Florida Department of Children and Families An Analysis of Increases in Out of Home Care: Executive Summary

Transcription

1 2016 Florida Department of Children and Families An Analysis of Increases in Out of Home Care: Executive Summary Mary Kay Falconer, Ph. D. Senior Evaluator Joe Anson, Ph.D. Senior Evaluator Ashley Schermerhorn, M.S.W., M.P.A. Research Assistant Kayla Sutherland, M.P.H. Research Assistant Research, Evaluation and Systems February 22, 2016

2 Analysis of Increases in Out of Home Care in Florida: Executive Summary National Trends and Challenges in Child Welfare Nationally, after a steady decline, an increase in the number of children entering care started around Fiscal Year (FY) Since then (between FY 2012 and FY 2014), the number of children entering foster care has increased by 5 percent. Other national data include: Between FY (most recent public data), the number of children involved in a screened in report of maltreatment remained relatively stable, although the number had been increasing steadily for several years prior. (Between FY 2009 and FY 2013, the number increased by 5%.) The rate of children involved in a screened in report of maltreatment is the highest it has been in many years (42.9 per 1,000 in FY 2013 compared to 40.3 per 1,000 in FY 2009). However, between FY 2009 and FY 2013 the number of children involved in a substantiated or confirmed screened-in report of maltreatment actually declined by 3%. As a result, the percent of children involved in screened in reports that are substantiated has declined. For example, in FY2009 about 23% of all children involved in a screened in report of maltreatment were substantiated. That percentage dropped to 21% in FY One of the reasons for the increase in children in foster care nationally is a slowed time to achieving legal permanency. For example, the timeliness of reunification measure from Round 2 of the Child and Family Services Reviews (C1.3) has steadily declined since FY 2009 (42%) and was 38% in the most recent year available (FY 2013). There are a number of challenges related to safely preventing the need for out-of-home placement, including the following (Pecora & Chahine, 2016): 1. Reporting laws: Ensuring that state laws regarding how to handle child maltreatment reports allow for agency discretion. 2. CPS hotline design and staffing: Enabling hotline staff members to use well-designed safety and risk assessment strategies to screen cases properly. 3. Retaining staff who have been trained and coached so that they can use skills effectively for engaging families, assess immediate safety and risk of future maltreatment, and select intervention strategies that could address family needs. Effective safety and risk assessments depend on the ability of CPS professionals to obtain accurate and timely factual information, as well as agency processes that promote critical analysis of that available information, increased leadership attention to staffing, career ladders, and staff development are essential because the new practice models and evidence-based interventions require a more highly skilled workforce with reasonable caseloads. 4. Community-based services: Ensuring that there is an adequate supply of the community-based services that are most needed by families at risk of child maltreatment such as behavioral health, domestic violence, home-visiting, housing, parent coaching and pre-natal services. Ounce of Prevention Fund of Florida 1

