Leisure-time Physical Activity Recodes
Physical activity recodes are generally not included on NHIS data files. However, SAS and Stata code are provided here to allow data users to construct recodes for adult leisure-time physical activity consistent with the indicators used to track progress toward national health objectives.
Since 1997, questions about usual light-moderate and usual vigorous leisure-time physical activity have been included on the NHIS Sample Adult core questionnaire. The question wording differed slightly from January-June 1997, but the questions have remained essentially unchanged since July 1997. In 2004, the phrase “leisure-time” was inserted in EACH question, (previously it appeared only in the introductory phrase for the LTPA question series.)
In addition to being used to track Healthy People Objectives, estimates based on the Sample Adult Core questions are routinely published in "Early Release of Selected Estimates from the National Health Interview Survey," "Health, United States," and "Summary Health Statistics for U.S. Adults (beginning in 2007)."
- SAS code for Healthy People 2020 LTPA recodes [2008 Physical Activity Guidelines] (SA_PA2020.sas)
- SAS code for Healthy People 2010 LTPA recodes (SA_PA2010.sas)
- Stata code for Healthy People 2020 LTPA recodes [2008 Physical Activity Guidelines] (SA_PA2020.do)
Prior to 1997, physical activity questions were asked in supplement questionnaires in selected years. Major Health Promotion and Disease Prevention/Healthy People Supplements were included in 1985, 1990, 1991, 1995, and 1998. Each contained a series of questions asking frequency, duration, and intensity of more than 20 leisure-time sports, exercises, and physically active hobbies. Using this information and applying estimates of kilocalorie expenditures associated with each activity, two recodes are created:
- KCAL - a estimate of average total kilocalories per kilogram of body weight per day over a recent two-week period, and
- PA_LEVEL - a categorical variable based on kilocalorie (kcal) expenditure: sedentary (kcal < 1.5), moderately active (1.5 ≤ kcal < 3.0), and very active (kcal ≥ 3.0).
The questionnaires for 1985 and 1990 were identical. Similarly, the questionnaires for 1991, 1995 and 1998 were identical. Caution should be used when studying trends that bridge the 1990 to 1991 questionnaire change. Age restrictions that were imposed for many of the leisure-time physical activities in the 1985 and 1990 questionnaires were removed beginning in 1991. Although the questions remained the same within these two time periods, valid value ranges and coding of "unknown" and "not asked" categories differed across the data years.
Description of the logic underlying the calculation of KCAL across all data years:
- KCAL = the sum of the kilocalorie expenditures for all mentioned activities for which both frequency and duration have valid values. [If heart rate is missing, it is imputed to the lowest value for purposes of calculating the recode.]
- Kilocalorie values for individual activities [KCALPI(I)] are calculated based on frequency, duration, and heart rate intensity during the activity. The individual KCALPI(I) values, rounded to one decimal, are then summed to create the recode KCAL, the total average daily kilocalorie expenditure over the two–week period. The values for individual activities [KCALPI(I)] are stored in the data file and may be of methodological interest but are not in themselves considered recodes.
- If an activity is mentioned but frequency or duration is missing and no other activities are mentioned, then KCAL=99999.9.
- If all activities are "no," then KCAL=00000.0.
- If some activities are "no" and others are unknown, then KCAL=99999.9
- If some or all activities were not asked (due to age or handicap), and the respondent did not volunteer that he or she engaged in the activities, "not asked" is treated as "no" for purposes of calculating KCAL. If the respondent volunteered one or more activities, his or her KCAL value is based on those volunteered activities.
A summary of the Energy Expenditures assigned in these recodes is provided.
SAS code for generating the LTPA recodes for the Health Promotion and Disease Prevention (1990)/Year 2000 Objectives Supplement data files for each year:
- SAS code for 1998 Recodes (HP98_KCAL.sas)
- SAS code for 1995 Recodes (HP95_KCAL.sas)
- SAS code for 1991 Recodes (HP91_KCAL.sas)
- SAS code for 1990 Recodes (HP90_KCAL.sas)
- SAS code for 1985 Recodes (HP85_KCAL.sas)
NOTE: The recodes KCAL and PA_LEVEL exist on the 1998 NHIS Year 2000 Objectives public use data file. For consistency across data years, input variables for 1995, 1991, 1990, and 1985 have been renamed to be consistent with the 1998 naming conventions.
In 1998, the core and supplement questions on leisure-time physical activity were both asked of the same adult respondents during the same interview. Using the recodes specified here, the Comparison of Two Sets of Adult Leisure-Time Physical Activity Questions: 1998 National Health Interview Survey compares the two sets of questions in terms of population prevalence (Table A) and in terms of their associations with selected health status indicators (Table B). This comparison is limited to the recode for the Healthy People 2010 physical activity indicator and has not been replicated with the indicators for the Healthy People 2020 [2008 Physical Activity Guidelines] indicator.