Reflecting on digital contact tracing in Singapore: Lessons for the global community
Amidst the global outbreak of coronavirus disease 2019 (COVID-19), the Singapore government announced that by 2021, entry to public spaces would require the government’s TraceTogether mobile application or token [1]. Both the application and token are digital tools for contact tracing, tracking Bluetooth signals between devices to register: who a person has had contact with, for how long, and at what distance [2]. When an individual tests positive for COVID-19, his or her contacts can then be identified and isolated through these records.
Now that public movement has been tied to digital contact tracing, 70% of the population reportedly uses either the application or token [3]. In turn, government officials expect the widespread adoption to support an open economy while managing the epidemic curve in “Phase 3” of the government’s COVID-19 management plans [3]. Despite this optimism, however, there remains limited evidence for the efficacy of digital contact tracing [4].
In a recent systematic review, Braithwaite and colleagues identified 15 studies attempting to quantify the impact of digital contact tracing. Of the 15, 7 were hypothetical studies modelling whether Bluetooth-based applications could contain the spread of COVID-19. The remaining 8 were case studies, observational studies, or proof-of-concept studies involving other forms of automation (from previous infectious disease outbreaks). Based on these studies, the authors concluded that pending further research, large-scale manual contact tracing should remain the order of the day due to implementation concerns about digital methods (e.g., low uptake, costs, or privacy concerns) [4].
Reflecting on Braithwaite et al.’s review, Singapore’s contact tracing plans are timely and offer key lessons for the global community. First, it is indicative that although Singapore’s manual contact tracing was judged early on to be the ‘gold standard’ for case detection [5], the country nonetheless pursued automation. Quite simply, it is near-impossible to rely on manual contact tracing (as the authors of the review suggested)[4]. Take for instance an illustration using case numbers in the United States. In the month of August, the lowest number of daily cases was 3,712 in California. Given that an average of 36 contacts need to be traced for each positive case [6], this means that 133,632 individuals would have to be contacted for a one day increase in new cases. Since manual contact tracers aim to contact 3 individuals every hour and work shifts of up to 4 hours [7], this works out to 44,554 hours of contact tracing for the day, with 11,136 contact tracers to staff. Attaining Singapore’s ‘gold standard’ likely requires further resources: for example, in the first quarter of 2020, Singapore organized interviews for each COVID-19 case, tracing their contacts through intensive measures such as checking receipts or reviewing surveillance footage [7]. In one instance, when a taxi driver tested positive, the local police reportedly traced a passenger by examining footage from 40 surveillance cameras, then knocked on every door within a housing estate to identify the contact [8]. In short, manual contact tracing is logistically challenging and near-impossible to scale up. From these numbers alone, one would conclude that “large-scale manual contact tracing” is — in most cases — an oxymoron.
Having concluded that manual contact tracing cannot be scaled up, two questions remain. First, in the best case scenario, can digital contact tracing replace manual tracing methods? In other words, if digital contact tracing can achieve the adoption rates advocated in statistical modelling studies (≥56% of the general population) [9], will this solution — at minimum — be as effective as manual contact tracing? Worldwide, two case studies now allow this question to be answered: Qatar’s contact tracing application Ehteraz is used by >90% of Qatar’s resident population [10], while Singapore’s TraceTogether system has attained a 70% take-up rate [11]. Consequently, it is now possible to evaluate the efficacy of Bluetooth tracing solutions, and we await the results of these studies.
At the same time, widespread adoption remains elusive at the global scale, with an average take-up rate of merely 9% [6]. In other words, in most countries where digital contact tracing has been introduced, implementation difficulties prevail [10]. This leads to the second question: how can a government boost the adoption of digital contact tracing to allow efficacy? To address this question, it is useful to examine the range of techniques that governments have tried (Table 1).
As might be expected, legislation has yielded the highest adoption rates to date. Both Qatar and Singapore have made it mandatory to use digital contact tracing: Qatar by imposing heavy penalties for non-usage (a fine of $55,000 or imprisonment of 3 years), and Singapore by mandating usage for entry into public buildings. While no other country has made similar national-level stipulations, limited forms of mandatory usage have been implemented. For example, the Philippines government requires all travelers to use the contact tracing application Traze App for border entry [12], while several workplaces have — independent of government requirements — mandated usage for site entry [13].
Although legislation is effective, public concerns regarding privacy have led many governments to encourage voluntary adoption. Before usage was tied to building entry, Singapore’s TraceTogether phone application had the highest voluntary adoption worldwide, with 42% of the general population having downloaded the application [14]. This comes close to the optimal usage rate (≥56%), suggesting that reliance on voluntary adoption could be sufficient.
To date, voluntary use of digital contact tracing has been encouraged through: financial incentives (e.g. lucky draws for users) [15], clear messaging (e.g., Ireland) [16], or by offering other information and resources in the contact tracing application (e.g., France) [17]. In academic circles, the use of behavioral ‘nudges’ has also been discussed (e.g. changing from an opt-in to opt-out consent) [18], as has the customization of advertising based on sociodemographic characteristics that predict usage [19]. Since contact tracing applications only require the once-off act of downloading the software (following which the application functions independently in the background), it seems plausible to achieve the necessary take-up rates through careful implementation informed by the behavioral sciences.
