Parent–child interaction after home‐visiting intervention for children born extremely preterm–A randomised clinical trial

Abstract

Aim

To determine whether a strength‐based home‐visiting program enhances parent–child interaction during the first year at home for children born extremely preterm (gestational age < 28).

Methods

A randomised controlled trial of 130 infants born extremely preterm 2018–2021 in Stockholm, allocated to either the Stockholm preterm interaction‐based intervention (SPIBI) (n = 66) or an extended follow‐up program (n = 64). The intervention group received ten home visits during the first year by a trained interventionist following SPIBI guidelines: an interaction‐based intervention supporting sensitive parental responses to infants' signals. The primary outcome of emotional availability was assessed at 12 months corrected age (CA) using the emotional availability scales (EAS).

Results

At 1‐year CA, data were collected from 115/130 (89%) of the included children. There were no significant group differences in emotional availability at 12 months CA. A secondary analysis showed an effect modifier in families with mothers self‐rated as depressed at discharge, with the outcome favouring intervention in the EAS dimension of child involvement.

Conclusion

The SPIBI had no significant main effect on emotional availability at 12 months CA. Children of self‐reported depressed mothers displayed superior involvement behaviour in the intervention group, prompting further research on risk groups and potential modifications of post‐discharge interventions.

Abbreviations

CA

corrected age

EAS

emotional availability scales

EPT

extremely preterm

NICU

neonatal intensive care unit

SPIBI

Stockholm preterm interaction‐based intervention

TOP

Transmurale Ontwikkelingsondersteuning voor Prematuurgeboren kinderen en hun ouders

Key Notes.

• Improved neonatal intensive care targeting extremely preterm‐born infants has not been matched with similar progress of post‐discharge interventions.

• In this randomised clinical trial including 130 children born extremely preterm, no significant differences were found regarding overall parent–child emotional availability at 12 months corrected age between the intervention group receiving an interaction‐based home‐visiting program during the first year at home and controls.

• Children of depressed mothers displayed superior involvement behaviour in the intervention group.

1. INTRODUCTION

Children born before 28 weeks of gestation are classified as extremely preterm (EPT); although neonatal intensive care has improved rapidly during the last two decades, they risk suboptimal development compared to term‐born peers. ¹ With a proactive approach to neonatal intensive care, the EPT survival in Sweden is over 80%, ² but the risk for neurodevelopmental sequelae is high. ³ Disabilities include neurodevelopmental impairments such as executive function deficits, intellectual dysfunction, learning difficulties, autism spectrum disorder, attention deficit hyperactivity disorder and cerebral palsy. ⁴ There is also a high occurrence of special educational needs in students born EPT. ⁵

Families of infants born EPT may spend 4 months in the neonatal intensive care unit (NICU), which may negatively affect parental mental health. ⁶ Poor parental mental health and relationship instability have been associated with developmental challenges for children born preterm. ⁷ Moreover, EPT birth influences relationships between parents and their infants as well as between parents unfavourably. ⁶ , ⁸ Parents may not be psychologically prepared for discharge ⁹ and hence continued support to the EPT‐born infant and family is highly motivated.

The attachment theory was established by Bowlby and Ainsworth. ¹⁰ Attachment security has been associated with positive child development. ¹¹ Concurrent with the attachment theory, Brazelton proved the infant to be an active agent already at birth and these ideas were further adapted to preterm‐born infants by Als. ¹² In their legacy, several early intervention developmental programmes have emerged, ¹³ , ¹⁴ , ¹⁵ focusing on cue‐based responsive interaction. ¹⁶ The aim of early intervention is to prevent developmental difficulties throughout childhood, by assisting parents to support the resilience of their child. Early interventions targeting children born preterm have reported effects on motor function, parental stress and emotional availability, but long‐term effects have been difficult to demonstrate. ¹³ , ¹⁴ , ¹⁵ , ¹⁶

The present study tested if a post‐discharge strength‐ and interaction‐based intervention, during the first year of life, for children born EPT, affected the quality of parent–child interaction at 12 months corrected age (CA) measured with the emotional availability scales (EAS). ¹⁷ The primary research question was if the Stockholm preterm interaction‐based intervention (SPIBI) improved emotional availability at one‐year CA, and hence EAS is the primary outcome of current RCT. We hypothesised higher scores on all six EAS dimensions in the intervention group. A secondary research question was to investigate the extent to which predefined medical, psychological and social variables ¹⁸ may modify the intervention's effect on emotional availability.

