Physical Activity Changes During Behavioral Weight Loss by Latinx Patients with Obesity with/without Binge-Eating Disorder

Abstract

Objective:

This secondary analysis examined PA changes and their prognostic significance among Latinx patients with obesity – with and without binge-eating disorder (BED) – who participated in randomized placebo-controlled trial (RCT) testing addition of orlistat to behavioral weight loss (BWL) treatment in a “real-world” clinical setting.

Methods:

In this RCT at a community mental-health center serving economically-disadvantaged Spanish-speaking-only Latinx patients, 79 patients with obesity (40 with BED, 39 without BED) received BWL and were randomized to orlistat or placebo. PA, weight, depression, and binge-eating were assessed at baseline, posttreatment (end-of-treatment; 4 months), and 6-month follow-up (10 months after baseline).

Results:

PA was low at baseline (9.3% categorized “active”), increased during treatment (32.9% categorized “active” at post-treatment), and declined from posttreatment to 6-month follow-up (28.2% classified “active”). At baseline, PA was lower among patients with BED than without BED. Changes in PA during and after treatment did not differ by BED status or medication condition. PA change was associated with reduced depression but not weight loss.

Conclusions:

Latinx patients with obesity receiving BWL achieved significant, albeit modest, increases in PA. Although PA changes were not associated with weight loss, they were associated with reduced depression. Identifying methods to increase PA further is necessary.

Introduction

Physical inactivity represents a major public health concern, as it is associated with increased risk for multiple forms of morbidity and all-cause mortality (1). Particularly low levels of physical activity (PA) are observed among individuals with obesity, and especially among those with binge-eating disorder (BED) (2,3). Increasing PA is a key component of lifestyle behavioral weight-loss (BWL) treatment for obesity (4) which has also received empirical support for treating BED (5). BWL reliably produces weight loss among patients in treatment studies for obesity (6) and BED (7). With respect to BED, BWL produces greater weight loss than other leading evidence-based approaches for BED, such as cognitive-behavioral therapy (CBT) which focuses on binge eating and body image but does not integrate nor emphasize the importance of regular PA (8,9). It appears, however, that the amount of weight loss achieved through BWL in studies with BED tends to be less than the amounts typically reported in treatment studies for obesity without BED (10). Research on BWL effects on PA generally indicates that PA levels improve and are associated with better longer-term weight outcomes (11). The very limited research literature with BED suggests that PA-specific interventions can increase PA, but little is known about the nature and prognostic significance of PA changes during and following BWL treatment for BED (12,13).

Latinx individuals, the largest and fastest growing ethnic minority group in the United States (14), are disproportionately affected by obesity (15) but have similar rates of BED as non-Latinx White individuals (16). Some research indicates that Latinx individuals are less physically active than non-Latinx White individuals (17), although the literature is mixed (18). Latinx individuals are markedly under-represented in treatment research for both obesity (19) and BED (20). The few studies that have evaluated efficacy of BWL among Latinx individuals have reported small weight losses (21–23). Some treatment research specifically with Latinx patients reported BWL produces statistically significant increases in PA levels (24) and secondary analyses found increased PA is associated with improved weight losses in Latinx with type 2 diabetes (25).

