Using Linked Administrative Data to Inform Decisions Actionable - - PowerPoint PPT Presentation

Using Linked Administrative Data to Inform Decisions

1

Actionable Intelligence for Social Policy (AISP) APDU June 18, 2018

Build Buildin ing the technic ical l and huma uman capacit ity for for imp improv

• ved soc
• cia

ial l policy cy so go gove vernment ca can work better, smarter and faster to se serve communiti ties. s.

AISP’S MISSION

Data Data sh shar aring is s as as relati ational al as as it t is s te technical al. We We don’t just need eed to integ egra rate e data, we we need to

• integrate peop
• ple.

OUR PHILOSOPHY

WHAT WE DO

q Convene a prof profession ional al network

• rk for local and state

governments working on data integration to share best practices and problem-solve together q Engage in adv advoc

• cac

acy on behalf of data sharing at the federal, state, and local level q Provide free re resourc urces an and d sam ampl ple do docum uments on data governance, legal considerations, data standards, and linkage technologies q Offer a formal train rainin ing an and d technic ical al as assis istan ance program to help interdisciplinary teams build capacity

4

qLink administrative data across at least three domains/agencies (not just across several programs within an agency) qServe as a public utility (not research for research’s sake) qHave a clear organizational home & defined governance structure (not one-off projects)

OUR FOCUS IS ON EFFORTS THAT:

AISP NETWORK 2018

6

Learning Community TA

Currently working with 15 sites across 2 cohorts to develop IDS capacity

AISP NETWORK 2020

8

What can IDS do?

9

IDS Help Governments & Research Partners:

Support a Master Client Index and Record Reconciliation Across Departments Link Individuals Within a Family Unit or Household Create Longitudinal Cohorts Understand & Address Complex Social Problems

10

Complex Social Problems States are Addressing with IDS:

Superutilizers in Healthcare Educational Achievement Gaps The Opioid Crisis Two-Generational Poverty

11

Intervention model building Descriptive epidemiological study of a social problem Intervention model testing

Stages of IDS Use

Example: IDS in Action

13

City of Philadelphia Mayor’s Office of Data Management

Better understand the cumulative impact of early childhood risk factors on children in Philadelphia

PHASE 1: ANALYSIS

Getting the “Right” Data

Vital Statistics Child Welfare Homeless Shelter System Public Health School District Child maltreatment Inadequate prenatal Care Low maternal education Teen mother Low birth weight/ preterm birth Lead exposure Shelter stays Academic & Behavioral Outcomes

Top 3 most harmful risks for educational

• utcomes

Reading

• 1. Low maternal education

2.High lead exposure

• 3. Child maltreatment

Classroom Engagement

• 1. Child maltreatment

2.Low maternal education

• 3. Homelessness

Truancy

• 1. Low maternal education

2.Child maltreatment

• 3. Teen mother

Rouse & Fantuzzo (2009); Rouse, Fantuzzo & LeBoeuf (2011); Fantuzzo, LeBoeuf & Rouse (2014)

Unique Influence of Risks

• n 3rd Grade Outcomes

Wi With ea each addit itio ional l ris isk a a child is…

30% less likely to meet reading proficiency in 3rd grade. 30-40% less likely to be engaged in the classroom in 3rd grade. 50-100% more likely to be truant in 3rd grade.

Rouse & Fantuzzo (2009); Rouse, Fantuzzo & LeBoeuf (2011); Fantuzzo, LeBoeuf & Rouse (2014)

Cumulative Influence of Risks

Intervention model building Descriptive epidemiological study of a social problem Intervention model testing

Map the distribution of early childhood risk factors across the city Compare that distribution to the distribution

• f publicly-funded high quality early

childhood programs

PHASE 2: ANALYSIS

20

IMPACT

21

City was distribute new pre-K seats in high-risk, low-supply neighborhoods

Intervention model building Descriptive epidemiological study of a social problem Intervention model testing

• Negotiate separate legal agreements to allow a data

pull to flag individual children who qualify as high-risk and live in neighborhoods with new pre-K slots

• Encourage case workers or others who may be in

routine touch with family to provide referral Important caveat: Workers are not told why a child is flagged as eligible

PHASE 2: OPERATIONAL DEPLOYMENT

Evaluate the impact of Philadelphia’s pre-K expansion across domains, with a special focus on children with cumulative risk factors (ongoing)

PHASE 3: ANALYSIS

Intervention model building Descriptive epidemiological study of a social problem Intervention model testing

Common Challenges & Opportunities in IDS

26

AISP Expert Panel Reports:

Published March, 2017 Challenges to Address:

• Scalability, Security, Standardization,

Automation, Personnel, Community Engagement, Comparability

Emerging Best Practices for:

• Legal Issues
• Governance
• Data Standards
• Technology & Security

AISP Network Survey 2018

28

Wh What do you u consider r to be stre rengths

• f
• f you
• ur IDS

DS? ?

Top responses:

• Im

Impac pacting po policy an and d pr program am de decisions

• Data partner engagement
• Linking to county & state data

Wh What lingeri ring challenges do you u face as as an an IDS DS? ?

Top responses:

• Su

Sustai ainabl able fundi ding mode del

• Metadata management
• Enabling access for external researchers
• Linking to federal data

Wh What policy are reas do you u anticipate wi will be of highest intere rest to yo your IDS DS and local policy-ma makers o

• ver t

the n next t three y years?

Top responses:

• Early childhood ed

educati tion

• K-12 education
• Juvenile justice
• Housing & homelessness
• Social determinants of health

On the Horizon: AISP Field-Building 2018-2020

• Future cohorts of sites building IDS capacity
• New cohorts working collaboratively on use cases
• Researcher training IRB module
• Civic Data Infrastructure Working Group
• Communications toolkit

29

Questions?

30

adeliaj@upenn.edu