OBJECTIVE: Parental involvement in the care of preterm infants in NICUs is becoming increasingly common, but little is known about its effect on infants' length of hospital stay and infant morbidity. Our goal was to evaluate the effect of a new model of family care (FC) in a level 2 NICU, where parents could stay 24 hours/day from admission to discharge.
METHODS: A randomized, controlled trial was conducted in 2 NICUs (both level 2), including a standard care (SC) ward and an FC ward, where parents could stay from infant admission to discharge. In total, 366 infants born before 37 weeks of gestation were randomly assigned to FC or SC on admission. The primary outcome was total length of hospital stay, and the secondary outcome was short-term infant morbidity. The analyses were adjusted for maternal ethnic background, gestational age, and hospital site.
RESULTS: Total length of hospital stay was reduced by 5.3 days: from a mean of 32.8 days (95% confidence interval [CI]: 29.6–35.9) in SC to 27.4 days (95% CI: 23.2–31.7) in FC (P = .05). This difference was mainly related to the period of intensive care. No statistical differences were observed in infant morbidity, except for a reduced risk of moderate-to-severe bronchopulmonary dysplasia: 1.6% in the FC group compared with 6.0% in the SC group (adjusted odds ratio: 0.18 [95% CI: 0.04–0.8]).
CONCLUSIONS: Providing facilities for parents to stay in the neonatal unit from admission to discharge may reduce the total length of stay for infants born prematurely. The reduced risk of moderate-to-severe bronchopulmonary dysplasia needs additional investigation.
WHAT'S KNOWN ON THIS SUBJECT:
Although advances in technology and medical treatment have allowed more infants to survive, morbidity remains high. The NICU environment and early parent-infant interaction have been associated with infant health and length of hospital stay.
WHAT THIS STUDY ADDS:
Data from this study indicate that parents staying in the NICU from admission to discharge may reduce the total length of stay for infants born prematurely. An individual-room NICU design could have a direct effect on infant stability and morbidity.
Today's neonatal intensive care environments reflect the advances in technology and medical treatment of preterm and sick newborns. For these infants, the environment may be very stressful, with long periods of light, unnatural noise, and repeated disturbances from caretaking procedures. Separation from the mother may cause stress for the infant1 as well as the mother.2 It has been suggested that a humane neonatal care initiative should be launched that focuses on the psychological needs of the sick and preterm newborn, and where 1 of the steps consists of giving the mother the opportunity to stay in the hospital with her infant 24 hours a day.3
Even if parental involvement in the NICU is now high on the international research agenda, most parents are still separated from their infants during the major part of the hospital stay. A survey of NICUs in 8 European countries conducted in 2004–2005 revealed that parents could visit their newborn at any time over the 24 hours in units located in the northwestern part of Europe, but not in Spain and Italy.4 In Turkey, 1 NICU had 4 individual rooms where mothers could stay with their preterm infant if the infant did not require mechanical ventilation or close monitoring. This model was associated with reduction of rehospitalization after discharge, as well as postnatal acute care visits.5 A NICU in Pakistan involved mothers in all basic nursing care of their prematurely born infants before discharge, and this practice significantly reduced lengths of hospital stay without raising readmission rates.6
Our aim was to evaluate the effect of this new model of family ward. The primary outcome was total length of hospital stay and the secondary outcome was short-term infant morbidity.
The study was conducted as a randomized, controlled trial, comparing outcomes of prematurely born infants randomly assigned to the family wards in 2 NICUs (both level 2) in Stockholm with those of infants randomly assigned to standard NICU care in the 2 hospitals. Recruitment to the study took place between September 2006 and March 2008.
One NICU was located at Sachs' Children's Hospital and the other at the Karolinska University Hospital Danderyd, both serving large maternity units with 6700 and 9000 deliveries per year, respectively. Both NICUs could provide nasal continuous positive airway pressure (CPAP), but infants born before 27 weeks of gestation, or those who were in need of mechanical ventilation for more than a few hours, were referred to the level 3 unit at the Karolinska University Hospital. The 2 study NICUs had both recently increased their capacity by opening a family oriented ward with 13 and 10 infant beds, respectively, where there were separate rooms for all families, including beds for both parents, a private bathroom, a bed for the infant, and equipment for supplementary oxygen and telemetric wireless monitoring. Each family ward had a 4-bed intensive care room, with staff continuously present. From admission, at least 1 of the parents was expected to stay 24 hours a day during the entire hospital stay. Infants in the intensive care rooms moved into the family rooms as soon as they reached a stable state: for the very preterm infants this was usually around 32 to 33 weeks' postmenstrual age.
