Artificial Intelligence

The importance of linkage: lessons from one pandemic to another

September 21, 2022
editor


In this matter, The Lancet’s digital health publishes a series entitled “Translating data in a pandemic” consisting of two series articles and one commentary. This series highlights the need to raise robustness standards for research conducted rapidly in a pandemic and improve data sharing systems for future outbreaks. But what challenges have existed since COVID-19 first broke out, and how likely are they to impact future epidemics like monkeypox?

Commentary by Lauren Gardner and colleagues highlights that pandemic models did not adequately inform responses to outbreaks and that projections were constrained by poor dissemination to policymakers and the public. For example, poor data reporting infrastructure in the US limited comparative analysis between states. The series paper by the same authors found that studies modeling COVID-19 were prone to misinterpretation and potentially harmful misuse. This problem was amplified as news outlets covered COVID-19 preprints extensively, expanding their reach far beyond the academic audience to include conspiracy theorists. The series paper also highlights systemic problems with the current information exchange systems in the United States. The authors question whether preprints or peer-reviewed journals are fit for purpose during a pandemic, as neither offer the required speed or quality for information sharing. The Lancet’s digital health advocates that models should not be used to make predictions for the public interest, but to provide experts with insight into disease epidemiology.

The series paper by Louis Dron and colleagues discusses the insufficient standardization of routinely collected medical data for COVID-19 cases. The series paper details a lack of transparency about how data is encoded and the limitations of current healthcare systems that limit the sharing of real-time data. In a 2021 House of Commons report on lessons learned from the COVID-19 response, early efforts to analyze the pandemic were hampered by the UK’s national public authorities’ failure to share COVID-19 data. Another Bureau for Statistics Regulation report on COVID-19 education emphasized that data sharing and linking can have life-saving implications. Investing in the necessary infrastructure for data sharing must be a priority for governments beyond the pandemic. To ensure data consistency across healthcare systems, Dipak Kotecha and colleagues developed the CODE-EHR minimum standards framework, which aims to improve the design and reporting of research studies using structured electronic health records. Studies adhering to the CODE-EHR framework will support productive data sharing and expand the impact of outbreak research.
During the pandemic, the exchange of electronic health records has played an important role in healthcare decisions across all medical specialties. For example, the RECOVERY study linked datasets to study participants while ensuring data security and quality. The study recruited over 47,000 participants in six countries to discover four effective COVID-19 treatments. Investing in the infrastructure needed for multi-site studies that can collect data quickly could help efforts to discover therapies to fight the monkeypox epidemic while access to vaccines remains limited.
This series has shown that digital health research has failed to accurately and responsibly disseminate during the pandemic. Better communication methods are needed that can handle the required speed of publication and balance that speed with the quality of research results. For example, modeling research should only be shared with the public by an experienced translator who can fully constrain the interpretations. To ensure that modeling research is reported correctly, Gardner and colleagues advocate using the EPIFORGE 2020 guidelines. These guidelines help researchers provide a clear definition of the study purpose and model goals, and full reporting of the data. Publishers have a great responsibility to support better reporting of data to the public and policy makers in a crisis. As such, The Lancet’s digital health invites you to attend the Lancet Summit on Big Data and Artificial Intelligence in Pandemic Preparedness, where multidisciplinary experts will review the global response to the COVID-19 pandemic and discuss how technology and data can be better leveraged to create equitable and accurate tools for future pandemic responses.
Analysis of financial forecast charts. The chart reflects data behavior over a historical period. Statistical research concept. 3D rendering

article information

publication history

ID

DOI: https://doi.org/10.1016/S2589-7500(22)00175-3

Copyright ©

© 2022 The author(s). Issued by Elsevier Ltd.

user license

Creative Commons Attribution (CC BY 4.0) |

ScienceDirect

Access this article on ScienceDirect

Linked Articles

  • Real-time COVID-19 Prediction: Challenges and Opportunities of Model Performance and Translation

    • The COVID-19 pandemic pushed mathematical modeling into the spotlight as scientists rushed to use data to understand transmission patterns and disease severity, and to predict future outcomes of the epidemic. However, the use of COVID-19 modeling has been criticized, in part for some particularly flawed projections early in the pandemic.1 More than 2 years after the start of the pandemic, models continue to face serious obstacles as tools for informing the response to the outbreak.1 Population-level health outcomes are difficult to predict accurately, particularly cases and hospitalizations,2 as discussed in the International Institute of Forecasters blog.