The Oregon Eligibility System Appropriately Determines Eligibility, But Input Errors Continue to Occur
Learn how the AI-generated research projects were createdOverall Conclusion
The audit found that the ONE system largely determines eligibility accurately for Medical and SNAP programs, and automated eligibility determinations are generally effective. However, input errors, manual data entry, and inconsistent override monitoring create risk for incorrect determinations and benefit calculations, particularly as the system and policies evolved during the COVID-19 Public Health Emergency. While change management is formal and well-documented, testing lacks a formal, unified plan, relying on informal controls. The report recommends automation, stronger override governance, and a formal UAT plan to mitigate these issues and improve input accuracy and overall reliability.
Source Document
Audit Scope
The audit evaluated the Oregon Eligibility System (ONE), a centralized system used by the Oregon Department of Human Services (ODHS) and the Oregon Health Authority (OHA) to determine eligibility and benefits for Medicaid, CHIP, and SNAP. The scope included automated system processes for processing applications and determining eligibility for Medical and SNAP programs. The test period ranged from eligibility determinations authorized on or after July 1, 2021, through May 31, 2023, though some tests evaluated ongoing eligibility beyond that period. The audit also assessed data input accuracy in ONE and the appropriateness of overrides by eligibility workers. It evaluated the sufficiency of ODHS and OHA procedures to analyze and prioritize changes to ONE; it did not assess Deloitte LLC's change management procedures. It also evaluated the sufficiency of user acceptance testing for system changes.
Key Findings Summary
ONE system largely determines eligibility appropriately for Medical and SNAP programs, but some automated functions did not work properly or reliably for a small number of determinations, including interface errors and social security number verification against federal records.
Most errors in eligibility determinations and benefit calculations were the result of manual input errors, not automated processes, with complexity and COVID-19 policy changes increasing manual inputs and input errors.
Manual overrides by eligibility workers, while rare, have reversed appropriate automatic determinations; there are policies but they are not consistently followed and are only recently being reviewed by management; overrides documentation is inconsistent and monitoring is limited (e.g., 0.50% of Medical and 0.08% of SNAP overrides); SNAP overrides review fou…
View the Findings tab to see all 6 findings
AI-Assisted
AI Scope Summary
Assess whether Oregon's ODHS and OHA have designed and implemented sufficient automated controls in the ONE eligibility system to accurately determine and maintain Medicaid and SNAP eligibility, prevent and detect input errors, monitor and document overrides, and control code/configuration changes, during the period July 1, 2021 to May 31, 2023, with emphasis on data quality, override governance, and test planning.
AI-Generated Insight
This audit underscores a critical tension in large, integrated benefits systems: precision in automated logic is high, but human data entry remains a major vulnerability. The COVID-19 emergency amplified this risk by expanding policies and increasing manual inputs. The findings suggest that significant benefits could be gained from targeted automation (e.g., income data entry prompts, mandatory reconciliations on information changes) and strengthened governance over overrides and testing. Implementing formal user acceptance testing and more rigorous change controls would reduce the risk of incorrect eligibility decisions, improve program integrity, and safeguard public funds, while maintaining the ONE system's efficiency gains.
Audit Objectives
Determine whether ODHS and OHA have designed and implemented sufficient controls to: Completely and accurately determine and maintain eligibility and benefits for the Medical and SNAP programs in the ONE system in accordance with rules and laws.
Prevent, detect, and correct manual input errors associated with benefit applications and maintenance activities.
Ensure overrides to eligibility determinations and benefit amounts are done for approved reasons and are appropriately documented and monitored in the ONE system.
Ensure that changes to computer code and configurations for the ONE system are appropriately controlled to ensure the integrity of the system.
Audit Findings (6)
ONE system largely determines eligibility appropriately for Medical and SNAP programs, but some automated functions did not work properly or reliably for a small number of determinations, including interface errors and social security number verification against federal records.
Most errors in eligibility determinations and benefit calculations were the result of manual input errors, not automated processes, with complexity and COVID-19 policy changes increasing manual inputs and input errors.
Manual overrides by eligibility workers, while rare, have reversed appropriate automatic determinations; there are policies but they are not consistently followed and are only recently being reviewed by management; overrides documentation is inconsistent and monitoring is limited (e.g., 0.50% of Medical and 0.08% of SNAP overrides); SNAP overrides review found 40% lacked case notes and 90% were not approved by policy.
Change management for ONE is formal and well documented, but state tests of changes rely on informal controls and user acceptance testing lacks a formal, comprehensive plan.
Input controls exist to ensure required fields are completed, but accuracy of input data not controlled by automated checks is challenging, relying on staff; training materials exist but are primarily instructional; multiple internal units perform data accuracy reviews, yet prevent/detect errors remains insufficient, especially under the PHE.
During testing, residency and income data errors were found; 156 residency-related instances where residency status should have ended eligibility after change; 29 cases where workers did not rerun eligibility after residency change; various income data issues, and instances where high reported income was entered manually; test results showed inaccuracies in SNAP benefit calculations and Medicaid eligibility due to manual input errors (e.g., misstatement of household size, incorrect income entries, misrecorded expenses); 156 residency errors could lead to continued eligibility when not appropriate once PHE ended.
Recommendations (3)
Evaluate opportunities to improve input accuracy through automation and propose changes to the ONE system to enforce them, including requiring entry into income fields necessary for eligibility and requiring rerun/authorization when information changes.
Refine override processes to ensure overrides are restricted to authorized personnel, consistently documented, and monitored: limit override rights to leads/managers or require approval, implement automated note requirement, formal override review schedule, and procedures to correct detected override errors.
Develop a formal user acceptance testing plan for changes, consolidating existing practices into an overarching plan that defines test types, priorities, risk assessment, coverage, resources, documentation of decisions, and a target of implementing a comprehensive UAT plan by 9/30/2025.