Trend Analysis Using NAMCS and NHAMCS Drug Data
In order to analyze NAMCS and NHAMCS drug data across several years, it is important to understand the changes that were made in the processing of these data starting with the 2006 surveys.
Prior to 2006, drugs were coded twice. The first coding was the drug “as entered” on the form, using an NCHS-developed classification that reflected exactly what was written on the Patient Record form. For example, if “acetaminophen” was entered, it received a code of “00260” but if “Tylenol” was entered, it received a code of “32905”, even though they are generically equivalent. These codes appeared in the public use data under the MED1-MED8 fields in 2005 and previous years.
Drugs were coded a second time, using a separate NCHS classification, to reflect their generic components. This coding system was entirely independent of the entry classification described above. In this second coding, using the previous example, a code reflecting “acetaminophen” would be used for any drug whose generic component was acetaminophen regardless of whether it was entered as Tylenol or some other name. Such entries appeared on the public use files for 2005 and past years in the GEN1-GEN8 fields.
Drugs with multiple ingredients received a generic code that merely reflected that it was a “combination product”, and up to 5 ingredients were listed separately for those drugs (using the NCHS generic code classification). Drugs were also assigned up to three therapeutic classes based on the FDA’s National Drug Code Directory. (Prior to 2002, only a single therapeutic class was assigned.)
Starting with the 2006 surveys, a new classification system was adopted to code the characteristics of drugs reported on the Patient Record forms. The NAMCS and NHAMCS no longer use the NCHS-developed 5-digit code to classify drugs according to their generic components (the second coding described above), nor do they use the National Drug Code Directory to code therapeutic class. However, the first coding of the drug “as written” using the NCHS-developed classification is still used.
Drugs are now coded in terms of their generic components and therapeutic classes using Lexicon Plus®, a proprietary database of Cerner Multum, Inc., also used by the National Health and Nutrition Examination Survey, NCHS. The Lexicon Plus is a comprehensive database of all prescription and some nonprescription drug products available in the U.S. drug market. For additional information, see the Multum Lexicon Drug Databaseexternal icon.
In Multum, multiple-ingredient drugs are assigned a single generic drug code encompassing all of a drug’s ingredients, rather than being assigned generic drug codes for each ingredient, as in past years of NAMCS and NHAMCS drug data.
For example, a written response on the Patient Record form of “hydrocodone w/acetaminophen” in the 2005 NAMCS would have received a drug entry code of “89039”, a generic name code of “51380” from the generic classification, simply meaning “combination product”, and ingredient codes of “50005” for acetaminophen and “70217” for hydrocodone bitartrate, again both from the NCHS generic classification. (Note that if it had been written on the Patient Record form as “Vicodin”, a brand name, the drug entry code would have been “34110” but the other codes would have been the same.)
In contrast, an entry of “hydrocodone w/acetaminophen” in the 2006 NAMCS would have received the same drug entry code “89039” as in 2005 but a DRUGID code of “d03428”. The DRUGID code (which replaces our previous GEN code) reflects the Multum code for the combination of acetaminophen w/hydrocodone. Therefore, separate ingredient codes for each drug are not included on the public use files. However, a separate file containing ingredients for each medication is available and is described below.
Multum’s therapeutic class structure is different than that of the National Drug Code Directory. For more information, go to the 2006 NAMCS and NHAMCS public use file documentation. Multum assigns multiple therapeutic classes for each drug, and up to four classes are included on the public use data files. However, for multiple-ingredient drugs, Multum assigns a “combination” therapeutic class which does not reflect the specific therapeutic classes for each ingredient. As a result, data users who are interested in specific therapeutic classes corresponding to each of a drug’s ingredients would need additional information not on the public use files.
As an example, a combination-product cold remedy might be assigned a therapeutic class of “upper respiratory combinations” in the Multum database. Only by going to a separate ingredient file (which includes therapeutic classes for each ingredient of a drug), would a researcher be able to identify that the product contained an antihistamine, antitussive, and a decongestant. It is necessary to merge the public use file with the ingredient file in order to get this information.
Q. How can I compare drug data across years when the classification system is different starting with 2006 data?
A. NCHS has developed a SAS program that will read in a previous year of NAMCS or NHAMCS data, match medication codes from that dataset with codes from the current ambulatory care drug database, and then drop the old drug characteristics associated with that medication code, while replacing them with the new Multum characteristics. The original program DRUGCHAR_MULTUM_2006.SAS can be downloaded from the Ambulatory Health Care Data website, along with the ASCII text file MEDCODE_DRUGID_MAP_2006.DATA that is used to replace the drug characteristics. Programs and databases for 2007 and 2008 are also available, and will continue to be added for each new year of data.
It is recommended that you always use the MOST RECENT database and program to add drug characteristics backwards, but programs and databases for prior years are being retained in separate folders for historical purposes. The SAS programs are ready to use, only requiring that you change path names in the program to reflect your own situation. Each new program should work for ALL previous years of data. For example, the 2008 program and database are currently the latest version, and can be used to update drug characteristics on all previous years of drug data. The only caveat is that by updating drug characteristics to the current year, you may lose older characteristics that may be of interest. For example, if a drug was a prescription-only preparation in 1995 and switched to over-the-counter in 2007, by using the 2008 characteristics, the drug would be coded as over-the-counter for all years of data.
Q. How can the ingredient data be combined with the basic public use file?
A. NCHS has also developed a SAS program that will add the ingredient file to the basic public use file, matching on the DRUGID variable. This program and the corresponding ingredient file were first developed for use with 2006 data (ADD_DRUG_INGREDIENTS_2006.SAS and DRUG_INGREDIENTS_2006.DATA). These programs and files continue to be added as each new year of data is released.
All of these resources can be downloaded at the Ambulatory Health Care Data website in the Dataset Documentation section, under a directory on the ftp server called ‘drugs’.
IMPORTANT NOTE: The ingredient programs and files are year-specific. In other words, if you want to add ingredients to the 2006 file, use the 2006 program and ingredient file. It is not recommended that you use the 2007 program and ingredient file with non-2007 survey years, or vice versa. The programs only require that you change path names in the programs to reflect your own situation.
SAS exercises (including examples of how to use the programs above) and a PowerPoint presentation, which were developed for a workshop at the 2008 NCHS Data Users Conference, are also available for downloading at our website under Staff Presentations. Updated SAS exercises and the PowerPoint presentation from the recent 2010 National Conference on Health Statistics should be available at the website in the near future.
If you need more information or technical assistance, please contact the Ambulatory and Hospital Care Statistics Branch at 301-458-4600