Fit & Strong! Plus Trial Outcomes for Obese Older Adults with Osteoarthritis

Abstract

Background and Objectives

We compared the effectiveness of standard Fit & Strong! (F&S!; targets physical activity [PA]) to Fit & Strong! Plus (F&S! Plus; targets PA and dietary weight loss) on weight, diet quality, and PA outcomes.

Research Design and Methods

We randomly assigned 413 overweight older adults with OA to the F&S! or F&S! Plus programs and assessed outcomes at 2 and 6 months.

Results

The F&S! Plus group lost −2.0 ± 0.2 kg (mean ± SE, 2% of starting weight) at 2 months that was maintained at 6 months. Two- and 6-month BMI and waist circumference improved significantly in the F&S! Plus group (p < .001). Diet quality at 2 months showed greater improvement in the F&S! Plus group: 4.6 ± 0.7 versus 2.0 ± 0.7, p = .006, with no significant difference between groups at 6 months. The F&S! Plus group differentially improved on PA engagement at 2 months and at 2 and 6 months in joint pain (6-month mean ± SE: −1.5 ± 0.3 vs −0.6 ± 0.3, p = .02), function (−4.7 ± 0.9 vs −1.5 ± 0.9, p = .01), and 6-min walk test (29.5 ± 5.1 m vs 14.1 ± 5.2 m, p = .04).

Discussion and Implications

Adding a dietary weight loss component to F&S! achieved weight and waist circumference benefits that were maintained at 6 months. Importantly, the weight loss was accompanied by clinically meaningful improvements in OA symptoms and mobility. Future work should investigate minimum thresholds for weight reduction that improve long-term function in this population.

Arthritis affects 54.4 million adults in the United States, a majority of whom are older adults (Barbour, Helmick, Boring, & Brady, 2017). Osteoarthritis (OA) is its most common form, and OA in lower extremity (LE) joints is a risk factor for mobility disability (Centers for Disease Control and Prevention [CDC], 2013; Szoeke, Cicuttini, Guthrie, Clark, & Dennerstein, 2006). Given projected increases in the older adult population, the need to reduce mobility disability is urgent (CDC, 2013).

Obesity is a risk factor for incidence and progression of knee OA (Szoeke et al., 2006). Therefore, promoting healthy lifestyles in this population is critical (Chase, 2015; Greenwood-Hickman, Renz, & Rosenberg, 2016; Hawley-Hague et al., 2014). Obese older adults are advised to decrease sedentary behavior (Palmer et al., 2018), increase physical activity (PA) (Hu et al., 2018; Matz-Costa, Howard, Castaneda-Sceppa, Diaz-Valdes Iriarte, & Lachman, 2018) and lose weight (Hawley-Hague et al., 2014); however, this population is 44% more likely to be physically inactive than individuals who are obese without OA (Hootman, 2010).

Many studies have examined the impact of PA on older adults, including those with OA; however, only a few trials have tested the additive impact of weight loss and moderate PA in obese older adults. Villareal and colleagues (2011) found that weight loss combined with PA improved function more than either intervention alone. A follow-up study found that weight loss combined with aerobic and resistance exercise was most effective in improving function among obese older adults (Villareal et al., 2017).

Two trials examined the same issue in obese adults with OA. Messier and colleagues (2004) assessed the impact of PA and weight management alone or in combination and found that participants who received both treatments improved the most in pain, mobility and function. However, the dietary intervention was intensive and lasted for 18 months, presenting scalability challenges. The follow-up IDEA trial (Messier et al., 2013) found that an intervention combining PA and weight loss produced better weight loss, function, and quality of life outcomes than either treatment alone; however, the use of meal replacements could limit the appeal of the intervention. These trials had limited racial/ethnic minority participation (20%) (Messier et al., 2013; Villareal et al., 2017).

To address the need for scalable PA and weight loss programs for obese adults with OA, we combined two evidence-based interventions. The first, Fit and Strong! (F&S!), has been tested successfully in both efficacy (Hughes et al., 2004) and effectiveness trials (Hughes et al., 2006). F&S! has been tested in diverse populations and focuses on increasing PA and reducing OA symptoms among older adults with OA (Hughes et al., 2006). The second intervention, Obesity Reduction Black Intervention Trial (ORBIT), was developed to meet the cultural and dietary preferences of obese African American women (Fitzgibbon et al., 2008; Stolley et al., 2009). The combined intervention, Fit & Strong! Plus (F&S! Plus), built upon the core PA and OA components from F&S!, and added material from ORBIT (Fitzgibbon et al., 2010; Smith-Ray et al., 2014). Both interventions were based on Social Cognitive Theory (SCT), which helps individuals to develop needed self-regulatory skills (Bandura, 1989). OA, like other chronic conditions, can be better managed through behavior change interventions, including those focused on increasing PA and decreasing weight through dietary change (Davies & Garbutt, 2010; Painter, Borba, Hynes, Mays, & Glanz, 2008).

