Abstract
Background
Sex influences health and related behaviors due to biological and psychosocial/socioeconomic factors. Assessing sex-specific responses to integrated treatment for comorbid obesity and depression could inform intervention targeting.
Purpose
To test (a) whether sex moderates the effects of integrated collaborative care on weight and depression outcomes through 24 months and (b) whether treatment response at 6 months predicts 12 and 24 month outcomes by sex.
Methods
Secondary data analyses on weight and depression severity (SCL-20) measured over 24 months among 409 adults with obesity and depression in the Research Aimed at Improving Both Mood and Weight trial.
Results
Men achieved significantly greater weight reductions in intervention versus usual care than women, whereas women achieved significantly greater percentage reductions in SCL-20 than men at both 12 and 24 months. In logistic models, at 80% specificity for correctly identifying participants not achieving clinically significant long-term outcomes, women who lost <3.0% weight and men who lost <4.1% weight at 6 months had ≥84% probability of not meeting 5% weight loss at 24 months. Similarly, at 80% specificity, women who reduced SCL-20 by <39.5% and men who reduced by <53.0% at 6 months had ≥82% probability of not meeting 50% decrease in SCL-20 at 24 months.
Conclusions
Sex modified the integrated treatment effects for obesity and depression. Sex-specific responses at 6 months predicted clinically significant weight loss and depression outcomes through 24 months. Based on early responses, interventions may need to be tailored to address sex-specific barriers and facilitators to achieving healthy weight and depression outcomes at later time points.
Clinical Trial Registration
NCT02246413 (https://clinicaltrials.gov/ct2/show/NCT02246413).
Keywords: Sex difference, Treatment effect, Integrated collaborative care, Obesity, Depression
With integrated behavior therapy for comorbid obesity and depression, men had greater weight loss and women had greater depression reduction over 24 months.
Introduction
Sex as a biological variable that also has socioeconomic correlates has an impact, beyond genetics, on health and disease and represents a potentially important focus in the practice of personalized medicine in primary care [1]. Comorbid obesity and depression is a major public health problem for both sexes. Obesity similarly affects women (41%) and men (38%) in the USA [2]. Lifetime prevalence of major depressive disorder is significantly higher in women (26%) than men (15%) [3], as is the prevalence of subclinical depression [4]. Furthermore, a reciprocal positive relationship between obesity and depression exists and is especially common among women [5–7]. Women with obesity are 36% more likely to have depression than those with normal weight, while men with obesity are only 8% more likely to have depression than their normal-weight counterparts [7]. Similarly, obesity rates are higher in women (47%) and men (37%) with depression compared to those without depression (33% in each sex) [6]. In addition to differences in the prevalence and association of obesity and depression, women and men may differ in their preferences for behavioral interventions to address these conditions and may face different socioeconomic barriers to health behavior change [8, 9]. Given these differences, a better understanding of how the outcomes of integrated treatments for obesity and depression may differ by sex will inform the development of targeted interventions.
Because the coexistence of obesity and depression is associated with poorer adherence and response to standard treatment of either condition among patients with the comorbid condition versus those without [10], research is needed on integrated treatment for simultaneously addressing both conditions within one coordinated care protocol. The recent Research Aimed at Improving Both Mood and Weight (RAINBOW) trial is one of the few randomized clinical trials that tested an integrated collaborative care intervention and demonstrated its superiority to usual care in reducing body mass index (BMI; difference −0.7, 95% confidence interval [CI] −1.1 to −0.2; p = .01) and severity of depressive symptoms (−0.2, 95% CI −0.4 to −0.0; p = .01) at 12 months, the primary treatment endpoint [11]. Importantly, sex significantly modified the intervention effect at 12 months, with men achieving significantly greater BMI reductions than women and women achieving significantly greater reductions in depressive symptoms than men. These results suggest that targeted intervention strategies by sex may be warranted. However, this analysis of sex difference was limited to a duration of 12 months. Prior studies reported that early treatment response predicted long-term maintenance of weight loss [12–16] and reduced depressive symptoms [17, 18], but how this differs by sex is unknown. To inform intervention targeting by sex, it is important to further examine the role of sex in treatment effects over time.
Leveraging long-term follow-up data in the RAINBOW trial, the goal of this study is to examine [1] whether sex modifies the treatment effects on weight loss and depressive symptoms through 24 months and [2] whether a short-term response to the integrated collaborative care intervention at 6 months predicts clinically significant outcomes at 12 and 24 months according to sex.
Methods
The Institutional Review Board for Sutter Health approved the study. The trial protocol and baseline participant characteristics were previously published [19, 20].
Study Participants
Adult primary care patients were eligible if they had obesity (i.e., BMI ≥30 for non-Asians or ≥27 for Asians) and clinically significant depressive symptoms (i.e., nine-item Patient Health Questionnaire [PHQ-9] score ≥10) [21, 22]. The World Health Organization has suggested a BMI of 27.5 or greater as representing high health risks (e.g., cardiovascular disease) in Asian populations [23]. Given the ethnic differences in BMI and disease risk, participants of Asian descent were included in the present study if they had a BMI ≥27. Exclusion criteria included significant medical (e.g., diabetes or cardiovascular disease) or psychiatric comorbidities (e.g., psychosis, bipolar, or suicidality) or other major barriers to study participation (e.g., pregnancy, relocation, or non-English speaking).