3 5. Leadership continuity: Ensuring continuity in agency policy and practice model as politically appointed leadership changes. 6. Judicial excellence: Having judges that understand the complexities of CPS practice and prioritize family preservation as long as child safety can be maintained. 7. Helping children placed in foster care achieve permanency. Once children are placed in out-ofhome care, they deserve to be helped to be reunified with family members, adopted or placed with a family with legal guardianship as quickly as possible. Across the country, agencies are striving to implement concurrent planning where more than one strategy is pursued, link families to the services that will most directly address the safety issues that brought their children into care, and address other barriers to securing permanency. This requires understanding why children are placed, where service gaps are located, and what sub-groups of children are experiencing delayed permanency and why. 8. Review of agency performance data. Regular review of agency quality and outcome data can provide feedback on areas of strength and areas for improvement. Performance data should be brought to bear to establish benchmarks and track improvement or make corrections to stimulate improvement. Trend data help clarify patterns and counteract distractions that might otherwise result from an occasional tragic event, bad quarter, or extraordinary year. When bolstered by sophisticated geo-mapping, heat mapping, and multivariate analyses (including predictive analytics), greater understanding of the true dynamics of system performance occurs. Unfortunately, media coverage of negative case events can often distract or de-rail careful planning and implementation of reforms as well as intimidate staff to place more children if the crises are not handled well by agency leaders (Turnell, Munro and Murphy, 2013). Agencies need a carefully designed communications plan, including a crisis management plan for such predictable exigencies. Development of prior political and media support is also essential to buffer human services administrators and staff from the day-to-day controversies of this work (Pecora, Whittaker, Barth, Maluccio and DePanfilis, 2009). Child welfare and community advocacy organizations can be excellent allies and sources of new ideas if communication is clear and flows both ways. Legislators can call for change. In addition, government-sponsored and independent policy research centers can provide objective analyses to help inform and drive change. (See the Washington Institute of Public Policy for examples of insightful reports at Even if many of these challenges are addressed well by a state or county, factors may be present that can affect the rate of foster care placements. This report documented increases in removals and the out-of-home care (OHC) population in Florida over a recent 24-month period (June 2013-May 2015). In doing so, it explored underlying causes and identified possible strategies to lower the rates. Investigation of Increases in Out-of-Home Care in Florida In order to monitor the performance of the system of care during the implementation of the Safety Methodology, the Florida Department of Children and Families (DCF) has been reviewing statewide and Ounce of Prevention Fund of Florida 2

4 circuit trends related to active investigations and child placements. Trends in investigations were examined as part of a Child Protection Investigator (CPI) workload analysis (Recent Increase in Number of Active Child Protective Investigations, issued in June, 2015). Preliminary analysis of the trends in investigations was conducted to identify some of the root causes and other contributing factors to the observed changes. Some of the factors explored were increases in incoming workload (new CPS investigations), caseworker turnover and resulting vacancies, caseworker inexperience, variation in the implementation of the Safety Methodology, allocation of CPIs by region and circuit that might not be meeting new caseload increases, and workload management from assignment through completion. In addition, child placement trends were presented and reviewed in a DCF report entitled, Community Based Care Lead Agency Trends and Comparisons: Caseloads and Use of Placement Resources (dated June 26, 2015). This report documented increases in removals and out-of-home care (OHC) population over a recent 24-month period (June 2013-May 2015). That early report also provided the following observations: 1. Removal rates and discharge rates are variable across the state, thus geographic targeting of strategies is essential 2. Circuits and lead agencies vary in terms of how often they protect children in their own homes vs. using out-of-home care, which can be more expensive 3. Circuits and lead agencies vary in terms of how often they place children in kinship care versus placing them in licensed non-relative foster care, which is often more expensive. 4. Circuits and lead agencies vary in terms of how often they place children in foster family homes vs. more expensive residential treatment and other forms of group care. Due to the increases in out-of-home care documented statewide since June 2013 and the variations observed across circuits, DCF was interested in expanding this area of inquiry. Specifically, DCF was interested in continuing the review and analysis of data available in Florida Safe Families Network (FSFN), as well as collecting additional information from CPIs and other child welfare professionals in the system of care. A more comprehensive approach (that included quantitative and qualitative research methods) was considered an opportunity to confirm preliminary reviews and analyses of trend data, and gain valuable insights from child welfare practitioners. Two primary research questions for the proposed study were the following: 1. What are some of the root causes and systemic factors in Florida in the last 24 months that are related to increased numbers in out-of-home care? 2. What strategies can be adopted by DCF, community-based care organizations (CBC) and other partners to safely address the number of children in out-of-home care and improve the system of care? As originally proposed, but with a few modifications made during the course of the project, the research methods for this study were the following: Ounce of Prevention Fund of Florida 3

