**Urgent Protection** *versus* **Chronic Need: Clarifying the Dual Mandate of Child Welfare Services across Canada**

#### **Nico Trocmé, Alicia Kyte, Vandna Sinha and Barbara Fallon**

**Abstract:** This study analyzed data from the 1998, 2003 and 2008 Canadian Incidence Study of reported child abuse and neglect (CIS) and compared the profile of children who were reported for an urgent protection investigation *versus* any other investigation or assessment. As a proportion of all investigations, urgent protection cases have dropped from 28% of all investigations in 1998, to 19% in 2003, to 15% in 2008. Results from the CIS-2008 analysis revealed that 7% of cases involved neglect of a child under four, 4% of cases involved sexual abuse, 2% of cases involved physical abuse of a child under four and 1% of cases involved children who had sustained severe enough physical harm that medical treatment was required. The other 85% of cases of investigated maltreatment involved situations where concerns appear to focus less on immediate safety and more on the long-term effects of a range of family related problems. These findings underscore the importance of considering the dual mandate of child welfare mandates across Canada: intervening to assure the urgent protection and safety of the child *versus* intervening to promote the development and well-being of the child.

Reprinted from *Soc. Sci.* Cite as: Trocmé, N.; Kyte, A.; Sinha, V.; Fallon, B. Urgent Protection *versus* Chronic Need: Clarifying the Dual Mandate of Child Welfare Services across Canada. *Soc. Sci.* **2014**, *3*, 483–498.

#### **1. Introduction**

Rates of reported child abuse and neglect have been rising across Canada: the number of investigated reports increased from 135,000 in 1998 to over 235,000 in 2008 [1]. In Ontario, the largest Canadian province and the only jurisdiction for which earlier data are available, the number of investigation investigations nearly tripled from just under 47,000 in 1993 to close to 129,000 in 2008 [2,3]. This expansion has been driven by a broadening of child welfare mandates, in particular with respect to the inclusion of: (1) emotional maltreatment [4]; (2) exposure to intimate partner violence [5]; (3) cases where risk of future maltreatment is the primary concern [6]; as well as by (4) professionals becoming more aware of the emotional and cognitive effects of child maltreatment [4,5,7,8].

In response to these changes, concerns are being raised that child welfare practice has shifted in Canada from an approach that had emphasized family support and family preservation towards more intrusive child protection models [9–11]. This shift is attributed in part to the introduction of risk assessment tools [11–13]. Several jurisdictions have sought to re-balance child welfare practice by introducing a range of "differential" or "alternate" response policies, by streaming lower-risk cases to family support services that do not focus as narrowly on protection concerns. Implementing a broader range of child welfare responses has proven to be difficult, especially in a risk adverse environment [14–18]. Part of the difficulty is that the concept of risk in child welfare remains

relatively uni-dimensional and does not distinguish clearly enough between different types of risk, in particular, the difference between: (1) situations where the primary concern is to prevent any form of recurrence, such as in cases where escalating physical or sexual abuse poses an acute threat to the safety of a child; compared to (2) situations where the concern is to the risk of the development and well-being of a child exposed to chronic family dysfunction. The purpose of this paper is to explore the difference between the concepts of risk and harm in the context of Canadian child welfare services.

#### **2. Disentangling Protection and Well-Being**

Expanding conceptualization of what constitutes child maltreatment poses a particular challenge to defining the central concepts that shape child welfare services. While concepts such as "protection", "safety", "harm" and "risk" are relatively clear in the context of specific forms of maltreatment, such as child sexual abuse or severe physical abuse, the meaning of these terms is less clear in reference to an increasingly broad array of forms of maltreatment such as neglect and emotional maltreatment, where the focus of concern shifts from protection from a specific abusive incident to long-term exposure of dysfunctional family interactions that lead to psychological harm [19,20]. A seminal critique of risk assessment in child welfare identified three very different ways that risk assessment was being used in child welfare at the time: as the "likelihood that a given person (usually a parent) will harm a child in the future" ([21], p. 486), as a way of distinguishing between levels of severity of maltreatment and lastly as a need's assessment [21]. Likelihood of recidivism, severity of abusive or neglectful incidents and a child and family's need for supportive services represent, however, three very different assessments. Despite the development of a significant body of empirical research on the predictive validity of risk assessment tools [22,23], the use of the concept of risk in child welfare practice remains as confused as ever [13,24]. This confusion became apparent in a series of validation tests completed at the end of the 2003 cycle of the Canadian Incidence Study of Reported Child Abuse and Neglect (CIS) which revealed that child welfare workers were force coding "maltreatment investigations", where there was no specific concern about an abusive or neglectful incident having possibly occurred, because of concerns about risk of maltreatment [25]. Following the addition of a "risk investigation only" option in the 2008 cycle of the study, 26% of all investigations were categorized as risk assessments [26].

