Detecting infants in need: Are complicated ... - Wiley Online Library

A

R

T

I

C

L

E

DETECTING INFANTS IN NEED: ARE COMPLICATED MEASURES REALLY NECESSARY? KAIJA PUURA

Tampere University and University Hospital, Finland ANTOINE GUEDENEY

Hôpital Bichat-Claude Bernard, AP/HP Paris MIRJAMI MÄNTYMAA AND TUULA TAMMINEN

Tampere University and University Hospital, Finland The object of this paper was to study how the Baby Alarm Distress Scale 共ADBB兲, developed as a simple screening tool for front line professionals working with infants, correlates with the more detailed assessment method of the Global Rating Scale 共GRS兲 for Mother-Infant Interaction at two and four months. A sample of 127 eight- to eleven-week-old infants was videotaped in free interaction with their mothers, and infant interaction behavior was rated with both methods by independent researchers. Compared to the GRS infant scales the sensitivity of the ADBB, using the recommended cutoff point of 5 or more, was 0.77 and specificity 0.80. In further analyses it was found that deviant ratings of two items of the ADBB, the quality of eye contact between the infant and the caregiver and assessment of the sense of relationship between the infant and the caregiver, were the items most strongly associated with poor interaction skills of the infant on the GRS. Mothers of infants found deviant in the ADBB performed more poorly in the interaction with their infants when compared to mothers of infants found healthy in the ADBB. For the purpose of detecting deviations in infant interaction skills as signs of possible problems in early parent-infant interaction the ADBB seems to be a sufficiently sensitive and specific instrument. However, the results of this study still need to be tested with larger samples and against other observation methods.

ABSTRACT:

RESUMEN: Resumen Este artículo tiene como objetivo estudiar cómo la Escala de Alarma de Trastorno del Bebé 共ADBB兲, desarrollada como una simple herramienta explorativa para los profesionales que trabajan con infantes en primera instancia, se correlaciona con el más detallado método de evaluación de la Escala de Puntuación Global para la Interacción Madre-Infante a los dos y cuatro meses 共GRS兲. Un grupo muestra de 127 infantes de ocho a once semanas de nacidos fue objeto de una videograbación This project was financially supported by the European Union 共Leonardo da Vinci Programme, EL 98/98/1/68123/ P1/1.1.1.b/FPC兲, the Ministry of Social Welfare and Health in Finland, the Medical Research Fund of Tampere University Hospital, and the Eemil Aaltonen Foundation. We should like to thank all the primary health care professionals involved in the project and the families and children. We also gratefully acknowledge the help of Raili Salmelin, MSc, PhD, in preparing the tables and figure. Direct correspondence to: Kaija Puura, MD, PhD, Assistant Chief of Child Psychiatry, Tampere University Hospital, P.O. Box 2000, 33521 Tampere, Finland; telephone ⫹358 3 3116 6111; fax ⫹38 3 3116 4375; e-mail: [email protected] INFANT MENTAL HEALTH JOURNAL, Vol. 28(4), 409–421 (2007) © 2007 Michigan Association for Infant Mental Health Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/imhj.20144

409

410

•

K. Puura, A. Guedeny, M. Mäntymaa, and T. Tamminen

durante la interacción libre con sus madres, y la conducta de interacción de los infantes recibió puntuaciones según ambos métodos por parte de investigadores independientes. En comparación con las escalas del infante del GRS, y usando la recomendada línea divisoria de 5 puntos o más, la sensibilidad de ADBB fue de 0.77 y la especificidad de 0.80. En análisis posteriores se determinó que las puntuaciones fuera de lo normal en el caso de dos categorías de ADBB, la calidad del contacto visual entre el infante y quien lo cuidaba, y la evaluación del sentido de relación entre el infante y quien lo cuidaba, fueron los asuntos más fuertemente asociados con las débiles habilidades de interacción del infante en el GRS. Aquellas madres cuyas puntuaciones estaban fuera de lo normal en ADBB tuvieron resultados aún más débiles en la interacción con sus infantes cuando se les comparó con las madres con puntuaciones fuertes en ADBB. Para el propósito de detectar desviaciones de lo normal en las habilidades de interacción de los infantes, como señales de posibles problemas en la temprana interacción entre madre e infante, ADBB parece ser un instrumento suficientemente sensible y específico. Sin embargo, los resultados de este estudio aún necesitan ser examinados con grupos muestras más grandes y comparados con otros métodos de observación. RE´ SUME´ :

