Centers for Disease Control and Prevention, Atlanta, Georgia
Arthritis Program Division of Adult and Community Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, 4770 Buford Highway NE, Mailstop K-45, Atlanta, GA 30341
The health effects of regular physical activity (PA) are well established and include decreased mortality and morbidity related to several common chronic diseases, such as cardiovascular disease, diabetes, and cancer, as well as contributing to improved mental health, physical functioning, and weight control (1–3). Despite this, US adults continue to get inadequate PA (4). In fact, physical inactivity was felt to be such a large public health problem that in 1996 the US Surgeon General released the landmark report Physical Activity and Health: A Report of the Surgeon General (3). This report, as well as an earlier joint report from the Centers for Disease Control and Prevention and the American College of Sports Medicine (CDC-ACSM) recommend all US adults participate in regular, moderate-intensity, leisure-time PA (2). The CDC-ACSM recommendation specifically states “Every US adult should accumulate 30 minutes or more of moderate intensity activity on most, preferably all, days of the week.” Details of these and other (5) public health recommendations and guidelines for developing and maintaining cardiorespiratory fitness are summarized in Table 1.
Table 1. Summary of physical activity and cardiorespiratory fitness recommendations from various public and private health organizations
Centers for Disease Control and Prevention and the American College of Sports Medicine
•Reduce to <20% the proportion of adults who engage in no leisure-time physical activity.
•Increase to 30% the proportion of adults who engage regularly, preferably daily, in moderate physical activity for at least 30 minutes per day.
•Increase to 30% the proportion of adults who engage in vigorous physical activity that promotes the development and maintenance of cardiorespiratory fitness 3 or more days per week for 20 or more minutes per occasion.
•Increase to 30% the proportion of adults who perform physical activities that enhance and maintain muscular strength and endurance.
•Increase to 43% the proportion of adults who perform physical activities that enhance and maintain flexibility.
Monitoring national PA levels is important to define the extent of and to identify trends in physical inactivity, evaluate national health objectives, and to identify subpopulations where interventions should be targeted (6). Populations known to be at risk for physical inactivity include women, ethnic/racial minorities, older persons, disabled persons, overweight/obese persons, and persons with chronic diseases, including cardiovascular disease, diabetes, and arthritis (2, 3, 6, 7). The high prevalence of arthritis in the US (43 million) makes persons with arthritis a large part of the population targeted for PA improvements (8). Participation in PA delays the onset of functional limitation (9, 10), prevents obesity (11), and is essential for normal joint health (12, 13). In addition, PA has been shown to reduce pain and disability among persons with arthritis and increase their physical performance and self efficacy (14, 15). Yet, persons with arthritis often resist PA messages because activity may initially increase pain or because they were told, inappropriately, not to be physically active.
The purposes of this study are to 1) describe the sources of data used for surveillance and special studies of PA among persons with arthritis, 2) compare levels of physical inactivity between 3 national health surveys, 3) report the prevalence of PA levels among the general US adult population and among adults with and without self-reported arthritis using data from the Behavioral Risk Factor Surveillance System (BRFSS), and 4) identify areas for future research regarding physical activity and arthritis.
The 3 nationwide surveys that are used to monitor PA are also used to collect data on the prevalence of arthritis. To judge the usefulness of using these data sources to conduct surveillance of and investigate issues related to PA and arthritis, we describe the sampling frame, define the physical activity and arthritis measures (Table 2), and provide estimates of the prevalence of physical inactivity for all 3 surveys (Figure 1).
Table 2. Details of the 3 nationwide surveys used to collect surveillance data on physical activity and arthritis*
Mode of administration
Measure of arthritis
Physical activity recall period
Measures of physical activity
NHANES = National Health and Nutrition Examination Survey; NCHS = National Center for Health Statistics; CDC = Centers for Disease Control and Prevention; T = type; F = frequency; I = intensity; D = duration; NHIS = National Health Interview Survey; BRFSS = Behavioral Risk Factor Surveillance System.
Household interview and mobile examination
Civilian, noninstitutionalized adults and children
Radiographs, physician exam results, and self-report
T, F, I, D
Civilian, noninstitutionalized adults
Previous 2 weeks
T, F, I, D
Civilian, noninstitutionalized adults
T, F, I, D
Third National Health and Nutrition Examination Survey (NHANES III).