5 1. Review evaluation documents available for other states that have implemented a practice model similar to the Action for Child Protection Model and examined changes in out-of-home care after implementation. If significant changes in child placements occurred in these states after implementation of the practice model, this review would inquire if there were recommendations developed to address these changes. Appropriate contacts in child welfare agencies or departments in five states were contacted and four states were interviewed. 2. Review Action for Child Protection s and Children s Research Center case reviews that were recently completed for DCF to compile added insights about the appropriateness of worker decision-making, and if the application of practice is in any way a factor in the rising number of out-of-home care placements. 3. Conduct an analysis of entries to out-of-home care in Florida in the past 24 months using quantitative and qualitative research methods to identify root causes and systemic factors that have had an impact on increases in out-of-home care placements. The project team was also asked to examine increases in the number of children in care due to delayed exits from foster care or discharges. In summary, the study utilized a mixed method design to enable the various data collection and analysis strategies to work together to: Inform the implementation of subsequent methods (exploratory) Explain the impact of factors that are identified in another method (explanatory) Allow comparisons of findings based on different methods (triangulation) Overall, it was believed that this design would lead to a more comprehensive understanding of trends for in-home and out-of-home care. The sequence and flow of the methodologies is displayed in Figure 1. There were four methodological phases in the project: (1) quantitative, (2) qualitative, (3) quantitative and (4) qualitative. Figure 1. Methodology Phases for Analysis of Increases in Out-of-Home Care PHASE 1 - QUANTITATIVE Regional and Circuit Trend graphs based on spinner data for: o In-Home/Out-of-Home Care o Removals o Discharges Regional percent change tables and figures for out-ofhome care Multivariate model that includes rates for removals, investigations and discharges to identify statistically significant impacts on change in out-of- home care PHASE 2 - QUALITATIVE Focus groups with multiple child welfare professionals with various roles from each region represented State Interviews Review of Case File Reviews PHASE 3 - QUANTITATIVE Develop a hierarchical logistic regression model that includes factors such as demographics, assessments, safety planning, CPI experience, safety management services, implementation of Safety Methodology and maltreatment FSFN data file that captures appropriate data items PHASE 4 - QUALITATIVE Interviews with judiciary Analysis of judicial interview data Complete a technical report Ounce of Prevention Fund of Florida 4

6 It is important to note that this project was not an evaluation of the Safety Methodology or an attempt to measure the model fidelity of the recent implementation of the Safety Methodology. However, there are many references in this analysis to the Safety Methodology and its components. The reason for this is to understand where possible root causes, contributory factors, and systemic factors are being manifested, to explain the relationships between root causes and the methodology, and to provide a platform for the development of recommendations that are relevant to current child welfare practice. Trend Analysis In Phase 1, this study examined and analyzed trends in out-of-home care, in-home care, removals and discharges over the past 24 months: July 2013 through April or June, The source for the data was the spinner reports available on the Center for Child Welfare website. The trends examined were for DCF regions, circuits and community-based care organizations in Florida. It is also important to note that Sheriff s offices conduct child protection investigations in six counties (Broward, Seminole, Pasco, Pinellas, Manatee and Hillsborough). The percentage change in out-of-home care, in-home care, removals and discharges over 6-month (January 2015 June 2015), 12-month (July 2014 June 2015), and 24-month (July 2013 June 2015) time periods were identified as appropriate preliminary indicators of trends across the six regions. Percentage change displays used categories for increases and decreases that identify small (0-10%), moderate ( %), and large ( 20.1%) changes. A list of summary observations appears below: 1. With one exception, increases in out-of-home care were observed in all of the regions for all time periods (24, 12 and 6 months). The exception was the Southern region for the most recent 6 months in which a small decrease was observed. It is also noteworthy that the increases for all but one region (Suncoast) for the 24 month time period were substantial ( 20.1%). 2. When reviewing removal and discharge changes, the picture shifts. Substantial increases ( 20.1%) in removals were observed for 4 of the 6 regions for the 24-month time period, and only 3 regions for the 12-month time period. In the 6 month time period, only one region, Northwest, had a substantial increase in removals but the other regions experienced a small increase or decrease (moderate and substantial). This could be one indication that there were successful attempts to address increases in removals during the latter part of the 24-month time frame for this examination. 3. Substantial increases in discharges occurred in all 6 regions for the most recent 6 months. This might indicate an elevation in attention focused on children in out-of-home care in order to accelerate legal permanency. With the variations in trends observed across time periods and across regions, the qualitative methods were viewed as a very valuable addition to this study. It was determined that the regional focus groups could produce data to describe and explain how a plethora of factors affected removal decisions and efforts to discharge children from out-of-home care. The quantitative method designed to analyze the statistical impact of several possible factors in a multivariate model was also considered an asset. Ounce of Prevention Fund of Florida 5

