comment

Digital technology and COVID-19

The past decade has allowed the development of a multitude of digital tools. Now they can be used to remediate
the COVID-19 outbreak.

Daniel Shu Wei Ting, Lawrence Carin, Victor Dzau and Tien Y. Wong

T
he year 2020 should have been in 2003. These range from extreme Using three global databases―the
the start of an exciting decade in quarantine measures in China (e.g., locking Official Aviation Guide, the location-based
medicine and science, with the down over 60 million people in Hubei services of the Tencent (Shenzhen, China),
development and maturation of several province) to painstaking detailed contact and the Wuhan Municipal Transportation
digital technologies that can be applied tracing with hundreds of contact tracers Management Bureau―Wu et al.
to tackle major clinical problems and (e.g., Singapore, Hong Kong, South Korea). performed a modeled study of ‘nowcasting’
diseases. These digital technologies include However, these measures may not be and forecasting COVID-19 disease activity
the internet of things (IoT) with next- effective in 2020 for tackling the scale of within and outside China that could be
generation telecommunication networks COVID-19. Could new digital technology used by the health authorities for public-
(e.g., 5G)1,2; big-data analytics3; artificial be used for COVID-19? In this Comment, health planning and control worldwide8.
intelligence (AI) that uses deep learning4,5; we explore the potential application of four Similarly, using the WHO International
and blockchain technology6. They are highly inter-related digital technologies (the IoT, Health Regulations, the State Parties
inter-related: the proliferation of the IoT big-data analytics, AI and blockchain) to Self-Assessment Annual Reporting Tool,
(e.g., devices and instruments) in hospitals augmenting two traditional public-health Joint External Evaluation reports and the
and clinics facilitates the establishment of strategies for tackling COVID-19: Infectious Disease Vulnerability Index,
a highly interconnected digital ecosystem, (1) monitoring, surveillance, detection Gilbert et al. assessed the preparedness
enabling real-time data collection at scale, and prevention of COVID-19; and and vulnerability of African countries in
which could then be used by AI and deep (2) mitigation of the impact to healthcare battling against COVID-19; this would help
learning systems to understand healthcare indirectly related to COVID-19 (Table 1). raise awareness of the respective health
trends, model risk associations and predict authorities in Africa to better prepare for
outcomes. This is enhanced by blockchain Monitoring, surveillance, detection the viral outbreak9.
technology, a back-linked database with and prevention of COVID-19 Third, digital technology can enhance
cryptographic protocols and a network First, the IoT provides a platform that public-health education and communication.
of distributed computers in different allows public-health agencies access to data In Singapore, the government has partnered
organizations, integrating peer-to-peer for monitoring the COVID-19 pandemic. with WhatsApp (owned by Facebook)
networks to ensure that data are copied in For example, the ‘Worldometer’ provides to allow the public to receive accurate
multiple physical locations, with modified a real-time update on the actual number information about COVID-19 and
algorithms to ensure data are secured of people known to have COVID-19 government initiatives (https://www.form.
but traceable6. worldwide, including daily new cases gov.sg/#!/5e33fa3709f80b00113b6891).
However, 3 months into 2020, the of the disease, disease distribution by Multiple social-media platforms (e.g.,
world is facing an existential global health countries and severity of disease (recovered, Facebook and Twitter) are currently used
crisis: the outbreak of a novel coronavirus– critical condition or death) (https://www. by healthcare agencies to provide ‘real-time’
caused respiratory disease (COVID-19)7. worldometers.info/coronavirus/). Johns updates and clarify uncertainties with the
As the knowledge of COVID-19 evolves, Hopkins University’s Center for Systems public. Additionally, some facial-recognition
increasing evidence suggests that it seems Science and Engineering has also developed companies (e.g., SenseTime and Sunell)
to be less deadly than initially thought a real-time tracking map for following cases have adopted the thermal imaging–enabled
(with a mortality rate of approximately 2%), of COVID-19 across the world, using the facial recognition to identify people with an
although more contagious (89,779 cases in data collected from US Centers for Disease elevated temperature at various screening
70 countries, with over 3,069 deaths as of Control and Prevention (CDC), the World points in China (https://apnews.com/PR%20
2 March 2020) (https://www.worldometers. Health Organization (WHO), the European Newswire/354aae0738073bc95331ee72a45
info/coronavirus/). The impact of COVID-19 Center for Disease Prevention and Control, 8cb50).
will probably be greater than that of severe the Chinese Center for Disease Control and Fourth, AI and deep learning can
acute respiratory syndrome (SARS) in Prevention (China CDC) and the Chinese enhance the detection and diagnosis of
2003, given globalization and the relative website DXY, which aggregates data from COVID-19. The need to provide access
importance of China in 2020 in terms of China’s National Health Commission and to accurate and low-cost tests for the
world trade and travel. the China CDC (https://gisanddata.maps. diagnosis of COVID-19 is a challenge.
How can this new crisis in 2020 be arcgis.com/apps/opsdashboard/index.html#/ Many peripheral hospitals in China and
tackled? How does it differ from the bda7594740fd40299423467b48e9ecf6). other developing countries in Asia, the
SARS epidemic in 2003? Many countries Second, big data also provides Middle-East and Africa do not have the
have relied on an extrapolation of classic opportunities for performing modeling tests or resources to accurately distinguish
infection-control and public-health studies of viral activity and for guiding COVID-19 from the ‘common flu’. In
measures to contain the COVID-19 individual country healthcare policymakers Indonesia, which has only two reported
pandemic, similar to those used for SARS to enhance preparation for the outbreak. case thus far, despite substantial exposure
Nature Medicine | www.nature.com/naturemedicine
comment