Le but de cet article était d’étudier la manière dont l’échelle alarme détresse bébé ADBB, développée comme un simple outil de détection pour les professionnels travaillant avec les bébés, met en corrélation la méthode d’évaluation la plus détaillée de la Global Rating Scale for Mother-Infant Interaction à deux et quatre mois 共GRS兲. Un échantillon de 127 bébés âgés huit à onze semaines a été filmé à la vidéo durant des interactions libres avec leur mère, et le comportement d’interaction du bébé a été évalué par des chercheurs indépendants avec les deux méthodes. Comparée aux échelles GRS bébé, la sensibilité ADBB 共en utilisant la limite recommendée de 5 ou plus兲 était de 0,77 et la spécificité de 0,80. Nous avons trouvé avec des recherches supplémentaires que les taux déviants de deux aspects de l’ADBB, la qualité du contact visuel entre le bébé et le mode de soin et l’évaluation du sens de la relation entre le bébé et le mode de soin, étaient les aspects les plus fortement liés à des compétences interactives faibles pour le bébé évalués par la GRS. Les mères de bébés établis déviants par la ADBB ont plus mal réussi dans l’interaction avec leur bébé, comparée aux mères de bébé en bonne santé dans l’ADBB. Pour ce qui est de détecter les déviations dans les compétences d’interaction des bébés en tant que signe de problèmes possibles dans l’interaction précoce parent-bébé, l’ADBB semble être un instrument suffisamment sensible et spécifique. Cependant, les résultats de cette étude doivent encore être testés avec de plus grands échantillons et comparés aux résultats obtenus à partir d’autres méthodes d’observation. ZUSAMMENFASSUNG: Zusammenfassung Das Ziel dieser Arbeit war der Vergleich zwischen dem Fragebogen genannt “Baby Alarm Distress 共ADBB兲”, der als simples Werkzeug für die Fachleute gedacht ist, die mit Kleinkindern arbeiten und dem detaillierten Diagnoseinstrument des globalen Fragebogens zur Mutter Kind Interaktion 共GRS兲 im Alter von zwei und vier Monaten korreliert. Eine Stichprobe von 127 acht bis elf Wochen alten Kleinkindern wurde mittels Video in freier Interaktion mit ihrer Mutter aufgenommen und das interaktionelle Verhalten des Kindes wurde mittels beider Fragebögen von verschiedenen Untersuchern beurteilt. Die Sensitivität der ADBB, verglichen mit der GRS, wenn man den empfohlenen Trennpunkt mit fünf oder mehr annahm war 0,77 und die Spezifität 0,80. Bei der weiteren Analyse fanden wir, dass die unterschiedliche Beurteilung zweier Einheiten, nämlich die Qualität des Augenkontakts und die Beurteilung der Art der Bindung zwischen dem Kind und seiner Betreuungsperson am meisten mit der Aussage der schlechten Interaktionsfähigkeiten im GRS korrelierten. Mütter und Kinder, die abweichend im ADBB beurteilt wurden, waren schlechter in der Interaktion, wenn man sie mit den Müttern der Kinder verglich, die im ADBB als gesund beurteilt wurden. Um Abweichungen bei den Interaktionsfähigkeiten des Kindes als einem Zeichen möglicher Probleme in der frühen Eltern Kind Interaktion zu finden, scheint der ADBB ein ausreichend sensitives und spezifisches Instrument zu sein. Jedoch ist es erforderlich die Ergebnisse dieser Studie mit einer größeren Stichprobe und anderen Beobachtungsmethoden zu überprüfen.

Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.

Detecting infants in need

*

*

•

411

*

Infants are born with both biologically determined capacities and urge to participate in human interaction 共Emde, 1983; Zeedyk, 1998; Trevarthen & Aitken, 2001兲, and the early parent-infant interaction provides the scaffolding necessary for the infant to develop. Pain, illnesses, and psychological distress can change the interaction behavior of an infant, diminishing his or her capacity to engage and maintain interaction with the parent 共e.g., Tronick & Weinberg, 1997; Guedeney & Fermanian, 2001兲. The quality of parent-infant interaction is affected by the qualities and behavior of both. The parents are sensitized and attuned to meet the needs of their newborn 共Emde, 1983; Brazelton & Cramer, 1990兲, but psychological distress, mental illnesses, and drug and alcohol abuse have been shown to impair their ability to engage and interact with their infants in a positive way 共Murray, Fiori-Cowley, Hooper, & Cooper, 1996; Zeanah, Boris, & Larrieu, 1997; Riordan, Appleby, & Faragher, 1999兲. Unsatisfying mother-infant interactions have long-term consequences for the child, affecting the quality of attachment 共Crittenden, 1995; Teti, Gelfand, Messinger, & Isabella, 1995兲 and restricting the child’s cognitive and socioemotional development 共Murray, 1992; Murray, Sinclair, Cooper, Ducournau, & Turner, 1999; Crockenberg & Leerkes, 2000; Carter, GarrityRokous, Chazan-Cohen, Little, & Briggs-Gowan, 2001; Luoma et al., 2001兲. Since early parent-infant interaction affects both the physiological and psychological development of an infant 共Nelson & Bosquet, 2000; Schore, 2001兲, it would seem reasonable to try to detect possible problems in the parent-infant interaction already in early infancy, in order to intervene for a better outcome for the child. Because the quality of parent-infant interaction depends on both, the assessment of interaction always needs to include an assessment of the infant’s actions and behavior as well as those of the parents 共Brazelton & Cramer, 1990; Guedeney, 1997; Zeanah, Larrieu, Heller, & Valliere, 2000兲. In fact, impaired social behavior of the infant may not only be an indicator of pathology in the infant, but show the first signs of a psychosocial difficulty, like postnatal depression or bonding difficulties which the parent Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.