The standard care (SC) wards in the 2 NICUs had 15 and 14 infant beds, respectively, with a 4-bed intensive care room in each ward and intermediate care rooms with 2 to 4 beds. There were a couple of rooms for parents but the opportunity to stay overnight was limited to a few days before discharge, provided a room was available. During the daytime, the parents were advised to be with their infant as much as possible.
In each hospital, medical guidelines and nursing practices were the same in the family ward as in the SC ward.
Infants in the study were admitted from the delivery ward of the respective hospital or from the level 3 NICU at the Karolinska University Hospital. Inclusion criteria were preterm birth (<37 weeks), no major congenital malformation diagnosed on admission, and a need for in-hospital care in a level 2 NICU. The infants were allocated to the family care (FC) ward or SC ward on admission to the respective NICU. Before the trial, the placement of new infants was primarily dependent on where an empty bed was available. During the trial period, an additional inclusion criterion was 1 available bed space in both the FC and SC ward, to allow random allocation.
Randomization was undertaken immediately on admission by sealed, opaque, and numbered envelopes in blocks of 6 (3 with FC and 3 with SC). When an infant met the inclusion criteria, a staff nurse opened the next envelope in the box. At this time point it was not considered ethical to give information and ask parents for consent to participate in the study because many of them were still overwhelmed by worry about their infant. The information was thus postponed to a more appropriate time within the first 3 days after admission. Altogether, consent was obtained for 329 infants. In 17 cases, the parents were not approached because the infant was transferred out of the NICU soon after the birth, and in 20 cases the parents did not consent to fill in questionnaires (not analyzed for the purpose of this article). However, in this study, all infants who were randomly assigned were included because ethical approval was obtained to collect information about the defined outcomes (ethical regional board in Stockholm, approval numbers 2006/532-31/3, 2008/1392-32).
Power calculation was on the basis of the assumption that FC would reduce the total length of hospital stay, and that a difference of 5 days would be of clinical importance. This would require a total sample of 314 infants (P < .05; 80% power).
Data Collection and Definitions
Data were extracted from hospital records and the National Perinatal Quality Register7 during a period ranging from birth to hospital discharge. In cases of extremely long hospital stay, a limit was set at 180 days after birth. Gestational age at birth was defined as the number of completed weeks and days estimated by first trimester ultrasound. The discharge criteria did not include any weight limits, but the infants had to be clinically well. Gavage feeding did not exclude from being discharged, but these infants became part of a domiciliary care program, lasting until they were able to breastfeed or bottle feed with satisfactory weight gain.
Total length of stay in the hospital was defined as number of days from birth to discharge. Length of stay in intensive care was also analyzed, both the number of days in the level 3 NICU at the Karolinska University Hospital Solna and the number of days in the intensive care rooms of the level 2 NICUs in the 2 study hospitals.
Ventilatory assistance included respiratory support, either by mechanical ventilation or CPAP, and oxygen supplementation.
Severity of illness was measured by the Clinical Risk Index for Babies score.8 Small for gestational age was defined as a birth weight less than −2 SD according to reference data for normal fetal growth in a Swedish population.9 Intraventricular hemorrhage (IVH) was classified as grade 1 to 4.10 Patent ductus arteriosus (PDA) was diagnosed by using echocardiography in combination with clinical symptoms. Sepsis was defined as a positive blood culture in the presence of clinical signs,11 and necrotizing enterocolitis (NEC), according to Bell et al.12 Retinopathy of prematurity (ROP) was classified in 5 stages.13 Bronchopulmonary dysplasia (BPD) was diagnosed on the basis of the need for supplemental oxygen at 28 days postnatally. The severity of BPD was determined at 36 weeks and classified as mild (no supplementary oxygen), moderate (<30% oxygen), or severe (≥30% supplemental oxygen and/or CPAP or ventilator).14
Data were analyzed by using SPSS 17.0 for Windows (SPSS Inc, Chicago, IL). Participants from the 2 hospitals were merged and all infants allocated to FC were compared with infants allocated to SC, regardless of hospital location. Analyses were conducted according to the principle of “intention-to-treat” for each randomized group regardless of postrandomization transfers. Analyses were also conducted according to gestational age, with groups defined at <30 weeks, 30 to 34 weeks, and 35 to 36 weeks. Subgroup analyses were conducted in which we excluded infants who were given a diagnosis of a congenital disease after they had been included in the study.
Differences between FC and SC in length of stay were calculated by using analysis of variance and the Mann-Whitney U test, and differences in infant morbidity were calculated by using odds ratios (ORs) with 95% confidence intervals (CIs). Analyses of the complete FC and SC groups were adjusted for maternal non–Swedish-speaking background, gestational age, and hospital site (Sachs' Children's Hospital or Karolinska University Hospital Danderyd). Analyses by gestation were adjusted for maternal non–Swedish-speaking background and hospital site.
