour diaper change group did not compromise skin health. There were no group differences in skin pH, TEWL or NSCS of the buttocks. Using the chest as a within subject comparison site, only the presence of stool predicted a higher pH on the buttocks compared to the chest. As would be expected in healthy skin, the pH decreased over time.[35] The pH of the buttocks is likely impacted by exposure of the buttocks to stool and the health of the gut microbiome.[36] The presence of stool in the diaper prior to skin health data collection was associated with increased TEWL on the buttocks. Since TEWL data were collected before cleansing, a higher TEWL may reflect altered skin hydration based on the recent presence of moist stool.

Conclusions

Findings from this study add to our understanding of the relationship between providing developmentally appropriate bundled care and neonatal skin health. Many caregiving activities are routinely performed because they are assumed to be non-intrusive and important to the care of infants in the NICU. Still, it is unclear what activities were part of bundled care that made them more intrusive than the neurobehavioral assessment. Little is known about what activities included in bundled caregiving lead to the greatest amount of stress or how similar and different levels of intrusiveness of activities contribute to cumulative stress experiences of preterm infants over their hospitalization.[8] Future research should examine the intrusiveness of all caregiving activities typically included in bundled care to understand changes in vital sign stability and other indicators of stress. The grouping or clustering of certain care activities that may increase stress should also be explored. The use of developmental care interventions to minimize the impact of stress on infants is appropriate, especially for essential activities that cannot be altered. However, if we could individualize when to provide bundled care and the type and number of activities to include, caregiving stress may be reduced. Currently, given the findings in our study, the frequency of diaper changes when stool is absent should be considered to decrease the stress of routine caregiving and promote infant sleep.