You are here

  • Home
  • Online First
  • Very preterm/very low birthweight infants’ attachment: infant and maternal characteristics

Article Text

Article menu

other Versions

Download PDFPDF

Original article
Very preterm/very low birthweight infants’ attachment: infant and maternal characteristics
  1. Dieter Wolke1,2,
  2. Suna Eryigit-Madzwamuse1,
  3. Tina Gutbrod3
  1. 1Department of Psychology, Lifespan Health and Wellbeing Group, University of Warwick, Coventry, UK
  2. 2Division of Mental Health and Wellbeing, Warwick Medical School, Coventry, UK
  3. 3Department of Psychology, University of Hertfordshire, Hatfield, UK
  1. Correspondence to Professor Dieter Wolke, Department of Psychology, Lifespan Health and Wellbeing Group and Division of Mental Health and Wellbeing, Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK; D.Wolke{at}warwick.ac.uk

Abstract

Objective To investigate whether there are differences in attachment security and disorganisation between very preterm or very low birthweight (VP/VLBW) (<32 weeks gestation or <1500 g birthweight) and full-term infants (37–42 weeks gestation) and whether the pathways to disorganised attachment differ between VP/VLBW and full-term infants.

Design The sample with complete longitudinal data consisted of 71 VP/VLBW and 105 full-term children and their mothers matched for twin status, maternal age, income and maternal education. Infant attachment was assessed with the Strange Situation Assessment at 18 months of age. Maternal sensitivity in the VP/VLBW and full-term samples was rated by neonatal nurses and community midwives in the neonatal period, respectively, and mother-infant interaction was observed at 3 months. Infant difficultness was assessed by maternal report at 3 months and infant's developmental status was assessed with the Bayley Scales (BSID-II).

Results Most VP/VLBW (61%) and full-term (72%) children were found to be securely attached. However, more VP/VLBW (32%) than full-term children (17%) had disorganised attachment. Longitudinal path analysis found that maternal sensitivity was predictive of attachment disorganisation in full-term children. In contrast, infant's distressing cry and infant's developmental delay, but not maternal sensitivity, were predictive of disorganised attachment in VP/VLBW children.

Conclusions A third of VP/VLBW children showed disorganised attachment. Underlying neurodevelopmental problems associated with VP/VLBW birth appear to be a common pathway to a range of social relationship problems in this group. Clinicians should be aware that disorganised attachment and relationship problems in VP/VLBW infants are frequent despite sensitive parenting.

  • Child Psychology
  • Neurodevelopment

https://doi.org/10.1136/archdischild-2013-303788

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    What is already known on this topic

    • Very preterm or very low birthweight (VP/VLBW) infants are at risk for a range of developmental impairments.

    • Neurological abnormalities might lead to disorganised attachment.

    What this study adds

    • VP/VLBW infants are at higher risk for disorganised attachment.

    • Disorganised attachment in VP/VLBW infants may occur despite sensitive parenting.

    • Neurological impairment (indicated by distressing cry and developmental delay) predicts VP/VLBW infants’ attachment disorganisation.

    Introduction

    Infant-caregiver attachment refers to the emotional bond the infant forms with consistent caregivers who are sensitive and responsive in social interactions.1–3 Securely attached children use attachment figures as a secure base to explore from and return to. Insecurely attached infants either avoid their caregiver or fail to act independently in stressful situations. Secure, avoidant and ambivalent attachment are organised strategies that are adaptive to the infants’ environment.4 In contrast, some infants show contradictory or misdirected behaviours (ie, a breakdown of organised attachment behaviours).5 ,6 Attachment disorganisation predicts psychopathology and poor peer relationships in later childhood.7 ,8

