Are behavioral interventions for arthritis effective with minorities? Addressing racial and ethnic diversity in disability and rehabilitation




To determine whether research evaluating the effectiveness of behavioral interventions for arthritis demonstrates that these interventions are effective with, and appropriately utilized by, minority participants.


A systematic review was conducted of arthritis intervention research from 1997 to 2008. For each article, information was gathered on the percentage of participants who were from different racial/ethnic groups, whether interventions were shown to be effective for minority participants, whether differential attrition analyses were conducted, whether efforts were reported in minority recruitment and retention, and whether attempts were made to make interventions culturally appropriate.


We identified 25 randomized intervention studies. Of these, only 2 reported on whether the intervention was similarly effective for white and black patients (equal effectiveness was found), and 6 studies reported examining differences in attrition by race (higher attrition in nonwhites was found in 1 study). Most studies did not report the percentage of participants from specific minority groups, and in many studies the percentage of minority participants was small. No studies reported making systematic efforts to assure that interventions were culturally appropriate for minority participants.


Minority patients with arthritis are at risk for higher levels of disability than white patients, but little is known about whether evidence-based interventions for arthritis are effective for culturally diverse patients. In addition, minority patients appear to be underrepresented in intervention research, and too little attention has been paid to minority recruitment and assuring that interventions are culturally appropriate for diverse patients.


Health disparities are increasingly recognized as an important issue. In the area of arthritis, blacks report a higher prevalence of activity/work limitations and severe pain (1), more functional impairment (1), lower utilization of surgical treatments (e.g., joint replacement therapy) (2), and less likelihood of ever participating in an arthritis self-management program (3) compared with whites.

Addressing the needs of minority groups with arthritis is an urgent public health issue. Interventions developed and tested on predominantly white samples may not be appropriate for, or effective with, minority populations (4). Although there is evidence that arthritis interventions are efficacious in reducing pain and disability (5), we are not aware of a review that has systematically examined whether behavioral interventions for arthritis are effective for patients from racial/ethnic minority groups. Reviews on other health-related topics, including treatment of depression and anxiety (6), demonstrate that minority patients are underrepresented in intervention research. Yet, when evidence-based mental health interventions are adapted to be culturally sensitive to meet the needs and expectations of minority populations, they are generally found to be equally effective across diverse racial/ethnic groups (6, 7).

The purpose of this review is to examine published research on behavioral interventions for arthritis, and to identify 1) rates of inclusion of minority patients, 2) analyses conducted to examine whether treatment efficacy is comparable for white and minority participants, 3) information on differential attrition by minority participants, 4) special efforts in recruitment and retention of minority participants, and 5) information showing efforts to assure that interventions were culturally appropriate to meet the needs and preferences of minority participants. Finally, we offer suggestions for future work, provide suggestions for ways to successfully recruit and retain minorities in existing arthritis interventions, and discuss models of nonarthritis interventions that successfully addressed racial/ethnic diversity.


We conducted a systematic review of arthritis interventions from 1997 to 2008 using PubMed. We concentrated on psychosocial interventions (e.g., coping, self-management) and exercise interventions conducted in the US that employed randomized controlled trials (RCTs). Only articles that included outcome data (e.g., pain, disability, depression, self-efficacy) from RCTs were included in our review. We primarily reviewed osteoarthritis (OA) interventions but included interventions where some participants had OA and others had another type of arthritis (e.g., rheumatoid arthritis). Search terms included “randomized controlled trial,” “osteoarthritis,” and “intervention,” in combination with either “psychosocial,” “self-management,” or “exercise.” We also included articles that we were aware of from other sources, including the reference lists of articles we examined. We identified 25 intervention studies. The results from our review are shown in Table 1.

