NAHDO has done an extensive review of state data release practices and policies. Most states used a multi-layered approach to the release of public use files:
1. Changing or deidentifying the data: suppression, aggregation, and restricting or denying the release of confidential/sensitive data elements.
2. Regulatory and management controls: Data Use Agreement/request form, penalties for non-compliance.
3. Limiting access through data products: web query systems (permitting anonymous queries for aggregate statistics without accessing the detailed data; public use files which are de-identified; research files with more sensitive data usually for bona fide research purposes and requiring IRB approval; and custom data releases---where the agency works with the user to release just the data fields that can be justified for the purpose---making tradeoffs between the potential universe of data elements (e.g. dates of discharge without zip or other patient variables, but not all).