The NBER previously posted version of the Healthcare Cost Report Information System (HCRIS) Form 2552-96 and 2552-10 files covered 1996-2018 and may not be updated. The NBER versions of these files were advantageous for some users because the were in .dta, or .csv formats, and were not in database normalized form. The normalized form may have been chosen by CMS because there are several multiplicative variables in the files, however it was an obstacle for users with no interest in those variables. The normalized files are not the only current release by CMS. There is a CMS webpage linking to a SAS version of the non-normalized 2552-10 files These are available for 2010 on, although the beginning and ending years are incomplete. The non-normalized files have more than 5,000 variables, but do not include all variables.

On this page we offer conversions of the non-normalized annual files to .dta and .csv formats, as well as panel versions. The panels add "fileyear" as an additional variable. The variable names differ from our 1996-2018 extracts in that names are punctuated with "_" rather than "." and the coverage may not be the same.

Data Downloads


The Andrews paper is required introductory reading.

last update 13 November 2020 by data@nber.org