November 15, 2010, 5:07 PM — The U.S. Office of Personnel Management (OPM) has decided to push back the planned launch of a controversial health claims database by one month.
The new database, which will eventually contain detailed health information on millions of Americans, was originally set to launch this morning. But in a notice issued today in the Federal Register, the agency said it was delaying the deadline to Dec. 15 so it can accommodate more comments from the public.
The OPM also said it may revise its original systems of records notice (SORN) about the database to better explain its authority to maintain such a system and to clarify its intent to "significantly limit" how health claims data will be shared. The OPM may also provide a more detailed explanation of planned security and privacy controls for the new database, the notice said.
It offered no details on when a revised SORN will be published.
The decision to push back the launch comes after privacy groups expressed considerable alarm over the OPM's planned Health Claims Data Warehouse. Much of the concern stemmed from what the groups said is a serious lack of details about why the new database is needed, with whom information will be shared, and how data will be protected.
According to the OPM, the database is designed to help the agency more cost-effectively manage the Federal Employee Health Benefit Program (FEHBP), the National Pre-Existing Condition Insurance Program and the Multi-State Option Plan.
The agency has said it will establish direct data feeds with each of these health plans and pull in data such as names of participating members, their addresses, Social Security numbers and dates of birth, plus the names of spouses and information about dependents and their healthcare coverage, procedures and diagnoses. In its original records notice, the OPM said it would share this data as required, for use in law enforcement, judicial or administrative proceedings and with third-party researchers and analysts. The OPM has noted that the data it collects would be de-identified in many instances, before any analysis takes place.