Preferences for arthritis interventions: Identifying similarities and differences among African Americans and whites with osteoarthritis




To determine if there are differences or similarities in arthritis intervention preferences and barriers to participation between African Americans and whites with osteoarthritis (OA).


Using a needs assessment survey, intervention preferences and barriers to participation in arthritis interventions among African American (n = 60) and white (n = 55) adults with self-reported doctor-diagnosed OA were examined. T-tests, chi-square tests, and multiple regression analyses adjusting for covariates were examined to determine race effects.


While there were many similarities, African Americans were more likely to report cost (P < 0.01), lack of trust (P = 0.04), fear of being the only person of their race (P < 0.001), lack of recommendation from their doctor (P = 0.04), and lack of recommendation of a family member or friend (P = 0.02) as barriers to participating in a community-based self-management arthritis intervention. After adjusting for covariates, African Americans preferred interventions that provide information on arthritis-related internet sources (P = 0.04), solving arthritis-related problems (P = 0.04), and talking to family and friends about their condition (P = 0.02) in comparison to whites. African Americans also preferred an intervention with child care services provided (P < 0.01), with instructors and participants of the same race (P < 0.01 and P < 0.001, respectively) or sex (P < 0.001 and P = 0.03, respectively), that allows a friend (P = 0.001) or family member (P = 0.02) to attend, that is offered at a local church (P = 0.01) or clinic (P < 0.01), or that is mailed (P < 0.01).


Findings suggest that similar interventions are preferred across racial groups, but some practical adaptations could be made to existing arthritis interventions to minimize barriers, increase cultural sensitivity, and offer programs that would be appealing to African Americans and whites with arthritis.


Arthritis is the leading cause of disability in the US (1), with osteoarthritis (OA) being the most common type. OA often results in pain, disability, and poor psychological well-being. Arthritis interventions (e.g., coping, self-management, exercise) have been shown through randomized controlled trials to be effective ways to manage OA symptoms (2–5). Effective interventions include, but are not limited to, the Arthritis Foundation Aquatics Program (6), the Arthritis Self-Management Program (ASMP) (7–9), Walk With Ease (4), and psychological interventions for symptom management (10). Such interventions have been shown to decrease pain, depression, medical visits, disability, discomfort, and medical expenses, and to increase relaxation, quality of life, self-efficacy, functional ability, self-care behavior, and knowledge about arthritis (2, 10). However, the positive effect is only true for those populations reached by the interventions.

Research consistently highlights the lack of inclusion and lack of participation of African Americans in health-related research and clinical trials (11–13). Similar problems have been found with arthritis behavioral intervention programs and research. The majority of arthritis intervention studies have been conducted with predominantly white populations without any special attention to minorities (e.g., examining effectiveness, recruitment and retention, social or structural needs, preferences or expectations) (2, 5, 10). The paucity of studies focused on African Americans in arthritis interventions emphasizes the need to understand the present challenges (e.g., preferences of the individual, social determinants, health care delivery system structure) that may serve as a barrier to the inclusion of African Americans in arthritis interventions, particularly when considering the potential impact of the lack of inclusion on health outcomes.

Culture-related barriers are a factor that deserves some attention in this field. Race is a multidimensional social construct that provides an overarching category for individuals who may have shared norms, values, experiences, and histories that may uniquely contribute either proximally or distally to the lack of inclusion in health care options of this type (14, 15). Exploring how race differences in preferences and barriers might contribute to the lack of inclusion of African Americans in arthritis interventions is an important topic. In addition, designing effective interventions that can be successfully implemented and disseminated among African Americans with arthritis is predicated on understanding the potential influence of the cultural aspect of race.

