Complex issues — such as those encountered in child welfare — require complex solutions, often necessitating cross-system data sharing and collaboration. Improving the safety and well-being of children and families requires a vision and tools beyond the scope of child protection. It requires a concerted and coordinated effort across multiple agencies — as well as a much higher level of partnership with impacted communities — to address the root causes of family stress that lead to higher risks of child maltreatment. Using data to inform this strategy is paramount.^1,2

Traditional child welfare practice focuses on individual-level accountability, often largely ignoring community conditions such as poverty, crime, and inequities in education that impact the well-being of children and families, particularly those in traditionally underserved communities including communities of color. If child welfare continues to function in a silo, community-level problems will continue to impact children and families and high rates of child welfare involvement will persist. 

Sharing data can inform a public health approach to child welfare, situating child protective services (CPS) within a broader continuum of family support and aligning the efforts of multiple agencies. Utilizing data from across systems — such as vital records, healthcare, housing, public benefits, education, and the courts — can help communities better understand areas of strength and need, informing community-based approaches to increase child safety. 

Foundational elements

Implementing data sharing across systems can be a complicated, time-consuming process. Building collaborative relationships, developing safe data transfer and storage protocols, and creating and finalizing data sharing agreements all are necessary steps. Starting with a relatively small, simple project using shared data can lay the groundwork for successful larger, more complicated projects in the future. 

Build in time for creating strong working relationships 

Sharing data with other agencies and systems can feel risky. It naturally involves some degree of vulnerability, as data may reveal areas where processes may not be as effective and outcomes may not be as positive as envisioned. Taking time to build effective working relationships with representatives from other agencies — including the development of shared norms, a joint vision, and common goals — is both a critical first step and an ongoing process. Maintaining trust involves ensuring that products created using shared data are relevant, explains researcher Hilary Shager, associate director of programs and management for the Institute for Research on Poverty at University of Wisconsin–Madison. “The time, effort, and investment has to go into building the relationships, building the structures where partners have an opportunity to review results,” she says. “We have learning exchanges with (the child protection agency) where we have faculty who are using the data for research come and present. It’s really a partnership. We are always leading with, ‘How is this going to be useful for you?’ We design the evaluation and research to be responsive to that from the get-go.”

Develop data sharing agreements and transfer processes

Developing data sharing agreements and data transfer processes — and getting those agreements and processes approved and implemented — can take a long time. Project timelines should reflect that. Similar to child protection agencies, some partners may have legal restrictions of their own (for example, educational institutions are all bound by the federal privacy provisions of FERPA). Once one data sharing agreement is developed, however, developing subsequent data sharing agreements with similar partners may occur more quickly. 

Connecticut’s report on legal issues in interagency data sharing includes practices to facilitate data sharing and provides summaries of federal laws impacting data sharing in child welfare, criminal justice, drug and alcohol use disorders, early childhood, education, health, homelessness, mental health, social services, and workforce development. Based on an analysis of survey data and existing data sharing policies from across state agencies, a review of relevant laws and regulations, and consultation with state and national experts, the report offers two primary recommendations to facilitate a data-driven approach:

• Establish a statewide structure for sharing data across agencies.
• Develop data sharing agreements that offer flexibility to protect confidentiality and increase efficiency. 

Safeguard data and remove identifying information when possible

Concerns about confidentiality are valid when linking to datasets and systems outside of child welfare. When sharing data across systems and making connections between datasets, certain safeguards are critical. These include: ensuring that data are securely stored; limiting the number of fields collected and shared; and limiting who can access the data. Obscuring birthdates and removing contact information are two straightforward protections. In addition, creating new identifiers can help protect personal information — new identifiers can be added to each data file that is to be shared, and prior to sharing each data file, the original identifiers can be deleted. This allows for tracking over time, as data can be updated and integrated using the common identifier. 

Wisconsin offers a compelling model for cross-system data sharing, including its practices related to confidentiality. The Wisconsin Administrative Data Core (WADC) at the University of Wisconsin-Madison’s Institute for Research on Poverty uses a multi-layer secure data system that stores identifiable data from state agencies. Programming staff, who are the only people with access to these files, remove all identifying information before sharing files with researchers.

Consider starting with a pilot project

Starting with a small project can allow partners to identify and overcome challenges and experience success before embarking on more complicated projects. When bringing on a new partner, it can be helpful to begin with a small, concrete pilot project. A one-project data sharing agreement should be structured similarly to a broader agreement, making it easier to expand the scope if the partnership continues. WADC began with a series of individual projects that required data sources from multiple partners. Over time, as the number of collaborations increased and the value of shared data was repeatedly demonstrated, WADC was formally established as a resource with dedicated staff and regular data maintenance.