    Very preterm or very low birthweight (VP/VLBW) infants are at increased risk for a range of developmental impairments.9–11 This has been linked to a predisposition to disorganisation.12 Indeed, a meta-analysis4 reported that 35% of infants with neurological abnormalities were classified as disorganised compared to 15% of infants in the general population. Surprisingly, most studies found no differences in secure-insecure attachment classifications compared to full-term infants at 12–18 months.3 ,13–16 However, these were mostly small sample studies of low risk infants.10 Furthermore, a lack of association between maternal sensitivity and infant attachment has been found in some13 ,14 but not in other preterm studies.17 ,18 Overall, there remains uncertainty whether VP/VLBW children are at higher risk for insecure and, in particular, disorganised attachment and whether this can be accounted for by parenting or underlying neurodevelopmental problems. This prospective study investigated, first, whether VP/VLBW children are at higher risk for insecure or disorganised attachment at 18 months of age compared to a full-term control group. Second, we compared alternative models of attachment formation between VP/VLBW and full-term samples and hypothesised that infant factors are more important drivers of attachment status than maternal factors in the VP/VLBW sample.

    Methods

    Participants

    Infants who were born VP/VLBW (<1500 g or <32 weeks of gestation) and admitted to one of three Neonatal Intensive Care Units (NICUs) between January 1998 and July 1999 in southeast England and remained in the NICUs were potential participants (N=270). Fifty-six infants died. Of the 214 survivors, 112 were randomly approached and parents of 90 infants consented to participate in the study (76 mothers). They spent an average of 58.1 days (SD=26.3) in the hospital, 3.8 days (SD=7.5) on ventilation, 8.2 days (SD=15.3) using continuous positive airway pressure, received 15.3 days (SD=13.1) of total parenteral nutrition and 37.7 days (SD=18.9) of nasogastric tube feeding; 24% of infants were small for gestational age (<10th percentile birthweight) and 17% had brain abnormality (ventricular enlargement, cysts) on ultrasound.

    Full-term infants (37–42 weeks gestation) from the same maternity units were approached in order to match to the VP/VLBW sample characteristics (N=116). The parents of 115 infants consented to participate in the study (98 mothers).

    Sample matching. Full-term and VP/VLBW samples were matched for infant sex, twin status, maternal age, income, maternal education and relationship status (table 1). More VP/VLBW infants (N=19) than full-term infants (N=11) dropped out of the study (χ2=4.59, p<0.05). The mothers who dropped out were significantly younger (t(203)=2.01, p=0.05), the infants weighed less (t(201)=2.44, p<0.05) and were born earlier (t(203)=2.02, p<0.05) than those whose families remained in the study.

    View this table:
    Table 1

    Infant and maternal sample characteristics

    Measures

    We identified cry duration, distressing cry, developmental delay and twin status as relevant infant factors, and maternal sensitivity at term and at 3 months as the relevant maternal factor. Assessment ages were corrected for prematurity.

    Cry duration at 3 months was measured using infants’ average daily cry.19 Mothers reported how long their baby cries during an average day. Median cry duration for VP/VLBW and full-term infants was found to be 85 min (range=10–440 min) and 60 min (range=3–360 min), respectively. VP/VLBW infants cried significantly longer than their full-term counterparts (H(1)=153.452, p<0.001).

    Distressing cry at 3 months was measured using mothers’ perception of cry.20 Mothers rated the sound of their babies’ crying on a ‘1 (not at all)’ to ‘9 (very much)’ point scale for 15 different aspects of cry sound, which were frantic, highly pitched, loud, angry, pained, ear-piercing, aggressive, worried, unpleasant, desperate, sad, heartbreaking, spoilt, whimper and whiny. Factor analysis yielded a single factor for both VP/VLBW and full-term samples excluding three items (ie, spoilt, whimper and whiny). The resulting factor explained 62% of total variance and was found to be reliable (Cronbach's α=0.82 for VP/VLBW and 0.86 for full-term samples). Mean comparisons showed no significant difference between VP/VLBW (mean=3.11, SD=1.37) and full-term infants (mean=3.01, SD=1.36).

    Developmental delay at 18 months was assessed with the Bayley Scales of Infant Development (BSID II21).i Examiners were trained to 95% agreement. The MDI and PDI were significantly correlated (r=0.70, p<0.001), loaded onto a single factor (explained variance: 81%) and total scores were thus averaged and reversed to indicate developmental delay in the path analysis.

    Twin status was noted at birth.

