Arthritis Foundation community-based physical activity programs: Effectiveness and implementation issues



There is a growing body of evidence that regular, moderate physical activity not only is safe for people with arthritis but also can significantly reduce pain and disability. Nevertheless, physical inactivity remains common among people with arthritis. Community-based, group exercise classes are one approach to promoting physical activity. Existing physical activity programs for people with arthritis appear to be beneficial but they are underutilized. This article describes 2 such programs developed by the Arthritis Foundation: the Arthritis Foundation Aquatic Program (AFAP) and the People with Arthritis Can Exercise (PACE®) program. The available evidence of their effectiveness and implementation issues are reviewed. Limitations of the research evidence and research needs are also discussed.

Statement of the problem

Arthritis and related musculoskeletal conditions are the leading cause of disability in the United States, making them a significant public health issue (1). More than 7 million Americans report limitation of activity due to arthritis (2) and there is a significantly higher rate of no leisure-time physical activity among people with arthritis as compared with the general population (3). Prolonged inactivity can result in much of the morbidity attributed to arthritis—problems such as fatigue, reduced joint flexibility and muscle strength, loss of independence, poor endurance and overall fitness, and depression. Inactivity also increases the risk for comorbid conditions, such as coronary heart disease, diabetes, and osteoporosis (4). On the other hand, several randomized, controlled trials have shown that people with arthritis who participate in appropriate land-based or aquatic exercise programs can experience significant improvements in their physical and psychosocial status without exacerbating their disease (5–11). Based on such evidence, the Surgeon General's report on physical activity and health concluded that people with osteoarthritis (OA) and rheumatoid arthritis (RA) can improve their pain and function by participating in regular moderate aerobic or resistance training exercise programs (12). In the past, patients with arthritis were often cautioned against exercising, however, the current American College of Rheumatology guidelines for OA management include exercise as a mainstay of treatment (13).

The National Arthritis Action Plan: A Public Health Approach (14), a blueprint for population based approaches to reduce the burden of arthritis, and Healthy People 2010 Summary of Objectives (15) highlight the need for interventions that promote physical activity in people with arthritis and other disabilities. The Arthritis Foundation has developed 2 physical activity programs, the AFAP and the PACE® program, that have been evaluated in different settings and appear to provide some positive benefits to participants. Unfortunately, these programs are not widely utilized, reaching less than 1% of people with arthritis (Arthritis Foundation, unpublished data).

Program overview

People with arthritis can get caught in a cycle of pain, inactivity, depression, and social isolation. Both AFAP and PACE® are recreational, community-based programs that are designed to address these problems in multiple ways. Offered in a group format, these programs encourage peer interactions and socialization that can help counteract feelings of depression and isolation. And although these are nonclinical programs and will not replace an individually prescribed regimen of therapeutic exercise, they do include exercises that can result in physical benefits. In addition to traditional exercises, the programs contain a variety of other components, such as breathing and relaxation techniques and socialization games, to help decrease symptoms like pain, stress, depression, and fatigue. In addition, PACE® includes health education segments and behavioral strategies to promote physical activity between classes. Table 1 lists the components that are available in each program.

Table 1. Program content*
ComponentAF Aquatic ProgramPACE®
  • *

    AF = Arthritis Foundation; PACE® = People with Arthritis Can Exercise.

Range of motionXX
StrengtheningX (against water)X (against gravity, light weights, or theraband)
Endurance/aerobicX (optional)X (optional)
Weight-bearing activities X
Balance and coordinationXX
Posture/body mechanicsXX
Body awarenessXX
Breathing/chest expansionXX
Relaxation techniques X
Health education segments X (lectures on principles of exercise, making exercise a routine, ways to manage stress and pain, osteoporosis)
Special activities (games to encourage socialization, promote sense of well being)XX
Strategies to promote self care and self efficacy X (goal setting and contracting, diaries, discussion of barriers to home exercising, buddy systems, written participant's manual)