All participants provided written informed consent and were randomized to receive the Integrated Coaching for Better Mood and Weight (I-CARE) intervention (n = 204) or usual care (n = 205). They received continued medical care from their personal physicians, information on routine obesity and depression care services, and wrist-worn activity trackers. The I-CARE intervention was yearlong and integrated the self-directed Group Lifestyle Balance (GLB) program for weight loss and the Program to Encourage Active, Rewarding Lives (PEARLS) for depression [19]. The GLB [24] program was adapted from the Diabetes Prevention Program [25], employed videos for self-study, and was previously demonstrated to be effective for weight loss and cardiometabolic risk reduction in primary care [26]. The PEARLS program [27, 28] used Problem Solving Therapy (PST) combined with behavioral activation strategies as first-line, plus as-needed intensification with antidepressant medications. The intensive treatment phase of the I-CARE intervention included 9 individual face-to-face sessions and 11 home-viewed GLB videos over 6 months. The maintenance phase included monthly phone sessions for six more months. The health coach administered the PHQ-9 at the beginning of each intervention session.
Measures
Participants were assessed at baseline and 6, 12, 18, and 24 months by study coordinators who were trained on the same protocols and blinded to random assignment [19]. The two coprimary outcomes were changes in BMI and 20-item Depression Symptom Checklist (SCL-20) scores. BMI was calculated as weight (kilogram) divided by height squared (meter2). Height was measured at baseline only, and weight was measured at baseline and each follow-up according to a standardized protocol [19]. In the case of missing study-measured weight, the closest-in-time electronic health record (EHR) weight within 3 months of the due date of a missed study visit or self-reported weight (if no EHR weight) was used. The SCL-20 is a valid, reliable measure of depression severity, with scores ranging from 0 to 4 [29]. Of the 409 randomized participants, 390 (95%), 367 (90%), 354 (87%), and 357 (87%) had weight data and 356 (87%), 344 (84%), 326 (80%), and 326 (80%) had SCL-20 data at 6, 12, 18, and 24 months, respectively (Supplementary Fig. S1 and Supplementary Table S1). Demographic variables (e.g., age, sex, race/ethnicity, and education), as well as self-reported binge eating disorder (BED) [30] and obesity-related psychosocial problems [31], were measured at baseline.
Statistical Analysis
Descriptive analyses used chi-square tests for categorical variables and Student’s t-tests for continuous variables to examine differences in baseline characteristics by sex. Analyses of between-group differences in outcomes over time by sex included participants with follow-up data at either 6, 12, 18, or 24 months for weight (203 of 204 intervention and 201 of 205 usual care) and SCL-20 (186 of 204 intervention and 189 of 205 usual care), respectively. Participants were analyzed based on the group to which they were randomly assigned. Estimates of the treatment effects for each time point by sex were obtained from Sex × Group × Time interactions in repeated-measures mixed-effects linear models. The fixed effects of each model included baseline values of the outcome, randomization covariates (i.e., clinic, age, sex, race/ethnicity, education, currently taking any antidepressant medication, and the number of hospitalizations in the past 12 months), group (intervention or usual care), time point (6, 12, 18, and 24 months), the two-way interactions of Group × Time and Group × Sex, and the three-way interaction of Sex × Group × Time. The random effects accounted for repeated measures with an unstructured covariance matrix and clustering of patients within primary care physicians. Analyses used all available data for each outcome from participants who had outcome data at one or more follow-up time points, and missing data on the outcomes were handled directly through maximum-likelihood estimation via mixed modeling. As exploratory analyses, we also included the baseline socioeconomic variables that were significantly different by sex (p < .05 in Table 1) in the mixed models to investigate whether any observed associations between sex and outcome are independent of or explained by these variables.
Table 1.
