Access to raw data would need legislation

Editor—Tony Delamothe argues that there should be universal access to the raw data of medical research.1 I agree, subject to practicality: some records are very bulky and cannot be kept indefinitely. Contrary to Delamothe's assertion, electronic files rarely are raw data—they are data that have been processed, and in the handling they have been open to error and filtering.

Delamothe does not mention that legislation would be required in the United Kingdom and the United States, because lobbying by the pharmaceutical industry has ensured that the raw data of their treatment trials have to remain secret unless a litigant can obtain a court order. The Freedom of Information Act in the United States does not give access to those copies of raw data from drug companies that are held within the Food and Drug Administration. For the past decade those copies have in any case been limited to the records of patients who died or entered hospital during a treatment trial.

Datasharing would cause problems in nuclear industry

Editor—The article by Tony Delamothe advocating researchers' unrestricted access to anonymised raw data1 might lead one to think the author naive, arrogant, and authoritarian. We do not hold such views, but we nevertheless …