2. Platz Fabian Schwendinger (Universität Basel):
Accelerometer-Based Physical Activity Surveillance in Healthy and Diseased Adults
Background: Accelerometers provide previse physical activity (PA) assessment, yet data quality depends on valid wear protocols and interpretability. Continuous, cut-point-free metrics capture the full 24-hour intensity spectrum, distinguishing PA volume from intensity. However, their interpretability remains limited without reference values, and relative intensity measures are lacking. This PhD project aimed to advance accelerometer-based PA surveillance by refining methodology, enhancing interpretability through reference values, and exploring how PA volume, intensity, and duration relate to health outcomes across healthy and diseased populations. Methods: Data from the COmPLETE study (healthy adults and patients with heart failure) involved 14-day wrist-worn accelerometry and cardiopulmonary exercise testing. Additional analyses used 7-day accelerometer data from the 2011–2014 NHANES cohort. Metrics from the R-package GGIR included average acceleration (AvAcc, PA volume) and intensity gradient (IG, PA intensity), expressed in absolute (_ABS) and relative (_REL) terms, plus PA fragmentation (MXRATIO). Associations with cardiorespiratory fitness and mortality were examined using regression and Cox proportional hazards models. Results: Cut-point-based metrics frequently misclassified PA intensity. Both AvAcc_ABS and IG_ABS independently predicted cardiorespiratory fitness, with intensity showing stronger associations than volume. Reference values and centile curves were generated for PA metrics. IG_ABS showed curvilinear inverse associations with all-cause and cardiovascular mortality, with intensity emerging as the primary driver. Conclusion: This thesis advances accelerometer methodology, supporting cut-point-free metrics for accurate, interpretable, and comparable PA assessment. Intensity appears key to health benefits, and new reference values enable practical application in research, clinical, and public health settings.
3. Platz Bryan Charbonnet (Universität Bern):
Problem-oriented talent research: Between striving for understanding and utility
The global rise of elite sport over the last 50 years has markedly heightened the importance of talent selection and development processes, prompting many nations to invest substantial resources in corresponding strategies. These processes typically begin in childhood or adolescence, aiming to allocate limited resources to the most promising athletes. Inefficient systems not only waste resources and weaken international competitiveness but also impose considerable personal and financial burdens on families and individuals pursuing unlikely careers. Addressing the challenge of predicting and optimizing human development, this dissertation adopts an interdisciplinary and problem-oriented research approach toward a “better sport,” pursuing both the quest for fundamental understanding and the consideration of use. The first paper advances theoretical understanding by developing a metatheoretical framework for talent research. It demonstrates how implicit fundamental assumptions, i.e., the underlying worldviews about “how we think”, shape methodological choices, define epistemic boundaries, and reveal blind spots in scientific inquiry. The second paper bridges theory and practice by critically revisiting the enduring dichotomy between early specialization and sampling. By reframing this debate, it exposes divergent goals and perspectives that have hindered consensus and introduces a new theoretical model of youth talent development, empirically illustrated and offering practical implications for sport federations. The third paper emphasizes the consideration of use, showing that algorithmic corrections for biological age distortions can have prognostic validity and thus serve as a valuable source of information for improving talent identification. Overall, the dissertation contributes both scientifically and socially by critically questioning prevailing assumptions and advancing conceptual, methodological, and practical discourse toward a “better” talent identification and development.
