SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Objective

To examine arthritis impact among US adults with self-reported doctor-diagnosed arthritis using the International Classification of Functioning, Disability and Health (ICF) framework (including the impairments, activity limitations, environmental, and personal factors domains and social participation restriction [SPR] as the outcome) overall and among those with and without SPR, and to identify the correlates of SPR.

Methods

Cross-sectional 2009 National Health Interview Survey data were analyzed to examine the distribution of the ICF domain components. Unadjusted and multivariable-adjusted prevalence ratios (PRs) and 95% confidence intervals (95% CIs) were estimated to identify the correlates of SPR. Analyses using SAS, version 9.2 survey procedures accounted for the complex sample design.

Results

SPR prevalence was 11% of adults with arthritis (5.7 million). After initial multivariable adjustment by ICF domain, serious psychological distress (impairments domain; PR 2.5 [95% CI 2.0–3.2]), ≥5 medical office visits (environmental domain; PR 3.4 [95% CI 2.5–4.4]), and physical inactivity (personal domain; PR 4.8 [95% CI 3.6–6.4]) were most strongly associated with SPR. A combined measure (key limitations [walking, standing, or carrying]; PR 31.2 [95% CI 22.3–43.5]) represented the activity limitations domain. After final multivariable adjustment incorporating all ICF domains simultaneously, the strongest associations with SPR were key limitations (PR 24.3 [95% CI 16.8–35.1]), ≥9 hours of sleep (PR 1.6 [95% CI 1.3–2.0]), and income-to-poverty ratio <2.00 and severe joint pain (PR 1.4 [95% CI 1.2–1.6] for both).

Conclusion

SPR affects 1 of 9 adults with arthritis. This study is the first to use the ICF framework in a population-based sample to identify specific functional activities, pain, sleep, and other areas as priorities for intervention to reduce negative arthritis impacts on disability, including SPR. Increased use of existing clinical and public health interventions is warranted.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Arthritis is common (affecting 50 million adults in the US [1]), costly (exceeding $128 billion annually [2]), and has been the most common cause of disability among US adults for more than 15 years ([3]). Despite this staggering impact, the complex process through which arthritis leads to disability is not fully understood ([4]). The International Classification of Functioning, Disability and Health (ICF) is a system developed by the World Health Organization (WHO) to “provide a unified and standard language and framework for the description of health and health- related states” ([5]). This approach considers the anatomic/ physiologic (impairment), individual (activity limitation), and societal (participation restriction) consequences of health conditions in the context of personal and environmental factors and reflects a continuum between positive functioning (functional and structural integrity) and negative functioning (impairments, activity limitations, and participation restriction), with the negative aspects indicating disability. At the societal level, participation is “involvement in a life situation,” while participation restriction is “problems an individual may experience in involvement in life situations” and reflects difficulties in activities such as running errands, visiting friends, and leisure activities ([5]). The WHO has proposed using the ICF to investigate the consequences of health conditions; the comprehensive scope makes it ideal for studying arthritis ([5, 6]).

Although the ICF has been mapped to different clinical outcome measures for arthritis ([7-9]) and ICF core sets have been developed for osteoarthritis, rheumatoid arthritis, and other chronic musculoskeletal conditions ([10]), no population-based studies have applied an inclusive view of all ICF domains to assess participation restriction in adults with arthritis ([11]). While population-based and clinical studies have often focused on the disease impacts considered activity limitations or impairments in the ICF, examining participation restriction using all of the ICF domains, including the contextual personal and environmental factors, provides an opportunity to explore a more comprehensive approach to assessing and describing one aspect of disability among adults with arthritis.

Recently, there has been growing interest in participation restriction ([12-17]) partly due to the recognition that the social consequences of musculoskeletal conditions (e.g., difficulty shopping or visiting relatives) may be of greater concern to individuals than impairments (e.g., pain) or specific activity limitations (e.g., walking greater than a half mile). Participation is an important outcome for examining the effectiveness of clinical (e.g., joint replacements) and public health (e.g., psychoeducational courses) interventions, even when participation may not be the target; the secondary benefits of these interventions may be improvements in the ability to run errands, look after friends and family, work, and be a part of the community. Importantly, even in the presence of ongoing signs (radiographic change), symptoms (pain), and activity limitation (walking limitation), participation can be maintained ([14]).

Despite the increased attention paid to participation, to our knowledge there are no ICF-based studies of participation restriction among the US adult population with arthritis ([11]). Furthermore, in most existing studies, the conceptualization of participation has been limited (e.g., measurement specification [17], hypothesis testing [16], and psychosocial aspects of role value and performance [15]). The purpose of this study was to examine arthritis impact among US adults with self-reported doctor- diagnosed arthritis using the entire ICF framework 1) overall and among those with and without social participation restriction (SPR), and 2) to identify the correlates of SPR.

Box 1. Significance & Innovations

  • This is the first nationally representative application of the International Classification of Functioning, Disability and Health among US adults with arthritis.
  • An estimated 5.7 million US adults with arthritis report having social participation restriction.
  • We observed a novel association between social participation restriction and sleep among adults with arthritis.
  • An income-to-poverty ratio of <2.00 and other measures of assets and public assistance suggest financial resources may play a key role in the process of arthritis disability.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Study sample

The National Health Interview Survey (NHIS) is an annual health survey including people of all ages. The survey uses a complex sample design to select a sample representing the US civilian noninstitutionalized population ([18]). Data are collected through in-home interview conducted by trained interviewers. We studied adults (ages ≥18 years) who in 2009 (the most recent year with all relevant variables) had records in the sample adult core, family, household, and person files (n = 27,731); the conditional and final response rates in the sample adult core were 80.1% and 65.4%, respectively ([18]).

Doctor-diagnosed arthritis

The sample was limited to people with self-reported doctor-diagnosed arthritis (hereafter referred to as arthritis; n = 6,696). Arthritis was identified by answering “yes” to “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?”

SPR

Respondents who answered “very difficult” or “can't do at all” to either of the following were classified as having social participation restriction: “By yourself, and without using any special equipment, how difficult is it for you to…” and ending with “… go out to things like shopping, movies, or sporting events,” and “… participate in social activities such as visiting friends, attending clubs and meetings, going to parties?”

