Public Use Data Documentation Updates

This page contains a list of updates that have been made to the NAMCS and NHAMCS public-use file documentation since their initial release. Periodically, typographic errors and other inaccuracies are corrected by uploading revised files to the ftp server. Data users can check here for the revisions that have been made and to ensure that they are using the most accurate version of the survey documentation. Minor corrections, if any (such as spelling or formatting), that have no substantive bearing on the documentation content are not listed.  Updates are arranged by survey and year.

• 2015 NAMCS Public Use Data File

  9/11/2017 – The label used for the YEAR variable was corrected from “Survey year (2014)” to “Survey year (2015)” in the SAS label statement, the SPSS syntax file, and the Stata do file. The item itself is correct (2015 is the value shown in the data), but anyone using the provided labels would notice that the name of the variable was not.

• 2014 NAMCS and NHAMCS-ED Public Use Data File Documentation

  6/2/2017 – The Hispanic and Non-Hispanic population estimates on pages 22-25 (NAMCS) and pages 30-33 (NHAMCS-ED) have been replaced to show 2014 estimates. The original versions showed 2013 estimates by mistake.

• 2010 NAMCS Public Use File Documentation

  9/11/2012 – On page 12 (Table II) and pages 14 and 16 (Tables III), the columns of population estimates for ages 5-14, 15-24, and 25-34 were misprinted. In the original version, the column labeled 5-14 years actually showed data for 5-9 years, the column labeled 15-24 years actually showed 10-14 years, and the column labeled 25-34 actually showed 15-24 years. These pages have been corrected.

• 2009 NAMCS Public Use Data File Documentation

  11/7/11: On page 73, the item description for RACERETH was updated by reversing categories 3 and 4. Category 3 is Hispanic and Category 4 is Other Race/Multiple Race, Non-Hispanic. The SAS format statement is not affected

  6/16/11 – Typographical errors corrected: On pages 75 and 76, the names of the numeric recode variables for write-in procedures were updated. These are correct in the corresponding SAS documentation.

• 2009 NAMCS SAS Statements

  8/31/11 – 2009 NAMCS SAS Statements were corrected to reflect the naming of the PASTVIS variable. This variable was written as PASTVISR in the SAS code, but should be PASTVIS, to match the 2009 Public Use File Documentation and past years of data.

• 2007-2009 NAMCS and NHAMCS Public Use Data File Documentation

  8/31/11 – 2007, 2008, and 2009 NAMCS and NHAMCS Public Use File Documentation was updated to correct the categories for INJDET in the NAMCS and OPD Record Format sections. The corresponding SAS formats were already accurate, however. Note that the item (Injury/Poisoning/Adverse Effect) was changed in 2007 with the addition of a category for ‘Injury/Poisoning – Unknown Intent.’

• 2008 NAMCS and NHAMCS Public Use Data File Documentation and SAS Statements

  4/29/11 – Dataset documentation and SAS files were reissued to rename the PATYPE variable to PATYPER (Recoded Primary Expected Source of Payment). This is intended to emphasize the fact that PAYTYPER is a recoded item which uses a hierarchy to assign a primary expected source of payment based on the collection of multiple expected sources of payment. More information can be found in the annual documentation.

  In addition, dataset documentation for both files had the following text added:
  “Users of NAMCS imputed race and ethnicity data should note that the initial release of this documentation included the following recommendation: ‘Analyses that treat the imputed values as if they were true values will tend to underestimate standard errors. An adjustment that would often be conservative would be to multiply the estimated standard error by [1 over the square root of c], where c is the proportion of records in the analysis having complete data on race and ethnicity. Note that this adjustment is analysis-specific; there is not one inflation factor that applies to all analyses’.

  This was put forward as a conservative approach for data users. However, it has subsequently been determined, based on much additional research by NCHS staff concerning imputation of NAMCS race and ethnicity data, that this adjustment appears not to be necessary in many cases, particularly when the design effect is already large. Therefore, NCHS is no longer making a specific recommendation for variance inflation adjustment with imputed NAMCS race and ethnicity data.”