2. PATIENTS AND METHODS

2.1. Study design

The SPIBI was a longitudinal randomised controlled trial with consecutive inclusion. Recruitment took place from 20 September 2018 until 8 November 2021, reaching a target of 130 children. Sample size was based on the assumption of moderate effect size of the intervention on EAS. ¹⁸ , ¹⁹ The home visits in the intervention group took place between October 2018–October 2022. Follow‐up visits for filming parent–child interactions at 1 year of age took place between 4 November 2019 and 3 January 2023.

2.2. Participants

This research project is a joint cooperation between Stockholm University and Karolinska Institutet with a recruiting base from the three neonatal wards at the Karolinska University Hospital and one neonatal ward at the Södersjukhuset in Stockholm, Sweden. The eligibility criteria were infants born <28 gestational weeks in any of the aforementioned hospitals with parents speaking Swedish or English. Exclusion criteria were infants with chromosome aberration, life‐threatening malformation or malformation in need of acute surgery and families without Stockholm residency. A flow chart of the study, following consolidated standards, is displayed in Figure 1.

FIGURE 1.

Consolidated standards of reporting trials flow‐chart.

The participants were assigned in a 1:1 ratio to either the SPIBI (n = 66) or treatment as usual with an extended follow‐up program (TAU⁺) (n = 64) through a cluster randomization scheme with blocks of 4–10 participants through the internet‐based random generator randomization.com, administered by a researcher not involved in recruitment. The intervention was completed before the pandemic outbreak in 10 (15%) of the 66 children allocated to the intervention group, 35% had initiated the intervention and 50% were recruited after COVID‐19 was declared a pandemic in March 2020. Hence the majority of participants were offered pandemic adjustments.

2.3. Procedure

NICU staff assessed 164 patients born EPT eligible for inclusion, seven of them were not asked for participation due to ethical reasons, such as uncertainty of survival, non‐permanent Swedish residence or personal parental circumstances. Parents of 157 children were invited to participate by EB or AP; after being thoroughly informed, parents of 130 children from 122 families consented. Four baseline questionnaires concerning discharge mental health were filled out by the included parents: Hospital Anxiety and Depression Scale, ²⁰ Spielberger State–Trait Anxiety Inventory, ²¹ Resilience Scale ²² and Parental Self‐Efficacy scale ²³ (Appendix). Families were thereafter randomised and informed about their allocation by EB. The intervention group received ten home visits by a SPIBI interventionist (Appendix S2); the initial meeting was held in the NICU, if possible. Due to the pandemic, home visits were sometimes replaced by video visits through a secure internet platform. The controls received extended treatment as usual including extra follow‐up visits relating to this study (Figure 2). All participants received a home visit by a physiotherapist at 3 months CA for data collection and a home visit for data collection at 12 months CA (±2 weeks) when the parent–child interaction was videotaped. Participants from both groups followed the general high‐risk follow‐up program. ¹⁸ Recruitment and data collection were identical in both groups.

FIGURE 2.

Study outline.

2.4. The intervention

SPIBI is a strength‐based intervention, intended to consolidate the positive parent–child interplay and facilitate child development through interaction‐based stimulation, described in the protocol. ¹⁸

2.5. Outcome measure

When the child reached 12 months CA, there was a follow‐up home visit. Questionnaires relating to child development and parental mental health were sent to families in advance. All infant parental dyads were filmed for 12 min (± 2 min) during a free play session at home according to standard procedure. ¹⁷ Later an EAS‐trained psychologist, blinded to the allocation, performed the scoring using EAS 4th ed. ¹⁷

2.5.1. Emotional availability scales

The EAS is the primary outcome of the SPIBI trial. The EAS includes six dimensions. Four dimensions are adult‐related: sensitivity, structuring, non‐intrusiveness and non‐hostility and two relate to the child: child responsiveness and child involvement. ¹⁷ , ²⁴ Each dimension is constructed by seven subscales (items) and can assume a value between 7 and 29, where higher scores indicate higher emotional availability. Scores from all six EAS dimensions from the intervention group were compared to the scores from the controls.