Methods

Design

We used a randomized comparative effectiveness trial to test whether F&S! Plus produced significantly better weight loss, dietary changes, diet quality, waist circumference, PA engagement, OA symptoms, functional performance, and anxiety and depression when compared with standard F&S! among overweight adults. The study design has been published (Smith-Ray et al., 2014). This project was approved by the Institutional Review Board at the University of Illinois at Chicago, and all participants provided written informed consent. The trial is registered at clinicaltrials.gov (NCT03180008). Participants were followed for 18 months; this article presents findings for 2 and 6 months.

Setting

Both interventions were conducted at the same 16 community sites.

Participants

Inclusion criteria

We used American College of Rheumatology criteria to define OA presence (Altman et al., 1986). Eligible participants reported pain in or around one or both knees or hips, most days in the past month and/or pain or stiffness in or around hips, knees ankles, feet or lower back on most days of at one least 1 month during the last 6 months, were 60 or older, had a BMI of 25–50 kg/m², and were willing to participate in research.

Exclusion criteria

Participants were excluded if they had rheumatoid arthritis, medical conditions that precluded exercise participation, engaged in ≥150 min/week of aerobic activity, had three or more errors on the Mental Status Questionnaire (MSQ), or had other LE joint issues previously described (Smith-Ray et al., 2014).

Interventions

Both interventions were conducted in 90-min sessions held three times per week over 8 weeks (24 sessions total). All aspects of the interventions were linked to our conceptual framework of SCT, so that there was a continual focus on enhancing self-regulatory skills in an effort to enhance social and environmental supports and self-efficacy in both F&S! and F&S! Plus. The first 60 min of both interventions consisted of PA that included stretching, low-impact aerobics, and LE strength training. The health education portion of both programs lasted 30 min and incorporated self-efficacy, social support, and self-regulation (e.g., goal setting planning), following a structured, scripted curriculum (see Supplementary Appendix 1 for a description of the curricula for both programs).

Fit & Strong!

The health education component of standard F&S! was manual-based and incorporated concepts of social support and goal setting to help participants build efficacy for managing OA symptoms through PA. The customized manual had no content on dietary weight loss or improving diet quality.

Fit & Strong! Plus

The curriculum retained the core PA and OA materials from the F&S! manual and added 16 weight and diet-related topics to the curriculum, with weight loss and diet quality concepts discussed in 22 of the 24 sessions. The diet quality content was based on the Group Lifestyle Balance curriculum that was adapted from the Diabetes Prevention Program and the 2010–2015 Dietary Guidelines for Americans (Dietary Guidelines for Americans, 2010) and USDA MyPlate. Participants were weighed weekly and were assigned homework including maintaining a daily food diary, reading food labels, planning healthy food shopping lists and goal setting. At session 2, participants received a USDA MyPlate eating plan tailored to produce a 5% weight loss at 6 months and improve diet quality. Together, the in-class curriculum, homework, and weekly weight checks were designed to enhance social support (e.g., group problem solving) and self-regulation (i.e., goal setting, planning, and self-monitoring) to build SE for weight loss and improving diet quality. Additional details regarding the interventions have been published elsewhere (Smith-Ray et al., 2014).

Instructors

Instructors in both study arms were Certified Exercise Instructors (Institute for Credentialing Excellence, 2018). F&S! instructors received the customary 8-hr in-person training, and F&S! Plus, instructors received 8 hr of customary training plus an additional 8 hr of training in the dietary weight loss/management curriculum. All trainings emphasized the critical role of enhancing SE to promote behavior change along with techniques to bolster efficacy.

Two research staff attended one session of each program between weeks 2 and 4 at each site to monitor intervention fidelity using an observational checklist, followed by a meeting with the instructor to review the instructor’s performance. If problems were identified, the issue was discussed and a follow-up visit was conducted to ensure that the issue had been corrected.

Individual telephone reinforcement sessions with health educators in months 4, 8, and 15 encouraged maintenance of behavior change, with a continued focus on self-monitoring, planning, and seeking social support from family and friends. Educators assigned to the F&S! group focused on strategies for PA change and maintenance, while those assigned to F&S! Plus focused on both PA and dietary change and maintenance. All educators were trained in motivational interviewing techniques.

Measures

All measures were administered by trained interviewers. Demographic characteristics and self-reported chronic conditions were assessed at baseline (Hughes, Edelman, Chang, Singer, & Schuette, 1991).

Primary outcomes

Anthropometrics

Weight was measured using a calibrated digital scale (Tanita Worldwide). Height was measured using a portable stadiometer (Seca, UK). Waist circumference was measured just above the ileac crest at the mid-axillary line (CDC, 2013) using a Gulick 150-cm anthropometric tape (Country Technology, Inc.).

