Effectiveness of two evidence-based programs in participants with arthritis: Findings from the active for life initiative

Authors


  • The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Robert Wood Johnson Foundation or institution affiliated with the authors.

Abstract

Objective

Arthritis is the leading cause of disability in the US. Strong evidence suggests that physical activity (PA) is beneficial to those with arthritis. This study examined whether changes in PA and PA-related outcomes from two general evidence-based PA programs (Active Choices [AC] and Active Living Every Day [ALED]) differed in participants with and without arthritis.

Methods

Active for Life was a 4-year, multicenter, translational initiative that evaluated the effects of AC and ALED on PA and PA-related outcomes. Participants self-reported arthritis, PA, depressive symptoms, perceived stress, and satisfaction with body appearance and function. A subset of participants completed functional fitness tests.

Results

Participants (n = 2,413 AC, n = 3,191 ALED) completed at least one outcome measure; 619 ALED participants completed at least one functional fitness test. Significant improvements in all PA and PA-related outcomes were seen for people with and without arthritis with the exception of depressive symptoms and perceived stress, which only improved in ALED participants. In general, effect sizes were similar for those with and without arthritis.

Conclusion

ALED and AC, both general behavioral PA programs, produced positive and meaningful changes in PA and a number of PA-related behaviors in participants with and without arthritis, pointing to the potential appropriateness of these evidence-based programs for people with arthritis.

INTRODUCTION

An estimated 50 million adults in the US had self-reported doctor-diagnosed arthritis in 2007–2009 (1). By 2030, this number is projected to increase to 67 million (2). Among adults ages ≥65 years, 50% reported an arthritis diagnosis in 2007–2009, and of these, 45.4% reported arthritis- attributable activity limitations (1). Arthritis and other rheumatic conditions remain the leading cause of disability among adults in the US (3).

There is strong evidence supporting the beneficial effects of physical activity (PA) for individuals with arthritis (4). The American College of Rheumatology highlights the importance of PA for the management of arthritis (5, 6). Despite this, a majority of individuals with arthritis do not engage in sufficient amounts of PA necessary to experience such benefits, and are less active than those without arthritis (7). The Centers for Disease Control and Prevention (CDC) Arthritis Program recommends 6 evidence-based PA programs for individuals with arthritis (8). The two initial programs (i.e., Arthritis Foundation Aquatics Program, People with Arthritis Can Exercise [9]) were developed specifically for people with arthritis. However, there is growing evidence that programs not designed exclusively for people with arthritis may be beneficial and appropriate for people with arthritis (8).

Two non–arthritis-specific evidenced-based PA programs, Active Living Every Day (ALED) and EnhanceFitness, were recently added to the CDC Arthritis Program's list of recommended interventions. Such programs are appealing in that a more “generic” intervention can be offered to the entire general population, yet still provide benefits to those with arthritis. This may be particularly advantageous for organizations that do not have resources available to offer a wide range of disease-specific interventions (8).

Active for Life (AFL) was a translational initiative that successfully implemented two evidence-based PA behavior modification programs (Active Choices [AC] and ALED) into “real-world” community settings with diverse populations. AFL resulted in significant improvements in PA and a number of other PA-related variables (10). A large proportion (64%) of participants in AFL self-reported doctor-diagnosed arthritis. The purpose of this study was to examine whether changes in PA and PA-related outcomes in participants in AC and ALED, both general PA programs, differed in participants with and without arthritis.

Significance & Innovations

  • General physical activity programs (i.e., Active Choices and Active Living Every Day) may be appropriate for people with arthritis.

  • Improvements in physical activity, physical functioning, and other physical activity–related behaviors were similar for those with and without arthritis.

MATERIALS AND METHODS

The methodology and major findings of AFL are described in more detail elsewhere (10, 11). AFL was a 4-year initiative that evaluated the effects of two evidence-based behavioral PA interventions, ALED and AC, on self-reported PA and other PA-related outcomes. The interventions were implemented by 9 public health, senior service, health care, and grassroots organizations across 12 sites. The programs were typically implemented by non–health care professional staff members from each organization (11).

Interventions.

In AFL, AC was a 6-month telephone-based program (12–14). Participants received an initial face-to-face orientation, followed by up to 8 one-on-one telephone counseling calls. The counseling was tailored to the participant's readiness for change (15) and emphasized key behavioral constructs in the social cognitive theory (16). During each call, the health educator discussed the participant's current level of PA, their readiness for change, a specific behavior change principle, and modified PA goals, if necessary.

