A scale for home visiting nurses to identify risks of physical abuse and neglect among mothers with newborn infants

Child Abuse & Neglect 28 (2004) 321–337

A scale for home visiting nurses to identify risks of physical abuse and neglect among mothers with newborn infants Hans Grietens∗ , Liesl Geeraert, Walter Hellinckx Centre for Disability, Special Needs Education, and Child Care, Katholieke Universiteit Leuven, Vesaliusstraat 2, B-3000 Leuven, Belgium Received 4 February 2003; received in revised form 17 February 2003; accepted 3 October 2003

Abstract Objective: The aim was to construct and test the reliability (utility, internal consistency, interrater agreement) and the validity (internal validity, concurrent validity) of a scale for home visiting social nurses to identify risks of physical abuse and neglect in mothers with a newborn child. Method: A 71-item scale was constructed based on a literature review and focus group sessions with social nurses and paraprofessionals who had experience with underprivileged families. This scale was applied in a random sample of 40 home visiting social nurses, who collected data in a sample of 373 nonabusive and 18 abusive/neglectful mothers with a newborn child. Results: Items with prevalence rates below 5% and items making no significant difference between maltreating and non-maltreating mothers were omitted. The final version contained 20 items. This scale showed high internal consistency (α = .92) and high interrater reliability (r = .97). Exploratory factor analysis yielded a three-factor solution: Isolation (8 items, explaining 62.17% of the common variance), Psychological complexity (6 items, 18.86%), and Communication problems (6 items, 8.41%). Scores on Communication problems and Isolation significantly predicted scores on a social deprivation scale, which significantly distinguished maltreating from non-maltreating mothers. Mothers scoring high on Communication problems or Isolation obtained higher scores for social deprivation than low-scoring mothers. Conclusions: Home visiting nurses can identify risks for physical abuse and neglect among mothers with a newborn infant by focusing on signs of social isolation, distorted communication and psychological problems. © 2004 Elsevier Ltd. All rights reserved. Keywords: Risk; Physical abuse; Neglect; Infancy

∗

Corresponding author.

0145-2134/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.chiabu.2003.10.011

322

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

Introduction To prevent intrafamilial child physical abuse and neglect, instruments are needed to identify early risks. Measuring parents’ potential for maltreatment, such instruments may be helpful for practitioners to make well-considered decisions concerning interventions in high-risk families in order to prevent child physical abuse and neglect (Milner, Murphy, Valle, & Tolliver, 1998; Pecora, 1991). During the last three decades, many efforts have been undertaken to screen the potential for abuse and neglect in parents and families (Browne, Davies, & Stratton, 1988; Farnell, 1980; Starr, 1982). Some instruments proved to be successful in making predictions of future harm in the child and of abusive or neglectful acts by the parent (Browne & Herbert, 1997). Many instruments, however, produced large numbers of so-called “false positives” by classifying nonabusive caregivers as at risk for abuse (Rodwell & Chambers, 1992). In addition, many instruments lack sufficient reliability and validity (Lutzker, Van Hasselt, Bigelow, Greene, & Kessler, 1998; Milner et al., 1998; Wald & Woolverton, 1990). Being developed mainly for practical purposes, only a few instruments encompassed the growing body of knowledge on etiological factors; moreover, few instruments were based on theoretical models supported by empirical findings (Baartman, 1996; Grietens, Hellinckx, Van Assche, Baartman, & Geeraert, 1999; Milner, 1986). In the present study, we report on a scale for home visiting social nurses to identify risks on child physical abuse and neglect in families with 0- to 3-month-old infants. The term “social nurses” is the position title of nurses who do home visits under the auspices of Child & Family, a government agency that is commissioned by law with the promotion of health care and well-being of families with 0- to 3-year-old children in the Flemish Community of Belgium. In addition to their medical formation, these nurses are trained in different aspects of social work (e.g., conducting interviews with parents). The work the nurses do can be compared with that done in some countries by community or public health nurses. Underlying the scale is an interactional view on child maltreatment (Grietens et al., 1999). This view adheres to ecological models on the determinants of parenting (Belsky, 1997; Belsky & Vondra, 1989) and to psychological (Cerezo, 1997) and pedagogical (Baartman, 1996) theories on child maltreatment. It considers child physical abuse and neglect as extreme manifestations of parenting problems, expressing severe problems in the relationship between parent and child. Further, it stresses the cumulative effects long-standing negative interactions with parents may have for the child, as well as the mediating or buffering role the parents’ person may play against outer stressors (e.g., unemployment) or predisposing child factors (e.g., prematurity). Many studies have demonstrated that early manifestations of inadequate parenting, problematic and disturbed parent-child relationships are precursors to child physical abuse and neglect (Ammerman, 1990; Becker-Lausen & Mallon-Kraft, 1997; Rogosch, Cicchetti, Shields, & Toth, 1995). Synthesizing the literature on parenting and child maltreatment, Baartman (1996, see also Grietens & Hellinckx, 2003) re-introduced the construct “parental awareness.” This construct was used earlier by Newberger (1980) and denoted as “an organized knowledge system with which the parent makes sense out of the child’s responses and behavior and formulates policies to guide parental action” (p. 47). According to Baartman’s synthesis, the dynamics underlying parental awareness are the parent’s capacity to take the child’s perspective and the parent’s willingness to serve the child’s needs and claims. The cognitions and behaviors of parents at-risk for child physical abuse and neglect may be characterized by a lack of perspective-taking abilities and a decreased willingness to serve the child’s needs and claims. At-risk parents have inappropriate or forced expectations with regard to the meaning the child can have for their own private well-being. Further, they have negative emotions towards their child, regardless of the child’s