2.5.2. Interrater reliability

All 115 films were scored by an EAS‐trained psychologist (VS), blinded to the treatment allocation. In addition, a total of 20% (n = 23) were independently scored by another EAS‐trained psychologist (EB) for interrater reliability analysis. Intraclass correlation coefficients demonstrated very high agreement ²⁵ : sensitivity = 0.98 [95% CI, 0.95–0.99], structuring = 0.95 [95% CI, 0.88–0.98], non‐intrusiveness = 0.97 [95% CI, 0.92–0.99], non‐hostility = 0.94 [95% CI, 0.85–0.98], child responsiveness = 0.95 [95% CI, 0.89–0.98] and child involvement = 0.95 [95% CI, 0.87–0.98].

2.6. Data analysis

Statistical analyses were conducted using SPSS version 29 [IBM Corp, New York, USA]. ²⁶ The relationships among background characteristics and perinatal data were explored using Pearson's Chi‐Square test for categorical versus t‐test for continuous data. Interrater reliability was calculated using intraclass correlation coefficients. ²⁵ The effect of SPIBI on the six EAS dimensions was investigated using Mann–Whitney U tests. All continuous variables were explored through a Pearson correlation matrix. Analysis of covariance for each of the six EAS dimensions, including medical, social and psychological independent variables was conducted. For missing parental baseline values, simple mean imputation of missing single items was used.

2.7. Ethics

Ethical approval was obtained from the Swedish Ethical Review Authority (2017/1596‐31). Participating parents received verbal and written information and written consent was obtained. The full trial protocol was pre‐published ¹⁸ and the study was registered in clinicaltrials.gov (NCT03714633).

3. RESULTS

Concerning the participants, no significant differences were found between the two groups with regard to sociodemographic and perinatal characteristics, except for a significantly larger proportion of posthaemorrhagic ventricular dilatation in the intervention group (Table 1).

TABLE 1.

Sociodemographic and perinatal characteristics of the two groups.

Sociodemographic, psychological and perinatal characteristics
	Total N = 130	Control n = 64	Intervention n = 66	p‐value a
Social background
Maternal age at birth, mean (SD)	33.4 (4.9)	33.0 (4.8)	33.8 (4.9)	0.396
Not speaking Swedish at home, n (%)	6 (5)	1 (2)	3 (5)	0.918
Dual languages at home, n (%)	63 (48)	34 (53)	29 (44)	0.415
Maternal education <12 years, n (%)	17 (13)	10 (16)	7 (11)	0.531
Paternal b education <12 years, n (%)	25 (19)	16 (25)	9 (14)	0.183
Perinatal factors
Sex, male, n (%)	69 (53)	32 (50)	37 (56)	0.489
Cases of twins, n (%)	21 (16)	7 (11)	14 (21)	0.112
Gestational weeks at birth, mean (SD)	25.9 (1.3)	25.9 (1.4)	26.0 (1.3)	0.693
Gestational age < 25 weeks at birth, n (%)	32 (25)	16 (25)	16 (24)	0.920
Birthweight (g), mean (SD)	817 (207)	800 (213)	833 (200)	0.358
Birthweight z‐score < −2, n (%)	30 (23)	15 (23)	15 (23)	0.923
Antenatal steroid use, n (%)	124 (95)	60 (94)	64 (97)	0.166
Apgar‐score at 5 min, mean (SD)	7.1 (2.2)	7.4 (2.1)	6.9 (2.2)	0.195
Days on ventilator, mean (SD)	12.8 (15.6)	12.0 (15.1)	13.5 (16.1)	0.592
First pregnancy, n (%)	71 (55)	38 (59)	33 (50)	0.283
Caesarean delivery, n (%)	82 (63)	41 (64)	41 (62)	0.972
Respiratory distress syndrome, n (%)	106 (82)	52 (81)	54 (82)	0.933
Bronchopulmonary dysplasia, n (%)	46 (35)	21 (33)	25 (38)	0.546
Intraventricular haemorrhage grade 1 or 2/ grade 3 or 4, n (%)	37 (28)/ 7 (5)	19 (30)/ 3 (5)	18 (27)/ 4 (6)	0.911
Posthaemorrhagic ventricular dilatation, n (%)	5 (4)	0 (0)	5 (8)	0.025 c
Retinopathy of prematurity grade 1 or 2/ 3 or 4, n (%)	47 (36)/ 34 (26)	23 (36)/ 16 (25)	24 (36)/ 18 (27)	0.938
Septic bouts before discharge, 1/2/3, n (%)	28 (22)/ 7 (5)/1 (1)	14 (22)/ 3 (5)/1 (2)	14 (21)/ 4 (6)/0 (0)	0.764
Patent ductus arteriosus, n (%)	56 (43)	32 (50)	24 (36)	0.118
Necrotising enterocolitis, n (%)	15 (12)	10 (16)	5 (8)	0.151
At discharge
Oxygen support at 36 postconceptional age, n (%)	58 (45)	26 (41)	32 (48)	0.367
Breastfed children, totally, n (%)/ Breastfed children, partly, n (%)	10 (8)/ 55 (42)	5 (8)/ 25 (39)	5 (8)/ 30 (45)	0.724

