Previously we have described our methods of analysis [2
]. We analyzed data from the Sample Core component and the Alternative Health Supplement to the 2002 National Health Interview Survey (NHIS). The NHIS is an in-person household survey conducted by the Census Bureau for the National Center for Health Statistics, and is the principal source of information in the U.S. on the health of the civilian, non-institutionalized household population. One adult was randomly selected from each household to complete this portion of the survey. There were 31,044 completed interviews, with a 73.4% response rate. The sampling methods for the NHIS are described elsewhere [18
For the NHIS, respondents were specifically asked, "Some people use natural herbs for a variety of health reasons. Some people drink an herbal tea to remedy a flu or cold. Others take a daily pill to help with a health condition or just stay healthy. Have you ever used natural herbs for your own health or treatment (for example ginger, echinacea or black cohosh including teas, tinctures, and pills) [18
]?" Those respondents that said yes were asked "During the past 12 months, did you use natural herbs for your own health or treatment?" For the purpose of this analysis, we will refer to any herb use as having occurred in the prior 12 months. Of those that said yes to this question, respondents were asked a series of questions about individual HDS use and disclosure of that use to conventional health care providers.
Because we were interested in high risk HDS behaviors, we excluded use of vitamins and minerals (which may have been recommended by health professionals), and focused on non-vitamin/mineral supplements. Respondents chose from a list of 35 non-vitamin/mineral supplements dietary supplements (29 supplements were plant based and 6 supplements not plant based these included: s-adenosylmethionine (SAM-e) progesterone cream, melatonin, bee pollen, fish oil, glucosamine and chondroitin).
We defined a young adult age range between 18 years of age to 30 years of age based on education and earning potential. We considered socio-demographic factors including respondents' age (18–22, 23–30) gender, education (<high school, high school graduate, some college, college graduate), annual family income (<$15,000, $15,000–34,999, $35,000–64,999, ≥ $65,000), race/ethnicity (Hispanic, non-Hispanic white, non-Hispanic black non-Hispanic other (Asian, American Indian/Alaskan native, Asian Indian, Chinese, and Filipino) and region of U.S. residence (Northeast, Midwest, South, West).
We also included lifestyle/behavioral factors that we thought could interact with or cause adverse events with HDS. These included smoking (current, former, nonsmoker); alcohol use in the past 12 months (abstainer or former drinker, current infrequent or light drinker (< 3 drinks a week), current moderate to heavy (>3 drinks per week to >14 drinks a week), or unknown drinking status);
Although physical activity level, in itself, is not a high risk behavior, adolescent and adult athletes have been reported to use herbs and other performance enhancing supplements more frequently than others of their age groups. Therefore, we analyzed a physical activity level variable [high (vigorous activity 2 times/wk or moderate activity 4 times/wk), moderate (vigorous activity 1 time/wk or moderate activity 1–3 times/wk), or sedentary (no vigorous or moderate activity/week)] in the model to assess for an association high physical activity with HDS in a national population of youth [19
Respondents were asked if they used prescription and over the counter (OTC) medications in the last 12 months. Respondents were asked if they disclosed natural herb use to conventional medical professionals including physicians, nurse practitioners, physician assistants, psychiatrists, and dentists.
Population estimates were calculated using NHIS weights, which are calibrated to U.S. 2000 census totals for gender, age, and race/ethnicity of the 2002 U.S. population. Descriptive statistics were used to examine the prevalence of HDS use, the most common HDS used among smokers and alcohol drinkers, and disclosure to medical professionals. We used multivariate logistic regression analysis to assess which variables were significantly associated with HDS use. We selected lifestyle and behavioral variables for testing in our logistic model adjusting for age, sex, race, income, education and entered these variables into the final model simultaneously. All analyses were performed using SAS- callable SUDAAN version 8.1 (Research Triangle Institute, Research Triangle Park, NC) [20
] to account for the complex sampling design of the NHIS.