USING STATA There are two ways to read the 1993-2001 NAMCS public-use data files into Stata: Option 1 - Use the self-extracting file in the Stata folder to open a complete Stata dataset of the NAMCS, NHAMCS-ED, or NHAMCS-OPD public-use files. In addition to the README93-01.TXT file, the Stata folder contains a Stata dataset (*.dta) which was compressed into a self-extracting file. Option 2 - Use the DO file (*.do) and dictionary file (*.dct) provided in the “Stata Documentation” section of the Ambulatory Health Care Data website, (under Public Use File Documentation) along with the flat data file (*.EXE) in the “Downloadable data files” section, to create a Stata dataset. The file that is created will be identical to the file described in Option 1 above. Two important notes when using Option 2 to create a Stata dataset: 1) The flat data file will have no extension once it is uncompressed. The INFILE command in the DO file requires that the data file has an extension, so the uncompressed datafile will need to be renamed with an extension, for example, NAMCS93.DAT. 2) The INFILE line in the DO file needs to be edited to correctly reflect both the name of the re-named dataset, and the exact locations of both the data and the dictionary file. A note on labels: It should be pointed out that Stata will not allow value labels for character (string) format variables. In the case of NAMCS and NHAMCS data, there are a number of character or string variables, such as DIAG1 (physician's first-listed diagnosis), etc. For most character variables we have been able to redefine them as numeric without changing their basic structure and the labels should work. However, for the diagnosis variables (DIAG1, DIAG2, DIAG3) which actually contain alpha characters (eg "V") in the string, this was not able to be done. We do also have a numeric recode version of the DIAG variables (DIAG1R, DIAG2R, DIAG3R), but these will not match with the character format of the ICD-9-CM codes in many cases, and we do not have a set of labels that match with the numeric recode variables for diagnosis codes. If data users experience difficulties or have suggestions for future releases of these products, please contact the Ambulatory Care Statistics Branch at 301-458-4600.