Background
In 2015, arthritis affected 23% (54.4 million) of U.S. adults, [
1] and is estimated to affect 78.4 million by 2040 [
2]. The prevalence of disability due to arthritis is expected to increase from 22.7 to 34.6 million Americans by 2040, a 52% increase from 2012 [
2]. Between 2013 and 2015, the average prevalence of arthritis was similar between Whites and African Americans (22.6% vs 22.2%) but lower among Latinos (15.7%) and Asians (11.8%) [
1]. However, arthritis-related activity limitations and pain were significantly higher among African Americans, Latinos, mixed-race, and American Indian/Native Alaskan populations compared to Whites [
1]. Effective interventions to address racial/ethnic disparities in symptom burden are warranted.
Physical activity (PA) is a safe, non-pharmacologic, evidence-based intervention to reduce arthritis symptoms. PA is effective at reducing pain and improving function among persons with moderate-to-severe arthritis and compared to the short-term benefits of nonsteroidal anti-inflammatory drugs and opioids, regular exercise provides more durable benefits [
3‐
8]. Also, exercise improves mood and quality of life in those with chronic pain due to arthritis [
9,
10]. Although some evidence suggests that strengthening exercise has a larger effect on pain reduction, [
11] both aerobic and strengthening exercises are efficacious in reducing pain and improving function [
7]; offering patients a choice between the two may improve adherence [
7]. Musculoskeletal organizations such as the American Academy of Orthopedic Surgeons (AAOS) and the Osteoarthritis Research Society International (OASRI) have issued consistent recommendations for regular PA for persons with arthritis [
12].
Despite strong evidence that PA is beneficial for people with arthritis, almost one-third of persons with arthritis are completely inactive, and only one-quarter adhere to national PA recommendations [
13]. The picture is even bleaker for minorities. African Americans and Latino adults with arthritis were less likely than Whites to meet aerobic physical activity guidelines [
14]. In one study among persons with or at risk of knee osteoarthritis, African Americans were 72–76% less likely than Whites to meet PA guidelines [
15].
Receiving advice from a physician to exercise is associated with a higher likelihood of meeting aerobic PA guidelines among those with arthritis [
16]. In 2011, only 60% of persons with arthritis received a physician recommendation to exercise for relief of arthritis symptoms [
16]. Several studies have demonstrated that African Americans and Latinos with arthritis were less likely to receive PA advice from physicians than their White counterparts [
12,
17]. It is plausible that physician implicit bias could affect whether a physician gives an exercise recommendation for patients from certain racial/ethnic groups because evidence shows African Americans are less likely to receive appropriate pain management compared to White Americans [
18]. Also, patients most likely to benefit (obese/overweight and those with higher pain levels, comorbidities, and activity limitations) were less likely to have received a physician recommendation to exercise [
19]. Yet, studies support recommending PA to reduce arthritis symptoms regardless of patient profile, including radiologic severity and pain levels [
19]. Disparities in receipt of physician recommendation for PA for arthritis symptoms could help explain disparities in symptom burden.
Studies examining the role of physician recommendation for PA among persons with arthritis have been limited in that they only asked about PA in general, rather than differentiating between aerobic and muscle strengthening activities. It is important for physicians to know if there is systematic variation in patient preferences in type of PA, as this information could be used to increase adherence [
7].
Using data from the National Health Interview Survey (NHIS), the objectives of this study were to examine: 1) the association of participant race/ethnicity with meeting physical activity guidelines (aerobic and strengthening) and arthritis symptoms (arthritis-associated activity limitations and severity of joint pain); and 2) the association of receipt of a physician recommendation to exercise with meeting PA guidelines (aerobic and strengthening) and arthritis symptoms (arthritis-associated activity limitations and arthritis-related pain), and whether race/ethnicity moderates these associations.
Methods
Study design and sample
This study included 27,877 adults; 64.3% were women and the mean age was 60.9 years (SD = 15.1) (Table
1). NHIS data was pooled 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 [
20]. 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.