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

323

characteristics or actual behavior. Finally, they display a lack of sensitivity and responsiveness towards the child’s needs for care and safety, due to the permanent imbalance between the parent’s and the child’s claims. Other risk factors of child physical abuse and neglect considered in the interactional view are the parent’s or family’s history of maltreatment, a lack of social support, parental personality problems and certain infant characteristics (Baartman, 1996; Coohey & Braun, 1997; Kolko, 1996; Rogosch et al., 1995). Many studies stress the importance of intergenerational dynamics of child physical abuse and neglect (Kaufman & Zigler, 1989), in particular the psychological sequel of maltreatment (Knutson, 1995), the way parents actually cope with the negative experiences of their own childhood (Rogosch et al., 1995), attachment to maltreating parents (Crittenden, 1988; Zuravin, McMillan, DePanfilis, & Risley-Curtiss, 1996), and the witnessing of domestic violence (Cerezo, 1997; Coohey & Braun, 1997). A lack of social support is repeatedly reported as being a significant precursor to child maltreatment (Cerezo, 1997; Erickson & Egeland, 1996; Milner, 1993). Particularly vulnerable are socially isolated mothers with a young infant who are deprived from support by their spouse or family. Further, numerous parental personality characteristics not directly related to child-rearing were found to have more or less predictive power with regard to child maltreatment, for instance depression, alcohol or drug abuse, generalized hyperreactivity and oversensitivity, attributional styles, and a lack of coping or problem-solving abilities (reviewed by Bugental, Mantyla, & Lewis, 1989; Cantos, Neale, O’Leary, & Gaines, 1997; Casanova, Domanic, McCanne, & Milner, 1992; Kolko, 1996; Milner & Dopke, 1997). Finally, the child’s active contribution to abusive and neglectful interactions with parents and caregivers is stressed by empirical studies on the direct influence of child characteristics (e.g., difficult temperament) on parenting processes and parental perceptions (Ammerman, 1990; Harrington, Black, Starr, & Dubowitz, 1998; Houldin, 1987). Since child physical abuse and neglect are considered to be extreme manifestations of parenting problems, a cultural perspective should be included in the measurement. Indeed, parenting cannot be disentangled from its cultural context. Many studies showed significant differences between cultures in definitions of child maltreatment, identification of risk factors, and interpretation of parenting practices (Kapitanoff, Lutzker, & Bigelow, 2000; Korbin, 1997, reviewed cultural perspectives on child maltreatment; see also Harkness & Super, 1995, for a discussion on the relationship between culture and parenting). Further, the measurement of risk factors is strongly tied to the setting in which measurement will take place. The target group of the instrument we developed were social nurses employed by Child & Family. Usually, a nurse visits the mother in the hospital shortly after delivery. She explains the services provided by the agency. If the mother agrees, four home visits within 3 months after birth in case of a first child, and three home visits within this period in case of a next child, are planned. Within the first 3 months, there are three ambulatory consultations. Agreement implies that the mother gives consent to all actions undertaken by the agency (medical as well as psychosocial actions). Each home visit lasts about 40 minutes and offers the possibility to obtain detailed information on the parenting process. However, most nurses lack an evidence-based framework on child-rearing to interpret their observations and have to base their reports on “intuition.” Child & Family reaches about 90% of all families with a newborn infant (Child & Family, 2001). Services provided by the agency are voluntary. All parents can at any time terminate the contacts. Recently, Child & Family focuses on parenting issues (e.g., by providing information on child development and best practices with regard to parental investment in infants, or by offering early parenting support). Helping the nurses to identify early signs of problems between parents and children by providing them useful screening tools, is currently high on the agency’s agenda. The cultural embeddedness of parenting, the nurses’ work setting, their specific needs, and the agency’s recent attention

324

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

for parenting issues and parent-child dynamics made that we could not simply translate and implement existing scales developed in other countries, for example in the United Kingdom (Browne, 1995; Browne et al., 1988). Objectives of the present study The present study’s objectives were: (1) to develop a risk scale, containing items that operationalize the aforementioned sets of interrelated risk factors (disturbed parent-child relationships, history of maltreatment in parent or family, lack of social support, parental personality problems, negative infant characteristics), for home visiting nurses in families with newborn children; (2) to implement this scale in a group of nonabusive mothers and in a group of mothers, who were diagnosed as abusive or neglectful, in order to reduce the number of risk items by examining their prevalence and discriminative power; (3) to examine the reliability of the reduced scale (utility, internal consistency, interrater agreement); and (4) to initiate the examination of the reduced scale’s validity (internal validity, concurrent validity). By adopting an interactional view on child physical abuse and neglect and by concretizing symptoms of early mismatches between parents and children, we tried to develop a risk scale that was tailored to the needs of the social nurses doing home visits under the auspices of Child & Family, offering them an evidence-based framework on child-rearing and helping them to interpret their observations. The study was approved by the Interdisciplinary Research Board of Child & Family.

Method Measures First version of the risk scale. The selection of items was based on a literature review and on information from practitioners. The literature review focused on the integration of different etiological models into a comprehensive theoretical framework and on the search for the most predictive risk factors. In addition, advantages and limitations of various risk assessment procedures and scales were summarized (Grietens et al., 1999). Information from practitioners employed by Child & Family was gathered using focus groups (Morgan, 1993). Three focus groups were composed, two consisting of social nurses (n = 18) and one consisting of paraprofessionals, who worked as volunteers for the agency and accompanied the social nurses during home visits in underprivileged and ethnic minority families, because they had experience with these families’ way of living and their everyday problems (n = 9). The group interviews focused on the practitioners’ perceptions of risk factors related to parenting problems during hospital visits, home visits and consultations and on their professional experiences with regard to issues of risk assessment (identification, decision-making, and intervention). The narratives provided concrete information on the appearance of risk factors for early parenting problems and child maltreatment in the daily work of social nurses and paraprofessionals. Furthermore, remarkable similarities were found between the practitioners’ accounts and the prevailing literature on risk factors and etiological models (Grietens et al., 1999). Based on the literature review and the practitioners’ narratives, a scale containing 71 items was developed, covering four topics: (1) interaction between social nurse and mother (9 items), (2) mother-child interactions (35 items), with items on inappropriate expectations, negative emotions, and lack of sensitivity and responsiveness, (3) maternal and family characteristics (22 items), with items on history,

