Data Science: Earning Trust in Public Policy
As society moves down the backside of the COVID curve, many feel we have an opportunity to create a NEW IMPROVED normal, focused on establishing systems that are more inclusive, data-driven, and ethical.
During this difficult time, citizens have relied on local government for guidance, perhaps more than ever before, and agencies have had to rapidly respond and adapt. While horrific, this pandemic has allowed local government agencies to show that they can move quickly, work remotely, and change how they offer services in ways that surprised not only citizens, but even the agencies themselves. This is a major accomplishment for organizations that are typically perceived as being incapable of change or agility.
However, as things stabilize and agencies move from “emergency mode” into “maintenance phase,” we have a moment now to reflect on how we meant our ‘new normal’ to look. Agencies have an opportunity to ask some very key questions:
How do we gather data to inform better policy whilst protecting people’s privacy?
How do we know what to reopen when and where?
MILLION DOLLAR QUESTION: How do we instill citizen TRUST in our decisions???
Let's start with having localized public policies driven by data.
The Data
There are many publicly available datasets run on ESRI, a GIS Mapping tool, that demonstrate the scale and spread of COVID through contact tracing. By layering on additional datasets like population data and current event data one can start to form some key predictions with respect to spikes in demand for health services. This would allow agencies to make better decisions and preemptively dispatch resources to the areas that will need them most.
Given the need for rapid response, one might assume that states would be running to embrace smartphone-enabled Contact Tracing, which allows for people to be automatically alerted should they come into contact with someone who has been infected, however, many states are opting instead to hire armies of people trained to do contact tracing due to concerns regarding its effectiveness and potential privacy issues. Tech companies are scrambling to address these issues. On May 27, a NC-based company called Diveplane announced the release of a promising new smartphone-enabled contact tracing application that addresses many of the data privacy concerns.
Larger sociological concerns
In addition to the technology concerns (hacking, data privacy, etc), there are larger sociological and political issues surrounding smartphone contact tracing as noted in this Wired article demonstrating that it cannot be a standalone solution for any county government. “Any effective contact tracing will require testing for Covid-19 to ramp up far past current levels. Diagnosed or exposed individuals need economic freedom and space to self-quarantine. And many low-income or older folks—those who appear to be most at risk—are less likely to have smartphones." Thus, smartphone contact tracing must be part of a more holistic solution that incorporates other elements too.
Neighborhood tracing through sewage
Many counties are turning towards a method that ensures data privacy, does not require a massive ramp up in tests and can be rolled out today. These counties are using COVID testing at Sewage treatment plants as research has shown that virus particles are shed through stool and other bodily fluids.
This method may be able to overcome several of the current barriers to testing. First, per this article in stat news, “The new research comes at an unprecedented moment in public health: The difficulty and expense of obtaining individual tests for millions of people combined with the virus’ rapid transmission means that public health officials are looking for other ways to grasp the scale of the spread. Clinical testing largely is for those with more severe symptoms, meaning those who are asymptomatic or have milder symptoms — but can still be contagious — often are missed.”
Additionally, by testing at sewage treatment plants those that live in disadvantaged communities will have equal representation in datasets. This is vitally important as many inequities have been identified. For example, in this article in MedPage Today, an example was given of Louisiana state. When the state initially launched drive-through testing, inequalities became clear -- when one 90-year-old woman walked a mile in the heat to get tested -- because as a low-income individual, she didn't have a car.
Government agencies need to be aware of and be able to access all of these methods of collecting data about where COVID is most impacting their communities.
Population Data
Population data is another key component that cannot be ignored. An example to consider is how the global pandemic has hit some communities much much harder than others. According to MedPage, predominantly black U.S. counties are experiencing a three-fold higher infection rate and a six-fold higher death rate than predominantly white counties.
“Comorbidities like hypertension and diabetes, which are tied to COVID-19 complications, disproportionately affect the black community. But the alarming rates at which COVID-19 is killing black Americans extends beyond these comorbidities and can be attributed to decades of spatial segregation, inequitable access to testing and treatment, and withholding racial/ethnicity data from reports on virus outcomes.”
