- Email: [email protected]
Child maltreatment: The collaboration of child welfare, mental health, and judicial systems
Child Abuse & Neglect, Vol. 19, No. 3. pp. 355-362, 1995 Copyright © 1995 ElsevierScience Ltd Printed in the USA. All rights reserved 0145-2134/95 $9.50 + .00
Pergamon
0145-2134(94)00136-7
SPOTLIGHT
ON
PRACTICE
CHILD MALTREATMENT: THE COLLABORATION OF CHILD WELFARE, MENTAL HEALTH, AND JUDICIAL SYSTEMS STEPHEN BUTLER, LESLIE ATKINSON, MICHAEL MAGNATI'A,
AND ERIC
HOOD
Family Court Clinic, Clarke Institute of Psychiatry, Toronto, Ontario, Canada
Abstract--The alliance of child welfare, mental health, and legal systems has received little empirical attention, despite the magnitude of its impact on children and families. We examined the congruence of child protection agencies legal positions, court clinic recommendations, and judicial dispositions in a sample of 59 contested child maltreatment cases. Placement recommendations/decisionsamong all three systems were highly correlated, although the relationship was not so strong as to undermine the independence of any one system. Where there was disagreement between successive evaluations, it was in the direction of enhancing family integrity and parental access rights. We advanced three hypotheses to account for our findings: (a) changes in successive recommendations reflect the increasing sophistication of the assessment process; (b) changes reflect increasing distance from the family's ecology and are therefore increasingly ill informed; and (c) the changes are purely probabilistic, reflecting a drift toward the societal status quo.
Key Words--Judicial disposition, Child maltreatment, Mental health assessment.
INTRODUCTION THE ALLIANCE OF a jurist and mental health worker is a tenuous one, given fundamental systemic differences (Poythress, 1977; Ziskin, 1981). Substantively, the mental health professional, operating within a scientific framework, sees human behavior as determined; individuals do not necessarily have control of their actions. Jurists, on the other hand, are less concerned with science than morality. They view individuals as ethically culpable unless proven otherwise (Freckleton, 1986). Methodologically, the mental health professional is an empiricist, a disinterested observer. But in his/her quest for justice, the jurist depends on the adversarial system, a process that is far from dispassionate (e.g., Freckleton, 1986). Furthermore, most basic to the positivist approach is the concept of continued refinement of knowledge (Fromm, 1980). Mental health is a field of provisional diagnosis. Assessment is an ongoing process and a final verdict need never be reached. Unlike the clinic, "Courts must make decisions when cases come before them; they cannot wait for certitude" (Bazelon, 1977, p. 292). Received for publication August 9, 1993; final revision received February 8, 1994; accepted February 10, 1994. Requests for reprints should be addressed to Stephen Butler, Ph.D., 250 College Street, First Floor, Toronto, Ontario, Canada M5T 1R8. The order of authorship for Stephen Butler and Leslie Atkinson has been determined randomly. 355
356
S. Butler, L. Atkinson, M. Magnatta, and E. Hood
Of particular relevance here, impediments to the successful conjunction of mental health and legal systems arise with regard to child maltreatment (Melton & Limber, 1989). In this area, the judiciary regularly solicits recommendations from mental health professionals regarding the need and possibility of effective intervention; the advisability of out-of-home placement, and, where relevant; the probability of sexual abuse having occurred. The context within which the courts consider the recommendations provided is complex. On the one hand, factors that imply continued danger to the child, like poor prognosis for maltreating families (Melton, Petrila, Poythress, & Slobogin, 1987), ineffective treatment programs (Azar & Siegal, 1990; Dubowitz, 1990), and limited social service resources, press toward radical disposition. On the other hand, difficulties in defining child maltreatment, lack of stability in foster care placements, and sociolegal values pertaining to family integrity and parental liberty, push toward maintenance of the status quo. The decision to terminate parental rights can be one of the most difficult the court has to render (Melton et al., 1987). Despite theoretical discussion of the difficulties inherent in the alliance between mental health clinics and court, and despite the complexity of this collaborative endeavor, we are aware of no attempt to empirically evaluate the efficacy of this relationship. Specifically, there has been no effort to study the impact of mental health professionals' expert opinions on judicial dispositions in contested child maltreatment cases. This is surprising, given that mental health evaluations are considered crucial to these proceedings (Melton et al., 1987). The current investigation attempts to redress this failure. In addition, we study the issue of collaboration between child protection agencies and mental health facilities, and between child protection agencies and the courts. Again, we are aware of no empirical literature discussing these issues.
METHOD
Setting In the Toronto area, Children's Aid Societies are responsible for the protection of children exposed to abuse and/or neglect. In certain cases, danger to children is so great that involuntary removal from parental custody and placement under the protection of the province is sought through the jurisdiction of the Family Court. Once the child is taken into custody, the parents must make sufficient changes to justify the child's return home. If these changes are not forthcoming, the case continues through the Family Court system, with the child protection agency seeking permanent guardianship and the parents requesting that the child be returned to their custody. In a number of these contested child maltreatment cases, the judiciary seeks a mental health assessment from the Family Court Clinic (FCC). The FCC is a multidisciplinary division of a large urban psychiatric hospital. Under child welfare legislation, the FCC conducts court-ordered assessments of the parent(s) and children to assist the court in rendering disposition with respect to the children's custody and care (i.e., in the child's best interest). By law, a decision regarding the child's permanent custody must be made within 2 years of removal from the home. The FCC assessment usually transpires in the second year of this period after several attempts to improve parental care.
Subjects We reviewed clinical records of all cases assessed at the FCC, Clarke Institute of Psychiatry, during the period 1985-1990 involving a dispute for legal custody between a maternal caretaker and the child protection agency under the Child and Family Services Act. We have no reason to suspect that these cases were unrepresentative of contested maltreatment cases before the
Child maltreatment
357
court, but we cannot rule out referral bias. Ethical approval to review the charts was obtained from the Institute's Human Subjects Review Committee. Of the 71 assessments undertaken, 59 cases contained complete information. Of these 59 mothers (mean age = 28.7 years, SD = 7.7), 73% were applying for custody as single parents and 62% were on social assistance. In all cases, the children had been taken into custody by a child protection agency due to physical abuse, sexual abuse and/or neglect and all were residing in foster care. Of the 59 target children (mean age -- 5.8 years, SD = 3.4), 33 were female and 26 were male. This was a high-risk sample with elevated rates of family instability and transiency (60.3%), spousal violence (67.6%), alcohol abuse (maternal 33.7%; paternal 50%), drug abuse (maternal 22.4%; paternal 23.5%), and criminality (maternal 17.5%; paternal 67.6%). A substantial minority (35.6%) of mothers were noncompliant with the FCC assessment.