412

•


has not been able to disclose 共Small, Brown, Lumley, & Astbury, 1994兲. Observation methods are now commonly used in the assessment of parent-infant interaction and development of infants and toddlers. Natural observations by independent observers provide reliable and valid data on the behavior of the child and are valuable when the child’s problems vary across different situations, whereas observations in standardized settings minimize the amount of variation caused by situational factors 共Reid, Patterson, Baldwin, & Dishion, 1988; Keren & Feldman, 2002兲. Some of the methods developed so far work on the microanalytic level, require videotaped material, and are time consuming and costly 共Clark, 1985; Ainsworth, Blehar, Waters, & Walls, 1978; Murray et al., 1996兲. They may be able to answer important questions about the process of interaction, but may not be of practical interest for many practitioners or for research projects with larger samples. On the other hand, other methods, such as the Home Observation for Measurement of the Environment Inventory 共Bradley & Caldwell, 1979兲, are more global and may not give sufficient detail of clinical significance. In helping front line workers in closest contact with infant families, such as primary health care personnel, to detect signs of problems in early parent-infant interaction and those of distress in infants, reliable and user-friendly observation methods would be of use, particularly if the method is based on the assessment of a dimension of behavior closely related to psychopathology. The Baby Alarm Distress Scale 共ADBB, Guedeney & Fermanian, 2001兲 is a new method intended to meet the needs of both clinicians and many researchers. The purpose of the present study was to see how the simpler ADBB method would correlate with a more detailed and better studied method for rating interaction behavior of young infants, such as the Global Rating Scale for Mother-Infant Interaction 共GRS兲共Murray et al., 1996兲. The hypothesis of the study was that the ADBB would distinguish those infants with good interaction behavior from infants with very little or no positive engagement in play and inert or flat behavior according to the GRS infant scales. I. MATERIAL AND METHODS

This study analyzes the Finnish subsample of a larger longitudinal early intervention study, the European Early Promotion Project 共EEPP兲, which has been described in detail in an earlier issue of this journal 共Puura et al., 2002兲. The mother-infant dyads within normal population were invited by public health nurses working at well-baby clinics to participate in the study. The public health nurses contacted mothers six weeks prior to the delivery and four weeks postpartum, and assessed the need status of the families with the help of a questionnaire devised for that purpose. The public health nurses recruited alternately risk and nonrisk families for the EEPP study. After recruiting a family with a need for support, the public health nurses invited the next nonrisk family attending the well-baby clinic to participate. Thus, although the sample was drawn from a normal population, it was enriched with families with some psychosocial risks. However, mothers and infants with severe risk factors such as serious physical problems of the infant, parents with previous or current psychotic problems, or families in need of child protection services, were excluded from the evaluation. Written informed consent was obtained from all participants 共N = 165兲. Half of the nurses had received special training in identifying families with psychosocial risks for the child’s development, in supporting these families, and in promoting motherinfant interaction. Mothers invited to participate in the study by these nurses formed the Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.


•

413

TABLE 1. Characteristics of the Sample (N⫽ 127) % Gender of the child Boy Girl Family structure Mother married or cohabiting Mother single Age of mother Age of partner Household’s net weekly income 共euro兲 Birth weight of the child 共kg兲