Program content was developed by multidisciplinary teams including physical therapists, occupational therapists, rheumatologists, health educators, nurses, fitness professionals, movement therapists, and people with arthritis. Because the intent was to reach as many people with arthritis as possible, PACE® and the AFAP were designed to accommodate groups with varying levels of limitations. In either program, each class session typically includes about 25 exercises, but instructors can choose from more than 70 exercises, which include multiple options for each major joint motion. For instance, in the PACE® program, there are options for doing knee flexion exercises while sitting, standing, or lying on the floor. Both programs are also offered in 2 different levels: a basic program for people who are currently sedentary or who have limited joint mobility, and an advanced level for individuals who have minimal joint involvement or whose primary issues are decreased muscle strength and fatigue.

Both programs are reviewed and updated as needed: PACE® was revised in 1999 and AFAP in 2002. Changes in the programs over the years have reflected insights gained from the growing exercise research literature. As an example, because of data about the benefits of resistance training, the use of light weights was added as an option when the PACE® program was last revised.

Both programs are taught by trained instructors who have successfully completed a 1-day training workshop provided by the Arthritis Foundation or by a partnering organization such as the Young Men's Christian Association/Young Women's Christian Association (YMCA/YWCA). The instructors are usually employed by the facilities in which the program is offered, such as YMCAs, YWCAs, community parks and recreation centers, senior centers, senior residence facilities, religious organization buildings, and sports/fitness clubs. Facilities have the option of offering these programs from once to several times a week and as either a class series (typically once or twice a week for 6–8 weeks) or on an ongoing basis.

Program costs vary by site and include the expense of training the instructors, program marketing, recruitment expenses, instructor fees, and the facility or pool rental or maintenance costs. A cost effectiveness study of the AFAP by Donald Patrick and his colleagues at the University of Washington estimated that each participant's average annual class registration fees were $291. In addition to these class fees, they calculated each participant's average annual time costs ($894) and transportation expenses to and from the class ($404), which added considerably to overall intervention costs (16).

Effectiveness data

The safety and efficacy of both aquatics therapy and land- based exercise programs for people with arthritis have been established through multiple controlled experimental studies. Participants in clinical trials have demonstrated improvements in such outcomes as pain, morning stiffness, tender joint counts, muscle strength, cardiovascular fitness, and mood (5–11).

Evaluation studies of AFAP and PACE® have primarily focused on determining their effectiveness within real-life community settings as opposed to controlled clinical sites. Few of these studies have been published; many are unpublished masters theses or conference abstracts. Nevertheless, these effectiveness studies provide useful insights about the potential value of these programs as described in the following section.

AFAP data

A small pilot study of AFAP that utilized a pretest–posttest design showed significant improvements in pain and functional status at 4 months (Arthritis Foundation, unpublished data). A small, randomized, controlled trial of AFAP with women with arthritis demonstrated significant positive changes in joint range of motion and muscle strength (17). In the cost-effectiveness study by Patrick et al (16), 249 men and women with OA were randomized into a delayed treatment control group or a treatment group that was asked to attend at least 2 aquatics classes per week for 20 weeks. Using data obtained from weekly diaries and questionnaires, participants in the AFAP had fewer physician visits, equal quality of well being, and better functional status and perceived quality of life. Of this treatment group, those adhering to the minimum attendance of 2 classes per week (29%) had significantly higher quality of life and functional status. Although direct medical care costs did decrease slightly in the treatment group, these savings were offset by the program costs, which as noted earlier, not only included the expense of recruitment but also what participants paid for registration fees, transportation costs, and the value of their time associated with participation in the exercise classes.

PACE® program data

Several studies have been conducted on behalf of the Arthritis Foundation that utilized nonrandomized groups and a pretest–posttest design. An initial pilot study revealed significant improvements at 4 months in frequency of self care behaviors, level of pain, and the belief that self help is efficacious (Arthritis Foundation, unpublished data). A larger study, utilizing 176 adults with arthritis enrolled in PACE® courses at various sites around the country, showed a significant decrease in depression level and increased functional ability and self efficacy (18). After the last program revision, Quitoni evaluated 119 PACE® participants from 7 states (unpublished data). Significant improvements were found for arm, hand, and finger function, from pretest to posttest. Mood improved for those who attended classes twice a week or more.