Dietary intake

We used the 110-item Block 2005 Food Frequency Questionnaire (FFQ; Block et al., 1986) to estimate diet quality using the Healthy Eating Index 2010 (HEI 2010; Guenther et al., 2013). HEI-2010 total scores range from 0 to 100 points, with a higher score signifying superior diet quality. Nutrition Quest (Berkeley, CA) scored the FFQ data and calculated HEI 2010 total and component scores.

Secondary outcomes

Physical activity

PA was assessed using the Physical Activity Scale for the Elderly (PASE; Washburn & Ficker, 1999). A higher score indicates greater PA.

Osteoarthritis symptoms

We used the WOMAC to assess LE joint pain and impairment. Higher scores indicate more severe OA symptoms.

Performance measures

Mobility was assessed using the 6-min walk test. We asked participants to walk as far as possible in 6 min (Guyatt et al., 1985; Naylor et al., 2014). Functional LE strength was measured using the 30-s chair stand. For both measures, higher scores indicate better performance.

Depression and anxiety

These outcomes were measured using the GERI-AIMS (Hughes et al., 1991).

Random assignment

Participants who completed the baseline interview were randomly assigned to F&S! Plus or F&S! using the Research Electronic Data Capture (REDCap) randomization module (Harris et al., 2009). Randomization was stratified by site/iteration, WOMAC physical function score (0–19 vs 20–68) and BMI (<35 kg/m² vs ≥35 kg/m²), antecedent variables that could predispose participants to be more responsive to the interventions.

Statistical Analysis

We used t tests and chi-square tests for between-group differences at baseline and Wilcoxon tests for differences in attendance. To test the effects of the interventions on primary and secondary outcome measures, we used repeated-measures linear models in SAS PROC MIXED with a fully specified (unstructured) covariance matrix and the baseline value included in the outcome vector (Mallinckrod, Lane, Schnell, Peng, & Mancuso, 2008). All models included intervention group, visit as a set of indicator variables with baseline as the reference group, and a group × visit interaction term. Stratification variables were included as covariates. If the baseline or a closely related value was included in the outcome vector (the weight, BMI, and WOMAC physical function models), that covariate was omitted from the model. For dietary outcome variables, records with implausible values for energy (<500 or >5,000 kcal) were excluded.

In addition to testing for an overall group × visit interaction, we used SAS LSMESTIMATE statements to estimate change from baseline in the adjusted mean within each intervention group at 2 and 6 months to test whether the within-group changes were significantly different from 0. We also tested for between-group differences in change at 2 and 6 months. To estimate the adjusted percentage change at 2 and 6 months, we divided the estimated change by the adjusted baseline mean and multiplied by 100.

A repeated-measures analysis using restricted maximum likelihood and unstructured covariance has been found to give unbiased estimates in cases with missing data, as long as the data are missing at random (Mallinckrod et al., 2008). Therefore, the analysis described above is an intention-to-treat analysis. All statistical analyses were performed using SAS v9.4 (SAS Institute, Cary, NC).

Results

As shown in Figure 1, we screened 1,379 older adults to recruit 433 persons who were eligible and interested in study participation. We ultimately randomized 413 participants: 203 to F&S! Plus and 210 to F&S!

Figure 1.

Participant flow.

Baseline demographic data are shown in Table 1. There were no between-group differences on any of the variables described. The total sample mean age was 68 years (SD = 6); 92% were African American and 86% were female. A majority (79%) attended some college or had a college degree. Median annual household income was $25,000.

Table 1.

Participant Characteristics at Baseline

	F&S! Plus,aN = 203		F&S!, N = 210		All, N = 413b
	Mean or %	SD or N	Mean or %	SD or N	Mean or %	SD or N
Age, years	68	6	68	6	68	6
Sex
Female	87%	177	85%	178	86%	355
Male	13%	26	15%	32	14%	58
Race
Black or African American, not Hispanic	92%	187	92%	193	92%	380
White, not Hispanic	4%	8	4%	9	4%	17
Hispanic	1%	3	1%	2	1%	5
Multiracial/other	2%	5	3%	6	3%	11
Education, years	14	2	14	2	14	2
Not HS graduate, n (%)	4%	9	7%	15	6%	24
HS graduate/GED	13%	27	16%	34	15%	61
Some college or technical school	45%	91	40%	84	42%	175
College graduate	37%	76	37%	77	37%	153
Employed full or part-time	15%	30	11%	23	13%	53
Marital status
Married or member of unmarried couple	26%	52	26%	54	26%	106
Divorced or separated	34%	70	39%	82	37%	152
Widowed	24%	48	20%	43	22%	91
Never married	16%	33	15%	31	16%	64
Income, medianb	25,000		25,000		25,000
<$20,000	37%	64	34%	58	35%	122
$20,000–<$40,000	26%	45	28%	48	27%	93
≥$40,000	37%	64	38%	66	38%	130
Health insurancec
Medicare	67%	136	70%	147	69%	283
Medicaid	16%	33	18%	38	17%	71
Private/supplemental	44%	89	40%	83	42%	172
Chronic conditions, of 17d	3	2	3	2	3	2
High blood pressure	76%	155	76%	160	76%	315
Diabetes	25%	51	26%	55	26%	106
Heart disease	10%	20	12%	26	11%	46
Vision problems	26%	53	28%	58	27%	111
Hearing problems	10%	20	12%	26	11%	46
Thyroid problems	10%	20	10%	22	10%	42
BMI category, kg/m2
Overweight (25–<30)	21%	43	19%	39	20%	82
Obese (≥30)	79%	160	81%	171	80%	331