    Maternal sensitivity at term was measured using the Boston City Hospital Assessment of Parental Sensitivity (BCHAPS),22 a 12-item questionnaire given to nurses to assess how well the mother cares for, interacts with, understands and enjoys her infant. It is scored on 5-point scales from ‘poor’ to ‘very competent’. Neonatal nurses who cared for the VP/VLBW infants in the NICU covering an average of 47 days or community midwives who visited the mothers on five occasions in the first 10 days following discharge from hospital in the case of full-term infants completed the BCHAPS. Factor analysis yielded one factor which explained 74.4% of the total variance and was found to be internally consistent (Cronbach's α=0.95 for both samples).

    Maternal sensitivity at 3 months was measured with the Mother-Infant Structured Play Assessment. During the assessment, babies lay supine on a blanket and mothers sat on the floor opposite their infants. Mothers were first asked to play freely with their infants for a 2 min period, and then they were asked to get their babies to look at them for an additional minute. Maternal behaviour was rated using 5-point scales of maternal positive emotional expression, sensitivity and stimulation level adapted from the Emotional Availability Scales23 and Infant and Caregiver Engagement Phases24 by trained researchers unaware of group membership. The inter-rater reliability (κ; N=20) was high for each scale (κPositiveEmotion=0.76, κSensitivity=0.76, κStimulationLevel=0.78); the composite of the scales had acceptable internal consistency (Cronbach's αPhase 1=0.72; Cronbach's αPhase 2=0.73), correlated across the two episodes (r=0.58, p<0.001) and loaded onto a single sensitivity factor, which explained 79% of the variance.

    Attachment security/disorganisation at 18 months were assessed using the Strange Situation Assessment (SSA).25 Infants are challenged through a series of separations and reunions with the caregiver, which reveal the infants’ reactions to being left alone, being left in the presence of a stranger and being reunited with their caregiver. Mother and infant behaviour were recorded during seven 3 min episodes.ii First, the A, B, C and D division was used to classify secure (B) versus insecure (A and C) and organised versus disorganised (D) attachment. Attachment disorganisation scores were calculated according to Main and Solomon's5 continuous scale of attachment disorganisation on a 9-point scale, where ‘1’ represented no signs of disorganisation and ‘9’ represented definite signs of attachment disorganisation. For organised versus disorganised attachment categories, those scoring ≥6 were classified as disorganised4; those scoring 5 were given either a primary or a secondary disorganised classification depending on the particular case; and those scoring <5 were classified as having organised attachment.

    Statistical analysis

    Sample characteristics were compared between the VP/VLBW and full-term subsamples using one-way ANOVA for mean comparisons, the Kruskal-Wallis test for variables that were not normally distributed and χ2 analysis for nominal variables. We also compared infant and maternal factors across 2 (secure vs insecure) × 2 (VP/VLBW vs full-term) groups and 2 (organised vs disorganised) × 2 (VP/VLBW vs full-term) groups. All comparisons were conducted using SPSS V.21.

    Multiple paths that link infant and maternal factors to each other and to attachment disorganisation (higher scores indicated more disorganisation) were examined simultaneously in VP/VLBW and full-term samples with AMOS 19. The goodness of fit of the data to the models was evaluated by examining the χ2 statistics, the root mean square of approximation (RMSEA) and the comparative fit index (CFI). AMOS also allows comparison of the models across samples (ie, VP/VLBW and full-term samples) evaluating the differences in goodness of fit statistics between unconstrained models and models where paths were constrained to be equal across samples. Path analyses were repeated after the cry duration at 3 months variable was log-transformed due to its non-normal distribution. The results were not altered by transformation. Therefore, we decided to report models which included non-transformed cry duration to aid interpretation.

    Results

    Figure 1 reports the full distribution of attachment classifications in our sample. Most infants were securely attached (VP/VLBW=61%; full-term=72%), and the traditional A and C insecure classifications were rare among both VP/VLBW (7%) and full-term children (11%). Using the fine-grained attachment categories (A, B, C and D) no overall significant difference was found between VP/VLBW and full-term controls (χ2=5.91, p=0.12). Comparing along the primary attachment dimensions (ie, secure/insecure, organised/disorganised) significantly more VP/VLBW (32%) than full-term infants (17%) were classified as disorganised (χ2=5.91, p<0.05).