Baseline characteristics by sexa
| All | Female | Male | p value | |
|---|---|---|---|---|
| n = 409 | n = 287 | n = 122 | ||
| Age, year | 51.0 ± 12.1 | 51.4 ± 12.1 | 49.8 ± 11.9 | .21 |
| Race/ethnicity, % | .01 | |||
| Non-Hispanic White | 70.6 | 66.9 | 79.5 | |
| Non-Hispanic Black | 1.5 | 2.1 | 0.0 | |
| Asian/Pacific Islander | 9.8 | 9.0 | 11.5 | |
| Hispanic | 13.7 | 17.1 | 5.7 | |
| Other | 4.4 | 4.9 | 3.3 | |
| Education, % | .17 | |||
| High school/GED or less | 6.9 | 8.0 | 4.1 | |
| Some college | 24.0 | 26.1 | 18.9 | |
| College graduate | 36.7 | 35.2 | 40.2 | |
| Postcollege | 32.5 | 30.7 | 36.9 | |
| Annual family income, %, n = 365 | .03 | |||
| <$75,000 | 25.5 | 28.7 | 18.0 | |
| $75,000- <$150,000 | 32.0 | 33.1 | 29.7 | |
| ≥$150,000 | 42.5 | 38.2 | 52.3 | |
| Employment status, %, n = 407 | .02 | |||
| Full-time | 57.2 | 52.8 | 67.8 | |
| Part-time | 14.5 | 16.1 | 10.7 | |
| Unemployed | 28.3 | 31.1 | 21.5 | |
| Health insurance, % | .29 | |||
| Preferred provider organization | 65.5 | 63.1 | 71.3 | |
| HMO | 21.3 | 23.7 | 15.6 | |
| Medicare fee for service | 9.0 | 9.4 | 8.2 | |
| Other (Medicare HMO, medical, or self) | 4.2 | 3.8 | 4.9 | |
| Marital status, %, n = 406 | .054 | |||
| Married/living with a partner | 60.6 | 57.5 | 67.8 | |
| Single/separated/divorced/widowed | 39.4 | 42.5 | 32.2 | |
| Household size, %, n = 400 | .10 | |||
| <2 | 19.0 | 18.9 | 19.2 | |
| 2 | 36.0 | 32.9 | 43.3 | |
| 3+ | 45.0 | 48.2 | 37.5 | |
| Weight, kg | 103.3 ± 21.0 | 98.9 ± 19.0 | 113.8 ± 21.9 | <.001 |
| BMI, kg/m2 | 36.7 ± 6.4 | 37.0 ± 6.2 | 35.7 ± 6.6 | .06 |
| SCL-20 score | 1.5 ± 0.5 | 1.5 ± 0.5 | 1.4 ± 0.5 | .04 |
| Binge eating disorder, % | 40.8 | 44.3 | 32.8 | .03 |
| Obesity-related problem raw scoreb, n = 408 | 2.0 ± 0.7 | 2.1 ± 0.7 | 1.7 ± 0.7 | <.001 |
BMI body mass index; HMO health maintenance organization; SCL-20 Symptom Checklist-20.
aValues are mean ± standard deviation or percentages unless otherwise noted.
bAverage of eight questions with a range of 0–3. The higher the score, the more obesity-related psychosocial problems.
Logistic regression was used to predict clinically significant treatment response percentages (defined as a percentage decrease of baseline severity), that is, ≥5% weight loss and ≥50% decrease in SCL-20 scores [27, 28, 32] at 12 and 24 months separately; missing data were not imputed. Given the evidence on early treatment response predicting later treatment outcome [33, 34], the model predicting clinically significant weight loss at 12 and 24 months included the percentage of weight loss at 6 months, and the model predicting depression response included the percentage of SCL-20 change at 6 months without other covariate adjustments. As a sensitivity analysis, the percentage of change of intervention participants’ PHQ-9 scores closest to the 6 month time point was also used to predict ≥50% decrease in SCL-20 scores at 12 and 24 months separately in logistic models.
The logistic models were used to generate receiver operating characteristic (ROC) curve, sensitivity, specificity, positive predictive value, and negative predictive value. Then, specificity was set at 80% to assess sex-specific thresholds of weight loss (%) and SCL-20 reduction (%) at 6 months that reliably predicted clinically significant weight loss (i.e., ≥5% of baseline weight) and depression response (i.e., ≥50% decrease in SCL-20 scores) at 12 and 24 months separately. A specificity of 80% was chosen to ensure a good probability of correctly identifying participants not achieving longer-term outcomes of clinical significance [35].
All analyses were conducted using SAS, version 9.4 (SAS Institute Inc., Cary, NC). Statistical significance was set at p < .05 (two sided).
Results
Study Participants
Table 1 shows baseline characteristics by sex. Participants were primarily middle-aged (mean = 51.0, standard deviation [SD] = 12.1, years), women (70%), non-Hispanic White (71%), and at least college educated (69%). On average, participants had moderately severe obesity (mean BMI = 36.7, SD = 6.4) and moderate depression (mean PHQ-9 = 13.8, SD = 3.1; mean SCL-20 = 1.5, SD = 0.5). Compared with women, men were more likely to be non-Hispanic White and have higher annual family income. Men also had lower depressive symptoms than women. Compared with men, a higher percentage of women reported BED (44.3% vs. 32.8%, p = .03) and obesity-related psychosocial problems (2.1 vs. 1.7, p < .001).