Prior to analysis, we reviewed the NHIS documentation to identify all variables that could be coded to ICF domains. We used the following 3 criteria to select variables for analysis: 1) conceptual relevance, 2) sufficient sample size (≥50 cases), and/or 3) meaningful prevalence in the population (≥5%). Several variables (e.g., housing density, excess alcohol consumption, and US census region) were initially considered but failed to meet these criteria. All of the analyzed variables are shown in Figure 1. The detailed ICF codes for each measure are provided in Supplementary Appendix A (available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21977/abstract), with the exception of variables representing personal factors (e.g., age, race/ethnicity) because, per the WHO, personal factors are not assigned codes ([5]).

image

Figure 1. International Classification of Functioning, Disability and Health domains with codes and National Health Interview Survey variables used in the analysis. Adapted, with permission, from ref.[5].

Download figure to PowerPoint

Impairments

The respondents rated the severity of their joint pain in the past 30 days on a scale ranging from 0 = none to 10 = as bad as it can be; those with ratings ≥7 were classified as having severe joint pain ([19, 20]). Body mass index (BMI; weight in kg/height in m2) was calculated from self-reported weight and height. The respondents were asked whether they have ever been diagnosed with a series of chronic conditions. We examined hypertension, heart disease (coronary heart disease, angina, myocardial infarction, and other heart disease), stroke, emphysema, asthma, cancer, diabetes mellitus, chronic bronchitis, weak/failing kidneys, and liver disease. The responses were scored (yes = 1 and no = 0), added (range 0–10), and categorized to represent the number of selected comorbid conditions (0, 1–2, or ≥3). Serious psychological distress was measured using the Kessler 6 (K6), a scale developed to identify and monitor population-level prevalence and trends of nonspecific serious psychological distress ([21]). The K6, consisting of 6 questions rated on a scale ranging from 0 = none of the time to 4 = all of the time, asks how often in the past 30 days the respondent felt each of sad, worthless, nervous, restless, hopeless, that everything was an effort. The values were added for a total score (range 0–24); scores ≥13 identified respondents with serious psychological distress ([21]). In an article examining the relationship of sleep to several health outcomes, Knutson commented on the U-shaped curve observed in self-reported and clinically measured sleep and noted that, across studies included, “short sleep” is generally <6 hours per night while “long sleep” is generally >8 hours ([22]). Therefore, we categorized the average number of sleep hours in a 24-hour period in this study as 1–5, 6–8, and ≥9 hours.

Activity limitation

Respondents reported their ability, using a 5-point scale (not at all difficult; only a little difficult; somewhat difficult; very difficult; can't do at all; do not do this activity), to do the following 9 specific activities critical to independent functioning: “By yourself, and without using any special equipment, how difficult is it for you to …” and the endings are “Lift or carry something as heavy as 10 pounds such as a full bag of groceries,” “Walk (climb) up 10 steps without resting,” “Use your fingers to grasp or handle small objects,” “Push or pull large objects like a living room chair,” “Reach up over your head,” “Sit for about 2 hours,” “Stand or be on your feet for about 2 hours,” “Stoop, bend or kneel,” and “Walk a quarter mile or 3 city blocks.” Each limitation was coded as a 2-level variable, with those indicating “very difficult” or “can't do at all” classified as answering “yes” and all others as “not limited.” Very few respondents answered “do not do this activity” (range n = 14 [0.2%] grasp to n = 432 [6.5%] push); because it was not possible to determine whether these individuals chose not to perform these activities or were unable to do them, the most conservative approach was to include these respondents in the “not limited” group. For programmatic and surveillance purposes, 2 summary limitation variables were created reflecting ≥1 limitations (versus none) and ≥3 limitations (versus 0–2).

Anticipating high multicollinearity among the above activity limitations, we created a combination variable for the regression analyses. The 2 authors with clinical backgrounds (RW [physiotherapy] and JMH [sports medicine/athletic training]) nominated each of lower extremity, upper extremity, mobility, and endurance as the most important capacities to include in a combined variable. These capacities were represented by 3 variables: walk (lower extremity mobility), carry (upper extremity mobility and strength), and stand (overall endurance). The presence of key limitations was defined as a response of “very difficult” or “can't do at all” for any of the 3 variables.

Contextual factors per the ICF

The ICF distinguishes contextual factors as environmental (i.e., external) or personal (i.e., internal) factors (not all categories of contextual factors [e.g., attitudes] were available in the NHIS). The environmental categories include one's immediate interpersonal and physical environment, use of services, and support and relationships. The personal factors include individual characteristics (e.g., age, gender, education, social background, past experience, and coping style) ([5]).

Environmental factors

Respondents were queried about homeownership, household size (number of people in household), and marital status. The participants were also asked a series of questions about their number of medical office visits in the past year, if they had health insurance, or if they had delayed health care because of cost in the past year. Receipt of government assistance was indicated by a “yes” response to receiving ≥1 type of government assistance in the previous calendar year (i.e., welfare, food stamps, and other assistance). The NHIS provides a calculated variable of family income divided by the relevant poverty threshold (poverty thresholds account for family size and are therefore more informative than income alone) ([18, 23]). Depth of poverty can be assessed by the income-to-poverty ratio. Although the US Census Bureau does not use this term, ratios >1.00 and ≤1.99 are often cited as near poverty ([23]). For the present study, the income-to-poverty ratio was categorized as <2.00 and ≥2.00. Finally, the most educated household adult was classified as less than high school; GED/high school graduate; some college, no college degree; and college degree or more.

Personal factors

The personal factors identified from the NHIS were 1) sociodemographic characteristics (age, sex, race/ethnicity, respondent's education level [defined as above], and employment status), 2) health behaviors (aerobic physical activity level per the 2008 criteria [meets physical activity guideline recommendations, insufficient, and inactive] [24], current smoker, and retrieve health information from the internet), and 3) whether the respondent had ever spent ≥24 hours homeless or in jail. Current smoker, retrieve health information from the internet, and ever spent ≥24 hours homeless or in jail were coded as 1 = yes and all others = no.

The number of missing values for the nondichotomous variables ranged from 1 (employment) to 250 (BMI; 0.0% to 3.7% of the sample, respectively), with the exception of the income-to-poverty ratio, which had 712 values missing (10.6% of the sample).

Statistical analysis

To address our first aim of describing the profile of people with arthritis using the ICF framework, we calculated distributions of all study variables across US adults with arthritis overall and among those with arthritis with and without SPR. The analyses were conducted using SAS, version 9.2 survey procedures ([25]); sampling weights were applied to derive nationally representative estimates, and the SE calculations accounted for the complex sample design. All of the reported estimates had a relative SE <20%.