Confusion about the meaning of risk in child welfare is compounded by a lack of clarity with respect to the notion of harm. Legal definitions often do not specify what is meant by harm, allowing discretion for interpretation by professionals and variability among jurisdictions [27]. Harm is often used as generic concept that combines both physical and emotional harm. For instance, the harm measure used in the U.S. National Incidence Study of Reported Child Abuse and Neglect distinguishes between a "harm standard" and "endangerment standard" that classifies harm to the child as fatal, serious, moderate or inferred, but does not distinguish between physical injuries, other health conditions and emotional harm [28].

In cases of investigated maltreatment, physical harm primarily involves minor injuries and in the most extreme cases severe or fatal injuries [29]. Only 8% of substantiated cases of maltreatment documented in the 2008 cycle of the CIS involved physical harm, and most of these cases involved bruises or other minor injuries that had not required medical attention [26]. Psychological harm, not physical harm, is the primary concern in most situations of child maltreatment. Longitudinal studies consistently demonstrate that for most victims of maltreatment the primary concern is the effect of chronic exposure to maltreatment on their social, emotional and cognitive development [19,30,31]. Even with respect to the neurodevelopmental sequelae of maltreatment, one needs to distinguish between injuries caused by shaking or other forms of physical abuse [32] and damage associated with chronic psychological and emotional dimensions of maltreatment [33].

The key distinction in the context of child welfare interventions is not so much between physical and emotional harm, but between acute harm (or the risk of acute harm) typically associated with severe physical abuse, sexual abuse, and neglect of very young children, and long-term psychological harm associated with chronic exposure to neglect, emotional maltreatment, and less severe forms of physical abuse, as well as with many situations involving acute harm. Most child welfare statutes address this distinction in describing the intent and scope of mandated child welfare services. Canadian Provincial and Territorial child welfare statutes generally make reference to notions of both "protection" and "well-being", the first referring to concerns about immediate safety from acute maltreating incidents, the second referring to the effects of chronic exposure to maltreatment (Table 1). In British Columbia the legislation states that both "safety and well-being of children are the paramount considerations" [34]. This notion of well-being and safety being equal and central considerations is also articulated in Ontario, Manitoba, and Newfoundland and Labrador [35–37]. Child welfare statutes in Québec, New Brunswick, and Alberta use the terms "security or development" also giving equal importance to both notions [38–40]. Both terms are also used in Saskatchewan, although perhaps in the more limited sense of promoting the well-being of children in need of protection [41], while Nova Scotia, the Yukon and the North West Territories and Nunavut make broader reference to "best interests" [42–44]. Prince Edward Island is the only jurisdiction that does not include well-being, development or best interests in setting forth the purpose of their legislation [45].

#### **3. Methods**

The Canadian Incidence Study of Reported Child Abuse and Neglect-2008 (CIS-2008) is the third nation-wide study to examine the incidence of reported child maltreatment and the characteristics of the children and families investigated by child welfare authorities [26]. The CIS-2008 tracked 15,980 maltreatment-related investigations conducted in a representative sample of 112 child welfare organizations across Canada in the fall of 2008. Data from the 1998 and 2003 cycles of the study were included as well to examine changes in types of investigations conducted in Canada (Table 2). The 1998 study tracked 7672 child maltreatment investigations conducted in a representative sample of 51 child welfare organizations across Canada and the 2003 study tracked 11,562 investigations in a sample of 55 child welfare organizations in Canada, excluding Quebec where missing data on harm precluded inclusion for the analyses presented in this paper [1,26].