Treatment Effects in the Overall Sample and by Sex
In the overall sample, the effects of the integrated collaborative care intervention as compared with usual care on BMI and depressive symptoms, which were significant through the end of intervention at 12 months as previously reported [11], did not persist at 18 and 24 months (Supplementary Table S2). Sex significantly modified the treatment effects on weight loss and depression outcomes over time (Fig. 1). Compared with women, men achieved significantly greater reductions in BMI at 6, 12, and 24 months in the intervention relative to usual care. Also, compared with usual care control, the intervention resulted in significant treatment effects in men but not in women at all time points. The differences in BMI change from baseline between the intervention and usual care groups were −1.6 (95% confidence interval [CI]: −2.2 to −1.0) in men and −0.2 (95% CI: −0.6 to 0.2) in women at 6 months, −1.8 (95% CI: −2.6 to −0.9) in men and −0.2 (95% CI: −0.7 to 0.4) in women at 12 months, −1.1 (95% CI: −2.2 to −0.0) in men and 0.1 (95% CI: −0.6 to 0.8) in women at 18 months, and −1.3 (95% CI: −2.5 to −0.1) in men and 0.2 (95% CI: −0.5 to 1.0) in women at 24 months (Fig. 1 and Supplementary Table S2). Similarly, the intervention led to significant differences in the percentage of weight loss compared with usual care control in men but not in women at all time points (Fig. 1 and Supplementary Table S2). Compared with men, women achieved significantly greater reductions in SCL-20 at 12 months in the intervention relative to usual care. Compared with usual care control, the intervention resulted in significant treatment effects in women at 6 and 12 months but not in men at any time point. The differences in SCL-20 score change from baseline between the intervention and usual care groups were −0.2 (95% CI: −0.4 to 0.1) in men and −0.3 (95% CI: −0.5 to −0.2) in women at 6 months, 0.1 (95% CI: −0.2 to 0.4) in men and −0.3 (95% CI: −0.5 to −0.1) in women at 12 months, 0.1 (95% CI: −0.2 to 0.4) in men and −0.1 (95% CI: −0.4 to 0.1) in women at 18 months, and 0.2 (95% CI: −0.1 to 0.5) in men and −0.2 (95% CI: −0.4 to 0.1) in women at 24 months (Fig. 1 and Supplementary Table S2). Similarly, the intervention led to significant differences in the percentage of SCL−20 change compared with usual care control in women at 6, 12, and 24 months but not in men at any time point (Fig. 1 and Supplementary Table S2). As only employment status and annual family income differed significantly by sex (Table 1), exploratory analysis results showed that including these additional covariates did not change the significance of the coefficient of sex differences in the mixed models for all the outcomes (results not shown).
Fig. 1.
Differences between the intervention and usual care control groups in body mass index change (A), the percentage of weight change (B), SCL-20 change (C), and the percentage of SCL-20 change (D) across follow-up time points among all participants and among men and women separately. Error bars indicate standard errors. *<0.05, **<0.01, ***<0.001 indicates significant sex differences in treatment effect at each follow-up time point.
Short-Term Thresholds at 6 months for the Prediction of Clinically Significant Weight Loss and Depression Response at 12 and 24 Months
Figure 2 shows ROC curves derived from the logistic models of the percentage of weight change and the percentage of SCL-20 change at 6 months in predicting clinically significant weight loss (≥5%) and depression response (≥50% SCL-20 reduction), respectively, at 12 and 24 months for both sexes. The areas under the ROC curve were similar for both sexes: 0.82 in women and 0.87 in men at 12 months and 0.67 in women and 0.65 in men at 24 months for predicting clinically significant weight loss; 0.72 in women and 0.77 in men at 12 months and 0.72 in women and 0.69 in men at 24 months for predicting clinically significant depression response. At 80% specificity, the threshold of the percentage of weight loss at 6 months to predict clinically significant weight loss at 12 and 24 months was 2.5% and 3.0%, respectively, in women and 3.4% and 4.1%, respectively, in men (Table 2). For example, if a woman did not achieve 2.5% weight loss at 6 months, there was a 90% probability (i.e., negative predictive value) that she would not achieve 5% weight loss at 12 months; whereas if a man did not achieve 3.4% weight loss at 6 months, there was a 94% probability that he would not achieve 5% weight loss at 12 months. Similarly, the threshold of the percentage of SCL-20 reduction at 6 months was 37.5% and 39.5% in women for predicting depression response at 12 and 24 months and 43.5% and 53.0% in men at 12 and 24 months, respectively. For example, if a woman did not achieve 37.5% SCL-20 reduction at 6 months, there was an 84% probability that she would not achieve clinically significant depression response at 12 months; whereas if a man did not achieve 43.5% SCL-20 reduction at 6 months, there was an 85% probability that he would not achieve clinically significant depression response at 12 months. Additionally, Supplementary Fig. S2 shows the linear relationship of the percentage of weight changes at 6 and 12 months and that at 6 and 24 months by sex. Supplementary Figure S3 shows the corresponding results for the percentage of SCL-20 changes at 6 and 12 months as well as at 6 and 24 months. The sensitivity analysis results showed that, among intervention participants only, the threshold of the percentage of PHQ-9 reduction at 6 months was 66.9% and 66.7% in women, and 70.1% and 70.0% in men, for predicting depression response (≥50% SCL-20 reduction) at 12 and 24 months, respectively (Supplementary Fig. S4 and Supplementary Table S3).
Fig. 2.
Receiver operating characteristic curves for the percentage of weight change or the percentage of SCL-20 change at 6 months in predicting the binary outcome of clinically significant weight loss or depression response at 12 and 24 months. Panel A: 6 month treatment response predicting clinically significant weight loss (i.e., ≥5% weight loss) at 12 months. Panel B: 6 month treatment response predicting clinically significant weight loss (i.e., ≥5% weight loss) at 24 months. Panel C: 6 month treatment response predicting clinically significant depression response (i.e., ≥50% decrease in SCL-20 scores from baseline) at 12 months. Panel D: 6-month treatment response predicting clinically significant depression response (i.e., ≥50% decrease in SCL-20 scores from baseline) at 24 months. AUC, areas under the receiver operating characteristic curve.
Table 2.