To address our second aim of examining associations with SPR and identifying correlates, regression modeling proceeded in 4 steps. First, independent associations between each variable and SPR were estimated with unadjusted prevalence ratios (PRs) and 95% confidence intervals (95% CIs).

Second, we identified eligible variables for the multivariable-adjusted models by examining 1) the potential modifiability (possibly amenable to intervention), 2) known relationship with SPR from other studies, 3) statistical significance in the univariable regression stage, and 4) if there was a moderate to strong association (defined as a Spearman's correlation coefficient ≥0.4). For example, if the Spearman's correlation coefficient between 2 variables was ≥0.4, the variable that was not modifiable or had no known relationship to SPR was excluded. Statistical significance in univariable regression was defined as having CIs that did not include 1.0; statistically significant differences in PRs were defined as nonoverlapping 95% CIs. This is a more conservative approach than significance testing and more appropriate for our study given the large sample size, number of variables, and multiple comparisons ([26]).

Third, once the candidate variables for each ICF domain were identified, the entire group of variables in each domain was analyzed simultaneously. We ran a series of multivariable models for each domain and tested for potential collinearity using the condition index (SAS PROC REG). For each model, we identified the variable with the largest individual condition index ≥30.0 ([27]). Then, the model was rerun excluding this variable with the final model including variables with a condition index of <30 only.

Fourth, once the multivariable model for each ICF domain was finalized, a multivariable “metamodel” incorporating all ICF domains was created using the same process described in step 3. This multivariable metamodel was created to identify the strongest statistically significant independent associations with SPR when variables across all ICF domains were examined simultaneously.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

Study population

Detailed characteristics of the study population (US adults ages ≥18 years with arthritis) are shown in Table 1 and Supplementary Appendix B (available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21977/abstract). The prevalence of SPR was 11% of adults with arthritis (5.7 million). Compared with adults without SPR, adults with SPR reported >5 times the prevalence of serious psychological distress (22.6% versus 4.1%) and 3–10 times the prevalence of activity limitations. Nearly all respondents with SPR (96.9%) reported ≥1 limitations and 85.5% reported ≥3 limitations, compared with 35.8% and 15.7% reporting ≥1 and ≥3 limitations, respectively, for those without SPR. Those with SPR reported having double the prevalence of an income-to-poverty ratio <2.00 than those without SPR (56.3% versus 27.5%). Most respondents with SPR were non-Hispanic whites (70.5%) and women (68.2%); 46.1% were age ≥65 years. Respondents with SPR tended to have a low education level (34.2% less than high school and 31.3% GED or high school graduate). Employment was the same across groups.

Table 1. Distribution of ICF characteristics by domain among US adults ≥18 years of age with self-reported doctor-diagnosed arthritis, National Health Interview Survey, 2009*
 All adults with arthritis (n = 52,106,717)Adults with arthritis but not SPR (n = 46,382,370)Adults with arthritis and SPR (n = 5,724,347)
Weighted no., thousands% (95% CI)Weighted no., thousands% (95% CI)Weighted no., thousands% (95% CI)
  1. Numbers may not add up to 100.0 because of rounding. ICF = International Classification of Functioning, Disability and Health; SPR = social participation restriction; 95% CI = 95% confidence interval; BMI = body mass index; GED = general educational development.

  2. a

    Defined as “very difficult” or “can't do at all.”