Procedure All files were coded as part of a larger study on child maltreatment (Butler, Radia, & Magnatta, 1994). The legal position of the child protection agency, FCC recommendations, and judicial disposition were extracted from the files. In families where more than one child was before the court, the disposition on the oldest child was selected for analysis. We coded these data on an ordinal scale from least to most restrictive disposition for the parent: I. Return home with a supervision order (child placed with mother subject to a period of supervision by the child protection agency); 2. Society Wardship (ongoing placement in foster care); 3. Crown Wardship With Access (child made a permanent ward of the state with the parent having visitation rights); 4. Crown Wardship Silent on Access (child made a permanent ward of the state where parental visitation is at the discretion of the child protection agency); 5. Crown Wardship No Access (child made a permanent ward of the state with no parental visitation; i.e., the child is legally qualified for adoption).
Analyses Spearman rho correlations were used to test agreement among the child protection legal position, FCC recommendation, and judicial disposition. We tested differences between the correlations with Fisher zr transformations for dependent correlations. To evaluate the degree to which combined child protection and FCC recommendations influenced judicial dispositions, we conducted a multiple regression analysis. We calculated binomial probabilities to determine the direction of changes in recommendations/dispositions (i.e., toward more or less restrictive interventions as cases moved from child protection to FCC to judiciary).
Resul~ The child protection agencies legal position, the FCC recommendation and the judicial disposition regarding the children's placement are shown in Table 1. In comparing outcomes across systems, the findings show large (Cohen, 1988) and significant correlations between child protection and FCC recommendations (rho -- .62, p < .0005), child protection recommendations and judicial dispositions (rho = .47, p < .0005), and FCC recommendations and judicial dispositions (rho = .79, p < .0005). The correlation between child welfare recommendations and court dispositions was significantly smaller than the correlation between FCC
S. Butler, L. Atkinson, M. Magnatta, and E, Hood
358
Table 1. Evaluation Outcome Regarding Child's Plaeement
Child Protection Agency Legal Position (N = 59)
Mental Health Assessment Recommendation (N = 59)
Recommendation
n
%
n
%
n
%
Return Home with Supervision Society Wardship Crown Wardship (with access) Crown Wardship (silent on access) Crown Wardship (no access)
7 5 18 2 27
11.9 8.5 30.5 3.4 45.8
I1 2 25 0 21
18.6 3.4 42.4 0 35.6
14 0 24 2 19
23.7 0 40.7 3.4 32.2
Judicial Disposition (N = 59)
recommendations and judicial dispositions, t(56) = 3.47, p < .01, 2-tailed. The correlation between child welfare and FCC recommendations was significantly larger than the correlation between child welfare recommendations and judicial dispositions, t(56) = 2.77, p < .01, 2 tailed. There was no significant difference when we compared the correlation between FCC and child protection recommendations with the correlation between FCC recommendations and judges' dispositions, t(56) = 1.47, ns. The analysis regressing FCC and child welfare recommendations on judicial disposition yielded R 2 = .63, F(2,56) = 46.74, p < .00005. FCC recommendations accounted for practically all of this variance, with the child welfare variable contributing a change in R 2 of .0008. Where there were disagreements between child protection, FCC, and judicial systems, there was a strong tendency for inconsistencies to be in the direction of preserving family integrity with successive assessments (e.g., moving from Crown Wardship No Access to Crown Wardship With Access). O f 15 disagreements between child welfare and FCC, 13 (87%) were in the direction of more conservative recommendations (p < .004). Similarly, when we compared child welfare recommendations and judicial dispositions, 17 of 21 (81%) disagreements were in the direction of maintaining parent-child relations (p < .004). Also, 7 of the 10 disagreements between FCC and the judiciary were in the conservative direction. However, the binomial probability test is insensitive here, with 9 of 10 findings in the same direction necessary for statistical significance.
DISCUSSION
Child Welfare and Mental Health The findings show a large and significant correlation between child protection and FCC recommendations, accounting for 38% of the variance. The correspondence between these two systems is not surprising; child welfare and mental health share similar philosophies about children's needs and are often staffed by professionals from similar disciplines. Additionally, child protection personnel complete their own assessments of each maltreating family before the court. This information is available to mental health professionals and is presumably influential in their evaluations. Nevertheless, 62% of the variance in FCC recommendations remains unexplained by access to child protection data. Several factors may explain the discrepancies in child welfare and FCC recommendations. Child welfare and mental health practitioners have different roles in the assessment process and their involvement with maltreating families occurs under different conditions. Child protection workers are mandated to investigate allegations and take immediate
Child maltreatment
359
protective action. These professionals are heavily influenced by factors associated with imminent and continued danger to the child such as ongoing maltreatment, poor prognosis for maltreating families, and limited social service resources. Such considerations often indicate apprehension of the child. Evidence must then be accumulated to justify out-of-home placement. These activities put child protection personnel in direct conflict with the parents' interests. This often results in a predominantly negative relationship between the parties, culminating in a legal dispute over the child at the dispositional phase of the court proceedings. The adversarial system may also serve to promote a state of cognitive dissonance for the child protection professional. By the dispositional phase, he/she may be reluctant to make recommendations that promote the relationship between parent and child when he/she has had to justify apprehension and ongoing foster placement. In contrast to the child protection worker, mental health practitioners undertake court-ordered assessments after the child has been placed in foster care. Under the auspices of the court, they engage the parent(s) as "neutral" professionals who perform a comprehensive assessment to assist the judiciary in determining what is in the child's best interests. With less emphasis on the child's immediate safety, the mental health practitioner is able to give greater consideration to factors that either mitigate the desirability of separating child and family (e.g., instability of foster care) or strengthen a less intrusive option (e.g., the value of family integrity). A significant preponderance of incongruities in child welfare and FCC recommendations is in the direction of preserving parent-child relations. This is consonant with the mental health practitioner's decreased emphasis on imminent risk to the child. Thus far, we have implied that the differences between FCC and child protection recommendations are attributable to the advantages enjoyed by FCC personnel and, perhaps, the resultant superiority of their assessments. This, after all, is why court clinics were established in the first place. Alternatively, it might be argued that the FCC mental health evaluation is conducted under artificial conditions, with parent and child forcibly separated, and with parents under legal counsel, and that these circumstances undermine a valid assessment. Thus, the discrepancies between child protection and FCC recommendations may be due to the artifactual, or ecologically invalid, nature of the latter.