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

325

social support and non-child related maternal personality characteristics, and (4) child characteristics (5 items) (a full list of items is available upon request from the first author). Items were operationalized as concrete as possible. Draft versions of the scale were commented on comprehensibility and relevance by the participants of the focus groups. All items were negatively formulated and had to be scored on a four-point rating scale, ranging from 0 = never observed or reported to 3 = very often (at least one time during each home visit) observed or reported. High scores pointed to the presence of risk factors. Scoring criteria and item-specific scoring rules were written down in an accompanying manual (Hellinckx, Baartman, Grietens, Van Assche, & Geeraert, 1999). To score the items, the nurses had to rely on observations, except for the maternal and family characteristics which had to be scored based upon verbal reports. Social deprivation scale. Six variables, measuring families’ socioeconomic living conditions were included: financial problems (in particular, a monthly income below 750 Euro), poor housing conditions (for instance, poor hygiene, lack of central heating, lack of safety), unemployment (in particular, at least one parent being unemployed for more than 6 months), health problems in one or more family members (in particular, physical illness or a history of mental illness), developmental problems in one or more children (for instance, growth or speech delay, mental retardation), and poor parental education (in particular, no secondary education degree). These variables were scored by the social nurses, using a “yes (=1)/no(=0)” format. Annual reports provide information on these variables (e.g., Child & Family, 2001). The scores were summed into a social deprivation scale, going from 0 to 6. Sample of social nurses The risk scale and the social deprivation scale were implemented in a new sample of social nurses. Out of the total population of nearly 600 social nurses, a sample of 40 female nurses was selected randomly. Each selected nurse was asked to follow a training program consisting of four sessions. The program consisted of three parts: (1) the provision of information on disturbed parenting and child maltreatment, (2) the presentation of the scale items, and (3) training sessions, during which the accompanying manual with the scoring rules was studied, general interview and observation skills were practiced, videotapes on normal and disturbed mother-child interactions were rated, and case vignettes of problematic parenting were discussed. The aim of the training sessions was to facilitate administration of the scale and to increase interrater agreement. In the fourth session the social nurses could give feedback on the clinical usefulness of the scale. Next, all selected social nurses were asked to register information on risks for child maltreatment and social deprivation for all births during a 3-month period. The scales had to be completed at the end of the regular home visits, this means after four visits in case of a first child and after three visits in case of a next child. Sample of mothers During the 3-month registration period, the 40 social nurses served 903 mothers with a newborn child. In this sample of births, 391 scales (43.3%) could be fully completed. This meant that for these births the regular home visits ended within the registration period. Within the sample of births for which regular home visits ended within the registration period, a subsample of abusive/neglectful mothers

326

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

(n = 18, or about 2% of the total sample of births) could be identified. Mothers were diagnosed as abusive or neglectful by multidisciplinary Confidential Doctors’ Teams working under the auspices of Child & Family. Confidential Doctors’ Teams consist of a child psychiatrist or pediatrician, a child psychologist or pedagogue, and social workers. They register reports of child maltreatment, examine reported cases and provide treatment. The diagnosis of child physical abuse or neglect was based on the general procedures used by the teams (Child & Family, 2001), not on scores on the risk scale. The social nurses involved in the study were blind to the diagnosis of abuse or neglect. Only the officials of the agency who coordinated the data collection knew about the mothers’ abuse/neglect status. All data were collected from October 1999 to March 2000. Within this period the agency reached 90.3% of all newborn infants in the Flemish Community (Child & Family, 2001). So, a nonresponse of less than 10% can be assumed. The characteristics of the nonabusive (n = 373; 190 boys, 183 girls) and the abusive/neglectful mothers (n = 18; 9 boys, 9 girls) are shown in Table 1. Of the nonabusive mothers, 81.8% had the Belgian nationality. Nonabusive mothers with non-Belgian nationality were from European (e.g., the Netherlands, Italy), African (e.g., Algeria, Congo, Morocco) or Asian (e.g., the Philippines, India) origin. Of the abusive/neglectful mothers, 77.8% had the Belgian nationality. Abusive or neglectful mothers with non-Belgian nationality were from European origin, in particular from the United Kingdom, Spain and former Yugoslavia. The mean duration of pregnancy in the group of abusive/neglectful mothers was

Table 1 Sample characteristics

Sample size Birth order Range M SD

Nonabusive mothers

Abusive/neglectful mothers

373

18

1–7 1.94 .24

1–4 2.17 1.42

Mothers’ age Range (years) M (years) SD (years)

22–44 29.77 5.33

19–41 29.46 4.47

Duration of pregnancya Range (weeks) M (weeks) SD (weeks)

28–42 39.07 3.50

35–40 36.56 1.29

3.15 .43

2.55 .87

.12 .61

1.50 1.98

Birth weight M (kg) SD (kg) Social deprivationa M SD

A significant difference between both groups was found. For duration of pregnancy: t = 3.04, p < .01, df = 17, for social deprivation: t = −2.96, p < .01, df = 17. a

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

327

significantly lower (M = 36.56 weeks) as compared to the nonabusive mothers (M = 39.07 weeks) (t = 3.04, p < .01, df = 17). On the social deprivation scale (Cronbach’s alpha = .85), the nonabusive mothers’ mean score was .12. The abusive/neglectful mothers’ mean score for social deprivation was significantly higher (M = 1.50; t = −2.96, p < .01, df = 17). Although the services provided by Child & Family are voluntary, all mothers gave informed consent for the home visits and the actions undertaken by the social nurse, including for the completion of the risk scale. From the beginning, the study was followed by the agency’s research committee. Each step was prepared in close co-operation with the committee. Statistical analyses The data were analyzed by means of SAS, release 6.12 for Windows (SAS Institute, 1996). A two-step procedure was used to reduce the initial item pool. At first, frequency distributions of risk items were computed. Items with prevalences lower than 5% were omitted. One reason for omission was that, although these items probably measured problems and risks related to child physical abuse and neglect, their low prevalence indicated that they cannot be observed or scored sufficiently within the context of the social nurses’ home visits. Another reason for omission was that, when trying to explain the intercorrelations between items by means of factor analysis, these items were too rarely endorsed to obtain meaningful factor loadings. Secondly, the discriminatory power of the risk items was examined by means of a comparison between the ratings of maltreating and non-maltreating mothers. Risk items for which no significant difference between both groups was found, were omitted. The discriminatory power of the risk items was tested by means of a logistic regression analysis (Freund & Littell, 2000) with two response levels (maltreating vs. non-maltreating), using the logit link function and the maximum likelihood method to estimate parameters. The logistic regression estimated the probability that abuse occurred as related to scores on the risk items. Wald χ2 and odds ratios were used as test statistics to determine the discriminatory power of risk items. An alpha level of .05 was used for all statistical tests. The reliability of the reduced scale was examined by computing (1) the internal consistency (Cronbach’s alpha coefficient) and (2) the level of agreement between raters (Pearson correlation coefficient). In order to test the agreement between raters, three couples of social nurses were formed, each of them working in the same center. They were given instructions to visit mothers simultaneously, to let one of them be the leading home visitor (and switch leadership the next visit), and to complete the scale independent of each other after the home visits. Due to practical and ethical issues, each couple could only visit three mothers. To examine the factor structure underlying the reduced scale, an exploratory factor analysis was performed, using the principal axis method and the mineigen = 1 criterion to extract factors. The latter option caused only factors with eigenvalues greater than 1 to be retained. The maximum likelihood method could not be used, since the assumption of multivariate normal distribution of risk items was severely violated. The initial factor solution was rotated by means of the promax procedure, allowing factors to be intercorrelated. Finally, the concurrent validity of the reduced scale was examined by relating scores for social deprivation to scores on the subscales as derived by factor analysis. Use was made of a stepwise linear unweighted multiple regression analysis, with scores on the subscales as the independent variables and scores on the social deprivation scale as the dependent variable. The alpha level for entry in the regression model was .05.