The NHANES III (1988–1994) is a nationally representative sample of the civilian, noninstitutionalized population and consists of a home interview and a clinical examination in a mobile examination center. To ensure reliable estimates for special groups, NHANES III oversampled ethnic minorities, young children, and the elderly. NHANES III asked about the type and frequency of selected activities engaged in during the previous month but did not collect data on the duration and intensity of PA, and thus, NHANES III can only be used to monitor physical inactivity. The prevalence of knee and hand osteoarthritis (among participants age 60 and older) in the NHANES III sample can be determined from radiographs and/or the examining physician's evaluation of hand, wrist, and knee pain and swelling. The adult questionnaire also asked questions about doctor-diagnosed arthritis and the presence of joint symptoms.
Beginning in 1999, NHANES became an annual survey that continually collects comprehensive PA data (including leisure-time, transportation, household, and occupational) and directly measures cardiorespiratory fitness with a submaximal treadmill test. Several questions specifically addressing sedentary behavior are also included. After 1999, arthritis can only be determined from self-report questions about doctor-diagnosed arthritis or the presence of chronic joint symptoms and not from radiographs. Additional details regarding the design and administration of the NHANES can be found at the National Center for Health Statistics (NCHS) Web site: http://www.cdc.gov/nchs/nhanes.htm.
National Health Interview Survey (NHIS).
The NHIS is an annual household interview of a national sample representative of civilian, noninstitutionalized adults (≥18 years). The year 1998 was the most recent year the NHIS included comprehensive PA questions. Each respondent is asked about the type, frequency, intensity, and duration of 24 specific activities engaged in during the 2 weeks prior to the interview. In the 1996 and earlier NHIS surveys, to identify arthritis status, respondents were given a list from which they were to report the presence of specific musculoskeletal conditions during the preceding 12 months, including arthritis and other rheumatic conditions. After 1996, participants were asked only about joint symptoms, not about physician-diagnosed arthritis. Further information regarding the NHIS survey, including sample design and questionnaires, can also be accessed through the NCHS Web site: http://www.cdc.gov/nchs/nhis.htm.
Behavioral Risk Factor Surveillance System (BRFSS).
The BRFSS survey is conducted annually by state health departments using random-digit dialing telephone methods. The survey consists of a set of core questions asked in all states, and standard optional modules selected by individual states. Leisure-time PA is measured using information about the type, frequency, and duration of the 2 activities in which respondents engaged most often outside of work in the past month. Respondents also reported distance for running, jogging, walking, or swimming. Data are scored, weighted to account for the sampling frame, and summed to produce combined state prevalences. In 2000, the BRFSS PA questions were included on the core survey and thus were asked of respondents in all 50 states. During the same year, arthritis status was obtained in 36 states via an optional module, and included questions about self-reported and doctor-diagnosed arthritis and chronic joint symptoms. Detailed technical information on the BRFSS surveys, as well as prevalence estimates for various behavioral risk factors and self-reported conditions can be found at the Centers for Disease Control and Prevention's Web site: http://www.cdc.gov/nccdphp/brfss/.
For simplicity, this analysis will report only detailed prevalences of PA levels among all noninstitutionalized US adults, as well as among adults with and without self-reported arthritis (35 states) using the most recent data available from the 2000 BRFSS survey. Illinois, 1 of the 36 states using the arthritis module in 2000, used a split-sample technique; because of difficulties in obtaining accurate subject weights, we did not include them in combined estimates for this analysis. The BRFSS defines moderate-intensity PA (3–6 mets; 1.0 met is equivalent to the energy expended at rest) as activities that cause some increase in heart rate and breathing, such as brisk walking and gardening; vigorous PA (>6 mets) activities are defined as those that result in large increases in breathing and heart rate, such as running, aerobic dance, or tennis. For analytic purposes, we created 3 levels of PA. Inactive persons reported no leisure-time PA. Adults classified in the insufficient activity category reported some moderate and/or vigorous activity but not for the recommended duration or frequency. Adults classified in the recommended activity category engaged in at least 30 minutes of moderate-intensity PA on 5 or more days per week or 20 or more minutes of vigorous-intensity PA on at least 3 days per week.
Respondents were defined as having arthritis if they had either doctor-diagnosed arthritis or chronic joint symptoms (responding “yes” to both of the following questions: “During the past 12 months, have you had pain, aching, stiffness or swelling in or around a joint?” and “Were these symptoms present on most days of at least one month?”). The prevalence of PA was stratified by various risk factors such as sex, age (18–29, 30–44, 45–64, 65–74, ≥75 years), racial/ethnic group (white non-Hispanic, black non-Hispanic, Hispanic, other), body mass index (<25, 25–30, >30), and years of education (≤8, 9–11, 12 or equivalent, 13–15, >15 years). A total of 175,246 respondents were included in the analysis of the total population (49 states and Washington, DC) and 40,650 respondents were included in the analysis of 35 states.
Physical activity levels among US adults.