Child Welfare and Judiciary The correlation between the legal position of child protection agencies and the judicial decision was large and significant, accounting for 22% of the variance. The child protection agency shares a mandate with the court to protect children and therefore this finding was reassuring. It is possible that the relationship was not stronger because, as mentioned above, the mandate and constraints of social service agencies (e.g., immediate protection of the child) contrast with those of the courts (e.g., balancing the rights of child and parent). Such an interpretation is consistent with the data showing that a significant proportion of the disagreements between child protection recommendations and judicial decisions are in the direction of less intrusive intervention. Specifically, disagreements between the systems were primarily from termination of parental rights to long-term foster care with access to the parents, and from foster care with access to return home with child protection agency supervision. This shift toward a more conservative disposition may reflect an attempt to balance the protection concerns of the child with preservation of the parent-child relationship. As such, it may be in accordance with child welfare legislation that attempts to promote the least detrimental alternative for the child. Another reason for the disagreements between child protection agencies and the court may lie in the fact that the court has additional, sometimes incompatible, recommendations available
360
S. Butler, L. Atkinson, M. Magnatta, and E. Hood
tO it (in the form of a mental health assessment). Thus, the correlation between child welfare recommendations and judicial dispositions was significantly lower than the correlation between child welfare recommendations and FCC recommendations. An alternative explanation for disagreements between child welfare agencies and the courts is that the judge evaluates the evidence under ecologically artificial circumstances. Judicial evaluation takes place, after parent and child have been separated for some time, in the chambers and courtroom. The assessment process is mediated by lawyers; a small proportion of the judge's evaluation is based on direct interaction with the litigants. Moreover, the aim of lawyers is to win the case, not to help the court discover facts (Marshall, 1980); this fact lies at the heart of the adversarial system. These factors may result in a decision that is less conducive to the child's welfare than that recommended by the social service agency and may explicate the discrepant recommendations of the two systems.
Mental Health and Judiciary The correlation between FCC recommendations and judicial dispositions was large and significant, explaining 62% of the variance. Furthermore, the correlation was significantly higher than that obtained between child protection recommendations and judiciary dispositions. This suggests that FCC mental health evaluations have an impact on judicial decision-making beyond information provided by child protection agencies. One factor that may account for this finding is the legally mandated role of court clinics vis-?t-vis the judiciary. The mental health evaluation is the "court's report" and receives "special evidentiary status." All parties must receive copies of the report and any party may summon the author as a witness and cross-examine that person, presumably as the court's witness (Wilson & Tomlinson, 1986, p. 112). It is possible that judges seriously consider mental health recommendations when the assessor is perceived (rightly or wrongly) to hold no bias toward any party and has developed specialized clinical expertise. Conversely, judges may perceive that social service personnel have less training and expertise, and they may place less value on their testimony in court. Although the correlation between the mental health and judicial systems was impressive, it was not so high as to suggest that judges indiscriminately accepted mental health opinion. This concern has been reported in other areas of forensic psychiatry such as civil commitment (Poythress, 1977) and is believed to undermine the legal process. Of particular relevance to the present context, "This danger becomes more real when the witness is actually called by the court, not by one of the partisan parties to the proceedings" (Freckleton, 1986, p. 180). Our data suggests that the judiciary is heavily influenced by FCC recommendations, but that a substantial proportion of the variance (38%) in court dispositions is attributable to factors beyond the mental health assessment. Importantly, the regression analysis showed that child protection recommendations contribute virtually no unique variance once FCC recommendations are accounted for. This is not to say that child protection recommendations do not influence the judiciary. Rather, their influence is indirect, mediated through court clinic reports that are themselves influenced by child protection data. Again, this demonstrates that the judiciary maintains substantial independence from expert testimony. The variance in judicial decision-making, which remains unexplained by expert recommendation may be accounted for by factors previously mentioned, such as philosophical differences between mental health and judiciary, differential weighting of child protection concerns and parental liberty, disparate ecological circumstances under which decisions are reached, or additional data available to judges at the time their ruling is made. Furthermore, the child protection agency and parent(s) may negotiate at the latter stages of the legal proceedings, considering the mental health recommendations and/or unforeseen developments that affect their legal positions.
Child maltreatment
361
Probabilistic Factors
Analyses showed that disagreements, as one moves from child welfare to mental health to justice, were consistently in the direction of maintaining parent-child relations. An alternative explanation to those offered above is that the changing pattern of recommendations may reflect a purely probabilistic, sociopsychological phenomenon, whereby a drift toward a common societal expectation or norm occurs with successive opinions regarding a decision. A professional may " a g r e e " with the preceding assessor's opinion, but find the recommendation slightly harsh. Changing decisions, then, may reflect an actuarial phenomenon similar to "regression toward the mean," rather than increasingly sophisticated, or unsophisticated, decision-making. In these contested cases of child maltreatment, the drift is toward maintenance of mother-child kinship, albeit in altered form. S u m m a ~ and Research Directions
We round significant agreement among child welfare, mental health, and judicial systems regarding recommendations/dispositions for child custody in serious cases of child maltreatment. The strongest agreement emerged between recommendations from the FCC mental health assessment and the eventual judicial disposition. Where disagreements occurred among the three systems, they were most frequently in the direction of protecting parental access rights. We suggested three hypotheses to account for these findings including; (a) The systems interact in well-balanced manner. Each system has discrete functions and specialized expertise, and at each level more extensive knowledge is available for decision-making. Thus, decisions improve as cases move from child welfare to court clinic to judiciary. (b) The correlations reflect increasing distance from the family and its natural ecology, such that decisions become increasingly ill-informed. (c) Changes in recommendations/dispositions represent a probabilistic drift toward the societal status quo. All these possibilities require further research. Detailed consideration should also be given to other issues emerging from this investigation. For example, it is important to examine the criteria upon which the judiciary base their decisions in contested child maltreatment cases. In this context, a recent study of a large sample (N -206) of serious child maltreatment cases before the courts found that permanent removal from the home was most strongly predicted by parental noncompliance with court-ordered services (Jellinek, Murphy, Poitrast, Quinn, Bishop, & Goshko, 1992). Limitations in this kind of knowledge have also been identified in other areas of forensic mental health evaluation such as juvenile delinquency (Niarhos & Routh, 1992) and contested divorce custody arrangements (Grisso, 1988; Kunin, Ebbesen, & Konecni, 1992). A better understanding of risk factors most relevant to children's welfare may facilitate the growth of more standardized child protection, mental health, and judicial assessments.