ALED is a 20-week lifestyle PA intervention (17, 18). Each week, participants attended a 60-minute group-based session that focused on behavior change principles consistent with the social cognitive theory (16). The goal of ALED was to help participants accumulate a minimum of 30 minutes of PA on most days of the week. Participants were encouraged to provide support and share successes and challenges with one another.

Both AC and ALED are behavioral programs that do not include structured PA as a part of the intervention sessions. In AFL, 4 organizations implemented AC and 5 organizations implemented ALED.

Participants.

The 9 participating organizations recruited participants through their networks and from the community at large. Recruitment strategies were tailored to the community they targeted. The most successful recruitment efforts were onsite staff recruitment, presentations at the facilities, direct mail, flyers, newsletters, and local media. Participants were ages ≥50 years, sedentary or underactive (≤2 days/week and <120 minutes/week), and free of medical conditions or disabilities that required higher levels of supervision, as determined by lead staff (e.g., unstable angina) (10).

Research design and procedures.

This study used a pre-post design, with data collected from 2003 to 2007. All of the participants completed an informed consent form approved by the Institutional Review Boards at the University of South Carolina and Texas A&M Health Science Center, and review boards or legal departments of the lead organizations.

Self-report outcomes.

Comprehensive surveys were administered to all participants in year 1 (∼100 per site) and to the first 100 participants per site in years 3 and 4. This decision was made a priori, as it was not necessary for adequate statistical power to collect survey data on all participants across all years. A shortened version of the survey that collected several key outcomes was given to the rest of the participants in years 3 and 4. Although the intervention was delivered during year 2, outcome data were not collected.

After year 3, the lead organizations taking part in AFL requested that the ALED program be shortened to 12 weeks. Although the program was shortened in year 4 (i.e., lasted 12 weeks, not 20 weeks), steps were taken to ensure that essential elements were maintained (10). Participants taking part in the ALED intervention during year 4 completed a 12- (all short versions) and 20-week posttest survey. This study only uses the 20-week data from year 4.

Sociodemographic, health-related, and other variables.

Participants self-reported height and weight to compute body mass index (BMI), date of birth, sex, marital status, race/ethnicity, and education level. Self-reported absence or presence of arthritis was assessed by asking participants, “Have you ever been told by a doctor, nurse, or other health professional that you had arthritis?” Dropout rates were tracked by AFL staff.

PA.

The 41-item Community Healthy Activities Model Program for Seniors questionnaire assessed PA (19). This self-report measure assesses the frequency and duration of various activities typically undertaken by older adults for exercise, activities undertaken in the course of their day that are physical in nature, and recreational activities that provide PA. Hours/weeks spent in moderate- to vigorous-intensity PA (MVPA; metabolic equivalents value of ≥3.0 [20, 21]) were calculated. This measure is valid (22), has acceptable test–retest reliability (22), and is sensitive to change (13, 19, 21, 23, 24).

PA was also assessed using a secondary 3-item measure from the Behavioral Risk Factor Surveillance System that assessed participation, frequency, and duration of moderate intensity PA (25). Participants were classified as meeting (≥30 minutes/day, ≥5 days/week) or not meeting PA recommendations (26).

Depressive symptoms.

The 10-item Center for Epidemiologic Studies Depression Scale (27–29) assessed symptoms of depression during the past week. Scores range from 0 to 30, with higher scores indicating greater depressive symptoms.

Perceived stress.

A 4-item version of the Perceived Stress Scale (30, 31) measured the degree to which situations in one's life are appraised as stressful. Scores range from 0 to 16, with higher scores indicating greater perceived stress.

Body satisfaction.

Satisfaction over the past 4 weeks with 9 aspects of body appearance and function was assessed on a 7-point scale. Two subscales were derived from factor analysis (32): satisfaction with body appearance (3 items) and satisfaction with body function (6 items). Scores range from −9 to +9 (appearance) and from −18 to +18 (function), with higher scores indicating greater satisfaction.

Physical performance outcomes.

One ALED site administered 4 performance-based physical functioning tests. Each test was administered during all 4 years of AFL with the exception of the Chair Sit-and-Reach Test, which was only administered during year 3. The methods used in administering physical functioning measures are reported in greater detail elsewhere (33). Lead staff was trained by an expert in physical functioning in how to administer the performance-based measures. The test manual and videotape provided by Human Kinetics were used to aid trainings (34).