328

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

Results Reduction of the initial item pool Twenty-four items (33.8%) had prevalence rates lower than 5% and were omitted from the analyses. Next, the discriminatory power of the remaining 47 risk items (Table 2) was tested by a comparison between ratings of maltreating and non-maltreating mothers, using logistic regression analysis. The Wald χ2 and associated odds ratios are presented in Table 2. The estimated odds ratios represent the probability of being classified as maltreating for each unit increase in the risk items. For instance, the estimated odds ratio for the item “Mother does not keep to the appointments regarding home visits and consultations” is 2.80, which means that for each unit increase in this item the odds of being classified as maltreating is increased 2.80 times. Significant Wald χ2 statistics were found for 20 items (Table 2), with odds ratios varying from 1.84 (“Mother has but few contacts outside the family and is dissatisfied with it”) to 6.49 (“I have the feeling that the information mother is giving on how she deals with the baby, is incomplete or does not tally”). Since in the abuse subsample eight children had birthweights lower than 2,500 g, we controlled for the possibility that any differences between both subsamples that were found were reflecting birthweight rather than the abuse/neglect status difference. At first, birthweights were dichotomized (<2,500 g vs. ≥2,500 g). Next, a univariate ANOVA was used to test the effects of birthweight on the 20-item scale score. No significant effect was found [F(1, 378) = .05, p = .82]. Finally, after having omitted all infants with birthweights lower than 2,500 g, we tested the discriminatory power of the 20 items by comparing ratings of maltreating (n = 10) and non-maltreating mothers (n = 343), using logistic regression analysis. Significant Wald χ2 statistics were found for all items, except for the item “Mother has but few contacts outside the family and is dissatisfied with it” (Wald χ2 = 2.96, p = .0851). Reliability of the reduced risk scale The utility of the risk scale was examined during the different phases of construction. The bottom-up approach of scale construction meant that the social nurses had the opportunity during each training session to give feedback on the scale content, the item formulation and the specification of the guidelines written down in the manual (Hellinckx et al., 1999). After each training session, the draft versions of the scale were adapted, taking into account the nurses’ comments. The internal consistency of the reduced 20-item scale was high (α = .92), as was the interrater reliability (across all items r = .97). Factor structure of the reduced risk scale Exploratory factor analysis produced a three-factor solution (eigenvalue of the unreduced correlation matrix for Factor 1 = 7.57, eigenvalue Factor 2 = 2.30, eigenvalue Factor 3 = 1.02). A first factor explained 62.17% of the common variance between items, a second factor 18.86%, and a third factor 8.41%. Promax rotation of the prerotated solution yielded a simple structure with eight items correlating .30 or higher—taking into account the sample size, this appeared to be the most appropriate criterion to determine salient correlations (Stevens, 1996)—on Factor 1, eight items correlating .30 or higher on Factor 2, and six items correlating .30 or higher on Factor 3. Table 3 shows the reference structure

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

329

Table 2 Comparison of ratings on risk items between maltreating and non-maltreating mothers Risk item Interaction between social nurse and mother Mother does not keep to the appointments regarding home visits and consultations It is difficult to get in touch with this mother Mother does not take advice, or only in part, on how to take care of the baby I believe mother is giving incorrect information on the baby’s behavior and development There is an atmosphere of secrecy in this family I feel uncomfortable in this family I have a feeling that the information mother is giving on how she deals with the baby, is incomplete or does not tally This family does not give much elbowroom Mother-child interaction Mother expects the baby to behave conveniently Mother expects the baby to give abundant love Mother believes this baby will be perfect Pregnancy and delivery are very negatively spoken of Taking care of the baby gives rise to tension and nervousness Taking care of the baby is considered to be a nuisance There is no warm bodily contact between mother and baby Mother considers the baby to be a restraint of her freedom of movement The baby is taken care of perfunctory, without much emotional involvement Mother does not give evidence of taking pleasure in the baby Mother and baby do not have much eye contact The way the baby is spoken about, does not betray much love Mother does not wonder what may be the reason for baby’s behavior Mother does not check her behavior’s effect on the baby Mother does not play with the baby Mother deals with the baby in consideration of general beliefs on how to handle babies and does not take into account that this child is peculiar Mother is easily put out of patience when dealing with the baby There is no consideration for baby’s feelings Mother’s reaction to baby’s behavior is very unstable Mother sets few limits and does not give much structure to baby’s life Mother speaks often about herself and not about the baby There is not much reaction to baby’s signs for social contact (crowing, smiling) Maternal and family characteristics Mother intimates that as a child she did not get much love from her parents or family Mother has already gone through several crises and it seems that she has difficulties in getting over it Mother is dissatisfied with contacts with family/friends Mother intimates that she is alone facing the problems There is not much support from the partner Mother is not able to adequately seek help or support Mother has but few contacts outside the family and is dissatisfied with it In a stress situation mother soon turns out to be helpless Mother intimates that she feels unhappy Mother does not show much self-confidence

Wald χ2

Odds ratio

18.34∗ 3.52 14.71∗ 3.26 10.13∗ 8.67∗ 11.27∗

2.80 1.97 3.66 2.13 3.45 3.08 6.49

.01

1.05

.81 4.12∗ <.01 7.81∗ .57 3.75 <.01 .28 .02 .01 <.01 <.01 3.81 .70 1.20 .47

1.58 2.32 <.01 2.81 1.47 2.45 .95 1.42 .89 1.11 1.04 <.01 3.14 1.74 1.51 1.52

.01 <.01 1.37 10.25∗ 5.68∗ <.01

1.09 <.01 1.94 2.71 2.31 .96

7.75∗ 20.91∗

2.42 2.98

5.86∗ 13.99∗ 17.55∗ 7.85∗ 4.17∗ 7.28∗ 11.36∗ 4.69∗

1.94 2.38 2.67 2.97 1.84 2.53 2.47 2.00

330

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

Table 2 (Continued ) Risk item Mother comes across as listless Mother is touchy and irritable Mother has gloomy expectations Mother easily looses self-control Child characteristics When baby cries, nobody is able to comfort it In my own opinion this is a “difficult” baby This baby cries an unusual lot ∗

Wald χ2

Odds ratio

3.76 .17 6.87∗ <.01

1.99 1.25 2.60 1.02

.21 2.08 .06

1.32 2.14 .81

p < .05.