REFERENCES Azar, S., & Siegal, B. (1990). Behavioral treatment of child abuse: A developmental perspective. Behavior Modification, 14, 279-300. Bazelon, D. L. (1977). Can psychiatry humanize the law? Psychiatric Annals, 7(5), 29-39. Butler, S., Radia, N., & Magnatta, M. (1994). Maternal compliance to court-ordered assessment. Child Abuse & Neglect, 18(2), 203-211. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. New York: Academic Press. Dubowitz, H. (1990). Cost and effectiveness of interventions in child maltreatment. Child Abuse & Neglect, 14, 177186. Freckleton, I. (1986). Court experts, assessors and the public interest. International Journal ¢~f Law and Psychiatry, 8, 161-188.
362
S. Butler, L. Atkinson, M. Magnatta, and E. Hood
Fromm, E. (1980). Greatness and limitations of Freud's thought. Toronto, Canada: Fitzhenry & Whiteside. Grisso, T. (1988). Evaluating competencies: Forensic assessments and instruments. New York: Plenum Press. Jellinek, M. S., Murphy, J. M., Poitrast, F., Quinn, D., Bishop, S. J., & Goshko, M. (1992). Serious child mistreatment in Massachusetts: The course of 206 children through the courts. Child Abuse & Neglect, 16, 179-185. Kunin, C. C., Ebbesen, E. B., & Konecni, V. J. (1992). An archival study of decision-making in child custody disputes. Journal of Clinical Psychology, 48(4), 564-573. Marshall, J. (1980). Law and psychology in conflict (2nd ed.). Indianapolis, IN: Bobbs-Merrill. Melton, G. B., Petrila, J., Poythress, N. G., & Slobogin, C. (1987). Psychological evaluations for the courts: A handbook for mental health professionals and lawyers. New York: Guilford. Melton, G. B., & Limber, S. (1989). Psychologists' involvement in cases of child maltreatment: Limits of role and expertise. American Psychologist, 44(9), 1225-1233. Niarhos, F. J., & Routh, D. K. (1992). The role of clinical assessment in the juvenile court: Predictors of juvenile dispositions and recidivism. Journal of Clinical ChiM Psychology. 21(2), 151 - 159. Poythress, N. G., Jr. (1977). Mental health expert testimony: Current problems. Journal of Psychiatry and Law, 5, 239-243. Wilson, J., & Tomlinson, M. (1986). Wilson: Children and the Law. (2nd ed.). Toronto, Canada: Butterworth & Co. Ziskin, J. (1981). Coping with psychiatric and psychological testimony (3rd ed.). Manna Del Rey, CA: Law and Psychology Press.
REsumE----Les autoritEs responsables du bien-Etre de l'enfance, de la sant6 mentale et du syst~me judiciaire constituent un partenariat qui a rarement fait l'objet d'une analyse empirique. Cette Etude a donc examine la convergence dans les prises de positions 16gales des agences de protection de l'enfance, les recommandations de cliniques affectEes aux tribunaux et, ultimement, les dispositions judiciaires qui en dEcoulent. Pour ce faire, les chercheurs ont eu rcours un Echantillon de 59 cas contestEs de maltraitance. On a note une grande correlation entre les trois syst~mes au niveau des recommandations et des decisions de placer les enfants, mais cette correlation n'Etait pas 6troite au point de mettre en question l'autonomie de chaque syst/:me. Lh oil il y avait dEsaccord, c'Etait habituellement dans le sens des droits d'acc~s des parents et dans l e s s e n s de preserver l'intEgritE familiale. Les auteurs postulent trois hypotheses pour expliquer leurs dEcouvertes : (a) lorsque les decisions et les recommandations changent au long du processus judiciaire, ceci est sans doute du au fait que le processus de l'Evaluation est tr~s sophistiquE et qu'il Evolue h mesure que le cas progresse; (b) les changements refl~tent le fait que les syst~mes se distancient de la famille et les personnes qui y travaillent prennent des decisions basEes sur des donnEes inExactes et (c) les changements son purement une question de probabilitE, indiquant une tendance vers le statut quo dans la sociEtE. R e s u m e n - - L a alianza de los servicios de bienestar infantil, la salud mental, y los sistemas legales han recibido poca atenci6n empfrica a pesar de la magnitud de su impacto en los nifios y las familias. Examinamos la congruencia de las posiciones de las agencias legales de protecci6n infantil, las recomendaciones de la corte clfnica, y las disposiciones judiciales en una muestra de 59 casos procesados de maltrato a los nifios. Las recomendaciones/decisiones sobre la colocaci6n de los nifios fue la que obtuvo la m~is alta correlaciEn de los tres sistemas, a pesar de que la relaci6n no era lan fuerte como para disminuir la independencia de cualquiera de los sistemas. Donde habia desacuerdo entre sucesivas evaluaciones, fue en direcci6n a resaltar la integridad familiar y el acceso a los derechos parentales. Presentamos tres hip6tesis para explicar nuestros hallazgos: (a) Los camblos en las recomendaciones sucesivas reflejan la avanzada sofisticaci6n del proceso de evaluaci6n; (b) Los cambios reflejan aumento en la distancia de la ecologia familiar y por lo tanto est~in cada vez mas mal informados; (c) Los cambios son puramente probabilisticos, reflejando un movimiento hacia el status quo social.