**Table 1.** Purpose or paramount principles guiding child welfare legislation across Canada.

Act; 5 Ontario Child and Family Services Act; 6 Quebec Youth Protection Act; 7 Newfoundland and Labrador Child, Youth and Family Services Act; 8 New Brunswick Family Services Act; 9 Nova Scotia Child and Family Services Act; 10 Prince Edward Island Child Protection Act; 11 Northwest Territories Child and Family Services Act; 12 Yukon Child and Family Services Act.

Information was obtained directly from child welfare workers using a three-page data collection form describing child, family and investigation related information that workers routinely gather as part of their investigation. For the 2008 cycle of the study maltreatment investigations were classified under five major categories, with 32 specific forms of maltreatment subsumed under each category. Investigations where no specific incident had been reported and where the concern was risk of future maltreatment were classified as risk-only investigations [6]. For the purposes of this paper, maltreatment categories were classified on the basis of the primary form of maltreatment, the age of the child and the presence of a severe injury, as either urgent protection investigations or other maltreatment related investigations or assessments. Investigations were classified as urgent protection if a child was younger than four and was investigated for neglect or physical abuse, if the primary concern was sexual abuse, or if a child had sustained physical harm requiring medical treatment. Investigations involving severe injuries were classified as potentially requiring an urgent investigation given that such cases can lead to an escalating pattern of maltreatment and the importance of forensic evidence [46]; sexual abuse investigations were included as requiring an urgent investigation because of the importance of forensic evidence and the possibility that the offender might threaten or pressure the child to recant [47]; and abuse and neglect cases involving children under the age of four were included as urgent because of the increased likelihood of severe injury [48,49], their limited verbal skills, and the possibility that harm may escape scrutiny from daycare and school programs.


**Table 2.** Types of child welfare investigations and assessments conducted in Canada, Canadian Incidence Study of Reported Child Abuse and Neglect (CIS) 1998, 2003 and 2008.

Notes: The italicized rows represent the totals for cases categorized as either "urgent protection" or "other investigations and assessments"; The last row in bold presents the combined "urgent protection" and "other investigations and assessments" resulting in the annual weighted estimate; The annual weighted estimates are based on samples of 7672 in 1998, 11,562 in 2003 and 15,980 in 2008; investigations from the province of Quebec are not included in the 2003 estimates; \* *p* < 0.01, test of significance compares incidence per 1000 children for each type of investigation in 1998 and 2008.

Urgent protection investigations were first compared to other investigations and assessments over the 1998, 2003, and 2008 cycles of the study. Table 2 presents weighted annual estimates of investigations for 1998, 2003 and 2008, as well as rates of investigations per 1000 children for 1998 and 2008; 2003 rates are not presented because of missing harm data in one province. Using data from the 2008 cycle, urgent protection investigations were then compared to other investigations and assessments in terms of harm, selected parent, household and child characteristics and short-term service outcomes.

The harm variables used for the study included (1) whether no harm had occurred, (2) whether harm was in the form of broken bones, a head trauma, bruises/cuts/scrapes, burns and scalds or (3) whether harm was of a mental or emotional nature. There was the possibility of one child sustaining multiple forms of injury. Primary caregiver risk factors included: substance abuse (*i.e.*, alcohol or drug abuse), mental health concerns, cognitive impairment, and lack of social supports. Child functioning concerns which cover a check list of 18 different issues ranging from attachment issues, drug/solvent abuse to academic difficulties were grouped together under a general dichotomous variable of "at least one child functioning concern". For each parent and child functioning variable the four choice ratings of "confirmed", "suspected", "no" or "unknown" were collapsed into dichotomous variables "confirmed or suspected" or "no or unknown". A "housing problems" variable was created based on whether the family had experienced more than two moves in the past year, whether the home was overcrowded or whether there were household hazards present in the home. The two investigation categories were also compared on two additional risk factors: whether the "household regularly runs out of money for basic necessities" and whether the family had previously received child welfare services. Service response documented during the investigation, typically within the first four to six weeks of contact, included whether the case was being transferred for on-going services, whether a referral had been made to specialized services (community based or child welfare agency run service), whether the child had been placed in out-of-home placement (including foster care, kinship care, group home or residential care), whether an application for a child welfare court order had been initiated and whether maltreatment was substantiated. Chi square tests were used to compare differences between the two types of investigations in terms of harm, parent, child or household characteristics and in service responses, whereas independent sample T-tests were used to compare differences in the incidence per 1000 children in 1998 and 2008 for each type of investigation. Both procedures were adjusted to take into consideration the CIS sampling design and the use of weighted data (see [26]).