Table 1
Characteristics of U.S. Adults Aged ≥ 18 years with Arthritis by Race/Ethnicity, National Health Interview Survey, 2002, 2006, 2009, and 2014, N = 27,887
Demographic characteristics |
Age (in years) |
18–44 | 4138 (14.8) | 2793 (13.8) | 674 (15.7) | 581 (20.7) | 90 (14.6) | <.0001 |
45–64 | 11,652 (41.8) | 8288 (41.1) | 1941 (45.3) | 1209 (43.1) | 214 (34.6) | |
≥ 65 | 12,097 (43.4) | 9100 (45.1) | 1669 (39.0) | 1014 (36.2) | 314 (50.8) | |
Sex |
Female | 17,922 (64.3) | 12,646 (62.7) | 2990 (69.8) | 1906 (68.0) | 380 (61.5) | <.0001 |
Male | 9965 (35.7) | 7535 (37.3) | 1294 (30.2) | 898 (32.0) | 238 (38.5) | |
Education |
< high school | 5769 (20.8) | 3061 (15.2) | 1294 (30.5) | 1276 (46.0) | 138 (22.5) | <.0001 |
high school graduate | 8236 (29.7) | 6215 (30.9) | 1242 (29.3) | 642 (23.1) | 137 (22.4) | |
Technical or some college | 8018 (28.9) | 6123 (30.5) | 1169 (27.5) | 592 (21.3) | 134 (21.9) | |
> College degree | 5692 (20.5) | 4685 (23.3) | 539 (12.7) | 265 (9.5) | 203 (33.2) | |
Health characteristics |
Smoking Status |
Never | 13,449 (48.6) | 9197 (45.9) | 2205 (52.0) | 1634 (58.6) | 413 (67.2) | <.0001 |
Former | 8889 (32.1) | 6941 (34.6) | 1083 (25.5) | 719 (25.8) | 146 (23.7) | |
Current | 5355 (19.3) | 3911 (19.5) | 954 (22.5) | 434 (15.6) | 56 (9.1) | |
Body mass index |
< 18.5, underweight | 395 (1.5) | 305 (1.6) | 47 (1.1) | 21 (0.8) | 22 (3.6) | <.0001 |
18.5–24.99, normal weight | 7248 (27.1) | 5631 (29.1) | 795 (19.4) | 527 (19.7) | 295 (48.8) | |
25–29.9, overweight | 9067 (34.0) | 6667 (34.5) | 1222 (29.8) | 974 (36.4) | 204 (33.8) | |
≥ 30, obese | 9989 (37.4) | 6716 (34.8) | 2036 (49.7) | 1154 (43.1) | 83 (13.7) | |
Self-rated health |
poor/fair | 8555 (30.7) | 5290 (26.2) | 1891 (44.2) | 1200 (42.8) | 174 (28.2) | <.0001 |
good/very good/excellent | 19,308 (69.3) | 14,873 (73.8) | 2390 (55.8) | 1602 (57.2) | 443 (71.8) | |
Number of co-morbidities1 |
0 | 6686 (24.6) | 5002 (25.4) | 756 (18.1) | 758 (27.9) | 170 (28.8) | <.0001 |
1–2 | 15,104 (55.5) | 10,843 (55.0) | 2456 (58.9) | 1469 (54.0) | 336 (56.9) | |
≥ 3 | 5419 (19.9) | 3883 (19.7) | 957 (23.0) | 494 (18.2) | 85 (14.4) | |
Psychological distress2 |
none to mild (0–4) | 19,407 (71.3) | 14,381 (72.9) | 2868 (68.6) | 1714 (62.9) | 444 (74.9) | <.0001 |
moderate (5–12) | 6037 (22.2) | 4182 (21.2) | 1000 (23.9) | 734 (26.9) | 121 (20.4) | |
severe (≥ 13) | 1774 (6.5) | 1156 (5.9) | 312 (7.5) | 278 (10.2) | 28 (4.7) | |
Outcomes |
Received clinician recommendation to exercise to help arthritis |
Yes | 15,799 (56.9) | 11,068 (55.1) | 2645 (62.0) | 1713 (61.3) | 373 (60.4) | <.0001 |
No | 11,982 (43.1) | 9031 (44.9) | 1623 (38.0) | 1083 (38.7) | 245 (39.6) | |
Severity of point Pain |
Mild (0–3) | 11,503 (41.7) | 8785 (44.0) | 1400 (33.3) | 1013 (36.6) | 305 (50.3) | <.0001 |
Moderate (4–7) | 10,817 (39.2) | 8107 (40.6) | 1494 (35.5) | 991 (35.8) | 225 (37.1) | |
Severe (8–10) | 5243 (19.0) | 3086 (15.4) | 1314 (31.2) | 767 (27.7) | 76 (12.5) | |
Arthritis-associated activity limitations |