Mean

SD

29.3 31.6 446 3.6

4.9 5.9 178 0.5

50 50 94 6

intervention group. The rest of the public health nurses worked in a traditional way and mothers recruited by them formed the control group. The effects of the intervention were analyzed after a two-year follow-up. The present study analyzes data from the initial assessment at the beginning of the intervention. At this time point there were no differences between the intervention and the control group regarding background factors; thus, in the present study the groups were assessed as one. After the initial semistructured interview designed for EEPP 共Puura et al., 2002兲, motherinfant interactions were videotaped when the infants were 8–11 weeks of age. The video recording took place either at home or in the laboratory, when the infant was alert and not hungry. One hundred twenty seven mothers were videotaped and analyzed using the Global Rating Scales for Mother-Infant Interactions 共Murray et al., 1996兲. The two researchers 共M.M. and E.V.兲 who analyzed the videotapes were trained for the Global Ratings method by the Winnicot Research Unit with an inter-rater reliability as follows: at least 90% of the scores within one point of the original score and at least 45% exactly the same. They were blind throughout the study to any information on the mother-infant dyads. For the current study the first author 共K.P.兲 rated infants from the tapes using the ADBB scale. The ratings were made after training for reliability in ADBB with two training tapes sent by the developer of the method 共A.G.兲, and after consensus discussion with him about the training tapes. According to the instructions, the first author first studied the criteria for rating each of the eight items thoroughly. She then watched and rated the five cases on the first training tape. After comparing her ratings with the correct ratings of the training tape, she watched and rated the first training tape one case at a time. The second training tape was worked through case by case until there was perfect agreement of caseness and no more than one point difference in ratings for less than three items per case. After the initial training the first author watched and rated three to five random cases from the training tapes at the beginning of each rating session. The first author was blind throughout the study to any information on the infants. These 127 infants assessed by both observation methods form the sample for this study. The characteristics of the sample are given in Table 1. There were no significant differences between the analyzed sample and those 38 mother-infant dyads that were not. Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.

414

•


A. Global Rating Scales for Mother-Infant Interaction

The Global Rating Scales for Mother-Infant Interaction were initially developed for distinguishing depressed from nondepressed mother-infant interaction in research conditions, and have shown good discriminant validity for a number of clinical groups such as depression and schizophrenia, social adversity, and low-risk/high-risk groups 共Murray et al., 1996; Riordan et al., 1999; Gunning, Murray, & Lawson, 2002兲. In this procedure mothers and infants interact face to face for 5 minutes without using toys. The video recording captures a full-face image of the infant, the infant’s upper limbs and trunk, and also the mother’s full face through its reflection in a mirror placed adjacent to the infant. Three maternal dimensions 共describing the mother’s behavior and her depressive symptoms兲, two infant dimensions 共describing the infant’s behavior兲, and an interaction dimension 共describing the goodness of fit within the dyad兲 are scored. The first infant dimension 共Inf1兲 describes the quality of engagement of the infant on a scale ranging from 1 to 5, with 1 representing avoidance of engagement and 5 very good engagement. The items in the Inf1 include attentiveness to mother, active communication, and positive vocalizations. Each item is scored from 1 to 5, and the average score is calculated. The higher the average is the better. The second infant dimension 共Inf2兲 includes inert subscore and fretful subscore. The inert subscore includes items of engagement with environment, liveliness, and attentiveness, which are scored from 1 to 5, with 1 representing the most inert end. An average inert score is calculated from the ratings of the individual items. The fretful subscore is formed by items happy-distressed and nonfretful-fretful, scored from 1 to 5, with 1 representing the most fretful end. An average fretful score is calculated from the ratings of the individual items. Finally, the averaged fretful score is subtracted from the averaged inert score to obtain the final Inf2 score. The Inf2 score ranges from −2 representing inertness to ⫹2 representing fretfulness. Scores close to 0 represent normal behavior. The maternal dimensions 共mother’s interactive behavior, mother’s intrusiveness or remoteness, and mother’s affective behavior兲 consist of four to five items, each of which is scored on a 5-step scale from 1 共poor兲 to 5 共good兲. The mean score of the items is calculated for mother’s interactive and affective behavior with a possible range from 1 to 5. For mother’s intrusiveness or remoteness the difference between intrusiveness 共mean score of the two items describing intrusive behavior and intrusive speech兲 and remoteness 共mean score of the two items describing remote and taciturn behavior兲 is calculated and then divided by two, with a possible range from −2 共intrusive兲 to +2 共remote兲 and 0 representing the optimal state. The interaction dimension consists of five items: smooth/easy versus difficult, fun versus serious, mutually satisfying versus unsatisfying, much engagement versus no engagement, and excited engagement versus quiet engagement. Each item is scored on a 5-point scale, with 5 being the best and 1 the worst, and the mean score calculated. B. The Baby Alarm Distress Scale

The Baby Alarm Distress Scale 共Guedeney & Fermanian, 2001兲 was designed to be a part of a medical examination in a well-baby clinic. It consists of eight items concerning the behavior and features of the baby 共facial expression, eye contact, vocalization, overall level of activity, self-stimulating behavior, briskness of response to stimulation, attraction towards the infant, relationship between the infant and the observer兲, which the observer rates on a 5-point scale ranging from 0 to 4. For each item 0 represents best functioning or normality of the infant and Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.