The present study is a secondary analysis designed to explore the nature and prognostic significance of PA changes among Latinx patients with obesity – with and without BED – who participated in RCT testing the addition of a weight-loss medication (orlistat versus placebo) to a BWL lifestyle modification intervention in a “real-world” clinical setting (26). Here, we summarize the study and previously-published primary outcomes (26) to provide context for the present analysis. The RCT was performed at an urban community mental-health center serving educationally/economically-disadvantaged patients with psychiatric conditions. In addition to addressing marked health disparities pertaining to obesity and BED in Latinx individuals, the RCT was designed to address the gap between research and practice in “real-world” settings. Specifically, the RCT was designed to examine whether evidence-based behavioral/pharmacological interventions for obesity/BED supported in research centers can be successfully delivered in “real-world” clinics with patients using minimal exclusionary criteria. Low educational attainment, complex/multiple forms of psychiatric comorbidity, and use of psychiatric medications were not exclusionary (except in instances of medical safety). Overall, 78% of patients completed treatments, completion rates did not differ significantly by medication or BED status, and significant reductions in binge-eating, weight, eating-disorder psychopathology, and depression levels were observed. The addition of orlistat to BWL, relative to placebo, resulted in significantly greater percent weight-loss in the non-BED group (4.9% versus 1.9%) but not the BED group (3.9% versus 2.1%). Of the patients with BED, binge-eating abstinence rates were 65% at end-of-treatment and 50% at 6-month follow-ups. These weight-loss outcomes for non-BED patients are similar to the 4.7% reported in the Latin-American orlistat trial (27) and fall in the 3%−7% range reported for similar BWL trials using the Diabetes Prevention Program (28) while the binge-eating and weight-loss outcomes are similar to those reported for similar RCTs in BED (29) with much more restrictive exclusionary criteria.

The present study is a secondary analysis designed to explore changes and prognostic significance of changes in PA among Latinx patients with obesity – with and without BED – who participated in a RCT (26). Specifically, analyses were designed to examine: [1] engagement in PA, [2] change in PA from baseline to posttreatment (end-of-treatment at 4 months) and 6-month follow-up after completing treatment (10 months after baseline), and [3] the association between changes in PA and changes in weight and depression.

Methods

Participants

Participants were Spanish-speaking (monolingual) Latinx patients with obesity recruited from a community mental-health center serving economically-disadvantaged Latinx individuals with psychiatric needs who participated in a RCT. The RCT tested the addition of a weight-loss medication (orlistat versus placebo) to BWL delivered in Spanish for 79 Latinx patients with obesity (BMI >30 kg/m²), with BED (N=40) and without BED (N=39). Minimal exclusionary criteria were applied (serious mental illness or anti-psychotic medication, suicidality, unstable medical conditions or unstable medication regimens, cardiac and neurologic disease) but did not exclude most psychiatric conditions. The study was approved by the Yale Institutional Review Board and all participants provided written informed consent.

Participants (65 women, 14 men) had a mean age of 46.32 years (SD=9.68) and a mean BMI of 37.57 (SD=6.62). Educational attainment was limited: 55% of participants reported less than high-school education, consisting mainly of grammar or middle-school education. Participants were diverse in their country of origin, representing Puerto Rico and several countries from Central/South America. Sixty-six percent (n=59) had at least one current additional psychiatric disorder and 75% (n=67) had at least one lifetime additional psychiatric disorder (based on diagnostic interviews; see below); 81% (n=72) were taking one psychiatric medication and 58% (n=52) were taking at least two psychiatric medications.

Treatment Conditions, Randomization, and Blinding Procedures

Behavioral Weight-Loss (BWL) Treatment.

All participants received four months of BWL in individual 60-minute weekly sessions. BWL was an adaptation of the Diabetes Prevention Program (28,30) four-month lifestyle behavioral treatment, which has been modified for broad implementation in community settings, available on public-domain website with both English and Spanish versions of materials, and found to be effective with most studies reporting 3%−7% weight-loss (28). BWL focuses on goal-setting for achieving sustainable lifestyle changes in healthy eating behaviors and nutrition (i.e., reducing dietary fat and total calories to a moderate deficit level), regular physical activity (i.e., 150 minutes/week), monitoring and tracking of eating and physical activity behaviors, and problem-solving to overcome barriers to achieve those lifestyle changes and achieving reasonable weight-loss.