An independent study conducted at 3 sites in Minnesota evaluated 45 people with RA or OA who attended PACE® classes twice a week for 7–8 weeks (19). This study demonstrated a significant increase in self efficacy and a decrease in depression from pretest to posttest. Kennedy et al conducted a randomized, controlled trial of 6 months of PACE® classes with 74 women with severe RA (20). Because she wanted to determine if PACE® was safe for people with moderate to severe RA, she restricted her study population to women with functional class II or III RA. Thirty-nine of the women were in the intervention group and 35 were in the control group. The intervention group participated in PACE® classes 3 times a week for 6 months. Kennedy et al concluded that the program was safe and effective. Comparisons of the intervention and control groups found that those in the intervention group had significant positive changes in health status and an increase in social activity as compared with the control group.

Another recent, unpublished study of the PACE® program provides data on the effectiveness of the program in a managed care setting (Arthritis Foundation, New York Chapter, unpublished data). Through a collaborative effort between the Arthritis Foundation, New York Chapter and the Health Insurance Plan (HIP) Health Plan of New York, 551 participants in 25 PACE® classes were evaluated in a quasiexperimental, pretest–posttest design that utilized a delayed treatment control group consisting of 133 persons randomly selected from a wait list for the program. Compared with the control group, PACE® participants who attended at least 5 weeks of classes reported significant improvements in use of self care techniques and ability to perform everyday tasks, arthritis symptoms in the past month, knowledge of arthritis care, self efficacy, and level of depression.

Summary of existing research

Table 2 describes the study populations, study designs, and results of the various studies conducted on the AFAP and the PACE® program. When available, information about the type of disease and demographic characteristics of the participants are indicated. As shown, there is evidence that suggests that both programs improve psychosocial status, including decreased depression or improved quality of life, and may have positive impact on perceived function. Improvement in pain is less consistent.

Table 2. Evaluations of the Arthritis Foundation Aquatic Program and the PACE® program*
Study/authorSampleDesignInterventionStatistically significant results
BeliefsBehaviorsPainPhysical/ function and health statusPsychosocial
  1. * PACE® = People with Arthritis Can Exercise Program; AF = Arthritis Foundation; VAS = Visual Analog Scale; HAQ = Health Assessment Questionnaire; OA = osteoarthritis; RA = rheumatoid arthritis; ROM = range of motion; CHD = Current Health Desirability Rating Scale; PQOL = Perceived Quality of Life Scale; QWB = Quality of Well-Being Scale; HS = high school; CES-D = Center for Epidemiologic Studies-Depression Scale; AIMS = Arthritis Impact Measurement Scale; HIP = Health Insurance Plan.

