Opportunities, risks, and the tension between privacy and public interest during the COVID-19 outbreak.
Contact tracing during a global pandemic
Background
The Australian Department of Health website states “Our vision is better health and well being for all Australians, now and for future generations” (D. o. Health 2020b).
Governments worldwide are scrambling to find responses to the global COVID-19 pandemic. As of 13th April 2020 there are over 180 countries with 1,773,084 confirmed cases (W.H.O., 2020). A pharmaceutical intervention does not currently exist for COVID-19. Australia (and other countries) has adopted blanket social distancing (D. o. Health 2020a) as a mitigation strategy. Whilst necessary, this dramatic action has many impacts (Rashid et al. 2015) e.g.;
· Individuals experiencing isolation, anxiety, loss of income.
· Closure of businesses that are not able to operate under social distancing.
Governments are faced with the unenviable choices of maintaining blanket social distancing until a widely available vaccine is developed or relaxing social distancing and accepting the health impacts. This creates tensions between health outcomes, economic outcomes, and protecting civil liberties. Journalist Jeremy Cliffe has coined this the “Coronavirus Trilemma”.
“…They can pick two of three things but cannot have them all: limit deaths, gradually lift lockdowns, or uphold cherished civil liberties.”
(Cliffe 2020)
Figure 1 “Coronavirus Trilemma” (Future Crunch 2020)
Some countries (including Australia) are in the early phases of trialing “Bio Surveillance” as another lever to pull in the struggle to contain COVID-19.
“In its least intrusive form, bio-surveillance involves using phone data — and other tools such as CCTV — to monitor how populations as a whole are behaving.” (Cliffe 2020)
My paper examines Bio Surveillance in the context of public health. Specifically, the risks and opportunities when used as a part of a contact tracing strategy.
What are the challenges?
Social Distancing
On 29 March the Australian National Cabinet limited non-essential gathering to 2 people (D. o. Health 2020a). Social distancing works by breaking the transmission network by limiting contact between individuals, therefore reducing opportunities for infection.
Figure 2 Confirmed cases in Australia. Data sourced from (Max Roser 2020)
Prior to COVID-19 large scale social distancing had not been attempted. (Rashid et al. 2015) found moderate effectiveness for social distancing with influenza and acceptable in the short term. Early studies such as (Abouk 2020), (Lewnard & Lo 2020), (Prem et al.) show a causal relationship between social distancing and reducing interaction. Social distancing is not without drawbacks.
“Although the scientific basis for these interventions might be robust, ethical considerations are multifaceted.9 Importantly, political leaders must enact quarantine and social-distancing policies that do not bias against any population group. The legacies of social and economic injustices perpetrated in the name of public health have lasting repercussions.10 Interventions might pose risks of reduced income and even job loss, disproportionately affecting the most disadvantaged populations: policies to lessen such risks are urgently needed. Special attention should be given to protections for vulnerable populations, such as homeless, incarcerated, older, or disabled individuals, and undocumented migrants. Similarly, exceptions might be necessary for certain groups, including people who are reliant on ongoing medical treatment.”
(Lewnard & Lo 2020)
Additionally there are major economic impacts, the OECD estimates a reduction of 22% in Australian GDP due to COVID-19 (OECD 2020). The Australian Bureau of statistics shows 66% of business had cash flow impacted, overseas visitor arrivals were down, and labour market impacts where workers had their hours reduced (ABS 2020).
Table 1 Percent of employed who worked fewer hours than usual, February 2020 (ABS 2020)
Networks and Epidemiology
Networks are a powerful tool in epidemiology, connections between individuals that allow a disease to propagate naturally form a network.
“..understanding of the structure of the transmission network allows us to improve predictions of the likely distribution of infection and the early growth of infection..”
(Danon et al. 2011).
“In homogeneously mixing populations, the relationships between key epidemiological quantities are generally well understood.”
(Pellis et al. 2015).