^a Pearson chi‐square for categorical data and t‐test for continuous data.

^b Paternal or non‐birthing parent.

^c Significant difference between the groups at a 0.05‐level.

3.1. Primary outcome: Emotional availability of SPIBI versus controls

Post‐intervention data at 12 months CA were available for 115 (89%) parent–infant dyads (intervention group 56 vs. 59 controls). The child was accompanied by the mother in 76 (66%) of the recordings, by the father/non‐birthing parent in 29 (25%) of the recordings and by both parents in 10 (9%) of the recordings. There were no significant differences between the intervention and control groups in any EAS dimension (Table 2). The primary hypothesis was rejected.

TABLE 2.

Means and standard deviation on the six dimensions of EAS for the intervention group and control group.

EAS outcome at 12 months	Intervention SPIBI a mean (SD)	Control mean (SD)	p b	Effect size c η 2
Sensitivity	26.1 (2.4)	25.9 (3.3)	0.831	0.0004
Structuring	26.3 (2.0)	25.8 (3.5)	0.828	0.0004
Non‐intrusiveness	25.4 (2.9)	25.5 (2.9)	0.751	0.0009
Non‐hostility	27.0 (1.9)	27.1 (2.2)	0.387	0.0066
Child responsiveness	25.7 (1.9)	25.5 (3.1)	0.438	0.0053
Child involvement	24.9 (2.4)	24.3 (3.5)	0.818	0.0005

^a SPIBI Stockholm preterm interaction‐based intervention.

^b Mann–Whitney U test.

^c Effect size calculated through Z ²/ (N−1).

3.2. Secondary analyses

To investigate to which extent predefined variables modify the intervention effect on emotional availability, perinatal data were explored with a Pearson correlation matrix. Analysis of covariance including the infant medical, family, social and parental psychological moderators suggested in the pre‐published Theory of Change ¹⁸ was run and refitted to best describe the data. The refitted model presented in Table 3 was significant in all six EAS dimensions. The model with the best fit was one for the EAS dimension of child involvement (Table 3). Included variables were checked for collinearity.

TABLE 3.

Analysis of covariance concerning the dependent variable EAS dimension of child involvement.

Type of data	Source	β	SE	p	Partial η 2
	Corrected model			<0.001**	0.554 a
Intervention	Allocation (intervention/control)	0.259	0.528	0.010*	0.117
Medical	Days on ventilator	−0.039	0.018	0.039*	0.077
	Intraventricular haemorrhage grade 0 versus 1–2 versus 3–4	−2.389/ 0.928	0.897/ 0.586	0.005*	0.179
	Necrotizing enterocolitis	−2.508	1.124	0.034*	0.088
Psychological	Maternal depression (Hospital Anxiety and Depression Scale clinical cut‐off at discharge)	−3.341	0.823	0.012*	0.112
	Maternal Resilience Scale at discharge	−0.050	0.019	0.014*	0.107
	Maternal Spielberger State–Trait Anxiety Inventory Trait	−0.006	0.056	0.916	0.015
	Maternal Spielberger State–Trait Anxiety Inventory State	−0.052	0.057	0.364	0.000
Social	Language spoken at home (Swedish/multilingual)	0.732	0.533	0.175	0.034
Interactional	Interaction allocation × maternal depression (Hospital Anxiety and Depression Scale at discharge)	3.150	1.343	0.023*	0.092