We adapted the BWL based on the DPP (28,30) further to meet the specific cultural/ethnic characteristics of the Latinx population in New Haven (i.e., more heavily Puerto-Rican and South-American versus the more Mexican-focus of the DPP Spanish-version materials). BWL adaptations included handouts geared towards the Latinx community and culture-specific food props to teach healthy portion sizes and nutritional combinations; additionally, relevant cultural constructs/factors including the roles of family and community in eating informed the delivery of the intervention. BWL was delivered in Spanish by bilingual masters’/doctoral-level clinicians at the community center after being trained (by Drs. Grilo and White) and having collaborated in process of adapting the BWL to make it culturally responsive. In cases where reading ability was a concern, clinicians took flexible approaches and would read materials (plus involve family members) and develop other methods to track weekly goals.

Medication.

Participants were randomly assigned (stratified by BED diagnosis) to either orlistat (120 mg three times daily, fixed dose, [er previous studies (26)) or placebo (three times daily) delivered in double-blind fashion and using identical-appearing capsules prepared by the Yale Investigational Drug Service. Medication management meetings were performed by a bilingual psychiatrist and involved minimal clinical management (dosing compliance and adherence, side effects assessment and management) without any additional psychotherapeutic components.

Randomization to medication condition followed exact order after completing all baseline evaluations and medical clearance following the randomization schedule (with BED stratification) developed/maintained independently from investigators by Yale Investigational Drug Service. Medication double-blind was maintained throughout the study until after the last 6-month follow-up assessment was completed.

Diagnostic assessments and outcome measures

Bilingual masters/doctoral-level trained and monitored clinicians administered assessments in Spanish independently (and blinded) from the interventions. Participants completed Spanish-language versions of self-report measures. In instances of inadequate reading levels, research clinicians read the self-report questionnaires verbatim to participants and their responses were coded “as is” without further clarifications (unlike the case for semi-structured investigator-based interviews where probing was used as needed). Weight and height were measured at baseline and weight was measured throughout the study at all time points using a calibrated medical balance beam scale. Participants were assessed at baseline, monthly throughout treatment, at posttreatment (i.e., 4 months after randomization), and 6-month follow-up after completing treatment (i.e., 10 months after randomization).

Interviews.

Participants were categorized with or without current BED using DSM-IV criteria. BED diagnosis was based on Spanish-Structured Clinical Interview for DSM-IV Axis I Psychiatric Disorders (S-SCID-I/P) (31) and confirmed with Spanish-Eating Disorder Examination interview (S-EDE) (32). S-EDE, a semi-structured investigator-based interview with good psychometric characteristics (32), assesses frequency of binge-eating episodes (i.e., consuming unusually large quantities of food while experiencing subjective sense of loss of control). In the present study, inter-rater reliability (rho) coefficient, based on N=20 cases rated independently, was .85 for binge-eating.

Self-Report Questionnaires.

Spanish-Godin-Shepard Leisure-Time Exercise Questionnaire (S-GLTEQ) (33) assessed PA. S-GLTEQ asks participants to report the frequency of engaging in bouts of at least 15 minutes of mild (e.g., “yoga, easy walking”), moderate (e.g., “fast walking, easy bicycling”), and strenuous (e.g., “jogging, vigorous swimming”) activity during a typical week. Total scores are derived by an algorithm combining these categories (3 * Mild + 5 * Moderate + 9 * Strenuous); the multipliers for each category correspond to the metabolic equivalent values for activities falling into that category (the resulting score, while based on metabolic equivalent values, is in arbitrary units). Individuals’ scores can be used to classify PA level (<14 is insufficiently active, 14–23 is moderately active, and >23 is active). GLTEQ has displayed adequate test-retest reliability and validity for estimating PA levels and classifying adults into sedentary/active categories (33,34) including good support from validation studies using objective measures (activity monitors and fitness tests) (35).

Spanish-Beck Depression Inventory-II (S-BDI-II) is a 21-item self-report measures of depression symptoms/levels (36). Items are scored on 4-point scale (0=absent to 3=severe) to yield a total score ranging from 0 to 63, with higher scores reflecting greater severity. Established cutpoints are as follows: 0–13 minimal depression, 14–19 mild depression, 20–28 moderate depression, and 29–63 severe. S-BDI-II has good reliability and validity (36,37); in the present study, internal consistency was excellent (α=.91).

