Introduction

About one in ten infants delivered annually in the United States will spend time in a Neonatal Intensive Care Unit [1]. About 50% of NICU infants are born prematurely, i.e., delivery occurring less than 37 weeks gestation, which may result in a variety of medical complications, however, some infants born full-term also may experience medical conditions requiring a NICU stay (e.g., gastroschesis, omphalocele, and seizures) [1]. Common NICU diagnoses include respiratory distress syndrome, chronic lung disease, congenital heart disease, necrotizing enterocolitis (NEC), periventricular leukomalacia (PVL), intraventricular hemorrhage (IVH), pulmonary hypoplasia (PHP), sepsis, transient tachypnea, and infections [1, 2]. Most NICU admissions are singleton births and about 15% are multiples [1]. The average length of stay for a NICU admission is two weeks, but admission stays range from several days to several months depending on the gestational age and medical complications of the infant [1, 3]. Parents typically describe the NICU experience as an emotional “roller coaster” because they experience a variety of emotions, including feeling overwhelmed, sad, stressed, tired, relieved, happy, angry, and helpless [4]. Because of the psychological toll that the NICU experience has on parents, it is essential to understand how infant medical issues during admission and post-discharge influence family outcomes. Given the complexity in co-morbid health conditions among infants, an examination of how NICU infant health is captured in psychosocial research and the implications for parental and child outcomes is warranted.

This paper first addresses variability in NICU infant health and gaps in the current literature. Next, the methods used to conduct this systematic review are described followed by the results of the review. Then, the results of the review are discussed as well as the research and clinical implications.

Variability in infant health

Infants with a NICU hospitalization are a unique pediatric population with a wide range of medical complications, diagnoses, and developmental outcomes. This population includes infants with chronic medical issues and those without any long-term impairment. Efforts to understand the range of infant health issues are timely because advances in perinatal and neonatal care over the last 10–20 years have led to increased rates of survival among infants born at earlier gestational ages and among infants with severe medical complications [5]. However, infants born at earlier gestational age are at increased risk for recurrent hospitalizations, complex medical conditions, and neurodevelopmental impairments [6]. Infants with more medical issues and developmental delays require increased time, attention, and resources from parents which may contribute to higher levels of parental stress and family burden [7]. In fact, infant health severity accounts for a significant portion of variance in family psychological adjustment [8]. As a result, infant medical complications during and after NICU hospitalization are key variables to examine familial adjustment. Improved understanding of how infant medical issues influence infant development, parental mental health, parent–child interactions, and couple and family functioning is needed to identify families and children who require additional support or services during and following NICU hospitalization.

Gaps in current literature

Understanding how infant health severity has been measured in the psychosocial literature is important because infant medical issues greatly impact child development and family outcomes. It is essential to be aware of relevant and validated scoring assessments and the limitations of these assessments. This information may shed light about relationships between infant severity and psychosocial outcomes (e.g., parental stress, mental health, parent–infant attachment, family functioning) and implications for future research and clinical interventions. Although it is clear that more severe medical complications negatively impact parental stress and family burden [7], the impact of specific medical indices of the child (e.g., birth weight, gestational age, medical devices, diagnoses, medical specialists) on infant, parental, and family outcomes remain unclear and warrant attention.

There are several issues that have likely contributed to this gap in the literature. First, the developmental trajectory of NICU infants can change drastically based on physical, developmental, psychological and social factors, suggesting that the change in medical status over time is important to consider. However, there is no gold standard measure to capture NICU infant health severity across time. The majority of scoring assessments measure infant health status only during the first few days following birth rather than infants’ longer term medical trajectory (e.g., Score for Neonatal Acute Physiology (SNAP); Clinical Risk Index for Babies (CRIB)). Further, given the diversity of infant health diagnoses, it is difficult to determine an aggregate scoring system to effectively assess the full range of medical complications. In fact, there is great variation among scoring systems for infant health/medical severity. For example, some psychological researchers measure infant health severity with a single index such as gestational age or birth weight [9,10,11], whereas others ask parents to report severity of infant health on subjective rating scales (e.g., how would you rate your infant’s medical severity) [12,13,14,15]. Research focused on family emotional responses can be enhanced by collaboration with medical professionals. The current validated assessments for infant medical issues typically require indices of infant health that medical staff or charts report (e.g., blood pH, temperature, blood pressure, respiratory rate, oxygenation index, PaO2). Inclusion of medical information about a child who has spent time in the NICU provides a more comprehensive picture of which factors contribute to infant development and psychosocial outcomes than parental self-report alone.