^a R ² = 0.554 (adjusted R ² = 0.463).

^*Significant result at a 0.05 level.

^**Significant at a 0.001 level.

Medical variables significantly contributing to explaining variance in outcome were days on ventilator, grade of intraventricular haemorrhage and occurrence of necrotizing enterocolitis (Appendix [Link], [Link]). Furthermore, maternal baseline data explained more variance than paternal, and the model was adjusted accordingly. The social factor of unilingual or multilingual homes was not significant. From a psychological perspective, maternal self‐rated depression ²⁰ at discharge, defined through Hospital Anxiety and Depression Scale score above clinical depression cut‐off ≥11, was a significant risk factor for low child involvement. However, maternal self‐rated anxiety in the Spielberger State–Trait Anxiety Inventory ²¹ was not a significant risk factor. Maternal resilience was a significant negative factor in the model. A significant effect modifier (F(1, 65) = 5.499, p = 0.023) of maternal self‐assessed depression at discharge and intervention allocation on EAS child involvement showed that children of depressed mothers had significantly higher scores on child involvement in the intervention group than the controls.

Due to some attrition in parental mental health variables, the model was tested also with simple imputation of missing data. Variables with systematic attrition, in particular the Resilience Scale and Parental Self‐Efficacy scale which were only available in Swedish, were omitted from the model. In the refitted model including 112 children, some independent variables became non‐significant but the interaction effect (allocation x maternal depression at discharge) remained significant, with a medium effect size. Children of self‐assessed depressed mothers were assessed as more emotionally available concerning their involvement (M = 25.6, SD = 2.1) if they had received SPIBI, compared with the controls (M = 21.5, SD = 3.8). No difference was found between children in the intervention group (M = 24.7, SD = 2.5) and controls (M = 24.8, SD = 3.3) if the mothers did not self‐assess as depressed at discharge, accounting for the non‐significant main effect of SPIBI. However, when adjusting for multiple outcome measurements using Bonferroni's correction, the level of significance was adjusted to 0.01 and hence the interaction effect was only borderline significant.

4. DISCUSSION

The objective of the present study was to investigate whether SPIBI, a strength‐based home‐visiting program, would enhance parent–child interaction during the first year at home in infants born EPT. Our results showed no significant overall effect of SPIBI on EAS at 12 months CA. However, children of self‐reported depressed mothers displayed more involvement behaviour after intervention as compared to controls. Predefined target inclusion number was calibrated to detect a medium effect size was reached. ¹⁸ There was a larger number of infants with posthaemorrhagic ventricular dilatation in the intervention group, possibly attenuating an intervention effect. However, results were similar when excluding infants with posthaemorrhagic ventricular dilatation.

SPIBI was built on a large body of research indicating that cue‐based interventions ¹⁶ focusing on mother–infant relationship ¹⁵ and offering maternal psychosocial support ¹⁴ may have long‐term positive effects. Despite this, SPIBI failed to show an overall effect on EAS. The study showed that some medical risk factors and maternal psychological characteristics at discharge explain a significant proportion of the EAS outcomes at 12 months CA. This means that the presence of medical risk factors in addition to the EPT birth was of great importance when it comes to the quality of the parent–child interaction. This was in line with what Brisch et al. have reported on attachment quality in infants born prematurely with neurological impairment. ²⁷