Table 3 Semipartial correlations ≥.30 between the risk items and the three factors (after promax rotation of the initial solution)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Risk item

Factor 1: Isolation

Mother intimates that she is alone facing the problems Mother has but few contacts outside the family and is dissatisfied with it Mother is not able to adequately seek help or support Mother is dissatisfied with contacts with family/friends There is not much support from the partner Mother has gloomy expectations Mother intimates that she feels unhappy Pregnancy and delivery are very negatively spoken of Mother expects the baby to give abundant love Mother speaks often about herself and not about the baby In a stress situation mother soon turns out to be helpless Mother has already gone through several crises and it seems that she has difficulties in getting over it Mother does not show much self-confidence Mother intimates that as a child she did not get much love from her mother or family I have a feeling that the information mother is giving on how she deals with the baby, is incomplete or does not tally Mother does not keep to the appointments regarding home visits and consultations There is an atmosphere of secrecy in this family I feel uncomfortable in this family Mother does not take advice, or only in part, on how to take care of the baby Mother sets few limits and does not give much structure to baby’s life

.65 .63

Factor 2: Psychological complexity

Factor 3: Communication problems

.53 .52 .48 .44 .40 .30 .64 .59 .54 .50 .46 .44 .71 .70 .63 .63 .59 .46

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

331

of the rotated solution, presenting the salient semipartial correlations between risk items and factors. All items had salient semipartial correlations with at least one factor. There were two items correlating above .30 on more than one factor: Item 20 “Mother sets few limits and does not give much structure to baby’s life” (semipartial correlation with Factor 2 = .37, with Factor 3 = .46) and Item 7 “Mother intimates that she feels unhappy” (semipartial correlation with Factor 1 = .40, with Factor 2 = .34). As could be expected, the intercorrelations between all factors were substantially high (p < .01): .54 between Factor 1 and Factor 2, .42 between Factor 1 and Factor 3, and .24 between Factor 2 and Factor 3. Factor 1 was named “Isolation,” since most items correlating high with this factor measured perceived problems with regard to social relationships, for instance Item 1 “Mother intimates that she is alone facing the problems” (.65), Item 2 “Mother has but few contacts outside the family and is dissatisfied with it” (.63), Item 3 “Mother is not able to adequately seek help or support” (.53) and Item 4 “Mother is dissatisfied with contacts with family/friends” (semipartial correlation = .52). Factor 2 was named “Psychological complexity” since most items correlating high with this factor measured maternal psychological problems which could foster a mismatch between mother and baby, for instance Item 9 “Mother expects the baby to give abundant love” (.64), Item 10 “Mother speaks often about herself and not about the baby” (.59), Item 11 “In a stress situation mother soon turns out to be helpless” (semipartial correlation = .54) and Item 12 “Mother has already gone through several crises and it seems that she has difficulties in getting over it” (.50). Factor 3 was named “Communication problems” since most items correlating high with this factor measured problems in mothers’ communication with the outer world, for instance Item 15 “I have a feeling that the information mother is giving on how she deals with the baby, is incomplete or does not tally” (.71), Item 16 “Mother does not keep to the appointments regarding home visits and consultations” (.70), Item 17 “There is an atmosphere of secrecy in this family (semipartial correlation = .63) and Item 18 “I feel uncomfortable in this family” (.63). Based on the outcome of the exploratory factor analysis, three subscales (Isolation, Psychological complexity, Communication problems) were constructed. Each subscale contained items correlating at least .30 with a factor. In case two correlations were above .30, the highest correlation was taken into account when assigning the item to a subscale. Item 7 “Mother intimates that she feels unhappy” was assigned to the subscale Isolation, Item 20 “Mother sets few limits and does not give much structure to baby’s life” to the subscale Communication problems. All subscales showed high internal consistency (Cronbach’s alpha = .89 for Isolation, .83 for Psychological complexity, and .82 for Communication problems). Concurrent validity of the reduced risk scale Scores of non-maltreating mothers on Isolation (M = 1.01, SD = 2.87), Psychological complexity (M = .65, SD = 1.88) and Communication problems (M = .68, SD = 1.77) were entered stepwise as independent variables in the linear unweighted multiple regression analysis. Scores on Isolation significantly predicted scores on the social deprivation scale [F(2, 370) = 36.33, p < .01, R2 = .05], whereas scores on Psychological complexity did not [F(3, 369) = 2.68, ns, R2 < .01]. Scores on Communication problems also significantly predicted scores on the social deprivation scale [F(1, 371) = 246.84, p < .01, R2 = .39] (Table 4). The total adjusted multiple correlation squared was .44. Standardized regression coefficients were .50 for Communication problems and .20 for Isolation, with both corresponding t values being significant at the .01 level. Mothers scoring high on Communication problems and Isolation were scoring high on the social deprivation scale.

332

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

Table 4 Linear unweighted multiple regression results for predicting scores on the social deprivation scale

Isolation Psychological complexity Communication problems ∗

Standardized regression coefficient

t value

R2

.20 .09 .50

3.55∗ 1.64 11.53∗

.05 <.01 .39

p < .01.