#### **4. Results**

The number of child maltreatment investigations conducted in Canada has increased from an estimated 135,260 in 1998 to 235,871 in 2008; during the same period the rate of investigation per capita has increased from 21.47 per 1000 children to 39.16 in 2008 (Table 2). Readers should note that the data for 2003 in Table 2 excludes investigations from Quebec because injury data was not available from Quebec in the 2003 study. Canada-wide investigation estimates published in Table 3-2 of the CIS-2008 Major Findings report [26] show that the increase in investigations occurred entirely between 1998 and 2003, with no significant change in the Canada-wide rate of investigations between 2003 and 2008. Despite the overall increase in investigations, the number of investigations that we classified as urgent has not changed significantly, an estimated 38,274 in 1998, compared to 36,299 in 2008, or 6.08 per 1000 children in 1998 compared to 6.03 per 1000 in 2008. As a proportion of all investigations, however, urgent protection cases have dropped from 28% of all investigations in 1998, to 19% in 2003 to 15% in 2008. The number of cases involving physical injuries and

physical abuse cases involving children under 4 has not changed significantly, while the number of sexual abuse investigations has decreased and the number of neglect cases involving children under 4 has increased. As shown in Table 2, the overall increase in investigation is primarily accounted for by exposure to intimate partner violence and risk assessments, which accounted for 42% of all investigations in 2008. Although exposure to intimate partner violence and risk assessments were not explicitly tracked as primary categories of maltreatment in 1998, such investigations would have been included under the neglect or emotional maltreatment categories.

Thirteen percent of investigations categorized as urgent protection involved some type of documented physical harm and four percent of other investigations or assessments involved physical harm not requiring medical attention; keeping in mind that all cases involving harm that was serious enough to require medical attention were re-coded for the present analysis as urgent protection (Table 3). Estimates of fatal harm could not be calculated because the number of fatalities in the study sample was too small to derive estimates. Most investigations where physical harm was noted involved either some other form of a health condition and bruising, cuts or scrape. Ninety percent of the other investigations and assessments that involved a minor bruise, cut or scrape, involved allegations of physical abuse (injury rates by sub-type of investigation not included in Table 4, but is available from the first author). In 16% of urgent investigations and 12% of other investigations and assessments, the investigating worker had documented some type of emotional harm, such as nightmares, withdrawal, or aggression that could be attributed to the alleged maltreatment.


**Table 3.** Urgent protection investigations and other investigations and assessments by harm (CIS-2008).

\*\*\* *p* < 0.001; \* *p* < 0.05; rows add up to more than 100% because a child may have sustained several types of injuries.

In comparison to other investigations and assessments, cases reclassified as urgent protection generally involve moderately more parent and household risk factors, including more substance abuse, cognitive impairment, lack of support, and housing and financial problems (Table 4). In contrast, investigations and assessments that did not meet our urgent protection classification involved a significantly larger proportion of children with noted internalizing or externalizing problems.


**Table 4.** Urgent protection investigations and other investigations and assessments by parent, household and child concerns (CIS-2008).

\*\*\* *p* < 0.001; \*\* *p* < 0.01; \* *p* < 0.05.

Table 5 compares cases classified as urgent protection to other investigations and assessments in terms of investigation outcomes and whether cases had been previously investigated. Rates of opening for on-going services, placement and use of court were higher for cases classified as urgent, while rates of previous child welfare services and rates of substantiation were higher for other investigations and assessments.


**Table 5.** Urgent protection investigations and other investigations and assessments by service response (CIS-2008).

\*\*\* *p* < 0.001.

#### **5. Discussion**

An estimated 36,299 child maltreatment related investigations completed in Canada in 2008 involved situations where a child: (1) had sustained a severe injury or health condition; or (2) was a possible victim of sexual abuse; or (3) was young enough (under four) to be at high risk of a serious injury as a result of abuse or neglect. From 1998 to 2008 the number of investigations that met one of these three criteria for urgent protective investigation has remained virtually unchanged, at a little over six investigations per 1000 children. In contrast, other maltreatment related investigations have more than doubled, going from a rate of 15.39 investigations per 1000 children in 1998 to

33.13 investigations per 1000 children in 2008, an increase that has been driven by investigations of children exposed to intimate partner violence and risk assessments where there were no specific abuse of neglect allegations [4,5]. As a result, the proportion of investigations that met our urgent protection classification has dropped from 28% in 1998 to 15% in 2008.