    Figure 1
    Figure 1

    Distribution of A, B, C and D attachment classifications in very preterm/very low birthweight and full-term infants.

    The mean comparisons for the infant and maternal predictors are reported in table 2 (attachment security) and table 3 (attachment disorganisation). The results revealed few differences across groups. Within the insecurely attached group mothers of VP/VLBW infants were more sensitive at term than mothers of full-term infants (table 2). Developmental delay significantly differed across VP/VLBW and full-term groups when both attachment security and attachment disorganisation were considered. VP/VLBW infants who had insecure or disorganised attachment had the lowest mental and motoric development scores compared to others (tables 2 and 3). No significant differences were found in terms of cry duration, distressing cry and maternal sensitivity at 3 months.

    View this table:
    Table 2

    Attachment model factors for VP/VLBW and full-term children who were classified as securely versus insecurely attached

    View this table:
    Table 3

    Attachment model factors for VP/VLBW and full-term children who were classified as organised versus disorganised in their attachment to mother

    Attachment disorganisation

    In path analysis we aimed to examine whether differences in VP/VLBW and full-term infants’ attachment disorganisation would relate to differences in the extent to which attachment disorganisation was explained by developmental delay, other infant characteristics and maternal factors. Figure 2 shows the model tested. The overall model fit for the unconstrained model had a good fit to data χ2(18, N=90, 115)=20.774, p=0.283; CFI=0.967; RMSEA=0.028. Attachment disorganisation was predicted by distressing cry at 3 months (B=0.36, SE=0.16, β=0.28, p<0.05) and developmental delay at 18 months (B=0.78, SE=0.18, β=0.44, p<0.001) in the VP/VLBW sample (explained variation: 34%).

    Figure 2
    Figure 2

    Attachment disorganisation model for very preterm/very low birthweight and full-term infants (regression weights and R2 on the left side of the vertical line are for preterm infants, and regression weights and R2 on the right side of the vertical line are for full-term infants).

    In the full-term sample, only maternal sensitivity at term significantly predicted attachment disorganisation (B=−0.57, SE=0.25, β=−0.21, p=0.019). Children with mothers who were less sensitive at term reported significantly higher levels of attachment disorganisation at 18 months (explained variation: 6%).

    The model where paths were constrained to be equal in VP/VLBW and full-term samples showed an unacceptable fit to data χ2(27, N=90, 115)=53.481, p=0.002; CFI=0.700; RMSEA=0.070. Moreover, the test for model equality yielded an unacceptable model invariance (Δχ2(9)=32.67, p<0.001), indicating that paths in this model did not operate similarly between VP/VLBW and full-term samples. The VP/VLBW and full-term data are best captured by distinct models.

    Discussion

    The present study investigated the prevalence and aetiology of attachment insecurity and disorganisation in VP/VLBW infants. Consistent with most previous results, VP/VLBW infants in our study were not more often insecurely attached than the full-term sample.3 ,13 ,14 ,16 However, VP/VLBW infants were found to have disorganised attachment more often than full-term infants. Furthermore, these differences emerged despite mothers of VP/VLBW children being more (at term) or equally (at 3 months) sensitive in comparison with mothers of full-term infants.

    Low maternal sensitivity at term was the most important predictor of attachment disorganisation in full-term infants only. This was consistent with the findings of a recent study18 which also reported a significant link between maternal sensitivity and disorganised attachment only in a full-term sample but not in a preterm sample. In contrast, for VP/VLBW children, infant characteristics were more important to the developing attachment relationship. For these infants, distressing cry and developmental delay were strong predictors of attachment disorganisation. The risk of disorganised attachment appears to be one of a range of adverse outcomes associated with preterm birth and neurodevelopmental problems are at the core of a spectrum of developmental deficits found in VP/VLBW children.9 ,11 ,26–28