Thresholds of the percentage of weight loss and the percentage of SCL-20 reduction at 6 months to achieve clinically significant weight loss and depression response at 12 and 24 months
| Predicting outcome | Threshold at 6 months | Sensitivity | Specificity | Positive predictive value | Negative predictive value |
|---|---|---|---|---|---|
| Female | |||||
| ≥5% Weight loss at 12 months | −2.5 | 0.68 | 0.80 | 0.49 | 0.90 |
| ≥5% Weight loss at 24 months | −3.0 | 0.51 | 0.80 | 0.45 | 0.84 |
| ≥50% decrease in SCL−20 at 12 months | −37.5 | 0.54 | 0.80 | 0.48 | 0.84 |
| ≥50% decrease in SCL-20 at 24 months | −39.5 | 0.52 | 0.81 | 0.49 | 0.82 |
| Male | |||||
| ≥5% Weight loss at 12 months | −3.4 | 0.81 | 0.80 | 0.50 | 0.94 |
| ≥5% Weight loss at 24 months | −4.1 | 0.39 | 0.80 | 0.28 | 0.87 |
| ≥50% decrease in SCL-20 at 12 months | −43.5 | 0.64 | 0.80 | 0.56 | 0.85 |
| ≥50% decrease in SCL-20 at 24 months | −53.0 | 0.48 | 0.81 | 0.40 | 0.85 |
Discussion
To our knowledge, this is the first study that has investigated sex differences in treatment effects over time of an integrated collaborative care intervention for coexisting obesity and depression. The results showed that sex significantly modified the treatment effects on weight and depression outcomes over time. The first hypothesis was confirmed with men showing greater weight loss and women showing greater depression reduction through 24 months. The second hypothesis was confirmed with sex-specific short-term responses at 6 months identified that were predictive of clinically significant weight loss and depression outcomes at 12 and 24 months.
The results regarding sex differences in weight loss are in agreement with previous obesity studies showing that men tend to lose more weight than women in behavioral weight-loss interventions [36, 37]. It is also worth noting that healthy behavioral changes (e.g., healthy diet and physical activity), some of which were observed in the RAINBOW trial at 6 months [38], can result in health benefits (e.g., improved blood pressure and lipid levels) even in the absence of weight loss [39–41]. To a lesser extent, previous studies of cognitive-behavioral therapy or PST for depression also reported that women had better depression treatment response than men [42, 43]. This study extends the available evidence to indicate that sex may be an important targeting variable when designing integrated collaborative care interventions for comorbid obesity and depression.
Although the study participants were asked their biological sex and the analyses adjusted for commonly measured sociodemographic variables, it could not be determined whether and how the different treatment responses were related to sex as a biological variable that also has socioeconomic effects or psychosocial factors associated with gender. Additional research is needed to advance the understanding of these complex constructs in clinical interventions. The current intervention under study integrated the self-directed GLB program for weight loss and PEARLS for depression. A prior clinical trial of the GLB program showed that, among women, those randomized to a coach-led group intervention had significantly greater weight loss than those randomized to a self-directed video-based intervention similar to the weight loss component of the current study intervention, whereas men did comparatively well in both the coach-led and self-directed interventions [26]. These data suggest that women may benefit more from group-based weight-loss interventions and highlight the relative importance of peer support in helping women during the weight loss process. Therefore, future research could examine treatment delivery preferences, or choice, for a group-based compared with a self-directed weight loss approach or the impact of integrating peer support to help women lose weight. In addition, women reported a higher rate of BED and more obesity-related psychosocial problems than men at baseline in this study. As reported previously, a post hoc analysis in the RAINBOW trial found that BED significantly modified the intervention effect on BMI [11], suggesting that the collaborative care intervention tested may not have been as effective in participants with BED, with respect to weight loss [44]. Future work may consider adding cognitive-behavioral therapy strategies to reduce BED symptoms and obesity-related psychosocial problems and examine whether this helps this subset of women lose weight. Furthermore, 19% of women in this study were Hispanic or non-Hispanic Black compared to 6% in men. Prior studies have shown that Hispanics and Blacks exhibit smaller weight loss in behavioral interventions compared with Whites [45, 46]. Additional research is needed to investigate sex differences in weight loss within racial and ethnic subgroups.
Little is known about why men might have demonstrated worse depression treatment response than women in this integrated collaborative care intervention. In this study, men had significantly lower SCL-20 scores than women at baseline, which might leave less room for men to improve. In addition, internal consistency measured by Cronbach’s alpha was consistently high in sex and race/ethnicity subgroups (Supplementary Table S4); however, it is unclear whether the SCL-20 scale is able to reflect the different symptom profiles of depression between men and women. A previous study hypothesized that women benefit more from psychotherapy because they are generally more willing to express their problems and feelings with a coach than men, and men might benefit more from behavioral activation [43]. The study intervention used PST combined with behavioral activation strategies; however, the behavioral activation component may need to be strengthened to produce more benefits in men. Future studies should continue to investigate the reasons for this sex difference to inform how cognitive-behavioral interventions for depression can be targeted for men.
