We want to share with you what it means and how it feels to be under quarantine in Asia. We asked our friend and partner Jason Alleyne, co-founder of ELITE and Director Asia-Pacific of Besurance Corp, how he experienced the outbreak of the pandemic in Asia. In the interview, Jason also shares his insights from an actuarial science perspective and tells us what he learned from the pandemic.
How did you become interested in COVID-19?
The short answer is – I live in China!
And the long story?
I remember talking to my cousin-in-law, who is an internationally renowned Saxophonist, when the early signs of trouble began. He said, “I don’t know why everyone is making such a fuss about this. Tens of thousands of people die from the flu each year”. At that point, I didn’t really know anything about the topic. Later, a friend of mine sent me a text that read, “Be careful. I know you can’t read Chinese, but this virus is very dangerous”. This was two days before Chinese New Year, and I was in Guangzhou then. So, of course, I began to get curious about the virus.
What did you do next?
Across the country, New Year festivities were greatly impacted. My family didn’t go to the temple and we avoided public gatherings voluntarily. That was when I began to pay more attention. I started to investigate data points that my cousin-in-law had pointed out: in France, the SARS cases in Hong Kong, the swine flu, and Ebola. I already knew certain data points on the Spanish flu from actuarial literature on extreme events. Shortly after the Lunar New Year Day itself, we decided to return to Hong Kong. We were able the cross the border just in time before train travel was halted on January 28th.
Wow, you were lucky! What did you do with all the data points?
I had about 8 days of reports in front of me. By simply looking at how quickly the growth numbers were changing, it became clear that the exponential growth had been disrupted. I decided to test my own hypothesis; I used those 8 data points as a proxy for the “start” of the containment of the spread, then predicted the next 5 days. To my surprise, I was correct to within 1% of the actual reported numbers. I still remember showing these figures to my wife, who is an Actuary, and she, too, was surprised to see my results.
What did you do with these findings?
I shared them with a few professionals from the insurance industry, other Actuaries, and some blockchain techies and investors. I received a range of responses, from cynicism to curiosity. I organized a small gathering of health insurance observers to explain what I believed was the effect of “social distancing”. As I explained, the human network that underpins the exponential spread was being disrupted in a dramatic and impactful way. Also, my understanding of the situation and intuitive analysis helped calm friends and family.
What’s important to note is that the spread of the virus and fatality is so much lower outside Wuhan. The country did not panic; people voluntarily obeyed rules for the greater good of the community, and I feel this greatly reduced the burden on the health care system. Really, all people did was stay at home. It was as simple as that. More importantly, they all knew why they were staying away from other people. For the most part, the country had the right mindset and the human network chain was effectively cut off. This is when I took out time to build my crude 3-stage model based on human network methodology and simulation-based thinking.
How would you explain this model to someone outside the industry?
For our pandemic, I hypothesized that there are 3 aspects in our control, and I was proven correct. The first is the exponential stage, where there is no awareness. People get infected from the source and unknowingly infect others. At this stage, there is a community-wide spread before the authorities act. In the second stage, the community responds by closing branches of the network, such as transportation, air travel, and malls. This is when “social distancing” begins. Furthermore, as the Chinese government explained the human-to-human network, I was compelled to think about how to model that concept instead of trying to fit it in an un-intelligent exponential function. This allowed me to adjust the network connectivity in real-time as it happened in China.
The third aspect that I used is a Monte Carlo simulation to teach myself what was happening. For example, the Diamond Princess cruise ship is a perfect test case of unmitigated spread in a closed population over 45 days from a single source. By using reinforcement learning, I was able to guestimate that the transmission rate was likely 3.5% in person-to-person daily interactions. I also found the mortality rate increase per day is 1 in every 2,000 while infected, but every day recovery is still very possible. The Monte Carlo simulation helps to fill in data gaps, improves intuition about the human process, and enhances communication rather than talking about the “R0” and using emotive terms, such as “highly transmissible” and “deadly”.
You mentioned “R0”. Could you explain what this is?
My point exactly. There isn’t a parent that would tell their 7-year old, “Honey, you can’t play outside because there is a virus out there with an R0 of 2.3”. All these terms – “R0”, “highly transmissible”, and “deadly virus” – are from a corporate-centric mindset and are derived from an equation-centered analysis lens. To answer your question, R0 is a mathematical expression for how fast the virus is spreading, assuming that the spread is in fact exponential. The Statisticians and Actuaries I met with did not care to consider using human-centered language in their analysis and social media posts. They used terms like “R0” and unintentionally caused more panic and fear. What I tried to do within my own circle and friends and family was to share correct facts and insights from what they were seeing for themselves, and the data that was being reported. This, in my opinion, was more helpful.
What does a human-centric fact or insight look like?
Human-centric thinking is realizing that viral spread happens in human-to-human networks, and that every country is different. For example, the Spanish flu was spread in clusters of soldiers in cramped quarters on ships returning from WWI. In our case, Wuhan is an industrial and manufacturing city. There are typical Chinese practices; for example, a vast majority of the elderly live with their immediate family, and there are few religious gatherings, if any at all. This would be very different from what happened in Korea, Iran, and possibly Italy. Hospitals were never overcrowded in Wuhan because of their remarkable efforts to build new temporary ones quickly. However, during the Spanish flu, hospitals were overcrowded. This remains a real risk in rural America, for example, and is where Italy is possibly headed.
Another example is how in China, very few people would ever hug or kiss one another upon meeting, but in Europe, the practice is different. So, the human-to-human network has a different frequently, closeness, and public gathering size based on unique cultures and cityscapes. Age and cohabitation norms are another big factor. There are few elderly care centers in China, so transmission from one elderly individual to another is likely to be different from transmission in European countries. These are just a few examples.
How does this approach fit in with your work on innovation and with Cookhouse Labs?
I’m glad you asked – this is very central to our work on innovation. There are 3 major themes that run here.
The first lesson comes from our actuarial forefathers; they taught us that deep analysis starts with investigating the underlying human reality, rather than simply assuming that an equation will explain a phenomenon. It is this deep investigation and human-centered analysis that creates the insight that leads to the mathematics, rather than the other way around. In our recent research in the Lab, we know our customers want this type of insight from our industry.
Secondly, we need to examine what “customer-centric” means versus “corporate-centric” language and products. In the case of our pandemic, fear of the virus, in my opinion, is deadlier. When I see experts announce in the news that their predictive models indicate 30% to 70% of the population will be infected, that is corporate-centric speak. The range is so wide and the numbers so large that, of course, this will generate fear and panic. What we’ve done in our own company is to first identify the most vulnerable (the economically vulnerable workers) and then develop a product to help them stay at home in quarantine, if infected. This is customer-centric thinking. We do understand the risk very deeply, but we also know that people want insight and actionable tips rather than big brother dictation.
Thirdly, we need to ask as an industry, “What is innovation?” Innovation is thinking differently. I previously mentioned equation-focused analysis and the lack of a customer-centric approach. Where is the social impact that was at the origin of this industry?
Millennials and younger generations have a passion for social impact. To become a credible voice to them, the insurance industry needs to revisit its origins in social impact. The industry exists to be a force of good in society and in communities. How can we use human-centered analysis to share insight and joy with our communities? How can we use a customer-centric and data-driven approach to develop truly impactful products and services, while making a difference to people’s lives? Innovation is magnet to our younger generation, because they want to make a social impact!