The maternal psychological baseline data explained more of the variance in child involvement at 12 months CA than the paternal ditto. This was in keeping with previous reports that preterm birth is associated with maternal depression ⁶ and systematic reviews have stressed the importance of psychosocial support targeting mothers. ¹⁴ The result also confirmed that maternal depression was important when it comes to parent–child interaction in preterm dyads. ²⁸ Maternal depression at discharge also showed an interaction effect with the intervention. Interestingly, Evans and colleagues suggested that maternal psychological symptoms should be included in post‐discharge studies of children born preterm. ¹⁶ Our secondary result suggested that certain subgroups may indeed benefit from SPIBI. This may be of clinical and research interest, especially since parental mental health has been shown to affect the development of premature children. ⁷ Further studies focusing on subgroups of children born EPT with mothers suffering from ill mental health seem warranted and may eventually lead to a tiered intervention approach. Maternal depressive symptoms may affect mother–child interaction more negatively than maternal anxiety since the symptomatology of depression includes sadness, lack of energy and psychomotor inhibition. ²⁸ Maternal resilience was negatively affecting the outcome, which may indicate the need for validated resilience questionnaires particularly targeting emerging motherhood. ²⁹

There has been a paucity of data on EAS in infants born EPT and no reference data of this subgroup are available. In the present study, many dyads had high scores on EAS dimensions ¹⁷ , ²⁴ despite the vulnerability of the population. This could be taken at face value as a positive outcome for our Swedish EPT‐born population, or that prematurity actually does not considerably affect the emotional availability, given that EAS has been reported to have good research properties. ²⁴ Since we believe emotional availability to be the fundamental vehicle for child development, as presented in the pre‐published protocol, the EAS provides important information concerning not only the parents but also the child and the relationship between them. However, the relatively high scores may just as well reflect that EAS is not sufficiently sensitive to detecting the interactional challenges faced by EPT‐born children at 12 months CA and their parents within a naturalistic home environment. Increasing the stress level during the recorded interaction might have given a wider range of ratings, but for obvious ethical reasons, the parent–child dyads should be protected from toxic stress. Access to EAS reference values specifically regarding children born prematurely might improve the validity of the assessment. The need for more diverse empirical EAS data has previously been pointed out. ³⁰ All six EAS dimensions correlated in the current study, which is in line with psychometric evidence for a unidimensional factor structure of EAS. ³⁰

Another reflection is that the high quality of neonatal care in Sweden, as well as generous parental benefits, may paradoxically have been an obstacle to show the effect of further interventions. The high‐quality maternal–child healthcare is included in the welfare system and is freely accessible to all residents. In the NICUs, where recruitment took place, families already receive extensive support in ‘cue‐based parenthood’ through the care inspired by the Newborn Individualised Care and Assessment Program, ³¹ followed by neonatal home care including neonatal nurse home visits and an extensive follow‐up program. Additionally, Sweden offers 480 paid parental leave days per child plus 60 days during the hospital period (a universal social intervention for all parents at the NICU). These provisions and high‐quality services for all families with a child born EPT might make it more difficult to prove the effect of early interventions, compared to settings where the families are not entitled to/cannot afford similar services. In other contexts, research has shown that children born preterm living in poverty have more cognitive and educational difficulties than children born preterm living in economically stable households, ³² underlining the importance of the socioeconomic situation for this group of children.

Some circumstances relating to the COVID‐19 pandemic may have affected the results. Several pandemic adaptions of healthcare were made at this time, leading to increased loneliness and reduced access to care, ³³ which in turn may have contributed to an increased rate of maternal postpartum depressions in general. The pandemic impacted the randomised controlled trial with a slower pace of recruitment (recruitment rate 6.3 vs. 3.7 children/month pre‐ vs. post‐pandemic breakout), 5% drop‐out in the intervention group due to fear of home visits, partial transition to a digital format and parental/societal pandemic concerns. Adjustment to these pandemic‐induced challenges may have affected the outcome negatively, as the intervention was not entirely conveyed as planned during this period. ³⁴

The debate of multiprofessional interventions versus mono‐professional interventions applies to this study. The benefit of a collaborative care chain is clear, ³⁵ and future analyses of collected outcomes will clarify if the multiprofessional approach (Appendix S2) had an effect on the motor, cognitive, parental mental health and/or preschool behaviour (Appendix S1). The SPIBI differs in this respect from the Dutch trial of infant behavioural assessment and intervention, later developed into the TOP intervention, with infant physiotherapists as interventionists and demonstrated effects on motor development. ³⁶