^aThere were no significant differences between groups at baseline.

^b N = 345 for income.

^cPercentage of participants reporting each type of insurance; some participants reported more than one type of insurance.

^dChronic conditions: number of self-reported conditions currently affecting health (of 17): arthritis, high BP, heart disease, mental illness, diabetes, cancer, alcohol or drug abuse, lung disease, kidney disease, liver disease, stomach disease, blood disease, stroke or other neurologic problems, vascular disease, vision problems, hearing problems, thyroid.

F&S! Plus participants attended a mean 72% of classes (median 83%), and F&S! participants attended a mean 68% (median 79%). Two-thirds of F&S! Plus participants attended at least 75% of classes, compared with 60% of F&S! participants. There was no significant between-group difference in attendance.

Of 413 participants randomized, 191 (94%) in the F&S! Plus arm and 192 (91%) in the F&S! arm completed 2-month post-tests. Of the total baseline sample, 175 (86%) in the F&S! Plus arm completed a 6-month post-test, compared with 173 (82%) of the F&S! arm.

Baseline characteristics are shown in Table 1.

Adjusted mean changes from baseline to 2 and 6 months for all outcomes are shown in Table 2.

Table 2.

Adjusted^a Mean Outcome Measures at Each Visit and Mean Change From Baseline to 2 and 6 Months