Discussion Significance of the findings Professionals working with families with newborn children need user-friendly, reliable and valid scales to assess risks for child abuse and neglect. Such scales can set the basis for well-considered decision-making with regard to early intervention and prevention of future harm in families at-risk for child maltreatment. The theoretical underpinnings of the scale presented in this study can be found in ecological, psychological and pedagogical views on the etiology of child physical abuse and neglect (Baartman, 1996; Belsky, 1997; Belsky & Vondra, 1989; Cerezo, 1997; Milner, 1993), all of which are strongly supported by empirical findings (see for reviews, Ammerman, 1990; Rogosch et al., 1995). Tailoring the scale to the professionals’ needs and working contexts was realized by using a bottom–up approach for scale construction. The items were based on direct information obtained from practitioners. This offered us the possibility to match the information from the literature with the “in vivo” narratives of practitioners, to use the words and phrases of focus group members when formulating scale items, and to let them give comments on the comprehensibility and the clinical usefulness of draft versions. The final scale contained 20 items and showed high internal consistency and interrater agreement. All items had prevalence rates higher than 5% and significantly estimated the probability that abuse occurred. Including a sample of substantiated cases of abuse or neglect is considered to be a very tough criterion to test the discriminative power of items used for child maltreatment evaluations (Milner et al., 1998). Exploratory factor analysis revealed a simple structure with three interrelated factors: Isolation, Psychological complexity, and Communication problems. The factor solution almost equalled a simple structure and stressed the major issues that may raise concerns in home visiting nurses evaluating early risks of physical abuse and neglect in families with a newborn child. Maternal isolation may be a primary reason for concern. It may appear to the social nurse either in a direct way (for instance, when mothers complain that they received no help or support after the baby’s birth, when they acknowledge to be dissatisfied with relationships with family or relatives) or in an indirect way (e.g., when mothers intimate that they are unhappy, show a lack of help-seeking behavior, or speak only in negative terms about pregnancy and delivery). The maternal psychological condition may be a second reason for concern. Signs that may raise concerns are distorted expectations about the child, a lack of self-confidence, a lack of coping mechanisms and problem-solving strategies, self-oriented speech, a lack of loving relationships during childhood, and not getting over psychological crises. A third reason for concern may be mother’s style of communication. Analyzing patterns of communication starts by analyzing the interactions during home visits. Elements to be focused on are the atmosphere in the family during the home visit, mother’s way of dealing with appointments and advices on how to take care of the baby, and the information mother gives

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

333

on her relationship with the baby. Alternatively, the third factor may be interpreted as a defensiveness factor, reflecting a stance of trying to hide from the nurse rather than being open with her. Indeed, mothers who during the birth to 3-month registration period were diagnosed as abusive or neglectful, may have become defensive and suspicious with regard to the home visits. Items measuring child characteristics (e.g., difficult temperament, crying infant) showed little or no discriminative power and were not retained in the final scale. This seems to be somewhat contrary to findings from empirical studies stressing the active contribution of children to abusive or neglectful interactions with parents and caregivers (Ammerman, 1990). One reason for the lack of discriminative power of these characteristics may be the children’s age (0 to 3 months). It is plausible to assume that infant temperament and crying behavior play a more prominent role as active contributors to parent-child interactions at later ages, for instance during the second half of the first year of life. Scores on a social deprivation scale were significantly predicted by the subscales Communication problems and Isolation, but not by the subscale Psychological complexity. It was not surprising that the subscale Communication problems was the most significant predictor of social deprivation. Socially deprived mothers at-risk of child physical abuse and neglect are often described as displaying distorted communication patterns and having troublesome relationships with professionals (Garbarino, 1997). Neither was it surprising that scores on the subscale Isolation significantly added to the prediction of social deprivation scores. Many studies found a lack of support from partner, family and relatives to be a significant precursor to child physical abuse and neglect (Baartman, 1996; Cerezo, 1997; Erickson & Egeland, 1996; Milner, 1993). The lack of significant predictive power of the subscale Psychological complexity was contrary to empirical evidence pointing to parental personality characteristics as precursors of physical child abuse and neglect (see for reviews Baartman, 1996; Milner & Dopke, 1997). It seems plausible that the items measure aspects of child abuse risk potential which are not covered directly by the social deprivation scale, for instance parental awareness (e.g., Item 9 “Mother expects the baby to give abundant love” and Item 10 “Mother speaks often about herself and not about the baby”) or feelings of being unloved and a lack of warm relationships during childhood (e.g., Item 14 “Mother intimates that as a child she did not get much love from her mother or family”). The scale measures perinatal experiences (e.g., psychological well-being, social support) and parenting issues (e.g., maternal expectations, limit-setting). Unlike other risk scales for home visiting nurses (e.g., Browne, 1995), it does not measure static risk factors (e.g., age of mother, child with a handicap). Although comparative research on the predictive validity of scales focusing on static risk factors and the present scale is necessary, we believe that combining these scales will increase the accuracy of screening procedures and be helpful to identify families in need of intervention. The present study has some limitations. At first, the number of maltreating mothers in our sample was rather small, although all cases included were substantiated cases, diagnosed by multidisciplinary teams independent from scale scores. Second, due to practical and ethical problems, examination of the scale’s interrater agreement was limited to a small sample. It was very difficult to organize home visits with two social nurses. Most mothers knew well the agency’s procedure. Announcing a visit with two nurses made them feel anxious, distressed, or suspicious, regardless of the nurses’ explanation. More research on the interrater agreement of the scale is needed, however. Third, future research should focus on the construct validity of the scale by using other measures than social deprivation and on the scale’s sensitivity and specificity in a new sample of maltreating and non-maltreating mothers. Fourth, little is known about the long-term predictive validity of the scale, the study of which is a major challenge to all researchers developing instruments for child maltreatment evaluations (Milner et al., 1998). Finally, father reports