In addition, this study identified sex-specific short-term responses for predicting clinically significant outcomes in the long term. Previous behavioral weight loss studies have similarly reported early response thresholds (e.g., 2% weight loss at 1 month, 2.5% at 5 weeks, and 3% at 2 months) for predicting future clinically significant weight loss [12, 15]. Regarding depression, compared to our study, previous pharmacological studies reported lower thresholds of the percentage of decrease in the 17-item Hamilton Depression Rating Scale total score early in treatment (e.g., 20% decrease at 1–2 weeks and 15%–25% decrease at 4 weeks) for predicting future depression response or remission [17, 18]. However, none of these studies examined sex-specific thresholds, and few psychotherapeutic studies have examined this relationship. The current findings can inform future development of targeted treatment optimization strategies for men and women separately based on their early responses, which is the goal of this analysis. It is recognized that behavioral intervention should not be a “one-size-fits-all” solution. In this study, women lost weight (i.e., lower mean BMI compared to baseline) within both the intervention and control groups, but the between-group differences were not significant at any of the follow-up time points. Likewise, men improved their depression symptoms (i.e., lower mean SCL-20 scores compared to baseline) within both the intervention and control groups, but the between-group differences were not significant at any of the follow-up time points. Interventions that do not seem to benefit a specific group of people can be targeted before the treatment even begins or optimized at the earliest possible time during the treatment to address the group-specific barriers and, importantly, individual variability within the group to achieving better outcomes. Our results have practical implications because health coaches can identify men and women who do not achieve their sex-specific thresholds at 6 months, as well as those who are not on track to achieve these thresholds even earlier on in the intervention. Thus, health coaches could provide an augmented intervention with targeted strategies (e.g., peer support for women and behavioral activation for men) to minimize the risk of treatment failure with poor end outcomes. However, more research is still needed to investigate sex differences in treatment adherence and response. Future comparative effectiveness trials are also needed to evaluate the benefit of implementing targeted treatment optimization strategies early in the intervention.
This study has limitations. First, although sex was an a priori biological variable, the study was not adequately powered to detect treatment effect moderation by sex or treatment effect within sex. Second, the study sample was relatively homogeneous and had fewer male participants (30%). In addition, the RAINBOW participants were patients with comorbid obesity and elevated depressive symptoms but without serious comorbidities; therefore, the results may not be generalizable to those with other related chronic diseases (e.g., diabetes and coronary heart disease). Third, results regarding the sex-specific early treatment response predicting later clinically significant weight loss and depression outcome were derived from post hoc analyses. The findings are hypothesis generating, and future replication studies are needed. Finally, we assessed the thresholds only at 6 months and not at earlier time points. Future research could evaluate sex-specific thresholds at 1 or 2 months, as previous studies have suggested that the 1 or 2 month time frames are critical decision periods for treatment adaptations in order to enhance treatment efficacy for individual patients [12–14, 17, 18, 34].
Conclusion
In conclusion, this study reported that the treatment effects of an integrated collaborative care intervention for obesity and depression differed significantly by sex, with men showing greater weight loss and women showing greater depression reduction at 12 and 24 months. Sex-specific responses at 6 months were identified for predicting clinically significant weight loss and depression outcomes at 12 and 24 months. These results suggest that sex may be an important intervention-targeting variable for integrated weight loss and depression treatments.
Supplementary Material
Acknowledgments
We extend special thanks to the participants and their families who made this study possible.
Funding
Research reported in this publication was supported by the National Heart, Lung, And Blood Institute of the National Institutes of Health under Award Number R01HL119453 and UH2/UH3 HL132368. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Compliance With Ethical Standards
Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards J.M. is a paid scientific consultant for Health Mentor, Inc. (San Jose, CA). L.M.W. is on the Scientific Advisory Board for One Mind Psyberguide and the External Advisory Board for the Laureate Institute for Brain Research. O.A.A. is the cofounder of Keywise AI and serves on the advisory boards of Blueprint Health and Embodied Labs.
Authors’ Contributions N.L., L.X., and J.M. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: J.M., N.L., L.X., and L.G.R. Data acquisition and analysis: N.L., L.X., J.M. Data interpretation: All authors. Drafting of the manuscript: N.L., L.X., L.G.R., C.R.R., A.S.P., and J.M. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: L.X. Obtained funding: J.M., L.M.W., L.G.R., E.M.V., M.B.S., and M.A.L. Administrative, technical, or material support: J.M., L.X., L.G.R., N.L., M.B.S., E.M.V., M.A.L., and A.S.P. Supervision: J.M. and L.G.R.
Ethical Approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed Consent Informed consent was obtained from all individual participants included in the study.
Data Availability
We would make the deidentified participant data and associated documentation (e.g., data dictionaries and trial protocol) available to users only under a formal data sharing and use agreement that provides for a commitment to the following: (a) using the data only for research purposes and not to identify any individual participant, (b) securing the data using appropriate computer technology, (c) destroying or returning the data after analyses are completed, (d) accepting reporting responsibilities, (e) abiding by restrictions on the redistribution of the data for commercial purposes or to third parties, and (f) proper acknowledgment of the data resource. In addition, appropriate fees may be assessed upon mutual agreement on requests for information in a format other than that we intend to provide. We will not be responsible for providing any analytical support.