Arthritis Foundation Aquatic Program
 AF, 1986 unpublished60 adults 70% OA; 28% RA mean age 67 88% female 93% white 14 years schoolPretest; 4 month posttest8 weeks classes; variable frequency  ↓18% (VAS)↑25% (HAQ)
 Suomi et al, 199727 women 53% RA; 47% OA mean age 60Randomized controlled trial (n = 17 in experimental group; 10 in control)6 weeks classes; 3 classes per week   ↑joint ROM, muscle strength 13% to 17%
 Patrick et al, 2001Statewide sample;  multisites 249 adults with OA mean age 66 86% female 94% white 69% at least HSRandomized controlled trial (n = 125 in treatment; n = 124 in delayed treatment control group)Minimum 2 classes per week for 20 weeks   ↑10% (HAQ)↑ 6% CHD ↑ 11% PQOL ↑ 5% QWB  (adherers only)
 AF 1986, unpublishedNational; multisite 43 adults 65% OA; 19% RA mean age 76 98% female 97% white 13 years educationPretest; posttest at 4 months8 weeks classes; variable frequency↑efficacy of self help 17% ↑ confidence in ability to continue activities 22%↑ self care 213%↓ 24% (VAS)
 Doyle 1990, abstract3 sites in Minnesota 45 adults 44% RA; 42% OA mean age 61 84% female 98% white 14 years educationPretest; posttest at 8 weeks7–8 weeks 2 classes/week↑ self efficacy   ↓ depression
 Anderson 1991, masters thesisNational: multisite 176 adults 53% OA; 27% RA mean age 68 86% female 95% white 13.7 years  educationPretest; posttest at 8 weeks and followup at 4 months6–8 weeks 1–2 classes/week↑self efficacy 10%  ↑ 17% (HAQ) maintained at 4 months↓ depression 19% (CES-D) maintained at 4 months
 Kennedy 1992, abstractClinical sample 74 women with Functional Class II or III RA median age 60Randomized controlled trial (n = 39 treatment; n = 35 control)6 months classes 3 classes/week   ↑ health status 10% (AIMS)↑ social activity 27% (AIMS)
 Quitoni 2000, masters thesis7 states sample; 12  sites 76 adults mean age 71 81% female 89% white 50% at least HSPretest; posttest at 6 weeks6 weeks classes 1–5 classes/week (groups analyzed by frequency of attendance; those attending ≥ 2 times/week = frequent users)  ↑ satisfaction with pain level (in frequent users only)↑ arm, hand, finger function (AIMS2) ↑ satisfaction with mobility levels, walking and bending (in frequent users)↑ mood (AIMS2) and satisfaction with quality of social interactions in frequent users
 AF New York/HIPNew York City; 25 classes in multipleQuasi-experimental1 class/week, 8 weeks; adherers = minimum attendance at 5 classes↑ arthritis knowledge ↑ self-efficacy  20%↑ use of pain management techniques 63%↓ pain 60% ↓ stiffness 48%↑ 35% (HAQ)↓ depression 14% (CES-D)

Several limitations of these data should be noted, including sampling, study design, intervention, and measurement issues. The generalizability of the evaluation results is hindered by the lack of representative samples. Most of the studies already conducted on AFAP and PACE® had small sample sizes and utilized samples that were primarily white, high school educated women with OA or RA. The New York study in a managed care setting is a notable exception, in that 52.5% of the experimental group and 35% of the control group were African Americans (unpublished data).

Few of the study designs were randomized, controlled trials; rather most of the studies used a pretest–posttest, single group or quasiexperimental, delayed treatment control group. As with any studies involving volunteers, caution must be exercised to avoid overinterpreting the results. The experimental studies focused on limited geographic populations—the AFAP evaluations were conducted at 1 site in Wisconsin (17) and in the state of Washington (16) and Kennedy's controlled trial of PACE® was done in Colorado (20).

There is little consistency in the intervention methodology. The studies varied in the intervention dose—from as little as 5 sessions to 6 months of classes—and even within the same study, the actual program content could vary. As noted earlier, instructors have a lot of flexibility in the program format and frequency. The Quitoni study (unpublished data), which also included in-depth interviews with the PACE® instructors, revealed that the frequency of classes varied from 1 to 3 times a week. Most of these instructors reported that while they did include the exercise components of the class, they were not usually including the health education, special activities, and behavioral techniques, which may account for why the participants did not experience the same level of positive benefits seen in the studies where there was more control of the program delivery.

There is also variability in the evaluation methodology, including the variables and instruments used and length of evaluation. Consistent with the objectives of these programs, (that is, to reduce the pain–depression–inactivity–social isolation cycle), most of the studies included an assessment of pain, function, general quality of life, and/or psychosocial status. How these variables were operationalized varied—for instance, function was assessed via the Health Assessment Questionnaire, Arthritis Impact Measurement Scales (AIMS), or AIMS2. The length of the evaluations ranged from immediate posttest at 6 weeks to a 6-month followup.

Very little is known about the dropout and adherence rates of program participants and how this might affect the evaluation outcomes. Patrick et al (16) reported that 17% of the treatment group did not complete their study and of those who did, 29% did not adhere to attending a 2-class per week minimum schedule. They discovered no significant differences between the completers and dropouts and they incorporated an intent-to-treat approach in their analysis (i.e., they incorporated the dropouts in their treatment arm), enhancing the generalizability of their results.