334

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

should be collected in order to increase the possibilities for home visitors to identify risks for physical abuse and neglect in families with newborn children. Implications Developing user-friendly and evidence-based risk assessment procedures is a first but crucial step towards the preparation of successful interventions in at-risk families and towards the planning of successful preventive actions (Baartman, 1996; Milner et al., 1998; Wald & Woolverton, 1990). Implementing these procedures offers many possibilities for enhancing the professionalism of all people involved in the prevention of child physical abuse and neglect. Furthermore, it may reduce the risks for misidentifying potential maltreaters as well as the labeling effects misidentification may have for children and parents. In the present study, three issues home visitors should focus on when evaluating early risks of physical abuse and neglect were identified, using a bottom–up approach: maternal isolation, psychological problems, and distorted communication with professionals. This approach allowed the development of a training program on parenting support and prevention of child maltreatment, aimed at enhancing professionalism with regard to risk assessment and the delivery of preventive services by home visiting nurses in families with newborn infants. Item scoring, interpreting scores and preparing interventions based on the scores (e.g., helping mothers seek social support in their neighborhood or family) are core elements of the program. References Ammerman, R. T. (1990). Predisposing child factors. In R. T. Ammerman & M. Hersen (Eds.), Children at risk. An evaluation of factors contributing to child abuse and neglect (pp. 199–221). New York: Plenum Press. Baartman, H. (1996). Opvoeden kan zeer doen. Over oorzaken van kindermishandeling, hulpverlening en preventie [Raising children can hurt. About causes of child maltreatment, treatment and prevention]. Utrecht: SWP. Becker-Lausen, E., & Mallon-Kraft, S. (1997). Pandemic outcomes. The intimacy variable. In J. Jasinski & G. Kaufman Kantor (Eds.), Out of the darkness: Contemporary perspectives on family violence (pp. 49–57). Thousand Oaks, CA: Sage Publications. Belsky, J. (1997). Determinants and consequences of parenting: Illustrative findings and basic principles. In W. Hellinckx, M. Colton, & M. Williams (Eds.), International perpectives on family support (pp. 1–21). Aldershot: Arena, Ashgate Publishing Limited. Belsky, J., & Vondra, J. (1989). Lessons from child abuse: The determinants of parenting. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment. Theory and research on the causes and consequences of child abuse and neglect (pp. 153–202). Cambridge: Cambridge University Press. Browne, K. (1995). Preventing child maltreatment through community nursing. Journal of Advanced Nursing, 21, 57–63. Browne, K., Davies, C., & Stratton, P. (Eds.). (1988). Early prediction and prevention of child abuse. Chichester: John Wiley & Sons. Browne, K., & Herbert, M. (1997). Preventing family violence. Chichester: John Wiley & Sons. Bugental, D. B., Mantyla, S. M., & Lewis, J. (1989). Parental attributions as moderators of affective communication to children at risk for physical abuse. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment. Theory and research on the causes and consequences of child abuse and neglect (pp. 254–279). Cambridge: Cambridge University Press. Cantos, A. L., Neale, J. M., O’Leary, K. D., & Gaines, R. W. (1997). Assessment of coping strategies of child abusing mothers. Child Abuse & Neglect, 21, 631–636. Casanova, G. M., Domanic, J., McCanne, T. R., & Milner, J. S. (1992). Physiological responses to non-child related stressors in mothers at risk for child abuse. Child Abuse & Neglect, 16, 31–44. Cerezo, M. (1997). Abusive family interaction: A review. Agression and Violent Behavior, 8, 215–240.

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

335

Child & Family. (2001). The child in Flanders. Brussels: Kind & Gezin. Coohey, C., & Braun, N. (1997). Toward an integrated framework for understanding child physical abuse. Child Abuse & Neglect, 21, 1081–1094. Crittenden, P. M. (1988). Family and dyadic patterns of functioning in maltreating families. In K. Browne, C. Davies, & P. Stratton (Eds.), Early prediction and prevention of child abuse (pp. 161–189). New York: Wiley. Erickson, M., & Egeland, B. (1996). Child neglect. In J. Briere, L. Berliner, J. Bulkley, C. Jenny, & T. Reid (Eds.), The APSAC handbook on child maltreatment (pp. 4–20). Thousand Oaks, CA: Sage Publications. Farnell, M. B. (1980). Screening procedures for the detection or prediction of child abuse: An annotated bibliography. Canada: Centre of Criminology, University of Toronto. Freund, R. J., & Littell, R. C. (2000). SAS system for regression (3th ed.). Cary, NC: SAS Institute, Inc. Garbarino, J. (1997). The role of economic deprivation in the social context of child maltreatment. In M. E. Helfer, R. S. Kempe, & R. D. Krugman (Eds.), The battered child (5th ed., pp. 49–60). Chicago: The University of Chicago Press. Grietens, H., & Hellinckx, W. (2003). Predicting disturbed parental awareness in mothers with a newborn infant: Test of a theoretical model. Infant & Child Development, 12, 117–128. Grietens, H., Hellinckx, W., Van Assche, V., Baartman, H., & Geeraert, L. (1999). Rationale and development of a screening instrument for social nurses to identify risks on child abuse and neglect. International Journal of Child and Family Welfare, 4, 24–41. Harkness, S., & Super, C. (1995). Culture and parenting. In M. H. Bornstein (Ed.), Handbook of parenting (Vol. 1, pp. 211–234). Mahwah, NJ: Lawrence Erlbaum Associates. Harrington, D., Black, M. M., Starr, R. H., & Dubowitz, H. (1998). Child neglect: Relation to child temperament and family context. American Journal of Orthopsychiatry, 68, 108–116. Hellinckx, W., Baartman, H., Grietens, H., Van Assche, V., & Geeraert, L. (1999). Handleiding bij de schaal “Problematische opvoedingssituaties, met risico op kindermishandeling en verwaarlozing” [Manual accompanying the Early Risks of Physical Abuse and Neglect scale]. Leuven: Authors. Houldin, A. D. (1987). Infant temperament and the quality of the childrearing environment. Maternal-Child Nursing Journal, 16, 131–143. Kapitanoff, S. H., Lutzker, J. R., & Bigelow, K. M. (2000). Cultural issues in the relation between child disabilities and child abuse. Aggression and Violent Behavior, 5, 227–244. Kaufman, J., & Zigler, E. (1989). The intergenerational transmission of child abuse. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment. Theory and research on the causes and consequences of child abuse and neglect (pp. 129–150). Cambridge: Cambridge University Press. Knutson, J. F. (1995). Psychological characteristics of maltreated children. Putative risk factors and consequences. Annual Review of Psychology, 46, 401–432. Kolko, D. (1996). Physical abuse. In J. Briere, L. Berliner, J. Bulkley, C. Jenny, & T. Reid (Eds.), The APSAC handbook on child maltreatment (pp. 21–50). Thousand Oaks, CA: Sage Publications. Korbin, J. E. (1997). Culture and child maltreatment. In M. E. Helfer, R. S. Kempe, & R. D. Krugman (Eds.), The battered child (5th ed., pp. 29–48). Chicago: The University of Chicago Press. Lutzker, J. R., Van Hasselt, V. B., Bigelow, K. M., Greene, B. F., & Kessler, M. L. (1998). Child abuse and neglect: Behavioral research, treatment, and theory. Aggression and Violent Behavior, 3, 181–196. Milner, J. S. (1986). The child abuse potential inventory. Manual (2nd ed.). DeKalb, IL: Psytec, Inc. Milner, J. S. (1993). Social information processing and physical child abuse. Clinical Psychology Review, 13, 275–294. Milner, J. S., & Dopke, C. (1997). Child physical abuse: Review of offender characteristics. In D. A. Wolfe, R. McMahon, & R. dev. Peters (Eds.), Child abuse: New directions in prevention and treatment across the life span (pp. 25–52). Thousand Oaks, CA: Sage Publications. Milner, J. S., Murphy, W. D., Valle, L. A., & Tolliver, R. M. (1998). Assessment issues in child abuse evaluations. In J. R. Lutzker (Ed.), Handbook of child abuse research and treatment (pp. 73–115). New York: Plenum Press. Morgan, D. L. (1993). (Ed.). Successful focus groups. Advancing the state of the art. Newbury Park, CA: Sage Publications. Newberger, C. (1980). The cognitive structure of parenthood. The development of a cognitive measure. In R. L. Selman & R. Yando (Eds.), Clinical-developmental psychology. New directions of child development: Clinical developmental research (No. 7). San Francisco: Jossey-Bass. Pecora, P. J. (1991). Investigating allegations in child maltreatment: The strengths and limitations of current risk assessment systems. Child and Youth Services, 15, 73–92.