Pergamon
0145-2134(94)00136-7
SPOTLIGHT
ON
PRACTICE
CHILD MALTREATMENT: THE COLLABORATION OF CHILD WELFARE, MENTAL HEALTH, AND JUDICIAL SYSTEMS STEPHEN BUTLER, LESLIE ATKINSON, MICHAEL MAGNATI'A,
AND ERIC
HOOD
Family Court Clinic, Clarke Institute of Psychiatry, Toronto, Ontario, Canada
Abstract--The alliance of child welfare, mental health, and legal systems has received little empirical attention, despite the magnitude of its impact on children and families. We examined the congruence of child protection agencies legal positions, court clinic recommendations, and judicial dispositions in a sample of 59 contested child maltreatment cases. Placement recommendations/decisionsamong all three systems were highly correlated, although the relationship was not so strong as to undermine the independence of any one system. Where there was disagreement between successive evaluations, it was in the direction of enhancing family integrity and parental access rights. We advanced three hypotheses to account for our findings: (a) changes in successive recommendations reflect the increasing sophistication of the assessment process; (b) changes reflect increasing distance from the family's ecology and are therefore increasingly ill informed; and (c) the changes are purely probabilistic, reflecting a drift toward the societal status quo.
Key Words--Judicial disposition, Child maltreatment, Mental health assessment.
INTRODUCTION THE ALLIANCE OF a jurist and mental health worker is a tenuous one, given fundamental systemic differences (Poythress, 1977; Ziskin, 1981). Substantively, the mental health professional, operating within a scientific framework, sees human behavior as determined; individuals do not necessarily have control of their actions. Jurists, on the other hand, are less concerned with science than morality. They view individuals as ethically culpable unless proven otherwise (Freckleton, 1986). Methodologically, the mental health professional is an empiricist, a disinterested observer. But in his/her quest for justice, the jurist depends on the adversarial system, a process that is far from dispassionate (e.g., Freckleton, 1986). Furthermore, most basic to the positivist approach is the concept of continued refinement of knowledge (Fromm, 1980). Mental health is a field of provisional diagnosis. Assessment is an ongoing process and a final verdict need never be reached. Unlike the clinic, "Courts must make decisions when cases come before them; they cannot wait for certitude" (Bazelon, 1977, p. 292). Received for publication August 9, 1993; final revision received February 8, 1994; accepted February 10, 1994. Requests for reprints should be addressed to Stephen Butler, Ph.D., 250 College Street, First Floor, Toronto, Ontario, Canada M5T 1R8. The order of authorship for Stephen Butler and Leslie Atkinson has been determined randomly. 355
356
S. Butler, L. Atkinson, M. Magnatta, and E. Hood
Of particular relevance here, impediments to the successful conjunction of mental health and legal systems arise with regard to child maltreatment (Melton & Limber, 1989). In this area, the judiciary regularly solicits recommendations from mental health professionals regarding the need and possibility of effective intervention; the advisability of out-of-home placement, and, where relevant; the probability of sexual abuse having occurred. The context within which the courts consider the recommendations provided is complex. On the one hand, factors that imply continued danger to the child, like poor prognosis for maltreating families (Melton, Petrila, Poythress, & Slobogin, 1987), ineffective treatment programs (Azar & Siegal, 1990; Dubowitz, 1990), and limited social service resources, press toward radical disposition. On the other hand, difficulties in defining child maltreatment, lack of stability in foster care placements, and sociolegal values pertaining to family integrity and parental liberty, push toward maintenance of the status quo. The decision to terminate parental rights can be one of the most difficult the court has to render (Melton et al., 1987). Despite theoretical discussion of the difficulties inherent in the alliance between mental health clinics and court, and despite the complexity of this collaborative endeavor, we are aware of no attempt to empirically evaluate the efficacy of this relationship. Specifically, there has been no effort to study the impact of mental health professionals' expert opinions on judicial dispositions in contested child maltreatment cases. This is surprising, given that mental health evaluations are considered crucial to these proceedings (Melton et al., 1987). The current investigation attempts to redress this failure. In addition, we study the issue of collaboration between child protection agencies and mental health facilities, and between child protection agencies and the courts. Again, we are aware of no empirical literature discussing these issues.
METHOD
Setting In the Toronto area, Children's Aid Societies are responsible for the protection of children exposed to abuse and/or neglect. In certain cases, danger to children is so great that involuntary removal from parental custody and placement under the protection of the province is sought through the jurisdiction of the Family Court. Once the child is taken into custody, the parents must make sufficient changes to justify the child's return home. If these changes are not forthcoming, the case continues through the Family Court system, with the child protection agency seeking permanent guardianship and the parents requesting that the child be returned to their custody. In a number of these contested child maltreatment cases, the judiciary seeks a mental health assessment from the Family Court Clinic (FCC). The FCC is a multidisciplinary division of a large urban psychiatric hospital. Under child welfare legislation, the FCC conducts court-ordered assessments of the parent(s) and children to assist the court in rendering disposition with respect to the children's custody and care (i.e., in the child's best interest). By law, a decision regarding the child's permanent custody must be made within 2 years of removal from the home. The FCC assessment usually transpires in the second year of this period after several attempts to improve parental care.
Subjects We reviewed clinical records of all cases assessed at the FCC, Clarke Institute of Psychiatry, during the period 1985-1990 involving a dispute for legal custody between a maternal caretaker and the child protection agency under the Child and Family Services Act. We have no reason to suspect that these cases were unrepresentative of contested maltreatment cases before the
Child maltreatment
357
court, but we cannot rule out referral bias. Ethical approval to review the charts was obtained from the Institute's Human Subjects Review Committee. Of the 71 assessments undertaken, 59 cases contained complete information. Of these 59 mothers (mean age = 28.7 years, SD = 7.7), 73% were applying for custody as single parents and 62% were on social assistance. In all cases, the children had been taken into custody by a child protection agency due to physical abuse, sexual abuse and/or neglect and all were residing in foster care. Of the 59 target children (mean age -- 5.8 years, SD = 3.4), 33 were female and 26 were male. This was a high-risk sample with elevated rates of family instability and transiency (60.3%), spousal violence (67.6%), alcohol abuse (maternal 33.7%; paternal 50%), drug abuse (maternal 22.4%; paternal 23.5%), and criminality (maternal 17.5%; paternal 67.6%). A substantial minority (35.6%) of mothers were noncompliant with the FCC assessment.