•

415

TABLE 2. Range, Mean, and Standard Deviations of Individual ADBB Items and of the Total Score. Frequency of Scores of 1 or More is Also Given for Each Item and for the Total Score Item Facial expression Eye contact General level of activity Self-stimulating gestures Vocalizations Briskness of response to stimulation Existence of relationship Attraction toward infant Total sum score

Range

Mean

SD

Frequency of scores above 0 共%兲

0–2 0–4 0–1 0–3 0–4 0–1 0–3 0–3 0–13

0.41 0.75 0.09 0.44 0.65 0.03 0.54 0.79 3.67

0.66 1.05 0.29 0.81 0.57 0.18 0.71 0.60 2.83

31.4 43.3 9.4 27.6 62.2 3.1 42.5 70.1 85.0

4 severe abnormality. A score of 5 or higher is thought to be deviant and a sign of distress of the infant, as shown in a study in a well-baby clinic with 60 infants 共Guedeney & Fermanian, 2001兲. The rating is done immediately after observation in a live situation, or on videotape. In this study we examined the infant’s performance with his or her mother in a standardized setting provided by the GRS task of interacting face to face for 5 minutes. The rater watched a 5-minute recording of each infant in interaction with his or her mother once, and rated the infant’s behavior immediately after that. The items of eye contact and sense of relationship were rated from the tapes as they occurred between the infant and his or her mother. II. RESULTS

In this sample of 127 infants the mean for engagement on the GRS infant scale 共Inf1兲 was 2.98, with the minimum value of 1 and maximum of 5. The score of 2 distinguished the lowest 15%. The mean for fretful or inert behavior 共Inf2兲 was −0.20, with a value range from −2 to +2. The lowest 15% representing the inert end were distinguished with a score of −0.68 and the highest 15% representing the fretful end with a score of 0.30. Withdrawal behavior was considered a more significant symptom for distress in infants this young, and in the analyses an infant’s interaction behavior was considered to be deviant if he or she scored 2.0 or less in the Inf1 scale, or −0.70 or less in the Inf2 scale. In our sample the mean score of the ADBB was 3.67, with a range of 0–13. The highest 15% were distinguished by a score of 7. The mean scores, standard deviations 共STD兲, and range of each of the items can be seen in Table 2. The correlation between the ratings of the two methods was first studied with cross tabulation. Calculated from Table 3, using the original cutoff point of 5, the ADBB had a sensitivity of 0.77 and specificity of 0.80. The percentage of agreement between the two methods was 79.5%. With a higher cutoff point of 6, the sensitivity of the ADBB was 0.65 and specificity 0.89. With a cutoff point of 7, the sensitivity was dropped to 0.43 and specificity rose to 0.96. Figure 1 shows the development of sensitivity and specificity for different cutoff points of the ADBB. As can be seen from Figure 1, in this study the ADBB cutoff point of 5 or more proved to be the optimal cutoff point. Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.

416

•


TABLE 3. Number of Infants Classified as Healthy or Deviant With the Two Observation Methods of GRS Infant Scales and ADBB With Three Different Cutoff Points. Scores of GRS Infant Scale are Deviant if Inf1 is Below 2 or Inf2 is Below −0.70 GRS infant scales

ADBB score Below 5 5 or above Below 6 6 or above Below 7 7 or above

Normal 共N = 90兲

Deviant 共N = 37兲

Total 共N = 127兲

72 18 80 10 86 4

8 29 13 24 21 16

80 47 93 34 86 4

Sensitivity

Specificity

0.77

0.80

0.69

0.89

0.46

0.96

The differences in the GRS scales of maternal behavior, infant behavior, and interaction between infants scoring below and above the cutoff point of 5 in the ADBB are given in Table 4. Infants in the ADBB deviant group were significantly less engaged and more inert than infants scoring below the cutoff point. Mothers of ADBB deviant infants were also less engaged and more intrusive than mothers of infants scoring below the cutoff point, and the quality of mother-infant interaction was poorer. There was no significant difference between ADBB deviant infants and nondeviant infants in the maternal subscale for withdrawn behavior of the mother. Finally, a binary regression analysis was conducted with all eight items of the ADBB as explanatory variables. The items were categorized with 0 representing normality and all other

FIGURE 1. Sensitivity and specificity of the ADBB for different cutoff points when compared to deviant GRS infant scales. Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.