• 2008 NAMCS and NHAMCS Drug Documentation

  12/5/2012 – The program DRUGCHAR_MULTUM_2008.SAS was updated to replace an error in one line. The data file MEDCODE_DRUGID_MAP_2008.DATA was replaced to correct a missing drug code. Users are reminded that the DRUGCHAR program for a specific year should only be used in conjunction with the map database for that year. Each year’s map reflects the drug database at a specific point in time. The program and map are to be used for assigning drug characteristics based on the Multum Classification at a specific point in time to older years of drug data, especially to those years prior to 2006 when a different classification was used. Being able to compare drug data using a single classification across years facilitates trend analysis.

  Each year’s map may contain corrections to prior versions of the database. In general, it is recommended that the most recent version of the program and mapping database be used. However, researchers should keep in mind that a drug’s characteristics may have legitimately changed across years; for example, a drug may have been prescription only in an earlier year and changed to over-the-counter in a later year. By standardizing drugs in previous years according to the most recent characteristics, such differences will be lost and would have to be manually recoded.

• 2007 NAMCS and NHAMCS Public Use File Documentation

  2/8/12 – Replaced page 19 (NAMCS) and page 31 (NHAMCS). These pages show population estimates for Non-Hispanic persons in the age groups 35-44, 45-54, 55-64, 65-74, and 75 years and over. The original versions were found to be unintentional duplicates of page 15 (NAMCS) and page 27 (NHAMCS).

• 2007 NAMCS Public Use Data File Documentation

  10/6/09 – Marginal data for “Patient Age” (pages 83 and 86) and “Has patient been seen in your practice before?” (page 84) were revised slightly. Several late edits were made to the 2007 NAMCS data prior to its release that had not been reflected in the earlier version of the documentation.

• 2006 NAMCS

  5/29/07 – Corrected Marginal Data for NAMCS visits by physician specialty on page 84 and physician estimates by specialty, solo practice and specialty type on page 88.

• 2006 NAMCS Stata Documentation

  2/11/2013 – It was recently discovered that some of the variables in the Stata dictionary file were defined using an incorrect type. For example, YEAR had been defined as ‘byte’ which caused data for this variable to be missing in the Stata dataset. The dictionary file has been corrected and the compressed Stata dataset has been replaced.

• 2005 NAMCS and NHAMCS

  11/28/07 – Corrected Marginal Data for NAMCS visits by age and sex, and drug mentions by age. Corrected variable names HOMVISR, HOSVIR, TELCONR, and ECONR, page 25 (NAMCS). Modified description of “Not Applicable” for GESTWEEK on page 28 (NAMCS) and page 61 (NHAMCS).

• 2004 NAMCS and NHAMCS

  4/7/06 – Corrected labels for Asian, Native Hawaiian/Other Pacific Islander (NHOPI) and American Indian/Alaska Native (AIAN) categories in U.S. population tables on pages 10 and 11 in NAMCS documentation and pages 20 and 21 in NHAMCS documentation.

• 2003 NAMCS File – Stata Documentation

  10/18/05 – A new namcs03.dct file was added to the Stata documentation for the 2003 NAMCS public use file on 10/18/05. The new file corrects some problems that were found with the Stata definitions for AGEDAYS, PATCODE, and TIMEMD. If you have used these variables with an earlier version of the .dct file, please note that item values greater than 100 may have been dropped in your output. We apologize for the inconvenience.

• 2002 NAMCS

  5/11/05 – Revised response rates for individual physician strata on page 9.
  9/27/04 – Revised the SPSS syntax file to correct a number of problems.
  7/29/04 – Replaced marginal data for National Drug Code Directory drug classes (page 63) in the general documentation (DOC02.PDF) file.

• 2001 NAMCS

  12/15/03 – These changes pertain to the general documentation (DOC01.PDF) file. Added Appendix IV which contains comparisons of 2 different formats used to collect 3 data items in 2001. Also updated the Drug Entry List to include some missing drugs.
  9/22/03 – To date, no changes have been needed for the SAS documentation. Value labels were updated in the general documentation (DOC01.PDF) Record Format section for the following items: ETHNIC, RACE, PRIMCARE, REFER, EPISODE, RETYPOFF, LAB, and OWNS.

