A Multi-method Assessment of Risk and Protective Factors in Family Violence: Comparing Italian and Migrant Families

Child maltreatment is a major public health issue, which threatens psychophysical well-being of almost 100,000 minors in Italy. The prevalence and social burden of this issue clearly indicate the need for preventive efforts, early identification of victimized children, and a rapid risk assessment of families referred to Child Protection Services (CPS). The present chapter proposes an evidence-based multi-method assessment system, specific for highlighting risk and protective factors in family functioning. The assessment system is comprised of six modules: 
 
 
One in-depth protocol for identifying proximal and distal risk and protective factors of maltreatment and recidivism against children. 
 
 
One exploratory tool for assessing the residual social capital in an at-risk family. 
 
 
Four screening surveys for assessing parenting stress, parental child abuse potential, parents’ resilience, and traumatic symptoms in children.


Introduction
Child abuse and maltreatment are serious developmental adversities in the short and long term for children and adolescents. International reviews highlight the enormous developmental, social, and economic costs of child abuse and maltreatment (Autorità Garante per l'Infanzia e l'Adolescenza, CISMAI, & Terre des Hommes, 2015; Bunting et al., 2018;CISMAI, Terres des Hommes, & Bocconi, 2014;Di Blasio, 2000Di Blasio, Camisasca, & Procaccia, 2007;Fry et al., 2018;Gallo, Munhoz, de Mola, & Murray, 2018;Kimber, Adham, Gill, McTavish, & MacMillan, 2018). Research has led to improvements in knowledge about intervention strategies to buffer negative consequences of child maltreatment and abuse (cf. Masten, 2001;Masten, Cutuli, Herbers, & Reed, 2009;Ozer, Best, Lipsey, & Weiss, 2003). In this light, the concept of resilience has proved to be very useful to describe those developmental trajectories that lead to positive adaptation notwithstanding serious adverse conditions (Masten et al., 2009). Resilience is a key construct for scholars and professionals in the field of Child Protection Services (CPS), as it is the very foundation of the possibility of changing maladaptive developmental trajectories into more positive ones. Thus, professionals and practitioners in CPS are often required to make decisions about the best intervention to protect minors at risk, often with limited time and resources to collect the information needed for such a resolution.
Moreover, CPS professionals often deal with referred families that fall into "gray areas" of family functioning: they appear to be not completely dysfunctional -nor are they completely functional -but are characterized by (acute and/or chronic) difficulties and adversities that impact day-by-day functioning and parenting. In these instances, it is very important to assess the risk of harm/maltreatment in a timely manner, in order to implement effective protective measures and to avoid recidivisms. A key element in the decision-making process is for CPS to be able to rapidly discriminate situations that need a more decisive plan of intervention (e.g., removing the child from the family) from those that require a more cautious approach. However, Italian professionals cannot rely on national documents and manuals on assessment and decision-making process when dealing with at-risk families but only on local guidelines and on their professionalism and training. As a result, in many instances, decisions are made through intuition and not informed by evidentiary information.
Migrant families are often a challenge to CPS personnel because of the difficulty in gathering relevant information, assessing the residual capital in family resources (both from a material and social point of view) and gauging the potential for repairing/enhancing family functioning (cf. Milani, Grumi, Gagliardi, & Di Blasio, 2016). Moreover, some studies have highlighted that migrant families have both a higher risk of referral to CPS and risk of maltreatment (cf. Alink, Euser, van Ijzendoorn, & Bakermans-Kranenburg, 2013).
Our Research Centre is trying to provide an assessment protocol and other useful tools to assist professionals and operators in CPS and in the judicial system in the complex decisions they face when working with at-risk families. To our knowledge, this is the first attempt at proposing a specific assessment system for this purpose in Italy, given the fact that -among the limitations previously cited -the Italian CPS system is still lacking a unified registry about allegations and management of child maltreatment. The lack of uniformity in assessment protocols hinders the possibility of comparison between methods of assessment and case management across different locals.
As concerns the international literature, research on decision-making systems in CPS is steadily shifting toward the aim of helping professionals and operators rely less on intuition and more on objective indicators when planning interventions (cf. Bartelink, van Yperen, & ten Berge, 2015, for a review). However, there is a paucity of objective indicators with the exception of the "California Family Risk Assessment": an instrument designed to help practitioners in the decision-making about at-risk children and families, focusing on the strengths and needs of caregivers, children, and family resources, via different checklists and guidelines 1 (Barlow, Fisher, & Jones, 2012).
Our assessment system is conceptually similar -albeit less detailed -than the British Framework for the Assessment of Children in Need and their Families (FACNF), a set of guidelines and indications aimed at helping CPS professionals to assess complex situations of children and families (cf. Léveillé & Chamberland, 2010). To assist professionals in this task, our framework contains some standardized checklists focusing not only on risk assessment but also on strengths and resources. We believe our assessment protocol meets operators' and professionals' requirements in terms of comprehensiveness, modularity (e.g., selecting single or multiple modules to focus on specific areas of assessment), and scientific validity.
In particular, the system presented herein can be defined as a risk assessment tool (cf. Bartelink et al., 2015) based on an array of psychological instruments and measures that can be used to offer an accurate assessment of at-risk families. It is grounded in four pillars: (a) a process-oriented model; (b) clinical expertise of our Research Centre 2 in terms of child maltreatment and abuse; (c) extensive research efforts to validate and refine the assessment system; and (d) constant interaction with field knowledge (CPS professionals and Italian judicial system).

