and predictive modeling

Transcription

1 Everyday, This is about we bringing care for millions those clients of clients into focus using advanced with different analytical needs capabilities and predictive different modeling conditions techniques Advanced to reveal different patterns analytical backgrounds of behavior capacity and predictive modeling different and outcomes experiences over long periods of time David Mancuso, PhD Washington State Department of Social and Health Services Research and Data Analysis Division January 15, 2014 This is about Big Data.... 1

2 Washington is one of a few states in the nation with integrated social service client databases 2

3 That means we can identify costs, risks and outcomes at the state or community level.... 3

4 ... and at an individual or family level 4

5 The data inform internal and external decisions Focused Policy Research Management Decisions Public and Legislative Access Direct Care Management 5

6 The data is now linked to other sources of information, like crime, incarceration, school, and employment data Dental Services Medical Eligibility Medicaid, State Only Hospital Inpatient/ Outpatient Arrests Charges Convictions Incarcerations Community Supervision Managed Care Services Physician Services Prescription Drugs Hours Wages Births Deaths Homelessness Housing Assistance Washington State Patrol Administrative Office of the Courts Department of Corrections Health Care Authority Employment Security Department Department of Health Department of Commerce External Internal WASHINGTON STATE Department of Social and Health Services INTEGRATED CLIENT DATABASES DSHS Aging and Long- Term Support DSHS Children s Services DSHS Developmental Disabilities DSHS Behavioral Health and Service Integration Mental Health and Substance Abuse Services DSHS Economic Services DSHS Juvenile Rehabilitation DSHS Vocational Rehabilitation Nursing Facilities In-home Services Community Residential Functional Assessments Child Protective Services Child Welfare Services Adoption Adoption Support Child Care Out of Home Placement Voluntary Services Case Management Community Residential Services Personal Care Support Residential Habilitation Centers and Nursing Facilities Assessments Detoxification Opiate Substitution Treatment Outpatient Treatment Residential Treatment Child Study Treatment Center Children s Longterm Inpatient Program Community Inpatient Evaluation/ Treatment Community Services Food Stamps TANF and State Family Assistance General Assistance Child Support Services Working Connections Child Care Institutions Dispositional Alternative Community Placement Parole Medical and Psychological Services Training, Education, Supplies Case Management Vocational Assessments Job Skills Family Reconciliation Services State Hospitals State Institutions 6

7 Very detailed indicators allow us to identify risk and outcomes Health Emergencies Diagnoses Services Juvenile Rehab Long Term Care Work Earnings Age Voc Rehab Gender Demographics Foster Care TANF WorkFirst Child Welfare Medical Employment Hours Citizenship Race/Ethnicity Progress Issues Housing Stable Geography County Grades Behavioral Health Status School Outcomes Stability Attendance Special Needs Homeless Locale Alcohol/Drug Treatment School District Disabilities Crime Arrests Misdemeanors Felonies Family Relationships Chronic Conditions Medications Births Utilization Treatments Hospitalization Incarcerations Convictions Siblings Unemployment Legislative District Community Risk Factors Deaths 7

8 Focused policy research informs state and national policy Hundreds of formal, in-depth studies Focused Policy Research 8

9 Risk Factors Associated with Development of Behavioral Health Needs in Adolescence ODDS RATIOS, Youth age 12 to 17 enrolled in Medicaid in SFY 2008 Substance Abuse Risk Factors Child Welfare Involvement: Child or Parent Substance Abuse: Parent Arrest/Conviction: Parent Mental Health Need: Parent Domestic Violence: Parent Death of a Parent Homelessness: Child or Parent INCREASED RISK 4.2 Mental Health Risk Factors Child Welfare Involvement: Child or Parent Mental Health Need: Parent Substance Abuse: Parent Death of a Parent Domestic Violence: Parent Arrest/Conviction: Parent Homelessness: Child or Parent Focused Policy Research INCREASED RISK SOURCE: DSHS Research and Data Analysis Division, Paper in Progress. 9

10 TANF Adults and Behavioral Health Needs Behavioral Health Risks Among TANF Adults Identified using pooled SFY 2005 to SFY 2009 indicators Neither MH nor AOD Treatment Need n = 26,319 36% MH Treatment Need Only n = 25,049 34% AOD Treatment Need Only n = 6,916 9% Both MH and AOD Treatment Need n = 15,637 21% Alcohol and or Drug Treatment Need 30% Mental Health Treatment Need 55% Focused Policy Research SOURCE: DSHS Research and Data Analysis Division, Adults on TANF in Washington State, Risks and Outcomes for Leavers, Cyclers and Stayers, Mancuso, Ford Shah, He, Estee, Felver, Beall, Yette, Fiedler, Sandberg, August 2010, 10