7 Introduction Regional Focus Groups with Child Welfare Professionals The qualitative data collected in nine focus groups across six regions of Florida during the last week of October 2015 was extensive. It contained important insights, explanations and perceptions regarding removals and discharges with the implementation of the Safety Methodology. The focus groups included representation of multiple child welfare professionals working with families and participating in the decision-making processes. Some of the participants were more experienced in their professional roles than others and some had implemented the Safety Methodology over a longer period of time. Removal Findings Based on the qualitative data collected in the regional focus groups, there was no consistent evidence that the Safety Methodology, as a child safety practice model, has increased or decreased removals. However, challenges were noted with the implementation and application of the Safety Methodology. Overall, it was learned that effective in-home safety plans have the greatest potential to safely prevent removals. There were several challenges related to the development and implementation of inhome safety plans. Factors that will strengthen in-home safety plans and enhance their potential to safely prevent removals are the following: 1. Appropriate and sufficient monitoring of the safety plans with informal supports or formal safety management services 2. The availability of appropriate evidence-based mental and behavioral health programs and other services as soon as possible at the beginning of a case 3. The participation and cooperation of the parents in the development and implementation of the safety plan 4. The assistance of an integrated or team approach among several child welfare professionals with relevant expertise in safety decision making 5. Reasonable caseloads that do not cause shifts in burdens between CPIs and CMs and between removals and in-home non-judicial cases 6. A CPI s ability to engage a family and skills that strengthen communication and knowledge of the Safety Methodology as well as available services 7. More cross training among child welfare partners (judges, GALs, CPT, and other) in the Safety Methodology 8. On-going training in some aspects of the safety methodology that are not being implemented consistently (danger threats and thresholds, safety plans) Ounce of Prevention Fund of Florida 6

8 Discharge Findings Based on the data collected in the regional focus groups, there is no consistent evidence that the Safety Methodology, as a child safety practice model, has increased or decreased the number of discharges or the length of time children are in out of home care. According to the focus group data, at this time, discharge decisions might be less impacted by the Safety Methodology and more judicially driven. There are several approaches that have been developed by DCF, the CBCs and other community partners to augment and/or expedite discharge decisions but these are not necessarily embedded in Safety Methodology practice statewide. Compared to removal decisions, the qualitative data also suggest that there might be more limited statutory guidance regarding the implementation of the Safety Methodology in reunification decisions. Overall, it was learned that effective in-home safety plans, frequent staffings that focus on the reunification of families, and the availability of an array of services that effectively change behavior have the greatest potential to increase discharges for reunification. In addition, what can help strengthen efforts to reunify families while maintaining the safety of children are the following: 1. Appropriate and sufficient monitoring of safety plans with informal supports or formal safety management services 2. Frequent communication between service providers and child welfare staff on the progress of the families in changing behavior (assessments and status) 3. The availability of appropriate evidence-based treatments and other needed services as soon as possible at the beginning and during the continuation of a case, including provision of in-home services after reunification 4. The participation and cooperation of the parents in the development and implementation of the safety plan and in-home services 5. The assistance of an integrated or team approach among several child welfare professionals with relevant expertise to consult during permanency and other staffings focused on reunification 6. Reasonable caseloads that do not cause shifts in burdens between CPIs and CMs or between out-of-home care and in-home care after reunification, 7. A CPI and Case Manager s ability to engage a family and skills that strengthen communication and knowledge of the Safety Methodology as well as available services 8. More cross training among child welfare partners in the Safety Methodology, including dependency court judges/magistrates, GALs, parent attorneys and service providers 9. On-going training in some aspects of the Safety Methodology that are not being implemented consistently (conditions for returns, safety analysis, and safety plans) Ounce of Prevention Fund of Florida 7

