Learnings from ICODA initiative highlight importance of data science in addressing COVID-19 research challenges

| 18/01/2023

The International COVID-19 Data Alliance (ICODA) research studies made a significant contribution to the global pandemic response, harnessing the power of health data research approaches to address major research questions. 


Learnings from our research teams 


1) Bespoke prediction tools are a priority 

As the pandemic spread, new risk assessment tools were needed to help clinicians decide which patients needed emergency care. The PRIEST Driver Project developed specific tools appropriate for a wide range of global settings for triaging suspected COVID-19 cases arriving in emergency departments. This included a triage scoring system designed for use in healthcare facilities in low- and middle-income countries.  

Dr Gordon Fuller, who worked with colleagues across South Africa, believes that while the project brought significant progress, further refinements are essential. If this does not happen, he fears that: “Emergency care systems in low- and middle-income countries have the potential to be overwhelmed if new variants of concern cause future infection waves.

2) Large studies are vital 

To evaluate the efficacy and safety of potential COVID-19 treatments, Ms Névine Zariffa’s EFCT project team compared research findings from randomised control trials (RCTs) and real-world studies. Ms Zariffa found the real-world studies tended to report sooner than RCTs, but the results could be less reliable which may be because they were smaller or impacted by sample bias.  

Ms Zariffa said: “The real-world studies are quite variable in their results, compared to RCTs, that are more consistent and contain less variation in results.” 

Ms Zariffa believes that large accurate studies are vital to help ensure their scientific reliability and that researchers need to “really think carefully” when planning a small study. A paper by the team in Springer Nature, which looked at RCTs, says: “small underpowered trials are not productive, and worse can be unethical and misleading when viewed on their own”.  

3) New statistical approaches can help combat future infectious disease outbreaks 

Understanding transmission chains and reproduction rates – how disease spreads – is essential for planning control measures. As the pandemic unfolded, some countries ceased providing the detailed epidemiologic survey data that researchers such as Dr Xiaofan Liu needed for this. The CHAIN study developed new approaches using high level and more abstract statistics, such as daily case numbers.  

Dr Liu, who was studying countries including China, Indonesia, Singapore, and the Philippines, said: “These metrics are very important for the transmission of all diseases. Hopefully our methodologies, which allow us to infer basic metrics from statistics, could help future work in fighting large-scale transmission of disease.” 

4) More use of routinely collected data to be made 

In researching how lockdowns and confinement policies impacted health services and institutional mortality, Dr Catherine Arsenault’s REHCORD project team worked with lead researchers across 10 countries including Nepal, Laos, Ghana, Haiti and Mexico and made use of routinely collected patient data. The team see greater use of routine data (such as clinical notes and administrative information) as essential for health system planning and service delivery.  

Dr Arsenault said: “There should be more work done with these data. There are still data quality challenges for sure, but without better demand, and better use of data, quality is unlikely to improve.” 

5) Preterm births dropped during lockdowns 

Preterm birth is the leading global cause of infant death, but the causes are largely unknown. When reports emerged that levels were dropping during lockdown the international iPOP team, co-led by Professor Sir Aziz Sheikh, Dr Meghan Azad, Dr Sarah Stock and others used data on 2.4 million births across 40 countries (22 low and middle income) to investigate.  

Professor Sheikh said: “The headline finding was that lockdown was associated with a modest reduction in prematurity. But at a population level this will translate into substantial numbers of premature births being averted. Importantly, it was not associated with any increase in stillbirths. A really important finding.”  

The team identified changes brought on by lockdowns which may have contributed to the reduction in preterm births such as impact on maternal workloads, hygiene practices and air pollution.  

6) Stakeholder engagement maximises research benefits 

The IROC Driver Project identified COVID-19-related risk factors among pregnant and lactating women and their infants in Uganda. The researchers then worked with the Women’s Health Team at a large hospital to develop a dashboard to support decision making on service delivery. 

Researcher Dr Joseph Ouma said: “Engaging stakeholders improved the ownership of the process. We were able to realise excitement and acceptability among the service providers and this led to an ongoing commitment by the hospital team to use the dashboard.” 