Table 1. Summary of arthritis intervention studies and minority participation*
Authors (year)Program type% minorityReported analyses for treatment by race/ethnicityReported attrition analyses by race/ethnicityReported special efforts in minority recruiting and/or retentionReported addressing cultural appropriateness of intervention
  • *

    FAST = Fitness Arthritis and Seniors Trial; ASMP = Arthritis Self-Management Program; API = Asian/Pacific Islander; ADAPT = Arthritis, Diet, and Activity Promotion Trial; SMART = Self-Management Arthritis Treatment; PACE = People with Arthritis Can Exercise.

Ettinger et al (1997)FAST24% African American (aerobic exercise); 28% African American (resistance training); 26% African American (health education/control)YesYesNoNo
Fries et al (1997)Mail-delivered ASMPNo information provided on racial/ethnic composition of sampleNoNoNoNo
Suomi et al (1997)Arthritis Foundation Aquatics ProgramNo information provided on racial/ethnic composition of sampleNoNoNoNo
Lorig et al (1998)ASMP10% nonwhite (3-week program); 18% nonwhite (6-week program)NoNoNoNo
Keefe et al (1999)Spouse-assisted pain coping skills programNo information provided on racial/ethnic composition of sampleNoNoNoNo
Deyle et al (2000)Manual physical therapy and knee exercise programNo information provided on racial/ethnic composition of sampleNoNoNoNo
Messier et al (2000)Exercise and weight loss program for older adults with knee osteoarthritisNo information provided on racial/ethnic composition of sampleNoNoNoNo
Messier et al (2000)FAST8 black/25 white (aerobic); 8 black/28 white (health education); 8 black/26 white (weight training)NoNoNoNo
Penninx et al (2001)FAST28% African American (control group); 21% African American (resistance exercise); 25% African American (aerobic exercise)YesYesNoNo
Solomon et al (2002)ASMP4% nonwhite (intervention); 7% nonwhite (control)NoNoNoNo
Suomi et al (2003)National Arthritis Foundation aquatic and on-land exercise programsNo information provided on racial/ethnic composition in this sampleNoNoNoNo
Baird et al (2004)Guided imagery with progressive relaxationNo information provided on racial/ethnic composition in this sampleNoNoNoNo
Hughes et al (2004)Fit and Strong! exercise program (preliminary results)12.8% African American, 1.3% Hispanic, 1.3% API (treatment group); 12.9% African American, 4.3% Hispanic (control group)NoYesNoNo
Messier et al (2004)ADAPT21% nonwhite (healthy lifestyle group); 28% nonwhite (diet only group); 25% nonwhite (exercise only); 22% nonwhite (diet plus exercise)NoYesYesNo
Keefe et al (2004)Spouse-assisted pain coping skills programNo information provided on racial/ethnic composition of sampleNoNoNoNo
Lorig et al (2004)SMART program5% nonwhiteNoYesNoNo
Lorig et al (2004)SMART ASMP10% nonwhiteNoNoNoNo
Focht et al (2005)ADAPTNo information provided on racial/ethnic composition of sample; details are provided in prior articles (21, 32)NoNoNoNo
Baird and Sands (2006)Guided imagery with relaxation intervention (pilot study)No information provided on racial/ethnic composition of sampleNoNoNoNo
Hughes et al (2006)Fit and Strong! exercise program (final sample)27.8% African American, 1.9% Hispanic, 0.9% API (treatment group); 16.3% African American, 3.3% Hispanic, 3.3% API, 2.2% other (control group)NoYesNoNo
Mangani et al (2006)FAST26.2% nonwhiteNoNoNoNo
Miller et al (2006)Physical activity, inflammation, and body composition trial14.0% black, 0% Native American, 0% API, 0% Hispanic (weight stable group); 9.1% black, 4.5% Native American, 0% API, 0% Hispanic (weight loss group)NoNoNoNo
Brismee et al (2007)Tai chi programNo information provided on racial/ethnic composition of sampleNoNoNoNo
Wang et al (2007)Physical activity, inflammation, and body composition trial12.5% nonwhite (weight loss group); 6% nonwhite (weight stable group)NoNoNoNo
Callahan et al (2008)PACE25% nonwhiteNoNoNoNo


Inclusion of minority participants.