Recent research has acknowledged that the lack of inclusion of minorities in arthritis intervention research is problematic. Researchers have begun to make an effort to address this gap in the literature by including African Americans in studies of the ASMP (16, 17) and other community-based arthritis interventions (18), adapting the materials to become more culturally sensitive (16, 18), including African American intervention leaders and coordinators, and working with African American community-based organizations (e.g., churches, sororities) (16–18). However, in these studies, the process of modifying the materials and delivery of the program for cultural sensitivity were not described in detail (16, 17), and the effectiveness, satisfaction, and acceptability of the strategies among African Americans were not reported (18). Moreover, 17% of the participants (n = 458) in the study conducted by Goeppinger et al (17) were African American, and Altpeter et al (18) reported that 19% of a combined sample (n = 1,517) from 4 community-based arthritis interventions were African American; however, analyses specific to African Americans were not included. To our knowledge, only 1 study to date has examined the acceptability or the effectiveness of an arthritis intervention (i.e., the ASMP) among African Americans (16). Without randomized controlled trials for arthritis interventions including and reporting findings on minorities (e.g., African Americans), the question remains as to whether arthritis interventions are appealing, acceptable, feasible, or effective in racial/ethnic minority populations.

Therefore, the purpose of this research was to determine if there were racial differences or similarities in barriers to participating in arthritis interventions or in preferences of arthritis intervention content, structure, and delivery. Past research has examined the needs of patients with arthritis; however, samples included predominately white participants, and analyses did not examine race differences (19, 20). While limited work using needs assessments has been conducted in the area of arthritis and arthritis disparities, utilizing needs assessments to evaluate the care needs of a target population is not a novel concept (21, 22). Researchers have used needs assessments to gather information on health service needs and cultural diversity in other contexts (e.g., health service needs for Latino youth [21] and end of life care needs for African Americans [22]), and to use this information to make interventions appropriate for diverse individuals.

Our project addressed the following research questions: 1) Are there race differences in barriers to arthritis interventions between African Americans and whites with OA? and 2) Are there race differences in intervention preferences (i.e., content, structure, and delivery) between African Americans and whites with OA? Based on previous literature (12, 13, 23, 24), it was hypothesized that there would be significant race differences in barriers to participation in arthritis interventions. African American participants were predicted to be more likely to report that lack of trust in the health care system, family responsibilities, cost, and transportation were barriers to participating in arthritis interventions. Due to the lack of extensive previous literature on issues of preferences for arthritis interventions, racial differences in the other issues were explored without hypotheses.

Significance & Innovations

  • The findings highlight the need to address practical issues and make systematic adaptations in order to make existing arthritis interventions appealing to both African Americans and whites.

  • Arthritis intervention studies typically have not included a substantial amount or have not reported the inclusion of African American participants. This study highlights the importance of including minority groups in randomized controlled trials testing the effectiveness of arthritis interventions as a way to ensure feasibility, appeal, and effectiveness across all groups.

  • Culturally appealing arthritis interventions ultimately may increase the utilization of existing arthritis interventions.



Participants were community-dwelling adults recruited from multipurpose senior centers, church groups, and senior housing communities in the Tampa, Florida area. Participants were also recruited from an existing arthritis participant registry at the School of Aging Studies, University of South Florida. The individuals included participants with OA from previous studies who had agreed to participate in future studies. Flyers, presentations, and the snowball method were used to recruit and inform participants about the study. Those interested in the study were provided with additional information about the study and were prescreened to determine eligibility. Eligible participants included individuals ages ≥50 years self-reporting a doctor diagnosis of OA with no other complicating musculoskeletal diseases or observable cognitive impairments. Age 50 years was selected because the onset of OA often begins around middle age and is more common with older age. Reading ability was not a criterion, since participants who had trouble with reading due to ability or vision impairments were allowed to participate by having the questionnaire read to them. The self-report method used to identify individuals with OA is similar to the method used by the Centers for Disease Control and Prevention (CDC) to examine arthritis prevalence rates in the population (25). Moreover, this method has been shown to be a valid method among older adults (26). While recruitment efforts and data analyses centered on African American and white adults with OA, volunteers from other race groups were not excluded from participating if they showed an interest in the study. However, due to the small sample size, those who classified themselves as a racial/ethnic group other than African American or white were not included in the analyses. All of the participants provided written informed consent prior to enrollment.