30-second Chair Stand Test.

The 30-second Chair Stand Test assessed lower body strength (34). Participants sat in the middle of a 17-inch high chair with their back straight, feet flat on the floor, and arms crossed at the wrist and held against the chest. On the signal “go,” participants rose to a full stand and returned to a fully seated position. Participants completed as many stands as possible in 30 seconds.

Eight Foot Up-and-Go Test.

The Eight Foot Up-and-Go Test assessed agility and dynamic balance (34). Participants sat in the middle of a 17-inch high chair with their back straight, feet flat on the floor, and hands on the thighs. On the signal “go,” participants got up from the chair, walked as quickly as possible around a cone placed 8 feet from the chair, and sat back down. Measured in seconds, a shorter time is better.

Chair Sit-and-Reach Test.

The Chair Sit-and-Reach Test assessed lower body (primarily hamstring) flexibility (34). Participants sat on the front edge of a 17-inch high chair and extended one leg out as straight as possible in front of the hip, with the foot flexed and heel resting on the floor. Participants slowly bent forward at the hip joint, reaching as far forward as possible toward (or past) the toes. The distance from the tips of the middle fingers to the top of the shoe was measured. Measured in inches, a longer distance is better.

30-Foot Walk Test.

The 30-Foot Walk Test assessed gait speed (35). On the signal “ready, go,” participants started walking at their normal speed. The stopwatch did not start until the participant crossed the 10-foot line, and was subsequently stopped when the participant crossed the 40-foot line. The time, in seconds, to complete the test was converted to gait speed (meters/second). A faster gait speed is better.

Impairment status.

Rikli and Jones' (34) age- and sex-based normative data were used to classify participants as impaired (less than or equal to twenty-fifth percentile) or not impaired at baseline and postprogram on the 30-Second Chair Stand Test, the Eight Foot Up-and-Go Test, and the Chair Sit-and-Reach Test. Because normative data for the 30-Foot Walk Test are not available from Rikli and Jones (34), data from a recent meta-analysis (36) were used to classify participants' gait speed as impaired (less than or equal to twenty-fifth percentile) or not impaired at baseline and postprogram.

Statistical analyses.

Participants who completed any baseline outcome measure (or physical functioning test for subanalyses) and had complete model covariates were included in the baseline demographic analyses. Differences in postprogram outcome measure completion and dropout rates, as well as demographic, health-related, and other key outcome variables, among participants with and without arthritis were examined with the chi-square test or analyses of variance. Square root transformations were conducted on both baseline and postprogram scores for MVPA and the Eight Foot Up-and-Go Test, as these measures were skewed and/or had high kurtosis. The transformations successfully corrected the problem(s).

Repeated-measures analyses of covariance that controlled for site clustering (using SAS PROC MIXED) were conducted to examine whether changes in outcomes from baseline to postprogram differed by arthritis status (yes/no). Changes in dichotomous outcome variables were assessed using SAS PROC GLIMMIX. Arthritis × time interactions were tested in each model, and all models controlled for sex, age, race/ethnicity, and BMI. Using the adjusted means and the unadjusted baseline SD, effect sizes (d = [postprogram mean − baseline mean] / baseline SD) were computed for each of the continuous outcome variables (37). Using Cohen's effect sizes (38), d = 0.2 was considered small, d = 0.5 was considered medium, and d = 0.8 was considered large.

RESULTS

A total of 5,604 participants completed at least one baseline outcome measure (n = 2,413 AC, n = 3,191 ALED). Figure 1 shows the number of participants with baseline and followup data for each intervention group by arthritis status. Self-reported arthritis was reported in 56% (AC) and 61% (ALED) of participants. Table 1 shows the baseline demographic, health-related, and outcome measures for each intervention group, and for those with and without arthritis separately. In both groups, those with arthritis were more likely to be women and have higher BMIs; were older, less educated, less likely to be married, more depressed, and less satisfied with their body function; and were more likely to have higher levels of perceived stress. In the AC group, participants with arthritis engaged in less PA, were less likely to be African American, and were less likely to be Hispanic. In the ALED group, participants with arthritis were less likely to meet PA recommendations and less likely to be satisfied with their body appearance.