Because NICU infants are a heterogeneous group presenting with a wide range of diagnoses and co-morbid functional impairments, there is relatively limited research about medical fragility at various stages of NICU hospitalization and discharge that may affect later parent–infant relationships and the child’s development [16]. A better understanding of family adjustment to these medical stressors could positively impact infant development and lead to more effective family interventions. The current review highlights and summarizes what has been reported in the literature regarding how NICU infant health severity is captured and related to infant development, as well as parent and family psychological and social adjustment. The current review examined how infant health severity is assessed in extant research; how infant health severity is related to parent, family, and child psychosocial outcomes; and implications of these methods and findings for researchers and clinicians.

Methods of current review

A systematic review was conducted following PRISMA guidelines utilizing PubMED, PsycINFO, and Google Scholar. The following key words were searched in each database: “NICU infant,” “NICU illness severity,” “NICU scoring system,” “NICU developmental outcomes,” “parent adjustment to NICU,” “NICU parent outcomes,” “NICU family.” This review sought to identify empirical investigations that included any assessment of NICU infant health severity and infant, parental, or family psychosocial outcomes.

The literature searches identified 1045 articles. These articles were independently reviewed by two authors (VG and AB). Papers were excluded based on the following criteria: (1) the study focused on a population other than infants admitted to a NICU; (2) the topic did not address psychosocial issues; (3) the primary purpose of the paper was to report on measured development without including a discussion of outcomes; (4) only qualitative data were presented; or (5) assessment of parental or family psychosocial factors or measurement of infant health severity (i.e., birth weight) was not included. The independent reviewers were in complete agreement that: 61 articles were duplicates and 913 met exclusion criteria. Therefore, 70 studies were identified as relevant and included in the current review (see Fig.1).

Fig. 1
figure 1

PRISMA flow diagram

Results of review

Infant health assessments

The 70 empirical research studies, identified separately by two authors, utilized an index of infant health severity in the context of family psychosocial adjustment. The assessments utilized for infant health severity as well as the main findings in these studies were examined and are displayed in Tables 1-7. The most common assessments are reported below.

Table 1 Gestational age and birth weight

Gestational age and birth weight

Infant birth weight and gestational age are included in every article examined in the current review; ten articles included these indices as the only measures of infant health (see Table 1). Gestational age and birth weight are accessible and easy to include because parents can report them. Researchers use infant birth weight and/or gestational age in several ways, as: inclusion criteria, a control variable, or a predictor of outcomes. These indices are essential to include given the impact that low birth weight and earlier gestational age have on neurological development and family adjustment [9, 12, 17]. Although birth weight and gestational age are known risk factors for impaired neurological development, other factors (e.g., parental interactions with infants) also affect development [8].

Nursery neurobiologic risk score (NBRS)

The NBRS [18] was utilized in eight of the articles reviewed (see Table 2). The NBRS requires information from a medical chart and/or medical team. It was developed as a predictor of neurodevelopmental issues in very low birth weight infants and a revised NBRS version was adapted to include seven factors [19]: blood pH, hypoglycemia, intraventricular hemorrhage, periventricular leukomalacia, seizures, infection, and need for mechanical ventilation [20]. Scores greater than 5 are associated with developmental impairments at 24 months corrected age [20]. Research indicates that 80% of 77 infants (12 months of age) with a score of 8 or higher had significant disability [19].

