In order to cope with current disruptive technical and societal transfor-mations, e.g. through digitalization or AI, new competences, commonly called future skills, are indispensable for everyday as well as for professional life. Many organizations, large and small, work on defining a set of future skills. This might imply that future skills are generic and identical across all professions. In contrast, it is a consensus in pedagogical research that generic competences are spe-cifically shaped by the professional environment. Clearly, these two positions contradict each other. But what does this contradiction mean for future skills?
This research rests on the assumption that these context-sensitive generic competences including future skills are developed differently for each occupational field. In order to be able to offer target- and competence-oriented teaching, target competences must be known in the first place, taking into account the specific professional characteristics, currently and in the future.
This paper provides evidence that the initial assumption of context-specific competences is true by collecting and comparing qualitative research data. To do so, qualitative data was collected for different occupations, in particular software engineering and health care.
Our research shows that in fact each profession expresses competences specifically, and this applies to technical as well as non-technical competences.
The details of relevant competences need to be identified and characterized as a prerequisite for being able to devise and offer competence-oriented learning approaches.

BACKGROUND: FSGS affects the complex three-dimensional morphology of podocytes, resulting in loss of filtration barrier function and the development of sclerotic lesions. Therapies to treat FSGS are limited, and podocyte-specific drugs are unavailable. To address the need for treatments to delay or stop FSGS progression, researchers are exploring the repurposing of drugs that have been approved by the US Food and Drug Administration (FDA) for other purposes. METHODS: To identify drugs with potential to treat FSGS, we used a specific zebrafish screening strain to combine a high-content screening (HCS) approach with an in vivo model. This zebrafish screening strain expresses nitroreductase and the red fluorescent protein mCherry exclusively in podocytes (providing an indicator for podocyte depletion), as well as a circulating 78 kDa vitamin D-binding enhanced green fluorescent protein fusion protein (as a readout for proteinuria). To produce FSGS-like lesions in the zebrafish, we added 80 µ M metronidazole into the fish water. We used a specific screening microscope in conjunction with advanced image analysis methods to screen a library of 138 drugs and compounds (including some FDA-approved drugs) for podocyte-protective effects. Promising candidates were validated to be suitable for translational studies. RESULTS: After establishing this novel in vivo HCS assay, we identified seven drugs or compounds that were protective in our FSGS-like model. Validation experiments confirmed that the FDA-approved drug belinostat was protective against larval FSGS. Similar pan-histone deacetylase inhibitors also showed potential to reproduce this effect. CONCLUSIONS: Using an FSGS-like zebrafish model, we developed a novel in vivo HCS assay that identified belinostat and related pan-histone deacetylase inhibitors as potential candidates for treating FSGS.

OBJECTIVE: Applications of machine learning in healthcare are of high interest and have the potential to improve patient care. Yet, the real-world accuracy of these models in clinical practice and on different patient subpopulations remains unclear. To address these important questions, we hosted a community challenge to evaluate methods that predict healthcare outcomes. We focused on the prediction of all-cause mortality as the community challenge question. MATERIALS AND METHODS: Using a Model-to-Data framework, 345 registered participants, coalescing into 25 independent teams, spread over 3 continents and 10 countries, generated 25 accurate models all trained on a dataset of over 1.1 million patients and evaluated on patients prospectively collected over a 1-year observation of a large health system. RESULTS: The top performing team achieved a final area under the receiver operator curve of 0.947 (95% CI, 0.942-0.951) and an area under the precision-recall curve of 0.487 (95% CI, 0.458-0.499) on a prospectively collected patient cohort. DISCUSSION: Post hoc analysis after the challenge revealed that models differ in accuracy on subpopulations, delineated by race or gender, even when they are trained on the same data. CONCLUSION: This is the largest community challenge focused on the evaluation of state-of-the-art machine learning methods in a healthcare system performed to date, revealing both opportunities and pitfalls of clinical AI.