References
- 1. Bartz D, Chitnis T, Kaiser UB, et al. Clinical advances in sex- and gender-informed medicine to improve the health of all: A review. JAMA Intern Med. 2020;180:574–583. doi: 10.1001/jamainternmed.2019.7194. [DOI] [PubMed] [Google Scholar]
- 2. Hales CM, Fryar CD, Carroll MD, Freedman DS, Ogden CL. Trends in obesity and severe obesity prevalence in US youth and adults by sex and age, 2007–2008 to 2015–2016. JAMA. 2018;319:1723–1725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Hasin DS, Sarvet AL, Meyers JL, et al. Epidemiology of Adult DSM-5 major depressive disorder and its specifiers in the United States. JAMA Psychiatry. 2018;75:336–346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rodríguez MR, Nuevo R, Chatterji S, Ayuso-Mateos JL. Definitions and factors associated with subthreshold depressive conditions: A systematic review. BMC Psychiatry. 2012;12:181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. de Wit L, Luppino F, van Straten A, Penninx B, Zitman F, Cuijpers P. Depression and obesity: A meta-analysis of community-based studies. Psychiatry Res. 2010;178:230–235. [DOI] [PubMed] [Google Scholar]
- 6. Pratt LA, Brody DJ. Depression and obesity in the U.S. adult household population, 2005–2010. NCHS Data Brief. 2014;(167):1–8. PMID: 25321386. [PubMed] [Google Scholar]
- 7. Pereira-Miranda E, Costa PRF, Queiroz VAO, Pereira-Santos M, Santana MLP. Overweight and obesity associated with higher depression prevalence in adults: A systematic review and meta-analysis. J Am Coll Nutr. 2017;36:223–233. [DOI] [PubMed] [Google Scholar]
- 8. Liddon L, Kingerlee R, Barry JA. Gender differences in preferences for psychological treatment, coping strategies, and triggers to help-seeking. Br J Clin Psychol. 2018;57:42–58. [DOI] [PubMed] [Google Scholar]
- 9. Hernandez EM, Margolis R, Hummer RA. Educational and gender differences in health behavior changes after a gateway diagnosis. J Aging Health. 2018;30:342–364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Nigatu YT, Bültmann U, Schoevers RA, Penninx BWJH, Reijneveld SA. Does obesity along with major depression or anxiety lead to higher use of health care and costs? A 6-year follow-up study. Eur J Public Health. 2017;27:965–971. [DOI] [PubMed] [Google Scholar]
- 11. Ma J, Rosas LG, Lv N, et al. Effect of integrated behavioral weight loss treatment and problem-solving therapy on body mass index and depressive symptoms among patients with obesity and depression: The RAINBOW Randomized Clinical Trial. JAMA. 2019;321:869–879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Unick JL, Hogan PE, Neiberg RH, et al. ; Look AHEAD Research Group . Evaluation of early weight loss thresholds for identifying nonresponders to an intensive lifestyle intervention. Obesity (Silver Spring). 2014;22:1608–1616. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Unick JL, Leahey T, Kent K, Wing RR. Examination of whether early weight loss predicts 1-year weight loss among those enrolled in an Internet-based weight loss program. Int J Obes (Lond). 2015;39:1558–1560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Unick JL, Neiberg RH, Hogan PE, et al. ; Look AHEAD Research Group . Weight change in the first 2 months of a lifestyle intervention predicts weight changes 8 years later. Obesity (Silver Spring). 2015;23:1353–1356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Miller CK, Nagaraja HN, Weinhold KR. Early weight-loss success identifies nonresponders after a lifestyle intervention in a worksite diabetes prevention trial. J Acad Nutr Diet. 2015;115:1464–1471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Nackers LM, Ross KM, Perri MG. The association between rate of initial weight loss and long-term success in obesity treatment: Does slow and steady win the race? Int J Behav Med. 2010;17:161–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Funaki K, Nakajima S, Suzuki T, Mimura M, Uchida H. Early improvements in individual symptoms to predict later remission in major depressive disorder treated with Mirtazapine. J Clin Pharmacol. 2016;56:1111–1119. [DOI] [PubMed] [Google Scholar]
- 18. Gilaberte I, Romera I, Perez-Sola V, Menchon JM, Schacht A. Different levels of lack of improvement at 4 weeks of escitalopram treatment as predictors of poor 8-week outcome in MDD. J Affect Disord. 2013;146:433–437. [DOI] [PubMed] [Google Scholar]
- 19. Ma J, Yank V, Lv N, et al. Research aimed at improving both mood and weight (RAINBOW) in primary care: A type 1 hybrid design randomized controlled trial. Contemp Clin Trials. 2015;43:260–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Ma J, Xiao L, Lv N, et al. Profiles of sociodemographic, behavioral, clinical and psychosocial characteristics among primary care patients with comorbid obesity and depression. Prev Med Rep. 2017;8:42–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Arroll B, Goodyear-Smith F, Crengle S, et al. Validation of PHQ-2 and PHQ-9 to screen for major depression in the primary care population. Ann Fam Med. 2010;8:348–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. World Health Organization Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363:157–163. [DOI] [PubMed] [Google Scholar]
- 24. Kramer MK, Kriska AM, Venditti EM, et al. Translating the Diabetes Prevention Program: A comprehensive model for prevention training and program delivery. Am J Prev Med. 2009;37:505–511. [DOI] [PubMed] [Google Scholar]
