Listening For A Pandemic

Starting in the spring of 2009, (H1N1)pdm09 or swine flu was estimated by the CDC to have killed 151,700 - 575,400 people worldwide. CDC also determined that 80 percent of (H1N1)pdm09 virus-related deaths were estimated to have occurred in people younger than 65 years of age.

As a frequent traveler, I was one of those infected, and I was also one of the lucky ones who survived.

My H1N1 Experience

In May 2009, I was personally diagnosed with H1N1 while living in Hong Kong as an analytics consultant focusing on Risk Management and Anti-Money Laundering. I infected multiple cities and corporate offices all within a single week: HDFC Bank in Mumbai, India, ICBC Bank in Shanghai, China, and finally experiencing symptoms in Manila, Philippines.

I documented the whole experience here.

During my nearly two weeks in isolation, I analyzed all the failures that allowed me to infect so many.

The biggest failure was my own: successfully concealing I was sick as I traveled through multiple immigration points - India, China, Hong Kong, and the Philippines. At that time, the way to determine if one was sick was to give passengers a paper form to fill out and to simply ask: Do you have the following symptoms? Fever? Running nose?

To get around this, you simply used the pen at the immigration counter connected by a string or a chain and check no. Then the next person who uses the pen can become infected. 

The other failures happened when I was brought to be tested. The most damning of these was the lack of preparedness by staff, missing supplies at local hospitals and medical centers, and logistics for ambulances to route patients in real-time to adequately prepared locations.

Predicting Pandemics in 2020

Now as Coronavirus is dominating the headlines with the time of writing this article with nearly 650 deaths worldwide where crowdsourcing efforts put the Chinese deaths much higher than what the government publishes. At the same time, the common flu is having it's highest child mortality rate in decades. In Tennessee, more than a dozen school districts have closed due to widespread illnesses.

There are three general solutions to predict and trace signals for upcoming pandemics in relation to travel and each has its own flaws:

Search Trends

The first and simplest idea is simply using Google Search trends to highlight where queries are being asked about symptoms or related terms.

But the issue with this way is that ground zero for the Coronavirus was in Wuhan, China. There Google is restricted. And hot trends in search normally do not track Chinese characters when English is supplied. With Baidu being the number one search engine for China and often being restricted and reshaped by the Government, search trends can be redacted, removed, or redirected.

Hospital Data Integration

Second is simply data collection from doctors, general practitioners, or urgent care centers that report symptoms. Then collating this data with hospital admittance.

Although China has worked hard over the last decade to integrate all of its hospital systems - it leaves out Chinese or Eastern Medicine doctors and facilities. They are normally the first to be contacted before Western medicine or hospitals. So data signals around potential outbreaks could be incredibly difficult to pick up.

Port Of Entry Detection

Thirdly, is having a port of entry scan for symptoms. Airport exit screenings are a way to detect a crescendo in ailments that lead to a contagion. Using heat sensors that detect external body temperatures where green and yellow – the colors for normal body temperatures — are cleared for travel. Anyone with a red forehead signifies a temperature of 101 or higher is then stopped for further screening.

But the problem with this solution is that it's after the passenger was locked in a confined space with other passengers after two, four hours, or longer period of time. So someone leaving a plane has already infected a flight won't show symptoms immediately.

Other Enclosed Spaces Travel

The Diamond Princess, a liner belonging to California-based Princess Cruises, is currently on quarantine in waters off Yokohama's Daikoku Pier in Japan after it was discovered on Tuesday that a passenger had been diagnosed with the Wuhan coronavirus after traveling aboard the vessel for one leg of its scheduled 14-day cruise. Now the cruise ship has sixty-one confirmed coronavirus cases with eight infected being American.

On another cruise ship, the US terrority of Guam has denied entry to Holland America's MS Westerdam cruise ship, which made a stop in Hong Kong last week. Phillippines and Japan also turned away the MS Westerdam. However, there are no known Coronavirus infected passengers on board and the MS Westerdam is not in quarantine. However, there are online reports that say that there are.

Finally, the Hong Kong government will be testing all the passengers and crew aboard World Dream, which is owned by holding company Genting Hong Kong Limited. The only reason for the testing is that three passengers from a previous cruise on the same ship tested positive after they left the ship.

Listening For A Pandemic

So is it possible to use analytics or artificial intelligence to detect and prevent the spread of pandemics?

Yes.

Is it possible to detect and predict pandemics using new methodologies, technologies on confined travel such as cruise ships and airplanes?

Enter our iQ Surveillance solution.

What do all of these examples have in common?

Let's look at what the CDC website says:

Common human coronaviruses, including types 229E, NL63, OC43, and HKU1, usually cause mild to moderate upper-respiratory tract illnesses, like the common cold. Most people get infected with these viruses at some point in their lives. These illnesses usually only last for a short amount of time. Symptoms may include

• runny nose
• headache
• cough
• sore throat
• fever
• a general feeling of being unwell

Human coronaviruses can sometimes cause lower-respiratory tract illnesses, such as pneumonia or bronchitis. This is more common in people with cardiopulmonary disease, people with weakened immune systems, infants, and older adults.

Symptoms.

The key to detecting and predicting pandemics is zeroing in on the symptoms. If you read my own experience with H1N1 it began with a tickle in my throat. Then a full-on cough. What if we could train artificial intelligence on - coughing.

The Analytics of Standup Comedy Meets The Analytics of Contagion

In a previous lifetime, I was a standup comedian. And the successful formula for doing standup comedy is based on data science. Based on my comedy mentor Jami Gong of TakeOut Comedy Hong Kong is getting the crowd engaged within 30 seconds of you standing in front of the mic and then getting the crowd to laugh seven times a minute. A usual starting set is five to seven minutes.

I did a television commercial that delved deep into the analytics of comedy. I also starred in a documentary, "I Need You To Kill" now streaming on Amazon Prime that depicts the intricacies of building a comedy audience across Asia-Pacific.

What does comedy have to do with contagion?

Answer: Enclosed spaces and counting.

As my H1N1 infection grew, so did the cadence of my coughing. I hadn't yet developed a fever but the virus was growing inside me. The number of coughs grew from hour to hour. Also, my number of sneezes increased.

A.I. Surveillance

Enter the LingLong DingDong China's answer to Google Home Mini or Amazon's Echo. Or use all of the above, allowing customers an opt-in app that counts coughs, sneezes within 5-7 minute sample sizes and simply saying "Gesundheit" in German. Or asking, "你还好吗?" in Mandarin. Or "Are you okay?" in English.

If simply the person just choked up on drinking something or possibly choking on food that might require emergency assistance - opt-in A.I. can be a reactionary force to detect and prevent before a pandemic onslaught.

The key to this is allowing for opt-in to make recommendations for checkups, potential information about current viruses going around, and offering a calendar and locations of getting flu shots or vaccinations.

The core component of this type of opt-in NLP (Natural Language Processing) we built with our iQ Surveillance solution. This component is integrated into our iQ Gaming, iQ AML for Gaming, iQ Hospitality and iQ Sports solutions that focus on the Hotel, Casino, and Cruise Line industries.

As a frequent traveler, I was one of those infected, and I was also one of the lucky ones who survived.

Using my first-hand experience of being an H1N1 survivor and how most solutions use "look too late" technology which is simply defensive posturing.

The goal is to go on the offensive. Count symptoms via a time sample via A.I.

Literally, listen before it's too late.

“Death can be as common as the common cold. We have taken everything for granted, but we forget that we are only travelers here for a short time. So don't play the bus driver when you don't know how to drive.”

― Anthony T. Hincks