by Jeremy Say

The deadliest animal on Earth is this itsy bitsy tiny animal weighing in at 2.5 milligrams. 180 times more deadly than sharks, crocodiles and lions combined as far as casualties go in 2015, mosquitoes are truly mass murderers. With what is estimated to be over a million deaths a year attributed to these disease carrying blood suckers alone, it’s only going to get worse if we ignore the problem.

Spreading this awareness of what was originally a tropical disease, the Discovery Channel has created an interesting documentary on the every growing issue that is the Mosquito.

The documentary, aptly named Mosquito, will feature insights into the little death machines, with interviews from world’s leading experts- Tom Frieden, Director of the Centers for Disease Control and Prevention, leaders’ from the World Health Organization and even Bill Gates will make a rare documentary interview appearance.

Hosted by Jeremy Renner (The Hurt Locker and Avengers), the documentary is set to premiere Thursday, 6th July 2017 at 10:50 PM.

Not wanting to miss out on the chance to look into how mosquitoes have affected Malaysia, we spoke to a local expert – Dr. Dhesi Raja, Chief Scientific Officer of Artificial Intelligence in Medical Epidemiology and Public Health Medical Specialist at the Ministry of Health. He is also one of the creators of the Predict and Beat Dengue App for Android.

“Hello Doctor, can you please introduce yourself to our readers?”

Hi everyone, I’m Dr Dhesi and I’m a medical doctor (MD), and I did my Masters and DrPH (Doctor in Public Health) – which is a sub-division of epidemiology that consists of disease control.

DR. Dhesi Raja, Medical Doctor at the Institute for Medical Research, Ministry of Health

Besides that, I’ve been working closely with dengue in the past 4 years – also,my work revolves around disease predictions.I work more closely on artificial intelligence (AI) than statistical methods. So, in regard to dengue, I study how mosquitoes behave, how do human patterns contribute to dengue outbreaks, weather patterns, geographical settings, etcetera.

“You mentioned about not using statistical data, but more on artificial intelligence (AI). How does that work?”

Well, what we do is try to understand how a disease behaves primarily and then we actually teach a machine to understand how these variables contributes to a dengue outbreak. Globally we have been very reactive towards public health outbreaks.

We currently react to contain the disease, not predict and mitigate the problem. So, this was one of the reasons why we tried to understand how diseases behave and in fact, there are more than 270 variables that can contribute to an outbreak. No normal human mind would be able to digest that enormous amount of data quickly. So that was one of the reasons why AI machine learning is so important, to actually be able to interpret and understand all the variables in real time and of course improve itself.

Currently what our AI can do is predict in advance, 60-90 days in advance for an outbreak.

We currently have created two different platforms for the AI. One that is popular among the community is the mobile app . The mobile app is for us to engage with the community and for them to know when the next outbreak will occur. It acts quite similarly to Waze, which instead of telling you that the next 200m there is probably a police officers or speed traps, our app will tell you whether you are entering a high risk area, moderate area or a low risk area for Dengue.

Our second platform is web-based. This is where we give decision makers, in this case, the government, the ability to prevent outbreaks. We have already aided the government in Brazil during the Rio Olympics, the city of Manila, and are moving forward with the government of Malaysia. This is all possible due to our ability to provide real time analytics and predictions given by our AI for decision makers to understand and to better manage dengue outbreaks.

“Talking a little bit about dengue itself, how did it become such a big pressing concern?  As well as what is the biggest concern currently within the government or medical community regarding how dengue has been spreading?”

Firstly, it’s the mosquitoes itself. Its ability to actually adapt is actually quite fast and problematic. One of our first issues is that out of 100 mosquitoes, probably only 2 or 3 female mosquitoes carry the virus. To identify which of these mosquitoes that is actually carrying the virus takes a tremendous effort and is near to impossible to actually find.

Aedes agypti swells with human blood.

Second, the community’s behavior themselves. I would like to say that dengue is a community based disease. So while our focus is so much on the vector and we speak a lot on the mosquitoes themselves, but I think the major culprit here is the environmental ecosystem of the community.

We are currently tackling dengue as a mosquito disease, but in fact it is an environmental problem.

Everything falls back to who is responsible. Conventionally it’s believed that dengue should be contained and controlled by the Ministry of Health, but with environment playing a factor, I feel that it should no longer be under the jurisdiction of just the Ministry of Health. This means that municipalities, schools, parks, recreational centres, construction sites, and city or town planning all play a role to modify the environment against mosquitoes.

“Talking about how the community plays such an important role, are there any suggestions or tips for the community?”

I think the basic principles of any vector borne disease; dengue, zika, chikungunya, yellow fever – it’s about search and destroy. Meaning that if every human being or every Malaysian just takes 5 minutes or 10 minutes of their day daily, to explore their surroundings, their housing area, and their general environment, we can solve a lot of issues. Everyone should be searching for breeding sites, such as stagnant or calm water, for mosquito eggs or larva and destroy or clean them.

A garbage dump in Kisumu, Kenya. Plastics have been blamed for the spread of mosquito borne disease.