Statistical Analyses

Descriptive analyses examined PA levels across the timepoints (proportions categorized as: none, insufficiently active, moderately active, and active). Given zero-inflation in the PA variable, non-parametric tests were utilized where appropriate. Mann-Whitney U-tests examined differences in PA between the BED and non-BED groups at baseline. Wilcoxon signed rank test evaluated change in PA across study timepoints. Independent samples t-tests tested differences in changes in PA between the BED and non-BED groups and between those who received orlistat versus placebo. Associations between PA and outcome variables (percent weight-loss and depression (S-BDI-II) scores) were evaluated with repeated measures analyses of variance (ANOVAs) adjusting for BED status and medication condition. Baseline PA (dichotomized as no PA versus at least some PA) and change in PA (calculated continuously as baseline PA subtracted from posttreatment PA), as well as the interaction between these, were tested in each model. Missing PA data at posttreatment and follow-up were imputed using the last observed value (including from measurement at each monthly visit) where possible; individuals were excluded from analyses when this was not possible (n=5 individuals with missing PA values were dropped from analyses).

Results

Baseline Physical Activity and Associations with Clinical Variables

At baseline, overall PA levels (S-GLTEQ) were low, with 55.4% engaging in no PA (Figure 1). Baseline PA level did not differ significantly between the BED and non-BED groups (U=670.50, p=.89; Table 1). Within the BED group, PA was significantly negatively associated with BMI and binge-eating frequency, but not with depression score. In the non-BED group, PA was not significantly associated with BMI or depression score (Table 2).

Figure 1. Percentage of participants falling into four physical activity (PA) categories at each major assessment timepoint.

At baseline, more than half of the participants reported no physical activity (Spanish-Godin-Shepard Leisure-Time Exercise Questionnaire). The proportion reporting no physical activity was significantly reduced at posttreatment (i.e., 4 months after randomization) and at 6-month follow-up (i.e., 10 months after randomization) assessments. There were no significant differences between the groups with versus without binge-eating disorder in physical activity at any assessment timepoint.

Table 1.

Clinical variables at baseline, posttreatment (end-of-treatment), and at 6-month follow-up (10 months after baseline) overall and separately for groups with and without binge-eating disorder (BED).

	Overall				Binge-Eating Disorder				Non-Binge-Eating Disorder
	N	Mean	SD	Median	N	Mean	SD	Median	N	Mean	SD	Median
Phys Activity
Baseline	74	9.05	16.47	0.00	39	7.77	13.48	0.00	35	10.49	19.37	0.00
Posttreatment	77	25.47	28.81	18.00	39	28.38	29.71	21.00	38	22.47	27.92	15.00
Follow-up	77	17.59	23.17	10.00	39	16.87	20.71	9.00	38	18.33	25.72	11.00
BMI
Baseline	79	37.21	5.34	36.51	40	38.11	6.20	36.45	39	36.30	4.16	36.59
Posttreatment	74	36.16	5.55	34.65	38	36.83	6.05	35.76	36	35.44	4.95	34.42
Follow-up	72	36.35	4.76	35.91	37	36.59	5.56	35.88	35	36.08	3.77	35.93
Depression
Baseline	77	21.61	11.93	21.00	39	24.28	11.26	23.00	38	18.87	12.13	15.50
Posttreatment	74	12.19	12.27	8.50	38	14.53	12.26	11.00	36	9.72	11.96	6.00
Follow-up	72	12.24	11.71	9.00	37	15.73	12.17	13.00	35	8.54	10.11	4.00
Binge-eating
Baseline					39	12.49	9.34	10.00
Posttreatment					37	0.76	1.54	0.00
Follow-up					37	2.00	3.34	0.00

Notes: N = number; SD = standard deviation; Phys Activity = Physical Activity score on the Spanish-Godin-Shepard Leisure-Time Exercise Questionnaire (S-GLTEQ) (33); BMI = body mass index; Depression = depression score on Spanish-Beck Depression Inventory-II (S-BDI-II) (36); Binge-eating = binge-eating frequency (number of episodes past month) assessed by Spanish-Eating Disorder Examination interview (S-EDE) (33). Posttreatment = end-of-treatment (4 months after randomization); Follow-up = 6 months after completing treatments (10 months after randomization).