This study was conducted using a nonprobability convenience sample survey research design. Based on what was convenient for the participants, the surveys were self-administered in either a group setting or individual appointments, and as noted, instruments were read to individuals who required assistance. The participants were provided a $5 Target gift card as a token of appreciation for their participation.


The instrument used in this study was designed as a structured questionnaire. In addition, 1 open-ended question was included to supplement quantitative findings pertaining to barriers to participation (i.e., Do you see the need for an arthritis self-help program for people who have been diagnosed with arthritis, Yes or No?; Please explain why or why not). Specifically, the questionnaire was designed to assess demographics, socioeconomic status (SES), health status, barriers, and preferences.

Independent variables.

Our primary variable of interest was race/ethnicity. To account for factors that may influence race/ethnicity differences, we included age, marital status, sex, SES, and health status as covariates.


Race/ethnicity was assessed by asking participants to “please tell me your race or ethnicity,” with the choices being 1) white/Caucasian, 2) black/African American, 3) Latino/Hispanic, 4) Asian/Pacific Islander, 5) Native American, or 6) other. Age was based on date of birth and measured as a continuous variable. Participants were also asked to report their marital status and sex.


SES was assessed based on income and education level. Participants were asked to identify their total annual family income (i.e., wages, pensions, dividends, and any additional household income) by selecting an income starting at $5,000 dollars and increasing incrementally, with the final choice being more than $100,000. Education was assessed with a 1-item question that asks, “What is the highest grade of school or year of college you have completed?” Choices ranged from 0 to 17+.

Health status.

Overall health status, disability, and pain of the participant were assessed using single-item questions from the Behavioral Risk Factor Surveillance System Questionnaire (27). The participants were asked to rate their health in general, with choices ranging from 0 = poor to 4 = excellent. Disability was assessed by asking the participants to respond yes (1) or no (0) to the question, “Are you limited in any way in any activities because of physical, mental, or emotional problems?” Pain was measured by asking the participants to rate their average joint pain over the past 30 days using an 11-point Likert scale (range 0–10, where 0 = no pain and 10 = extreme pain) (27). The participants were also asked to report the number of years they have had arthritis-related symptoms and the number of years with an OA diagnosis.

Dependent variables.

Potential barriers.

Barriers were assessed by presenting the participants with 2 vignettes describing existing arthritis interventions. The participants were then asked to read each vignette and answer the questions that follow. Each vignette was designed to assess barriers to participation in both an existing group community-based arthritis intervention (i.e., the ASMP) (7) and an existing home-based mailed arthritis intervention (i.e., the Arthritis Self-Management Tool Kit) (17). The first vignette stated, “Imagine that you have been diagnosed with arthritis and have been asked to participate in an arthritis group education course. The course will be offered at the local community center and will be taught by a person who has arthritis and has been trained to offer this course. The course will cost you $30. You will be asked to come to the center once a week from 10:00 am– 12:00 noon for six weeks. In addition to the weekly meetings you will receive The Arthritis Helpbook (a textbook for the course), Arthritis Foundation brochures, and a year's subscription to the Arthritis Today magazine. During the six weeks you will develop an exercise program just for you, discuss arthritis medications, learn how to manage your pain, learn how to solve arthritis-related problems, and learn how to communicate with your health care providers. You will be responsible for getting yourself to and from the community center.” The second vignette was similar in nature, delineating the components of the home-based ASMP Tool Kit intervention (i.e., the ASMP packaged for independent use at home).

Drawing from previous literature that has identified barriers to recruiting and retaining minorities in health-related research, clinical trials, and interventions, a list of 11 home and 13 community potential barriers was developed by the authors. The list of potential barriers to participation (e.g., I believe my lack of trust in the health care system would prevent me from participating) was presented following each vignette. Participants were asked to determine how much they believed each statement using an 11-point Likert scale (range 0–10, where 0 = don't believe at all and 10 = definitely believe). A complete list of potential barriers can be found in Tables 1 and 2.