Per the 2000 BRFSS, 27.3% of US adults did not engage in any leisure-time PA (Table 3). The prevalence of inactivity was higher among women (29.4%) than men (25.1%), among Hispanics (40.5%) and blacks (34.1%) than whites (24.0%), and among obese (34.6%) and overweight (25.4%) than normal weight (24.0%) persons. The prevalence of inactivity increased with age, ranging from 21.9% among 18–29 year olds to 39.1% among persons over age 75, and decreased with increasing years of education (61% for ≤8 years to 18.5% for >15 years) (Table 3). BRFSS trend data also suggest the prevalence of inactivity has been resistant to change, remaining around 30% for all 50 states and the District of Columbia during the decade 1988–1998 (7). Both the NHIS (15) and NHANES III (16) reported a somewhat lower prevalence of physical inactivity than the 2000 BRFSS (Figure 1). These differences may be related, in part, to the shorter recall period (previous 2 weeks in NHIS), operational problems (seasonal and regional survey differences in NHANES III), and potentially real temporal trends.
Table 3. Age-adjusted prevalence of physical activity levels among the total US population and among persons with and without self-reported arthritis, 2000 Behavioral Risk Factor Surveillance Survey*
Physical activity level prevalence (%)
Data were self-reported. Total US population = 49 states (excludes Illinois) and the District of Columbia; Data on arthritis = 35 states (AL, AK, AZ, CA, CO, CT, FL, GA, HI, ID, IN, IA, KS, KY, MD, MN, MS, NE, NH, NJ, NM, NC, NY, ND, OH, OK, OR, RI, SC, TN, UT, VT, VA, WI, WY). Inactive = no reported physical activity; Insufficient = some physical activity but not for recommended duration or frequency; Recommended = moderate physical activity for at least 30 minutes per session, 5 times per week, or vigorous activity for at least 20 minutes per session, 3 times per week.
Age group (years)
Body mass index (kg/m2)
12 or equivalent
In 2000, 26.2% of BRFSS respondents reported engaging in recommended levels of PA, whereas 46.5% reported insufficient PA (Table 3). The prevalence of both recommended and insufficient PA levels were higher among men than women, and among whites than other racial/ethnic groups. The prevalence of insufficient PA decreased with age and was highest among overweight (47.9%) persons and about the same among normal weight (45.8%) and obese (46.1%) persons. For recommended PA levels, the prevalence generally increased with age but decreased as body weight increased, from almost 30.2% among normal weight persons to 19.3% among obese persons. The prevalence of recommended PA also increased as education increased.
Physical activity levels among persons with and without arthritis.
Among the 35 states that reported both physical activity status and arthritis information on the 2000 BRFSS, the combined prevalence of arthritis was 32.3% (data not shown). Among persons with arthritis, 30.8% were inactive compared with 25.8% of persons without arthritis (Table 3). Among persons with arthritis, a prevalence of inactivity greater than 33% was found in women, persons aged ≥45 years, blacks, Hispanics, those of “other” race/ethnicity, those who were obese, and those with 12 or fewer years of education. In contrast, among persons without arthritis, only those who were aged ≥75 years or Hispanic had such a high prevalence.
As expected, given the higher prevalence of inactivity, persons with arthritis had a lower prevalence compared with those without arthritis of both recommended (24.3% versus 27.4%) and insufficient PA (44.9% versus 46.8%) (Table 3). These differences persisted across almost all demographic subgroups. Among persons with arthritis, blacks had the lowest prevalence (19.9%) of recommended PA of any race/ethnicity, more than 5.3 percentage points below the prevalence for whites and “other” race/ethnicities. Racial/ethnic disparities also were prominent among persons with arthritis reporting insufficient PA, where Hispanic and “other” persons had a prevalence more than 9.0 percentage points below those for whites.
The proportions of persons achieving recommended levels of PA through vigorous PA only and a combination of moderate and vigorous PA were about the same for persons with and without arthritis. Slightly more persons without arthritis achieved recommended levels of PA through moderate PA only than did persons with arthritis (Figure 2). Among those persons with arthritis who achieved recommended levels of PA through moderate PA only, walking, gardening, bicycling for pleasure, swimming laps, golf, and unspecified home exercises were the activities most often reported (data not shown).