The Process-Oriented Model
Our assessment system is grounded in the "process-oriented model" of developmental trajectories and children adjustment (cf. Cummings, Davies, & Campbell, 2000), which posits that human development is characterized by mutual influences between different factors and environments, which lead to adaptive or potentially maladaptive patterns. This model has been modified and articulated to fit the clinical literature on risk assessment in CPS work (cf. Cirillo & Di Blasio, 1989;Di Blasio, 1997;. The final model is presented in Fig. 1 The left side of the model highlights the contribution of genetic, biological, and psychological factors of individuals, which interact with familial functioning and environmental affordances into determining parental attitudes toward offspring. The central part of the model focuses on the "here-and-now" psychological functioning that mediates between contextual factors and adaptive/maladaptive trajectories. The right side of the model presents the outcomes of this process: social competence on one side and maladjustment on the other. In our vision, social competence can be thought as a developmental outcome related to those families where there is a preponderance of proximal protective factors (PPF). In these instances, parents are able to express positive parenting and to buffer the impact of eventual negative events (e.g., loss of job, death of a relative) by tapping into personal and contextual resources.
On the other hand, maladaptive outcomes can be characterized by those instances when parents are not able to buffer the effects of negative events -or of personal difficulties and inabilities -and expose their children to situations of vulnerability or clear harm via a significant prevalence of distal and proximal risk factors (DRF/ PRF). These can take the form of a sporadic, infrequent parental inadequacies that last for a limited period of time and can be addressed by "lower-impact" interventions, such as parental monitoring or training, or can take the form of a chronic impairment that poses a serious threat to children and that needs to be tackled with more impactful interventions such as limiting parental responsibility or removing the child from the family.

Distal and Proximal Factors
In our opinion, the most appropriate way to conceptualize and to address potentially maltreating parents and sub-par parenting skills is an examination of risk and protective factors (cf. Camisasca & Di Blasio, 2002;Di Blasio, 2000;Di Blasio, Camisasca, Procaccia, & Verrocchio, 2005). In this light, high-risk conditions are characterized by (chronic or acute) exposure to adversities that can slowly (or swiftly, in case of traumatic events) erode, consume, and ultimately deplete physical and psychological resources of individuals and families. However, high-risk conditions are not necessarily equivalent to maladaptive outcomes. In fact, it is often difficult to assign a "clear-cut" score of damaging potential to many life events, due to the complexity of the mutual interaction of the factors in play. Thus, to better discriminate those factors that exert a direct influence on developmental trajectories in at-risk families and those that are indirect, we refer to the distinction by Baldwin, Baldwin, & Cole (1990) between distal and proximal factors. The term "distal" is used because these factors are supposed to have an indirect effect on children and can be thought of as a humus on which more proximate events and factors build their influence. Proximal factors have a direct influence on family processes and can overturn frail points of balance, often exacerbating covert tensions and conflicts. Distal risk factors can lead a given family to a condition of vulnerability; however their mere presence does not equate to harm or damage to children. Nonetheless, when distal risk factors are intertwined with proximal risk factors, the situation is very likely to evolve into a clear harm for children. For example, literature shows that maternal lack of knowledge or interest relating to child development (distal risk factor) is associated with worse health outcomes in children due to difficulties in understanding medical prescriptions. However, a lack of knowledge or interest alone is seldom related to maltreatment; other proximal factors such as family conflict and difficult temperaments in children need to be present in order to generate harm. It should be noted that proximal factors include both risk and protective factors. Protective factors are those proximal instances and resources that have the power -if recognized and adequately taken into consideration when implementing an intervention -to reduce the negative impact of distal and proximal risk factors. When a protective factor comes into play, if services are able to turn it into a resilience process via work with the family, it is possible that a maladaptive trajectory's likely negative outcomes will be buffered and will not cause harm to the child.