• 1995 and 1996 NAMCS and NHAMCS-OPD SAS Input Statements

  3/10/11 – It has recently come to our attention that the SAS input statements for these files contained a missing dot in the entries for four variables: BLODLEAD, CHOLEST, MRI, and OTHCOUN. For example, in the line identifying the BLODLEAD item on the 1995 and 1996 NAMCS input statements:
  @101 BLODLEAD 1. /* Blood lead level */
  The dot after ‘1’ was missing. Such an error would not abort the program, but would result in incorrect data being displayed for these four variables. Corrections have been made to the SAS input statements on the FTP server. We apologize for any inconvenience this may have caused.

• Revised Population Estimates for the 1995 and 1996 NAMCS CD-ROMs and the 1996 NHAMCS CD-ROM

  This notice is for anyone who still uses NAMCS or NHAMCS data on CD-ROM. It was discovered that population estimates in the pop.dat file of the 1995 and 1996 NAMCS CD-ROMs and the 1996 NHAMCS CD-ROM omit persons 100 years and over. (The 1995 NHAMCS CD-ROM population figures are correct.) The discrepancy is small (0.01 percent of the total population estimate) and does not impact most rate calculation using NAMCS or NHAMCS data. However, users focusing on visits made by very elderly subgroups may be getting incorrect rate results with the SETS software.

  The missing figures for persons 100 years and over are listed below:

  Year	Age	Total	Total Male	Total Female	White male	White female	Black male	Black female	Other male	Other female
  1995	100+	29,196	6,731	22,465	4,933	17,669	1,412	3,963	386	833
  1996	100+	31,766	7,174	24,592	5,243	19,388	1,478	4,223	453	981

  Note that SETS sums individual population estimates in the pop.dat file to use as denominators in rate calculation for larger groups. The denominators will, therefore, be off by the small amounts shown above, but this omission should have little effect on visit rates for all but the most elderly.

NHAMCS

• 2009 NHAMCS – Emergency Department Public Use File Documentation and SAS Format file

  1/12/12 – On page 35 of the File Documentation, Patient Residence was updated to reflect changes on the 2009 Patient Record form. The correct label for category 3 is Homeless. The correct label for category 4 is Other. There is no category 5 as such (because entries of unknown are converted to -8 for all items during data processing). These changes were also made on the SAS Format file.
  9/20/11 – The 2009 NHAMCS Emergency Department Public Use File has been updated to include a change to the visit disposition variable ADMITHOS (admit to this hospital). Original file data are correct, but it was discovered that some researchers were not aware that the disposition OBSHOS (admitted to observation unit, then hospitalized) should also be used to capture visits resulting in hospital admission. To simplify matters, we have edited ADMITHOS to Yes whenever OBSHOS is Yes. Therefore, ADMITHOS will now reflect all visits resulting in hospital admission, either directly or through the observation unit. A note regarding this change has also been added to the Public Use Data File Documentation. Researchers using the original version of the file who are interested in hospital admissions need only select records where ADMITHOS is Yes; or OBSHOS is Yes to capture all hospital admissions.

• 2007-2009 NAMCS and NHAMCS Public Use Data File Documentation

  8/31/11 – 2007, 2008, and 2009 NAMCS and NHAMCS Public Use File Documentation was updated to correct the categories for INJDET in the NAMCS and OPD Record Format sections. The corresponding SAS formats were already accurate, however. Note that the item (Injury/Poisoning/Adverse Effect) was changed in 2007 with the addition of a category for ‘Injury/Poisoning – Unknown Intent.’