Table 2.

Correlations between clinical variables separately for groups with binge-eating disorder (BED) and without binge-eating disorder (non-BED) at baseline.

Group	Physical Activity	Body Mass Index (BMI)	Depression (S-BDI-II) Score
BED
BMI	−.33**
Depression	−.16	−.05
Binge eating	−.39**	.19	.71**
Non-BED
BMI	−.23
Depression	.05	.05

Notes: BED = binge-eating disorder; BMI = body mass index;Physical Activity = Physical Activity score on the Spanish-Godin-Shepard Leisure-Time Exercise Questionnaire (S-GLTEQ) (33); Depression = depression score on Spanish-Beck Depression Inventory-II (S-BDI-II) (36); Binge-eating = binge-eating frequency (number of episodes past month) assessed by Spanish-Eating Disorder Examination interview (S-EDE) (33)

^**denotes p < .01

Changes in Physical Activity

Figure 1 summarizes the percentage of participants falling into four physical activity categories at each major assessment timepoint overall and separately for the BED and non-BED groups. At baseline, 55.4% of participants were categorized with no PA with an additional 18.9% as insufficiently active; only 9.5% were classified as active. Descriptively, this translates to participants reporting an average of 1.25 (SD = 2.33) bouts (i.e., episodes of 15 or more minutes of activity at that intensity level) of mild activity, 0.85 (SD = 1.98) bouts of moderate activity, and 0.10 (SD = 0.57) bouts of strenuous activity in a typical week at baseline.

At post-treatment, 22.9% of participants were categorized with no PA with an additional 12.9% classified as insufficiently active; 32.9% were classified as active. Descriptively, this translates to participants reporting an average of 3.03 (SD = 3.12) bouts of mild activity, 2.16 (SD = 2.64) bouts of moderate activity, and 0.74 (SD = 1.97) bouts of strenuous activity in a typical week at posttreatment.

At 6-month follow-up, 38.0% of participants were categorized with no PA with an additional 14.1% classified as insufficiently active; 28.2% were classified as active. Descriptively, this translates to participants reporting an average of 2.10 (SD = 2.91) bouts of mild activity, 1.71 (SD = 2.64) bouts of moderate activity, and 0.39 (SD = 1.31) bouts of strenuous activity in a typical week at follow-up.

There were no significant differences at any study timepoint between BED and non-BED groups in proportion of participants classified in PA categories (χ²=4.87, p=.18 at baseline; χ²=4.06, p=.26 at posttreatment; χ²=3.00, p=.39 at follow-up) (Figure 1). Analyses revealed significant increases in PA (activity category) from baseline to posttreatment (z=4.93, p<.001) and significant declines in PA from posttreatment to 6-month follow-up (z=−2.44, p=.02), although PA levels at follow-up remained significantly greater than at baseline (z=2.93, p=.003).

Table 1 summarizes the levels of physical activity considered continuously at each major assessment point overall and separately for BED and non-BED groups. Changes in PA from baseline to posttreatment (M_BED=20.64 ± 32.00 and M_non-BED=12.40 ± 31.71, t(72)=−1.11, p=.27) and from posttreatment to follow-up (M_BED=−11.51 ± 26.21 and M_non-BED=−4.06 ± 24.66, t(72)=−1.29, p=.20) did not differ significantly between those with and without BED. Changes in PA from baseline to posttreatment (M_Med=14.41 ± 32.16 and M_Placebo=19.03 ± 31.96, t(72)=0.62, p=.54) and from posttreatment to follow-up (M_Med=−6.05 ± 23.21 and M_Placebo=−9.65 ± 28.11, t(72)=−0.61, p=.54) did not differ significantly between orlistat and placebo.