Of the 652 preterm infants admitted to the 2 NICUs during the recruitment period, 366 met the inclusion criteria and were randomly assigned to FC and SC, respectively (Fig 1). The exclusions were explained by lack of an available bed in 1 of the models, which made random allocation impossible.
Of the 366 infants randomly assigned into the study, 2 in the FC group and 7 in the SC group were transferred to a level 3 NICU within 1 to 2 days after randomization. One more infant in SC was transferred after 4 months. The reason for the early transfers in the FC group was respiratory distress (n = 2); in the SC group the reasons were heart surgery (n = 1, and 1 twin sibling), intestinal surgery (n = 1), congenital liver disease (n = 1), myotubular myopathy (n = 1), and respiratory distress (n = 2). The late transfer in the SC group was caused by severe BPD. The infant with myotubular myopathy was still in the hospital 6 months after birth, and 1 infant with respiratory distress in SC died.
Of the 356 infants who continued in their allocated model of care, 4 in FC were later transferred to SC and 7 in SC to FC (Fig 1). The reasons for the crossover from FC to SC were as follows: a family with twins could not cope (n = 2); maternal medical reason (n = 1); and a family moved to a hospital closer to home in another city (n = 1). The reasons for crossover from SC to FC were as follows: admission to the wrong ward by mistake after surgery at a level 3 NICU (n = 1); medical indication to improve infant stability by creating a quieter environment (n = 1); and to stimulate mother-infant contact (n = 5). All these decisions were made by the attending neonatologist. In sum, 345 infants were exposed to the allocated model of care: 177 (97%) in FC and 168 (92%) in SC.
Infant and maternal characteristics did not differ statistically between the FC and SC groups except for a higher median postconceptional age at randomization in the 30 to 34 group, and a lower proportion of women with a non–Swedish-speaking background in the FC group (Table 1). A comparison between infants exposed to FC and SC gave almost identical results.
Total length of hospital stay was a mean of 5.3 days shorter in the FC group (mean: 27.4 days compared with 32.8 days in the SC group [P = .05]) (Table 2). The median difference, which was not adjusted for possible confounders, was −2 days (P = .25). The reduction of total length of stay was mainly explained by differences between the most preterm infants: mean −10.1 days (P = .26) and median −16 days (P = .02). Table 2 also shows that the smaller number of days in the hospital in the FC group was explained by a shorter stay in intensive care: mean −4.7 days (P = .02) and median −3 days (P = .02). Table 3 provides details of care in a level 3 NICU, such as admission directly from the delivery ward or later, and length of stay. The total length of stay was also calculated after excluding infants with congenital diseases diagnosed after randomization (2 infants in the FC group: Down syndrome, nonketotic hyperglycinemia; and 5 in the SC group: Williams syndrome, Prader-Willi syndrome, myotubular myopathy, congenital pulmonary stenosis, and congenital liver disease), and was 27.3 days in the FC group and 31.8 days in the SC group, respectively (P = .05).
The proportion of infants given respiratory support was 49% in the FC group and 60% in SC (complete groups: P = .04; when late diagnosed congenital diseases were excluded: P = .06). However, after adjusting for maternal non–Swedish-speaking background, gestational age, and hospital site, the mean duration of treatment days in the respective group was 0.6 (95% CI: 0.4–1.7) vs 1.3 (95% CI: 0.6–2.1; P = .29) for mechanical ventilation; 6.5 (95% CI: 3.2–9.8) vs 8.7 (95% CI: 6.2–11.2; P = .30) for nasal CPAP; and 11.0 (95% CI: 6.6–15.4) vs 12.2 (95% CI: 9.0–15.6; P = .63) for supplemental oxygen.
Sixty-eight percent of infants in the FC group and 72% in the SC group attained their birth weight before discharge (P = .34). Table 4 shows no statistically significant difference in infant morbidity, except for a lower occurrence of moderate-to-severe BPD in FC: 1.6% vs 6.0% in SC (adjusted OR: 0.18 [95% CI: 0.04–0.8]) (Table 4). These findings did not change when late detected congenital diseases were excluded.
We found a reduction of 5 days of total hospital stay in preterm neonates admitted to a level 2 NICU where parents could stay from admission to discharge, compared with standard NICU care. This finding should be interpreted in light of the definition of our sample, where also infants born at 35 to 36 weeks' gestation were included, 29% of the FC and 26% of the SC. The mean length of stay was only 6 and 8 days, respectively, and these infants were, therefore, less exposed to the different models of care. Our findings suggest that FC in NICU has a stronger impact on length of stay in the more preterm infants, and during the initial period of care. Analysis by nonparametric methods showed a more modest reduction of total length of stay by a median of 2 days, but a larger reduction of 16 days in the most preterm infants of <30 weeks' gestation. These estimates, however, were not adjusted.