C.Ri.d.e.e. Multi-method Assessment System
In our experience, in order to sustain and mobilize residual resources in families at risk, it is important to not only focus on risk factors but also on personal, familial, and social strengths. Our proposal of a multi-method assessment system is specifically tailored to help professionals of CPS and judicial systems to promote resilience processes by identifying areas of vulnerability and areas of strength in the families at risk. The general aim of our assessment system is to integrate the specific know-how and guidelines of social workers, psychologists, lawyers, and professionals involved in the decision-making process in order to protect children at risk of abuse/maltreatment/neglect due to parenting deficits in families referred to CPS/ judicial system. Thus, the different modules of our assessment system focus mainly on parenting skills and the resources in the family/social environment.
The assessment system is comprised of six stand-alone modules: one in-depth protocol for identifying proximate and distal risk and protective factors of maltreatment and recidivism against children; one explorative tool for assessing the (residual) social capital in an at-risk family; and four screening surveys for assessing parenting stress, parental child abuse potential, parents' resilience, and traumatic symptoms (in children).
As regards the surveys, three have been validated by C.Ri.d.e.e. research team (PSI-SF, CAPI, and TSCC), and one is a well-known measure for resilience (RSA) that we recommend for its reliability and accuracy. 3 Each of the six modules can be applied alone or in conjunction with the others, and each of them provides information about one key area of family functioning: • The Risk and Protective Factors Protocol and Graph can assist in collecting information about several aspects of family functioning and can help operators focus on risk factors as well as potential resources of the family. The array of risk and protective factors can be arranged as a list or as a graph, and it can be used to assess a single parent or the whole family. In this case, our recommendation is to assign the presence of a given factor to the whole family even if it is reported only for one parent (e.g., parental psychopathology reported for the mother only but assigned to the whole family). The rationale is that the presence of a risk and/or protective factor influences not only a given parent but also the whole family. • The Todd Map can assist in detecting social capital of a given individual/family and deepen the knowledge about relational bonds between individuals. As in the case of the Risk and Protective Factors Protocol and Graph, the Map can be used to assess an individual or a family. • The four screening instruments can be used to obtain quantitative data about the level of stress related to the parenting role, the risk of maltreatment/abuse by a parent, resources that can lead to resilience in parents, and the eventual traumatic consequences on the child. The precise indications about how to use each instrument can be found in the validation papers.
In our proposal, professionals should aim to assess all of the above areas as a rule of thumb, as the array of information gathered could help in subsequent decisionmaking. However, the modularity of the system allows professionals to arrange and personalize the assessment in order to meet time constraints and priority. In this light, some indications may be helpful: • The Risk and Protective Factors Protocol and Graph can be completed in different sessions and can be updated as the knowledge about the family increases.
Indications are to use this module whenever possible, as it is also a useful tool to monitor progress. • The Todd Map can be used as a "resource-oriented" tool to complete the representation about the social capital of the family. As such, it can be of invaluable help to gain a broader picture of the resources that can be engaged in a support system for the family. As for the Risk and Protective Factors Protocol, it can be completed in different sessions and can be used as a monitoring tool. • We think of the four instruments as sort of practical "alarm tools" to be administered when the operator suspects the presence of: -Deep parental strain that may impact child-rearing or lead to increased risk of maltreatment (PSI + CAPI + RSA) -Risk of parental maltreatment (CAPI + TSCC to child) -Risk of psycho-traumatic damage to the child (TSCC to child) -Need for understanding resilience resources in parents (RSA)