Procedure All files were coded as part of a larger study on child maltreatment (Butler, Radia, & Magnatta, 1994). The legal position of the child protection agency, FCC recommendations, and judicial disposition were extracted from the files. In families where more than one child was before the court, the disposition on the oldest child was selected for analysis. We coded these data on an ordinal scale from least to most restrictive disposition for the parent: I. Return home with a supervision order (child placed with mother subject to a period of supervision by the child protection agency); 2. Society Wardship (ongoing placement in foster care); 3. Crown Wardship With Access (child made a permanent ward of the state with the parent having visitation rights); 4. Crown Wardship Silent on Access (child made a permanent ward of the state where parental visitation is at the discretion of the child protection agency); 5. Crown Wardship No Access (child made a permanent ward of the state with no parental visitation; i.e., the child is legally qualified for adoption).
Analyses Spearman rho correlations were used to test agreement among the child protection legal position, FCC recommendation, and judicial disposition. We tested differences between the correlations with Fisher zr transformations for dependent correlations. To evaluate the degree to which combined child protection and FCC recommendations influenced judicial dispositions, we conducted a multiple regression analysis. We calculated binomial probabilities to determine the direction of changes in recommendations/dispositions (i.e., toward more or less restrictive interventions as cases moved from child protection to FCC to judiciary).
Resul~ The child protection agencies legal position, the FCC recommendation and the judicial disposition regarding the children's placement are shown in Table 1. In comparing outcomes across systems, the findings show large (Cohen, 1988) and significant correlations between child protection and FCC recommendations (rho -- .62, p < .0005), child protection recommendations and judicial dispositions (rho = .47, p < .0005), and FCC recommendations and judicial dispositions (rho = .79, p < .0005). The correlation between child welfare recommendations and court dispositions was significantly smaller than the correlation between FCC
S. Butler, L. Atkinson, M. Magnatta, and E, Hood
358
Table 1. Evaluation Outcome Regarding Child's Plaeement
Child Protection Agency Legal Position (N = 59)
Mental Health Assessment Recommendation (N = 59)
Recommendation
n
%
n
%
n
%
Return Home with Supervision Society Wardship Crown Wardship (with access) Crown Wardship (silent on access) Crown Wardship (no access)
7 5 18 2 27
11.9 8.5 30.5 3.4 45.8
I1 2 25 0 21
18.6 3.4 42.4 0 35.6
14 0 24 2 19
23.7 0 40.7 3.4 32.2
Judicial Disposition (N = 59)
recommendations and judicial dispositions, t(56) = 3.47, p < .01, 2-tailed. The correlation between child welfare and FCC recommendations was significantly larger than the correlation between child welfare recommendations and judicial dispositions, t(56) = 2.77, p < .01, 2 tailed. There was no significant difference when we compared the correlation between FCC and child protection recommendations with the correlation between FCC recommendations and judges' dispositions, t(56) = 1.47, ns. The analysis regressing FCC and child welfare recommendations on judicial disposition yielded R 2 = .63, F(2,56) = 46.74, p < .00005. FCC recommendations accounted for practically all of this variance, with the child welfare variable contributing a change in R 2 of .0008. Where there were disagreements between child protection, FCC, and judicial systems, there was a strong tendency for inconsistencies to be in the direction of preserving family integrity with successive assessments (e.g., moving from Crown Wardship No Access to Crown Wardship With Access). O f 15 disagreements between child welfare and FCC, 13 (87%) were in the direction of more conservative recommendations (p < .004). Similarly, when we compared child welfare recommendations and judicial dispositions, 17 of 21 (81%) disagreements were in the direction of maintaining parent-child relations (p < .004). Also, 7 of the 10 disagreements between FCC and the judiciary were in the conservative direction. However, the binomial probability test is insensitive here, with 9 of 10 findings in the same direction necessary for statistical significance.
DISCUSSION
Child Welfare and Mental Health The findings show a large and significant correlation between child protection and FCC recommendations, accounting for 38% of the variance. The correspondence between these two systems is not surprising; child welfare and mental health share similar philosophies about children's needs and are often staffed by professionals from similar disciplines. Additionally, child protection personnel complete their own assessments of each maltreating family before the court. This information is available to mental health professionals and is presumably influential in their evaluations. Nevertheless, 62% of the variance in FCC recommendations remains unexplained by access to child protection data. Several factors may explain the discrepancies in child welfare and FCC recommendations. Child welfare and mental health practitioners have different roles in the assessment process and their involvement with maltreating families occurs under different conditions. Child protection workers are mandated to investigate allegations and take immediate
Child maltreatment
359
protective action. These professionals are heavily influenced by factors associated with imminent and continued danger to the child such as ongoing maltreatment, poor prognosis for maltreating families, and limited social service resources. Such considerations often indicate apprehension of the child. Evidence must then be accumulated to justify out-of-home placement. These activities put child protection personnel in direct conflict with the parents' interests. This often results in a predominantly negative relationship between the parties, culminating in a legal dispute over the child at the dispositional phase of the court proceedings. The adversarial system may also serve to promote a state of cognitive dissonance for the child protection professional. By the dispositional phase, he/she may be reluctant to make recommendations that promote the relationship between parent and child when he/she has had to justify apprehension and ongoing foster placement. In contrast to the child protection worker, mental health practitioners undertake court-ordered assessments after the child has been placed in foster care. Under the auspices of the court, they engage the parent(s) as "neutral" professionals who perform a comprehensive assessment to assist the judiciary in determining what is in the child's best interests. With less emphasis on the child's immediate safety, the mental health practitioner is able to give greater consideration to factors that either mitigate the desirability of separating child and family (e.g., instability of foster care) or strengthen a less intrusive option (e.g., the value of family integrity). A significant preponderance of incongruities in child welfare and FCC recommendations is in the direction of preserving parent-child relations. This is consonant with the mental health practitioner's decreased emphasis on imminent risk to the child. Thus far, we have implied that the differences between FCC and child protection recommendations are attributable to the advantages enjoyed by FCC personnel and, perhaps, the resultant superiority of their assessments. This, after all, is why court clinics were established in the first place. Alternatively, it might be argued that the FCC mental health evaluation is conducted under artificial conditions, with parent and child forcibly separated, and with parents under legal counsel, and that these circumstances undermine a valid assessment. Thus, the discrepancies between child protection and FCC recommendations may be due to the artifactual, or ecologically invalid, nature of the latter.