9 Multivariate Analysis of Removals Introduction The Phase 3 quantitative analysis used multilevel (hierarchical) logistic regression. This is a form of multivariate analysis appropriate for situations where the outcome is binary and the data has a hierarchical structure. In this case, the binary outcome is whether a removal occurred within sixty days of the investigation start date. The hierarchical structure of the data occurs because investigations fall under supervisors, while supervisors are situated within judicial circuits. In other words, investigators are nested under supervisors who are nested under judicial circuits. Based on the focus group study, we expected that some of the variation in removal decisions might be associated with supervisors and some might be associated with judicial circuits. Multilevel analyses take this situation into account. The investigations studied in this analysis all had received (start) dates between July 2013 and June 2015, a period during which the Safety Methodology was gradually implemented statewide. This analysis allowed us to compare investigations where the Safety Methodology was used to those where it was not, while controlling for other factors that would be expected to predict removal decisions, especially verified maltreatment findings and child demographics. The following caveats apply to the present analysis: 1. It is not an evaluation of the Safety Methodology, its fidelity of implementation, or of any aspect of the performance of DCF or the CBCs. 2. It is correlational and cannot definitively establish causal relationships. 3. It cannot establish whether or to what extent the increase in removals during the period studied was desirable or beneficial. For example, sometimes a high quality safety assessment model results in better assessments, greater specificity in documentation, and more children being placed if the needed community-based services are not available. Findings Findings of the multivariate analysis of removals suggest the following factors contribute to predicting removal: (1) implementation of the Safety Methodology; (2) maltreatment findings and types; (3) child demographics; (4) supervisors, and (5) judicial circuits. It was found that: 1. Use of the Safety Methodology was associated with the chances of removal increasing by between 8 percent and 26 percent. This association was stronger in the presence of verified neglect, substance misuse or family violence. This means that the Safety Methodology seems to have magnified the importance of these types of maltreatment on removal decisions. 2. Verified maltreatment was the most powerful predictor of removal in our models. Ounce of Prevention Fund of Florida 8

10 Abandonment was the maltreatment type most closely associated with removal, followed by neglect, substance misuse, physical abuse, sexual abuse and family violence. 3. Infants were about four times as likely to be removed as older children, controlling for other factors including verified maltreatment. (Infants special vulnerability may be a key factor; and the overall pattern is mirrored by national statistics.) 4. Black children were 51% more likely than other children to experience removal during the period studied. This is consistent with the disproportional representation of children of color in some foster care systems nationwide. The association between race and removal was smaller in cases where the Safety Methodology was used. In these cases, black children were only 29% more likely to experience removal than other children. 5. The association between verified family violence and removal was stronger for black children than for other children. Family violence increased the chances of removal more for black children than it did for other children. 6. The association between verified substance misuse and removal was weaker for black children than for other children. Substance misuse increased the chances of removal less for black children than it did for other children. 7. Judicial circuits and CPI supervisors were statistically significant predictors of removal in all of the models we tested. Almost twice as much variation in removal decisions was associated with CPI supervisors as was associated with judicial circuits. Our analysis suggests that, if there is ongoing interest in reducing unwanted variation in removal decisions, it would be more fruitful to focus on CPI supervisors than judicial circuits. Introduction Interviews with Dependency Court Judges As part of the comprehensive analyses conducted for this project, obtaining input from dependency court judges was considered essential. Qualitative data collected in the regional focus groups confirmed repeatedly that the judiciary is a key partner in child protection. There are many steps at which the CPIs, case managers, CLS attorneys and other partners come together before, during and after court hearings that have implications for removals to and discharges from out-of-home care. Interviews were conducted during the first half of December Six circuit judges and two magistrates participated in the interview sessions and represented six judicial circuits out of 20. There is a general concern in drawing conclusions that have implications for the dependency courts statewide from a small sample of dependency court judges. However, the sessions were a valuable supplement to the regional focus groups and helped clarify what was learned in the regional focus groups. Ounce of Prevention Fund of Florida 9