In summary, most of the reviewed studies were not randomized, controlled trials and were not published in peer-reviewed journals. Nevertheless, when viewed in aggregate, the preliminary evidence suggests that both AFAP and PACE® appear to be have positive benefits on psychosocial status, including depression or quality of life, and physical functioning. Impact on pain is less clear.

Implementation issues

Key implementation issues are quality control and low utilization. Assuring the quality of these programs is very difficult because they are primarily delivered through third parties. Current quality control measures include a standardized instructor training protocol, a certification process, and letters of agreement signed by facilities and instructors to help ensure adherence to program policies. Nevertheless, Arthritis Foundation chapters usually do not have the resources to monitor the classes and there is no ongoing way to measure program satisfaction or outcomes among program participants.

Perhaps the most meaningful measure of whether a community-based program is effective is the extent to which it is being utilized. In an editorial that accompanied the article by Patrick et al (16), Clark (21) identifies 2 aspects of utilization that are important in determining the effectiveness of any physical activity program: penetration, or the extent to which the intervention reaches those who might benefit, and participant adherence, described as the extent to which the participant followed the recommended frequency of classes. Information about penetration and adherence and how to increase utilization of AFAP and PACE® is still minimal. What is known is that the overall participation rate is low. The AFAP, which was codeveloped with the YMCA of the United States and released in 1983, is the oldest and most popular of the Foundation's physical activity programs. However, the program only reached 156,000 participants in 2001. PACE®, released in 1987, had 37,000 participants nationwide in 2001. The relatively greater success of the AFAP appears to be due to the Foundation's national collaboration with the YMCA, which has an infrastructure including more than 2,000 local service points, and ongoing training and program marketing that has sustained the program and promoted growth over time.

More recently, the Arthritis Foundation has tried to increase program penetration by focusing on reaching larger populations via collaborations with other community agencies, public health agencies, managed care organizations, and worksites to deliver its programs. The PACE® program implementation, which was designed to reach all of the HIP Health Plan members with arthritis, is an example of a successful population based intervention, which as defined by Brady and Sniezek in this issue (22), is an intervention designed to reach most members of a population group or subgroup. The net impact of such efforts on the national penetration rate of these programs has yet to be determined. Various reasons have been suggested for the low participation rates in physical activity programs. Qualitative research done by the Arthritis Foundation has identified lack of awareness of its programs and misperceptions about the value of exercise. The full name of the PACE® program (People with Arthritis Can Exercise) was designed purposefully to counteract the myth that people with arthritis should not exercise. Anecdotal information suggests that younger people with arthritis are not attracted to programs that primarily include older adults—as is common in most of the Arthritis Foundation's programs. Others have cited their desire to attend mainstream wellness physical activity programs rather than those, like the Foundation's programs, that label them as having arthritis. Another reported barrier is the lack of reimbursement because these programs are not considered therapeutic. However, as evidenced by the HIP Health Plan implementation of PACE®, some managed care organizations are beginning to provide these classes on a complementary or reduced cost basis to their members.

Other potential barriers for people with arthritis are suggested in the literature and include low self efficacy, the lack of awareness about the benefits of exercise, inaccessibility of exercise facilities, inconvenient schedules of classes, lack of support from significant others, and lack of time (23–25). The general literature on physical activity identifies lack of readiness to change and preference for engaging in physical activity outside of formal groups or classes as other obstacles (26, 27). The research team involved in the Patrick et al study of the AFAP (16) published a separate article on the challenges involved in recruiting for this statewide, community-based sample versus a clinical sample (28). Although the article does not describe the total size of the target population or how many were exposed to the recruitment messages, it does describe how the study population was successfully recruited. Successful recruitment strategies included a recruitment letter distributed by the Arthritis Foundation to its members, local television coverage, local radio and newspaper public service announcements, and physician referral (utilizing a list of physicians supplied by the chapter). Paid newspaper ads were not successful.