• 2008 NAMCS and NHAMCS Drug Documentation

  12/5/2012 – The program DRUGCHAR_MULTUM_2008.SAS was updated to replace an error in one line. The data file MEDCODE_DRUGID_MAP_2008.DATA was replaced to correct a missing drug code. Users are reminded that the DRUGCHAR program for a specific year should only be used in conjunction with the map database for that year. Each year’s map reflects the drug database at a specific point in time. The program and map are to be used for assigning drug characteristics based on the Multum Classification at a specific point in time to older years of drug data, especially to those years prior to 2006 when a different classification was used. Being able to compare drug data using a single classification across years facilitates trend analysis.
  Each year’s map may contain corrections to prior versions of the database. In general, it is recommended that the most recent version of the program and mapping database be used. However, researchers should keep in mind that a drug’s characteristics may have legitimately changed across years; for example, a drug may have been prescription only in an earlier year and changed to over-the-counter in a later year. By standardizing drugs in previous years according to the most recent characteristics, such differences will be lost and would have to be manually recoded.

• 2008 NAMCS and NHAMCS Public Use Data File Documentation and SAS Statements

  4/29/11 – Dataset documentation and SAS files were reissued to rename the PATYPE variable to PATYPER (Recoded Primary Expected Source of Payment). This is intended to emphasize the fact that PAYTYPER is a recoded item which uses a hierarchy to assign a primary expected source of payment based on the collection of multiple expected sources of payment. More information can be found in the annual documentation.
  In addition, dataset documentation for both files had the following text added:
  “Users of NAMCS imputed race and ethnicity data should note that the initial release of this documentation included the following recommendation: ‘Analyses that treat the imputed values as if they were true values will tend to underestimate standard errors. An adjustment that would often be conservative would be to multiply the estimated standard error by [1 over the square root of c], where c is the proportion of records in the analysis having complete data on race and ethnicity. Note that this adjustment is analysis-specific; there is not one inflation factor that applies to all analyses’.
  This was put forward as a conservative approach for data users. However, it has subsequently been determined, based on much additional research by NCHS staff concerning imputation of NAMCS race and ethnicity data, that this adjustment appears not to be necessary in many cases, particularly when the design effect is already large. Therefore, NCHS is no longer making a specific recommendation for variance inflation adjustment with imputed NAMCS race and ethnicity data.”

• 2008 NHAMCS Public Use Data File Documentation

  1/5/11 – Typographical errors corrected: The order of expected sources of payment categories used in developing the PAYTYPE hierarchy was corrected in the text to put Medicare first, followed by Medicaid/SCHIP, on pages 33 and 74. This textual order was changed to reflect the change in codingof the data as described on those pages.

• 2007 and 2008 NHAMCS Emergency Department Public Use File Documentation

  7/11/12 – On page 44 (for 2007) and page 42 (for 2008), two labels in the Visit Disposition variable list were found to be mistyped. Item 135, [LEFT] Left without being seen, should be [DOA] Dead on arrival. Item 136, [DOA] DOA/Died in ED, should be [DIEDED] Died in ED. Both pages have been corrected in the downloadable documentation. The SAS code for both years was correct and has not been changed.

• 2007 and 2008 NHAMCS Stata Documentation

  2/11/2013 – It was recently discovered that some variables, such as CSTRATM and the recoded diagnosis variables DIAG1R -DIAG13, were defined as type ‘int’ when they should have been type ‘long’. This error caused data for the affected variables to be missing in the Stata dataset. The dictionary files have been corrected and the compressed Stata datasets have been replaced.

• 2007 NAMCS and NHAMCS Public Use File Documentation

  2/8/12 – Replaced page 19 (NAMCS) and page 31 (NHAMCS). These pages show population estimates for Non-Hispanic persons in the age groups 35-44, 45-54, 55-64, 65-74, and 75 years and over. The original versions were found to be unintentional duplicates of page 15 (NAMCS) and page 27 (NHAMCS).

• 2007 NHAMCS Public Use Data File Documentation

  10/20/09 – The total weighted number of emergency department visits on page 22 and marginal data for emergency department visits on pages 112-116 were updated. Several late edits were made to the 2007 NHAMCS data prior to release that had not been reflected in the earlier version of the documentation.

• 2005 and 2006 NHAMCS – OPD

  12/1/11 – An error was recently discovered in the labeling of the values for 1 and 2 in the USETOBAC item on page 61 of the 2005 documentation and page 72 of the 2006 documentation. The correct labels are as follows:
  0 = Blank
  1 = Not current
  2 = Current
  3 = Unknown
  The SAS formats for this item were always correct, however. NAMCS is not affected.