11 Findings The major findings based on four sessions with judges and magistrates in early January 2015 are summarized in three categories: (1) removals to out-of-home care, (2) discharges from out-of-home care (reunifications) and (3) training in the Safety Methodology. Removals to Out-of-Home Care 1. All judges interviewed asserted that statutory law is followed in removals to out-of-home care. The key references in statutory law are a determination of probable cause and a determination of whether there have been reasonable efforts to keep the child in the home prior to removal. 2. There were a variety of services mentioned by the judges that are considered appropriate as reasonable efforts to keep a child in a home prior to removal. However, there were differing opinions on whether the information in an affidavit regarding services was sufficient to make the determination of whether there were reasonable efforts. It was also not clear if a determination of reasonable efforts prior to removal consistently contributed to an increase or decrease in removals across all circuits. 3. According to some of the interviewed judges, there is more information about the family available with the Safety Methodology. 4. Inconsistency between the timing of updates in assessments (Family Functioning Assessment or FFA) and the timing of the arraignment was noted as a possible challenge. 5. Knowledge of the Safety Methodology varied across the participating judges. The judges were focused on the statutory requirements, and their knowledge of the Safety Methodology was limited. None of the judges claimed to be familiar with the details in the assessments regarding danger threats (present danger, impending danger or family functioning assessment). These documents are not filed with the court. Discharges from Out-of-Home Care (reunifications) 1. All judges interviewed asserted that statutory law, specifically substantial compliance with case plans, is followed in discharges out of out-of-home care which includes reunifications. 2. A focus on behavior change of the biological parents was mentioned by multiple judges. However, this focus on behavior change was not always considered a factor contributing to earlier reunification across circuits. 3. In one jurisdiction, the availability of a set of services referred to as a family reunification team was considered a factor contributing to earlier reunifications. 4. More frequent permanency staffings was mentioned as an effective strategy by more than one judge. It was agreed that staffings allowed for more attention and focus on a case. However, it was not clear that this was a factor leading to earlier reunifications across circuits. Ounce of Prevention Fund of Florida 10

12 5. The Safety Methodology term, conditions for return, was not recognized by multiple participants. Instead, the judges referred to substantial compliance. However, during the session discussions, conditions for return were thought to be very similar to conditions that are currently considered in reunifications. Training in Safety Methodology 1. There was unanimous support from the interviewed judges for more training and/or education of circuit judges and magistrates in the Safety Methodology. 2. There were a variety of suggestions offered regarding the amount of information that a judge or magistrate should know about the Safety Methodology. There were general references to a need for a better understanding of how decisions are made and the terminology that is being used. None of the participants thought that the judges and magistrates needed the level of detail required for those actually implementing the Safety Methodology. 3. If there is formal training, venue options for training were discussed. 4. If there is formal training, the interviewed judges agreed that the training team should include a circuit judge and a DCF representative. Summary of Findings and Recommendations Relying on both qualitative and quantitative methods, this study found that there were multiple causes and systemic factors that contributed to increases in out-of-home care. It also identified factors with the most noteworthy impacts and explained how they were manifested in the gradual implementation of the Safety Methodology over the past two years. The multiple causes interact in a system of care that has multiple partners, including investigators, case managers, service providers, the judiciary, and other stakeholders. In addition, decisions to safely remove children to out-of-home care and discharge children from out-of-home care, primarily for reunification, depend on a variety of resources, supports and skills. Lastly, some of the factors impact key decisions more directly while others have impacts that are indirect. In order to develop recommendations to improve and further inform how the removal and discharge decisions occur, some of the more important factors were highlighted. Beginning with the multivariate hierarchical analysis, it was determined that the implementation of the Safety Methodology had a modest impact on removals. While controlling statistically for several factors that were expected to predict removals, the model estimated that a proxy for the implementation of the Safety Methodology was associated with the chances of removal increasing by between 8 and 26 percent. The association between Safety Methodology and removal was larger in the presence of verified neglect, substance misuse or family violence. Child demographic factors were also important, with infants being about four times more likely to be removed and black children being 51 percent more likely to be removed. This association between race and removal was smaller (29%) in cases where the Safety Methodology was used. While enlightening in a number of respects, the multivariate analysis did not suggest that the Safety Methodology was the major driver in removal decisions. Ounce of Prevention Fund of Florida 11