Similarly, the New York Chapter of the Arthritis Foundation utilized multiple strategies to recruit PACE® participants from HIP Health Plan subscribers. Targeted mailings sent to 20,000 HIP members with arthritis were supplemented with articles in the HIP newsletters (reaching a total of 760,000 readers), a mailing to 250 HIP providers whose patients were being asked to participate, an article in the chapter's newsletter (circulation of 55,000), and information on the chapter's Website. After the first year of the project, the chapter had received 1,714 inquiry calls; 445 HIP members had participated in PACE® classes; and another 1,095 were on a waiting list.

The level of participant adherence also should be considered in any evaluation of a program's effectiveness. Both the Patrick study of AFAP (16) and the Quitoni study of PACE® (unpublished data) showed that those who participated in the programs at least twice a week had better outcomes. However, getting people to class more frequently is a significant challenge. Patrick et al (16) suggest the use of motivational incentives such as telephone reminders, reinforcements, and rewards.

While getting people into classes and to stay physically active remains a challenge, there is also evidence that many people with arthritis continue these programs on a long-term basis. In a recent, cross-sectional survey of 178 PACE® participants from 19 classes around the country, the majority reported attending classes twice a week—of these, 41% indicated they were physically active at least 4 times a week. Fifty-seven percent of those surveyed had been attending PACE® classes for 1 or more years (Hinchman, unpublished data). An improved understanding of these more active participants may provide insights regarding strategies to improve adherence in others.

Suggestions for future research

There is a clear need for more published studies of the PACE® program and AFAP that build on the work that has already been done regarding effectiveness. Additional experimental studies that utilize different settings and different populations, including with people with varying diseases and level of severity, younger adults, men, and low-income and minority populations are necessary. As the Arthritis Foundation has increased its efforts to disseminate its programs to populations of people through public health agencies, managed care organizations, and employers, it has discovered the need to have more evidence on the cost effectiveness of its programs. The Patrick et al (16) study of AFAP is an excellent example of a well-done, comprehensive, cost effectiveness study; however, more of these studies are needed. Research is also needed to determine if these programs are effective in increasing physical activity outside of the classes and what is the relative impact of class attendance and these other physical activities. Longitudinal studies are needed to determine the long-term impact of these programs.

Clark recommends studies that not only evaluate the cost effectiveness of the intervention, but which also collect penetration data and compare different approaches to improving adherence (21). The programs appear to be beneficial in reducing the cycle of problems of pain, inactivity, depression, and social isolation; further research is warranted to determine which program elements are most critical to success.

Market research is also needed to answer questions that would help identify who are the most appropriate targets for these programs. What are the characteristics of people who do not choose to participate? What are the characteristics of the active participants—what got them started, what sustains them, and how have they overcome barriers? What factors impede or stimulate class participation, dropping out, or continued attendance?

Practical implications for clinicians

Additional research is warranted to further identify which patients will benefit the most from the AFAP and PACE® classes. However, the preliminary evidence suggests that these programs are safe and beneficial for many people with arthritis. Clinicians are therefore encouraged to refer their patients to these programs by contacting the Arthritis Foundation ( or 1-800-283-7800). As noted earlier, physician referral has been demonstrated to be an effective strategy for getting people into aquatics classes. Minor points out that to achieve optimal benefit, patients should be given specific recommendations about type, intensity, duration, and frequency of exercise (29). The research cited here suggests that patients be advised to attend PACE® or AFAP classes at least twice a week as part of the overall recommendation to increase their physical activity. Many patients may need their fears about the safety of exercise allayed; asking about their concerns and perceived barriers to exercise can help guide this discussion (30). Brady et al (31) suggest a 4-step approach to promote physical activity and other self management behaviors: 1) Reinforce the value of increasing physical activity; 2) Recommend strongly that the patient start the activity; 3) Refer to specific resources; and 4) Reconsider your approach if your patient doesn't become physically active, matching your exercise recommendations to their situation and reasons for not exercising. Since ongoing research is critical, clinicians are also encouraged to support their patients' participation in studies of these programs.