SPR5,72411.0 (10.0–12.0)00.05,724100.0
Impairments      
Severe joint pain, score ≥7 of 1014,11627.1 (25.6–28.6)10,94023.6 (22.1–25.1)3,17655.5 (51.9–59.1)
BMI, kg/m2      
Under- or normal weight (<25.0)13,71527.4 (26.0–28.7)12,12827.1 (25.7–28.5)1,58829.4 (25.9–33.0)
Overweight (25.0–29.9)17,24734.4 (33.0–35.8)15,69635.1 (33.5–36.6)1,55128.8 (25.4–32.1)
Obese (≥30.0)19,18338.3 (36.8–39.8)16,92837.8 (36.2–39.4)2,25541.8 (38.1–45.6)
No. of selected comorbid conditions      
012,41223.8 (22.5–25.1)11,82025.5 (24.1–26.9)59210.3 (8.5–12.1)
1–227,56952.9 (51.4–54.4)24,89153.7 (52.9–55.3)2,67846.8 (43.1–50.5)
3–1012,12623.3 (21.9–24.6)9,67120.9 (19.4–22.3)2,45442.9 (39.0–46.8)
Serious psychological distress8,50116.3 (15.1–17.5)1,9004.1 (3.4–4.8)1,29222.6 (19.2–25.9)
Sleep in 24-hour period, hours      
1–56,51812.7 (11.7–13.7)5,34311.6 (10.6–12.7)1,17421.5 (18.2–24.7)
6–838,78775.4 (74.1–76.7)35,75277.8 (76.4–79.2)3,03555.4 (51.5–59.4)
≥96,13911.9 (10.9–13.0)4,87410.6 (9.5–11.7)1,26523.1 (20.2–26.0)
Activity limitationsa      
Carry/lift something as heavy as 10 pounds6,46512.4 (11.5–13.4)2,9996.5 (5.7–7.2)3,46660.6 (57.0–64.1)
Climb up 10 steps without resting7,98515.3 (14.3–16.3)4,2989.3 (8.4–10.2)3,68764.4 (61.1–67.7)
Grasp/handle small objects with fingers2,8485.5 (4.8–6.1)1,5303.3 (2.8–3.8)1,31723.0 (19.6–26.4)
Push/pull large objects (e.g., living room chair)9,48918.2 (17.0–19.5)5,51911.9 (10.8–13.0)3,97069.4 (66.0–72.7)
Reach up over head3,6677.0 (6.3–7.8)1,9284.2 (3.5–4.8)1,73930.4 (27.0–33.8)
Sit for about 2 hours5,1159.8 (8.9–10.8)3,2287.0 (6.1–7.8)1,88733.0 (29.6–36.3)
Stand or be on feet for about 2 hours14,22627.3 (25.8–28.8)9,41620.3 (18.8–21.8)4,80984.0 (81.5–86.5)
Stoop, bend, or kneel14,15027.2 (25.7–28.6)9,75521.0 (19.7–22.4)4,39576.8 (73.6–79.9)
Walk a quarter mile or 3 city blocks10,93321.0 (19.7–22.3)6,44113.9 (12.7–15.1)4,49278.5 (75.5–81.5)
Summary limitation variables      
≥1 limitations22,16042.5 (40.9–44.1)16,61435.8 (34.2–37.5)5,54696.9 (95.4–98.3)
≥3 limitations12,19023.4 (22.0–24.8)7,29715.7 (14.5–17.0)4,89385.5 (82.8–88.2)
Key limitations (walk/stand/carry)17,02032.7 (31.1–34.2)11,65025.1 (23.6–26.7)5,36993.8 (92.1–95.5)
Environmental factors      
Homeownership      
Own/being bought39,58776.0 (74.6–77.5)35,93477.5 (76.0–79.0)3,65363.9 (60.8–67.1)
Rent/other12,48524.0 (22.5–25.4)10,42222.5 (21.0–24.0)2,06336.1 (32.9–39.2)
Household size (no. of people in household)      
112,30823.6 (22.5–24.8)10,54122.7 (21.5–23.9)1,76730.9 (28.0–33.7)
222,73243.6 (42.1–45.2)20,56344.3 (42.7–46.0)2,16937.9 (34.5–41.3)
≥317,06632.7 (31.1–34.4)15,27832.9 (31.2–34.7)1,78831.2 (27.6–34.9)
Marital status      
Married/living with partner32,83763.1 (61.6–64.6)30,11765.0 (63.5–66.5)2,72047.5 (43.6–51.4)
Divorced/separated/widowed14,87828.6 (27.2–29.9)12,42926.8 (25.5–28.2)2,44942.8 (39.3–46.3)
Never married4,3368.3 (7.5–9.1)3,7818.2 (7.3–9.0)5559.7 (7.8–11.6)
Medical office visits in past year      
0–17,45914.5 (13.4–15.6)7,05515.4 (14.2–16.6)4047.2 (5.6–8.9)
2–426,68451.9 (50.4–53.5)24,59253.7 (52.0–55.4)2,09237.5 (34.6–40.5)
≥517,25233.6 (32.0–35.1)14,17530.9 (29.3–32.6)3,07755.2 (51.9–58.6)
Health insurance, no4,4988.6 (7.7–9.6)4,0078.7 (7.7–9.7)4908.6 (6.7–10.5)
Delayed health care due to cost in past year, yes8,07415.5 (14.5–16.5)6,81514.7 (13.6–15.8)1,25922.0 (19.0–25.0)
Received government assistance in past calendar year, yes4,5318.7 (7.9–9.5)3,2687.0 (6.2–7.9)1,26322.1 (19.0–25.2)
Income-to-poverty ratio      
<2.00 (below or near poverty)14,35830.6 (29.1–32.2)11,47627.5 (25.9–29.1)2,88256.3 (52.9–59.7)
≥2.0032,50469.3 (67.8–70.9)30,26672.5 (70.9–74.1)2,23843.7 (40.3–47.1)
Most educated adult in household      
Less than high school5,18910.0 (9.1–10.8)4,0308.7 (7.9–9.5)1,15920.3 (17.4–23.2)
GED or graduated high school13,30225.6 (24.2–27.0)11,57425.0 (23.6–26.4)1,72830.3 (27.2–33.4)
Some college, no degree10,89320.9 (19.6–22.3)9,70821.0 (19.6–22.4)1,18620.8 (17.9–23.7)
College degree or more22,63743.5 (41.9–45.1)20,99945.3 (43.7–47.0)1,63728.7 (25.6–31.8)
Personal factors      
Age, years      
18–448,96317.2 (15.9–18.5)8,16417.6 (16.2–19.0)79914.0 (11.4–16.6)
45–6423,84445.8 (44.0–47.5)21,55946.5 (44.6–48.4)2,28639.9 (36.3–43.5)
≥6519,29937.0 (35.3–38.7)16,66035.9 (34.1–37.7)2,63946.1 (43.0–49.2)
Sex      
Male20,77539.9 (38.5–41.3)18,95740.9 (39.4–42.4)1,81831.8 (28.4–35.1)
Female31,33260.1 (58.7–61.5)27,42559.1 (57.6–60.6)3,90768.2 (64.9–71.6)
Race/ethnicity      
Non-Hispanic white40,54977.8 (76.4–79.2)36,51378.7 (77.3–80.2)4,03670.5 (67.5–73.5)
Non-Hispanic black5,72811.0 (9.9–12.0)4,86510.5 (9.4–11.5)86315.1 (12.8–17.3)
Hispanic3,7937.3 (6.5–8.0)3,3067.1 (6.3–8.0)4878.5 (6.9–10.1)
Non-Hispanic other2,0373.9 (3.3–4.5)1,6983.7 (3.0–4.3)3405.9 (4.1–7.7)
Education      
Less than high school8,72016.8 (15.7–18.0)6,77514.7 (13.5–15.8)1,94534.2 (31.2–37.3)
GED or graduated high school16,37631.6 (30.1–33.0)14,59431.6 (30.1–33.1)1,78231.3 (28.3–34.4)
Some college, no degree10,53020.3 (19.0–21.6)9,59120.8 (19.4–22.2)93816.5 (13.7–19.3)
College degree or more16,22731.3 (29.7–32.9)15,20632.9 (31.3–34.6)1,02218.0 (15.1–20.8)
Employment status      
Working27,80253.4 (51.7–55.0)24,74853.4 (51.6–55.2)3,05453.4 (49.8–56.9)
Not working24,30046.6 (45.0–48.3)21,63146.6 (44.8–48.4)2,67046.6 (43.1–50.2)
Current smoker, yes10,86820.9 (19.5–22.2)9,30120.1 (18.6–21.5)1,56727.4 (23.6–31.2)
Aerobic physical activity level      
Meets recommendations18,45136.2 (34.5–38.0)17,85539.4 (37.6–41.3)59610.6 (8.5–12.7)
Insufficient11,26122.1 (20.7–23.5)10,42023.0 (21.5–24.5)84115.0 (12.4–17.5)
Inactive21,19541.6 (39.9–43.4)17,02237.6 (35.8–39.4)4,17474.4 (71.0–77.8)
≥24 hours homeless or in jail, yes3,5296.8 (5.9–7.6)2,8456.1 (5.3–7.0)68412.0 (9.1–14.8)
Retrieve health information from internet, yes24,58847.2 (45.6–48.8)23,15349.9 (48.2–51.6)1,43425.1 (21.9–28.2)

PRs, unadjusted

A detailed discussion of the unadjusted associations is included in Supplementary Appendix B (available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21977/abstract).

