TL;DR
This study used individual participant data from 15 trials to investigate if specific eating behavior traits (EBTs) predict how well Acceptance and Commitment Therapy (ACT)-based weight management interventions work. They found that individuals with medium levels of internal disinhibition and emotional eating may benefit more from ACT compared to those with high levels, suggesting a potential for personalized treatment.
What Problem Does This Paper Address?
Obesity is a major global health issue with significant consequences. While various weight management interventions exist, their effectiveness varies greatly among individuals. The concept of "precision medicine" aims to tailor treatments to individuals based on their unique characteristics. Eating behavior traits (EBTs) like emotional eating, external eating, disinhibition, and restraint are thought to influence how people respond to weight management programs. Acceptance and Commitment Therapy (ACT) is a promising intervention that promotes psychological flexibility. However, it's unclear if ACT is more effective for individuals with certain baseline EBT levels, or if there's a "sweet spot" where ACT is most beneficial. This paper addresses this gap by using a powerful meta-analysis approach to explore if EBTs act as effect modifiers for ACT-based interventions.
Methods
The researchers conducted a one-stage Individual Participant Data (IPD) meta-analysis. This is a robust method where raw data from multiple studies are pooled and re-analyzed, allowing for more nuanced research questions than traditional meta-analyses. They systematically searched eight databases for eligible trials of ACT-based interventions for adults with a BMI ≥ 25 kg/m². Data from 15 trials (N=2535) were obtained, checked, and harmonized. They used mixed regression models to investigate both continuous and categorical interactions between baseline EBTs and the intervention (ACT vs. control) on weight outcomes. Categorical analyses involved stratifying EBTs into "low," "medium," and "high" levels based on tertiles of the pooled sample. Sensitivity analyses were performed to check the robustness of findings under different conditions (e.g., risk of bias, attendance levels).
Datasets
The study utilized data from 15 randomized controlled trials (RCTs) that met their eligibility criteria. These trials provided individual participant data (IPD) on weight outcomes and various eating behavior traits. The specific EBTs included were: emotional eating, external eating, disinhibition (general, internal, external), restraint (general, flexible, rigid), and uncontrolled eating. The data were collected from studies conducted primarily in the US, UK, Canada, Finland, Portugal, and Spain.
Key Insights and Ideas
- Non-linear effects are key: The most significant finding is that the relationship between EBTs and ACT effectiveness isn't linear. It's not simply "more of this trait means less benefit." Instead, there seems to be a "medium" level of certain traits where ACT is most beneficial.
- "Medium" is the magic number: For internal disinhibition and emotional eating, individuals with medium levels of these traits showed a greater benefit from ACT compared to those with high levels. This suggests that very high levels of these traits might indicate a need for more intensive or different types of support.
- ACT is generally effective, but precision matters: While ACT appears to be a suitable intervention on average, these findings highlight the potential for tailoring it based on EBTs to maximize outcomes for specific individuals.
- External eating is complex: While not a strong moderating effect, there was a trend suggesting a greater benefit for lower levels of external eating. However, the authors note that ACT didn't significantly change external eating, which might explain this unexpected direction.
Key Results
- No continuous moderation: When EBTs were analyzed as continuous variables, no significant moderating effects of ACT were found on percentage weight change.
- Categorical moderation for internal disinhibition: At 12 months post-intervention, participants with medium levels of internal disinhibition showed a significantly greater weight loss (-4.47% difference) compared to those with high levels.
- Sensitivity analysis support: Similar moderating effects for medium vs. high levels were observed in sensitivity analyses for emotional eating (especially in trials targeting experiential avoidance or using the TFEQ) and internal disinhibition (with higher attendance).
- General intervention effects: ACT interventions showed significant effects on reducing emotional eating, internal disinhibition, and uncontrolled eating, and increasing restraint from baseline to intervention end.
Limitations and Future Work
- Confirmatory trials needed: The authors emphasize that these findings are exploratory and require replication in larger, confirmatory trials.
- Generalizability: The sample was predominantly female and from high-income countries, limiting generalizability to other demographics and regions.
- Data limitations: The study was constrained by the variables available in the original trials, preventing the exploration of other potential moderators like therapist expertise.
- Self-report measures: EBTs were measured via self-report, which has inherent limitations.
- Future directions:
- Conduct large confirmatory trials with simpler statistical models.
- Include a broader range of bio-psycho-social factors.
- Facilitate data sharing for future IPD meta-analyses.
- Explore how EBTs interact to impact weight loss and derive meaningful cut-off scores for intervention allocation.
- Develop strategies to ensure adequate support for individuals with high levels of emotional eating or internal disinhibition.
Summary updated