Yes | 11,487 (41.3) | 7893 (39.2) | 2051 (48.0) | 1312 (46.9) | 231 (37.5) | <.0001 |
No | 16,354 (58.7) | 12,257 (60.8) | 2225 (52.0) | 1487 (53.1) | 385 (62.5) | |
Met guideline for aerobic physical activity3 |
Yes | 9151 (33.9) | 7124 (36.5) | 1092 (26.2) | 716 (26.1) | 219 (36.0) | <.0001 |
No | 17,879 (66.1) | 12,387 (63.5) | 3076 (73.8) | 2027 (73.9) | 389 (64.0) | |
Met guideline for strengthening physical activity 4 |
Yes | 4500 (16.3) | 3507 (17.5) | 572 (13.5) | 313 (11.2) | 108 (17.6) | <.0001 |
No | 23,118 (83.7) | 16,478 (82.5) | 3666 (86.5) | 2470 (88.8) | 504 (82.4) | |
Measures
Covariates
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 m
2) categorized as underweight (< 18.5), normal weight (18.5–24.9), overweight (25.0–29.9), or obese (
> 30); self-rated health (excellent/very good/good versus fair/poor); and 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) [
21,
23].
Statistical analysis
Descriptive statistics were used to examine the distributions of demographic, health characteristics, and outcomes by race/ethnicity. Chi-square tests were used to assess bivariate differences in demographic factors, health characteristics, and outcomes by race/ethnicity.
Multivariable logistic regression was used to analyze the association of race/ethnicity and receipt of physician recommendation on the odds of meeting aerobic and strengthening PA guidelines, and arthritis activity limitation). Additionally, we assessed whether the association of receiving physician recommendation to exercise with the odds of meeting aerobic and strengthening PA guidelines and reporting arthritis activity limitations were moderated by race/ethnicity (included an interaction term for race/ethnicity x receipt of physician recommendation in the model). Covariates included demographic characteristics (age, sex, education, marital status, income, employment, health insurance, having a usual source of care, and region) and health characteristics (smoking status, BMI, self-rated health, number of comorbidities, and psychological distress).
Multivariable linear regression was used to analyze the association of race/ethnicity and receipt of physician recommendation on joint pain severity, and whether the association physician recommendation with pain severity was moderated by race/ethnicity (interaction term for race/ethnicity x receipt of physician recommendation was included in the model). Covariates included demographic characteristics (age, sex, education, marital status, income, employment, health insurance, having a usual source of care, and region) and health characteristics (smoking status, BMI, self-reported health, comorbidities, psychological distress).
With approximately 13–15% missing response rates on our outcome variables we used weighted hotdeck imputation to account for these missing responses while also accounting for study weights. Using weighted hotdeck imputation with 20 donors, SAS proc. MIAnalyze was used to account for the between-imputation variance.
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.