• 2005 NAMCS and NHAMCS

  11/28/07 – Corrected Marginal Data for NAMCS visits by age and sex, and drug mentions by age. Corrected variable names HOMVISR, HOSVIR, TELCONR, and ECONR, page 25 (NAMCS). Modified description of “Not Applicable” for GESTWEEK on page 28 (NAMCS) and page 61 (NHAMCS).

• 2004 NAMCS and NHAMCS

  4/7/06 – Corrected labels for Asian, Native Hawaiian/Other Pacific Islander (NHOPI) and American Indian/Alaska Native (AIAN) categories in U.S. population tables on pages 10 and 11 in NAMCS documentation and pages 20 and 21 in NHAMCS documentation.

• 2002 NHAMCS-ED and OPD

  9/27/04 – Revised the SPSS syntax file to correct a number of problems.
  7/29/04 – Replaced marginal data for National Drug Code Directory drug classes (pages 94, 98) in the general documentation (DOC02.PDF) file.

• 2001 NHAMCS – ED

  9/22/03 – To date, no changes have been needed for the SAS documentation. Value labels were updated in the general documentation (DOC01.PDF) Record Format section for the following items: ETHNIC, RACE.

• 2001 NHAMCS – OPD

  9/22/03 – To date, no changes have been needed for the SAS documentation. Value labels were updated in the general documentation (DOC01.PDF) Record Format section for the following items: ETHNIC, RACE. PRIMCARE, REFER, and EPISODE.

• 1995 and 1996 NAMCS and NHAMCS-OPD SAS Input Statements

  3/10/11 – It has recently come to our attention that the SAS input statements for these files contained a missing dot in the entries for four variables: BLODLEAD, CHOLEST, MRI, and OTHCOUN. For example, in the line identifying the BLODLEAD item on the 1995 and 1996 NAMCS input statements:
  @101 BLODLEAD 1. /* Blood lead level */
  The dot after ‘1’ was missing. Such an error would not abort the program, but would result in incorrect data being displayed for these four variables. Corrections have been made to the SAS input statements on the FTP server. We apologize for any inconvenience this may have caused.

• 1995 and 1996 NAMCS and NHAMCS-OPD SAS Input Statements

  3/10/11 – It has recently come to our attention that the SAS input statements for these files contained a missing dot in the entries for four variables: BLODLEAD, CHOLEST, MRI, and OTHCOUN. For example, in the line identifying the BLODLEAD item on the 1995 and 1996 NAMCS input statements:
  @101 BLODLEAD 1. /* Blood lead level */
  The dot after ‘1’ was missing. Such an error would not abort the program, but would result in incorrect data being displayed for these four variables. Corrections have been made to the SAS input statements on the FTP server. We apologize for any inconvenience this may have caused.

• Revised Population Estimates for the 1995 and 1996 NAMCS CD-ROMs and the 1996 NHAMCS CD-ROM

  This notice is for anyone who still uses NAMCS or NHAMCS data on CD-ROM. It was discovered that population estimates in the pop.dat file of the 1995 and 1996 NAMCS CD-ROMs and the 1996 NHAMCS CD-ROM omit persons 100 years and over. (The 1995 NHAMCS CD-ROM population figures are correct.) The discrepancy is small (0.01 percent of the total population estimate) and does not impact most rate calculation using NAMCS or NHAMCS data. However, users focusing on visits made by very elderly subgroups may be getting incorrect rate results with the SETS software.

  The missing figures for persons 100 years and over are listed below:

  Year	Age	Total	Total Male	Total Female	White male	White female	Black male	Black female	Other male	Other female
  1995	100+	29,196	6,731	22,465	4,933	17,669	1,412	3,963	386	833
  1996	100+	31,766	7,174	24,592	5,243	19,388	1,478	4,223	453	981

  Note that SETS sums individual population estimates in the pop.dat file to use as denominators in rate calculation for larger groups. The denominators will, therefore, be off by the small amounts shown above, but this omission should have little effect on visit rates for all but the most elderly.