Correlates of SPR, domain-specific multivariable models (Table 2).

Impairments

After multivariable adjustment, serious psychological distress continued to be the impairment most strongly associated with SPR (PR 2.5). There was ≥50% increased probability of SPR among those with severe joint pain, ≥3 selected comorbid conditions, and ≥9 hours of sleep.

Table 2. Unadjusted and adjusted associations of ICF characteristics by domain with SPR among US adults ≥18 years of age with self-reported doctor-diagnosed arthritis, National Health Interview Survey, 2009*
 Unadjusted associations with SPR, PR (95% CI)Multivariate ICF domain-specific models, PR (95% CI)Metamultivariate model, PR (95% CI)
  1. The multivariate ICF domain-specific models column shows the associations for each domain examined independently. The metamultivariate model column shows the associations for all ICF domains examined simultaneously in the same model. ICF = International Classification of Functioning, Disability and Health; SPR = social participation restriction; PR = prevalence ratio; 95% CI = 95% confidence interval; BMI = body mass index; GED = general educational development.

  2. a

    Defined as “very difficult” or “can't do at all.”

  3. b

    10.6% missing.

Impairments   
Severe joint pain, score ≥7 of 10 (ref. = no)3.4 (2.9–4.0)1.7 (1.4–2.0)1.4 (1.2–1.6)
BMI, kg/m2   
Underweight/normal (<25.0)1.0
Overweight (25.0–29.9)0.8 (0.6–1.0)
Obese (≥30.0)1.0 (0.8–1.2)
No. of selected comorbid conditions   
01.01.0
1–22.0 (1.6–2.7)1.4 (1.1–1.7)
3–104.2 (3.2–5.6)1.5 (1.2–1.9)
Serious psychological distress (ref. = no)4.5 (3.7–5.4)2.5 (2.0–3.2)
Sleep in 24-hour period, hours   
1–52.3 (1.9–2.9)1.3 (1.0–1.6)1.3 (1.1–1.6)
6–81.01.01.0
≥92.6 (2.1–3.2)1.5 (1.3–1.9)1.6 (1.3–2.0)
Limitations (ref. = not limited)a   
Carry/lift something as heavy as 10 pounds10.8 (9.2–12.7)
Climb up 10 steps without resting10.0 (8.5–11.8)
Grasp/handle small objects with fingers5.2 (4.4–6.1)
Push/pull large objects (e.g., living room chair)10.2 (8.5–12.2)
Reach up over head5.8 (4.9–6.8)
Sit for about 2 hours4.5 (3.9–5.3)
Stand or be on feet for about 2 hours14.0 (11.1–17.7)
Stoop, bend, or kneel8.9 (7.3–10.8)
Walk quarter mile or 3 city blocks13.7 (11.1–17.0)
Summary limitation variables   
≥1 limitations (ref. = no)41.9 (25.5–68.8)
≥3 limitations (ref. = no)19.3 (15.0–24.9)
Key limitation (walk/stand/carry; ref = no)31.2 (22.3–43.5)31.2 (22.3–43.5)24.3 (16.8–35.1)
Environmental factors   
Homeownership   
Own/being bought1.0
Rent/other1.8 (1.5–2.1)
Household size (no. of people in household)   
11.0
20.7 (0.6–0.8)
≥30.7 (0.6–0.9)
Marital status   
Married/living with partner1.01.01.0
Divorced/separated/widowed2.0 (1.7–2.4)1.4 (1.2–1.7)1.1 (0.9–1.3)
Never married1.6 (1.2–2.0)1.2 (0.9–1.5)1.3 (1.0–1.6)
Medical office visits in past year   
0–11.01.0
2–41.5 (1.1–2.0)1.6 (1.2–2.2)
≥53.3 (2.4–4.5)3.4 (2.5–4.4)
Health insurance status   
Have health insurance1.0 (0.7–1.3)
No health insurance1.0
Delayed health care because of cost (ref. = no)1.5 (1.3–1.9)
Received government assistance in past calendar year (ref. = no)3.0 (2.5–3.6)
Income-to-poverty ratiob   
<2.00 (below or near poverty)2.9 (2.5–3.4)2.5 (2.1–3.0)1.4 (1.2–1.6)
≥2.001.01.01.0
Most educated household adult   
Less than high school1.7 (1.4–2.1)
GED or graduated high school1.0
Some college, no degree0.8 (0.7–1.0)
College degree or more0.6 (0.5–0.7)
Personal factors   
Age, years   
18–441.01.0
45–641.1 (0.8–1.4)1.0 (0.8–1.3)
≥651.5 (1.2–2.0)1.2 (1.0–1.6)
Sex   
Male1.01.01.0
Female1.4 (1.2–1.7)1.3 (1.1–1.6)1.1 (1.0–1.3)
Race/ethnicity   
Non-Hispanic white1.01.01.0
Non-Hispanic black1.5 (1.2–1.8)1.2 (1.0–1.5)0.9 (0.7–1.1)
Hispanic1.3 (1.0–1.7)1.0 (0.7–1.3)1.1 (0.9–1.4)
Non-Hispanic other1.7 (1.2–2.4)1.7 (1.2–2.4)1.4 (1.0–1.9)
Education   
Less than high school2.1 (1.7–2.5)1.8 (1.5–2.1)
GED or graduated high school1.01.0
Some college, no degree0.8 (0.6–1.1)1.0 (0.8–1.3)
College degree or more0.6 (0.5–0.8)0.9 (0.7–1.1)
Employment status   
Working1.0
Not working1.0 (0.9–1.2)
Current smoker (ref. = no)1.4 (1.2–1.7)1.3 (1.1–1.6)1.2 (1.0–1.4)
Aerobic physical activity level   
Recommended1.01.0
Insufficient2.3 (1.6–3.3)2.0 (1.5–2.9)
Inactive6.1 (4.6–8.1)4.8 (3.6–6.4)
≥24 hours homeless or in jail (ref. = no)1.9 (1.5–2.4)
Retrieve health information from the internet (ref. = no)0.4 (0.3–0.5)
Limitations

Those with key limitations were 31 times more likely to report SPR compared with those without key limitations (PR 31.2 [95% CI 22.3–43.5]).