•

417

TABLE 4. Mean Scores of the GRS Scales for Infants Scoring Below and Above the Cutoff Point of 5 Points in the ADBB

GRS scales Mat 1 共engagement兲 Mat 2a 共intrusive behavior兲 Mat 2b 共withdrawn behavior兲 Mat 3 共warm affect兲 Inf1 共engagement兲 Inf2 共inert-fretful兲 Interaction

ADBB under 5 mean

ADBB 5 or more mean

Sig

3.75 0.15 0.15 4.07 3.44 −0.04 3.46

3.43 0.30 0.23 3.70 2.25 −0.41 2.42

0.006 0.05 ns 0.005 0.001 0.001 0.001

ratings 共from 1 to 4兲 representing deviance. The two items that were most strongly connected with a deviant score in either of the GRS infant scales were infant’s eye contact with mother 共OR 2.42, 95% CI 1.62–3.22, p = 0.003兲 and observer’s rating of the quality of the relationship between the infant and the mother 共OR 1.68, 95% CI 0.94–2.42, p = 0.02兲. These two categorized items together 共i.e., scoring a 1 or more on both items兲 had a sensitivity of 0.95 and specificity of 0.71 when compared to GRS infant scale ratings. III. DISCUSSION

In this study we aimed at finding out how well a simpler observation method designed to be used in primary services for infant families would correlate with a more detailed and better known method. The GRS was designed specifically to identify depressed and nondepressed mother-infant interaction for research purposes, and has shown predictive validity for subsequent performance 共Murray et al., 1996; Murray, Hipwell, Hooper, Stein, & Cooper, 1996兲 and good discriminant validity for a number of clinical groups such as depression and schizophrenia, social adversity, and low-risk/high-risk groups 共Murray et al., 1996; Riordan et al., 1999; Gunning et al., 2002兲. The sample of this study consisted of healthy full-term babies without any major risk factors, indicating a need for child protection or mental health services. The aim of the study was not to assess depressed parents particularly, but mother-infant dyads with various psychosocial problems. This may mean that the GRS was used in a nonoptimal context. However, infants’ interaction behavior shows similar deviant features regardless of the cause for unsatisfying parent-infant interaction 共Field et al., 1998; Powell & Bettes, 1992; Riordan et al., 1999; Guedeney, 1997兲, and therefore the use of GRS infant scales particularly can be justified. Despite the fact that the sample did not include infants living in severely adverse conditions, there were 37 infants who were deviant in their interaction behavior according to the GRS infant scales. The infants in the study were assessed quite early in life, at 8–11 weeks, with the effects of relationship disturbances probably not being so outstanding than if the assessment had taken place later. The nonclinical, fairly well-to-do sample and early assessment time may explain why the mean score of the sample in the ADBB was rather low, and Infant Mental Health Journal DOI 10.1002/imhj. Published on behalf of the Michigan Association for Infant Mental Health.

418

•


why the frequencies of most deviant scores were low or even nonexistent concerning some of the features observed, such as delayed response time or frequent self-stimulation. Used in its initial purpose the ADBB is supposed to be applied directly after live observation, and ratings should mostly refer to the infant’s behavior with a stranger instead of the mother. The interaction with the mother used in this study may have helped at least some of the infants to perform better and get better ratings in the ADBB than otherwise. On the other hand, infants being recorded in interaction with their own mothers in a standardized task most likely minimized the bias resulting from interacting with a stranger in varying situations 共Reid et al., 1988兲. The fact that the ADBB used with videotaped interaction readily detected 80% of those infants rated as having poor interaction skills on the GRS infant scales in this study suggests further applicability of the method, particularly since the number of infants rated as false positives in comparison to GRS infant scales remained small. As in the earlier study done with the ADBB 共Guedeney & Fermanian, 2001兲, the suggested cutoff point of 5 or more proved to be the optimal with satisfactory sensitivity and specificity. The items common to the two observation methods include observation of the quality of eye contact of the infant, and sense of engagement, which may partly explain why the eye contact and relationship item of the ADBB had the strongest correlation with GRS rating of deviant interaction behavior. Currently there is no gold standard for measuring infant interaction skills, and comparing different measures with each other is needed to establish some conception of their usefulness. In this study there were clear, significant differences in the interaction behavior on the GRS infant scales between the infants scoring above the cutoff point of 5 in the ADBB compared to infants who scored low, indicating that both measures did indeed distinguish infants with poor interaction skills. It may be that eye contact is of such crucial importance in normal infant development 共Farroni, Csibra, Simion, & Johnson, 2002兲, that it also is the first feature affected by any adversities the infant may experience. Eye contact is also the main mode of establishing an early relationship between a parent and an infant, and infants from birth prefer looking at faces that engage in mutual gaze 共Hains & Muir, 1996; Farroni et al., 2002兲. The items of eye contact together with the sense of relationship were the most sensitive in detecting infants with poor interaction behavior, and although the specificity was lowered to 0.70 compared to the full ADBB, a measure looking at only these two features might be sufficient for some purposes. In addition to differences in the infant interaction behavior in the ADBB deviant group, there were also clear differences in the engagement and affect of the mothers between infants scoring above the cutoff point of 5 and infants scoring below it. This indicates that the ADBB might well be used as a screening tool in primary health care, for detecting both infant distress and relationship disturbances. In line with a recent study 共Dollberg, Feldman, Keren, & Guedeney, 2006兲, mothers of infants scoring high on the ADBB were also more intrusive than mothers of infants scoring below the cutoff point on the ADBB. In conclusion it can be stated that the ADBB proved to have an acceptable sensitivity and specificity in detecting impaired infant interaction skills, and even in distinguishing poor parent-infant interaction in general. The fact that the researcher doing the ADBB was able to use the method after training done with specific training tapes speaks for the feasibility of the method. Being fairly simple to learn and use, it may prove to be a useful tool in the early detection of infant distress. This is of no small importance, as lowered infant sociability has been shown to be connected to both physical and psychological illness 共Tronick & Weinberg, 1997; Guedeney & Fermanian, 2001; Mäntymaa et al., 2003兲 and also to parental distress and