However, networks in human populations are more complicated. Links between humans are dynamic and can form and dissolve over time and space. Some people(nodes) may be more highly connected than others, links(edges) between nodes may have differing risks of transmission.
Contact tracing
“Contact tracing (i.e. real-time tracking of infected individuals and their exposed contacts) is a typical network-based intervention”
(Pellis et al. 2015)
Information for entire populations is rare, traditionally contact tracing is only performed after a person has been diagnosed with an infection and their contacts with others traced. This information is rarely dynamic and usually only records the presence or absence of contact, and not its frequency or strength. For example, when contact tracing STIs an individual could name their partners and then the process could be repeated for their partners in a ‘snowball’ process. However, this may not capture all networks present within the population and would miss asymptomatic individuals. Contact tracing after diagnosis, whilst useful for understanding transmission and behaviour of past epidemics are less useful in the early stages of new outbreaks and have little predictive power (Danon et al. 2011). However if high risk individuals could be automatically identified contact tracing can be highly effective as a preventative or control strategy, and is particularly useful for asymptomatic infections. (Pellis et al. 2015, Sec. 8). Public health interventions can aim to reduce transmission along network edges without fundamentally altering the network topology (face masks, hand washing), or can have local and population-scale effects on the topology of the contact network (e.g. school closures, social distancing or vaccination, which reduce contacts by removing network edges) (Pellis et al. 2015, Sec. 8).
A proactive approach enabled by modern location data collection and machine learning techniques could be used to predict (and respond to) how an outbreak might play out (Abbas & Michael 2020). By automatically identifying high-risk individuals, it can be highly effective as a preventative or control strategy, and is particularly useful for asymptomatic infections (Pellis et al. 2015, Sec. 8).
What are the opportunities?
Augmenting traditional contact tracing with mobile data
Traditional approaches to contact tracing which rely on manual process can be augmented by automated data collection such as;
· Proactive Self Reporting by individuals via smartphone app to a central agency.
· Location data through smartphones and other devices (cell phone towers, GPS, Wi-Fi).
· Nearby device data collected by smartphone apps using Bluetooth or NFC.
(Danon et al. 2011; internetnz 2020; Pellis et al. 2015; Zastrow 2020)
This would allow networks to be built in real time and used predictively to identify individuals and locations which are at higher risk. This could be used to target or relax existing intervention as risk assessment changes.
Having a real time view of transmission networks would also provide a framework to test what interventions are effective and potentially deployable elsewhere.
What are the issues implementing this innovation?
The COVID-19 environment is highlighting the tension between privacy more than ever. In the face of a global outbreak rapid access to data is important to fighting the outbreak (Hao 2020). The drastic impacts of blanket social distancing on individuals, society, and the economy raise the stakes for inaction or poorly thought out interventions (Wetsman 2020).
“You might as well ask yourself, has history ever shown that once the government has surveillance tools, it will maintain modesty and caution when using them?” (Hao 2020)
There is also the risk that authoritarians will use the crisis as an opportunity to erode civil liberties, Hungary has put forward a bill allowing “rule by decree”.
“COVID-19 contact-tracing apps can reinforce existing social biases, thus stigmatizing locations and communities.” (Millar 2020)
Discrimination has been observed towards Asian communities in Australia and Canada. Information such as location or individual publicly flagged as high risk may lead to stigmatisation singling out of those flagged. Communities of existing disadvantage such as Low Socio Economic groups are impacted earliest and most significantly during a time of crisis. Care must be taken to acknowledge the asymmetric impact of COVID-19 extends beyond health concerns (Davis 2020; Millar 2020; Wetsman 2020).
“This allows all of these issues of biases and worldviews to be written off as “just the AI.”” (Davis 2020)
Advanced prediction and modelling techniques can appear like a magical black box to the lay person or even the decision maker.
The manner in which the analysis (e.g. risk assessment of individuals) is present back to the community needs to be in a manner which gives the public the information they need to protect their health and nothing more. There is a very real risk that information being collected can be used to identify individuals. E.g. South Korea published very detailed information including travel histories of confirmed patients, this could be used to de-anonymise individuals.