Outcomes over time	F&S! Plus					F&S!					Significance of difference in mean change between groups	Significance of overall group × visit interaction
	Outcome measure	Change from baseline			(N)	Outcome measure	Change from baseline			(N)
	Mean ± SE	Meanb ± SE	%c	p d		Mean ± SE	Meanb ± SE	%c	p d		p e	p
Weight, kg												<.0001
Baseline	93.2 ± 1.1				(N = 203)	93.7 ± 1.1				(N = 210)	—
2 months	91.2 ± 1.1	−2.0 ± 0.2	−2.1%	***	(N = 191)	93.2 ± 1.1	−0.5 ± 0.2	−0.5%	**	(N = 191)	<.0001
6 months	91.1 ± 1.1	−2.0 ± 0.2	−2.2%	***	(N = 175)	93.0 ± 1.1	−0.8 ± 0.3	−0.8%	**	(N = 171)	.0004
BMI, kg/m2												<.0001
Baseline	34.7 ± 0.4				(N = 203)	35.0 ± 0.4				(N = 210)
2 months	34.0 ± 0.4	−0.7 ± 0.1	−2.1%	***	(N = 191)	34.9 ± 0.4	−0.2 ± 0.1	−0.5%	**	(N = 191)	<.0001
6 months	34.0 ± 0.4	−0.8 ± 0.1	−2.2%	***	(N = 175)	34.7 ± 0.4	−0.3 ± 0.1	−0.8%	**	(N = 171)	.0004
Waist circumference, cm												<.0001
Baseline	114.0 ± 0.5				(N = 201)	111.9 ± 0.5				(N = 210)
2 months	111.1 ± 0.5	−2.9 ± 0.4	−2.5%	***	(N = 191)	111.4 ± 0.5	−0.5 ± 0.4	−0.4%		(N = 191)	<.0001
6 months	110.5 ± 0.6	−3.5 ± 0.5	−3.1%	***	(N = 175)	110.9 ± 0.6	−1.0 ± 0.5	−0.9%		(N = 171)	.0004
Healthy Eating Index-2010f												.02
Baseline	65.8 ± 0.7				(N = 197)	66.7 ± 0.7				(N = 203)
2 months	70.5 ± 0.7	4.6 ± 0.7	7.0%	***	(N = 183)	68.8 ± 0.7	2.0 ± 0.7	3.0%	**	(N = 185)	.006
6 months	67.8 ± 0.8	2.0 ± 0.7	3.0%	**	(N = 166)	67.6 ± 0.8	0.9 ± 0.7	1.3%		(N = 160)	.29
Selected HEI component scores f
Whole fruit (0–5)												.02
Baseline	3.8 ± 0.1				(N = 197)	4.0 ± 0.1				(N = 203)
2 months	4.5 ± 0.1	0.7 ± 0.1	17.7%	***	(N = 183)	4.3 ± 0.1	0.3 ± 0.1	7.3%	**	(N = 185)	.006
6 months	4.4 ± 0.1	0.6 ± 0.1	15.1%	***	(N = 166)	4.2 ± 0.1	0.2 ± 0.1	6.3%	*	(N = 160)	.03
Total vegetables (0–5)												.008
Baseline	4.0 ± 0.1				(N = 197)	4.0 ± 0.1				(N = 203)
2 months	4.5 ± 0.1	0.4 ± 0.1	11.1%	***	(N = 183)	4.1 ± 0.1	0.1 ± 0.1	2.8%		(N = 185)	.004
6 months	4.2 ± 0.1	0.2 ± 0.1	5.8%	**	(N = 166)	3.9 ± 0.1	−0.1 ± 0.1	−2.2%		(N = 160)	.01
Whole grains (0–10)												.07
Baseline	4.5 ± 0.2				(N = 197)	4.3 ± 0.2				(N = 203)
2 months	4.4 ± 0.2	−0.1 ± 0.2	−1.4%		(N = 183)	4.2 ± 0.2	−0.1 ± 0.2	−2.2%		(N = 185)	.91
6 months	3.8 ± 0.2	−0.7 ± 0.2	−14.6%	**	(N = 166)	4.3 ± 0.2	0.0 ± 0.2	−1.0%		(N = 160)	.06
Total protein foods (0–5)												.09
Baseline	4.8 ± 0.0				(N = 197)	4.7 ± 0.0				(N = 203)
2 months	4.7 ± 0.0	−0.1 ± 0.1	−1.5%		(N = 183)	4.7 ± 0.0	0.1 ± 0.1	1.2%		(N = 185)	.08
6 months	4.7 ± 0.1	0.0 ± 0.1	−0.7%		(N = 166)	4.6 ± 0.1	−0.1 ± 0.1	−1.2%		(N = 160)	.81
Refined grains (0–10)												.03
Baseline	9.0 ± 0.1				(N = 197)	9.1 ± 0.1				(N = 203)
2 months	9.4 ± 0.1	0.4 ± 0.1	4.1%	**	(N = 183)	9.0 ± 0.1	0.0 ± 0.1	−0.5%		(N = 185)	.01
6 months	9.3 ± 0.1	0.3 ± 0.1	3.3%	*	(N = 166)	9.0 ± 0.1	0.0 ± 0.1	−0.3%		(N = 160)	.08
Empty calories (0–20)												.01
Baseline	12.0 ± 0.3				(N = 197)	12.5 ± 0.3				(N = 203)
2 months	13.8 ± 0.3	1.7 ± 0.3	14.4%	***	(N = 183)	13.2 ± 0.3	0.7 ± 0.3	5.6%	*	(N = 185)	.01
6 months	12.2 ± 0.4	0.2 ± 0.3	1.7%		(N = 166)	12.8 ± 0.4	0.3 ± 0.3	2.5%		(N = 160)	.82
PASE physical activity scoreg												.07
Baseline	99.4 ± 4.2				(N = 199)	99.2 ± 4.2				(N = 210)
2 months	128.8 ± 5.0	29.4 ± 4.8	29.6%	***	(N = 184)	113.6 ± 5.0	14.4 ± 4.7	14.5%	**	(N = 184)	.03
6 months	116.2 ± 5.2	16.9 ± 5.0	17.0%	***	(N = 172)	113.8 ± 5.3	14.6 ± 5.0	14.7%	**	(N = 169)	.75
WOMAC pain (0–20)h												.03
Baseline	5.4 ± 0.2				(N = 203)	5.7 ± 0.2				(N = 210)
2 months	4.1 ± 0.2	−1.4 ± 0.2	−25.3%	***	(N = 191)	5.2 ± 0.2	−0.5 ± 0.2	−8.9%	*	(N = 192)	.01
6 months	4.0 ± 0.3	−1.5 ± 0.3	−26.8%	***	(N = 174)	5.1 ± 0.3	−0.6 ± 0.3	−10.2%	*	(N = 173)	.02
WOMAC physical function (0–68)h												.003
Baseline	17.7 ± 0.9				(N = 203)	17.8 ± 0.9				(N = 210)
2 months	12.1 ± 0.9	−5.6 ± 0.8	−31.5%	***	(N = 191)	15.9 ± 0.9	−1.9 ± 0.7	−10.9%	*	(N = 191)	.0007
6 months	13.0 ± 1.0	−4.7 ± 0.9	−26.6%	***	(N = 174)	16.3 ± 1.0	−1.5 ± 0.9	−8.5%		(N = 173)	.01
Six-min walk, m												<.0001
Baseline	360.2 ± 6.2				(N = 201)	357.7 ± 6.2				(N = 208)
2 months	411.1 ± 6.8	50.9 ± 4.4	14.1%	***	(N = 190)	380.8 ± 6.8	23.0 ± 4.4	6.4%	***	(N = 189)	<.0001
6 months	389.7 ± 7.2	29.5 ± 5.1	8.2%	***	(N = 173)	371.9 ± 7.3	14.1 ± 5.2	4.0%	**	(N = 165)	.04
Chair stands in 30 s												.19
Baseline	8.8 ± 0.2				(N = 203)	8.9 ± 0.2				(N = 209)
2 months	10.6 ± 0.2	1.8 ± 0.2	20.5%	***	(N = 187)	10.2 ± 0.2	1.4 ± 0.2	15.3%	***	(N = 188)	.10
6 months	10.4 ± 0.3	1.7 ± 0.2	18.8%	***	(N = 172)	10.0 ± 0.3	1.2 ± 0.2	13.3%	***	(N = 165)	.10
Anxiety and depression (0–10)i												.23
Baseline	2.5 ± 0.1				(N = 203)	2.6 ± 0.1				(N = 210)
2 months	2.0 ± 0.1	−0.5 ± 0.1	−19.9%	***	(N = 191)	2.2 ± 0.1	−0.4 ± 0.1	−13.8%	***	(N = 190)	.26
6 months	2.1 ± 0.1	−0.4 ± 0.1	−15.4%	***	(N = 174)	2.4 ± 0.1	−0.1 ± 0.1	−5.9%		(N = 173)	.09