Table 1. Barriers to group community-based participation*
 African American participants (n = 60)White participants (n = 55)P
  • *

    Values are the mean ± SD. The table includes imputed values.

  • By t-test.

  • Item is reverse coded.

Cost6.01 ± 4.203.72 ± 3.81< 0.01
Difficult to find time5.23 ± 3.774.65 ± 3.740.41
Hard to find the drive or determination4.99 ± 3.704.48 ± 3.770.46
Only if recommended by my doctor4.58 ± 4.153.05 ± 3.750.04
Other health concerns4.27 ± 3.653.29 ± 3.330.14
Lack of trust in the health care system3.95 ± 3.752.59 ± 3.340.04
Difficult to find transportation3.82 ± 4.043.15 ± 3.800.36
Family responsibilities3.40 ± 3.632.38 ± 3.450.13
Only if family and friends thought it was a good idea3.25 ± 3.991.67 ± 3.190.02
Physically unable3.16 ± 3.642.29 ± 3.130.17
Fear of being the only person of my race2.78 ± 3.780.67 ± 1.68< 0.001
My neighborhood is not safe2.45 ± 3.481.38 ± 2.620.07
Believe the program will be helpful1.84 ± 2.702.65 ± 2.910.12
Table 2. Barriers to home-based participation*
 African American participants (n = 60)White participants (n = 55)P
  • *

    Values are the mean ± SD. The table includes imputed values.

  • By t-test.

  • Item is reverse coded.

Cost5.85 ± 4.273.18 ± 3.79< 0.01
Only if recommended by my doctor3.82 ± 3.932.73 ± 3.500.12
Hard to find the drive or determination3.79 ± 3.564.05 ± 3.510.68
Difficult to find time3.73 ± 3.873.31 ± 3.400.54
Other health concerns3.45 ± 3.573.02 ± 3.350.50
Physically unable3.33 ± 3.702.33 ± 3.330.13
Believe the program will be helpful3.16 ± 3.102.04 ± 3.270.06
Lack of trust in the health care system3.02 ± 3.721.38 ± 2.350.01
Only if family and friends thought it was a good idea2.78 ± 3.671.14 ± 2.390.01
Family responsibilities2.58 ± 3.561.60 ± 2.700.10
Fear of being the only person of my race2.16 ± 3.460.67 ± 1.61< 0.01

Arthritis intervention preferences.

Intervention preferences were assessed by asking about components of an existing arthritis intervention and about components that could be included in potential arthritis interventions. The questionnaire developed by the authors was modeled after a questionnaire used to assess preferences in bereavement service utilization (28). The questionnaire assessed preferences for the content (e.g., information about medications, internet sources for arthritis care, and reducing pain; range 0–10, where 0 = not helpful at all and 10 = extremely helpful), structure (e.g., individual, group, family), and delivery (e.g., in-home program, mailed program; range 0–10, where 0 = not at all and 10 = extremely) of arthritis interventions by having the participant rate the value of intervention components on an 11-point Likert scale.

Statistical analysis.

Exploratory data analysis using descriptive statistics was used to assess relationships between variables and identify outliers and missing data. Independent-samples t-tests, chi-square tests, and multiple regression analyses were used to assess race/ethnicity effects and to adjust for additional covariates. Although missing data were minimal in this particular study (i.e., ∼2% of the dependent variables), imputed values for missing dependent variables were used. Significance levels were rounded to 2 digits in the study.

Independent-samples t-tests and chi-square analyses.

Frequencies were examined to assess demographics, barriers, and intervention preferences. First, comparisons between groups (African Americans versus whites) among the study variables were made with independent-samples t-tests and chi-square analyses. While several questions were listed under specific categories (e.g., barriers), each question within the category was treated as an individual dependent variable.