11 Outcomes of youth at risk Criminal justice and employment outcomes among youth Comparing youth who received Functional Family Parole to those who did not 31.9% No Functional Family Parole 43.9% Functional Family Parole 21.5% Functional Family Parole 29.5% No Functional Family Parole 35 of 163 arrested 52 of 163 arrested Percent Arrested During 9 months following JRA release 61 of 139 working 41 of 139 working Percent Employed During year following JRA release Focused Policy Research SOURCE: DSHS Research and Data Analysis Division, Effects of Functional Family Parole on Re-Arrest and Employment for Youth in Washington State, Lucenko, He, Mancuso, Felver, October 2011, 11

12 Dynamic, current data for program managers Use rates, costs, trends, distance from targets Operational Measures Q. Are social workers evaluating children s health and safety each month? Rate of Children/Youth Visited by Social Workers During the Calendar Month Management Decisions Build cleaner version of this chart, see if data goes back further 12

13 Allows us to reach clients during natural disasters Clients who are vulnerable during fires, floods, earthquakes, tsunamis, other natural disasters Elderly Medically at risk Living with disabilities Persons with limited mobility Depth of Potential Flooding Management Decisions 13

14 Provides granular detail for clients at extreme risk Indicators Homelessness Arrests, Convictions Medical Encounters Diagnoses Injuries Prescription Meds Management Decisions Number of instances per month Since 1998 NOTE: This level of detail is protected by law and is used for internal uses only 14

15 Off-the-shelf data for policy makers Public and Legislative Access Legislative District Detail Prepared tables Client counts Costs by district Use rates Specific services used 15

16 CLALLAM JEFFERSON GRAYS HARBOR PACIFIC MASON THURSTON KITSAP COWITZ CLARK LEWIS PIERCE KING SKAMANIA WHATCO M SKAGIT SNOHOMISH CHELAN YAKIMA KITTITAS KLICKITAT OKANOGAN FERRY STEVENS DOUGLAS GRANT BENTON FRANKLIN ADAMS LINCOLN WALLA WALLA COLUMBIA SPOKANE WHITMAN GARFIELD PEND OREIL LE ASOTIN County level service detail and costs Public and Legislative Access County Tables Service use rates Direct service costs By age Race/ethnicity 10 years of data 16

17 Trends in service use and risk factors by county Public and Legislative Access Prepared Reports Risk factors for infants, parents, youth, adults, elderly Treatment for injuries Alcohol/drug treatment need Mental health treatment need Employment Arrests, convictions Service history Adverse childhood experiences Homelessness Disabilities Demographics 17

18 Client Registry supports coordinated service delivery Desktop tool for Case Managers for cross-program coordination Client Registry Direct Care Management Connects directly to most DSHS client information systems Reduces workload redundancies, decreases costs, improves client outcomes Confidential, secure, updated nightly Shows when and where clients received services and who to contact to coordinate services 18

19 Cutting edge predictive risk application Hands-on tool for front line workers Predictive Risk Intelligence SysteM (P R I S M) Direct Care Management Decision support tool to support care management interventions for high-risk Medicaid patients Intuitive and accessible health and demographic information from administrative data sources Continues to evolve to meet changing program needs Over 700 users 19

20 Challenges we faced... Practical. Convincing management it could be done prior failures expensive and embarrassing Data sharing. Each organization had its own data sharing and privacy controls there was resistance Administrative. Government salaries for IT staff were not competitive and internal technology practices restricted innovation 3-year 1990 Development Cycle 2000 How did we get here? Leveraging investments over time Direct service dollars added Client data reporting begins, ongoing CLIENTS Needs Assessment Database Prototype built as proof-of-concept Logic developed to define services and match clients across multiple systems 600 SAS processes Mental Health Transformation Alcohol/Drug Treatment Expansion SSI Aged, Blind, Disabled Evaluation Client Services Database Development of a production system was funded Process logic from Needs Assessment prototype refined 5,000 services grouped into 80 standard report groups Data from 20 payment and case management systems Test-driven development dynamic and data exploration ongoing Homelessness and Housing Predictive Risk Modeling P R I S M DOLLARS 2010 School Outcomes Dual Eligibles GEOGRAPHY 20

21 Sustainability Lessons learned from our 20 year track record Funding is and will always be a challenge Our best marketing tool is our analytical capability Many vendors can organize data into databases with Business Intelligence tools that push the hard analytical work back on the end user Our competitive advantage is that we have the analytical capability to extract meaning from complex data With recent budget shortfalls, we ve become more entrepreneurial in developing internal funding while also developing more federal grant and private foundation funding Health and social service programs and technologies are always evolving We are always learning, always improving Maintaining an understanding of changing social and health service programs and source IT systems is main threat to sustainability Need to stay on top of technologies to manage Big Data Attracting and retaining talented staff is critical Our key investment is in our staff Where we are sometimes unable to offer competitive salaries, we can offer flexibility and the opportunity to make a real difference in peoples lives Our team translates data into real-life, practical analyses that directly inform policy 21

22 Questions? 22