^aFrom repeated-measures linear models with a fully specified (unstructured) covariance matrix and the baseline value included in the outcome vector. Covariates included in models: iteration, baseline WOMAC physical function (except for the WOMAC physical function model), and baseline BMI (except for the weight and BMI models).

^bEstimated mean change is the difference between the follow-up and baseline adjusted means.

^cPercent change is mean change divided by the adjusted baseline mean.

^dTest for within-group change different from 0.

^eTest for difference in change between groups at 2 months and 6 months.

^fA higher score indicates higher diet quality.

^gA higher score indicates greater physical activity.

^hA higher score indicates greater difficulties due to OA.

ⁱA higher score indicates greater anxiety/depression.

*p < .05; **p < .01; ***p < .001.

Primary Outcome: Body Weight, Adiposity

There were no significant between-group differences for body weight, BMI, or waist circumference at baseline. Mean BMI at baseline was 34.8 kg/m² (SD = 5.5) and mean waist circumference was 113.1 cm (SD = 13.5). At 2 months, F&S! Plus participants experienced a mean 2.0 kg weight loss (SE = 0.2), or 2.1% of starting weight, that was maintained at 6 months, compared with a statistically but not clinically significant weight loss of <1 kg in the F&S! group, adjusted for iteration and baseline physical function (p < .001). Similarly, BMI decreased by 0.7 ± 0.1 kg/m², 2.1% of baseline BMI, at 2 months and 0.8 ± 0.1 kg/m² at 6 months for F&S! Plus compared with much smaller losses for F&S! (0.2 and 0.3 kg/m²), p < .001. Mean waist circumference also decreased in the F&S! Plus group: 2.9 ± 0.4 cm, 2.5% of baseline, at 2 months and 3.5 ± 0.5 cm, 3.1%, at 6 months, compared with no significant change in the F&S! group (−0.5 cm and −1.0 cm), p < .001.

Diet Quality

The HEI 2010 total score increased significantly more in the F&S! Plus group than in the F&S! group at 2 months, but the difference between groups was not maintained at 6 months. At 2 months, HEI 2010 total scores increased by 4.6 (SE = 0.7), or 7.0% over baseline in F&S! Plus, compared with 2.0 ± 0.7 in F&S!, p = .006. At 6 months, there was no longer a significant difference between groups. For whole fruit intake, there was a significant difference between groups in change from baseline at both 2 (0.7 ± 0.1 vs 0.3 ± 0.1, p = .006) and 6 months (0.6 ± 0.1 vs 0.2 ± 0.1, p = .03). Total vegetable intake increased significantly for the F&S! Plus group at both 2 and 6 months but did not change significantly in the F&S! group (2 months: 0.4 ± 0.1 vs 0.1 ± 0.1, p = .004; 6 months: 0.2 ± 0.1 vs. −0.1 ± 0.1, p = .01). Intake of refined grains improved (decreased) at 2 months for F&S! Plus, but not for F&S! (0.4 ± 0.1 vs 0.0 ± 0.1, p = .01). Intake of empty calories also improved (decreased) in both groups at 2 months, with more improvement in F&S! Plus (1.7 ± 0.3 vs 0.7 ± 0.3, p = .01), but this improvement was not maintained at 6 months in either group. Other HEI components did not show significant differences between groups: total fruit, greens and beans, dairy, seafood and plant protein, fatty acids, and sodium (data not shown).