Risk and Protective Factors Protocol and Graph
The Risk and Protective Factors Protocol is a theoretically grounded instrument that consists of 36 factors. The aim of the Protocol is to help professionals evaluate individual, familial, and contextual factors that come into play when parents are at risk of not attending to their children needs or becoming maltreating/abusive.
As it is clear from Fig. 1.2, distal risk factors mostly concern context, values, and resources that exert an influence on parenting activities and raise the level of vulnerability of families. Some of them are non-specific in terms of child maltreatment and abuse (e.g., chronic poverty, young maternal age, etc.). Three are very specific: Experience of neglect, violence, and/or abuse during the childhood; Approval of violence and punishments as educational practices; and Approval of child pornography. These are underlined in Fig. 1 B. Mother/father: Young maternal age, experience of neglect, violence and/or abuse during the childhood, parental psychopathology, parental social deviance, parental drug abuse, denial of responsibility about child maltreatment, compensation syndrome, 4 poor empathy skills, impulsivity, lack of frustration tolerance, separation anxiety, parental psychological elaboration of experience of neglect, violence and/or abuse during the childhood, empathy, assumption of responsibility about child maltreatment, desire to improve oneself, autonomy/independence, good level of self-esteem. C. Parents as a couple: Chronic poverty, low educational level, one-parent family, approval of violence and punishments as educational practices, approval of child pornography, unwanted pregnancy and maternity, dating conflict and domestic violence, discomfort related to the dependency on services, conflict management skills. D. Relatives, friends, community: Lack of interpersonal relationships, lack of network and social integration, distrust of social rules and institutions, problematic relationship with family of origin or partner's family, good relationship with at least one relative, supportive network of relatives and/or friends.
4 "Compensation syndrome" can be defined as a parental attitude in which the parent expects some sort of psychological compensation for experiencing adverse growing conditions as a child, and this compensation is expected from a third person (usually his/her child) or an institution (e.g., CPS, judicial system, social welfare, etc.).

Fig. 1.3 Risk and protective factors Graph
A useful complement is the risk and protective factors graph (see Fig. 1.3) that explodes the Protocol on a 2-dimensional space in order to help professionals quickly assess whose areas of family functioning are more worrying or needing attention. The operator may use it as a graphic representation of the array of risk and protective factors assessed in the family, by recording in the appropriate area the presence of a given factor. For example, a family characterized by Chronic poverty (DRF 1), Parental psychopathology (PRF 1), Separation anxiety (PRF 9), and Empathy (PPF 3) will have one note in the quadrant I-C, two in quadrant II-B, and one in quadrant III-B. This notation system allows the professional to quickly be aware of those areas that are more at-risk or more resourceful in a given family.

Todd Map
The Todd Map (1979) is a sociological instrument to assess relational capital in individuals. It allows the operator to measure with a graph the density and the rank of social capital of individuals, in different environments and contexts.
It can be represented as a series of concentric circles segmented in different sectors, while the parent under assessment is placed in the center of the graph. The operator should trace a dot for each person the parent is in relation with for the four relevant environments reported: family, friends, work, and public services. The distance from the center indicates how relationally close that person is to the parent: one is the most important (a close relative), while four is the least important (an acquaintance).
The Todd Map allows the operator to understand, with a quick glance, how dense the support system of a parent is in terms of proximity and distribution in the four environments. For example, a parent may have many contacts in the "work" subsystem but way less in the "friend" subsystem, signaling a potentially unbalanced distribution of social capital (as illustrated in Fig. 1.4).
The Todd Map allows also a "qualitative" assessment of the relations between the parent and his/her social capital. The operator may ask the parent to draw lines that connect the most relevant relations between him/her and the persons drawn on the map, or to differentiate the relations in terms of quality (e.g., positive relations -> solid lines; negative relations -> dotted lines, etc.). The social capital is considered a very important buffering factor in at-risk families, decreasing the likelihood of child maltreatment and abuse (cf. Fujiwara, Yamaoka, & Kawachi, 2016;Zolotor & Runyan, 2006). In this regard, we assess both a sense of community and cohesion in the neighborhood and the availability of friends, relatives, and close relationships in the family. The Todd Map can help assess how many resources the parent/family has and how close they are.
The following paragraphs will focus on four quantitative measures, aimed at providing professionals with some "alarm checks" about key areas of parenting that need immediate attention. In particular, we assess both parenting issues and children outcomes via four measures: one about parental stress, one about parental disposition to become maltreating or abusive toward their children, one about resilience resources eventually present in parents, and finally one about eventual psychotraumatic damage to children. As previously stated, C.Ri.d.e.e. research team has extensively validated three of them (PSI-SF, CAPI and TSCC), while the other is a very well-known measure for resilience (RSA) that we recommend for its reliability and accuracy.

