Study design and sample
This study used NHIS pooled data from 2002, 2006, 2009, and 2014, linking the Adult Core sample file and Person file for each year. The NHIS is an ongoing, multistage probability cross-sectional in-person household survey of a nationally representative sample of the U.S. noninstitutionalized population residing in all 50 U.S. states and the District of Columbia. (19) We selected adults (ages 18 years and older) with self-reported arthritis defined by a “yes” response to the item, “Have you EVER been told by a doctor or other health professional that you have some form of arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia?” Responses of “no” and those with unknown arthritis status were excluded. For the remainder of the manuscript, the phrase “doctor or other health professional” is referred to as physician. The authors utilized publicly available de-identified data, therefore, this study does not constitute human subjects research.
Meets Aerobic Physical Activity Guidelines. The Health and Human Services (HHS) 2008 Physical Activity Guidelines for Americans 2nd edition recommend 150 minutes/week of moderate aerobic PA (e.g., brisk walking) or 75 minutes/week of vigorous aerobic PA (e.g., running) or an equivalent combination for all adults.  Individuals reported weekly frequency and duration of moderate and vigorous aerobic PA. Total aerobic PA was assessed by combining moderate and vigorous aerobic PA, where 1 minute of vigorous aerobic PA is equivalent to 2 minutes of moderate aerobic PA. (14) Those reporting ≥ 150 minutes/week of total aerobic PA were considered to have met aerobic PA guidelines.
Meets Muscle Strengthening Physical Activity Guidelines. The HHS 2008 guidelines for muscle strengthening PA recommend ≥ 2 strengthening activities per week. Individuals that report lifting weights or doing calisthenics ≥ 2 times/week met muscle strengthening PA guidelines.
Arthritis-Associated Activity Limitation. This variable was assessed using the item “Are you now limited in any of your usual activities because of your arthritis or joint symptoms?” Response options were yes/no.
Severity of arthritis-related pain. Individuals were asked if they experienced symptoms of pain, aching, or stiffness in or around a joint in the past 30 days. Those responding “yes” were asked to rate their average joint pain during the past 30 days on a scale of 0 to 10; 0 = no pain or aching and 10 = pain and aching as bad as it can be. Those responding “no” to the initial question were considered to have no pain and assigned a value of 0 (no pain); thus scores ranged from 0–10 with a higher score indicating more pain.
Self-reported race/ethnicity. The race/ethnicity item used response options of: Hispanic, non-Hispanic White, Non-Hispanic Black, Non-Hispanic Asian, and Non-Hispanic/all other races. Hereafter, we refer to these groups as Latino, White, African American, and Asian. Respondents reporting non-Hispanic/all other races were dropped from the analyses.
Receipt of physician recommendation to exercise to help arthritis. This variable (yes/no) was assessed using the question, “Has a doctor or other health professional EVER suggested physical activity to help your arthritis or joint symptoms?” Those with missing responses were excluded.
Demographic covariates included age (18–44, 45–64, 65, or ≥ 65 years), sex (male or female), educational attainment (less than high-school degree, high-school graduate, technical college/some university, four year college degree or higher), marital status (yes/no), employment status (yes/no), annual household income (0-$34,999, $35,000–64,999, and ≥ $65,000), health insurance (yes/no), having a usual source of care (yes/no), and U.S. region (Northeast, Midwest, South, or West).
Health covariates included: smoking status (never, former, or current); body mass index (BMI calculated from self-reported weight and height as a continuous measure (weight in kg/height in m2) categorized as underweight (< 18.5), normal weight (18.5–24.9), overweight (25.0-29.9), or obese (≥ 30); and self-rated health (excellent/very good/good versus fair/poor); number of comorbidities (0, 1–2, or ≥ 3) as a count of the following conditions: asthma, cancer, chronic obstructive pulmonary disease, heart disease, hepatitis, diabetes, kidney disease, hypertension, psychological distress, and stroke; and psychological distress assessed with the Kessler-6 and categorized into 3 levels based on the sum score: none/mild (0–4), moderate (5–12), and severe (≥ 13). (20, 21)
Descriptive statistics were used to examine the distributions of demographic, health characteristics, and outcomes by race/ethnicity.
Multivariate logistic regression was used to analyze the independent effects of race/ethnicity and receipt of physician recommendation on the odds of meeting aerobic and strengthening PA guidelines, and having activity limitations. Additionally, we assessed whether the effects of receipt of physician recommendation to exercise on these outcomes was moderated by race/ethnicity (included an interaction term for race/ethnicity x receipt of physician recommendation) in each model. Covariates included demographic characteristics (age, sex, education, marital status, income, employment, health insurance, having a usual source of care) and health characteristics (smoking status, BMI, self-reported health, comorbidities, psychological distress) and region.
Multivariate linear regression was used to analyze the independent effects of race/ethnicity and receipt of physician recommendation on joint pain severity, and whether these effects were moderated by race/ethnicity. An interaction term for race/ethnicity x receipt of physician recommendation was included. Covariates included demographic characteristics (age, sex, education, marital status, income, employment, health insurance, having a usual source of care) and health characteristics (smoking status, BMI, self-reported health, comorbidities, psychological distress) and region.
The NHIS sampling weights for the Sample Adult Core were used for all analyses to generate nationally representative population estimates. SAS/STAT software PROC SURVEY SELECT was utilized to insure correct estimates and standard errors.