13 In order to form a more in-depth understanding of causes and systemic factors, the findings based on the qualitative methods were valuable. The qualitative methods included focus groups with multiple stakeholders in six regions of the state and interviews with several dependency court judges and magistrates. From the perspective of those participating in the qualitative methods, there was no consistent evidence that the Safety Methodology was a direct cause of the increase in out-of-home care, through either removal decisions or discharge decisions. There were, however, several findings that identified challenges and possible deficiencies or limitations in the overall system of care that resulted in a removal or a child remaining in out-of-home care, perhaps longer than necessary. Regarding removals, the key findings focused on an effective in-home safety plan. In many cases, the effectiveness of a safety plan was not achievable and a child had to be removed. Regarding discharges, particularly for reunification purposes, an effective in-home safety plan continued to be important but it was also learned that frequent staffings that focus on the reunification of families and the availability of an array of services that effectively change parent behavior have key impacts in preparing a family for reunification. The participation of professionals in several different child protection roles with different expertise is essential along with the participation and cooperation of the parents. Communication between case management and the mental and behavioral health program providers was a link identified that must also be in place. The findings noted several gaps between the Safety Methodology and the resources required to implement it with fidelity. High caseloads were identified as a concern that might have impacted removal or discharge decisions. It was widely agreed that the time required to complete a family functioning assessment (FFA), which is a key component of the Safety Methodology, was higher. However, it was unclear if the time invested to collect the needed information was a factor leading to removals. Obtaining more information about a family was viewed in a favorable light, and it was thought that it would likely improve the capacity for making informed decisions. As another possible cause explored in the qualitative methods, media attention to child deaths, it was determined that media attention to child deaths probably increased the number of reports and led to a heightened level of awareness and caution that was exercised by all of the stakeholders. Some of the impacts identified in this analysis might be temporary or may persist through the initial implementation phase of the Safety Methodology and beyond. The recommendations address several issues that appeared to be the most urgent and noteworthy of attention now. They should augment efforts to allocate resources and point to ways to make the Safety Methodology more effective as a practice model. It is also important to keep in mind that more examination of the possible causes and systemic factors will continue to enhance the understanding of how and how much the factors studied in this analysis as well as others weigh into key decisions about the safety of children. Based on the findings of this study, the following recommendations are offered for consideration: 1. Trends in out-of-home care, in-home care, removals and discharges should continue to be monitored by geographic sub areas, including region, circuit, and CBCs and different time frames (24, 12 and 6 months). Several measures should be used to monitor these trends, including measures that use number of investigations as a denominator and those that have Ounce of Prevention Fund of Florida 12