Parenting Stress Index Short Form (PSI-SF)
The Parenting Stress Index Short Form (PSI-SF; Abidin, 1995; Italian validation by Guarino, Di Blasio, D'Alessio, Camisasca, & Serantoni, 2008) is a self-report measure of 36 items with answers on a Likert 5-point scale (from 1 = strong disagreement to 5 = strong agreement) suitable for evaluating the stress level in the relationship between caregiver and child. The measure can be answered by each of the parents independently. The measure provides professionals with a score about a clinical level of parental stress and can be administered in about 10 minutes.
The Italian version of PSI-SF consists of three subscales such as Parental Distress (PD, 12 items; e.g., "Many things in my life disturb me"), Parent-Child Dysfunctional Interaction (P-CDI, 12 items; e.g., "My child seldom does things that gratify me"), and Difficult Child (DC, 12 items; e.g., "My child does some things that annoy me a lot") as well as a defensive responding subscale that consists of 7 items drawn from the PD subscale. The defensive responding subscale evaluates parental bias in reporting by quantifying the desire of parents to present a favorable impression of themselves and minimize problems in the parent-child relationships. The PD subscale focuses on the sense of competence/incompetence in rearing the child, marital conflict, lack of social support, and stress associated with the restrictions deriving from the role of parent. The P-CDI subscale measures parents' perceptions of the emotional quality of their relationship with their children. Finally, the DC subscale focuses on the parent's perception of the child in terms of temperament, hostile, non-collaborative, and provoking behaviors. The sum of the scores of the three subscales (PD, P-CDI, and DC) leads to the Total Stress score, which gives an indication of the overall level of the parental role-specific stress, not deriving from other roles or other events. The 90th percentile of the total PSI-SF score represents a "clinically significant" level of parenting stress (Abidin, 1995) and can be used as an indicator that counseling or other supports are required. The values of internal consistency of the Italian validation of the PSI-SF (Guarino et al., 2008) correspond to a α = 0.91 for the Total Stress scale, a α = 0.91 for the PD subscale, a α = 0.95 for the P-CDI subscale, and a α = 0.90 for the DC subscale.