Table 2 Nursery neurobiologic risk score (NBRS)

Perinatal risk inventory (PERI)

The PERI [21] was utilized in four of the reviewed articles (see Table 3). The PERI is an 18-item inventory that requires information from a medical professional. It includes characteristics such as Apgar scores, gestational age and appropriateness of the weight of the infant, congenital infection, seizures, infant’s health growth, nature of electroencephalogram, cranial computed tomographic and ultrasonography findings, sepsis and/or meningitis, duration of ventilation and polycythemia, hypoglycemia, hyperbilirubinemia, and long-term physical disabilities [21]. The items are scored from 0 to 3 using an ordinal scale and then totaled. Higher scores indicate greater neurological impairment. A score of 10 or higher on the PERI is considered at highest risk for neurological impairment (e.g., cerebral palsy) [21]. Moreover, Schemer and Sexton reported that PERI scores of 10 or higher explained 28% of the variance in infant’s intelligence quotient (IQ) or developmental quotient (DQ) scores.

Table 3 Perinatal risk inventory (PERI)

Clinical risk index for babies (CRIB)

One research study used the CRIB (see Table 4) [22]. The CRIB include items such as birth weight, gestational age, congenital malformation, maximum base deficit in first 12 h, sex, and FIO2 in first 12 h. The information for the CRIB is reported by a medical professional or extracted from a medical chart. The CRIB scoring system weights each item based on their statistical relation to likelihood of death. CRIB has mixed results with regard to ability to predict death and/or neurological impairments in infants. Doering and colleagues stated that the CRIB did not predict severe disability at 12 or 24 months of age [20]. Lago and colleagues reported that birth weight alone was similar and gestational age alone was better at predicting neurodevelopmental outcomes for infants, compared to the CRIB [23]. However, studies with smaller sample sizes have suggested that the CRIB has good predictive validity [24].

Table 4 Clinical risk index for babies (CRIB) and score for neonatal acute physiology (SNAP)

Score for neonatal acute physiology (SNAP)

The SNAP (see Table 4) [25]. an alternative to the CRIB, was utilized in four studies. The SNAP is based on 28 items in the first 24 h of life and is weighted based on expert opinion rather than statistics [20]. It requires information from a medical professional and/or medical chart. There are several versions of the SNAP and a few items include birth weight, Apgar score, mean blood pressure, lowest temperature, and seizures. The SNAP predicted infants with poorer motor and psychomotor development [20]; however, it requires a high number of complex indices.

Study specific assessments

Other researchers created their own assessment tools and/or utilized specific individual or multiple indices (e.g., diagnosis, medical devices, length of stay) of infant health beyond gestational age and birth weight. These indices ranged in terms of whether they required information from medical charts, staff, or parents. In this review, 18 of the studies included assessments of infant health severity developed by researchers for their studies (see Table 5). The assessments included a combination of indices and typically utilized: gestational age, birth weight, length of time in the NICU, Apgar scores, medical technology, specific chronic diagnoses (e.g., bronchopulmonary dysplasia, chronic lung disease), and/or neurological injuries (i.e., seizures) [12, 13, 26,27,28]. In seven of the 18 studies, researchers created categories of high and low risk groups of infants (see Table 6). Researchers operationalized high and low risk differently, but examples of high risk infants included those diagnosed with bronchopulmonary dysplasia (or a chronic illness throughout first year of life) and/or very low birth weight (<1500 g at birth), and those who required supplementary oxygen for more than 28 days (or being dependent on medical technology), and/or radiographic evidence of chronic lung disease [12,13,14, 16, 29]. In four of the 18 studies, researchers assigned one point for each health issue (e.g., presence of broncopulmonary dysplasia, respiratory distress, retinopathy of prematurity, apnea, or seizure) to create a severity rating [28, 30]. These aggregate scores were created without regard for variability in severity of each health issue (i.e., each issue equally weighted). Three of research teams developed their own study measures and scoring systems that were not validated [7, 15, 31].