    We speculated that perception of distressing cry29 may be an expression of early neurodevelopmental problems that predispose VP/VLBW infants to disorganised attachment problems. The current study showed that beyond the impact of neurological impairments at birth, later neurodevelopmental delay had a unique impact on explaining attachment disorganisation in the VP/VLBW sample. VP/VLBW birth increases the risk of insult to normal brain development. Even in the absence of identifiable brain injury, VP/VLBW children have been shown to have reduced cortical volume, size and complexity.30 Furthermore, these delays remain or tend to become more, not less, apparent as children mature.10 ,26 ,31–33 As shown here, disorganised attachment is another feature of this symptom complex related to neurodevelopmental problems, which occur despite sensitive parenting.4 ,34 ,35

    Conversely, this may raise the question whether the SSA is an appropriate assessment tool to classify attachment relationships in children with neurodevelopmental problems. Children with developmental delay and less mobility may not be able to express required attachment behaviours coded in the SSA. As suggested by Mangelsdorf et al,15 we conducted the SSA at 18 months of age. Furthermore, we found that even severely developmentally delayed children (average delay of 6 months) were able to express behaviours consistent with organised attachment. Thus, the findings cannot be attributed to methodological problems with the SSA. Additionally, to avoid miscoding behaviours in developmentally delayed infants (eg, repetitive behaviour and stilling) these were coded as signs of attachment only if they occurred exclusively in the presence of the caregiver and were not considered if they were observed throughout the SSA.34

    The strengths of the study are the large sample for attachment studies and the inclusion of a full-term control group drawn from the same hospitals that was closely matched to the VP/VLBW population on a range of socio-demographic indicators. The study was prospective and allowed for analysis of differential precursors of attachment disorganisation. All attachment ratings and classifications were conducted by an accredited independent research group. Limitations of the study are that maternal sensitivity ratings at term covered a longer observation period in the VP/VLBW sample than they were in the full-term sample; slightly more families with VP/VLBW children dropped out and dropouts were more socially disadvantaged. This has been found in most longitudinal at-risk studies23 but is unlikely to impair prediction as shown in simulations.31

    In conclusion, a third of VP/VLBW children showed disorganised attachment patterns in infancy despite sensitive parenting. Converging evidence suggests that the underlying neurodevelopmental problems associated with VP/VLBW birth are at the core of a variety of problems in social relationships of these children. Clinicians should be aware that in VP/VLBW infants, attachment disorganisation may be due to neurodevelopmental problems rather than poor parenting.

    Acknowledgments

    We would like to thank the advisors of the study David Messer, Jay Belsky, Jag Ahluwalia and Mike Thompson; Elizabeth Carlson (University of Minnesota) and team who analysed the strange situation videotapes; and the researchers who assisted in recruitment and data collection: Libi Rust, Karine Edme, Laura Golders, Sue Phillips, Stephanie Auge and Becky Segar. Special thanks are due to the participating hospitals (Addenbrookes Hospital, Cambridge; Luton and Dunstable Hospital, Luton; and Queen Elizabeth II Hospital, Welwyn Garden City) and the parents and their children who participated.