Child Abuse Potential Inventory (CAPI)
The Child Abuse Potential (CAP, Milner, 1986; Italian validation by Miragoli, Camisasca, & Di Blasio, 2015;Miragoli et al., 2016) is a self-report inventory of 160 items with a forced-choice format ("agree" vs. "disagree"), and it includes the Abuse scale (77 items) that is widely used as a child physical abuse screening tool. As regards the purpose of risk assessment about the physical potential abuse, only the Abuse scale (77 items) is habitually used (Milner, 1986). An even briefer version of the scale was developed for the Italian population (CAPI, 17 items; Miragoli, Traficante, Camisasca, & Di Blasio, 2017). Items were selected to maximize (a) discriminative function referred to the abuse condition; (b) predictivity of the Abuse scale on the total score; (c) factor stability; and (d) clinical validity of the selected items. A similar international brief version of the CAPI can be found in Ondersma, Chaffin, Mullins, and LeBreton (2005). The CAPI in full form can be administered in 15 minutes. The original CAP Inventory (Milner, 1986) contains a total of ten scales. The primary clinical scale is the 77-item physical child abuse scale. This Abuse scale can be divided into six factor scales: Distress (e.g., "I often feel very frustrated"), Rigidity (e.g., "Children should always be neat"), Unhappiness (e.g., "I am an happy person"), Problems With Child and Self (e.g., "I have a child who is slow"), Problems With Family (e.g., "My family fights a lot"), and Problems From Others (e.g., "Other people have made my life hard"). In addition, the CAP Inventory contains three validity scales: Lie Scale, Random Responses Scale, and Inconsistency scale. The factorial structure of the Italian version of the CAP Inventory presents factors consistent with the original version (accounting for 31% of the variance): "Loneliness and Distress" (16 items; e.g., "I often feel alone"; "People do not understand me"); "Rigidity" (10 items; e.g., "Children should always be neat"); "Impulsiveness and Anxiety" (13 items; e.g., "I find it hard to relax"; "Sometimes I fear that I will lose control of myself"); "Unhappiness" (7 items; e.g., "I do not laugh very much"); "Problems with Self, Child, and Family" (10 items; e.g., "I have a child who is bad"); and "Interpersonal Difficulties" (6 items; e.g., "People expect too much from me").
The Abuse scale of the Italian version of CAP Inventory shows adequate internal consistency (α = 0.87) and, with the cut-off of 166 as suggested in the original version of the CAP Inventory (Milner, 1986), when abusive and non-abusive groups are compared, it correctly classifies 70.3% of the abusive parents and 100% of the nonabusive parents. The cut-off of 166 derives from the scoring procedure of the CAP Inventory, which involves a series of weighted items ranging from 1 to 23 (cf. Milner, 1986).

Resilience Scale for Adults (RSA)
The Resilience Scale for Adults (RSA) is a self-report screening instrument used to determine the ability to overcome stress and hardships while maintaining positive psychological functioning. It is based on the work of Friborg, Barlaug, Martinussen, Rosenvinge, and Hjemdal (2005) and has been validated in Italian by Capanna, Stratta, Hjemdal, Collazzoni, and Rossi (2015). RSA is comprised of 33 items on a 7-point semantic differential scale that provides a measure on six resilience factors and a total score. The scale can be administered in about 15 minutes. The six factors identified by the authors are as follows: Perception of self (e.g., confidence in own abilities); Planned future (e.g., positive attitude toward future); Social competence (e.g., confidence in own abilities to create and maintain social bonds); Family cohesion (e.g., positive attitude toward own family shared values and support); Social resources (e.g., availability of positive social capital outside family); and Structured style (e.g., ability to self-direct and plan goal-oriented actions).
In the Italian validation of Capanna et al. (2015), the structure of the test proved to be good for five factors out of six while only adequate for the "Structured style" factor. Research papers indicate that a mean total score of the RSA higher than 6.0 can be thought as "high resilience" (cf. Friborg et al., 2006) and a score below 3.5 may be correlated with problems of adaptation like substance use disorder (cf. Bonfiglio, Renati, Hjemdal, & Friborg, 2016).

Trauma Symptom Checklist for Children (TSCC)
The Trauma Symptom Checklist for Children (TSCC; Briere, 1996; Italian validation by Di Blasio, Piccolo, & Traficante, 2011) is a self-report questionnaire designed to assess posttraumatic stress, dissociation, anxiety, anger, sexual concerns, and depression in children and adolescents (separately normed for boys and girls ages 8-12 and 13-16 years), who have been traumatized and/or abused. Children indicate on a 4-point scale (0 = never to 3 = almost all of the time) how often experiences such as "Feeling nervous or jumpy inside" happen to them.
The Italian version of TSCC is a self-report of 54-item questionnaire and consists of two validity scales, Underresponse (UND, 10 items) and Hyperresponse (HYP, 8 items), as well as six clinical scales: Anxiety (ANX, 9 items; e.g., "I feel scared"), Depression (DEP, 9 items; e.g., "I feel sad and unhappy"), Posttraumatic Stress (PTS, 10 items; e.g., "I can't stop thinking about something bad that happened to me"), Sexual Concerns (SC, 10 items; e.g., "I touch my private parts too much"), Dissociation (DIS, 10 items; e.g., "I feel I'm not in my body"), and Anger (ANG, 9 items; e.g., "I feel furious"). Each clinical scale yields subscale raw and T-scores, and higher scores designate a greater number of symptoms. 5 The clinical cut-off score is set at T = 65 for all subscales except sexual concerns where the clinical range is set at above T = 70. All components of the TSCC show good reliability and validity in the Italian population (ANX: α = 0.72; DEP: α = 0.77; PTS: α = 0.79; SC: α = 0.80; DIS: α = 0.71; ANG: α = 0.81). It can be administered in about 15/20 minutes.