Table 5 Study specific assessments developed by researchers
Table 6 High and/or low risk groups

Twenty-four of the studies in this review used specific indices (e.g., diagnosis, medical devices, rehospitalizations) to define infant health severity (see Table 7). These researchers did not create an overall index score, but used the indices as individual variables. Given the variability among diagnoses, medical devices, and time spent in the NICU, it is reasonable to keep these indices separate and not assume that more time, devices, or diagnoses is necessarily associated with increasing health impairment. These researchers included indices such as: infant sex, gestational age, length of stay, technology dependence (e.g., extracorporeal membrane oxygenation), severe central nervous system injury, congenital or chromosome abnormalities, specific diagnoses (e.g., pulmonary hypoplasia (PHP), necrotizing enterocolitis (NEC)), or Apgar scores [2, 30, 32,33,34,35,36,37,38,39,40,41].

Table 7 Study specific assessments (i.e., Individual or multiple indices of infant health)

Infant health and psychosocial findings

Various attempts to capture infant health severity were found within literature presented in this review, suggesting that this is an important construct to identify and measure. Understanding whether, and to what extent, infant health severity influences family dynamics, parental psychosocial adjustment, and infant development highlights the need to include this construct in research and clinical settings. These findings would assist in identifying families who would benefit from additional support and accommodations during and following NICU hospitalization.

Family impact

Infant health severity and ongoing medical complications greatly impact the family. Family impact was most often defined with the Impact on Family Scale, Revised which conceptualizes it as the perceived burden the child on the family (e.g., My child’s illness is a strain on our family). An aspect of the parent-child relationship, couple, or family functioning was assessed in 21 of the studies. Parents of infants with lower birth weights and increased medical complications report greater family impact [2, 7, 10, 42, 43]. Parents of very low birth weight (VLBW) with a neurodevelopmental impairment, functional disability, or low developmental quotient reported more family impact than developmentally normal VLBW infants [7, 10, 38]. Lower gestational ages, longer hospitalizations, more days on ventilator or oxygen, higher incidence of necrotizing enterocolitis, and longer duration of parenteral feeding were associated with increased impact on families [2, 44]. Overall, families of infants with more medical issues and greater functional impairments report greater family impact.

Family impact also varies as a function of family resources (e.g., money, access to basic needs). Families with less resources report greater stress and family burden [2, 42]. Caring for a medically complex infant adds an extra burden to already stressed parents. Parents must learn to manage their child’s health issues, the cost of care, and medical appointments, in addition to their existing parenting responsibilities. Siblings in the home, higher hospital fees, and increased need for support from the medical team are all associated with increased family stress [2, 44]. Interestingly, Taylor and colleagues reported that although decreased family resources is a risk factor for greater family burden, higher neonatal medical risk did not predict a more negative impact on the family among disadvantaged families compared to more advantaged families [43]. Further, high infant health issues predicted a more negative impact only among advantaged families [43]. Although families with fewer resources report a higher level of family burden, infant medical issues may play a larger role among families with more resources. Not surprisingly, mothers of preterm infants with greater social support, greater sense of family cohesiveness, and less stress reported less decline in family and marital functioning at 6 weeks postpartum [45]. Fathers of preterm infants with higher prenatal family sense of cohesiveness and family and marital functioning also reported better family and marital functioning at 6 weeks postpartum [45]. Increased infant responsiveness is associated with better cohesion and adaptability for the family [46]. Findings highlight how neurobiological, health, psychological, and social variables all play a role in family functioning.

Parental adjustment

Twenty-nine of the studies assessed parental mental health or parenting stress. Overall, parents of infants with lower gestational ages and birth weights reported increased stress, especially within the first three years following discharge, compared to parents of term and normal birth weight infants [14, 32, 47]. Mothers of premature infants reported higher levels of traumatic, depressive, and anxiety symptoms compared with controls [37]. Additional medical complications of a preterm child also influence parental functioning. When parents were asked to identify a primary stressor, a majority chose their infants’ health and physical separation due to hospitalization [48]. Central nervous system complications and infant weight have been associated with maternal anxiety whereas perception of the infant as problematic and infant biological sex impacted paternal anxiety levels [49]. Mothers of extremely low birth weight (ELBW) children with developmental impairments report greater perceived stress compared to mothers of ELBW children with typical development [15]. Longer length of stay in the NICU, extreme prematurity, and a cardiovascular diagnosis were consistent predictors of parenting stress [34]. Surprisingly, one study reported that maternal anxiety was not associated with infant severity of illness [50]. These results suggest that infant medical issues and development explain some portion of the variability in mental health outcomes among NICU parents.