We found that persons with arthritis have, not surprisingly, a worse PA profile than persons without arthritis. Despite the fact that multiple clinical trials have shown that PA and exercise are beneficial for people with arthritis, almost one-third of persons with arthritis are completely inactive, and only one-quarter get enough PA to comply with national PA recommendations (13–16). Yet, it is possible for persons with arthritis to meet the national recommendations by engaging in moderate types of joint friendly PA such as walking, gardening, bicycling, and swimming. In fact, of the 24% of persons with arthritis who met the national recommendations, well over half (13.1% of all persons with arthritis) met the recommendations by engaging in moderate PA only. Although we only present data from the 2000 BRFSS, 2 other national health surveys described here can be used to track trends in PA levels among persons with and without arthritis, to evaluate PA intervention programs, and to investigate future research questions regarding PA and arthritis.
The results from the 2000 BRFSS presented here are consistent with prior reports on PA levels among people with arthritis. Earlier data from the 1990–1991 NHIS reported a higher prevalence of physical inactivity among persons with arthritis than among persons without arthritis (17). Persons with arthritis in the NHIS also have a lower prevalence of recommended PA than people without arthritis. Because arthritis is the leading cause of disability, one benefit of increasing PA levels may be decreasing the development of arthritis-related disability (17). Evidence from randomized clinical trials suggests that PA, both aerobic and resistive exercise, benefits persons with knee osteoarthritis, contributing to modest improvements in disability, physical performance, and pain (13–16).
Data from the BRFSS, as well as other national health surveys, do have limitations. All information on the BRFSS is self-reported and subject to recall bias. However, due to the cross-sectional design, it is unlikely that respondent recall of PA would be different for people with or without arthritis. More important to this report of prevalence estimates is the fact that there is potential for misclassification bias because respondents are only asked to recall the 2 types of PA they engaged in most often in the past month. Some people who engage in 3 or more activities may be categorized into the wrong PA level because their total PA was not taken into account. Also, self-report of specific types of arthritis (e.g., osteoarthritis) has only been moderately correlated with clinical disease assessed by radiographs and, therefore, respondents may be misclassified as to arthritis status (18). Our case definition of arthritis attempts to capture the public health burden of all types of arthritis by including both diagnosed and undiagnosed arthritis and is not meant to capture a single clinical entity. However, to address this, validation studies of the BRFSS arthritis questions are currently being conducted by Centers of Disease Control and Prevention in both the community and managed-care settings.
The results reported here suggest there are 2 groups of persons with arthritis that may require different approaches to increasing their physical activity levels. First, there is a need to get the inactive persons with arthritis moving, preferably at recommended levels, but at least into the insufficient activity category. Even though the ultimate goal is for all adults to exercise at levels that develop and maintain cardiorespiratory fitness (2, 3), scientific evidence suggests that significant health benefits can be achieved by moving people from inactive to active (1). Furthermore, exercise need not be vigorous to improve health; even moderate daily activity can lower blood pressure, improve blood lipid and glucose profiles, and improve mental health (3). The challenge is to help persons with arthritis choose appropriate types of physical activity that will not excessively load joints and will be enjoyable and sustainable.
Second, there is a need to get persons with arthritis and insufficient PA to increase the frequency and duration of PA to recommended levels. This may involve targeting intervention programs to help them choose appropriate types of PA, counseling them to periodically alter the type and amount of PA to prevent overuse and boredom, teaching joint protection strategies, and instructing them on how to incorporate moderate PA into their usual day. As PA intervention programs for persons with arthritis are developed, tested, and implemented, established public health surveillance systems can be used to track changes in activity levels among this group.
Future research on PA and arthritis should focus on several areas: 1) identifying the types and amounts of PA that mitigate joint impact stresses but allow persons with arthritis to achieve national PA recommendations; 2) identifying factors that enable and inhibit persons with arthritis to achieve recommended PA levels; 3) developing, testing, implementing, and evaluating PA intervention programs for persons with arthritis; 4) identifying which PA interventions for the general population also reach persons with arthritis; and 5) determining whether the national PA recommendations are appropriate for persons with arthritis by examining the effects of recommended levels of PA on the symptoms, progression, and functional status of persons with arthritis.
In conclusion, the prevalence of both physical inactivity and arthritis are high among the US adult population. Physical inactivity is a modifiable behavior, and by increasing their PA levels, persons with arthritis may decrease their pain and the risk of functional impairment or disability while improving their quality of life and obtaining general health benefits. In fact, Healthy People 2010 specifically targets persons with arthritis for 3 PA objectives (Table 1) (19). Many persons with arthritis have comorbid conditions, such as heart disease, diabetes, and hypertension, conditions that may also be favorably impacted by increased PA levels. Persons with arthritis can safely achieve recommended levels of PA by choosing appropriate, joint-friendly types of moderate activity, and PA interventions for persons with arthritis should emphasize these types of activities.
Ms. Cher Dallal, intern, Division of Nutrition and Physical Activity, Centers for Disease Control and Prevention, assisted with analyzing the data used in this report.