Research Data on the C.Ri.d.e.e. Multi-method Assessment System
In order to improve the Protocol of risk and protective factors  and to test its applicability, a series of empirical studies were conducted. Studies adopted the "judgment analysis" approach (Stewart, 1988) to investigate association between the Protocol's factors and the level of minors' safety. This methodology requires retrospectively analyzing information cues, content, and characteristics of referred case files and to test their correlation with CPS workers' assessment and intervention (Benbenishty & Chen, 2003). In the following paragraphs, we summarize research data about the Risk and Protective Factors Protocol.

Exemplary Data on Italian Sample
The first preliminary study performed by Miragoli and Verrocchio (2008) focused on the collection of data on an Italian sample to verify the applicability and the discriminative power of the Protocol's risk and protective factors. The evaluation of factors' presence or absence was performed through the retrospective analysis of 400 referred families' social records. The most frequent maltreatment was neglect (35.8%), followed by physical maltreatment (15.0%). The sample was split up on the basis of the CPS workers' intervention, operationalized as child out-of-home placement (high risk) or monitoring and parental abilities support (low risk). Results showed that several factors, in particular distal factors, emerged as good predictors of workers' assessment for both mothers and fathers. In particular, it emerged that all distal risk factors but "One-parent family" and "Low educational level" were significant predictors of a high-risk intervention by CPSs. The most relevant proximal risk factors in predicting high-risk intervention were "Parental social deviance," "Parental drug abuse," "Denial of responsibility about child maltreatment," "Poor empathy skills," and "Unwanted pregnancy and maternity." Some factors were found significant only for mothers ("Parental psychopathology" and "Problematic relationship with family of origin or partner's family") and some only for fathers ("Impulsivity" and "Lack of frustration tolerance"). Regarding the interplay between distal and proximal risk factors, it is possible that the former shows more discriminant power because they have more objective indicators (e.g., poverty can be gauged with monthly wage of parents) and are consequently easier to detect.

Research Data on Comparisons Between Italian and Migrant Families
The risk and safety assessment could be more complex for CPS workers who operate in a context characterized by high rates of migrant families, and who often show a higher level of vulnerability (Milani, 2013). In fact, literature confirmed that the multiple challenges faced by migrant parents may negatively affect their level of well-being and the quality of their family relationships (Valtolina, 2013;Yu & Singh, 2012). Moreover, several studies identified the overrepresentation of migrant children in the rates of the cases reported to CPSs (Alink et al., 2013; Autorità garante per l'infanzia e l'adolescenza, Terre des homes, & CISMAI, 2015). This overrepresentation seems to be partially explained by socioeconomic disadvantage; however, literature did not provide univocal results about which factors predict the workers' safety assessment (LeBrun et al., 2015). Considering this gap in the literature, a series of studies about the Protocol's application was investigated with two questions (Grumi, Milani, & Di Blasio, 2017a;Milani et al., 2016;Milani & Gagliardi, 2013): • What are the differences in the prevalence rates of distal and proximal risk and protective factors between Italian and Migrant families?
• Which risk and protective factors significantly influence CPS workers' removal decision in case of Italian and migrant families?
About the first aim, results are consistent with previous studies and confirmed the vulnerability of migrant parents in many areas. Table 1.1 shows that migrant families are characterized by a higher prevalence of distal risk factors related to their socioeconomic status and by a lower prevalence of protective factors compared to Italian parents (Grumi et al., 2017a). Migrant families show lower prevalence than Italian families in many indicators: parental experiences of maltreatment, psychopathology, compensation syndrome, separation anxiety, problematic relations with family, and domestic violence. We attribute these counterintuitive results to two processes: (a) migrant families could be characterized by some resilience (they accept the risk of moving from their home country to gain better living conditions); and (b) migrant families could be more "under the spotlight" of social services due to their socioeconomic condition (as the higher prevalence of distal risk factors shows) and hence taken into care before proximal risk factors appear (or worsen).
About the second point, studies demonstrated that migrant and Italian families are characterized by two different patterns of factors (Tables 1.2 and 1.3), 6 with the exception of lack of knowledge or interest relating to child development, which con- stitutes a significant predictor for both groups. In particular, for Italian families the lack of knowledge or interest related to child development enormously increases the risk of child removal, and even though it exerts an indirect impact, it is linked to critical variables like parental monitoring and unrealistic expectations (Milner & Chilamkurti, 1991). Migrant and Italian families are characterized by a clear difference between Odds Ratio regarding this factor: we argue it could be possibly due to a sort of differential perception of its relevance by CPS personnel. It is possible that CPS professionals tend to overreact when this factor is detected in Italian families due to cultural reasons (i.e., "Bad parenting") while attributing it to cultural reasons when found in migrant families (i.e., "Different parenting"). Other relevant factors for migrant parents were past experience of neglect violence and/or abuse during the childhood that may affect parental well-being and quality of parent-child relationship and attachment (Ionio & Mascheroni, 2014) and the good level of selfesteem that supports change and allows migrant parents to ask for help in case of need, without denying obstacles and limits (Milani & Gagliardi, 2013). For Italian families other significant predictors were the parents' psychopathology and the social deviance, while discomfort related to the dependency on services and a good level of autonomy/independence reduced the level of risk perceived by social workers.