Multiple regression analyses with imputation.

Next, a general linear model approach was used to examine the associations between our independent and outcome variables (i.e., intervention barriers, content preferences, and structure and delivery preferences). Race/ethnicity was our primary independent variable of interest. All of the models were adjusted for the covariates age, education, marital status, sex, income, health status, pain, disability, years with arthritis symptoms, and years with OA diagnosis to account for intergroup variation associated with the covariates. All analyses were generated using SAS software, version 9.2 (29).


Sample characteristics.

Descriptive statistics for the sample (n = 115) are shown in Table 3. Analyses were conducted including only African American (n = 60) and white (n = 55) participants. There is variation in the sample size due to missing values, and the missing data vary per variable per race group. When missing data were present, most dependent variables had ≤3 missing observations. Income for African Americans and whites had the largest amount of missing data (n = 14 and 7, respectively). African American and white participants in this sample had similar education levels, with an average of 13.3 and 13.8 years, respectively. Less than 30% of the participants in each group reported being married. African Americans were more likely to be younger and more likely to have a lower annual household income than whites. African Americans reported significantly worse health in general and more arthritis-related pain than whites. However, there was no significant difference in the level of disability between African American and white participants. African Americans reported both having arthritis symptoms and being diagnosed with OA for a less amount of time than their white counterparts.

Table 3. Sample characteristics
 African American participants (n = 60)White participants (n = 55)P*
  • *

    By t-test or chi-square test.

Age, mean ± SD years67 ± 9.6077 ± 8.24< 0.001
Education, % high school or less47.532.10.10
Marital status, % married20.327.30.38
Female sex, %
Income, %  0.02
Health status, mean ± SD1.8 ± 0.902.2 ± 0.910.01
Pain, mean ± SD6.8 ± 2.375.5 ± 2.42< 0.01
Disability, % yes63.271.70.34
Duration of symptoms, mean ± SD years11.7 ± 9.7918.6 ± 15.22< 0.01
Arthritis diagnosis, mean ± SD years10.3 ± 7.8916.0 ± 14.430.01

Potential barriers.

Participants' perceptions of need or lack thereof for an arthritis self-help program could serve as a barrier to participation. Of those participants who identified themselves as being African American, 90% reported there was a need for an arthritis self-help program. In addition, of those who identified themselves as white, 100% reported there was a need for an arthritis self-help program. Although a statistically significant difference between the groups (P = 0.03) was found using Fisher's exact test, both groups strongly supported the need for an arthritis self-help program.

Table 1 reports the race/ethnicity differences in barriers to participating in the community-based ASMP. There were significant group differences in variables assessing potential barriers to participation in the community-based and home-based versions of the ASMP. Significant differences were found on 5 variables associated with the community-based intervention. African Americans were more likely to report that their lack of trust in the health care system, their fear of being the only person of their race in the program, and the cost of the program would be barriers to participating in comparison to whites. African Americans were also more likely to report that they would only participate in the program if family and friends thought it was a good idea and only if recommended by a doctor. Contrary to expectations, there were no group differences in transportation or family responsibilities as a potential barrier to participation in the community-based intervention. After adjusting for covariates, only one significant main effect remained between African American and white participants. African Americans were more likely to participate only if they were recommended by their doctor in comparison to whites (P = 0.03).

As shown in Table 2, the independent-samples t-test for barriers to participation in an at-home self-management program yielded similar results to that of the group community-based program. Four significant differences were found. African Americans were more likely to report cost, lack of trust in the health care system, fear of being the only person of their race, and lack of endorsement by family/friends as barriers. After adjusting for covariates, no group differences remained significant.

Arthritis intervention preferences.