In examining intervention studies and their attention to treatment of minority participants, articles generally fit into 1 of 3 categories. One group of studies (11 total) provided no information on racial/ethnic diversity (8–18). Other studies (8 total) reported the percentage of participants who were white and nonwhite, but did not provide more detailed racial/ethnic breakdowns of their participants (19–25). A third group of studies (6 total) provided detailed information on racial/ethnic diversity (26–31), of which 3 in particular included a substantial numbers of blacks (26, 28, 30). In many studies, the percentage of minority participants was small.

Analyses addressing effectiveness of interventions for minority participants.

We found only 2 studies that examined whether arthritis interventions were comparably effective for white and minority patients (26, 28). Ettinger and colleagues found, in subgroup analyses, that both blacks and whites in exercise intervention groups showed improved outcomes in pain, disability, and walking distance (26). Penninx and colleagues also examined differential treatment effectiveness by race and found none (28).

Analyses addressing differential attrition by minority participants.

Six studies reported analysis of attrition including race/ethnicity, with 1 showing greater attrition from the intervention study for nonwhites (22) and 5 showing no differences based on race/ethnicity (21, 26, 28–30). Two of these studies (29, 30) reported no differences in attrition, comparing responders and nonresponders on demographic characteristics without explicitly stating that they examined race. However, race was one of their demographic characteristics.

Efforts in recruiting and retention of minority participants.

We examined whether studies reported making any special efforts to recruit and/or retain minority participants (e.g., work with churches in minority neighborhoods). The Arthritis, Diet, and Activity Promotion Trial (ADAPT) intervention showed extensive efforts to recruit minorities (21). Notably, Messier and colleagues (21) reported in detail their efforts at minority recruitment (e.g., development of culturally sensitive recruitment materials), but the other 24 studies reported no special effort at minority recruitment or retention.

Efforts to assure that arthritis interventions were culturally appropriate for minority participants.

None of the articles that we reviewed mentioned efforts to assure that interventions were appropriate to meet the needs and preferences of minority patients.


Despite substantial evidence on the effectiveness of behavioral interventions for arthritis, there is insufficient evidence that many of these interventions are effective, or even appealing, to people from diverse minority groups. Although we applaud the extensive efforts of the ADAPT intervention (21, 32) to recruit individuals from minority groups, we found only 2 intervention studies that examined racial/ethnic differences in effectiveness of treatment (26, 28) and 6 studies that examined race differences in attrition rates (21, 22, 26, 28–30). Generally, the literature that we reviewed shows underrepresentation of minorities, lack of detail about the racial/ethnic makeup of samples (i.e., 19 of 25 studies provided no or limited information regarding race), and few results addressing whether minorities derive similar benefits as whites.

These problems are not unique to arthritis intervention research, and many researchers struggle with ways to increase minority participation. One review of clinical trials from 1989 to 2000 in selected areas of known health disparities (diabetes, cardiovascular disease, human immunodeficiency virus/acquired immunodeficiency syndrome, and cancer) published in the Annals of Internal Medicine, the Journal of the American Medical Association, and the New England Journal of Medicine found that 40% of trials did not report race, and <1% reported analyses by race/ethnicity (33). We also understand that, particularly in studies with small sample sizes, there may not be sufficient statistical power to examine the differential effects of the intervention by race. Studies may also have attended to diversity in recruitment or worked to assure that interventions were appropriate for minority patients and not reported this in their research articles. However, we offer the following recommendations for further research.

Researchers should report the percentages of their participants who come from specific racial/ethnic groups rather than simply reporting whether participants are white or nonwhite and whether the racial/ethnic makeup of participants is representative of the local population.