Decision Trees
A second set of studies used a decision tree analysis (CHAID) in order to identify the most relevant factors for the assessment of migrant and Italian parents (Grumi, Milani, & Di Blasio, 2017b;Milani, Di Blasio, & Grumi, 2017). The CHAID algorithm (Chi-squared Automatic Interaction Detector, Kass, 1980) is a form of recursive partitioning that begins with the "parent node" which is subsequently split into subgroups, "child nodes," by a predictor variable. The predictor variable at each stage was selected on the basis of Chi-square tests, and splitting continued until predetermined stopping criteria were met. The final nodes identify subgroups defined by different sets of independent variables. This data analysis approach addresses issues related to the high number of risk and protective factors included in the assessment and the small sample size of our study. Moreover, it allows us to identify multiple profiles of high and low risk that are associated with a specific outcome (child-out-of-home placement decision vs. monitoring and support). We maintain that this methodology and the results it has provided could be of use to professionals in order to have some landmarks to prioritize some factors over others when acquiring information on the case they handle.
In particular, two decision trees were performed separately for migrant (n = 177) and Italian (n = 163) families. As shown in Fig. 1.5, the decision tree for migrants was estimated to correctly classify 77.4% of the cases and identified 2 high-risk parental profiles (see nodes 3 and 12) and 2 low-risk parental profiles (see nodes 6 and 7), while one node is not sufficiently discriminative at this level of split. In particular, the good level of self-esteem acts as a protective factor reducing the risk of child removal, but it is not sufficient when risk factors, impulsivity and young maternal age, are present. Good level of self-esteem χ 2 = 38.56 * * * Lack of know. or int. for child developement χ 2 = 5.6 * Impulsivity χ 2 = 15.09 * * *  As shown in Fig. 1.6, the decision tree for Italian families correctly classified 73.3% of the cases. It identified one high-risk parental profile (see node 3) and two low-risk profiles (see nodes 7 and 8), while two nodes (4 and 5) were not sufficiently discriminative at this level of split. In particular, the absence of autonomy or independence combined with impulsivity emerged as the pattern with the highest probability of child removal, while a good level of autonomy together with the desire to improve oneself characterizes parents evaluated as low-risk, even if they have scarce interpersonal relationships.
To sum up, two different patterns of predictors -except for impulsivity -emerged for migrant and Italian parents, suggesting that CPS workers focused on different indicators when they are asked to assess parents with different sociocultural backgrounds. However, in line with previous results, distal risk factors that are overrepresented among migrants (as chronic poverty and lack of network and social integration) did not constitute significant predictors of the safety assessment.
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.