Synthetic healthcare data
What an inpatient claim gets paid, and how far that differs between commercial insurance, Medicaid and Medicare, across 1.2 million people.
Payment distributions vary by insurance type
The amount paid per claim for the inpatient care of 1.2 million people in 2016, representing $8.6 billion in health care costs, stated in 2022 dollars. The underlying dataset is synthetic — generated by the Agency for Healthcare Research and Quality to match the statistical shape of real discharge records without describing any real patient.
Ideas welcome. The next questions are large language models and algorithmic fairness metrics in health care settings; email [email protected] to work on them.
Data and code
- Data: AHRQ Synthetic Healthcare Database for Research codebook, Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services.
- Transform: insurance_plan_payment_histogram_inflation_adjusted.sql in onefact/synthetic-healthcare-data.