14 child population as a denominator. These actions can help identify communities who are bright spots for Florida from which others can learn. Communities that are struggling can be analyzed more deeply for action planning. 2. DCF and the CBCs should conduct a review of training needs for the implementation of the Safety Methodology. The method of delivery of the training should be considered with an emphasis on coaching and more direct hands-on training in the field. Practices that were mentioned during the regional focus groups that merit special attention included preparation of safety plans, interpreting danger threats and thresholds, conducting a safety analysis and specifying conditions for return. The need for training that will improve skills to engage families as well as other skills important in practice should also be considered in this review. (Also, see recommendation #7 below.) 3. A Safety Methodology workload analysis should be conducted. This analysis should allow the identification of reasonable caseloads for investigations and case management when implementing the Safety Methodology. 4. Efforts to explore, pilot and evaluate activities and protocols that are designed to decrease removals and increase reunifications while maintaining the safety of a child should be encouraged. Information on these efforts should be documented using a standard approach and time frame. Successful efforts should be shared throughout the state. Examples of pilot activities that are appropriate include, at a minimum, the following: a. Safety decision-making teams that include the participation of multiple roles and expertise in assessing present and impending danger assessments b. Safety management services that stabilize the family in their home, facilitate the development of an in-home safety plan, and monitor the safety plan c. Special permanency staffings that occur more frequently and are focused on reunifications for parents who are cooperative and have demonstrated desired changes in behavior prior to completing services d. Family reunification teams that facilitate reunification and continue to work with the family after reunification for extended periods of time (several months) 5. A study of reasonable efforts prior to the removal of a child should be conducted. The importance of ensuring reasonable efforts have occurred prior to removal was elevated in the sessions with dependency court judges. A comprehensive understanding of what constitutes reasonable efforts will inform the local community service providers and child protection agencies in their continuing quest to prevent removals while ensuring the safety of the child. This study should identify the types and durations of services that meet the statutory requirements as well as variations that might occur throughout Florida. 6. CBCs should conduct a comprehensive review of behavioral and mental health therapies, parent education and other services for families. The overall goal of this review is to ensure that required changes in parental behavior and mental health are addressed. The review should Ounce of Prevention Fund of Florida 13

15 include the development of an inventory of current behavioral and mental health services and their status as an evidence-based program or model. Within the available resources, adjustments that could be made to expand the number of evidence-based programs available should be identified. Evaluation of the implementation of the programs to determine model fidelity as well as effectiveness should be included. Approaches to improve the coordination and communication between case managers and therapists/service providers should also be identified. 7. The Statewide Dependency Court Improvement Panel should identify specific education and training needs regarding the Safety Methodology. The education and training needs identified might be appropriate for dependency court judges/magistrates as well as CPIs, CLS attorneys and Case Management. At a minimum, the focus should include a comparison of state and federal legal requirements and Safety Methodology practice requirements. Differences in terminology should be noted. Ways that the court can use information collected in the Safety Methodology FFA timely and efficiently should be highlighted. This coordinated review of training and education needs might uncover a need for statutory clarification or elaboration but, at this time, there is no assumption that amendments to the statutes are needed. (Also, see recommendation #2 above.) 8. Hierarchical logistic regression models that analyze investigation data from FSFN should continue to be estimated. The analysis should expand and improve the selection of variables and related data included in the model analyzed in this study. 9. Because of the ongoing interest in reducing unwanted variation in removal decisions, a special focus in current training of CPI supervisors might be fruitful. When comparing the variation in removals detected for CPI supervisors and judicial circuits in the hierarchical logistic regression model estimated in this project, the variation in CPI supervisors was greater. Conclusion Relying on both qualitative and quantitative methods, this study found that there were multiple factors that may have contributed to the increase in removals to out-of-home care and perhaps to delays in discharges from out-of-home care. There was evidence that the implementation of the Safety Methodology presented some challenges. However, there was no consistent evidence that the Safety Methodology as a practice model contributed to the increase of children in out-of-home care. Numerous factors associated and not associated with the implementation of the Safety Methodology were documented through the process of conducting regional focus groups with a diverse set of child welfare professionals and interviews with members of the judiciary. In addition, the multivariate hierarchical analysis confirmed a modest impact by the implementation of the Safety Methodology as well as other factors including certain demographic findings related to child age and race. Ounce of Prevention Fund of Florida 14

16 References Pecora, P.J. & Chahine, Z. (in press). Catalysts for child protection reform. Child Welfare. Pecora, P.J., Whittaker, J.K., & Maluccio, A.N., Barth, R.P., & DePanfilis, D. (2009). The child welfare challenge: Policy, practice, and research. (Third Edition.) Piscataway, NJ: Aldine-Transaction Books. Turnell, A., Munro, E., & Murphy, T. (2013). Soft is hardest: Leading for learning in child protection services following a child fatality. Child Welfare, 92(2), Ounce of Prevention Fund of Florida 15