Comparison with data from other jurisdictions is limited because few jurisdiction report rates of harm in cases of investigated maltreatment. The Fourth National Incidence Study of Reported Child Abuse and Neglect in the United States, which collected data in 2005 and 2006, reports a rate of 5.5 substantiated Child Protective Service investigations per thousand that met the study's harm standard, and 16.9 Child Protective Service investigations that met the study's broader endangerment standard (derived from Tables 8-1 and 8-8, [28]). While differences in definitions and methodologies limit the comparability of CIS investigation statistics and National Incidence Study (NIS) CPS statistics, it is nevertheless noteworthy that at 2.59 per 1000 children, the CIS rate of substantiated urgent protection cases was lower than the NIS rate of harm standard victims, while the overall rate of substantiated CIS investigations was higher, at 20.81 substantiated investigations per 1000 children (Table 5: 2.59‰ + 18.22‰ = 20.81‰). As a result, a third (33%) of the NIS-4 endangerment Child Protective Service cases met the NIS harm standard, whereas only one in eight (12%) of the CIS-2008 substantiated investigations met our urgent protection rating. The difference between the NIS and CIS can be explained in part by a broader inclusion of risk and exposure to intimate partner violence cases in Canada relative to the U.S. [50], which account for an important part in the overall increase in investigations in Canada [5]. A similar increase in investigations driven by cases of exposure to intimate violence had been documented in Australia [51].

The overall increase in child welfare investigations in Canada is puzzling, not only because investigations involving urgent protection have not been increasing, but also in light of growing evidence from several jurisdictions that rates of child victimization in the population have been declining. Population based victimization studies in the U.S., Australia and the UK point to a decrease in the incidence of physical and sexual abuse [52]. Declines were also observed in the U.S. for victimizations that are serious and indicative of more pathological circumstances like homicide [52]. Rates of child homicide have also been declining in Canada [53]. In contrast, cases of neglect and psychological maltreatment have remained relatively stable over time and in some jurisdictions may even be increasing [54,55]. Explanation for these declines can be attributable to a number of factors with no real consensus as to the driving cause (*i.e.*, reporting practices, definitional standards, increased social intervention). Whereas the stability of neglect trends may be the result of increased education and policy initiatives to include new forms of neglect, which in turn, may be masking a decline of more conventional forms [55].

The comparison between cases classified as urgent protection investigations and the other investigations and assessments show more similarities than differences. The vast majority of both urgent protection (88%) and other investigations (97%) involved situations where no physical harm had been noted, even though all severe harm cases were categorized as urgent protection. Urgent protection investigations involved moderately more situations where parent or household risk factors were present, whereas the other investigations and assessments involved more situations where child functioning concerns had been noted. Emotional harm was documented in more urgent protection cases than in other investigations or assessments, but, as with physical harm, the most noteworthy finding was that in well over 80% of investigations, there no immediate signs of emotional harm had been noted. The fact that relatively few investigations involved situations involving signs of emotional or physical harm is not surprising, given that child welfare statutes include situations where a child has been harmed or is "at risk of harm". For instance, legislation in Alberta makes reference to a child being in need of intervention services "*…if there are reasonable and probable grounds to believe…*" (Alberta Child and Family Enhancement Act, 2000, Section 2) a child has been a victim of maltreatment, and in Quebec for cases of neglect, sexual and physical abuse a "*situation in which there is a serious risk*" of these forms of maltreatment occurring also warrant intervention services (Quebec Youth Protection Act, 2007, Section 38 (b)2, 38 (d)2, 38 (e)2). The focus on risk of harm is certainly consistent with longitudinal studies showing that the effects of maltreatment manifest themselves over time [19,30,31], and is reflected in the general widening of the child welfare mandates to include maltreatment categories that go beyond physical injury and demonstrable harm [54].