Sensitive data would be collected on individuals, their location, who they associate with, and their health status.
Consideration should be paid to;
· What will happen to the data afterwards
· Who holds/owns the data?
· At what scale(region/country/city/global) is a contact tracing system deployed
· How transparent will the code base be?
· Is the data collection anonymised?
· Can anonymised data be deanonymized?
· Ability to see how it arrives at decisions (Singer & Sang-Hun 2020)
How will this innovation impact the organisation / sector?
Figure 1 illustrates the competing goals which must be traded against each other. An ideal solution which meets all three goals does not exist. That said, faster access to analysis and greater granularity around decision making as a result gives a huge opportunity to better mitigate the negative outcomes of COVID-19.
Hu Yong lays out 3 principles in seeking a balance of private and public interests during this crisis.
“First principle to adhere to is to treat public interest (concerns) as exceptions to (the protection of) privacy. “
Second principle: if it is really necessary to manage (restrict) privacy for the sake of public interest, then we must establish appropriate guarantees for basic civil rights and personal interests in the process of managing (restricting) privacy.
Third, insist on fair use of information.”
(Yong 2020)
Whilst I am receptive to the idea of automated data collection to assist the building of complete contact tracing networks, I want to see the following conditions to be met;
1. Data collected from individuals to be anonymised in a robust fashion.
2. Guidelines for data use to be clearly established and sunset clauses in place to expire and delete data.
3. Data to be held by a Government agency.
4. Code used to build tools for collection to be available for independent review. Preferably the code would be open sourced.
Another question to consider is “should automated contact tracing be compulsory for citizens?”.
Bibliography
Abbas, R. & Michael, K. 2020, The coronavirus contact tracing app won’t log your location, but it will reveal who you hang out with viewed 17 Apr 2020, <https://theconversation.com/the-coronavirus-contact-tracing-app-wont-log-your-location-but-it-will-reveal-who-you-hang-out-with-136387>.
Abouk, R.a.H., Babak, 2020, ‘The Immediate Effect of COVID-19 Policies on Social Distancing Behavior in the United States ‘.
ABS 2020, 6291.0.55.001 — Labour Force, Australia, Detailed — Electronic Delivery, Feb 2020, viewed 18 Apr 2020, <https://www.abs.gov.au/ausstats/abs@.nsf/Lookup/6291.0.55.001main+features3Feb%202020>.
Cliffe, J. 2020, The rise of the bio-surveillance state, viewed 17 Apr 2020, <https://www.newstatesman.com/science-tech/2020/03/rise-bio-surveillance-state>.
Crunch, F. 2020, FC96: Coronavirus and the Bio-Political Straitjacket viewed 18 Apr 2020, <https://futurecrunch.substack.com/p/fc96-coronavirus-and-the-bio-political>.
Danon, L., Ford, A.P., House, T., Jewell, C.P., Keeling, M.J., Roberts, G.O., Ross, J.V. & Vernon, M.C. 2011, ‘Networks and the epidemiology of infectious disease’, Interdiscip Perspect Infect Dis, vol. 2011, no. 2011, p. 284909.
Davis, H. 2020, A Dataset is a Worldview, self, Medium, viewed 17 Apr 2020, <https://towardsdatascience.com/a-dataset-is-a-worldview-5328216dd44d>.
Hao, K. 2020, Coronavirus is forcing a trade-off between privacy and public health, viewed 17 Apr 2020, <https://www.technologyreview.com/2020/03/24/950361/coronavirus-is-forcing-a-trade-off-between-privacy-and-public-health/>.
Health, D.o. 2020a, Social distancing for coronavirus (COVID-19) viewed 18 Apr 2020, <https://www.health.gov.au/news/health-alerts/novel-coronavirus-2019-ncov-health-alert/how-to-protect-yourself-and-others-from-coronavirus-covid-19/social-distancing-for-coronavirus-covid-19>.
