This habilitation thesis, divided into three parts, covers my main research areas: 1) factors influencing weight loss and energy compensation in lifestyle interventions, 2) technology-assisted methods for assessing and monitoring dietary intake, and 3) the health effects of large lifestyle interventions. The thesis encompasses a total of 14 publications. PART I – Predictors of Energy Compensation and Weight Loss in Lifestyle Interventions More than 1.9 billion adults worldwide are overweight, with 650 million having obesity. Excess body fat increases the risk of serious health conditions and premature mortality, creating public health and economic challenges. Lifestyle modifications involving caloric restriction and physical activity are the first-line approach to treat obesity. However, many individuals lose less weight than expected, particularly with exercise interventions, due to a phenomenon termed “weight compensation," where increased energy intake negates the exercise-induced energy deficit. Since weight loss is a key motivator, its absence often leads to program discontinuation. Identifying predictors of weight loss success is crucial to guide treatment and support patients to improve long-term outcomes. In Publication 1, we showed that biological and behavioral characteristics predict total and compensatory post-exercise energy intake differently in men (primarily fasting concentrations of appetite-regulating hormones) and women (primarily habitual exercise behavior). In Publication 2, we showed that lower baseline levels of moderate-to-vigorous physical activity are associated with less weight loss, higher compensation, and increased energy intake during a 24-week supervised aerobic exercise intervention. In Publication 3, we focused on dynamic intervention-specific factors as predictors of weight loss in an exercise intervention. We found that less initial (4-week) weight loss is associated with longer-term attenuated weight loss and greater compensation during 6 months of aerobic exercise training. In Publication 4, an invited review article, we aimed to answer the question of why exercise by itself is a relatively ineffective method for weight loss, discussing energy-related aspects of weight loss and exercise effects on eating behavior. In Publication 5, we tested if initial weight loss could predict longer-term weight loss in a 2-year primary care-based intervention. We found that greater initial (2-, 4-, and 8-week) weight loss, daily self-weighing adherence, and adherence to the expected weight loss trajectory predicted longer-term weight loss at 6, 12, and 24 months. In Publication 6, we went beyond predictor analyses and aimed to identify the mediators of weight change, showing that several psychological (dietary restraint, disinhibition) and behavioral (fat and fruit/vegetable intake, PA) variables mediated weight change during a 2-year pragmatic weight loss intervention. In Publication 7, we examined genotype-diet interactions on weight loss, demonstrating that with the current ability to genotype participants as fat- or carbohydrate-responders, evidence does not support greater weight loss on genotype-concordant vs. genotype-discordant diets.   PART II – Technology-Assisted Food Intake Assessment Accurate food intake measurement is crucial for assessing diet-health interactions, monitoring dietary changes in obesity treatment, and informing public health policies. Despite their inaccuracies, self-report methods (e.g., food records, recalls, and frequency questionnaires) remain common in research, often leading to energy and nutrient intake assessment errors. Image-assisted food intake assessments using automated or semi-automated analysis have recently gained popularity, addressing many limitations of traditional self-report methods, as we showed in a literature review (Publication 8). In Publication 9, we further showed that beyond improved accuracy, image-assisted methods of food intake assessment also have greater feasibility and acceptability in different research settings and for day-to-day dietary monitoring. Finally, in Publication 10, we reviewed continuous glucose monitoring-based approaches regarding their accuracy in meal detection as standalone and combined methods. We thereby evaluated the applicability of these methods for Just-In-Time Adaptive Interventions, aiming to detect dietary lapses and non-adherence in (near) real time and deliver intervention content when it is most needed, and the patient is most receptive. PART III – Other Work Part III highlights further research of my habilitation work and postdoctoral training, such as examining the health effects of large lifestyle interventions beyond weight loss, discussing challenges in defining adherence to calorie restriction goals in weight loss interventions, and gaining further expertise in food intake research. In Publication 11, we showed that a pragmatic and scalable obesity treatment program delivered by health coaches to over 800 primary care patients can elicit clinically meaningful improvements in cardiometabolic health. This finding suggests that such a program could offer a viable and more successful primary care-based treatment option for obesity-related comorbidities compared to the existing Medicare/Medicaid model in the United States, potentially influencing healthcare practices. In Publication 12, we delved deeper into the health effects of structured exercise, showing that exercisers who compensated (i.e., lost less weight than expected) during a 24-week exercise intervention exhibited unfavorable increases in waist circumference and visceral adipose tissue compared to those who did not compensate. In Publication 13, we aimed to tackle a common challenge of dietary weight loss interventions, namely that of defining successful adherence to calorie restriction goals. For this, we determined the level of calorie restriction associated with the zone of adherence (i.e., the expected weight loss trajectory) in CALERIE 2 by utilizing a validated weight loss calculator and assessed if participants' actual level of calorie restriction was within the zone of adherence by using the intake-balance method. Our results showed that the zone of adherence in CALERIE 2 considered calorie restriction far less than the 25% goal as being adherent, which should be considered in designing future calorie restriction interventions and strategies to promote adherence. Finally, in Publication 14, as part of my postdoctoral training in food intake research, I designed a randomized crossover study measuring food intake during a test meal with acute e-cigarette use (“vaping”) as a manipulator. We showed that while acute e-cigarette use increased subjective feelings of satiety and decreased subjective feelings of hunger, these subjective effects did not translate into reductions in acute energy intake, which differs from previous findings and warrants further research.      «

This habilitation thesis, divided into three parts, covers my main research areas: 1) factors influencing weight loss and energy compensation in lifestyle interventions, 2) technology-assisted methods for assessing and monitoring dietary intake, and 3) the health effects of large lifestyle interventions. The thesis encompasses a total of 14 publications. PART I – Predictors of Energy Compensation and Weight Loss in Lifestyle Interventions More than 1.9 billion adults worldwide are overweight...     »