Figure 1.

Sample size at baseline and followup, by program and arthritis status. a = participants with at least 1 outcome variable plus all covariates at baseline; b = physical functioning subsample; c = post data on at least 1 outcome variable; ALED = Active Living Every Day; AC = Active Choices; PF = physical functioning.

Table 1. Baseline characteristics of participants in AC and ALED by arthritis status*
 ACALED
No arthritisArthritisNo arthritisArthritis
  • *

    Values are the number (percentage) unless otherwise indicated. AC = Active Choices; ALED = Active Living Every Day; BMI = body mass index; MVPA = moderate- to vigorous-intensity PA; PA = physical activity.

  • Significant difference between participants with and without arthritis within AC and ALED.

  • Significant difference between participants with and without arthritis within AC.

  • §

    Significant difference between participants with and without arthritis within ALED.

  • Significant difference between participants with and without arthritis within AC and ALED. Possible range: 0–30.

  • #

    Significant difference between participants with and without arthritis within AC and ALED. Possible range: 0–16.

  • **

    Significant difference between participants with and without arthritis within ALED. Possible range: −9 to 9.

  • ††

    Significant difference between participants with and without arthritis within AC and ALED. Possible range: −18 to 18.

Sex    
 Male252 (23.9)225 (16.6)273 (21.9)265 (13.7)
 Female802 (76.1)1,134 (83.4)976 (78.2)1,677 (86.4)
Age, no. patients/mean ± SD years1,054/64.0 ± 9.71,359/67.2 ± 9.61,249/69.2 ± 9.81,942/71.3 ± 8.9
 50–64537 (53.3)505 (39.5)381 (33.2)418 (23.2)
 65–74306 (30.4)447 (35.0)391 (34.0)693 (38.4)
 ≥75164 (16.3)326 (25.5)377 (32.8)693 (38.4)
Race/ethnicity    
 White429 (40.7)558 (41.1)816 (65.3)1,256 (64.7)
 African American419 (39.8)590 (43.4)364 (29.1)592 (30.5)
 Hispanic127 (12.1)119 (8.8)40 (3.2)45 (2.3)
 Other79 (7.5)92 (6.8)29 (2.3)49 (2.5)
Education    
 Less than high school graduate118 (11.4)203 (15.1)108 (8.7)261 (13.5)
 High school graduate or GED183 (17.7)324 (24.1)326 (26.3)521 (27.0)
 Some college or more733 (70.9)820 (60.9)804 (64.9)1,150 (59.5)
Marital status    
 Married/partnered468 (44.8)505 (37.6)579 (46.9)784 (40.9)
 Not married577 (55.2)839 (62.4)655 (53.1)1,135 (59.2)
BMI, no. patients/mean ± SD kg/m21,054/29.8 ± 6.81,359/32.1 ± 7.91,249/28.3 ± 6.21,942/30.5 ± 6.9
 Underweight15 (1.4)7 (0.5)18 (1.4)23 (1.2)
 Normal weight243 (23.1)244 (18.0)360 (28.8)386 (19.9)
 Overweight351 (33.3)355 (26.1)472 (37.8)605 (31.2)
 Obese445 (42.2)753 (55.4)399 (32.0)928 (47.8)
MVPA, no. patients/mean ± SD hours/week547/2.9 ± 3.8695/2.4 ± 3.4688/2.8 ± 4.1963/2.4 ± 3.4
PA recommendations§    
 Does not meet979 (93.7)1,246 (92.6)1,022 (83.6)1,662 (87.1)
 Meets66 (6.3)99 (7.4)201 (16.4)247 (12.9)
Depressive symptoms, no. patients/mean ± SD540/5.7 ± 5.0690/6.8 ± 5.2684/4.7 ± 4.7959/6.5 ± 5.2
Perceived stress, no. patients/mean ± SD#501/4.3 ± 3.0666/4.8 ± 3.1687/4.3 ± 3.1961/4.7 ± 3.2
Satisfaction with body appearance, no. patients/mean ± SD**502/−2.7 ± 5.4667/−3.1 ± 5.2688/−1.4 ± 5.4964/−2.8 ± 5.5
Satisfaction with body function, no. patients/mean ± SD††1,006/−0.7 ± 10.41,326/−3.4 ± 9.71,244/0.7 ± 10.41,936/−2.6 ± 10.1

At least one physical functioning test was completed by 619 participants at baseline (from one ALED site). Of this sample, 68% self-reported arthritis. Table 2 shows the demographic, health-related, and outcome measures for those with and without arthritis separately. Participants with arthritis were more likely to be women, performed poorer on the Eight Foot Up-and-Go Test and 30-Foot Walk Test, and were more likely to be classified as impaired on the Eight Foot Up-and-Go Test than those without arthritis.