For interventions where there is a sufficient sample size, it is important to examine whether interventions show similar effectiveness across racial/ethnic subgroups.

A major priority should be to examine whether interventions shown to be effective in predominantly white samples (e.g., pain coping skills training [10, 15]) are also effective in diverse cultural groups. Such studies should oversample specific minority subgroups and examine whether these effective interventions are beneficial to culturally diverse participants.

Arthritis journals should consider adopting policies similar to those currently used by journals such as Stroke (34), which now includes the statement in their instructions to authors, “Please provide sex-specific and/or racial/ethnic-specific data, when appropriate, in describing outcomes of epidemiologic analyses or clinical trials; or specifically state that no sex-based or racial/ethnic-based differences were present.”

Research on whether existing arthritis interventions are culturally appropriate, or on modifying arthritis interventions in ways that would make them more appealing to and effective with minorities, should be a high priority for future research. Designing arthritis interventions to be culturally sensitive and appropriate will contribute to more successful recruitment and retention of members of racial/ethnic minority groups.

To achieve these recommendations, 2 issues deserve special attention: methods to successfully recruit and retain minorities, and assuring that interventions are appropriately flexible and culturally sensitive. These issues may be intertwined because successful retention of patients in research may be heightened by offering culturally competent and sensitive interventions.

When addressing the issue of recruitment and retention of minorities in arthritis interventions, a top priority must be to test the efficacy of existing interventions among underserved populations, and a major challenge will be recruitment and retention of minorities. Traditional recruitment methods, including health fairs, media advertisement, and cold calling and mailing may not be sufficient for the recruitment of minorities (35) and nontraditional modes may be necessary. Community outreach efforts may involve establishing trust, a long-standing commitment to the community, and inviting community members to serve on an advisory board (36, 37). Program-specific adaptations may include personalizing the invitation to participate (e.g., face-to-face recruitment), using culturally appropriate incentives, and inviting family members to participate (36, 37). Other suggestions include providing transportation and child care, providing the program in a safe, local, and familiar facility, incorporating the program with other scheduled activities (e.g., choir rehearsal), and offering a variety of classes at varying times (36, 37). For the investigator, this may involve engaging in actions that address the specific shared beliefs, values, and practices of the population (36, 38), which may diminish the one size fits all approach to intervention. For example, implementing an aquatics program among blacks may not be as feasible as with majority populations, because many blacks never learned to swim and fear drowning (37) and black women report hair maintenance as a barrier to swimming (39). This may not be known unless the investigator is willing to expend significant time and effort assessing the needs and values of the minority group.

The investigators in the ADAPT intervention (21, 32) used many of these methods to recruit members of racial/ethnic minority groups, and we believe that other arthritis interventions would benefit from adopting a similar approach. Some of their efforts included providing transportation if needed, seeking the endorsement of community leaders, working with church groups and community centers in minority neighborhoods, and developing culturally sensitive recruitment materials tailored for each specific minority group. Their final sample was 22% African American and also included participants who were Native American, Asian/Pacific Islander, or Hispanic, coming close to their minority recruitment goal of 30%, of which they expected 26% to be African American. We should also acknowledge the ADAPT investigators' efforts to examine racial/ethnic differences in adherence to diet and exercise interventions (40). They found no significant racial/ethnic differences in low attendance versus high attendance at sessions focused on improving diet and exercise in multivariate analyses (40). Future arthritis interventions would benefit from greater attention to these types of issues. We should also note that Lorig et al (19) reported that in a needs assessment conducted prior to their RCT, they found that coordinators of the Arthritis Self-Management Program ranked “cultural sensitivity of materials” as the least problematic aspect of the program. It would be instructive for future studies to ask minority participants to rank problematic and helpful areas with regard to the cultural sensitivity of interventions.