•

419

relationship issues between the parent and the infant 共Trevarthen & Aitken, 2001兲. In a longitudinal study by Keller & Zach 共1993兲 infants who averted their gaze from their parents were found to have poorer interaction with their parents at 2 years and developmental delays up to 6 years. Withdrawal behavior in a feeding situation also significantly differentiated clinical from nonclinical populations 共Keren, Feldman, & Tyano, 2001; Dollberg et al., 2006兲. By being able to pay attention to infant distress, the front line workers in primary care could have significantly better chances of preventing problems and also intervening early should problems be detected in the early parent-infant interaction. Early detection of poor infant interaction skills may also help in anticipating vulnerability in the physical health of the infants and help parents to provide a better environment for their infants in order to reduce the risk for illness 共Mäntymaa et al., 2003兲. Finally, a simple and standardized measure would make it possible to make relatively reliable repeated assessments, even with different assessors, for following up normal infant development or the effect of an intervention in primary care. This would increase the quality of care, as infants showing prolonged deviance in development or no improvement during an intervention could be referred to specialist services without delay. REFERENCES Ainsworth, M.D.S., Blehar, M.C., Waters, E., & Walls, S. 共1978兲. Patterns of attachment. Hillsdale, NJ: Lawrence Erlbaum. Bradley, R., & Caldwell, B. 共1979兲. Home observation for measurement of the environment: A revision of the pre-school scale. American Journal of Mental Deficiency, 84, 235–244. Brazelton, T.B., & Cramer, B.G. 共1990兲. The earliest relationships. Reading, MA: Addison-Wesley. Carter, A.S., Garrity-Rokous, E., Chazan-Cohen, R., Little, C., & Briggs-Gowan, M. 共2001兲. Maternal depression and comorbidity: predicting early parenting, attachment security, and toddler socialemotional problems and competencies. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 18–26. Clark, R. 共1985兲. The parent-child early relational assessment. Madison: University of Wisconsin, Medical School, Department of Psychiatry. Crittenden, P.M. 共1995兲. Attachment and risk for psychopathology: The early years. Journal of Developmental and Behavioral Pediatrics, 16, 12–16. Crockenberg, S., & Leerkes, E. 共2000兲. Infant social and emotional development in family context. In C.H. Zeanah 共Ed.兲, Handbook of infant mental health. New York: Guilford Press. Dollberg, D., Feldman, R., Keren, M., & Guedeney, A. 共2006兲. Sustained withdrawal behavior in clinic referred and nonreferred infants. Infant Mental Health Journal, 27, 292–309. Emde, R.N. 共1983兲. Pre-representational self and its affective core. Psychoanalytic Study of the Child, 38, 165–192. Farroni, T., Csibra, G., Simion, F., & Johnson, M.H. 共2002兲. Eye contact detection in humans from birth. Proc. Natl. Acad. Sci. U.S.A., 99, 9602–9605. Field, T.M., Scafidi, F., Pickens, J., Prodromidis, M., Pelaez-Nogueras, M., Torquati, J., et al. 共1998兲. Polydrug-using adolescent mothers and their infants receiving early intervention. Adolescence, 33, 117–43. Guedeney, A. 共1997兲. From early withdrawal reaction to infant depression: A baby alone does exist. Infant Mental Health Journal, 18, 339–349.