Table 2. Baseline characteristics for participants completing functional fitness testing by arthritis status*
 Active Living Every Day
No arthritisArthritis
  • *

    Values are the number (percentage) unless otherwise indicated. BMI = body mass index.

  • Significant difference between participants with and without arthritis.

Sex  
 Male40 (20.5)51 (12.0)
 Female155 (79.5)373 (88.0)
Age, no. participants/ mean ± SD years195/70.2 ± 9.3424/71.4 ± 8.7
 50–6456 (28.7)103 (24.3)
 65–7477 (39.5)166 (39.2)
 ≥7562 (31.8)155 (36.6)
Race/ethnicity  
 White162 (83.1)362 (85.4)
 Nonwhite33 (16.9)62 (14.6)
Education  
 Less than high school graduate20 (10.3)42 (10.0)
 High school graduate or GED55 (28.2)143 (33.9)
 Some college or more120 (61.5)237 (56.2)
Marital status  
 Married/partnered82 (42.3)184 (43.8)
 Not married112 (57.7)236 (56.2)
BMI, no. participants/ mean ± SD kg/m2195/30.9 ± 6.9424/32.1 ± 7.2
 Underweight3 (1.6)5 (1.2)
 Normal weight43 (22.5)67 (16.3)
 Overweight62 (32.5)122 (29.6)
 Obese83 (43.5)218 (52.9)
Income  
 <$30,00058 (63)83 (50.6)
 $30,000–$59,99923 (25)61 (37.2)
 ≥$60,00011 (12)20 (12.2)
30-second Chair Stand Test, no. participants/ mean ± SD155/10.9 ± 2.7274/10.4 ± 3.3
 Impaired85 (54.8)164 (59.9)
 Not impaired70 (45.2)110 (40.2)
Eight Foot Up-and-Go Test, no. participants/ mean ± SD seconds190/6.9 ± 1.8392/7.8 ± 2.7
 Impaired78 (41.7)204 (55.1)
 Not impaired109 (58.3)166 (44.9)
Chair Sit-and-Reach Test, no. participants/mean ± SD inches27/−2.0 ± 3.671/−1.3 ± 3.6
 Impaired15 (55.6)34 (47.9)
 Not impaired12 (44.4)37 (52.1)
30-Foot Walk Test, no. participants/mean ± SD meters/second191/1.1 ± 0.2414/1.1 ± 0.2
 Impaired101 (53.7)243 (61.7)
 Not impaired87 (46.3)151 (38.3)

Postmeasure followup and dropout rates are shown in Table 3. There were no significant differences in the percentages of participants with and without arthritis who completed postmeasures or dropped out, for any intervention group.

Table 3. Followup and dropout rates by intervention group among participants with and without arthritis
 No arthritis, no. (%)Arthritis, no. (%)P
Completed ≥1 posttest measure   
 Active Choices  0.36
  Yes657 (62.3)822 (60.5) 
  No397 (37.7)537 (39.5) 
 Active Living Every Day  0.96
  Yes880 (70.5)1,370 (70.6) 
  No369 (29.5)572 (29.5) 
 Physical functioning subsample  0.37
  Yes121 (62.1)247 (58.3) 
  No74 (38.0)177 (41.8) 
Dropped out   
 Active Choices  0.51
  Yes74 (7.0)105 (7.7) 
  No980 (93.0)1,254 (92.3) 
 Active Living Every Day  0.11
  Yes158 (12.7)210 (10.8) 
  No1,091 (87.4)1,732 (89.2) 
 Physical functioning subsample  0.94
  Yes23 (11.8)52 (12.3) 
  No172 (88.2)372 (87.7) 

AC.