Table 4 shows results on preferences for intervention content. African Americans were significantly more likely on 20 of the 22 comparisons to report that the variables would be helpful if included in an arthritis intervention. To ensure the differences were true differences and not a result of response bias, additional analyses were also conducted using z-transformed scores. The response results were unaffected by this procedure. Previous research has shown that this is a sound way to address potential response bias concerns (30). After adjusting for covariates, African Americans reported being more likely to prefer an intervention that will teach you about internet sources focused on arthritis care (P = 0.04), teach you about problem-solving skills for arthritis-related problems (P = 0.04), and teach you how to talk to your family and friends about arthritis (P = 0.02).

Table 4. Content and preferences*
 African American participants (n = 60)White participants (n = 55)P
  • *

    Values are the mean ± SD. The table includes imputed values. PCP = primary care provider.

  • By t-test.

Teach you helpful exercises9.23 ± 2.118.02 ± 2.730.01
Get the most out of the health care system9.23 ± 1.578.11 ± 2.33< 0.01
Teach you about your type of arthritis9.20 ± 1.717.81 ± 2.63< 0.01
Reduce your pain9.19 ± 2.088.27 ± 2.440.03
Problem-solving skills for arthritis-related problems9.13 ± 1.887.93 ± 2.810.01
Treatment decisions9.07 ± 1.968.02 ± 2.200.01
Problem-solving skills in general for health-related problems9.01 ± 1.967.91 ± 2.700.01
Nutrition and healthy eating8.95 ± 2.207.55 ± 2.87< 0.01
Improve physical functioning8.95 ± 2.258.33 ± 1.830.11
Arthritis health care options/resources other than PCP8.89 ± 1.927.87 ± 2.670.02
Make meals you eat already in a healthier way8.88 ± 2.427.22 ± 2.94< 0.01
Arthritis medications that are available8.88 ± 2.117.40 ± 3.14< 0.01
Various types of arthritis8.87 ± 2.076.79 ± 3.19< 0.001
Talk to professionals about your arthritis8.75 ± 2.467.29 ± 2.87< 0.01
Information about the use of medications8.62 ± 2.607.49 ± 3.000.03
Decrease negative emotions8.62 ± 2.686.80 ± 3.46< 0.01
Set personal treatment goals and a plan of action8.48 ± 2.346.97 ± 3.13< 0.01
Choose a doctor8.47 ± 2.697.15 ± 3.100.02
Nontraditional treatment decisions8.40 ± 2.767.85 ± 2.840.30
Talk to family and friends about your arthritis8.35 ± 2.775.87 ± 3.47< 0.001
Internet sources for arthritis care8.15 ± 3.065.15 ± 3.92< 0.001
Discrimination in health care8.05 ± 3.165.86 ± 3.97< 0.01

Regarding intervention structure and delivery, the groups differed significantly on 19 of 22 variables (Table 5). In comparison to whites, African Americans were more likely to see the need of the structure and delivery variables for an arthritis intervention. After standardizing the responses to address a potential response bias, the results were unaffected. After adjusting for covariates, African Americans were more likely than their white counterparts to prefer mailed interventions (P < 0.01), bringing a family member (P = 0.02) or friend (P < 0.01) to support and learn with you, interventions to be given by someone of the same race (P < 0.01), interventions given by someone of the same sex (P < 0.001), an intervention given at the local church (P = 0.01), an intervention given at a local health clinic (P < 0.01), interventions limited by sex (P = 0.03), interventions providing child services (P < 0.01), and interventions that include people of the same race (P < 0.001).

Table 5. Structure and delivery*
 African American participants (n = 60)White participants (n = 55)P
  • *

    Values are the mean ± SD. The table includes imputed values.

  • By t-test.

