BackgroundClosing the gap between research production and research use is a key challenge for the health research system. Stakeholder engagement is being increasingly promoted across the board by health research funding organisations, and indeed by many researchers themselves, as an important pathway to achieving impact. This opinion piece draws on a study of stakeholder engagement in research and a systematic literature search conducted as part of the study.Main bodyThis paper provides a short conceptualisation of stakeholder engagement, followed by ‘design principles’ that we put forward based on a combination of existing literature and new empirical insights from our recently completed longitudinal study of stakeholder engagement. The design principles for stakeholder engagement are organised into three groups, namely organisational, values and practices. The organisational principles are to clarify the objectives of stakeholder engagement; embed stakeholder engagement in a framework or model of research use; identify the necessary resources for stakeholder engagement; put in place plans for organisational learning and rewarding of effective stakeholder engagement; and to recognise that some stakeholders have the potential to play a key role. The principles relating to values are to foster shared commitment to the values and objectives of stakeholder engagement in the project team; share understanding that stakeholder engagement is often about more than individuals; encourage individual stakeholders and their organisations to value engagement; recognise potential tension between productivity and inclusion; and to generate a shared commitment to sustained and continuous stakeholder engagement. Finally, in terms of practices, the principles suggest that it is important to plan stakeholder engagement activity as part of the research programme of work; build flexibility within the research process to accommodate engagement and the outcomes of engagement; consider how input from stakeholders can be gathered systematically to meet objectives; consider how input from stakeholders can be collated, analysed and used; and to recognise that identification and involvement of stakeholders is an iterative and ongoing process.ConclusionIt is anticipated that the principles will be useful in planning stakeholder engagement activity within research programmes and in monitoring and evaluating stakeholder engagement. A next step will be to address the remaining gap in the stakeholder engagement literature concerned with how we assess the impact of stakeholder engagement on research use.Electronic supplementary materialThe online version of this article (10.1186/s12961-018-0337-6) contains supplementary material, which is available to authorized users.

A deep learning classifier for detecting seizures in neonates is proposed. This architecture is designed to detect seizure events from raw electroencephalogram (EEG) signals as opposed to the state-of-the-art hand engineered feature-based representation employed in traditional machine learning based solutions. The seizure detection system utilises only convolutional layers in order to process the multichannel time domain signal and is designed to exploit the large amount of weakly labelled data in the training stage. The system performance is assessed on a large database of continuous EEG recordings of 834h in duration; this is further validated on a held-out publicly available dataset and compared with two baseline SVM based systems.The developed system achieves a 56% relative improvement with respect to a feature-based state-of-the art baseline, reaching an AUC of 98.5%; this also compares favourably both in terms of performance and run-time.The effect of varying architectural parameters is thoroughly studied. The performance improvement is achieved through novel architecture design which allows more efficient usage of available training data and end-to-end optimisation from the front-end feature extraction to the back-end classification. The proposed architecture opens new avenues for the application of deep learning to neonatal EEG, where the performance becomes a function of the amount of training data with less dependency on the availability of precise clinical labels.

BackgroundPublic and patient involvement in healthcare research is increasing, but the impact of involvement on the individuals, on service delivery and on health outcomes, particularly in specialist population groups like critical care, remains unclear, as does the best way to involve people who have experienced critical illness. ObjectivesTo explore former patients' and family members' views and experiences of involvement in critical care research and/or quality improvement. MethodsUsing a qualitative methodology, semi-structured telephone interviews were conducted with seven former intensive care unit patients and three close family members, across England. Data were analyzed using a standard process of inductive thematic analysis. ResultsFour key themes were identified: making it happen; overcoming hurdles; it helps; respect and value.Findings centre on the need for flexibility, inclusivity and transparency. They further highlight the particular challenges faced by critical illness survivors and their family members in relation to research involvement, the importance of individualised support and training and the vital role that project leads have in making people feel valued and equal partners in the process DiscussionThis is the first study to explore patients' experiences of involvement in critical care research.Despite the small, homogenous sample, the study provides valuable and important data, to guide INVOLVEMENT IN RESEARCH AND QI 5 future practice. It highlights the need to enable and support people to make informed choices at a time when they are ready to do so. It further highlights the importance of gatekeepers, to avoid vulnerable people contributing before they are ready, a practice, which could negatively affect their heath status.