The National Cancer Institute has made increasing minority participation in clinical trials a high priority, and has funded 1 program of note, the Minority-Based Community Clinical Oncology Program (41). This program funds cooperative agreements between health care institutions that serve high numbers of minority patients as well as cancer centers conducting clinical trials. A recent report from this experience found dramatic improvements in minority patient participation in cancer clinical trials, and emphasizes the importance of a number of components, especially engaging minority physicians from the community.

At present there is almost no evidence base concerning the efficacy of arthritis interventions for minority patients, but there are some lessons from related areas of research that can be drawn upon. A distinction can be made between interventions that are designed to be culturally sensitive and appropriate versus interventions that are tailored to be used primarily with a subgroup. For example, interventions may be designed to appeal to diverse patients by incorporating diverse interventionists, training interventionists in cultural issues, and conducting intervention in community locations that will appeal to diverse patients (42). Some interventions are culturally tailored, i.e., focused specifically on meeting the needs of a special group and intended to be used primarily with this special group, such as low income urban minority populations (43). Based on their review of evidence-based mental health interventions, Miranda et al (44) suggested that when evidence-based mental health interventions are adapted to meet the needs and expectations of minority populations they are generally found to be effective, but that there was little evidence on the necessity of specific tailoring for different racial/ethnic groups.

Two common goals in arthritis rehabilitation are increasing physical activity and improving knowledge and psychological adaptation to arthritis. Excellent examples are available, in existing research on improving physical exercise (45) and improving family caregiver adaptation to Alzheimer's disease (42, 46, 47), of modification of interventions in order to make them appealing to and effective with culturally diverse individuals. Albright et al (45) emphasized surface level factors (e.g., using ethnically-matched interventionists or suggesting activities/locales known to local Latina women) and deep structure factors (e.g., respecting core cultural values including family responsibilities) in a study demonstrating increasing exercise in a sample of predominately Latina, low-income women. Using a culturally sensitive intervention, they found significant improvements in exercise among these women. The Resources for Enhancing Alzheimer's Caregivers' Health (REACH) project addressed family coping in over 1,200 white, black, and Hispanic caregivers (47). Extensive efforts were made to hire diverse intervention staff, offer training in cultural diversity issues to all intervention staff, schedule training workshops for blacks and whites in local churches rather than insisting that they attend sessions at the university, and utilize in-home educational followup sessions to avoid problems with transportation, with the goal of making interventions appropriate across different cultural groups. Intervention was of comparable effectiveness across race/ethnicity when these culturally sensitive practices were used, and the project successfully recruited and retained large numbers of whites, blacks, and Hispanics (47).

In conclusion, our review of the literature on arthritis interventions suggests that more attention is needed in terms of 1) recruiting and retaining members of racial/ethnic minority groups, 2) careful reporting on the percentages of participants from racial/ethnic minority groups, and 3) examining whether there are racial/ethnic differences in the effectiveness and appeal of arthritis interventions. Successful models of interventions that implement strategies to increase minority participation and the cultural sensitivity of interventions are available from other studies, which provide important methods that can be applied to arthritis interventions. In addition, investigators from the ADAPT study implemented a variety of strategies that can be extended to other arthritis interventions to increase the number of minority participants.

Health disparities are a major issue facing the field of arthritis care and research. Research on behavioral interventions for arthritis has come a long way, and we now have an evidence base demonstrating the effectiveness of arthritis interventions for improving disability and coping with arthritis. However, it is unclear whether these benefits also occur for minority participants, who may be among those in the greatest need. Research that leads to improved arthritis care for culturally diverse patients and families should become an important priority, and could advance the goal of bringing excellent arthritis care to everyone.


Dr. McIlvane 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 design. McIlvane, Baker, Mingo, Haley.

Acquisition of data. McIlvane, Baker, Mingo.

Analysis and interpretation of data. McIlvane, Baker, Mingo, Haley.

Manuscript preparation. McIlvane, Baker, Mingo, Haley.

Statistical analysis. McIlvane, Baker, Mingo.