Consider data hosting by a neutral third party

WADC is able to serve as a third party because it is hosted within the neutral Institute for Research on Poverty at the University of Wisconsin-Madison. State agencies, which may have concerns about sharing data directly with other state agencies, instead share data with WADC, which integrates and de-identifies the data. “There are a lot of times when we’re a very useful intermediary,” says Steve Cook, a researcher for the Institute. “The negotiating among state agencies — especially when it involves more than two agencies — from a bureaucratic standpoint, from a legal standpoint, and from a trust standpoint is often very hard to do. We serve a role there that would be hard for another state agency to do.” The Institute’s technical report on lessons learned in the development of the WADC can inform the development of similar databases in other jurisdictions.

Key strategies for impact

Sharing and using data across systems can help agencies develop upstream strategies that focus on the prevention of child maltreatment. Agencies may find it helpful to visualize data through maps to better understand areas of need and strength. Involving community members in the interpretation of findings and development of solutions is crucial and helps to ensure that the overall approach is relevant to the community. Sharing raw data — and questions posed by the data — is not enough, however. The findings and implications from second-level data analyses also should be shared with community members so they can continue to have influence over the direction of child welfare locally.

Extend data collection and analysis upstream

Current federal child welfare data collection practices, such as the Adoption and Foster Care Analysis and Reporting System (AFCARS) and the Child and Family Services Reviews, focus on individual-level data collected after families are known to CPS. While the collection of post-contact data is vital, it does not inform or advance prevention efforts. The collection and analysis of community-level data, however, accomplishes this goal. 

Community-level data can expose resource gaps and needs, helping communities plan for upstream investments —including providing economic supports to ensure a secure and adequate income, stable housing, quality education, child care, and safe outdoor spaces. These may be multi-level interventions, benefiting individuals, families, neighborhoods, and the larger community. 

The City of Philadelphia’s Department of Human Services (DHS) found that 93% of reports made to its child protection hotline were ultimately not accepted for ongoing formal safety service. The agency analyzed the data to determine why that rate was so high. As DHS was examining disproportionate reporting based on race/ethnicity, it found strong associations between reporting rates and poverty, vacant land (a proxy for community disinvestment), and historical redlining. The team shared these findings with stakeholder groups throughout the city and brought together partners from the Department of Behavioral Health, Department of Housing, Community Economic Opportunities, Public Health, the school district, and major children’s hospitals to plan next steps. 

“We didn’t stay at the individual level,” explains Allison Thompson, a senior research officer for the department. “We asked, ‘What structures do we need to adjust so that we don’t have such high rates of reporting to our hotline for concerns related to poverty stressors?’ We applied for and received the Family Support through Primary Prevention grant through the Children’s Bureau. We are funded to build up an alternative to the DHS hotline — the support line. We’re really excited that our study looking at neighborhood-level and structural-level factors led to a solution that changes the structural landscape and creates an alternative to the CPS hotline for families with non-safety, wellbeing concerns.”

In California, Safe & Sound’s Data Playbook for Prevention Action Planning was created to help county leaders and community partners develop effective prevention plans that are informed by data. It includes information on selecting a data framework, gathering and analyzing data, and sharing stories and results.

Visualize data through maps

Currently, resources that examine federal and state measures of child, family, and community well-being are limited. To bolster this base of data, child protection agencies and researchers could collect the addresses of families reported to CPS, based on census tract or ZIP code, as a method to analyze community risk factors and determine the most urgent well-being needs, Indicators of community risk and protective factors include measures of neighborhood economic and housing stability, social stability, and the built environment. A starting point for examining community-level indicators is Casey Family Programs’ Community Opportunity Map, which provides census tract-level data on demographic characteristics, child and family well-being, education, economy, housing, access to the internet, access to healthy food, and uptake of Supplemental Nutrition Assistance Program (SNAP) benefits.

New York City’s Citizens’ Committee for Children provides data on demographic characteristics, economic conditions, housing and homelessness, early care and education, K-16 education, youth and juvenile justice, child welfare, and community safety. The data can be broken down by age, race/ethnicity, and gender as well as by location (community district, school district, ZIP code, and police precinct). In addition, its map of community resources focuses on assets related to economic security, housing, health, education, youth, and family/community.

