NIHR Global Health Research Unit on Improving Health in Slums at University of Warwick

This unit will be led by the Warwick International Centre for Applied Health Research and Delivery (W-CAHRD), University of Warwick (UoW). Its staff and facilities will form the foundation of the proposed unit. W-CAHRD was set up in 2014 with core UoW investment, in line with the institution's strategic aim to increase international collaboration and support world-class global health research. Since inception, W-CAHRD has won 21 externally funded research projects worth £12.8 million and published 100+ peer reviewed global health papers, including a Lancet series on health in slums2,3. UoW provides core funding for a Global Research Priorities (GRP) programme to address the most challenging problems facing the world and to provide a platform for interdisciplinary research. W-CAHRD is represented on the programme and addresses 3 of the GRP's 11 themes: sustainable cities, health and international development. UoW will contribute matched funds and support unit sustainability (see justification of costs). Our unit will focus on how health services are delivered and used in slums, home to nearly a billion people, and identify options to improve affordable access for this group. Access to healthcare is a critical barrier to improved health in LMICs4. Universal Health Coverage (UHC) is a key element of the current global health agenda. It requires a set of physically accessible and financially affordable service providers, and a reduced burden of out-of-pocket expenditure5. All UN Member States have agreed to try to achieve UHC by 2030. Studying disease epidemiology is unhelpful if infrastructure is not in place to address it6. In the context of rapid urbanisation, poor economic growth, urban planning and regulation, slum populations are increasing. Slums, generally situated near to urban centres, are often physically closer to health services than rural settlements. Despite this, people in slums can have poorer health outcomes than their rural counterparts. This is due to the hazardous slum environment and poor access to appropriate healthcare2. In this vulnerable and marginalised group, child and maternal mortality remain high, as do infectious disease deaths, while non-communicable disease risk is growing2,7. Health emergencies are magnified in the slum environment. In the recent Ebola epidemic, the disease was concentrated in slums, and infected residents facilitated transmission throughout the city8. Poor people in LMICs are prone to catastrophic financial loss if they fall ill, and in slums, out of pocket expenses are essential to maintain life (e.g. to buy clean water)9. Observational studies of maternity care in Nairobi have documented a near absence of public facilities in slums. Private providers of varying quality fill the void8. Just 13.9% of people living in Dhaka's slums seek care from modern public providers, and coverage gaps exist10,11. Fragmented services may lead to poor co-ordination of care, reducing efficiency and resilience e.g. over-provision and/or under-provision of some services; reduced continuity of care; increased probability of developing drug resistance; reduced opportunities for disease surveillance. In slums, proximity to urban centres and population density mean that improvements to health service delivery could benefit many people simultaneously and have a large impact on health in LMICs. Our unit will seek to advance this issue. Objectives: Short-term To map geo-spatially current health service delivery arrangements and understand patterns of health service use (including equity of use) in slums in major and secondary cities. Medium-term To identify costs associated with different models of health service delivery arrangements in these slums, including by whom costs are incurred. Long-term To model options for health service delivery in slums, considering quality, cost-effectiveness and equity of provision. To develop capacity, communities of practice and a sustained research programme which exceeds the lifetime of the unit. Throughout To identify, synthesise and curate literature on potential models of health service delivery relevant to the slum context. To engage decision-makers and users in designing models of health service delivery in slums, with a view to subsequent evaluation of effectiveness and costs of viable options. To achieve the objectives work will be structured in 5 packages (WPs see Fig2,3). We have identified primary study sites, located in Kenya, Nigeria, Bangladesh and Pakistan (Fig1). These sites provide variety in social and physical factors, allowing us to examine diverse existing models of healthcare delivery. WP1: Geo-spatial mapping of health services in slums (Years 1-2). a)We will create accurate maps of the slums using participatory mapping instruments that combine local information to 'ground truth' data generated from Earth-observation satellite imagery. This methodology has worked in slums, with lay groups collecting data for the purposes of informing humanitarian aid organisations12. These maps have 2 functions: they will form the basis on which to overlay geo-spatial data on health service location (see b); they will inform our sampling frame in WP2. b)Local participatory mapping will be used to identify and add all health services present in the slum or available to people living in the slum to our maps, including hospitals, clinics, traditional healers, pharmacies etc. With our augmented map, we can calculate the area covered by mapped facilities and identify blind spots ; areas that are not covered or have a high distance/travel time to existing services. WP2: Household survey of health service use by slum residents (Years 1-2). A survey of health service use will be conducted with a stratified random sample of people living in the examined slums. Health service use will be investigated using questions adapted from a validated questionnaire (e.g.13). We will ask where healthcare has been sought by participants over the past year; for what health conditions; unmet healthcare needs; satisfaction with care received (a facet of 'quality'); and associated costs, including out-of-pocket costs. Data will be cross-referenced with WP1 to identify health services used by people living in the slums that are not recorded in the mapping, such as informal providers working from unmarked premises, services delivered at home (e.g. by community health workers) and those delivered outside the slums (e.g. secondary or tertiary care which is likely to be exclusively outside the slum area). Surveys will be georeferenced, so we can examine distance/travel time to health services. We will use Slum Dweller Association collected data14 to augment our findings. WP3: Curation and synthesis of the literature on models of health service delivery relevant to the slum context (Years 1-4). We will perform systematic reviews and overviews of the literature concerning relevant models of health service delivery with specific reference to delivery arrangements as defined in Lavis's taxonomy of structures and implementation strategies within health systems15. WP4: Simulation-based interactive design of optimal service models (Years 2-3). We will draw on insights from WP1-3 to develop simulation models representing alternative approaches to slum health service delivery. These will be used to evaluate affordability and efficiency of services as they currently operate, and provide a policy tool to predict the impact of adopting new approaches, or improving the organisation of existing ones, including estimates of resilience to seasonal or unexpected (e.g. conflict or epidemic related) demand surges. Models will take a broad economic perspective, populated by health service resource use data collected from providers identified in WP1, and patient/ household costs identified in WP2. We will estimate cost-effectiveness by modelling the impact of service configuration on health outcomes as far as possible given available data. The models will allow us to estimate the impact and value-for-money of existing and proposed service configurations, taking affordability and logistical constraints into account, in line with WHO Choosing Interventions that are Cost Effective (WHO-CHOICE) methodology16.The idea is to design optimal services according to resource and staff available and to provide a pathway of improvements, including the option of improving technical efficiency of existing models. WP5: Engaging stakeholders and implementing viable options for health service delivery in slums (Years 1-4). A series of workshops with stakeholders such as healthcare commissioners, providers and people who live in slums will be held in the cities studied. In year 1, workshops will seek to understand local policy-drivers, questions and concerns of local people relevant to the work and develop and prioritise research questions to be addressed by the unit. Year 4 workshops will examine viable models of health service delivery in slums informed by the evidence gathered in WP1-4. E.g. printed maps will facilitate discussion on the optimal location of future services; simulation models will provide a decision aid. WP5 will culminate with formulation of an action plan to implement favoured models of care with selected commissioners or providers of health services. Opportunities for pilot evaluation studies will be examined. While WPs are specified above, we have capacity to respond to emerging global health research requirements. Notably: WP3 is designed to be flexible to allow evidence synthesis activity to support our developing work programme; specific PhD projects will evolve in response to stakeholder priorities (including those of our academic partners) explicated in Year 1 workshops; in years 3-4 we have allocated flexible researcher time to exploit data collected and methodology developed, and to pursue promising leads arising during years 1-2.