IEEE Computational Intelligence Magazine - November 2020 - 60

quality checked and analyzed, [5]. MIDAS
therefore proposes a two-pronged approach to
ethical and governance assurance: the public as
partners and a system of robust ethical and scientific oversight from all parties involved in
the MIDAS platform. Public engagement is
core and needs to be meaningful. This requires
a program of engagement, education and support for the public. Obviously this also requires nuance, resource
and openness by science and government, as well as innovative
techniques for engagement, such as the Chatbot discussed previously, and a platform such as 'engage' [9] used throughout the
MIDAS platform development. In the time of SARS-CoV-2
(COVID-19) this may seem a luxury, but we need to plan
now for future outbreaks, pandemics, or other public health
emergencies. This public engagement and perhaps the use of
opt in/opt out models of data use for public health is a discussion that needs to take place urgently. A measure of control is
essential in managing public trust, expectation and compliance
in the use of any system. MIDAS mitigates the risk by creating
a system to manage this requirement: an Honest Broker Service model (HBS). This system creates an operational structure
for review, scientific justification and oversight drawn from all
interested parties, the public, government, academia and business. This is the ABCD model: Academia, Business, Client, and
Direction. These parties set the bar in respect of the scientific/
policy question at hand, allowing scrutiny of the hypothesis by
parties not directly invested in specific work, within a framework that allows review for quality assurance and feasibility.
This model creates a system that can be trusted, a regulatory
framework much as the ones that exist for devices and pharmaceuticals that can allow science to drive decision making.
Public messaging and education is integral to acceptance of
the model, and of course the operational integrity of the system is dependent on user engagement through data use, a
symbiotic relationship between the public and data investigators. A system that includes the public as contributor and gatekeeper, vouchsafed by independent review goes some way to
safeguarding this trust.

... a technological solution, such as the MIDAS
platform, allows analysis of heterogeneous datasets,
in an environment that allows relationships and policy
to be explored.
text. The free form responses were analyzed using IBM's Watson Natural Language Understanding tool, which in turn provided the campaign owners with a general sense of the
sentiment of the conversations they had with the chatbot, and
emotions expressed in the responses. The aggregated view of
these responses shows that the favorite stay at home activity of
38% of the public was the physical exercise, with video chatting
at 25%, and hobbies at 20%. When asked how technology has
shaped their life since the start of the COVID-19 pandemic a
large majority (65%) of people said they rely on [technology]
more than ever and a third of the respondents said they use it as
much as before. Only 2% of the respondents said they use technology less than before.
Topics trending in the campaign responses were family and
friends, health and fitness, resulting societal impacts of the disease, children and education. The scientific impact due to the
global efforts worldwide was also mentioned. The system can
identify the aggregated sentiment of free-form text responses
given by all survey participants and provide the prediction of the
mean probability of the emotions (within the categories: sadness,
joy, fear, disgust or anger) in the free form responses using IBM's
Watson NLU. Of all of the categories mentioned, only technology and computing were mentioned in a neutral context,
whereas the majority of responses were given in a positive context which can be seen in green in the graph in Figure 6.
Ethics in the Time of COVID-19

"May you live in interesting times" is often quoted as an
ancient Chinese curse, but dig deeper and this origin is erroneous. It is actually attributable to Joseph Chamberlain, a 19th
century British politician. The parallels with the current
COVID-19 crisis requires little imagination. The transitional
world in which we currently live has unheard of restrictions of
movements and freedoms normally available in democratic
societies [15]. These restrictions are driven by modelling and the
epidemiological evidence and, certainly to this point in May
2020, the public appears to have trusted the rationale and ap--
proach in large part [32]. Lockdown strategies are a matter of
choosing short term loss over long term gain; these are the policy questions that are being dealt with and, as such, require the
best evidence available. What is clear is that a technological
solution, such as the MIDAS platform, allows analysis of heterogeneous datasets, in an environment that allows relationships
and policy to be explored. A key output of the platform development was the realization that this environment should be
apolitical, in the sense that policy should be based on science
and the relationships of the data used in the system, robustly

60

IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE | NOVEMBER 2020

Conclusion

A substantial part of technology adoption in public health and
healthcare is the utility of the tools and the meaningfulness of
their outcomes. As a result of being co-created with stakeholders [8], undergoing regular impact evaluations [10], and having
usability formally evaluated by policy-makers [6], the MIDAS
platform has proved its usefulness and has led to the development of components driven by stakeholder requests. Significant
interest was established in the MIDAS platform in what it can
offer to new regions, cities and organizations.
Moreover, the platform comprises a representative set of
open, anonymized and synthetic data upon which the full
range of available analytics and corresponding visualizations
reside. This is valuable in an epidemic scenario, enriching the
proprietary data of the public health authority, with existing
IEEE Computational Intelligence Magazine - November 2020

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - November 2020
