We have completed work on the Discharge Abstract Database files that were released on May 20. The basic program used to enhance the clinical file required a small change to handle changes introduced with SPSS 23.
As with other cycles, the clinical file has been enhanced with flag variables to let users identify records of interest without looking at each of the 25 ICD-10 or 20 CCI codes separately, and the ICD-10 and CCI codes have been reformatted to match the documentation (e.g., punctuation inserted). Additionally, variables in both the clinical and geographic files that were stored as strings have been converted to numeric codes so that Stata users can have value labels attached, and to make them more useable (e.g., allowing age to be a ordinal variable rather than a nominal string).
I’ve noted a concern / question though – the number of records in the clinical file is WAY down this year. Was this once again a 2-year sample (as the read-me file indicates), or a 1-year sample this time?
Clinical file
2009-11 393,625 cases
2011-13 404,650 cases
2013-14 204,529 cases ???
It also appears to not be a 2-year sample because of the change in naming convention for this release.