Discussion
In this study, among adults with arthritis, we examined: 1) the association of participant race/ethnicity with meeting PA guidelines and arthritis symptoms; and 2) the association of receipt of a physician recommendation to exercise with meeting PA guidelines and arthritis symptoms, and whether race/ethnicity moderates these associations. Overall, we found that almost 60% of respondents reported at least moderate joint pain and over 40% reported activity limitations, while only 34 and 16% met aerobic and strengthening activity guidelines. Over 40% indicated they had not received a physician recommendation for exercise to relieve symptoms. Controlling for demographic and health characteristics, we found no racial/ethnic differences in the likelihood of meeting aerobic PA guidelines. Controlling for demographic and health factors, African Americans were more likely, and Asians were less likely to meet strengthening guidelines compared to Whites. In adjusted models, Latinos and African Americans were more likely and Asians were less likely than Whites to report severe pain, and there were no differences by race/ethnicity in reporting activity limitations. Receipt of physician recommendation was independently and positively associated with meeting aerobic and strengthening guidelines, having arthritis-associated activity limitations, and more severe joint pain; these associations of receipt of physician recommendation were ubiquitous across all race/ethnic groups except for a weak negative association observed among Latinos for meeting strengthening guidelines.
The proportions of those with arthritis who met aerobic and strengthening guidelines in our study are similar to those found in a previous study of the U.S. population ― approximately 36 and 18%, respectively, supporting the need for more attention to increasing PA in persons with arthritis to relieve symptoms [
14]. In our study, disparities among African Americans and Latinos in meeting aerobic PA guidelines observed in bivariate analyses were explained by demographic and health-related factors in the multivariable models. Although in unadjusted analyses, Asians were just as likely as Whites to meet aerobic and strengthening PA guidelines, disparities among Asians in meeting strengthening guidelines emerged after adjusting for demographic and health-related covariates. African Americans were more likely than Whites to meet muscle strengthening guidelines after accounting for demographic and health factors. Unmeasured cultural preferences for weight lifting/strengthening exercises or avoiding over-exertion might help explain these differences among African Americans [
24,
25].
In unadjusted analyses, African Americans and Latinos were more likely than Whites to report activity limitations and severe joint pain (and in adjusted analysis). Although African Americans and Latinos were more likely to report arthritis-associated activity limitations in unadjusted models, these differences were attenuated when controlling for demographic and health factors. African Americans and Latinos reported more, and Asians less severe pain than Whites, before and after adjusting for other factors, suggesting that factors other than health and demographic characteristics explain these racial/ethnic differences in levels of pain. Our results agree with prior literature on greater pain among African Americans and Latinos but differ in that we found no racial/ethnic differences on activity limitations [
26]. However, our analyses controlled for demographic and health characteristics while the referenced study only included demographics.
Smoking and obesity were negatively associated with meeting aerobic and strengthening guidelines and positively associated with activity limitations and severity of joint pain, in adjusted models. Physician counseling regarding these risk factors is important it has can promote a healthier lifestyle and improved quality of life [
27,
28]. Additionally, more severe psychological distress was positively associated with greater arthritis symptom burden (activity limitations and greater pain severity), in adjusted models. Because the relationship between distress and chronic pain is well-established [
29] and may be stronger among African Americans and Latinos, [
30] screening patients with chronic arthritis-related pain for psychological distress may be indicated.
Our findings are consistent with a 2013 study showing that receipt of physician recommendation to exercise was independently associated with meeting PA guidelines, corroborating the importance of physician advice [
16]. Although the proportion of adults with arthritis receiving a physician recommendation to exercise has improved from 28% between 1999 and 2000 [
31] to nearly 60%, [
16,
32] there is still considerable room for improvement. In our study, in unadjusted analyses, a greater proportion of African Americans, Latinos, and Asians received a physician recommendation to exercise, compared to Whites. In adjusted analyses, both reporting activity limitations and severe pain were associated positively with receipt of an exercise recommendation, which could explain the higher rates of physician recommendation in patients where the need is greatest.
As a retrospective cross-sectional study, this study is limited in that directionality and causality cannot be determined. This study uses secondary data from the NHIS which limits our ability to ask specific questions about cultural preferences and physician-patient interactions that may help explain our findings. Also, the survey item on which the case definition for arthritis was based included gout, lupus, or fibromyalgia in the item stem. Although the prevalence of these conditions is much lower than that of arthritis, we were unable to identify and drop cases of gout, lupus or fibromyalgia from our estimates. Further, our arthritis population had a 14% missing response rate on study outcomes, and it is possible that imputation may have introduced some selection bias.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.