Amidst statutory and non-statutory calls for effective patient and public involvement (PPI), questions continue to be raised about the impact of PPI in healthcare services. Stakeholders, policy makers, researchers, and members of the public ask in what ways and at what level PPI makes a difference. Patient experience is widely seen as an important and valuable resource to the development of healthcare services, yet there remain legitimacy issues concerning different forms of knowledge that members of the public and professionals bring to the table, and related power struggles. This paper draws on data from a qualitative study of PPI in a clinical commissioning group (CCG) in the UK. The study looked at some of the activities in which there was PPI; this involved researchers conducting observations of meetings, and interviews with staff and lay members who engaged in CCG PPI activities. This paper explores power imbalances when it comes to influencing the work of the CCG mainly between professionals and members of public, but also between different CCG staff members and between different groups of members of public. The authors conclude that a hierarchy of power exists, with some professionals and public and lay members afforded more scope for influencing healthcare service development than others-an approach which is reflected in the ways and extent to which different forms and holders of knowledge are viewed, managed, and utilized.

This study presents a novel end-to-end architecture that learns hierarchical representations from raw EEG data using fully convolutional deep neural networks for the task of neonatal seizure detection. The deep neural network acts as both feature extractor and classifier, allowing for end-to-end optimization of the seizure detector. The designed system is evaluated on a large dataset of continuous unedited multichannel neonatal EEG totaling 835 hours and comprising of 1389 seizures. The proposed deep architecture, with samplelevel filters, achieves an accuracy that is comparable to the state-of-the-art SVM-based neonatal seizure detector, which operates on a set of carefully designed hand-crafted features. The fully convolutional architecture allows for the localization of EEG waveforms and patterns that result in high seizure probabilities for further clinical examination.Index Terms-neonatal seizure detection, convolutional neural networks, support vector machine, EEG waveforms.

AimThis paper aims to explore patient and public representation in a NHS clinical commissioning group and how this is experienced by staff and lay members involved.BackgroundPatient and public involvement is believed to foster greater public representativeness in the development and delivery of health care services. However, there is widespread debate about what representation is or what it should be. Questions arise about the different constructions of representation and the representativeness of patients and the public in decision‐making structures and processes.DesignEthnographic, two‐phase study involving twenty‐four observations across two types of clinical commissioning group meetings with patient and public involvement, fourteen follow‐up interviews with NHS staff and lay members, and a focus group with five lay members.ResultsPerceptions of what constitutes legitimate representativeness varied between respondents, ranging from representing an individual patient experience to reaching large numbers of people. Consistent with previous studies, there was a lack of clarity about the role of lay members in the work of the clinical commissioning group.ConclusionsUnlike previous studies, it was lay members, not staff, who raised concerns about their representativeness and legitimacy. Although the clinical commissioning group provides resources to support patient and public involvement, there continues to be a lack of clarity about roles and scope for impact. Lay members are still some way from constituting a powerful voice at the table.

EEG is the gold standard for seizure detection in the newborn infant, but EEG interpretation in the preterm group is particularly challenging; trained experts are scarce and the task of interpreting EEG in real-time is arduous. Preterm infants are reported to have a higher incidence of seizures compared to term infants. Preterm EEG morphology differs from that of term infants, which implies that seizure detection algorithms trained on term EEG may not be appropriate. The task of developing preterm specific algorithms becomes extra-challenging given the limited amount of annotated preterm EEG data available. This paper explores novel deep learning (DL) architectures for the task of neonatal seizure detection in preterm infants. The study tests and compares several approaches to address the problem: training on data from full-term infants; training on data from preterm infants; training on age-specific preterm data and transfer learning. The system performance is assessed on a large database of continuous EEG recordings of 575[Formula: see text]h in duration. It is shown that the accuracy of a validated term-trained EEG seizure detection algorithm, based on a support vector machine classifier, when tested on preterm infants falls well short of the performance achieved for full-term infants. An AUC of 88.3% was obtained when tested on preterm EEG as compared to 96.6% obtained when tested on term EEG. When re-trained on preterm EEG, the performance marginally increases to 89.7%. An alternative DL approach shows a more stable trend when tested on the preterm cohort, starting with an AUC of 93.3% for the term-trained algorithm and reaching 95.0% by transfer learning from the term model using available preterm data. The proposed DL approach avoids time-consuming explicit feature engineering and leverages the existence of the term seizure detection model, resulting in accurate predictions with a minimum amount of annotated preterm data.

No abstract