The adjusted baseline and postprogram mean scores for all outcomes, as well as the calculated effect sizes for participants with and without arthritis in the AC intervention, are shown in Table 4. There was a significant arthritis × time interaction for MVPA (P = 0.01), meeting PA recommendations (P = 0.047), and satisfaction with body appearance (P = 0.04). Significant improvements were seen in all outcomes for both participants with and without arthritis; however, those without arthritis improved significantly more than those with arthritis. There were no other significant arthritis × time interactions for any of the other variables, indicating no difference in change over time for participants with and without arthritis. There was a significant main effect for satisfaction with body function (P = 0.001), indicating improvements over time regardless of arthritis status. There were no significant time effects for depressive symptoms (P = 0.16) or perceived stress (P = 0.27).

Table 4. Changes in study outcomes among participants in Active Choices with and without arthritis*
 No arthritisArthritisP (group × time)P (time)
PretestPosttestEffect size, dPretestPosttestEffect size, d
  • *

    Means are adjusted for race/ethnicity, age, body mass index, and sex. Effect sizes are computed using adjusted baseline and postprogram means and baseline unadjusted SDs. The physical activity (PA) measure was skewed and a square root transformation was conducted and used in all change analyses and in the effect size calculations. However, the untransformed means are shown using the original scales of measurement. MVPA = moderate- to vigorous-intensity PA; CHAMPS = Community Healthy Activities Model Program for Seniors; N/A = not applicable.

MVPA (CHAMPS), mean2.926.230.792.805.120.620.010.0002
Meets PA recommendations, %9.2437.11N/A10.4831.40N/A0.0470.007
Depressive symptoms, mean5.765.18−0.126.866.74−0.020.120.16
Perceived stress, mean4.324.19−0.054.914.83−0.030.790.27
Satisfaction with body appearance, mean−2.39−0.040.43−2.38−0.700.320.04< 0.0001
Satisfaction with body function, mean0.055.780.55−2.562.870.560.580.001

ALED.

The adjusted baseline and postprogram mean scores for all outcomes, as well as the calculated effect sizes for participants with and without arthritis in the ALED intervention, are shown in Table 5. There was a significant arthritis × time interaction for depressive symptoms (P = 0.008), where depressive symptoms significantly decreased in participants with arthritis but did not change in participants without arthritis. There were no other significant arthritis × time interactions for any of the other variables, indicating no difference in change over time for participants with and without arthritis. There were significant main effects for MVPA (P < 0.0001), meeting PA recommendations (P = 0.0006), perceived stress (P = 0.0001), satisfaction with body appearance (P < 0.0001), and satisfaction with body function (P < 0.0001), indicating significant improvements over time regardless of arthritis status.

Table 5. Changes in study outcomes among participants in Active Living Every Day with and without arthritis*
 No arthritisArthritisP (group × time)P (time)
PretestPosttestEffect size, dPretestPosttestEffect size, d
  • *

    Means are adjusted for race/ethnicity, age, body mass index, and sex. Effect sizes are computed using adjusted baseline and postprogram means and baseline unadjusted SDs. The physical activity (PA) measure was skewed and a square root transformation was conducted and used in all change analyses and in the effect size calculations. However, the untransformed means are shown using the original scales of measurement. MVPA = moderate- to vigorous-intensity PA; CHAMPS = Community Healthy Activities Model Program for Seniors; N/A = not applicable.

MVPA (CHAMPS), mean3.246.060.633.155.670.680.75< 0.0001
Meets PA recommendations, %17.9349.04N/A16.0844.81N/A0.560.0006
Depressive symptoms, mean4.664.28−0.086.305.31−0.190.0080.005
Perceived stress, mean4.123.36−0.244.553.92−0.190.450.0001
Satisfaction with body appearance, mean−1.010.810.33−1.520.380.350.74< 0.0001
Satisfaction with body function, mean1.747.600.57−1.084.980.600.64< 0.0001

Physical functioning.

The adjusted baseline and postprogram mean scores for all physical functioning outcomes, as well as the calculated effect sizes for participants with and without arthritis from the one ALED site that administered functional fitness tests, are shown in Table 6. There was a significant arthritis × time interaction for the 30-second Chair Stand Test (P = 0.004) and the percentage classified as impaired on the 30-second Chair Stand Test (P = 0.007). There were significant improvements in the 30-second Chair Stand Test and the percentage of participants classified as impaired in the 30-second Chair Stand Test for participants with and without arthritis; however, those without arthritis improved significantly more than those with arthritis. There were no other significant arthritis × time interactions for any of the other physical functioning variables, indicating no difference in change over time for participants with and without arthritis. There was a significant main effect for the Eight Foot Up-and-Go Test (P < 0.0001), the Chair Sit-and-Reach Test (P = 0.01), the 30-Foot Walk Test (P < 0.0001), and the percentage classified as impaired on the Eight Foot Up-and-Go (P < 0.0001), Chair Sit-and-Reach (P = 0.04), and 30-Foot Walk (P < 0.0001) tests, indicating significant improvements over time regardless of arthritis status.

