ProjectsNIHR Policy Reviews FacilityWork in progress
Policy Reviews Facility: Work in progress

These reviews are currently being conducted by the Policy Reviews Facility:

  • Interventions to increase active travel and improve health outcomes in different population groups

Active travel is defined as walking, cycling, wheeling (the use of mobility aids), or scootering activity, for the functional purpose of transport to a particular destination (i.e. getting from place to place), such as work, school or the shops (Public Health England 2016; Saunders et al, 2013). The overall aim of this research is to develop understanding of the potential public health impact of active travel interventions across different population groups, particularly those subject to disparities in health outcomes. This will be addressed via a multi-stage programme of work (see Figure overleaf). We will first produce a descriptive and high-level overview of the research evidence on differential impacts of active travel interventions (Evidence Map), followed by an in-depth assessment of their effectiveness (Effectiveness Review). The Evidence Map will help to refine the scope of the Effectiveness Review to make it optimally informative and efficient given available evidence, and contextualise the findings of the of the Effectiveness Review – e.g. by illustrating types of interventions for which data are not available. This programme of work aims to answer the following questions: (Evidence Map): What is the extent of current evidence evaluating active travel interventions targeting, or reporting outcomes for, different groups? (Effectiveness Review): What is the impact in different groups of evaluated interventions on i) active travel ii) health outcomes?

  • Generative Large Language Model-Based Tools for Health and Social Care Applications: A Living Map and Critical Review

Generative large language model- (LLM) based tools are a class of artificial intelligence (AI) tools capable of generating natural language text, images, audio1, or other media2 in response to user- or self-inputs (‘prompts’). These tools can generate content - that is, perform 'generative decoder operations’ - because generative LLMs can process (‘encode’) - and thereby ‘learn’ to predict (‘decode’) - the syntax, semantics, and patterns of natural language (or the patterns of images or audio): that is, they are capable of ‘next word prediction’.

Decisions about whether, and when, to adopt generative LLM-based tools for health and social care applications need to balance evidence for their benefits against the associated risks of harmful or unintended consequences. An informed and balanced understanding of the strengths and limitations of this technology is therefore essential to enable health and social care policy makers and practitioners to ask relevant questions and make informed adoption decisions about generative LLM-based tools, in order to unlock and effectively harness their benefits and safeguard against potential harms.  

This project will produce two main outputs. 

First, we will develop and continually maintain a living evidence and gap map (EGM) with the following objectives: 

  1. To maintain a surveillance of the landscape of accumulating research evidence (and gaps in the evidence base) for the use of generative LLM-based tools for health and social care applications; 

  1. To provide a regularly updated descriptive overview of the landscape of cumulative research evidence (and gaps in the evidence base) for the use of generative LLM-based tools for health and social care applications, classified in terms of its key features and characteristics; and 

  1. To make cumulative research evidence for the use of generative LLM-based tools for health and social care applications findable, accessible and reusable. 

  1. To compile a glossary of key terms and concepts relating to LLMs and LLM-based tools for health and social care applications 

Second, we will use the framework of a critical review to critically evaluate the literature and develop conceptual understanding of the use of LLM-based tools for specific tasks and emergent classes of health and social care applications.

Many reviews of the use of research evidence in public health decision-making have been suggestive of an underutilisation of research evidence in decision-making in local public health decision-making. Several potential barriers and facilitators to the utilisation of research evidence in decision-making have been identified. However, understanding how these can be addressed is challenging as we often lack detailed understandings of current practice and process models which identify evidence needs at different stages of decision-making for different types of decisions being made. Many of the issues surrounding the perceived underutilisation of research evidence may be as much a reflection of the ‘supply’ side and ensuring that research is produced in a way that can be used either instrumentally or for enlightenment, as much as a reflection of the ‘demand’ side and the need to stimulate engagement with research evidence among policy-makers. This work focusses on one type of ‘intervention’ – embedded researchers – and explores their potential in helping organisations to become more research active. We view organisations that are research active as those that are engaging with research evidence (i.e. using research evidence to inform their practice or decision-making) or generating research (i.e. producing research internally or commissioning research) or who are influencing research (i.e. influencing the conduct of research or contributing to influence research priorities).

See the report here (PDF).

Older people are more likely to be characterised by risk factors for loneliness including having poorer health, having a long-term illness or disability, living alone, and being widowed. During the current coronavirus crisis, millions of older people (70+) across the UK and elsewhere are advised to be particularly stringent about social distancing, and to avoid contact with those outside their household. This places older people at even higher risk of social isolation and loneliness. Social isolation and loneliness adversely affect quality of life, wellbeing and mental health, and are associated with physical ill health and mortality. However, what works to prevent or mitigate loneliness is less clear. The requirement for older people to restrict their activities during the COVID-19 pandemic has identified a need to understand how to minimise the impact of loneliness and isolation, at a distance.

A number of evidence reviews have highlighted the diverse range of interventions aimed to address and alleviate loneliness (and the consequences of loneliness) amongst older people in a variety of settings. In the main, these have been face-to-face interventions, either in groups or between individuals. Given the current ‘stay at home’ instructions from Government, these face-to-face interventions are not possible. Much of our social contact now has to be conducted over the telephone, or through use of videoconferencing tools.  We are conducting a rapid review exploring what are effective befriending, social support, and low intensity psychosocial interventions delivered remotely to reduce social isolation and loneliness among older people, and are attempting to understand how they ‘work. This work is being led jointly by the EPPI Centre and the Older People and Frailty PRU, with colleagues from the University of York/Hull York Medical School and National University of Ireland (Galway). We are conducting a review of reviews and are undertaking novel syntheses of the results including Narrative Synthesis, Intervention Component Analysis and Qualitative Comparative Analysis.

See the rapid review of systematic reviews report here.

If you wish to send us your comments on our work or be notified when these reviews are published please enter your details using the form below.

The Policy Reviews Facility is a collaboration between: EPPI Centre (Evidence for Policy and Practice Information Centre), UCL Institute of Education, University College LondonCRD (Centre for Reviews and Dissemination), University of York; and PHES (Public Health, Environments and Society), London School of Hygiene and Tropical Medicine.

  
Feedback
  
Copyright 2019 Social Science Research Unit, UCL Institute of Education :: Privacy Statement :: Terms Of Use :: Site Map :: Login
Home::About::Projects::Training::Research Use::Resources::Databases::Blog::Publications