    References

    1. ↵
      1. Bowlby J
      . Attachment and loss: attachment. Vol. 1. New York, NY: Basic, 1969.
    2. ↵
      1. Goldberg S,
      2. Gotowiec A,
      3. Simmons RJ
      . Infant-mother attachment and behavior problems in healthy and chronically ill preschoolers. Dev Psychopathol 1995;7:267–82.
      OpenUrlCrossRef
    3. ↵
      1. van Izjendoorn MH,
      2. Goldberg S,
      3. Kroonenberg PM,
      4. et al
      . The relative effects of maternal and child problems on the quality of attachment—a meta-analysis of attachment in clinical samples. Child Dev 1992;63:840–58.
      OpenUrlPubMedWeb of Science
    4. ↵
      1. van IJzendoorn MH,
      2. Schuengel C,
      3. Bakermans-Kranenburg MJ
      . Disorganised attachment in early childhood: meta-analysis of precursors, concomitants, and sequelae. Dev Psychopathol 1999;11:225–50.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Main M,
      2. Solomon J
      . Procedures for identifying infants as disorganised/disoriented during the Ainsworth Strange Situation. In: Greenburg M, Cicchetti D, Cummings EM, eds. Attachment in the preschool years. Chicago, IL: University of Chicago Press, 1990:121–60.
    6. ↵
      1. Carlson E
      . A prospective longitudinal study of attachment disorganisation/disorientation. Child Dev 1998;69:1107–28.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Gutbrod T,
      2. Wolke D
      . Attachment formation in very premature infants. In: Nocker-Ribaupierre M.ed. Music therapy for premature and newborn infants. Gilsum, NH: Barcelona Publishers, 2004:33–51.
    8. ↵
      1. Weinfield NS,
      2. Sroufe AL
      . Attachment from infancy to early adulthood in a high-risk sample: continuity, discontinuity, and their correlates. Child Dev 2000;71:695–701.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Hille ETM,
      2. den Ouden AL,
      3. Saigal S,
      4. et al
      . Behavioural problems in children who weigh 1000 g or less at birth in four countries. Lancet 2001;357:1641–3.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Moster D,
      2. Lie RT,
      3. Markestad T
      . Long-term medical and social consequences of preterm birth. N Engl J Med 2008;359:262–73.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Samara M,
      2. Marlow N,
      3. Wolke D
      . Pervasive behavior problems at 6 years of age in a total-population sample of children born at ≤25 weeks of gestation. Pediatrics 2008;122:562–73.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Spangler G,
      2. Grossmann K
      . Individual and physiological correlates of attachment disorganisation in infancy. In: Solomon J, George C.eds. Attachment disorganisation. New York, NY: Guilford Press, 1999:95–124.
    13. ↵
      1. Minde KK
      . The impact of prematurity on the later behavior of children and their families. Clin Perinatol 1984;11:227–44.
      OpenUrlPubMed
    14. ↵
      1. Zeanah CH,
      2. Smyke A,
      3. Koga SF,
      4. et al
      . Attachment in institutionalized and community children in Romania. Child Dev 2005;76:1015–28.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Mangelsdorf SC,
      2. Plunkett JW,
      3. Dedrick CF,
      4. et al
      . Attachment security in very low birth weight infants. Dev Psychol 1996;32:914–20.
      OpenUrlCrossRef
    16. ↵
      1. Brisch KH,
      2. Bechinger D,
      3. Betzler S,
      4. et al
      . Early preventitive attachment-oriented psychotherapeutic intervention program with parents of a very low birthweight infant: results of attachment and neurological development. Attach Hum Dev 2003;5:120–35.
      OpenUrlCrossRefPubMed
    17. ↵
      1. Goldberg S,
      2. Perrotta M,
      3. Minde K,
      4. et al
      . Maternal behavior and attachment in low birthweight twins and singletons. Child Dev 1986;57:34–46.
      OpenUrlCrossRefPubMed
    18. ↵
      1. Mijikovitch R,
      2. Moran G,
      3. Roy C,
      4. et al
      . Maternal interactive behavior as a predictor of preschoolers’ attachment representations among full term and premature samples. Early Hum Dev 2012;89:349–54.
      OpenUrlPubMed
    19. ↵
      1. St. James-Roberts I,
      2. Halil T
      . Infant crying patterns in the first year: normal community and clinical findings. J Child Psychol Psychiatry 1991;32:951–68.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Zeskind PS
      . Adult responses to cries of low and high risk infants. Infant Beh Dev 1980;3:167–77.
      OpenUrlCrossRef
    21. ↵
      1. Bayley N
      . Bayley scales of infant development. 2nd edn. San Antonio, Texas: The Psychological Corporation, 1993.
    22. ↵
      1. Zahr L,
      2. Cole J
      . Assessing maternal competence and sensitivity to premature infants’ cues. Issues Compr Pediatr Nurs 1991;14:231–40.
      OpenUrlCrossRefPubMed
    23. ↵
      1. Biringen Z
      . Direct observation of maternal sensitivity and dyadic interactions in the home—relations to maternal thinking. Dev Psychol 1990;26:278–84.
      OpenUrlCrossRef
    24. ↵
      1. Weinberg MK,
      2. Tronick EZ
      . Infant affective reactions to the resumption of maternal interaction after the still-face. Child Dev 1996;67:905–14.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Ainsworth MD,
      2. Bell SM
      . Attachment, exploration, and separation: illustrated by the behavior of one-year-olds in a strange situation. Child Dev 1970;41:49–67.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Johnson S,
      2. Marlow N
      . Positive screening results on the modified checklist for autism in toddlers: implications for very preterm populations. Pediatrics 2009; 154:478–80.
      OpenUrlCrossRefWeb of Science
    27. ↵
      1. Wolke D,
      2. Samara M,
      3. Bracewell M,
      4. et al
      . Specific language difficulties and school achievement in children born at 25 weeks of gestation or less. Pediatrics 2008;152:256–62.
      OpenUrlCrossRefWeb of Science
    28. ↵
      1. Hack M
      . Care of preterm infants in the neonatal intensive care unit. Pediatrics 2009;123:1246–7.
      OpenUrlFREE Full Text
    29. ↵
      1. Van den Boom DC
      . The Influence of temperament and mothering on attachment and exploration—an experimental manipulation of sensitive responsiveness among lower-class mothers with irritable infants. Child Dev 1994;65:1457–77.
      OpenUrlPubMedWeb of Science
    30. ↵
      1. Volpe JJ
      . Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances. Lancet Neurol 2009;8:110–24.
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Bohnert KM,
      2. Breslau N
      . Stability of psychiatric outcomes of low birth weight—a longitudinal investigation. Arch Gen Psychiatry 2008;65:1080–6.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Boyle MH,
      2. Miskovic V,
      3. Van Lieshout R,
      4. et al
      . Psychopathology in young adults born at extremely low birth weight. Psychol Med 2011;41:1763–74.
      OpenUrlCrossRefPubMed
    33. ↵
      1. Marlow N,
      2. Wolke D,
      3. Bracewell MA,
      4. et al
      . Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 2005;352:9–19.
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Pipp-Siegel S,
      2. Siegel CH,
      3. Dean J
      . Neurological aspects of the disorganised/disoriented attachment classification system: differentiating quality of the attachment relationship from neurological impairment. Monogr Soc Res Child Dev 1999;64:25–44.
      OpenUrlCrossRefPubMed
    35. ↵
      1. Jaekel J,
      2. Wolke D,
      3. Chernova J
      . Mother and child behavior in very preterm and term dyads at 6 and 8 years. Dev Med Child Neurol 2012;58:716–39.
      OpenUrl
    View Abstract