Table 1 | Digital technologies and their impact on public-health strategies

Digital technology
Public-health measures IoT Big data AI Blockchain
1. Monitoring, surveillance, +++ +++ ++ +
detection and prevention
of COVID-19 (directly
related to COVID-19)
Examples 1. Real-time tracking and live 1. Modeling of disease 1. Detection of COVID-19 1. Manufacturing and
updates in various online activity, potential growth from chest imaging (X-ray) distribution of COVID-19
databases in the USA, UK and areas of spread (Beijing Hospital) vaccines once they are
and China available
2. Live tracking of the 2. Modeling of the 2. Prognostication of disease 2. Insurance claims from
at-risk vicinity in Korea preparedness and progression via clinical data, COVID-related illness
(Coronamap.live; vulnerability of countries in imaging and AI and death
Wuhanvirus.kr) fighting a disease outbreak
2. Mitigation of impact +++ ++ +++ ++
(indirectly related to
COVID-19)
Examples 1. Virtual clinics (PingAn, 1. Business modeling on 1. AI to automatically diagnose 1. Distribution of patients’
China) pharmaceutical supplies for medical conditions unrelated regular medication to
various medications to COVID-19 (Zhongshan the local pharmacy or
Ophthalmic Eye Center, patients’ doorstep
China)
2. Public information 2. Modeling of the utility of 2. Medical ‘chat bots’ to address
dissemination via WhatsApp operating theaters and public inquiries on COVID-19
in Singaporea clinics with manpower
projections
The likely impact of digital technologies on (1) disease monitoring, surveillance, detection and diagnosis, and (2) mitigation of impact: +, low (no clear example yet in either official government website);
++, moderate (one clear example); +++, high (two or more examples). Gray shading indicates potential applications that are not described in the literature thus far but should be considered by
technology companies or research groups worldwide to help battle against COVID-19. Additional examples beyond those mentioned in the text are included in this table. ahttps://www.form.gov.sg/#!/
5e33fa3709f80b00113b6891.