Child Welfare and Judiciary The correlation between the legal position of child protection agencies and the judicial decision was large and significant, accounting for 22% of the variance. The child protection agency shares a mandate with the court to protect children and therefore this finding was reassuring. It is possible that the relationship was not stronger because, as mentioned above, the mandate and constraints of social service agencies (e.g., immediate protection of the child) contrast with those of the courts (e.g., balancing the rights of child and parent). Such an interpretation is consistent with the data showing that a significant proportion of the disagreements between child protection recommendations and judicial decisions are in the direction of less intrusive intervention. Specifically, disagreements between the systems were primarily from termination of parental rights to long-term foster care with access to the parents, and from foster care with access to return home with child protection agency supervision. This shift toward a more conservative disposition may reflect an attempt to balance the protection concerns of the child with preservation of the parent-child relationship. As such, it may be in accordance with child welfare legislation that attempts to promote the least detrimental alternative for the child. Another reason for the disagreements between child protection agencies and the court may lie in the fact that the court has additional, sometimes incompatible, recommendations available
360
S. Butler, L. Atkinson, M. Magnatta, and E. Hood
tO it (in the form of a mental health assessment). Thus, the correlation between child welfare recommendations and judicial dispositions was significantly lower than the correlation between child welfare recommendations and FCC recommendations. An alternative explanation for disagreements between child welfare agencies and the courts is that the judge evaluates the evidence under ecologically artificial circumstances. Judicial evaluation takes place, after parent and child have been separated for some time, in the chambers and courtroom. The assessment process is mediated by lawyers; a small proportion of the judge's evaluation is based on direct interaction with the litigants. Moreover, the aim of lawyers is to win the case, not to help the court discover facts (Marshall, 1980); this fact lies at the heart of the adversarial system. These factors may result in a decision that is less conducive to the child's welfare than that recommended by the social service agency and may explicate the discrepant recommendations of the two systems.
Mental Health and Judiciary The correlation between FCC recommendations and judicial dispositions was large and significant, explaining 62% of the variance. Furthermore, the correlation was significantly higher than that obtained between child protection recommendations and judiciary dispositions. This suggests that FCC mental health evaluations have an impact on judicial decision-making beyond information provided by child protection agencies. One factor that may account for this finding is the legally mandated role of court clinics vis-?t-vis the judiciary. The mental health evaluation is the "court's report" and receives "special evidentiary status." All parties must receive copies of the report and any party may summon the author as a witness and cross-examine that person, presumably as the court's witness (Wilson & Tomlinson, 1986, p. 112). It is possible that judges seriously consider mental health recommendations when the assessor is perceived (rightly or wrongly) to hold no bias toward any party and has developed specialized clinical expertise. Conversely, judges may perceive that social service personnel have less training and expertise, and they may place less value on their testimony in court. Although the correlation between the mental health and judicial systems was impressive, it was not so high as to suggest that judges indiscriminately accepted mental health opinion. This concern has been reported in other areas of forensic psychiatry such as civil commitment (Poythress, 1977) and is believed to undermine the legal process. Of particular relevance to the present context, "This danger becomes more real when the witness is actually called by the court, not by one of the partisan parties to the proceedings" (Freckleton, 1986, p. 180). Our data suggests that the judiciary is heavily influenced by FCC recommendations, but that a substantial proportion of the variance (38%) in court dispositions is attributable to factors beyond the mental health assessment. Importantly, the regression analysis showed that child protection recommendations contribute virtually no unique variance once FCC recommendations are accounted for. This is not to say that child protection recommendations do not influence the judiciary. Rather, their influence is indirect, mediated through court clinic reports that are themselves influenced by child protection data. Again, this demonstrates that the judiciary maintains substantial independence from expert testimony. The variance in judicial decision-making, which remains unexplained by expert recommendation may be accounted for by factors previously mentioned, such as philosophical differences between mental health and judiciary, differential weighting of child protection concerns and parental liberty, disparate ecological circumstances under which decisions are reached, or additional data available to judges at the time their ruling is made. Furthermore, the child protection agency and parent(s) may negotiate at the latter stages of the legal proceedings, considering the mental health recommendations and/or unforeseen developments that affect their legal positions.
Child maltreatment
361
Probabilistic Factors
Analyses showed that disagreements, as one moves from child welfare to mental health to justice, were consistently in the direction of maintaining parent-child relations. An alternative explanation to those offered above is that the changing pattern of recommendations may reflect a purely probabilistic, sociopsychological phenomenon, whereby a drift toward a common societal expectation or norm occurs with successive opinions regarding a decision. A professional may " a g r e e " with the preceding assessor's opinion, but find the recommendation slightly harsh. Changing decisions, then, may reflect an actuarial phenomenon similar to "regression toward the mean," rather than increasingly sophisticated, or unsophisticated, decision-making. In these contested cases of child maltreatment, the drift is toward maintenance of mother-child kinship, albeit in altered form. S u m m a ~ and Research Directions
We round significant agreement among child welfare, mental health, and judicial systems regarding recommendations/dispositions for child custody in serious cases of child maltreatment. The strongest agreement emerged between recommendations from the FCC mental health assessment and the eventual judicial disposition. Where disagreements occurred among the three systems, they were most frequently in the direction of protecting parental access rights. We suggested three hypotheses to account for these findings including; (a) The systems interact in well-balanced manner. Each system has discrete functions and specialized expertise, and at each level more extensive knowledge is available for decision-making. Thus, decisions improve as cases move from child welfare to court clinic to judiciary. (b) The correlations reflect increasing distance from the family and its natural ecology, such that decisions become increasingly ill-informed. (c) Changes in recommendations/dispositions represent a probabilistic drift toward the societal status quo. All these possibilities require further research. Detailed consideration should also be given to other issues emerging from this investigation. For example, it is important to examine the criteria upon which the judiciary base their decisions in contested child maltreatment cases. In this context, a recent study of a large sample (N -206) of serious child maltreatment cases before the courts found that permanent removal from the home was most strongly predicted by parental noncompliance with court-ordered services (Jellinek, Murphy, Poitrast, Quinn, Bishop, & Goshko, 1992). Limitations in this kind of knowledge have also been identified in other areas of forensic mental health evaluation such as juvenile delinquency (Niarhos & Routh, 1992) and contested divorce custody arrangements (Grisso, 1988; Kunin, Ebbesen, & Konecni, 1992). A better understanding of risk factors most relevant to children's welfare may facilitate the growth of more standardized child protection, mental health, and judicial assessments.