Health, D.o. 2020b, What we do, viewed 17 Apr 2020, <https://www.health.gov.au/about-us/what-we-do>.
internetnz 2020, Let’s talk contact tracing, viewed 18 Apr 2020, <https://internetnz.nz/let%E2%80%99s-talk-contact-tracing>.
Lewnard, J.A. & Lo, N.C. 2020, ‘Scientific and ethical basis for social-distancing interventions against COVID-19’, Lancet Infect Dis, pp. S1473–3099(20)30190–0.
Max Roser, H.R., Esteban Ortiz-Ospina and Joe Hasell 2020, Coronavirus Disease (COVID-19) — Statistics and Research, Published online at OurWorldInData.org., viewed 19 Apr 2020, <https://ourworldindata.org/coronavirus>.
Millar, J. 2020, A well-intentioned but unproven app could reinforce biases and create confusion and stress, something developers must take more time to consider. , policyoptions, viewed 17 Apr 2020, <https://policyoptions.irpp.org/magazines/april-2020/five-ways-a-covid-19-contact-tracing-app-could-make-things-worse/>.
OECD 2020, EVALUATING THE INITIAL IMPACT OF COVID CONTAINMENT MEASURESON ACTIVITY, viewed 19 Apr 2020, <https://read.oecd-ilibrary.org/view/?ref=126_126448-kcrc0cs6ia&title=FFEVALUATING_THE_INITIAL_IMPACT_OF_COVID_CONTAINMENT_MEASURES_ON_ECONOMIC_ACTIVITY>.
Organization, W.H. 2020, Coronavirus disease 2019 (COVID-19) Situation Report –84.
Pellis, L., Ball, F., Bansal, S., Eames, K., House, T., Isham, V. & Trapman, P. 2015, ‘Eight challenges for network epidemic models’, Epidemics, vol. 10, pp. 58–62.
Prem, K., Liu, Y., Russell, T.W., Kucharski, A.J., Eggo, R.M., Davies, N., Flasche, S., Clifford, S., Pearson, C.A.B., Munday, J.D., Abbott, S., Gibbs, H., Rosello, A., Quilty, B.J., Jombart, T., Sun, F., Diamond, C., Gimma, A., van Zandvoort, K., Funk, S., Jarvis, C.I., Edmunds, W.J., Bosse, N.I., Hellewell, J., Jit, M. & Klepac, P., ‘The effect of control strategies to reduce social mixing on outcomes of the COVID-19 epidemic in Wuhan, China: a modelling study’, The Lancet Public Health.
Rashid, H., Ridda, I., King, C., Begun, M., Tekin, H., Wood, J.G. & Booy, R. 2015, ‘Evidence compendium and advice on social distancing and other related measures for response to an influenza pandemic’, Paediatr Respir Rev, vol. 16, no. 2, pp. 119–26.
Singer, N. & Sang-Hun, C. 2020, As Coronavirus Surveillance Escalates, Personal Privacy Plummets, viewed 18 Apr 2020, <https://www.nytimes.com/2020/03/23/technology/coronavirus-surveillance-tracking-privacy.html>.
Wetsman, N. 2020, Personal privacy matters during a pandemic — but less than it might at other times, The Verge, Online, viewed 16 April 2020, <https://www.theverge.com/2020/3/12/21177129/personal-privacy-pandemic-ethics-public-health-coronavirus>.
Yong, H. 2020, The Public Interest and Personal Privacy in a Time of Crisis (Part II), viewed 17 Apr 2020, <English Translation https://docs.google.com/document/d/1Lvuox6N_g_j61L-ohD0WStaYmZl7x6wyxibqd7cA9cU/edit
Zastrow, M. 2020, South Korea is reporting intimate details of COVID-19 cases: has it helped? , viewed 18 Apr 2020, <https://www.nature.com/articles/d41586-020-00740-y>.