to Chinese tourists (Bali had 1.2 million disease; and the 5% who have severe disease, key hospital activities (e.g., research and
Chinese tourists in 2019), health authorities including those at high risk of mortality. education), virtual e-learning platforms are
decided against testing the 243 returning Finally, AI can also facilitate the discovery of increasingly being explored to eliminate
but asymptomatic citizens from Wuhan novel drugs with which to treat COVID-19. physical meetings.
because of cost of the test (the reagent Second, the utilization of various AI-based
was quoted as costing nearly US$75,000). Mitigation of COVID-19’s impact triage systems could potentially alleviate the
Alternative diagnostic and screening tests Although the focus of tackling the direct clinical load of physicians. An online medical
for COVID-19 will be extremely useful. impact of COVID-19 is important, in ‘chat bot’ could help patients recognize
In this context, China has large datasets many healthcare settings, it is important to early symptoms, educate people on the
of cases positive for COVID-19 (>70,000 maintain core and critical clinical service. importance of hand hygiene and refer people
cases). These are ideal datasets for deep AI The initial reaction in many countries is for medical treatment should symptoms
and deep learning (https://www.wired.com/ for healthcare facilities to reduce or even worsen. Additionally, phone-based software
story/chinese-hospitals-deploy-ai-help- cease many clinical services, including that detects and records patients’ data
diagnose-covid-19/). Such AI algorithms closure of clinics and postponement of (e.g., daily temperature and symptoms) may
can then be used as an initial screening tool medical appointments or elective surgeries. prevent unnecessary hospital consultations
for suspected cases (e.g., travel history to However, such strategies cannot be sustained for patients with mild flu-like symptoms.
China, Iran or South Korea, or exposure to indefinitely if the COVID-19 pandemic These data could also be developed into AI
confirmed cases) so that patients at higher extends beyond 6 months. algorithms for the detection of COVID-19.
risk could have confirmatory laboratory- Healthcare systems should plan to use Third, many hospitals in China are
based tests or be isolated. digital technology. For example, ‘virtual collaborating with blockchain companies
Although most patients have mild cases clinics’ could be set up through the use of and pharmacies to deliver patients’
of COVID-19, physicians have to apply the tele-medicine consultations with imaging medication to their doorsteps. Through
same level of intensive methods to isolate, data (e.g., chest X-ray and/or CT of the the use of blockchain, hospitals could
treat and monitor all patients. AI algorithms thorax) uploaded from peripheral sites and ensure timely delivery of medications with
could be developed to help physicians interpreted remotely. This would ensure that accurate tracking.
triage patients with COVID-19 into patients continue to receive standard clinical In summary, while the world continues
potentially three groups: the 80% who have care while reducing physical crowding of to rely on classic public-health measures
mild disease; the 15% who have moderate patients into hospital premises. For other for tackling the COVID-19 pandemic,

Nature Medicine | www.nature.com/naturemedicine

 comment

in 2020, there is now a wide range of As the saying goes, ‘a crisis provides Published: xx xx xxxx
digital technology that can be used to an opportunity’; this first great crisis https://doi.org/10.1038/s41591-020-0824-5
augment and enhance these public- of 2020 provides a great opportunity for
References
health strategies (https://www.vox.com/ digital technology. ❐ 1. Perkel, J. M. Nature 542, 125–126 (2017).
recode/2020/2/27/21156358/surveillance- 2. Ting, D. S. W., Lin, H., Ruamviboonsuk, P., Wong, T. Y. &
tech-coronavirus-china-facial-recognition). Daniel Shu Wei Ting   1 ✉, Sim, D. A. Lancet Digital Health 2, E8–E9 (2019).
3. Shilo, S., Rossman, H. & Segal, E. Nat. Med. 26, 29–38 (2020).
There is also a longer-term goal. The Lawrence Carin2, Victor Dzau2,3 and 4. LeCun, Y., Bengio, Y. & Hinton, G. Nature 521, 436–444 (2015).
immediate use and successful application Tien Y. Wong1 5. Ting, D. S. W. et al. Nat. Med. 24, 539–540 (2018).
6. Heaven, D. Nature 566, 141–142 (2019).
of digital technology to tackle a major, 1
Singapore Health Service (SingHealth), 7. Huang, C. et al. Lancet 395, 497–506 (2020).
global public-health challenge in 2020 Duke-NUS Medical School, Singapore, Singapore. 8. Wu, J. T., Leung, K. & Leung, G. M. Lancet 395, 689–697 (2020).
9. Gilbert, M. et al. Lancet 6736, 30411–30416 (2020).
will probably increase the public and 2
Duke University, Durham, NC, USA.
governmental acceptance of such 3
US National Academy of Medicine, Competing interests
technologies for other areas of healthcare, Washington, DC, USA. D.S.W.T. and T.Y.W. hold a patent on a deep learning
including chronic disease in the future. ✉e-mail: [email protected] system for the detection of retinal diseases.

Nature Medicine | www.nature.com/naturemedicine