With that, I don’t see it as a problem here, but if we could be complacent and point fingers at the authorities, then you know, it’s nearly impossible to actually contain this. Like I said, it’s a community-based disease. The mosquito breeding sites comes from the community, not from a hospital setting.

“Can you explain what a vector is?”

A vector is an agent that is actually transmitting/harbouring/transferring a disease to the human body. Vectors can be rats, bats, and in this case, mosquitoes. They can move the disease from a human being to another, or even to their offsprings.

“Mosquitoes borne diseases are usually in the warmer tropical climate. Though recently we see European and Northern American countries starting to see its spread. This must be quite a concern because that they weren’t there before.”

Dengue was first found in 1779, in Indonesia and in 1970s, there were just 9 countries that were affected. Today there are more than 28 countries affected. So approximately, there are now 2.5 to 3 Billion people at risk. With the main reason for the spread is air travel, because traveling time have shortened and is now shorter than the incubation period of certain diseases nowadays.

Reason number 2 would be global warming and climate change, so we see now zika, which is carried by the same Aedes Aegypti mosquito, spreading to Miami and Florida. We’ve even seen dengue cases now in Japan and South Korea. Diseases like dengue or zika are generally like you said around the tropical belt, but now we are seeing them in the northern hemisphere.

“With the scare of mosquitoes and how they could keep on spreading viruses like dengue and zika – are there any misconceptions or common myths about mosquitoes that you would like to debunk?”

A common example would be about fogging. If we see fogging happening in our area, we assume that there’s a dengue case, so that’s the reason why they’re fumigating and the dengue would stop there. But essentially, this is not true. Fumigation only kills the adult mosquitoes, not the eggs and larvaes.

Another thing would be that people think that mosquito breed on dirty waters and only in dirty areas. In reality, mosquitoes like to breed in clean areas and water. So clean stagnant waters are also breeding sites for mosquitoes.

“While I have you here, can you tell us a little bit about mosquito borne diseases in Malaysia?”

Basically, as of June 2017, today we have seen about 48,092 cases, and we have seen about 107 deaths. We can also see dengue changing from dengue serotype (variant) 2 to dengue serotype 1. We also see that 70%-80% of cases are happening in urban and suburban areas.

Miami Dade mosquito control teams look for aedes aegypti larvae.

I would say that Selangor has the highest cases of dengue because of the dense population in most urban areas. However, the dengue wave comes about every 4 years or so. It has spiked up last year and eased down this year, but it doesn’t mean that it’s not around us at present.

Our case fatality rate would be under 1% if it was detected early. If an outbreak ever were to happen to someone you know, bring them immediately for a medical checkups. Early medical intervention can severely reduce the fatality rate of our loved ones.

“Is there any specific symptoms that would be concerning and warrant a medical checkup?”

Usually, to define dengue, it would be fever with a few other symptoms like, muscle ache, bone ache or rashes. If you have these, then you should definitely seek a doctor’s help and we’ve also have this app called ‘Beat Dengue’ – it’s a free app.

The app not only can track your movements for a period of 14 days,to help you remember where you could have gotten dengue from, but will also ask you a few questions if you are unsure if you have dengue but have a fever. The AI will partially diagnose you and let you know whether you need a doctor’s consult or not.”

“May I ask what inspired you to create this AI called AIME – Artificial Intelligence in Medical Epidemiology (AMY)?”

In the global perspective, everyone is interested in curative medicine. People like to get well fast and see the outcome quickly. But in terms of preventive, people are a bit reluctant to invest in it because you don’t see the outcome. As said before, this is a real-time based application, we can’t use statistical studies because by the time we publish it, it would be outdated. So I envision it to be like a Bloomberg for health and medicine channel for the community. This idea was pitched in many competitions, and the prototype won top 5 project in the Silicon Valley. Whereby we received a special funding in Rio to predict zika and dengue for the Olympics, with an 84% accuracy.”

“Your predictive technology is showing results. Do you see it expanding in any way in Malaysia?”

We are working with the Ministry of Health, doing our best to go all out with this project for the national implementation. We already did a trial in Penang this year and had predicted 30 out of 37 outbreaks for an 81% accuracy. Hopefully by the end of the year we can have a system implemented.

“Do you see this AI type of platform being used for any other concerning medical disease?”

Yes, there are already many AI platforms in the clinical/medical sectors, but I think the beauty about ‘AIME’, the one that we’re using, is that we have moved from Research & Development to execution. We have been able to replicate it in different countries and have actual users using it.

So that is one difference about the AI development in healthcare. Secondly we’re looking into prediction of surgical modalities, prediction of skin diseases with image recognitions, and many more clinical insights. I think that many countries, universities, and various academia are very confident with our work in dengue as we have shown results. So we’re looking forward for ‘AIME’ to grow.

“Before we end, is there any shoutouts you’d like to make and share with us any future plans?”

I would like to firstly congratulate Discovery Channel for taking the effort to highlight neglected tropical diseases like malaria, dengue, chikungunya, and yellow fever. These are generally tropical diseases that affect a lot of middle and low income countries. So it’s a good initiative that they came up with this program.

Secondly, it is time for the World Health Organization or Global Health, to move from reactive mode towards predictive/in real-time mode in order to mitigate diseases outbreaks.