7) Public engagement can help with health inequalities 

Dr Maria Yury Ichihara and her colleagues in her IDS-COVID19 team in Brazil worked with communities in areas of social and economic deprivation with poor access to healthcare. They developed effective ways to monitor the pandemic’s spread and to provide healthcare information to vulnerable groups. Their work, across 5,500 municipalities, demonstrated the much greater severity of the pandemic in poorer areas.  

Dr Ichihara said: “We had an amazing experience of dialogue with different groups in 13 states in Brazil. We think this study can contribute to guiding public policies and help focus resources on public health and push forward the principle of nationalisation of our healthcare system.” 

8) New technologies can democratise care 

Many Driver Projects were based in low- to middle-income countries where gathering data could be challenging and the impact of COVID-19 was especially severe, due to lack of access to healthcare.  

Dr Fernando Bozza, whose EFFECT-Brazil team used mobile data to examine the vaccine rollout in Brazil, said this points to an urgent need to: “democratise the use of new technologies to provide access to healthcare in poor communities and to develop people-centred community-based research integrating data in primary health care.” 

9) Building data science capacity is vital 

The Accord project, in South Africa’s Western Cape province, researched issues including the severity of COVID-19 variants as they emerged which helped shape the pandemic response and provide public information.  

Project leader Professor Andrew Boulle said: “The main lesson was that having a data centre team – a group of scientists who work with routine individual level health data on a daily basis – was critical to being able to pivot to answer new questions quickly. Building the data science capacity within routine health services is a key priority for future readiness.” 

10) Data systems investment will pay dividends 

Accessing and cleaning data was a major challenge for the PIH-CoVCo team investigating the impact of COVID-19 on healthcare provision and service use by patients with chronic conditions in Haiti, Malawi, Mexico and Rwanda.  

As a result, the Partners in Health researcher, Stefanie Joseph, said: “there was a strong recognition of the benefits that would come from greater investment in electronic health records (EHRs) and the creation of systems making it easier to pull relevant data into curated databases for analysis.” 

11) Consistent data formatting improves utility for all 

The ISARIC project created a unified code allowing hospital records from more than 50 countries – which was in many formats and used different data points – to be reconciled in a user-friendly way. This made it simpler to compare the impact of COVID-19 in different times and places. The code was made publicly available so it could be downloaded and improved.  

Dr Luis Felipe Reyes Velasco said: “It’s not only important to gather data and put it on a website, it’s important that people can use it. You don’t need to be an expert data scientist to use this data with this approach. We teach people how to use it, ensuring the data has a greater impact.” 

12) Preparing for the next pandemic 

Showcase contributors repeatedly stated that COVID-19 should not be seen as a one-off and that ICODA’s work needs to be built on and extended.  

Dr Juliane Oliveira, whose RASUP project team researched how social inequalities affected COVID-19’s spread in Brazil, said: “We built capacities that can be adapted to all public health contexts and we also strengthened modelling – essential if we are to be prepared for the next pandemic.” 

Future initiatives 

ICODA, which completed its work at the end of  December 2022, involved 135 researchers from nineteen countries and studied data from over 60 countries – resulting in 37 publications from the research teams to date. 

HDR UK is building on this impressive legacy with HDR Global, a programme focussing on a wide array of global health challenges. It will maintain and strengthen the partnerships ICODA has established in Asia, in Africa and in Latin America and elsewhere and will build on learnings around trustworthy data sharing from the ICODA programme. 

Professor Agnes Binagwaho, who co-chaired ICODA’s Scientific and Strategic Advisory Council, said: “This is an example of what the future should be like, and how we should be sharing what we do as soon as we can for the benefit of patients – wherever they are. I am optimistic that we are on the right path.” 

For more about the impact of ICODA visit the publications and outputs pages. 

You can find a replay of ICODA’s Showcase event on the ICODA website.

For more information, please contact Clare Leahy, Projects Comms Lead, at clare.leahy@hdruk.ac.uk