Secondary Outcomes

Physical activity

Participants’ baseline mean score on the PASE was 97.2 (SD = 61.3), with no significant between-group difference. Although both groups improved on this measure at posttest, a statistically significant between-group difference favoring F&S! Plus was seen on the PASE at 2 months (p = .03) that was not maintained at 6 months. Scores in the F&S! Plus group increased by 29.4 (SE = 4.8), or 29.6% of baseline, at 2 months and 16.9 ± 5.0 (17.0%) at 6 months. Comparable increases for the F&S! group at the same time-points were 14.4 ± 4.7 (14.5%) and 14.6 ± 5.0 (14.7%).

Osteoarthritis symptoms

There were no significant between-group baseline differences on the WOMAC scales. Participants’ mean baseline score was 5.6 (SD = 4.0) for pain and 18.0 (SD = 12.9) for physical functioning, indicating a moderate amount of OA-related impairment. Pain decreased in both groups, but significantly more in F&S! Plus than in F&S! at 2 (p = .01) and 6 (p = .02) months. F&S! Plus participants reported a 1.4 (SE = 0.2), or 25.3%, decrease in pain at 2 months and a 1.5 ± 0.3 (26.8%) decrease at 6 months compared with decreases of 0.5 ± 0.2 (8.9%) and 0.6 ± 0.3 (10.2%) in the F&S! group at the same time-points. Functional impairment improved in both groups at 2 months, with greater improvement in F&S! Plus (5.6 ± 0.8, 31.5% vs 1.9 ± 0.7, 10.9%, p < .001). This between group improvement was maintained at 6 months (4.7 ± 0.9, 26.6%, vs 1.5 ± 0.9, 8.5%, p = .01).

Performance Measures

Six-minute walk test

Participants’ mean score at baseline was 356.3 m (SD = 97.1), or 1.0 m/s with no between-group differences. There were statistically significant between-group improvements favoring F&S! Plus on this outcome at both 2 (p < .001) and 6 (p = .04) months. Walking distance improved in the F&S! Plus group by 50.9 (SE = 4.4) m, or 14.1%, at 2 months and by 29.5 ± 5.1 m, 8.2%, at 6 months. Increases in the F&S! group were 23.0 ± 4.4 m, 6.4% and 14.1 ± 5.2, 4.0% at the same time-points.

Thirty-second chair stand

Participants’ baseline mean score was 8.7 (SD = 3.6) on the 30-s chair stand, with no significant between-group differences. Improvements on this measure were seen at 2 and 6 months in both groups, with no significant between-group differences.

Depression and anxiety

Participants’ mean baseline score for depression and anxiety was 2.5 (SD = 1.7), with no significant between-group differences. We saw no statistically significant between-group differences at 2 or 6 months.

Discussion

To our knowledge, this is the first study in the literature to test the impact of a combined PA/diet/weight loss intervention in older obese African American adults with OA. OA affects more than 30 million older adults in the United States, with African Americans disproportionately affected (Jordan, 2015). The disparity is driven largely by the higher rates of overweight and obesity among African Americans (Flegal, Kruszon-Moran, Carroll, Fryar, & Ogden, 2016). Findings at both 2 and 6 months showed improved BMI and waist circumference in F&S! Plus compared with F&S!. Diet quality, based on HEI 2010, also, showed greater improvement in F&S! Plus compared to F&S! at 2 months, but there was no significant difference between the two groups at 6 months. When compared with standard F&S!, F&S! Plus participants also had less joint pain, improved joint function, and faster walking speed that approximated findings in other studies that achieved greater amounts of weight loss (Christensen, Bartels, Astrup, & Bliddal, 2007; Messier et al., 2004).

Primary Outcomes

Although the goal of F&S! Plus was to achieve a 5% reduction in weight at 6 months, the actual impact was 2.2%, with similar effects on waist circumference. The IDEA trial (Messier et al., 2013) reported a 10% reduction in body weight at 6 months. Our smaller effect on weight may be a function of the dose and duration of our weight loss intervention. The IDEA trial’s active intervention phase was 6 months in duration, was implemented by nutrition interventionists, conducted one-on-one or in small groups, included an 800–1000 calorie/day deficit at the start of the trial, and provided daily meal replacements (Foy, Penninx, Shumaker, Messier, & Pahor, 2005). In contrast, F&S! Plus lasted for 2 months, was conducted in larger groups (10–15), targeted smaller caloric decreases and was implemented by exercise instructors. The Villareal trial (2017) found a 9% decrease in body weight at 6 months. The dietary intervention tested lasted for 26 weeks, incorporated an energy deficit of 500–750 calories/day, contained 1 g of high quality protein per kilogram of body weight per day, and was implemented through weekly visits with a dietitian.