Environmental factors

The highest frequency of office visits in the past year (≥5) and <2.00 income-to-poverty ratio were the strongest correlates of SPR in the multivariate environmental model; each was associated with a greater than double increase in the likelihood of SPR (PRs 3.4 and 2.5, respectively).

Personal factors

There were small increases in the likelihood of SPR for women and current smokers (PR 1.3 for both). Non-Hispanic others and those with less than a high school education had at least a 70% greater likelihood of having SPR. The strongest associations with SPR were for low physical activity (insufficient and inactive [PRs 2.0 and 4.8, respectively]).

Multivariable correlates of SPR, metamodel

No personal domain variables remained significant in the metamodel (Table 2). Also, the strength of association between SPR and key limitations was attenuated in the metamodel, dropping to PR 24.3 (95% CI 16.8–35.1). Nevertheless, key limitations remained the single strongest correlate of SPR after adjustment, with the remaining significant PRs demonstrating associations between a 30% and 40% higher likelihood of SPR.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

By identifying the characteristics of adults with arthritis who are most likely to have SPR, researchers can further refine the development and targeting of interventions that enhance quality of life and decrease disability and health care costs. Our results provide the first population-based examination of arthritis disability in US adults using all ICF domains. Our approach extends the literature by using both ICF domain–specific multivariate models and a metamodel to demonstrate associations with SPR. The domain-specific models showed numerous potentially modifiable characteristics, while the metamodel results can be viewed as identifying priority areas with the strongest and possibly most important relationships requiring immediate resolution.

The findings from the domain-specific impairment multivariate model are consistent with findings from existing studies. For example, using the same measure of serious psychological distress as in our study, Okoro et al found that adults with disability and serious psychological distress were worse off than those with just self-reported disability ([28]). In our study, the domain-specific multivariate association of serious psychological distress with SPR was quite strong (PR 2.5), and, when coupled with evidence from existing studies ([28-31]), suggests that people with arthritis could benefit from more aggressive and targeted control of mental health symptoms. Although the negative impacts of mental health effects and physical disability appear to be cyclical ([32]), it is reasonable to attempt to “break the cycle” through existing and effective but underused interventions, such as pharmacologic and cognitive–behavioral therapy, self-management education, and aerobic exercise, for the depression and anxiety components of serious psychological distress among those with arthritis ([30]).

Among the variables that remained statistically significant in the metamodel, the key limitations variable was by far the most strongly associated with SPR (PR 24.3). This finding reiterates the importance of targeting the component activities (walking a quarter mile, standing for about 2 hours, and carrying something weighing about 10 pounds) for improved performance among people with arthritis. Both aerobic and muscle strengthening exercise programs have been shown to improve pain, functional performance measures (6-minute walk, timed up-and-go, chair stand, etc.), self-reported physical function (e.g., Health Assessment Questionnaire score), cardiorespiratory fitness (endurance), strength, and balance in randomized controlled and comparative effectiveness trials among adults with arthritis ([33-36]). Improvements in impairments and limitations via exercise may delay or reduce the risk of disability. For example, the Fitness Arthritis and Seniors Trial reported an approximately 43% reduced risk of incident activity of daily living disability 18 months after a structured aerobic and muscle strengthening intervention among older adults with osteoarthritis ([37]).

The persistent association of severe joint pain with SPR after all adjustments in the metamodel (PR 1.4) was expected and is consistent with existing studies examining joint pain in people with arthritis. For example, Wilkie et al found that the highest level of knee pain severity was strongly associated with restricted mobility outside of the home (adjusted odds ratio 2.4) ([13]). Hawker et al determined that unpredictable, intense, and emotionally draining pain “resulted in significant avoidance of social and recreational activities” ([38]). The effect of osteoarthritis pain on sleep onset and continuation was also associated with greater disability, fatigue, and mood disturbances ([38]). These findings lead one to question whether the participants in our study reporting ≥9 hours of sleep per 24 hours were actually sleeping that entire time and what the quality of their sleep was; a low quality of sleep may explain the association between the high number of sleep hours and SPR. Unfortunately, quality of sleep was not assessed in the NHIS.

While arthritis pain is complex, it is treatable. Over the counter medications (acetaminophen and nonsteroidal antiinflammatory medications); topical preparations (capsaicin); thermal modalities (hot and cold packs); aerobic, aquatic, and muscle strengthening exercises; weight loss; assistive devices (e.g., cane or crutch); orthotics/braces; and self-management education have all been shown to reduce osteoarthritis pain ([39, 40]). In cases where pain is not controlled using these first-line treatments, intraarticular corticosteroid injections, hyaluronate injections, duloxetine, and opioids can be used ([39, 40]). From this evidence, it seems clear that uncontrolled pain among people with arthritis is substantially damaging to their functioning and quality of life, and better control of joint pain could have positive cascading effects on sleep, mental health, and disability.

Existing studies have demonstrated that poor socioeconomic status is associated with poor health outcomes in general ([41, 42]), and for specific condition groups including arthritis ([43, 44]). Our study showed univariate (PR 2.9), domain-specific multivariate (PR 2.5), and meta-model (PR 1.4) associations between <2.00 income-to-poverty ratio and SPR. In addition to income-to-poverty ratio, 2 other measures (delayed health care due to cost [PR 1.5] and received government assistance [PR 3.0]) had strong univariate associations with SPR, suggesting that financial resources may have a key role in the process of arthritis disability. These findings may have important policy implications from both the perspectives of reducing and addressing disability among adults with arthritis ([45]).

This study has at least 4 limitations. First, doctor-diagnosed arthritis was self-reported and may have been subject to recall bias. However, this case-finding question is considered valid for public health surveillance ([46, 47]). Second, cross-sectional study data cannot be used to infer causation. Third, there were conceptual limitations in the variables available to measure some elements; for example, marital status and household size were also proxies for the broader concept of a social network. Similarly, as described in the introduction, social participation can be conceptualized in many ways, so our measure of SPR assesses only those aspects captured in the NHIS questions, which represent the “capacity” aspect of participation. The ICF defines “capacity” as what an individual can do in a standard environment without barriers or facilitators to participation and defines performance as what an individual can do in their usual environment with barriers (e.g., no sidewalks) and facilitators (e.g., walking aids). If the NHIS measured the performance of social participation, the proportion of individuals with SPR may have been lower ([48]). Fourth, the NHIS does not measure all ICF elements, so some important concepts were not included. In particular, there were no available variables on specific environmental characteristics (e.g., built environment features such as sidewalks, curbs, and transportation access) the modification of which, especially in conjunction with assistive mobility technologies, could be expected to influence SPR ([49]).