In summary, given the scale of the pandemic, we suggest that the default containment strategy for COVID-19 should incorporate digital contact tracing protocols. As Singapore has shown, a manual approach can continue in parallel (or supported by) automated strategies, with efforts made to boost the conditions needed for automated approaches to succeed (e.g., increasing adoption of contact tracing applications, promoting trust in government, ensuring adequate infrastructure). Although research lags behind, the potential benefits of digital contact tracing outweigh the risks.
This commentary was written by Saw Young Ern, Tan Yi-Qin Edina, Madhumitha Ayyappan and Jean Liu, a team of researchers from Yale-NUS College who have been examining technological innovations during the COVID-19 pandemic. This commentary was originally written for submission to an academic journal.
References
1. Wong L. Use of TraceTogether app or token mandatory by end Dec. The Straits Times [Internet]. 2020 Oct 21; Available from: https://www.straitstimes.com/singapore/use-of-tracetogether-app-or-token-mandatory-by-end-dec
2. TraceTogether, safer together [Internet]. TraceTogether. 2020 [cited 2020 Dec 10]. Available from: https://www.tracetogether.gov.sg
3. Hermes. Target for 70% of residents to use TraceTogether met, Singapore News. 2020; published online Dec 24. https://www.straitstimes.com/singapore/target-for-70-of-residents-to-use-tracetogether-met (accessed Dec 29, 2020).
4. Braithwaite I, Callender T, Bullock M, et al. Automated and partly automated contact tracing: a systematic review to inform the control of COVID-19. The Lancet Digital Health [Internet]. 2020 Aug 19; Available from: https://www.thelancet.com/journals/landig/article/PIIS2589-7500(20)30184-9/abstract
5. Niehus R, Salazar PMD, Taylor AR, et al. Quantifying bias of COVID-19 prevalence and severity estimates in Wuhan, China that depend on reported cases in international travelers. medRxiv. 2020 Feb 18;2020.02.13.20022707. Available from: https://www.medrxiv.org/content/10.1101/2020.02.13.20022707v2
6. Chan, S. COVID-19 contact tracing apps reach 9% adoption in most populous Countries. SensorTower [Internet]. 2020 July 14; Available from: https://sensortower.com/blog/contact-tracing-app-adoption
7. Vaswani K. Coronavirus: The detectives racing to contain the virus in Singapore. BBC [Internet] 2020 Mar 19 [cited 2020 Sep 30]; Available from: https://www.bbc.com/news/world-asia-51866102
8. Mahmud, AH. “Like an invisible criminal”: How police helped find missing link between COVID-19 church clusters in a day [Internet]. [cited 2020 Sep 30]. Available from: https://www.channelnewsasia.com/news/singapore/police-missing-link-church-clusters-covid19-coronavirus-12509492
9. Hinch R, Probert W, Nurtay A, et al. Effective configurations of a digital contact tracing app: A report to NHSX. 2020 Apr 16;29; Available from: https://cdn.theconversation.com/static_files/files/1009/Report_-_Effective_App_Configurations.pdf?1587531217
10. Rivero N. Global contact tracing app downloads lag behind effective levels [Internet]. Quartz. 2020; Available from: https://qz.com/1880457/global-contact-tracing-app-downloads-lag-behind-effective-levels/
11. Kit TS. 70% of Singapore residents participating in TraceTogether programme: Vivian Balakrishnan. CNA. 2020; published online Dec 23. https://www.channelnewsasia.com/news/singapore/covid-19-tracetogether-adoption-singapore-crosses-70-percent-13829042 (accessed Dec 29, 2020).
12. Philippine Airlines. Must download Traze contact tracing app [Internet]. Philippine Airlines. 2020 [cited 2020 Dec 10]. Available from: https://www.philippineairlines.com/en/aboutus/newsandevents/advisory-covid19-27nov20-124-traze
13. Hoffman C, Litchfield JL. Employer use of contact tracing apps: The good, the bad, and the regulatory [Internet]. Foley. 2020 [cited 2020 Dec 10]. Available from: https://www.foley.com/en/insights/publications/2020/07/employer-use-of-contact-tracing-apps
14. Palma S. Singapore ‘strikes a balance’ with coronavirus response. Financial Times [Internet]. 2020 Oct 2; Available from: https://www.ft.com/content/062f901d-afd1-44b2-a0c5-56d35080807d
15. Win $55 vouchers when you use TraceTogether and SafeEntry [Internet]. TraceTogether. 2020 [cited 2020 Dec 10]. Available from: https://www.tracetogether.gov.sg/55/
16. Yamanis T. Clear, consistent health messaging critical to stemming epidemics and limiting coronavirus deaths. The Conversation [Internet]. 2020 Dec 2; Available from: https://theconversation.com/clear-consistent-health-messaging-critical-to-stemming-epidemics-and-limiting-coronavirus-deaths-134529
17. How is France’s new Covid app different from old one? The Connexion [Internet]. 2020 Oct 26; Available from: https://www.connexionfrance.com/Practical/Your-Questions/How-is-France-s-new-Covid-app-different-from-old-one
18. Mello MM, Wang CJ. Ethics and governance for digital disease surveillance. Science. 2020 May 29; 368(6494):951–4.
19. Saw YE, Tan EY, Liu JS, et al. Predicting public take-up of digital contact tracing during the COVID-19 crisis: Results of a national survey. medRxiv. 2020 Oct 5;2020.08.26.20182386.
