Summary This is Part I of a three-part series on community empowerment as a route to greater health equity. We argue that community ‘empowerment’ approaches in the health field are increasingly restricted to an inward gaze on community psycho-social capacities and proximal neighbourhood conditions, neglecting the outward gaze on political and social transformation for greater equity embedded in foundational statements on health promotion. We suggest there are three imperatives if these approaches are to contribute to increased equity. First, to understand pathways from empowerment to health equity and drivers of the depoliticisation of contemporary empowerment practices. Second, to return to the original concept of empowerment processes that support communities of place/interest to develop capabilities needed to exercise collective control over decisions and actions in the pursuit of social justice. Third, to understand, and engage with, power dynamics in community settings. Based on our longitudinal evaluation of a major English community empowerment initiative and research on neighbourhood resilience, we propose two complementary frameworks to support these shifts. The Emancipatory Power Framework presents collective control capabilities as forms of positive power. The Limiting Power Framework elaborates negative forms of power that restrict the development and exercise of a community’s capabilities for collective control. Parts II and III of this series present empirical findings on the operationalization of these frameworks. Part II focuses on qualitative markers of shifts in emancipatory power in BL communities and Part III explores how power dynamics unfolded in these neighbourhoods.

No abstract

Speech Emotion Recognition (SER) is an important and challenging task for human-computer interaction. In the literature deep learning architectures have been shown to yield state-ofthe-art performance on this task when the model is trained and evaluated on the same corpus. However, prior work has indicated that such systems often yield poor performance on unseen data. To improve the generalisation capabilities of emotion recognition systems one possible approach is cross-corpus training, which consists of training the model on an aggregation of different corpora. In this paper we present an analysis of the generalisation capability of deep learning models using crosscorpus training with six different speech emotion corpora. We evaluate the models on an unseen corpus and analyse the learned representations using the t-SNE algorithm, showing that architectures based on recurrent neural networks are prone to overfit the corpora present in the training set, while architectures based on convolutional neural networks (CNNs) show better generalisation capabilities. These findings indicate that (1) cross-corpus training is a promising approach for improving generalisation and (2) CNNs should be the architecture of choice for this approach.

Background Locally delivered, place-based public health interventions are receiving increasing attention as a way of improving health and reducing inequalities. However, there is limited evidence on their effectiveness. This umbrella review synthesises systematic review evidence of the health and health inequalities impacts of locally delivered place-based interventions across three elements of place and health: the physical, social, and economic environments. Methods Systematic review methodology was used to identify recent published systematic reviews of the effectiveness of place-based interventions on health and health inequalities (PROGRESS+) in high-income countries. Nine databases were searched from 1st January 2008 to 1st March 2020. The quality of the included articles was determined using the Revised Assessment of Multiple Systematic Reviews tool (R-AMSTAR). Results Thirteen systematic reviews were identified - reporting 51 unique primary studies. Fifty of these studies reported on interventions that changed the physical environment and one reported on changes to the economic environment. Only one primary study reported cost-effectiveness data. No reviews were identified that assessed the impact of social interventions. Given heterogeneity and quality issues, we found tentative evidence that the provision of housing/home modifications, improving the public realm, parks and playgrounds, supermarkets, transport, cycle lanes, walking routes, and outdoor gyms – can all have positive impacts on health outcomes – particularly physical activity. However, as no studies reported an assessment of variation in PROGRESS+ factors, the effect of these interventions on health inequalities remains unclear. Conclusions Place-based interventions can be effective at improving physical health, health behaviours and social determinants of health outcomes. High agentic interventions indicate greater improvements for those living in greater proximity to the intervention, which may suggest that in order for interventions to reduce inequalities, they should be implemented at a scale commensurate with the level of disadvantage. Future research needs to ensure equity data is collected, as this is severely lacking and impeding progress on identifying interventions that are effective in reducing health inequalities. Trial registration PROSPERO CRD42019158309

This paper explores public health policy implementation through partnership working at the local level by examining how local actors from public health and the wider workforce, make sense of and work on social inequalities in health. An ethnographic case study was used to examine policy implementation in one local strategic partnership in north-west England during a period of significant resource constraint. Semi-structured interviews were the primary method of data generation. Sensitising concepts from figurational sociology were used to develop a theoretical account of how local policy implementation directed at narrowing social inequalities in health tended to give rise to relatively fragmented and short-term services, projects and practices, which focused on lifestyle factors and behaviour change. Theorising partnership work as figurations goes some way to explaining the apparent paradox among participants who expressed a relatively detached appreciation of wider social influences, alongside emotional involvement in their work. This process of individualisation explains how local professionals tended to conceptualise health inequality and the social determinants of health as personal troubles. Individualisation meant that the social reality of working in partnerships on difficult issues was simplified. Thus, any scope for working on the social determinants of health tended to be overlooked. The extent to which this was intentional or a matter of struggling to see opportunities, or a mixture of the two, was difficult to discern. Although the policy landscape has changed, the findings give some insight into understanding how local collaborative processes reproduce local public health work underpinned by lifestyle choices.