RDC Xinjie Cui - Data Interoperability_II

Xinjie Cui BMed PhD MBA

Chief Analytics Officer

Building a Child Data Center with Linked Population Data

to Inform Policy

CYDL Link Analyze Inform 2

The Need and The Opportunity

Complex social, health, and education policies REQURE• More • Relevant • Timely Evidence for decision

Advancement of information technology ENABLES• More and better data collection and store (digital storage) • More, better, new methodologies for data discovery


Secondary Data and Secondary Analysis

Secondary analysis – utilization of existing data

Secondary data • Administrative data – collected to conduct business• Service delivery data – collected to provide services• Research data – collected to answer research questions

Benefits• Potentially faster • Relatively low cost


Usefulness of Secondary Data

• Monitor/track health and wellbeing of the population (identify trends and issues)

• Discover patterns and relationships of risk/resilient factors and outcomes

• Support development of strategies and policies • Evaluate effectiveness of policies, programs,

and services• Inform performance measurement and result

based budgeting


Child Data Centre of Alberta

A data and research facility –Wide variety of data holdings with

potential for linkage–Value added services–Broad access by qualified policy

analysts and researchers


Vision

Improved outcomes for children and family by informing social and health policy, practice, and service planning

Through New knowledge generation Timely and relevant evidence for policy and practiceCapacity building


ACCFCR Board

ACCFCR

Child Population

Data Repository

Child Research Data

Repository


Community Service

Agency DataRepository

Children First Act


Government Administrative DataPros:• Populations based (all Albertans)• Longitudinal• Linked data may address system-wide and policy relevant

questions well and provides opportunity to study broad determinants of health

Cons:• Lack details on some measures such as behavioral risk

factors, physical measurements etc.• Inconsistent measures; recording and coding errors• Subject to misinterpretation if not familiar with data

collection and manipulation process


Service Delivery DataFrom community agencies and service organizations

Pros• Detailed in depth data• Intervention and services focused• May contain risk factors and qualitative measures

Cons• Inconsistency across agencies/clinics

– standards, capacity, systems, services…

• Large number of organization of different kinds


Research Data• Public funded research data exist in isolation• Often times, when research project is done,

data gets destroyed or shelved – seldom gets reused or further analyzed

Modern Research

• Collaborative, multi-disciplinary, and answer complex questions

• Data are large in volume and digitized• Great value for repurpose and reuse


Research Data Sharing• Trend around the world: – Funding Agencies: NIH USA, MRC UK,

Tri-council Canada– Open access content: PLOS

• Lack of centralized place for storage, management and further utilization

• Especially lack of support around secondary use


Child Data Centre of Alberta• Securely stored, cleaned, catalogued,

enhanced, managed, and potentially linkable. (where and when it is possible and appropriate).

• Access and use of data facilitated, managed and supported – value add

• Information privacy protected


ACCFCR Board

ACCFCR

Child Population

Data Repository

Child Research Data

Repository


Community Service

Agency DataRepository

Children First Act


Child and Youth Data Lab

As the Administrative Data component of CDCA

• Links and analyzes government administrative data – cross-sector

• For policy relevant research • With a focus on children, youth and families

in Alberta


Partnering Ministries

• Human Services (Lead ministry)• Health (HUB for data matching)• Education • Innovation and Advanced Education• Justice and Solicitor General• Aboriginal Relations


CYDL Research Process• Partnership – working closely together • Process

– Identification of research priorities – Project development/Study design – Identification of data elements – Feasibility assessment (data and privacy)– Research Ethics approval– Data sharing agreements and Privacy Impact Assessment – Data linkage and transfer – Data validation, analysis, and interpretation – Report generation and dissemination/knowledge mobilization

• Involvement of ministries throughout the process


Characteristics of CYDL• Authentic Partnership – Ministries involvement throughout the full

research cycle

• Anonymous Linkage – Data hashed at source

• Policy Relevant – Research questions are policy driven

• Project Based Model– No population repository yet


Experience of Albertan Youth over Time • Longitudinal design• Six years of data 2005/06 to 2010/11

• Broad scope: • 5 ministries• 25 databases covering more program and service

areas • Over 250 variables

• Key indicators: • Demographic• Socio-economic status• Mental health status• Educational achievement • Health service use


0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

All programs

Health

Child intervention

Child support services

FSCD

Child health benefits

Child care subsidy

Education

Corrections

Offence charges

Advanced Education

Income support learners

Income support

AISH

Adult health benefits

PDD

Advancing Futures bursary

Age (years)

CYDL Longitudinal Project Sample Indicators

Young adulthood

Childhood

Youth

Early childhood

Court outcomes

Type of charge, charge history

Program and institution details

Household type, enrollment details

Client type, household type

Disability type, household indicators

Type of benefits

Type and amount of support

Enrollment and support details

Perinatal indicators, diagnoses, mental health status

Type of abuse/neglect, type of services provided, history

Type of order, reason for file closure

Type of disability, duration of services, reason for file closure

Type of benefit received, age at first use

Type of program, reason for care, duration

Achievement, school indicators, special programs, high school completion

Gender, age, socio-economic status, region of residence/service, Aboriginal status (if available)


Types of AnalysisTime trendsPolicy impact over timeService use patter over timeSpecific topics such as

• Youth transition • Resiliency• Early childhood

development• Poverty

• Persistent offenders• FASD• Aboriginal youth• Special education• Self-harm


Research Data Repository

• Stakeholders consultation• Best practice - technical, management,

ethical, metadata, and data governance • Principal Investigators of two cohort studies

on board and other datasets• Design of processes and structure


Research Data Repository

• Set up governance structure • Build technical infrastructure – enclave • Transfer initial datasets • Develop metadata • Pilot early adopter projects• Support access and use


Research Data Repository• Develop communication/marketing plan to

promote CDCA • Identify potential research datasets for

repository• Provide training opportunities for students

and trainees• Identify opportunities to develop Service

Delivery Data component • Explore feasibility of integration with CYDL

component


Issues for Consideration

Research Ethics • Consent, Re-consent, Agreements

Information Privacy and Confidentiality • Right to privacy and confidentiality as public good

Data Security • Technology (increased risk but also new tools to

protect)• Data linkage (increased risk for re-identifiability)

Original Researchers’ Involvement• Researchers’ input, flexibility


Success Factors• Independent

• Outside of government• Outside of universities

• Authentic Partnership• Engagement of research community• Political and Executive support• Enabling data sharing legislation• Sponsors and Champions at multiple levels• Strong Governance and Accountability structure

and process• Sustainable funding and Designated resources• Strong project management and coordination• Dedicated staff and technical/research expertise• Persistence and patience


Questions?

RDC Xinjie Cui - Data Interoperability_II

Science

Transcript of RDC Xinjie Cui - Data Interoperability_II