- 25. Knowler WC, Barrett-Connor E, Fowler SE, et al. ; Diabetes Prevention Program Research Group . Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346:393–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Ma J, Yank V, Xiao L, et al. Translating the diabetes prevention program lifestyle intervention for weight loss into primary care: A randomized trial. JAMA Intern Med. 2013;173:113–121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ciechanowski P, Wagner E, Schmaling K, et al. Community-integrated home-based depression treatment in older adults: A randomized controlled trial. JAMA. 2004;291:1569–1577. [DOI] [PubMed] [Google Scholar]
- 28. Ciechanowski P, Chaytor N, Miller J, et al. PEARLS depression treatment for individuals with epilepsy: A randomized controlled trial. Epilepsy Behav. 2010;19:225–231. [DOI] [PubMed] [Google Scholar]
- 29. Glass RM, Allan AT, Uhlenhuth EH, Kimball CP, Borinstein DI. Psychiatric screening in a medical clinic. An evaluation of a self-report inventory. Arch Gen Psychiatry. 1978;35:1189–1195. [DOI] [PubMed] [Google Scholar]
- 30. Spitzer RL, Kroenke K, Williams JB. Validation and utility of a self-report version of PRIME-MD: The PHQ primary care study. Primary Care Evaluation of Mental Disorders. Patient Health Questionnaire. JAMA. 1999;282:1737–1744. [DOI] [PubMed] [Google Scholar]
- 31. Karlsson J, Taft C, Sjöström L, Torgerson JS, Sullivan M. Psychosocial functioning in the obese before and after weight reduction: Construct validity and responsiveness of the obesity-related problems scale. Int J Obes Relat Metab Disord. 2003;27:617–630. [DOI] [PubMed] [Google Scholar]
- 32. Katon WJ, Lin EH, Von Korff M, et al. Collaborative care for patients with depression and chronic illnesses. N Engl J Med. 2010;363:2611–2620. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Yank V, Xiao L, Wilson SR, Stafford RS, Rosas LG, Ma J. Short-term weight loss patterns, baseline predictors, and longer-term follow-up within a randomized controlled trial. Obesity (Silver Spring). 2014;22:45–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Lv N, Xiao L, Majd M, et al. Variability in engagement and progress in efficacious integrated collaborative care for primary care patients with obesity and depression: Within-treatment analysis in the RAINBOW trial. PLoS One. 2020;15:e0231743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Power M, Fell G, Wright M. Principles for high-quality, high-value testing. Evid Based Med. 2013;18:5–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Williams RL, Wood LG, Collins CE, Callister R. Effectiveness of weight loss interventions—Is there a difference between men and women: A systematic review. Obes Rev. 2015;16:171–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Wing RR, Hamman RF, Bray GA, et al. ; Diabetes Prevention Program Research Group . Achieving weight and activity goals among diabetes prevention program lifestyle participants. Obes Res. 2004;12:1426–1434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Rosas LG, Xiao L, Lv N, et al. Understanding mechanisms of integrated behavioral therapy for co-occurring obesity and depression in primary care: A mediation analysis in the RAINBOW trial. Transl Behav Med. 2020:ibaa024. [Online ahead of print] doi: 10.1093/tbm/ibaa024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Appel LJ, Moore TJ, Obarzanek E, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med. 1997;336:1117–1124. [DOI] [PubMed] [Google Scholar]
- 40. Appel LJ, Sacks FM, Carey VJ, et al. ; OmniHeart Collaborative Research Group . Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: Results of the OmniHeart randomized trial. JAMA. 2005;294:2455–2464. [DOI] [PubMed] [Google Scholar]
- 41. Ross R, Janiszewski PM. Is weight loss the optimal target for obesity-related cardiovascular disease risk reduction? Can J Cardiol. 2008;24(suppl D):25D–31D. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Davidson KW, Rieckmann N, Clemow L, et al. Enhanced depression care for patients with acute coronary syndrome and persistent depressive symptoms: Coronary psychosocial evaluation studies randomized controlled trial. Arch Intern Med. 2010;170:600–608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Donker T, Batterham PJ, Warmerdam L, et al. Predictors and moderators of response to internet-delivered Interpersonal Psychotherapy and Cognitive Behavior Therapy for depression. J Affect Disord. 2013;151:343–351. [DOI] [PubMed] [Google Scholar]
- 44. Ma J, Venditti EM, Snowden MB. Collaborative care for adults with obesity and depression—reply. JAMA. 2019;322:368. [DOI] [PubMed] [Google Scholar]
- 45. McCurley JL, Crawford MA, Gallo LC. Prevention of type 2 diabetes in U.S. hispanic youth: A systematic review of lifestyle interventions. Am J Prev Med. 2017;53:519–532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. West DS, Elaine Prewitt T, Bursac Z, Felix HC. Weight loss of black, white, and Hispanic men and women in the Diabetes Prevention Program. Obesity (Silver Spring). 2008;16:1413–1420. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
We would make the deidentified participant data and associated documentation (e.g., data dictionaries and trial protocol) available to users only under a formal data sharing and use agreement that provides for a commitment to the following: (a) using the data only for research purposes and not to identify any individual participant, (b) securing the data using appropriate computer technology, (c) destroying or returning the data after analyses are completed, (d) accepting reporting responsibilities, (e) abiding by restrictions on the redistribution of the data for commercial purposes or to third parties, and (f) proper acknowledgment of the data resource. In addition, appropriate fees may be assessed upon mutual agreement on requests for information in a format other than that we intend to provide. We will not be responsible for providing any analytical support.