Service response patterns vary in an interesting way. Urgent cases were less likely to be substantiated, possibly showing that (1) professionals are more likely to report suspicions when situations involve serious injuries, very young children or possible sexual abuse and (2) that these cases are less likely to be screened out. In contrast, urgent cases were more likely to be open for ongoing service, to lead to and out of home placement and to proceed to court. In other words, the cases that we classified as urgent protection cases appear to be more likely to reported, screened in, open for service, and placed in out of home care. Nevertheless, given that only 15% of cases met our urgent protection classification, the majority of cases open for on-going services—26% of 199,973 other investigations and assessments compared to 31% of 36,299 urgent protection investigations—and the majority of children placed in out of home care—7% of 199,973 other investigations and assessments compared to 14% of 36,299 urgent protection investigations—involved situations where there were no serious injuries, the child was four or older and there were no allegations of sexual abuse. In addition, previous child welfare services were noted in more than half of both types of cases, with previous services being noted most often (63%) in situations involving other investigations and assessments. The vast majority of child maltreatment related reports investigated and eventually opened for ingoing services in Canada involve non-urgent situations where repeated exposure to neglect, emotional maltreatment and family violence jeopardize the well-being of children.

#### **6. Limitations**

A number of limitations must be taken into consideration in interpreting these findings. The CIS collects information directly from the investigating child welfare workers and the data collected is not independently verified. Child welfare workers are provided with training from the research team to increase consistency in the application of the study definitions. The CIS only examines cases at the point when they completed their initial investigation of a report of possible child abuse or neglect, or risk of future maltreatment. Therefore, the scope of the study is limited to the type of information available at that point, and in particular does not include information about longer-term emotional that may not be manifested at the time of the investigation. The CIS does not include information about unreported maltreatment nor about cases that were investigated only by the police. Comparisons across cycles of the CIS must be made with caution. The forms of maltreatment tracked by each cycle were modified to take into account changes in investigation mandates and practices. Comparisons across cycles must in particular take into consideration the fact that the CIS-2008 was the first to explicitly track risk-only investigations. Finally, it should be noted that the urgent protection and other investigation and assessment categories used in the article analyses are *post-hoc* classifications and do not represent a direct assessment of urgency made by the investigating workers.

#### **7. Conclusions**

The expansion of child welfare mandates across Canada has resulted in an increase of cases being reported to child welfare authorities where the concerns are long term and are associated with chronic exposure to an increasing range of family problems. While this expansion is consistent with the legislated mandates of child welfare authorities to *promote the protection and well-being of children, there continues to be controversy about the extent to which promoting child well-being is an appropriate function for* child welfare services. It is clear that whenever possible families should be referred to community organizations, however when community services are lacking or families are unwilling to seek services what is the alternative? In the meantime, we know that failure from child welfare agencies to respond to non-urgent cases often results in further deterioration of the family resulting in eventual child welfare response [56,57]. Several Canadian jurisdictions have developed differential response policies in an attempt to address this broader range of maltreatment related problems. The extent to which these policies are being implemented in practice and are truly leading to a better differentiated range of services, remains to be determined.

Hospital emergency rooms depend on clear triage protocols to distinguish between acute and chronic conditions to ensure that the appropriate level of service is provided in a timely fashion. While acute conditions are given priority for urgent care, chronic conditions are not as a result considered to be any less severe. In fact, in many instances the chronic conditions are the ones that in the long term will receive the most intense and extensive services. In response to the dramatic expansion in child welfare mandates across Canada, child welfare service providers and policy makers are looking for protocols and service delivery models that will similarly ensure that the appropriate level of services are provided in a timely fashion. In order to do so effectively, policies and services must be able to disentangle urgent protection from chronic need.

#### **Acknowledgment**

Research for the secondary analyses used in this paper was funded by a Social Sciences and Humanities Research Council Partnership Grant [# 895-2011-101].

#### **Author Contributions**

This paper was written by Nico Trocme, Alicia Kyte and Vandna Sinha. Conceptualization and collection of data of the original study was done by Nico Trocme, Barbara Fallon and Vandna Sinha. The secondary data analyses performed for this paper was completed by Barbara Fallon and Alicia Kyte. All authors read and approved the final manuscript.

### **Conflicts of Interest**

The authors declare no conflict of interest.

### **References**