Table 6. Changes in physical functioning outcomes among participants completing function fitness tests*
 No arthritisArthritisP (group × time)P (time)
PretestPosttestEffect size, dPretestPosttestEffect size, d
  • *

    Means are adjusted for race/ethnicity, age, body mass index, and sex. Effect sizes are computed using adjusted baseline and postprogram means and baseline unadjusted SDs. A square root transformation was conducted for the Eight Foot Up-and-Go Test and was used in all change analyses and effect size calculations. However, the untransformed means are shown using the original scales of measurement. N/A = not applicable.

30-second Chair Stand Test, mean no.10.912.90.7510.511.60.350.004< 0.0001
 Impaired, %59.926.9N/A66.148.3N/A0.007< 0.0001
 Not impaired, %40.173.1N/A33.951.7N/A0.007< 0.0001
Eight Foot Up-and-Go Test, mean seconds6.86.4−0.267.46.8−0.240.36< 0.0001
 Impaired, %47.635.4N/A59.146.6N/A0.89< 0.0001
 Not impaired, %52.464.6N/A40.953.4N/A0.89< 0.0001
Chair Sit-and-Reach Test, mean inches−2.2−1.00.34−1.3−0.20.320.920.01
 Impaired, %48.633.7N/A38.727.6N/A0.770.04
 Not impaired, %51.466.3N/A61.372.4N/A0.770.04
30-Foot Walk Test, mean meters/second1.11.20.321.11.10.280.89< 0.0001
 Impaired, %60.450.7N/A67.754.6N/A0.50< 0.0001
 Not impaired, %39.649.3N/A32.345.4N/A0.50< 0.0001

DISCUSSION

ALED and AC, both general PA programs, produced positive and meaningful changes in PA and a number of PA-related behaviors in participants with arthritis, pointing to the potential appropriateness of these evidence-based programs for people with arthritis. With the exception of depressive symptoms and perceived stress in the AC program (which did not improve in those without arthritis either), significant improvements in all PA, PA-related, and physical functioning outcomes were seen for people with arthritis. With a few exceptions, improvements (effect sizes) were similar in magnitude to those without arthritis.

Although PA is beneficial to people with arthritis, the symptoms of arthritis often deter individuals from engaging in regular PA (39, 40). Medium to large increases (effects sizes 0.62–0.68) in PA were seen for individuals with arthritis in both intervention groups. The percentage of participants with arthritis meeting PA recommendations also nearly tripled in both groups. Furthermore, there were no negative outcomes reported, lending support to the safety of ALED and AC for individuals with arthritis. It is difficult to compare our findings to other studies due to differences in study design. However, thus far, findings from arthritis-specific community-based PA interventions have found mixed results (41–43).

Depression is common among people with arthritis (44, 45). Depressive symptoms did not change for any participants in the AC program, and they did significantly decrease among ALED participants with arthritis (small effect size), but not among those without arthritis. Although participants with arthritis were more depressed at baseline than those without arthritis and therefore had more room for improvement, depression scores in all participants were quite low. ALED is a group-based program where social interactions occur often, and support from other group members is encouraged. Participants may have found reassurance in talking to other participants who were attempting to make similar lifestyle changes. This may have been especially useful and important to participants with arthritis who may experience unique challenges and barriers when becoming active. Similar to our findings for AC, no changes in depressive symptoms have been reported in other community-based arthritis-specific intervention studies (41, 42); however, as stated earlier, caution should be taken when comparing findings due to differences in study design.

Perceived stress (ALED only) and satisfaction with body appearance and function also significantly improved in participants with arthritis (small to medium effect sizes). These variables have not been measured in other studies targeting people with arthritis; therefore, it is hard to make comparisons across studies. However, these measures are important and pertinent to individuals with arthritis. Being satisfied with one's body appearance and function might be particularly important for people with arthritis, who may be more likely to experience problems with energy, stamina, strength, and functioning.