The existing PA and dietary weight loss interventions for overweight adults with OA have not focused simultaneously on decreasing calories while improving diet quality. The participants in F&S! Plus were instructed to follow an energy restricted USDA MyPlate eating plan to promote weight loss and improve diet quality. We chose to focus on diet quality because diet quality is associated with less weight gain over time (Fung & Pan, 2015), weight loss (Anderson, Green, & Payne, 2009), and lower systemic inflammation controlling for weight loss. (Mohammadshahi, Haidari, Karandish, Ebrahimi, & Haghighizadeh, 2014) Although improvements in overall diet quality were not maintained in the F&S! Plus group at 6 months, they did have significantly higher whole fruit and total vegetable HEI 2010 component scores at 6 month follow-up compared with the F&S! group. Whole fruits and vegetables are associated with greater intentional weight loss (Fuglestad, Jeffery, & Sherwood, 2012) and may have translated to the greater weight loss and improved function in the F&S! Plus group at 6-month follow-up.

Secondary Outcomes

Despite the modest impact on weight and dietary behavior, we saw differential improvement in the F&S! Plus group on secondary outcomes that examined LE function, including the 6-min walk test and LE joint pain and function. Distance covered in the 6-min walk test improved in the F&S! Plus group at twice the amount seen in the customary F&S! at both 2 and 6 months. Percent decrease in LE joint pain at both time-points averaged about 25%. The percent decrease in LE joint impairment was also substantial at both time-points (31.5% and 26.6%, respectively) in the F&S! Plus group. These decreases in WOMAC scores exceed the 20% improvement in these measures that is required for a clinically important difference (Bellamy, Buchanan, Goldsmith, Campbell, & Stitt, 1988; Dougados et al., 2000). The 26.8% reduction in LE pain scores achieved by F&S! Plus at 6 months compares favorably with the 31.4% decrease reported by the IDEA trial at the same time point (Messier et al., 2013); a promising finding since our intervention was 2 rather than 6 months in duration. The 26.6% decrease in functional impairment in our trial at 6 months also compares favorably with the 32.9% decrease in the IDEA trial at the same time-point.

The significant between-group increase in PA favoring F&S! Plus at 2 months declined over time and was similar to the increase achieved by the F&S! group at 6 months. Although we saw significant within-group improvement on the 30-s chair stand test in both groups at 2 and 6 months, there were no significant between-group differences on this measure. Failure to observe significant between-group differences may reflect the fact that in a comparative effectiveness design, both groups received the PA component of customary F&S!, including systematic LE strength training and coaching to boost PA engagement. Therefore, it is reasonable that both groups would improve on these outcomes. Finally, although we observed significant within-group improvements in depression/anxiety in the F&S! Plus group at 2 and 6 months, we did not find a significant between-group difference on this outcome.

Limitations

The current study is limited to individuals who are overweight or obese. We used self-reported LE pain and stiffness to define presence of OA, and our sample is primarily African American and female. We used an FFQ to capture dietary intake. This instrument may suffer from inaccurate recall or a tendency to provide socially desirable responses (Thomson et al., 2011).

Strengths

Strengths include our rigorous randomized comparative effectiveness design, the high degree of adherence to the interventions and our enrollment of an understudied minority population. Our study sample was 92% African American and 86% female. African American women have 2.5 times the prevalence of knee OA as non-Hispanic white women (Sowers, Lachance, Hochberg, & Jamadar, 2000) and are also disproportionately affected by obesity (Flegal et al., 2016).

We have previously documented the presence of obesity and OA-related symptoms in our study population at baseline (Fitzgibbon et al., 2018). At baseline, our study participants had an HEI 2010 total score, an indicator of diet quality, that was slightly lower (66.3, SD = 10.5) than the average for older U.S. adults (69.3, SE = 1.8), and fewer than 50% met recommended PA levels (USDA, 2012). The mean 6-min walk test score in our sample at baseline (356.3, SD = 97.1 m) is lower than mean scores of 488.1–500.3 m reported by Wilcox and colleagues (2015) and the 467 m reported in the IDEA trial (Messier et al., 2013).

The picture that emerges from these data is one of a sample that suffered from multiple obesity and OA-related symptoms at baseline but responded to a community-based intervention positively with reduced weight and improved diet quality, LE OA symptoms, and mobility at 2 months that were largely maintained at 6 months. In this context, our finding that a small amount of weight loss is associated with clinically significant reductions in LE pain and function is important.

Conclusions

When participants are limited to persons with LE OA who are also overweight or obese. F&S! Plus is more effective than customary F&S! in producing small weight and adiposity changes that are associated with clinically important improvements in LE joint pain, function and mobility at 2 months that are maintained at 6 months. We believe that the trial findings fill an important gap in the evidence base examining the impact of weight loss and PA interventions for overweight older African American women who have OA.

Funding

This work was supported by the National Institute on Aging of the National Institutes of Health under award number (R01AG039374), the American Cancer Society of Illinois (261775) and American Cancer Society Mentored Research Scholar (MRSG014-025-01-CNE) to Dr L. Tussing-Humphreys.

Conflict of Interest

None reported.