Mailed8.57 ± 2.415.84 ± 3.60< 0.001
Include discussion with other people who have arthritis8.20 ± 2.896.47 ± 3.28< 0.01
Offer a class to a group teaching you to manage arthritis7.90 ± 3.306.41 ± 3.400.02
Given at a community or senior center7.55 ± 3.395.49 ± 3.62< 0.01
Bring a friend for support and to learn with you7.54 ± 3.264.05 ± 3.55< 0.001
Led by someone who has arthritis themselves7.52 ± 3.196.68 ± 2.900.14
Videotape/DVD7.45 ± 3.545.76 ± 3.710.01
Given in your neighborhood7.43 ± 3.606.24 ± 3.520.08
Cassette tape/CD7.31 ± 3.655.25 ± 3.82< 0.01
Given in an instructional or lecture format7.28 ± 3.545.75 ± 3.150.02
Given at the local Arthritis Foundation7.17 ± 3.474.69 ± 3.35< 0.001
Material on spiritual beliefs7.17 ± 3.524.53 ± 3.76< 0.001
Given at a local health clinic7.04 ± 3.184.07 ± 3.17< 0.001
Bring a family member for support and to learn with you6.95 ± 3.864.00 ± 3.55< 0.001
Include people in your same age group6.63 ± 4.444.76 ± 3.720.02
Given at a local church6.45 ± 3.853.00 ± 3.16< 0.001
Include people of your race5.60 ± 4.472.55 ± 3.33< 0.001
Be given by someone of the same sex5.42 ± 3.952.05 ± 3.04< 0.001
Provide child care services5.25 ± 4.331.51 ± 2.67< 0.001
Be limited by condition5.13 ± 4.194.85 ± 3.440.69
Given by someone of the same race4.73 ± 4.230.94 ± 1.93< 0.001
Be limited by sex4.19 ± 3.991.38 ± 2.38< 0.001


The results suggest a number of differences between African Americans and whites with OA in their preferences and reported barriers to community- and home-based arthritis interventions. In general, African American participants were more likely to report being affected by barriers (e.g., cost, trust, race, lack of doctor's recommendation) to participating in both a community- and home-based intervention. Physicians may be a very important facilitator in getting African Americans to participate in behavioral arthritis interventions. This finding is interesting in that although we are not aware of any studies examining the doctor's recommendation as a barrier to participating in behavioral arthritis interventions, researchers who have examined this relationship in the context of arthritis joint replacement found that African Americans were less likely than whites to be willing to consider joint replacement, even with a recommendation from their doctor (31). It may be that there is some sort of skepticism around participating in any unfamiliar intervention procedure; however, a doctor's recommendation may facilitate participation in a health care option that is behavioral and not surgical.

Consistent with prior research, African Americans were found to express distrust in the health care system and to report this as a barrier (14). A lack of trust could hinder optimal health care. Cultural mistrust, which is the mistrust of whites by African Americans in politics, interpersonal relationships, education and training, and business and work (32), could be a potential explanation as to why African Americans were more likely to report distrust in the health care system as a barrier to participating. Although this study did not examine cultural mistrust and only examined trust as a barrier to participation with a single-item question, it does suggest that more research is warranted to analyze the relationship between trust and participation in the ASMP for African Americans with OA.

Contrary to our predictions, there were no significant differences between African Americans and whites in reporting that transportation or family responsibilities served as a barrier to participating in the arthritis programs. Many participants in this study were recruited from their church, senior housing community center, and multipurpose senior center, so our recruitment approach may have been biased toward individuals who had the ability to attend and participate in other programs. Future research is warranted with a more representative sample to fully understand how barriers may impact the utilization of an arthritis health care program.

The results clearly demonstrate that there are race differences in preferences regarding the content, structure, and delivery of an arthritis self-management intervention. After controlling for covariates, African Americans reported being more likely to prefer interventions that teach about arthritis-related internet sources, how to solve arthritis-related problems, and how to talk to family and friends about their condition. In addition, an intervention that provides child care services, includes instructors of the patient's race or sex and participants of the same race or sex, allows the patient to bring a family member or friend, is offered at a local church or clinic, or could be mailed were all components of an arthritis intervention that were preferred by African American participants. Such findings suggest that personal and social factors, some of which could be readily addressed, play a significant role in the appeal of an arthritis intervention. For example, given the fact that communities and family members often serve as gatekeepers to participation in health-related research or programs (13), it is not surprising that African Americans would prefer a program that was offered at a local church or clinic and allowed them to bring a family member or a friend.