Involve community members

Involving community members in sharing their experiences and in interpreting findings is crucial in understanding program strengths and weaknesses. Sharing data is necessary, but taking the next step to develop and implement strategies with community members to address problems — based on that data — is even more essential. In Louisiana, My Community Cares involves residents of neighborhoods that have the highest rates of child maltreatment reports and children in out-of-home care to develop programs and strategies to better support families.

Develop and use positive community indicators

Most indicators of child, family, and community well-being are deficit-based (for example, the percentages of children living in poverty or individuals with no health insurance). More research should be done on using strength-based indicators of protective factors. Data should be collected on positive elements of children’s well-being, including cognitive and physical development, emotional stability, and social connections. 

Plan for effective communication of findings

A lot of time is spent on accuracy and validity, but not enough time is often spent on communicating findings and getting the messages out. The City of Philadelphia gears its communications toward action. “Every time we present research to any audience, we have actionable implications. We consider our audience when selecting actionable implications to highlight, and we have pushed ourselves to take an ecological approach with our suggested solutions to include policy and systems-level changes in addition to child- or family-centered interventions.” explains Eliza Ziegler, project manager for Philadelphia’s Office of Children and Families.

Be curious and relentless

While setting up processes for a shared data system can be time-consuming and sometimes frustrating, the potential benefits for child, family, and community well-being far outweigh any logistical challenges. Learning from other jurisdictions can help. Actionable Intelligence for Social Policy, a think tank housed at the University of Pennsylvania, helps local and state governments collaborate through data sharing, provides descriptions of jurisdictions’ experiences with data sharing, and offers a training and technical assistance program to center racial equity in data integration. 

Selected resources

Data Sharing Across Child-Serving Sectors: Key Lessons and Resources (Nemours Children’s Health System and Mental Health America, 2019)

• While this issue brief focuses on data-sharing partnerships across health, education, and early childhood sectors, many of the key lessons learned and resources shared are relevant to data sharing in child welfare.

Data Sharing: Courts and Child Welfare (Children’s Bureau, 2018)

• This document discusses benefits, challenges, and considerations for sharing data between courts and child welfare agencies. Models for exchanging data and a sample memorandum of understanding are included. 

Data Sharing in Child Welfare (Quality Improvement Center for Adoption & Guardianship Support and Preservation, 2018)

• This report shares the experiences and lessons learned from eight sites that established data use agreements to support the evaluation of interventions to support permanency.

Data Sharing Resources (National Data Archive on Child Abuse and Neglect, 2018)

• This resource list links to resources relevant to data sharing in child welfare. 

Introduction to Data Sharing & Integration (Actionable Intelligence for Social Policy, 2020)

• This guide discusses the benefits, purposes, limitations, and risks of data sharing and integration, and includes considerations for agencies as they plan for and implement data sharing and integration.

Roadmap for Foster Care and K-12 Data Linkages (Data Quality Campaign and Legal Center for Foster Care and Education, 2017)

• This document discusses the importance of linking K-12 education data and foster care data, provides recommendations for how to implement data sharing, and includes case studies of jurisdictions that have done it successfully.

Secure Shouldn’t Mean Secret: A Call for Public Policy Schools to Share, Support, and Teach Data Stewardship (Georgia Policy Labs, 2019)

• This white paper discusses the importance of sharing data across agencies, describes necessary components to keep data safe, provides examples of secure data access models, provides a resource list of organizations that have successfully set up data sharing, and describes the importance of training researchers to be good stewards of data.

1 This brief is based on interviews with: Samantha Rivera Joseph, Director of Implementation Science, City of Philadelphia Office of Children and Families, Allison Thompson, Senior Research Officer, City of Philadelphia Department of Human Services, and Eliza Ziegler, Project Manager, City of Philadelphia Office of Children and Families, November 23, 2021; and Steve Cook, Researcher, Institute for Research on Poverty, University of Wisconsin-Madison and Hilary Shager, Associate Director, Institute for Research on Poverty, University of Wisconsin-Madison, August 2, 2022. Portions of this brief were abstracted from an unpublished report by the Child Welfare Data Leaders.
2 Content of this brief was informed by consultation with members of the Knowledge Management Lived Experience Advisory Team on April 18 and May 6, 2022. This team includes youth, parents, kinship caregivers, and foster parents with lived experience of the child welfare system who serve as strategic partners with Family Voices United, a collaboration between FosterClub, Generations United, the Children’s Trust Fund Alliance, and Casey Family Programs. Members who contributed to this brief include Aleks Talsky, Roberto Partida, and Keith Lowhorne.