Parents who spend most of their time with their newborn may have a greater opportunity to interpret and act on signs of distress and other needs of the infant compared with what is possible for the NICU staff, which has more than 1 infant to care for. In the FC wards, the parents quickly became the primary caregivers of their infants, and the greater continuity could possibly contribute to more individualized care. Research on individualized care interventions have revealed positive effects on infant stress and sleep,15,16 and also on length of hospital stay in very preterm infants.17,18 The most comprehensive program of this kind, the Newborn Individualized Developmental Care and Assessment Program, includes a range of different measures to decrease external stimuli and support comfortable positioning, but also to increase the involvement of parents.19,20 The amount of time parents were involved with their newborns was not presented in those studies, but it was definitely less than in FC in our study. On the basis of our findings, one may speculate whether providing facilities for parents to stay with their infants throughout the entire hospitalization might further enhance the effect of developmental supportive care.
The lower prevalence of BPD in the FC group suggests that parental involvement could have a direct effect on infant stability and morbidity. Mechanical ventilation was less common in the FC group, although not significantly, and it cannot be excluded that this contributed to the lower rate of BPD in this group because recent data suggest that less aggressive respiratory treatment in preterm infants may reduce the risk of BPD.21 Reports of lower incidence of pulmonary morbidity by the Newborn Individualized Developmental Care and Assessment Program18,22 support the interpretation that nonmedical interventions might reduce the incidence of BPD.
A limitation of our study was that the number of hours that 1 or both parents spent with their newborn during the intensive care period was not specifically recorded. However, our observation was that at least 1 of the parents in the FC group was present on the ward most of the time, except for a maximum of a few nights. The presence of parents in the SC group was limited by the fact that they usually could not stay overnight in the hospital, neither during the period of intensive care nor during rest of the stay. We did not find any evidence suggesting that SC infants were sicker, despite a slightly larger number of infants admitted to a level 3 NICU from the delivery ward.
The context of NICU care varies widely between countries. In some countries, where most mothers are housewives and members of their extended family can take care of siblings at home, a mother may be able to stay with her prematurely born infant in the NICU, as described in the publication from Pakistan.6 In a country like Sweden, where most women participate in the labor market, parental presence in the NICU is facilitated by a parental allowance regulation that entitles both parents to financial parental support when the newborn is seriously sick, corresponding to 80% of the yearly income.23 This support makes it possible for parents to share the time spent in the hospital, and at the same time take care of siblings at home. In other countries, potential cost savings associated with a reduced length of stay in the NICU could be balanced against the costs associated with parental financial support similar to the Swedish system.
This study demonstrated a reduction of total length of hospital stay for infants born prematurely by providing facilities for parents to stay in the NICU 24 hours/day from admission to discharge. Analyses of secondary outcomes also suggested a reduction in pulmonary morbidity, such as moderate-to-severe BPD. However, this finding needs additional investigation.
This study was supported by the Vårdal Foundation and Stiftelsen Samariten.
We thank Hans Pettersson at the Department of Clinical Science and Education at Södersjukhuset for valuable statistical advice. We are grateful to Erik Borgström, Evalotte Mörelius, and Birgitta Lundin for assistance with data gathering and input at various times during the process.
- Accepted September 9, 2009.
- Address correspondence to Annica Örtenstrand, RN, PhD, Sachs Children's Hospital, Södersjukhuset, SE-118 83 Stockholm, Sweden. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- CPAP =
- continuous positive airway pressure •
- SC =
- standard care •
- FC =
- family care •
- IVH =
- intraventricular hemorrhage •
- PDA =
- patent ductus arteriosus •
- NEC =
- necrotizing enterocolitis •
- ROP =
- retinopathy of prematurity •
- BPD =
- broncopulmonary dysplasia •
- OR =
- odds ratio •
- CI =
- confidence interval
- Bhutta ZA,
- Kahn I,
- Salat S,
- et al
- 7.↵National Register, PNQ. PeriNatal Quality Register. 2007. Available at: www.medscinet.com/pnq. Accessed May 30, 2009
- Als H,
- Gilkerson L,
- Duffy FH,
- et al
- Als H
- Westrup B,
- Kleberg A,
- von Eichwald K,
- et al
- 23.↵Government Offices of Sweden. Social insurance. Stockholm, Sweden: Ministry of Health and Social Affairs; 2007; fact sheet no. 8. Available at: www.regeringen.se/sb/d/574a/79449. Accessed December 4, 2009
- Copyright © 2010 by the American Academy of Pediatrics