Associations between Changes in Physical Activity and Clinical Outcomes

Change in PA from baseline to posttreatment were not associated significantly with weight loss but was associated significantly with changes in depression (greater increases in PA were associated with greater reductions in depression scores) (Table 3). This association was moderated by baseline engagement in PA; specifically, the association was strongest for individuals who reported no PA at baseline (Figure 2).

Table 3.

Physical activity predicting changes over time in the outcome variables.

	N	F-test	p
Percent Weight Loss a	68
Medication		0.37	.69
Binge-Eating Disorder		0.35	.71
Physical Activity Category		0.52	.60
Change in Physical Activity		0.05	.96
Physical Activity Category-by- Change in Physical Activity		0.87	.42
Depression Score a	67
Medication		0.56	.57
Binge-Eating Disorder		0.68	.51
Physical Activity Category		0.42	.66
Change in Physical Activity		5.54	.01
Physical Activity Category-by- Change in Physical Activity		4.16	.02

Notes:

^aFindings from repeated-measures ANOVAs using Time (baseline, posttreatment, and 6-month follow-up) as a within-subjects factor of the variables predicting changes in two outcome variables (Percent Weight Loss and Depression Score on the Spanish-Beck Depression Inventory-II) (36).

Change in PA = posttreatment PA – baseline PA; Physical Activity Category refers to baseline classification as either no reported engagement in PA (0) or at least one bout of PA (1) on the Spanish-Godin-Shepard Leisure-Time Exercise Questionnaire (S-GLTEQ) (33)

Figure 2. The interaction between baseline engagement in physical activity (PA) and changes in PA predicting changes in depression.

Participants were divided into those who reported no activity at baseline and those reporting at least one bout of activity (Spanish-Godin-Shepard Leisure-Time Exercise Questionnaire). Change in PA is represented as one standard deviation above (high change) and below (low change) mean change in PA across the sample. Change in depression score (Spanish-Beck Depression Inventory-II) over the three assessment points was predicted by the interaction between these variables.

It was not possible to explore the relationship between changes in PA and changes in binge-eating over time within the BED group. Binge-eating frequency was significantly and robustly reduced with minimal variability in the outcomes (Table 1). Similarly, analyses with categorical binge-eating outcomes could not be performed; only two participants at posttreatment and eight participants at 6-month follow-up reported once-weekly or greater binge-eating frequency.

Discussion

This secondary analysis of a RCT (26) is the first study to evaluate treatment effects of weight-loss interventions on PA and their prognostic significance among Latinx individuals with obesity diagnosed with BED versus without BED. PA increased significantly overall among participants during treatments, with a greater than three-fold increase in individuals (from 9.3% to 32.9%) reporting PA levels that meet “active” levels on the S-GLTEQ measure at posttreatment. Although PA declined significantly between posttreatment (4 months after randomization) and 6-month follow-up after completing treatments (10 months after randomization), the declines (from 32.9% to 28.2% classified as “active”) were modest and the PA levels remained significantly greater than at baseline. The changes in PA during and after treatments did not differ significantly between those with BED versus without BED nor between orlistat and placebo medication condition. Increases in PA were not associated significantly with weight loss but were associated significantly with reduced depression scores.

The findings suggest that significant increases in PA can be achieved with a 4-month BWL (an adaptation of the Diabetes Prevention Program; (28,30)) intervention delivered to disadvantaged Latinx in a “real-world” community mental-health center. Moreover, increases in PA did not differ between patients with BED and without BED; these findings from this RCT are encouraging and challenge some literature suggesting that persons with occurring BED and obesity might benefit less from weight-loss treatments than those with obesity without BED (10). The finding that PA was not associated significantly with weight-loss outcomes converges with research suggesting weight loss is primarily driven by decreased caloric intake during treatment, rather than increased caloric expenditure through PA (38) whereas research suggests interventions promoting PA may result in improved longer-term weight-loss maintenance (39).

