Actionable Intelligence for Social Policy (AISP)
AISP is a University of Pennsylvania initiative that focuses on the development, use, and innovation of integrated data systems (IDS) for policy analysis and program reform using administrative data. AISP brings together state and local governments and their university and non-profit partners in a professional learning network, engages in federal advocacy to support data sharing to inform evidence-based policy, and conducts research using administrative data.
Location: Philadelphia, PA
Summary
Housed at the University of Pennsylvania, AISP works with partners in state and local government to help them design data systems that improve the quality of social programs, policies and practices. The sites in AISP’s network are each unique in organizational model and scope, but each has developed the capacity to integrate data, in other words, to link individual level data across multiple government agencies across the data life cycle. These integrated data systems (IDS) provide de-identified, linked research data sets to government agencies and external researchers to enable cross-agency analysis of service utilization and program outcomes. AISP provides these sites with training and peer-to-peer support and engages in advocacy to enable better and more responsible data collection, sharing and use by state and local authorities and to support collaboration between the public and academic sectors. The AISP network is made up of sites that link data representing over 26% of the U.S. population, and includes leaders in the field of integrated data systems, such as the Center for Innovation through Data Intelligence (CIDI) in New York City, the South Carolina Office of Research and Statistics, and Chapin Hall at the University of Chicago.
AISP also works to promote the expansion of IDS sites in the United States and the setting up of new sites to the end of promoting “actionable intelligence — an iterative, evolving process of data-based decision making that produces more effective and ethical responses to high-priority problems at the local, state, or federal levels.” AISP currently supports 15 developing sites that are in the process of establishing Integrated Data Systems of their own by providing them training, legal and technical expertise, and by facilitating peer-to-peer learning about running an IDS site.
Website: https://www.aisp.upenn.edu/
Sectors in which AISP network sites often link data for research purposes: Juvenile justice, homelessness, health and vital statistics, adult justice, behavioral health, education, assisted housing, workforce development, employment and earnings, and child welfare.
Target Audience: AISP’s primary audience is state and local government agencies, as well as universities involved in securely linking administrative data from multiple sources to enable evaluation of government programs and policies. Their secondary audiences include policy makers, federal agencies, and researchers who use linked administrative data. Some network sites also engage in data linkage to support the social and non-profit sector.
Services Offered: AISP¹ 1) provides technical and legal guidance and strategic consulting to partner institutions (state or local governments and university partners) who have or are seeking to set up “Integrated Data Systems (IDS)” that link data across agencies; 2) convenes a professional network of IDS sites and provides peer-to-peer learning opportunities; and 3) conducts multi-site research projects that demonstrate real-world uses of integrated data for policy and program improvement.
Capacity-building activities: The AISP network endeavors to build data capacity to enable states and counties to securely share data with other agencies or external researchers. They develop data sharing standards, collect best practices and provide technical assistance to sites that need it. For example, after conducting a survey of advanced IDS sites in the country, AISP identified the most successful implementations and analyzed the common features among them. This understanding of best practices enabled AISP to recommend an innovative set of technology standards to set up an IDS. This includes details about security features, data linkage and disclosure control. The findings acknowledge the challenges of developing an IDS but suggest they’re not insurmountable. AISP has also worked out data standards for future IDS sites — making it easier to use existing administrative data to conduct program evaluation in a more timely and cost-effective manner.
Knowledge-sharing across network sites: In its capacity as a facilitator of peer-to-peer knowledge sharing, AISP encourages each site to share best practices, communicate common barriers and share technical expertise with one another. It also commissions best practice reports and expert panels and provides training programs in the areas of data quality, data integration, benefit cost analysis, and ethical and legal issues.
Responsible and Ethical Data Use: Sharing data across systems is challenging as a result of legal, technical and cultural impediments. Federal and state laws, or the interpretation of those laws by agency legal counsel, may impede data integration in a well-intentioned effort to safeguard individual privacy. Even for those localities that are willing and have the expertise to navigate the legal requirements, integration is sometimes limited by the absence of knowledge of ethical and responsible data handling practices. Therefore, privacy protection concerns, particularly in connection with personally identifiable health and human services data are addressed by AISP in great detail in their expert panel reports. For instance, one of the expert panel reports is focused on governance issues around the setting up of an IDS — in particular, the ethical use of personally identifiable administrative data. The report makes wide ranging suggestions, from laying down principles or shared values for data sharing to engaging various stakeholders and creating formal/informal governance structures to deal with the various aspects of responsible data usage. This includes procedures for making data requests and the technical aspects of data transfer, security and integrity.