This study has several strengths. First, the NHIS is a unique and rich data source for examining ICF-based correlates of disability represented by SPR, including personal and environmental factors frequently absent from clinical studies. Second, the study had a sufficiently large sample to estimate precise moderate associations in the metamodel. Third, this study is also the first nationally representative application of the ICF among adults with arthritis and the findings are generalizable to US adults with arthritis. This study has addressed a gap by providing an inclusive and descriptive application of the ICF to arthritis in the US. Fourth, our findings can be used to develop applied research questions to explore arthritis impacts and relationships to improve our understanding of and ability to modify adverse arthritis outcomes, including SPR.

Social participation represents an important life domain for many people. Social activity has longitudinal associations with decreased risk of incident disability among community-dwelling older adults ([50]), and a growing number of studies have demonstrated the potentially protective effects of “having and retaining favorite pastimes” ([32]). Our study findings empirically demonstrate some of the complex relationships across the ICF domains and provide priority areas for clinical and public health interventions to decrease pain, address negative mental health effects, control arthritis symptoms, and create environments in which people with limitations or impairments are still able to participate.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms Theis had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Theis, Hootman, Wilkie.

Acquisition of data. Theis, Hootman.

Analysis and interpretation of data. Theis, Murphy, Hootman, Wilkie.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. REFERENCES
  • 1
    Cheng YJ, Hootman JM, Murphy LB, Langmaid GA, Helmick CG.Prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitation: United States, 2007–2009.MMWR Morb Mortal Wkly Rep2010;59:12615.
  • 2
    Yelin E, Murphy L, Cisternas MG, Foreman AJ, Pasta DJ, Helmick CG.Medical care expenditures and earnings losses among persons with arthritis and other rheumatic conditions in 2003, and comparisons with 1997.Arthritis Rheum2007;56:1397407.
  • 3
    Brault MW, Hootman J, Helmick CG, Theis KA, Armour BS.Prevalence and most common causes of disability among adults, United States, 2005.MMWR Morb Mortal Wkly Rep2009;58:4216.
  • 4
    Woolf A, Akesson K.Understanding the burden of musculoskeletal conditions: the burden is huge and not refelcted in national health priorities.BMJ2001;5:107980.
  • 5
    World Health Organization.International Classification of Functioning, Disability, and Health.Geneva:World Health Organization;2001.
  • 6
    Kostanjsek N.Use of the International Classification of Functioning, Disability and Health (ICF) as a conceptual framework and common language for disability statistics and health information systems.BMC Public Health2011;11 Suppl 4:S3.
  • 7
    Prodinger B, Cieza A, Williams DA, Mease P, Boonen A, Kerschan-Schindl K, et al.Measuring health in patients with fibromyalgia: content comparison of questionnaires based on the International Classification of Functioning, Disability and Health.Arthritis Rheum2008;59:6508.
  • 8
    Weigl M, Cieza A, Harder M, Geyh S, Amann E, Kostanjsek N, et al.Linking osteoarthritis-specific health-status measures to the International Classification of Functioning, Disability, and Health (ICF).Osteoarthritis Cartilage2003;11:51923.
  • 9
    Pollard B, Johnston M, Dieppe P.What do osteoarthritis health outcome instruments measure? Impairment, activity limitation, or participation restriction?J Rheumatol2006;33:75763.
  • 10
    Schwartzkopf SR, Ewert T, Dreinhofer KE, Cieza A, Stucki G.Towards an ICF core set for chronic musculoskeletal condtions: commonalities across ICF core sets for osteoarthritis, rheumatoid arthritis, osteoporosis, low back pain and chronic widespread pain.Clin Rheumatol2008;27:135561.
  • 11
    Weigl M, Cieza A, Cantista P, Reinhardt JD, Stucki G.Determinants of disability in chronic musculoskeletal health conditions: a literature review.Eur J Phys Rehabil Med2008;44:6779.
  • 12
    Wilkie R, Peat G, Thomas E, Croft P.Factors associated with participation restriction in community-dwelling adults aged 50 years and over.Qual Life Res2007;16:114756.
  • 13
    Wilkie R, Peat G, Thomas E, Croft P.Factors associated with restricted mobility outside the home in community-dwelling adults ages fifty years and older with knee pain: an example of use of the International Classification of Functioning to investigate participation restriction.Arthritis Rheum2007;57:13819.
  • 14
    Wilkie R, Thomas E, Mottram S, Peat G, Croft P.Onset and persistence of person-perceived participation restriction in older adults: a 3-year follow-up study in the general population.Health Qual Life Outcomes2008;6:92.
  • 15
    Gignac M, Backman C, Davis A, Lacaille D, Mattison C, Montie P, et al.Understanding social role participation: what matters to people with arthritis?J Rheumatol2008;35:165536.
  • 16
    Machado GP, Gignac MA, Badley EM.Participation restrictions among older adults with osteoarthritis: a mediated model of physical symptoms, activitiy limitations, and depression.Arthritis Rheum2008;59:12935.
  • 17
    Botha-Scheepers S, Watt I, Rosendall FR, Breedveld FC, Hellio le Graverand MP, Kloppenburg M.Changes in outcome measures for impairment, activity limitation, and participation restriction over two years in osteoarthritis of the lower extremities.Arthritis Rheum2008;59:17505.
  • 18
    National Center for Health Statistics.Data file documentation, National Health Interview Survey, 2009 (machine readable data file and documentation).Hyattsville (MD):National Center for Health Statistics, Centers for Disease Control and Prevention;2009.
  • 19
