SARASWATI: Filling data black holes from humanitarian settings

Rapid epidemiological tool and integrated data platforms for infectious disease, near-misses and cause of death in humanitarian crises

Co-PIs: 

Prof. Gilbert Burnham, Centre for Humanitarian Health, Johns Hopkins Bloomberg School of Public Health

Prof. Debarati Guha-Sapir, Centre for Research on the Epidemiology of Disasters, UCLouvain

In 2021, 85million people were displaced due to conflict and disasters. This excludes those living in conflict or disaster zones with no or few functioning health facilities, and some agencies estimate that up to 2billion people currently live in these desperate conditions. Monitoring key health indicators and providing relevant services is challenging. Insecurity, access constraints, and fragile capacity mean civil registration of cause of death – a central health indicator – is nonexistent, and routine health records are unavailable. These settings are our “data black holes”; they are also powder kegs for infectious disease epidemics due to poor sanitation, environmental disruption, undernutrition, overcrowding, and lack of surveillance.

While advanced analytics allow estimates of mortality with limited datasets, the humanitarian community has been unable to identify causes of death and near misses – a critical blind spot. A major barrier is the inability to deploy field-friendly tools which can produce data that can be rapidly analysed. Other issues are the fragmented information on humanitarian action, and the inefficient use of the existing data, even if limited. Integrating data at regional and global levels on mortality in crisis-affected populations can visualize trends, patterns and ultimately bring assistance to thousands of people.

First and foremost, our solution targets populations living in humanitarian contexts. We envision our solution’s ultimate impact to be reducing their suffering through an improved evidence-base to detect and correctly respond to public health threats. We expect to identify their main disease risks and contribute to early recognition of a deteriorating health situation through the sentinel VASA, and allowing early intervention before outbreaks can spread and further increase mortality.

The solution itself is to be later used by people working in civil and registration, UN and agencies with an international mandate, health service providers, humanitarian agencies. It will allow them to conduct situation assessments to better tailor and evaluate interventions. Conducting public health programmes without a clear understanding of the burden of disease in a specific population carries the risk of designing ineffective interventions or worse, of planning projects that do more harm than good. The speed at which the data is compiled and made widely available is key in humanitarian settings. Our solution ensures a rapid turnaround of results and data for end-users such as humanitarian organizations and decision-makers, who will be able to adjust their actions in a timely manner according to the evidence produced with our solution. 

As firm supporters of open science, we will make sure all the products developed in the scope of this project will be globally accessible. This includes:

• A validated tool to ascertain causes of death in humanitarian settings will be publicly available which can promote understanding of the public health and political circumstances which are responsible for excess mortality
• Open access publications explaining the methodological steps needed to use the improved VASA tool and with results from our field studies
• Quick credible and curated data that will bring informative insights for containment, closer monitoring for the short term and forward planning for disease control and mortality reduction in the medium term.
• Public access integrated data hubs on disease and mortality for the health and humanitarian community. This hub will add massive added value by curating valuable data streams from different silos (data sources listed elsewhere) and making them useable to a wider disease control and response community. 
• Shared scripts of the analyses conducted in advanced statistical software

We envision the following impacts for the specified stakeholders:

- people affected by conflicts: better understanding of causes of death and disease patterns, allowing for evidence-based and tailored interventions

- humanitarian community: knowing burden of disease is key to plan resources for health in the short, mid, and long-term. A validated tool for the use of this community will allow agencies to better plan and evaluate their interventions.

- researchers: a validated but scientifically robust tool to ascertain cause of death in humanitarian settings. Having the potential to be applied in all humanitarian settings, we may unveil new knowledge on patterns of mortality and burden of disease that are specific to this setting.

- legal teams: responsibility on the impact of conflicts is commonly determined in the international criminal court. Reparations are often determined based on the number of people dead and their causes of death (direct violence or indirect deaths).