336

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

Rodwell, M. K., & Chambers, D. E. (1992). Primary prevention of child abuse: Is it really possible? Journal of Sociology & Social Welfare, 19, 159–176. Rogosch, F., Cicchetti, D., Shields, S., & Toth, S. (1995). Parenting dysfunction in child maltreatment. In M. H. Bornstein (Ed.), Handbook of parenting (Vol. 4, pp. 127–159). Mahwah, NJ: Lawrence Erlbaum Associates. SAS Institute. (1996). SAS® language and procedures: Introduction, version 6 (1st ed.). Cary, NC: SAS Institute, Inc. Starr, R. H. (1982). (Ed.). Child abuse prediction. Policy implications. Cambridge, MA: Ballinger Publishing Company. Stevens, J. (1996). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Wald, M. S., & Woolverton, M. (1990). Risk assessment: The emperor’s new clothes? Child Welfare, 69, 483–511. Zuravin, S., McMillan, C., DePanfilis, D., & Risley-Curtiss, C. (1996). The intergenerational cycle of child maltreatment. Continuity versus discontinuity. Journal of Interpersonal Violence, 11, 315–334.

Résumé Objectif: Le but était de construire et de tester la fiabilité (utilité, cohérence interne, l’accord entre estimateurs) et la validité (validité interne, validité concomitante) d’une échelle destinée aux infirmières visiteuses à domicile afin d’évaluer les risques de sévices physiques, et de négligence chez les mères ayant un nouveau-né. Méthode: On a construit une échelle de 71 items à partir d”une revue de la littérature et de rencontres de groupes centrées sur un thème réunissant des infirmières sociales et des paraprofessionnels ayant l’expérience des familles démunies. Cette échelle a été appliquée sur un échantillon choisi au hasard de 40 infirmières visiteuses à domicile qui ont récolté des données auprès d’un échantillon de 373 mères d’enfants nouveaux-nés non-maltraitantes et de 18 mères d’enfants nouveaux-nés maltraitantes et négligentes. Résultats: On a négligé les items avec un taux de prévalence inférieur à 5% et les items qui n’établissaient pas de différence entre les mères maltraitantes et non-maltraitantes. La version finale comportait 20 items. Cette échelle montrait une convergence interne élevée (= .92) ainsi qu’une fiabilité élevée entre les estimateurs (= .97). L’examen par une analyse factorielle a donné une solution à trois facteurs: Isolement (8 items, expliquant 62.17% de la variance commune); complexité psychologique (6 items, 18.86%) et problèmes de communication (6 items, 8.41%). Les scores des problèmes de communication et de l’isolement prédisaient significativement les scores d’une échelle de carence sociale qui distinguait significativement les mères maltraitantes des non-maltraitantes. Les mères qui avaient des scores élevés aux problèmes de communication ou d’isolement ont obtenu des scores plus élevés pour la carence sociale que les mères qui avaient les scores les moins élevés. Conclusion: Les infirmières visiteuses à domicile peuvent identifier les risques de mauvais traitements physiques et de négligence chez les mères d’enfants nouveaux-nés en se focalisant sur les signes d’isolement social, de communication troublée et sur les problèmes psychologiques.

Resumen Objetivo: El objetivo fue construir y comprobar la confiabilidad (utilidad, consistencia interna, acuerdo entre pruebas) y la validez (validez interna, validez concurrente) de una escala, para que las enfermeras sociales visitadoras del hogar pudieran identificar los riesgos de abuso f´ısico y negligencia en madres con un niño recién nacido.

H. Grietens et al. / Child Abuse & Neglect 28 (2004) 321–337

337

Método: Se construyó una escala de 71 items basada en una revisión de la literatura y sesiones de grupos focales con enfermeras sociales y para-profesionales que ten´ıan experiencia con familias marginadas. Esta escala fue aplicada en una muestra al azar de 40 enfermeras sociales visitadoras del hogar, quienes recog´ıan datos en una muestra de 373 madres no abusivas y 18 madres abusivas/negligentes con un niño recién nacido. Resultados: Los ´ıtems con puntajes de prevalencia por debajo de 5% y los ´ıtems que no arrojaron una diferencia significativa entre las madres maltratantes y las no maltratantes fueron omitidos. La versión final conten´ıa 20 items. Esta escala demostró una alta consistencia interna (a = .92) y una alta confiabilidad entre las pruebas (r = .97). El análisis del factor exploratorio arrojó una solución de tres factores: aislamiento (8 items, relacionados con el 62.17% de la varianza común); complejidad psicológica (6 items, 8.41%). Los puntajes sobre problemas de Comunicación y Aislamiento predijeron significativamente puntajes en una escala de privación social, que discriminó significativamente las madres maltratantes de las no maltratantes. Las madres que obten´ıan puntajes elevados en problemas de Comunicación o Aislamiento obten´ıan puntajes más elevados en privación social que las madres con puntajes bajos. Conclusión: Las enfermeras visitadoras del hogar pueden identificar los riesgos del abuso f´ısico y la negligencia en las madres con recién nacidos al observar signos de aislamiento social, comunicación distorsionada y problemas psicológicos.