Direct Verification Presentation to School Nutrition Association November

Direct Verification Presentation to School Nutrition Association November 29, 2007 Abt Associates Footer Information goes here

What is Direct Verification? • Using information from means-tested programs to verify school meal applications without contacting households Authorized means-tested programs: • • • Food Stamp Program (FS) Temporary Assistance for Needy Families (TANF) Food Distribution Program on Indian Reservations Medicaid (Title XIX) State Children’s Health Insurance Program (SCHIP) (Title XXI)

Goals For the Evaluation • Evaluate DV-M Implementation Ø Is it feasible? Ø What types of systems work? Ø What are the challenges and lessons? • Evaluate DV-M Effectiveness Ø What percentage of school districts use DV-M? Ø What percentage of applications are directly verified? Ø What do districts think of this tool? • Participating States: Indiana, Oregon, South Carolina, Tennessee, Washington [Georgia in 2007]

Direct Certification and Direct Verification Direct Certification Direct Verification When? Before/during 1 st month of school (can be repeated) After selecting verification sample (October 1) Data source? Food Stamp, TANF, FDPIR Food Stamp, TANF, Medicaid/SCHIP Purpose? Certify without application Verify application without (& exempt from verification) contacting household Scale? LARGE – attempt to match all children SMALL – attempt to match children in verification sample

Why Use Direct Verification with Medicaid (DV-M)?

Potential Benefits of Direct Verification 1. Reduce workload and hassle for school district staff 2. Reduce burden and intrusion on families selected for verification 3. Reduce number of non-respondents and rate of benefit termination for non-response 4. Improve program integrity by checking eligibility with programs that document income information

Advantages of Medicaid/SCHIP Data 1. Directly certified FS/TANF children are exempt from verification. • Thus, few applications will be directly verified with FS/TANF data. 2. Medicaid/SCHIP income limit exceeds Food Stamp income limits • Limit at or above NSLP-RP limit (185% FPL) in 46 States.

Number of States by Medicaid/SCHIP Eligibility Limits Medicaid/SCHIP income limit (%FPL)

How Does Direct Verification Work?

Guidelines for Direct Verification with Medicaid • Timing of data – Use latest available Medicaid/SCHIP information, no more than 180 days prior to NSLP application date* • Matching program data to NSLP applications – Use names and other identifiers of children listed on the NSLP application. – If Medicaid income limit <133% FPL, a match verifies NSLP-free eligibility. Else, Medicaid info about family income and family size (or income as %FPL) verifies eligibility for NSLP-free or RP. • Using match results – Match one child on the NSLP application and all children on the application are verified. * Or use data from the month prior to application through the month of verification.

Alternative Ways for States to Implement Direct Verification 1. Send Medicaid/SCHIP data to districts 2. Collect application data from districts and match at State level 3. Develop a “look-up system” on the State CN/Education website 4. Provide direct access to existing Medicaid/SCHIP program data system Level of effort: #1 – Low effort for State #2 – Most work for State (year after year) #3 – Upfront investment, low maintenance cost #4 – Low effort for State if available (depends on existing infrastructure)

Tennessee – Send Medicaid Data to Districts • State divided file of Medicaid children by county and posted Excel® files on secure website • Districts downloaded data file from State website and searched manually • Identifiers: SSN, name, DOB, guardian name, address • Districts verified NSLP-free eligibility by matching children to Medicaid file

Oregon – Send Medicaid Data to Districts • State provided statewide file of Medicaid children via secure e-mail • Districts downloaded data file, opened with their own software—usually Excel®—and searched manually • Identifiers: name, DOB, FS/TANF #, guardian name, address • Districts verified free/reduced-price eligibility with family income and household size from Medicaid file

Washington – Send State-Level Match Results to Districts • State matched Medicaid children with statewide student database by name and DOB, created F/RP indicator based on Medicaid information • State created Excel® files for selected districts and sent via email (web-based distribution planned for 2007) • Districts searched manually and checked F/RP indicator • Identifiers: name, DOB, gender, State student ID #, district student ID #, address, school code and name, Medicaid ID number

South Carolina – Collect NSLP application data from Districts and Match at State Level • Districts created files of verification sample using State template • State CN Agency collected disks from districts and sent file to Legislative Office of Research and Statistics (ORS) • ORS matched verification sample data with Medicaid data by SSN, name, date of birth, etc. • ORS sent verification sample files with match results to State CN Agency, which sent them to school districts

Indiana - Look-up System on State Website • State provided form-based interface on secure website • Query of FS, TANF and Medicaid Data (children eligible in July, August, September, or October) • Districts login to website and search for individuals using: – Student name & DOB (phonetic match) – FS/TANF case # – Parent/guardian name/SSN • Search returned identifiers, F/RP status, reference number

Georgia - District Access to Existing Medicaid Data System • Current Food Stamp/TANF/Medicaid eligibility data available via online inquiry system (“GO”) • School districts obtained login and installed software to access GO system • Query by child’s name, DOB (or age), and sex, or by case number, or parent’s SSN • Case record indicated FS/TANF/Medicaid eligibility and listed case members • Budget screen provided household income

Implementation Process 1. Meet with State Medicaid Agency Discuss NSLP verification, direct verification, and data needs 2. Determine how DV-M system will work What Medicaid data to provide to districts and how; how to protect confidential data 3. Establish data-sharing agreements Specify data elements, formats, timing of exchange; define authority for exchange; provide assurances for protection of confidential data 4. Implement State-level processes Disseminate instructions and/or provide training to districts; prepare data; “go live” with website or by distributing data to districts; ongoing support

Implementation Challenges in 2006 1. Getting access to Medicaid data Confidentiality issues in Indiana and South Carolina 2. Testing before going live Income data gap found in Washington; incomplete file for State in Indiana 3. Making it easy for districts to use Oregon file hard to use; batch matching helpful for large districts; include only the right amount of information 4. “Go live” by October 1 Districts need data and instructions before they start verification; State needs adequate lead time with room for delays

Results of 2006 -07 Survey of Districts

District Participation Varied Across States

Percent of Applications Directly Verified

Most Districts Found DV-M Easy

Mixed Ratings on Usefulness of DV-M

Most Districts Expected to Use DV-M in 2007

Verification Cost Per Application: All Districts

Verification Cost Per Application in Districts with Applications Directly Verified

Summary • States have demonstrated technically feasible approaches to DV-M • Challenges for implementation are mainly on the “soft side”— negotiating agreements, promoting district participation, setting and keeping schedule • If the State offers DV-M and makes it easy to use, school districts are likely to use it • Substantial percentage of applications may be verified if data are timely and complete • Effectiveness of DV-M is primarily influenced by district participation and Medicaid income limits • When DV-M is effective, it can save time for districts

Slides for Q&A

Time and Cost of Verification Activities Definitions • Direct verification time/cost includes: – Reading instructions and orienting to new process – Accessing system to download data or search – Searching for students listed on NSLP applications selected for verification – Documenting results • Household verification time/cost includes: – Sending initial letters to households – Answering queries from households – Processing household documents, determining eligibility, and following up if documents are incomplete – At least one follow-up contact with nonresponders

Minutes Per Application: Direct Verification Saves Time When It Works