Addressing Legal Concerns: Along with the question of ethical data use, there are also legal concerns that arise when dealing with personally-identifiable data including those protecting people’s health data, education records and other personally identifiable data — all of which serve as important checks and balances to ensure that no individual’s privacy is compromised. In order to help network sites address these concerns, AISP commissioned an expert panel whose report, titled “Legal Issues for IDS use: Finding a way forward,” explains each of these laws and suggests specific ways in which the concerns regarding them can be addressed. It also makes noteworthy points about the exemptions provided within these laws, for example, for using data for research purposes, and provides resources to “develop the foundational legal documents,” including MOU (Memorandum of Understanding) and DUL (Data Use License), required to set up an IDS. Many of the legal concerns around using electronic data systems have also been addressed in AISP’s best practices resources. The report recommends, for example, that “direct identifiers” must be removed from a health and human services dataset in order for researchers to use them in compliance with HIPAA. While the reports are not substitutes for legal counsel, they show that there are ways to comply with the law and use these datasets responsibly in a manner and that these laws aren’t hindrances to setting up an IDS — rather, they help ensure that paramount importance is attached to privacy.
Multi-site Research Projects: AISP also works to generate interest in multi-site research projects and sub-contracts with Network sites. For example, AISP coordinated a multi-site study funded by the MacArthur Foundation designed to help city policy-makers understand the patterns and factors that predict crossover from child welfare to juvenile justice services. AISP Network sites in New York City, Cook Co., IL (Chicago), and Cuyahoga Co., OH (Cleveland) each utilized linked data from their own IDS but followed the same study design across sites. All three found that children who entered foster care for the first time after age nine were at a remarkably higher risk for later juvenile justice involvement relative to children with their first placement in infancy. Findings from the study not only informed changes in New York City’s strategy for targeting services to dually involved youth and older foster care youth, but also led the three sites to engage in further research collaboration aimed at developing the methodology for creating a national estimate for dually involved youth.
AISP Research: AISP faculty and staff also conduct their own research using linked administrative data. Municipal leaders from across multiple public-service systems in the city of Philadelphia commissioned Professor John Fantuzzo, co-director of AISP, to conduct a comprehensive study of publicly monitored factors in administrative databases that may contribute to the achievement gap between boys of different racial backgrounds. Using the Kids Integrated Data System (KIDS) in Philadelphia, the study brought together data from School District of Philadelphia, Department of Public Health, Department of Human Services, and Office of Supportive Housing. The study revealed that African American boys were experiencing early risks at a higher rate than white boys in the sample when controlling for poverty. Findings also indicate that African American boys lag in reading and mathematical achievement if they were maltreated, experienced high lead exposure, had mothers who did not graduate from high school, or had mothers with inadequate prenatal care. Leveraging the results of the study, the city of Philadelphia has developed Early Childhood Risk Indicators that are maintained in their integrated data system to inform and assess service provision.
Network Research: AISP Network sites also conduct their own research projects to drive program and policy decisions. Local and state health and human services (HHS) agencies typically maintain several siloed databases for the services they offer such as public welfare, homeless services, mental health and so on. Integrating these databases is intended to provide valuable insights about users who access more than one of these services and to better target relevant services to them. At the University of Chicago, a research center called Chapin Hall integrates data from seven HHS databases to understand the characteristics and usage patterns of multiple-service users in order to better target services to them. At the University of North Carolina, Charlotte, researchers combine data from the Homeless Management Information System (HMIS) and Veteran Services to study the service utilization of veterans experiencing homelessness to the end of understanding how they could be connected easily to veteran services. Integrated Data Systems, when used responsibly and ethically, can not only help cut costs but also improve the quality of service provided to the public.
Disseminating Findings: Research findings from AISP’s multi-site research, as well as findings from individual Network site’s research and findings from AISP expert panel reports are available on the AISP website for public access.
Funding resources: AISP is grant funded and currently receives support from the MacArthur Foundation, the Annie E Casey Foundation, and a federal SIF grant, among others. A grant from the Arnold Foundation supported AISP’s development of the aforementioned expert panel reports. Each partner site in the AISP Network has its own sources of funding. Executive- and agency-based models seek funding through government contracts and foundation grants. University-based models often depend on project-driven grants.
Contact details: Adelia Jenkins (adeliaj@sp2.upenn.edu)