4.1. Strengths and limitations

The strength of this trial is the study group which was powered to detect medium‐sized effects in children born EPT. The randomization worked well, with balanced groups except for a higher incidence of posthaemorrhagic ventricular dilatation in the intervention group. Moreover, it is a strength that socially vulnerable families were included in the trial, reflecting the spread of social circumstances in the Swedish population 63 (48%) of the families were multilingual and 17 (13%) of the mothers and 25 (19%) of the fathers had less than 12 years of education.

A limitation of the trial was that at each home visit, SPIBI targeted the parent currently at home and not specifically the mother or father. This meant that the intervention sometimes could be conveyed to either the mother, father or both parents. Targeting both parents in neonatal interventions is common in Sweden. The EAS recording was done with the parent staying at home that day, which differed between families, and was not necessarily the parent who stayed at home the most during the 12‐month intervention period. When informing parents about the study, we encouraged both parents to learn from the program, and EAS results were the same regardless of which parent was present during the interaction assessment.

The trial may be generalised to the Nordic conditions with similar hospital and social security systems but it may be difficult to apply to completely different societal contexts with limited maternal and paternal parental leave. SPIBI is adapted to meet the needs of children born already at 22 weeks of gestation, which increases its generalizability to a broad population of children born EPT. Large but sparsely populated catchment areas complicate home visits and e‐health adaptation of the intervention may be required. No adverse effects were revealed during the study.

5. CONCLUSION

There was no significant effect of the SPIBI on our primary outcome, i.e. emotional availability at 12 months CA. An interaction effect of maternal self‐assessed clinical depression at discharge and group allocation upon child involvement was found, indicating that the subgroup of children with mothers depressed at discharge benefitted from the SPIBI. This highlights the relevance of including parental mental health challenges in post‐discharge planning and maybe a starting point for future post‐discharge intervention research targeting different subgroups.

AUTHOR CONTRIBUTIONS

Erika Baraldi: Writing – original draft; investigation; conceptualization; methodology; visualization; formal analysis; project administration; data curation. Mara Westling Allodi: Conceptualization; investigation; funding acquisition; methodology; validation; writing – review and editing; project administration; data curation; supervision; resources. Kristina Löwing: Resources; methodology; validation; writing – review and editing; funding acquisition; investigation; conceptualization. Noni Wadström: Writing – original draft; formal analysis. Ann‐Charlotte Smedler: Conceptualization; investigation; funding acquisition; methodology; validation; writing – review and editing; project administration; supervision; resources. Maria Örtqvist: Conceptualization; funding acquisition; writing – review and editing; methodology; resources. Björn Westrup: Conceptualization; investigation; writing – review and editing. Ulrika Ådén: Conceptualization; investigation; funding acquisition; methodology; validation; writing – review and editing; project administration; data curation; resources.

FUNDING INFORMATION

SPIBI was funded by the Stockholm university faculty funds (2016–2024), Region Stockholm in collaboration with Stockholm University (2016 and 2022), Region Stockholm clinical research appointment, Clas Groschinsky's Memorial Fund, Swedish state under the agreement between the Swedish government and the county councils (the ALF‐agreement), Philipson Foundation, Swedish Brain Foundation, Vetenskapsrådet, Frimurarorden Barnhuset, Queen Silvia Jubilee Fund for research on children with disability, Sällskapet Barnavård, H.K.H Kronprinsessan Lovisas förening för barnsjukvård, The Sven Jerrring Foundation, Promobilia Foundation, The Foundation of Sunnerdahls Disability Fund, Centrum för kompetensutveckling inom vård och omsorg and The Filéenska Fund.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest.

Supporting information

Appendix S1:

Appendix S2:

Appendix S3:

Baraldi E, Allodi MW, Löwing K, Wadström N, Smedler A‐C, Örtqvist M, et al. Parent–child interaction after home‐visiting intervention for children born extremely preterm–A randomised clinical trial. Acta Paediatr. 2025;114:74–82. 10.1111/apa.17399