We observed that increases in PA were associated with significant reductions in depression scores over time, which is consistent with the literature linking PA and lower depression (40) and reduced depression (41). Importantly, the effect of PA increases on depression decreases throughout the treatment study and follow-up was greatest among participants who were inactive (i.e., categorized as “no PA”) at baseline. This “moderator” finding parallels findings from the Diabetes Prevention Program that significant health effects (reduced incidence of diabetes) of PA were greater among the participants who were least active at baseline (42). These findings are consistent with views that increasing PA levels by those who are most inactive may potentially have greater benefits (1,42).

Among individuals with BED, at baseline, PA was negatively associated with binge-eating frequency. This is consistent with findings from community studies suggesting that among those with BED, higher PA is associated with lower binge-eating frequency (43). Importantly, the increases in PA by patients with BED did not differ from those without BED. The small number of individuals with BED reporting binge-eating at post-treatment and follow-up assessments, however, precluded analyses testing associations with changes in PA. Pending future research, it seems reasonable to suggest the utility of increasing PA in persons with BED given the observed positive effects on depression levels, which we found to be highly correlated with binge-eating frequency.

This study has several limitations that should be considered as context for the findings. We examined self-reported PA which likely involves more error than objectively measured PA, any may vary by culture or age (44). Future research should utilize objective measurements of PA to more definitively determine changes in PA and how such changes are related to other clinical targets and outcomes among Latinx individuals with obesity with and without BED. Additionally, our sample size was relatively small; thus, while we were able to observe some significant associations, we may have been underpowered to detect smaller effects. Finally, our findings pertain to a group of Latinx patients characterized by substantial sociodemographic hardships and with concurrent psychiatric/addiction disorders who were seeking treatment and willing to participate in research at a community mental-health center. Thus, our findings may not generalize to Latinx groups characterized by fewer hardships and comorbidities or to those who do not seek treatment or are unwilling to participate in research (45).

This study addresses several important areas that are underrepresented in the literature. Notably, this study focused on Spanish-speaking-only Latinx individuals who are underrepresented in treatment research on BED (20) and obesity (19). Understanding the effects of existing treatments in culturally-diverse samples is imperative so that treatments can be refined and effectively disseminated more broadly. Additionally, evaluating specific behavioral changes that are components of complex interventions may contribute to a better understanding of ways to enhance treatment outcomes and to guide more targeted dismantling studies.

Collectively, our findings provide evidence suggesting that BWL can increase PA among Latinx patients with obesity with BED similar to patients with obesity with co-occurring BED and that the increases in PA are associated with significant reductions in depression scores. Importantly, our analyses revealed that the benefits of increased PA might be greatest amongst those who are most inactive to start. Future research should continue to build on these findings with studies that specifically address the role of PA in weight management and associated depression symptoms in Latinx individuals and in improved recognition and management of BED across demographic groups (45).

Study Importance Questions.

What is already known about this subject?

• Physical activity is key component of behavioral weight-loss treatment and is associated with weight-loss maintenance after treatment.

• Behavioral weight-loss treatment effects on physical activity in people with binge-eating disorder and in Latinx groups is uncertain.

What are the new findings in your manuscript?

• Physical activity was low at baseline, increased during treatment, declined after treatment to 6-month follow-up but remained elevated relative to baseline.

• Significant increases in physical activity, which did not differ between groups with versus without binge-eating disorder or between medication conditions, were associated with reduced depression but not weight loss.

How might your results change the direction of research or the focus of clinical practice?

• Culturally-adapted evidence-based behavioral weight-loss treatment can be effectively delivered in “real-world” settings serving socioeconomically disadvantaged Latinx patients with binge-eating disorder and with obesity.

• Future research should investigate methods to increase physical activity in Latinx groups with obesity including those with binge-eating disorder.