Infant health also has important implications for parental interactions with the infant. Mothers of preterm infants and/or more neurologically impaired infants tend to engage in more intrusive behaviors and report more worries related to their child [51,52,53, 64]. Preterm mothers with high posttraumatic stress symptoms were more likely to engage in a “controlling” dyadic pattern of interaction [54]. Mothers of medically fragile premature infants spent more time interacting, touching, and looking at their infants compared to mothers of non-chronically ill premature infants [16]. Mothers with higher levels of depressive symptoms who had a preterm infant demonstrated a poorer quality of attachment with their infant, controlling for level of neonatal health complications [30]. Findings suggest that the infant gestational age, birth weight, neurological impairment as well as maternal mental health play a critical role in maternal-infant attachment. Notably, 14 attachment studies included in this review exclusively examined maternal-infant attachment and three studies included parental-infant interactions.

Infant development

Medical issues resulting in a NICU admission influence child development. Twenty-one of the studies included in this review assessed infant cognitive, behavioral, or socio-emotional outcomes. Gestational age, birth weight, length of hospitalization, and Apgar score all have been associated with infant cognitive outcomes, irritability, and language development [29, 33, 55]. However, there are several other important factors to consider including socio-demographic variables, parental mental health, and parent–infant interactions. With regard to socio-demographics, male infants are at increased risk for developmental impairment. Preterm birth, infant sex, and parental age accounted for 30–32% of the variance in infant cognitive and motor scores [9]. Further, triplets are at increased risk for negative cognitive outcomes, compared to twins and singletons [56]. Maternal IQ directly has been associated with infant cognitive outcomes and income and the home environment also mediated the relationship [11].

Parental mental health is another key variable that may have an influence on a NICU infant’s development. Zelkowitz and colleagues reported that when controlling for neonatal morbidity and maternal education, maternal anxiety was associated with adverse cognitive development and an increase in internalizing behaviors among children [50, 57]. Maternal depression and worry also have been associated with poorer developmental outcomes and increased irritability in premature infants [14, 36]. Additionally, medical complications, child’s cognitive status, and mothers’ interactive behaviors are significantly associated with low birth weight and children’s social competence [58].

Parent–infant attachment is critical for a child’s development. Poehlmann and colleagues reported that the quality of parent-infant interaction mediated the relation between neonatal risk and cognitive development [30]. Decreased maternal responsiveness to infant vocalization and irritability as well as controlling patterns of attachment were associated with later infant behavioral problems (i.e., eating problems) and social competence [54, 55]. Various combinations of birth status, medical issues, and parent–child relationships, explain about a third of the variance in the development of very low birth weight infants at 1–2 years of age [59]. Findings highlight the powerful role of parent-child interactions, suggesting that the combination of medical complications, parent reactions to stress, and parent–child synchrony are significant for explaining an infant’s cognitive and socio-emotional development.