According to the New York Times, ‘other communities like the Hasidic enclaves in New York were also left vulnerable to the coronavirus by a range of social factors, including high levels of poverty, a reliance on religious leaders who were in some cases slow to act and the insular nature of Hasidic society, which harbors a distrust of secular authorities that is born of a troubled history.‘
By overlaying population data into community datasets, we can also form predictions of where spikes in the pandemic will occur. Evidence-based interventions are only possible when there is data to inform them. Public health officials need to know who is at highest risk, and where, to deploy containment measures.
Claim-8 is the brainchild of Peter Raymond, the CEO of New Bureau. Claim-8’s dashboard above demonstrates how this type of platform would look - gathering outbreak data from around the world using data from sewage treatment plants, contact tracing, changes in localized regulation, even protests covered in the news to predict when and where spikes in COVID will occur.
By mining insights from datasets like these, counties can create dynamic effective public policy that are hyper targeted towards affected neighborhoods.
The Million Dollar Question: How to Glean Trust
The million dollar question is, how will local governments glean the trust of citizens? As counties have been rushing to hire thousands of COVID tracers to manually track the spread of the disease, online micro courses have been published to train people how to do this. John Hopkins recently published a free 5 hour course on Coursera to train people how to become a COVID tracer. The course is easy to follow, has a number of short videos explaining how the virus spreads and offers an intro into the data science behind COVID tracing. If high schools used online assets like this to teach about the real world as it happens in a micro course, it would actually serve many goals:
1) Offer training in a real world epidemic
2) Offer introduction into data science using our global pandemic
3) Offer training in a desired skill and competency that states are hiring for right now
4) Those high school students that have been trained in how COVID spreads and the data science behind it can teach their families.
By teaching their family members, perhaps they can also help to influence families to better trust the public policies being created by communities that use data science to inform decisions.
Extreme Collaboration
As states, counties, and cities around the country struggle to find ways to combat COVID, a holistic approach is required to ensure that all communities are being properly monitored for the spread of the disease and that data that is being used to inform public policy is timely, relevant and maintains the privacy of people. Extreme collaboration MUST be used by these government entities in order to create better and more trusted public policy for all.
The collaboration tools are all there. Slack, Box, Mural.ly, video conferencing tools, etc are all available for these entities to work together and share best practices. The challenge is not the tools- the challenge is the culture. If these government entities can begin to use the collaborative tools and learn to work together to share best practices, they will learn that they can glean vast efficiencies because they will not need to learn how to recreate the wheel.
In summary, counties need a holistic approach towards combatting COVID, and frankly county CIOs need to be embracing a form of extreme collaboration to work together. There is no reason for each CIO to recreate the wheel in this fight. There is not one sole solution but a myriad of solutions that when woven together can help inform better public policy that addresses all populations AND glean the hard earned trust of citizens.
Phaedra Boinodiris
Phaedra Boinodiris is a Fellow of the RSA and has focused on inclusion in technology since 1999. Phaedra is currently pursuing her PhD in AI and Ethics due to a generous scholarship from the European Union in collaboration with NYU.
Phaedra is the author of the book Serious Games for Business and is a regular public speaker and contributor to articles in Forbes, Fast Company, the National Academy of Engineering Journal, NPR and other publications. She is on the editorial board for the Journal of Future Robotic Life and on the board of Marbles Museum. She holds a BA and an MBA from the University of North Carolina at Chapel Hill
Julie Miller, PhD
Julie Miller, PhD is a behavioral scientist and innovation leader for Wake County government. She has extensive expertise in academia, working under world-renowned behavioral economist Dan Ariely, as well as applied settings within both government and technology. Julie’s strength lies in applying behavioral science to develop effective and engaging solutions to practical problems. Her interests cross multiple domains including: environmental protection and sustainability, policy-creation, innovation, financial decision-making, human-technology interaction, and health and chronic illness. Julie has taught at several universities, has co-led workshops for groups including government, Centene, EPA, and Google M. She enjoys mentoring local startups and serving on several advisory boards.