REFERENCES Azar, S., & Siegal, B. (1990). Behavioral treatment of child abuse: A developmental perspective. Behavior Modification, 14, 279-300. Bazelon, D. L. (1977). Can psychiatry humanize the law? Psychiatric Annals, 7(5), 29-39. Butler, S., Radia, N., & Magnatta, M. (1994). Maternal compliance to court-ordered assessment. Child Abuse & Neglect, 18(2), 203-211. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. New York: Academic Press. Dubowitz, H. (1990). Cost and effectiveness of interventions in child maltreatment. Child Abuse & Neglect, 14, 177186. Freckleton, I. (1986). Court experts, assessors and the public interest. International Journal ¢~f Law and Psychiatry, 8, 161-188.
362
S. Butler, L. Atkinson, M. Magnatta, and E. Hood
Fromm, E. (1980). Greatness and limitations of Freud's thought. Toronto, Canada: Fitzhenry & Whiteside. Grisso, T. (1988). Evaluating competencies: Forensic assessments and instruments. New York: Plenum Press. Jellinek, M. S., Murphy, J. M., Poitrast, F., Quinn, D., Bishop, S. J., & Goshko, M. (1992). Serious child mistreatment in Massachusetts: The course of 206 children through the courts. Child Abuse & Neglect, 16, 179-185. Kunin, C. C., Ebbesen, E. B., & Konecni, V. J. (1992). An archival study of decision-making in child custody disputes. Journal of Clinical Psychology, 48(4), 564-573. Marshall, J. (1980). Law and psychology in conflict (2nd ed.). Indianapolis, IN: Bobbs-Merrill. Melton, G. B., Petrila, J., Poythress, N. G., & Slobogin, C. (1987). Psychological evaluations for the courts: A handbook for mental health professionals and lawyers. New York: Guilford. Melton, G. B., & Limber, S. (1989). Psychologists' involvement in cases of child maltreatment: Limits of role and expertise. American Psychologist, 44(9), 1225-1233. Niarhos, F. J., & Routh, D. K. (1992). The role of clinical assessment in the juvenile court: Predictors of juvenile dispositions and recidivism. Journal of Clinical ChiM Psychology. 21(2), 151 - 159. Poythress, N. G., Jr. (1977). Mental health expert testimony: Current problems. Journal of Psychiatry and Law, 5, 239-243. Wilson, J., & Tomlinson, M. (1986). Wilson: Children and the Law. (2nd ed.). Toronto, Canada: Butterworth & Co. Ziskin, J. (1981). Coping with psychiatric and psychological testimony (3rd ed.). Manna Del Rey, CA: Law and Psychology Press.
REsumE----Les autoritEs responsables du bien-Etre de l'enfance, de la sant6 mentale et du syst~me judiciaire constituent un partenariat qui a rarement fait l'objet d'une analyse empirique. Cette Etude a donc examine la convergence dans les prises de positions 16gales des agences de protection de l'enfance, les recommandations de cliniques affectEes aux tribunaux et, ultimement, les dispositions judiciaires qui en dEcoulent. Pour ce faire, les chercheurs ont eu rcours un Echantillon de 59 cas contestEs de maltraitance. On a note une grande correlation entre les trois syst~mes au niveau des recommandations et des decisions de placer les enfants, mais cette correlation n'Etait pas 6troite au point de mettre en question l'autonomie de chaque syst/:me. Lh oil il y avait dEsaccord, c'Etait habituellement dans le sens des droits d'acc~s des parents et dans l e s s e n s de preserver l'intEgritE familiale. Les auteurs postulent trois hypotheses pour expliquer leurs dEcouvertes : (a) lorsque les decisions et les recommandations changent au long du processus judiciaire, ceci est sans doute du au fait que le processus de l'Evaluation est tr~s sophistiquE et qu'il Evolue h mesure que le cas progresse; (b) les changements refl~tent le fait que les syst~mes se distancient de la famille et les personnes qui y travaillent prennent des decisions basEes sur des donnEes inExactes et (c) les changements son purement une question de probabilitE, indiquant une tendance vers le statut quo dans la sociEtE. R e s u m e n - - L a alianza de los servicios de bienestar infantil, la salud mental, y los sistemas legales han recibido poca atenci6n empfrica a pesar de la magnitud de su impacto en los nifios y las familias. Examinamos la congruencia de las posiciones de las agencias legales de protecci6n infantil, las recomendaciones de la corte clfnica, y las disposiciones judiciales en una muestra de 59 casos procesados de maltrato a los nifios. Las recomendaciones/decisiones sobre la colocaci6n de los nifios fue la que obtuvo la m~is alta correlaciEn de los tres sistemas, a pesar de que la relaci6n no era lan fuerte como para disminuir la independencia de cualquiera de los sistemas. Donde habia desacuerdo entre sucesivas evaluaciones, fue en direcci6n a resaltar la integridad familiar y el acceso a los derechos parentales. Presentamos tres hip6tesis para explicar nuestros hallazgos: (a) Los camblos en las recomendaciones sucesivas reflejan la avanzada sofisticaci6n del proceso de evaluaci6n; (b) Los cambios reflejan aumento en la distancia de la ecologia familiar y por lo tanto est~in cada vez mas mal informados; (c) Los cambios son puramente probabilisticos, reflejando un movimiento hacia el status quo social.