- global public health community: COVID-19 has increased the awareness of the importance to global public health of identifying potential outbreaks. Since there is little information coming out of humanitarian settings, while the risk of outbreaks in those contexts is high, we will provide an important global public good.

We will have a specific dissemination function with the team. This task will involve press briefings, papers in scientific journals, roll out of tools in regional and international operational communities and side events in key global forums with the help of TTC. These forums can include U.N.- ECOSOC and the WHO General Assembly . We will create a website for this solution and social media profiles, which will be advertised in our institutional websites. Apart from seminars to high-level stakeholders, we will also ensure the results are properly presented to the beneficiaries and local health service providers / humanitarian actors.

We also plan to set up a steering group for this project which will ensure this project follows scientific rigour and scalability. This group will consist of the following individuals: Javier Teran (Humanitarian Data Exchange); Philip Settel (Vital Strategies); Daniel Chandramohan (WHO Verbal Autopsy Reference Group Chair), Tefera Darge (Professor University Addis Ababa) ; Susan Cutter (University South Carolina) ; and a TTC member.

Once the instruments are well developed by year 2, we will look for humanitarian opportunities in Africa or the Middle East to test the VASA in critical situations to strengthen its scalability.

We will monitor our impact using the following indicators:

- Number of twitter followers, tweets liked and shared

- Number of organizations attending our webinars and physical seminars

- Number  of website visits 

- Number of presentations of tool to stakeholders such as the health cluster, the humanitarian logistics, special session at the ECOSOC, OCHA meetings

- Citations of datasets and publications

- Number of opportunities in assisting humanitarian organizations in implementing collection of VASA data in difficult areas with limited data which have critical humanitarian needs

Our greatest risk is conflict and political insecurity which would limit access. We will focus in areas and countries where we have an established network, facilitating monitorization of the security situation and site selection.

We may find a low number of deaths to be able to run advanced analytics and generate mortality fractions for the entire population. To mitigate this, we will have a relatively long recall period and use a systematic approach to detect deaths such as conducting brief surveys of households to retrospectively identify recent deaths; and identifying community leaders that might serve as key informants regarding the occurrence of recent deaths.

These community leaders will also be interviewed about the acceptability of conducting VASA interviews and the best way to approach affected households, identify appropriate respondents and increase acceptability of the interview, since low acceptability may also be a barrier to our project.

IRB clearances for the field studies will be key to the progress in our work and authorisations  must be clearly uploaded on the regional data hub website. We will negotiate specific memoranda with the main data suppliers to ensure steady and automatic inflow of data to the hub. 

Gilbert Burnham: Center for Humanitarian Health, Johns Hopkins University Bloomberg School of Public Health.

Debarati Guha-Sapir: Centre for Research on the Epidemiology of Disasters, Université catholique de Louvain

Eran Bendavid: Stanford Health Policy; Stanford Infectious Diseases; Woods Institute for the Environment

We see this an exciting new direction, taking forward proven processes. TTC is attractive to us as it is clearly willing to step out of the traditional or risk averse approaches to find solutions in a neglected setting. Furthermore, the option to link up with advanced technical skills from the founding members    is also an exciting opportunity, something that we don’t find with other support organizations

Overall, having a project in the scope of the TTC is great opportunity to build relationships, and long-term partnerships with, governments, national universities, local and international humanitarian organizations. These partnerships can be with the construction of the diagnostic and data analytic process, and can be transitioned to the use of the data once the system is providing data in real time which can be used to manage disease outbreaks for the vulnerable groups. Such partnerships will help refine data collection methods, ensure indicators are aligned with the humanitarian response needs, and organize its analysis and dissemination in a way that will be relevant to the humanitarian response. In this way there will be the maximum potential for impact.

We would greatly appreciate to have professional support to set up the creation of global and regional platforms, to ensure interoperability of data. We would also benefit from expertise on spatial data processing e.g. Google, and potential use of satellite sensing