Cultural competence is a broad concept describing the importance of understanding social and cultural influences on health beliefs and behaviors and the ability for practitioners or health care providers to use this knowledge habitually to improve health care programs/services and eliminate health disparities (15, 33–36). Research has identified various strategies (e.g., cultural awareness, cultural tailoring, culturally based, culturally relevant, culturally appropriate, culturally sensitive, cultural leverage) through which cultural competency can be reached (15, 33, 34). This particular study has drawn from 2 of those constructs (i.e., cultural sensitivity and cultural tailoring). Based on the findings of this study, we believe that there is not a specific need to culturally tailor (i.e., design an intervention only for a specific racial/ethnic group, e.g., an African American ASMP) to replace existing arthritis self-management programs because African Americans and whites appear to agree on many facets of what makes a program desirable. However, practical issues should be addressed and systematic adaptations should be made to make these program more culturally sensitive (i.e., taking into account the shared preferences, beliefs, values, experiences, and practices in the design and delivery of the program) and appealing for African Americans. For instance, personal and social factors could be addressed by ensuring that those trained to deliver the ASMP are from diverse backgrounds, and that the programs are offered in areas that would attract both white and minority participants. The findings of our research are in line with findings from the research done by Goeppinger et al (16), in that there is a need to strengthen the content of the ASMP by including information on cultural dimensions (e.g., the importance of social support). It is likely that adaptations have begun to take place (16, 17) but are not systematically documented in the arthritis intervention literature. It is suggested that adaptations be documented and tested to determine whether such changes yield benefits in acceptability and successful recruitment and retention.

Our study had several limitations. Individuals in this study were asked to self-report a doctor-diagnosed case of OA. While self-report has been used as a method to determine national prevalence rates for arthritis of any type, the use of both clinical criteria and radiologic scales would be the ideal method to confirming an OA diagnosis. Using clinical criteria and radiologic scales would be considered a gold standard and would strengthen the validity of the research. These findings need to be examined to see if they are relevant to other interventions (e.g., cognitive–behavioral therapy, exercise, education) and with larger samples. Because our study was predominately female, sex differences should also be explored.

The American College of Rheumatology's most recent nonpharmacologic recommendations for individuals with OA include self-management and physical activity arthritis interventions (37). In addition, the CDC has placed an emphasis on identifying evidence-based arthritis interventions (38) and the importance of increasing the participation and availability of the interventions (39), but did not emphasize addressing cultural diversity and health disparities. This issue deserves greater attention in arthritis interventions. Reducing health and health care disparities has been identified as a national priority and has been the focus of the CDC, Department of Health and Human Services, and National Arthritis Action Plan (40, 41). We believe that our results can be useful in designing culturally sensitive and effective arthritis interventions. The CDC has also made an effort to address public health concerns around arthritis in general by funding state health departments to offer self-management and physical activity interventions (42). While these are laudable actions, efforts should be made to determine how funding coupled with a culturally sensitive approach could make a positive impact. Culturally sensitive interventions must promote accessibility and be designed in a manner taking into consideration the values, preferences, and needs of diverse populations. It is vital that we make strides toward being more culturally sensitive and take all steps possible to broaden the appeal of arthritis interventions.


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. Dr. Mingo 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. Mingo, McIlvane, Haley.

Acquisition of data. Mingo, McIlvane.

Analysis and interpretation of data. Mingo, McIlvane, Jefferson, Edwards, Haley.


The authors would like to gratefully acknowledge Drs. Tamara A. Baker, Francis J. Keefe, Victor Molinari, and Brent J. Small for serving as dissertation advisors for the project from which this article was derived; Shanti Herpal for data collection and data cleaning; and Ronna Metacalf and Billi Griffin for their invaluable assistance in recruiting volunteers for the research study.