    Footnotes

    • Contributors DW: conception and design, drafting and revising and approval of the final version. SE-M: statistical analysis, revising and approval of the final version. TG: conception and design, data collection, drafting and approval of the final version.

    • Funding This study was supported by grant 1590/611 from the Health Foundation, UK.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval University of Hertfordshire and the NHS ethical review boards of the Addenbrookes Hospital, Cambridge; Luton and Dunstable Hospital, Luton; and Queen Elizabeth II Hospital, Welwyn Garden City.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • ↵i The scores for the Mental Developmental Index (MDI) ranged from 50 to 109 in the VP/VLBW sample and from 50 to 123 in the fullterm sample; 17 VP/VLBW children and 7 fullterm children scored 2 standard deviations below the fullterm sample's mean (Mean=93.66, SD=15.27). The scores for the Psychomotor Developmental Index (PDI) ranged from 50 to 111 in the VP/VLBW sample and from 50 to 131 in the fullterm sample; 22 VP/VLBW children and 3 fullterm children scored 2 standard deviations below the fullterm sample's mean (Mean=100.75, SD=13.09).

    • ↵ii All tapes were scored at the Institute of Child Development, University of Minnesota (Elizabeth Carlson) blind to family characteristics and preterm status. Randomly selected tapes (38% preterm; 32% fullterm) were rated a second time by the 3rd author and intercoder reliability was acceptable for both the preterm (κ=0.74) and fullterm samples (κ=0.76).