Discussion

The current review examined how infant health severity is assessed in the literature and reviewed relevant findings regarding associations between infant health severity and psychosocial outcomes. Seventy articles were identified that include an index of infant health severity and relationship to psychological and social outcomes. Some researchers utilize validated measures of infant health [44, 60, 61], whereas others did not [29, 33, 49]. Studies included in this review utilized assessments such as the Nursery Neurobiologic Risk Score (NBRS), Perinatal Risk Inventory (PERI), specific indices of infant health severity (i.e., gestational age, birth weight, length of time in NICU, Apgar scores, medical devices), and/or created a new measure specific to their study. Within this literature, the researchers who created their own scoring systems is concerning given that these scoring systems were not validated and not comparable to existing measures. Finally, infant health can vary over time and these assessments typically captured infant’s health at one point in time. This gap could be addressed by combining infant medical severity at discharge with information from follow-up clinics that utilize developmental assessments (cognitive, social-emotional, motor). For example, the NBRS as neurological status at discharge and information about developmental milestones from Neonatal Follow-Up Clinics could be correlated to determine whether these cognitive, motor, and language assessments can be used to track child health status over time. The variability in the measures utilized for assessing infant health limits generalizability and eliminates the possibility of making comparisons across different studies. Barriers for measuring NICU infant health severity the complex nature of NICU infant health given diverse disorders, the changes in infant health over time, and confounding variables such as experience with the medical team, access to resources, and family psychosocial functioning.

Results of the review indicate that infant health severity has important implications for infant and family outcomes. Parents of infants with lower birth weights and increased medical complications report greater family impact [2, 7, 10, 42, 43]. Lower birth weights and increased functional and/or neurological impairments are associated with increased parenting stress and more intrusive parenting styles [15, 51,52,53]. Infant health severity (e.g., gestational age, birth weight, length of hospitalization), parental depression, anxiety, posttraumatic stress disorder (PTSD), and parent-infant interaction impact an infant’s cognitive and socio-emotional development [14, 29, 33, 36]. It is important to note that factors such as education, income, parental mental health, social support, and parent-child interaction also play important roles.

Research and clinical implications

The NBRS assessment holds the most promise as an assessment tool given its predictive validity and feasibility. The PERI also was effective at capturing an infant’s medical issues; however, it is slightly more intensive and may result in greater participant burden. Both of these assessments require information from the medical chart and/or assistance from a medical professional, which can be a challenge across settings as this information may not be readily available. A more accessible and valid tool that utilizes parental self-report while also capturing the infant health severity across time may be especially useful for researchers. An assessment that includes parental report would allow for the tool to be more accessible to researchers and more easily disseminated across multi-site studies. Parental report does not require that infants be at specific institutions and would allow for Internet-based studies to capture infant health severity across more diverse samples of infants from a variety of locations. Further, developing a composite score that correlates with NBRS, PERI, and infant neurological outcomes would ensure the tool’s validity. It is clear that infant health severity, parent mental health, family resources, parent-infant interactions all influence outcomes, but more research is needed to understand the relative impact of each of these factors and how the medical experience, support from medical staff, and access to care may contribute.

Clinically, the results suggest that the NICU experience and subsequent stressors are crucial elements that influence future family adjustment. Given the fact that both medical issues and psychosocial functioning affect outcomes, it seems optimal for physicians, nurses, psychologists, social workers, and other health care providers to work together in a collaborative care model within the NICU. The growing inclusion of psychologists and other dedicated mental health providers integrated in the NICU is a recent development and these results highlight the critical role of psychological support [62]. While parents are always important to consider in pediatric populations, parents are essential for a child’s prognosis within the NICU population. Psychological support in the NICU setting should be provided routinely for parents, along with continued monitoring following discharge [62, 63]. Integrating mental health services for parents in the NICU, as well as screening measures and the provision of resources following discharge, are pertinent to ensure optimal development for NICU children.

Conclusion

The NICU journey is a stressful experience for families, and infant medical issues are important to understand and predict parental and family adjustment. Current assessment tools addressing infant health severity are varied and generally include information that only medical staff can report. Preterm birth, medical issues, infant sex, family education and income, parent mental health, and parent–child interactions may explain the variability in psychosocial outcomes. Additional work is needed to develop an infant health severity measure that is feasible, valid, and easily disseminated. Researchers should identify and quantify the dynamic relationship among infant health and parental functioning. A multidisciplinary collaborative care model that includes a professional psychological support is recommended within the NICU given the contribution of medical, neurological, psychological, and social variables that impact parent and infant outcomes.