420

•


Guedeney, A., & Fermanian, J. 共2001兲. A validity and reliability study of assessment and screening for sustained withdrawal reaction in infancy: The alarm distress baby scale. Infant Mental Health Journal, 22, 559–575. Gunning, M., Murray, L., & Lawson, C. 共2002, July兲. Global rating scheme. Paper presented at the Eighth Congress of the World Association for Infant Mental Health, Amsterdam. Hains, S.M., & Muir, D.W. 共1996兲. Infant sensitivity to adult eye direction. Child Development, 67, 1940–1951. Keller, H., & Zach, U. 共1993兲. Developmental consequences of early eye contact behavior. Acta Paedopsychiatrica, 56, 31–36. Keren, M., & Feldman, R. 共2002兲. The role of feeding interaction assessment in the routine psychiatric evaluation of the infant. Devenir, 14, 1:5–16. Keren, M., Feldman, R., & Tyano, S. 共2001兲. Diagnoses and interactive patterns of infants referred to a community-based infant mental health clinic. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 27–35. Luoma, I., Tamminen, T., Kaukonen, P., Laippala, P., Puura, K., Salmelin, R., et al. 共2001兲. Longitudinal study of maternal depressive symptoms and child well-being. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 1367–74. Mäntymaa, M., Puura, K., Luoma, I., Salmelin, R., Davis, H., Tsiantis, J., et al. 共2003兲. Infant-mother interaction as a predictor of child’s chronic health problems. Child Care Health Dev., 29, 181–191. Murray, L. 共1992兲. The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry, 33, 543–561. Murray, L., Fiori-Cowley, A., Hooper, R., & Cooper, P. 共1996兲. The impact of postnatal depression and associated adversity on early mother-infant interactions and later infant outcome. Child Development, 67, 2512–2526. Murray, L., Hipwell, A., Hooper, R., Stein, A., & Cooper, P. 共1996兲. The cognitive development of 5-year-old children of postnatally depressed mothers. Journal of Child Psychology and Psychiatry, 37, 927–935. Murray, L., Sinclair, D., Cooper, P., Ducournau, P., & Turner, P. 共1999兲. The socioemotional development of 5-year-old children of postnatally depressed mothers. Journal of Child Psychology and Psychiatry, 40, 1259–1271. Nelson, C.A., & Bosquet, M. 共2000兲. Neurobiology of foetal and infant development: implications for infant mental health. In C.H. Zeanah 共Ed.兲., Handbook of infant mental health. New York: Guilford Press. Powell, G.F., & Bettes, B.A. 共1992兲. Infantile depression, non-organic failure to thrive and DSM-III-R: A different perspective. Child Psychiatry and Human Development, 22, 185–196. Puura, K., Davis, H., Papadopoulou, K., Tsiantis, J., Ispanovic, V., Rudic, N., et al. 共2002兲. The European early promotion project: A new primary health care service to promote children’s mental health. Infant Mental Health Journal, 23, 606–625. Reid, J.B., Patterson, G.R., Baldwin, D.V., & Dishion, T.J. 共1988兲. Observations in the assessment of childhood disorders. In M. Rutter, A.H. Tuma, & I.A. Lann 共Eds.兲, Assessment and diagnosis in child psychopathology. New York: Guilford Press. Riordan, D., Appleby, L., & Faragher, B. 共1999兲. Mother-infant interaction in postpartum women with schizophrenia and affective disorders. Psychol. Med., 29, 991–995. Schore, A.N. 共2001兲. Effects of a secure attachment relationship on right brain development, affect regulation, and infant mental health. Infant Mental Health Journal, 22, 7–66.



•

421

Small, R., Brown, S., Lumley, J., & Astbury, J. 共1994兲. Missing voices: what women say and do about depression after childbirth. Journal of Reproductive and Infant Psychology, 12, 89–103. Teti, D.M., Gelfand, D.M., Messinger, D.S., & Isabella, R. 共1995兲. Maternal depression and the quality of early attachment: An examination of infants, preschoolers, and their mothers. Developmental Psychology, 31, 364–376. Trevarthen, C., & Aitken, K.J. 共2001兲. Infant intersubjectivity: Research, theory, and clinical applications. Journal of Child Psychology and Psychiatry, 42, 3–48. Tronick, E.Z., & Weinberg, M.K. 共1997兲. Depressed mothers and infants: failure to form dyadic states of consciousness. In L. Murrayand P. Cooper 共Eds.兲, Postpartum depression and child development. New York: Guilford Press. Zeanah, C.H., Boris, N.W., & Larrieu, J.A. 共1997兲. Infant development and developmental risk: A review of the past 10 years. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 165–178. Zeanah, C.H., Larrieu, J.A., Heller, S.S., & Valliere, J. 共2000兲. Infant-parent relationship assessment. In C.H. Zeanah 共Ed.兲, Handbook of infant mental health 共2nd ed.兲. New York: Guilford Press. Zeedyk, M.S. 共1998兲. Parent-infant interaction: Interpreting meaning in infants’ actions. In C.A. Niven & A. Walker 共Eds.兲, Current issues in infancy and parenthood. Oxford: Butterworth Heinemann.
