Back

A Systematic Background Check of TRICARE Provider Names

Bychkov, D.

2023-08-22 health systems and quality improvement
10.1101/2023.08.14.23294073 medRxiv
Show abstract

During Covid-19, the Defense Health Agencys TRICARE insurance plan expanded its coverage to include 30.1% additional civilian healthcare providers. The DHAs Annual Report, however, states that TRICAREs provider directories are only 80% accurate. Although the DHAs 9.6 million beneficiaries need expanded access to care, they also require protection from misleading information, medical fraud, patient abuse, and identity theft. Since 2013, the Department of Health and Human Services Office of the Inspector General has excluded 17,706 physicians from federal health programs due to misconduct. Patients who receive care from excluded providers experience worse medical outcomes. To determine if any excluded provider names were found on TRICAREs website, we performed background checks on TRICARE Wests healthcare provider directory between January 1 and March 2023. Out of 39,463 provider names sampled from 22 states, there were 2,398 matches (6.08%) with individuals and businesses found in the OIG List of Excluded Individuals and Entities (OIG-LEIE), the GSA-SAM, the HHS HIPAA Breach Report, the International Trade Administrations Consolidated Screening List, the OIG-HHS Fugitive List, the FBIs January 6th Capitol Violence List of Charged Defendants, State Medicaid Exclusion Lists, and FDA Debarment Lists. Our study includes demographic analysis of the matching names and recommendations for an Insider Threat Management model. We recommend that DHA officials publish the National Provider Identification (NPI) numbers of all TRICARE providers. NPI numbers facilitate more accurate background checks of healthcare providers.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.