    Jensen M, Turner J, Romano J, Fisher L.Comparative reliability and validity of chronic pain intensity measures.Pain1999;83:15762.
  • 20
    Kapstad H, Hanestad B, Langeland N, Rustoen T, Stavern K.Cutpoints for mild, moderate and severe pain in patients with osteoarthritis of the hip or knee ready for joint replacement surgery.BMC Musculoskelet Disord2008;9:55.
  • 21
    Kessler RC, Andrews G, Colpe LJ, Hiripi E, Mroczek DK, Normand SL, et al.Short screening scales to monitor population prevalences and trends in non-specific psychological distress.Psychol Med2002;32:95976.
  • 22
    Knutson KL.Sleep duration and cardiometabolic risk: a review of the epidemiologic evidence.Best Pract Res Clin Endocrinol Metab2010;24:73143.
  • 23
    DeNavas-Walt C, Proctor B, Smith J.Current population reports, P60–238, income, poverty, and health insurance coverage in the United States: 2009.Washington, DC:US Government Printing Office;2010.
  • 24
    Department of Health and Human Services. 2008 physical activity guidelines for Americans. URL: http://www.health.gov/paguidelines/guidelines/default.aspx.
  • 25
    SAS Institute.SAS/STAT user's guide. Version 9.Cary (NC):SAS Institute;2004.
  • 26
    Schenker N, Gentleman J.On judging the significance of differences by examining the overlap between confidence intervals.Am Stat2001;55:1826.
  • 27
    Allison P.Logistic regression using the SAS system.Cary (NC):Wiley Interscience, SAS Institute;1991.
  • 28
    Okoro CA, Strine TW, Balluz L, Crews JE, Dhingra S, Berry JT, et al.Serious psychological distress among adults with and without disabilities.Int J Public Health2009;54:S5260.
  • 29
    Dunlop DD, Semanik P, Song J, Manheim LM, Shih V, Chang RW.Risk factors for functional decline in older adults with arthritis.Arthritis Rheum2005;52:127482.
  • 30
    Murphy LB, Sacks JJ, Brady TJ, Hootman JM, Chapman DP.Anxiety and depression among US adults with arthritis: prevalence and correlates.Arthritis Care Res (Hoboken)2012;64:96876.
  • 31
    Dunlop D, Manheim L, Song J, Lyons J, Chang R.Incident disability among preretirement adults: the impact of depression.Am J Public Health2005;95:20038.
  • 32
    Parmelee PA, Harralson TL, Smith LA, Schumacher HR.Necessary and discretionary activities in knee osteoarthritis: do they mediate the pain-depression relationship?Pain Med2007;8:44961.
  • 33
    Callahan LF, Mielenz T, Freburger J, Shreffler J, Hootman J, Brady T, et al.A randomized controlled trial of the People With Arthritis Can Exercise Program: symptoms, function, physical activity, and psychosocial outcomes.Arthritis Rheum2008;59:92101.
  • 34
    Callahan LF, Shreffler JH, Altpeter M, Schoster B, Hootman J, Houenou LO, et al.Evaluation of group and self-directed formats of the Arthritis Foundation's Walk With Ease Program.Arthritis Care Res (Hoboken)2011;63:1098107.
  • 35
    Kelley GA, Kelley KS, Hootman JM, Jones DL.Effects of community-deliverable exercise on pain and physical function in adults with arthritis and other rheumatic diseases: a meta-analysis.Arthritis Care Res (Hoboken)2011;63:7993.
  • 36
    Ettinger W Jr, Burns R, Messier S, Applegate W, Rejeski W, Morgan T, et al.A randomized trial comparing aerobic exercise and resistance exercise with a health education program in older adults with knee osteoarthritis: the Fitness Arthritis and Seniors Trial (FAST).JAMA1997;277:2531.
  • 37
    Pennix B, Messier S, Rejeski W, Williamson J, DiBari M, Cavazzini C, et al.Physical exercise and the prevention of disability in activities of daily living in older persons with osteoarthritis.Arch Intern Med2001;161:230916.
  • 38
    Hawker G, Stewart L, French M, Cibere J, Jordan J, March L, et al.Understanding pain experience in hip and knee osteoarthritis: an OARSI/OMERACT initiative.Osteoarthritis Cartilage2008;16:41522.
  • 39
    Hochberg MC, Altman RD, April KT, Benkhalti M, Guyatt G, McGowan J, et al.American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.Arthritis Care Res (Hoboken)2012;64:46574.
  • 40
    Zhang W, Nuki G, Woskowitz R, Abramson S, Altman R, Arden N, et al.OARSI recommendations for the management of hip and knee osteoarthritis: part III. Changes in evidence following systemic cumulative update of research published through January 2009.Osteoarthritis Cartilage2010;18:47699.
  • 41
    Marra C, Lynd L, Harvard S, Grubisic M.Agreement between aggregate and individual-level measures of income and education: a comparison across three patient groups.BMC Health Serv Res2011;11:69.
  • 42
    Franks P, Gold M, Fiscella K.Sociodemograhpics, self-rated health, and mortality in the US.Soc Sci Med2003;56:250514.
  • 43
    Jacobi CE, Mol GD, Boshuizen HC, Rupp I, Dinant HJ, van den Bos G.Impact of socioeconomic status on the course of rheumatoid arthritis and on related use of health care services.Arthritis Rheum2003;49:56773.
  • 44
    Harrison M, Tricker K, Davies L, Hassell A, Dawes P, Scott D, et al.The relationship between social deprivation, disease outcome measures, and response to treatment in patients with stable, long-standing rheumatoid arthritis.J Rheumatol2005;32:23306.
  • 45
    Yelin E.Health policy and musculoskeletal conditions.Curr Opin Rheumatol1992;4:16773.
  • 46
    Bombard J, Powell K, Martin L, Helmick CG, Wilson W.Validity and reliability of self-reported arthritis: Georgia senior centers, 2000-2001.Am J Prev Med2005;28:2518.
  • 47
    Sacks J, Harold L, Helmick CG, Gurwitz J, Emani S, Yood R.Validation of a surveillance case definition for arthritis.J Rheumatol2005;32:3407.
  • 48
    Katz PP, Morris A.Use of accommodations for valued life activities: prevalence and effects on disability scores.Arthritis Rheum2007;57:7307.
  • 49
    Theis KA, Furner S.Shut-in? Impact of chronic conditions on community participation restriction among older adults.J Aging Res2011;2011:759158.
  • 50
    James B, Boyle P, Buchman A, Bennett D.Relation of late-life social activity with incident disability among community-dwelling older adults.J Gerontol A Biol Sci Med Sci2011;66A:46773.