Nearly 40% of adults with arthritis report arthritis-attributable activity limitations (46). Functional declines are common in people with arthritis (47–50), although PA may help to prevent them. Significant improvements were seen in all physical functioning measures in the subsample of participants with arthritis performing the physical functioning tests. Effect sizes were in the small to medium range; applied across the population, improvements in function would be very meaningful. The percentage of participants with arthritis who were classified as “impaired” also significantly decreased between 11% and 18%. These findings are promising, as individuals with arthritis who may already be experiencing declines in physical functioning can also benefit from PA-based programs (i.e., ALED); it may be possible to reverse such impairments. A recent meta-analysis (51) of randomized controlled trials found that community-deliverable exercise interventions resulted in a small but statistically significant improvement in physical functioning.

Although MVPA, meeting PA recommendations, and satisfaction with body appearance significantly improved in AC participants with arthritis, the improvements were not as large as those seen in participants without arthritis. One possible explanation for these findings is that AC did not include pain management strategies (although ALED did not either); pain and other arthritis-specific symptoms can be unique barriers to PA in this population. Additional studies are needed to understand how these outcomes may be improved even more in individuals with arthritis. Furthermore, although improvements in PA and a number of PA-related variables, all of which are important and relevant for people with arthritis, resulted from the AC and ALED interventions, this study did not measure other variables that are more specific to individuals with arthritis (e.g., pain, fatigue, stiffness). Therefore, it is unclear whether ALED and AC would produce meaningful changes in such variables. Future studies are needed to explore this in more detail.

Although the effectiveness of AC among individuals with arthritis has not been evaluated (or at least reported), the effectiveness of ALED has. Fontaine and colleagues (52) conducted a pilot study in a sample of participants with fibromyalgia based “loosely” on ALED. However, information specific to fibromyalgia was incorporated into the intervention, and therefore the original ALED was not tested. Callahan et al (53) tested the effectiveness of ALED in a large sample of participants with arthritis. Although results from the trial have not yet been published, qualitative data indicate that ALED was safe and appropriate for people with arthritis, and modifications to the program to make it more relevant for people with arthritis were suggested by the participants and instructors (53).

This study had a number of strengths, including the large sample size, the diverse sample, and the objective measures of functional fitness on a subsample. We also recognize study limitations. Perhaps the largest limitation is the lack of a control group. However, the goal of AFL was not to test the efficacy of AC and ALED, but to test whether both programs could be successfully translated into the community. Therefore, although internal validity may be threatened, the focus on external validity adds to the overall science-based literature. Second, the presence of arthritis was self-reported; there was no validity check to confirm the presence of doctor-diagnosed arthritis. Third, the type and severity of arthritis were not measured; it is possible that our sample had low symptoms (i.e., pain, fatigue) that may have influenced our findings. Participants with more severe symptoms might not have done as well in the programs. Fourth, a number of self-report measures were used, which could be subject to biases. Finally, because only one site (ALED) administered functional fitness tests and because not all participants were required to complete the long version of the survey, sample sizes for some outcomes were significantly lower.

Arthritis and its related complications remain an enormous public health problem. The CDC recommends a number of PA interventions for people with arthritis, some of which were recently added (8). Although options are available to clinicians and public health officials, participation in these programs is very low (9). AC and ALED resulted in improvements in PA, physical functioning, and other PA-related variables in participants with arthritis that were similar to those without arthritis. PA programs such as AC and ALED that target the general population, but may also result in improvements in PA, functional fitness, and other PA-related behavior in individuals with arthritis, are promising, as a larger and perhaps more diverse group of individuals could be reached.

AUTHOR CONTRIBUTIONS

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. Baruth 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. Wilcox.

Acquisition of data. Wilcox.

Analysis and interpretation of data. Baruth, Wilcox.

Acknowledgements

We gratefully acknowledge the many participants who took part in the AFL program and evaluation. We also acknowledge the involvement and contribution of staff from the following organizations involved in AFL: Berkeley Public Health Department, Blue Shield of California, Church Health Center, Council on Aging of Southwestern Ohio, FirstHealth of the Carolinas, Greater Detroit Area Health Council, Human Kinetics, Jewish Council for the Aging of Greater Washington, Kaiser Permanente-Colorado, The OASIS Institute, San Mateo County Health Department, Texas A&M University System, University of South Carolina, and the YMCA of Metropolitan Chicago.

Ancillary