Analysing infectious diseases

COVID-19 has had a huge impact on a global scale – what has that meant for your line of work?

One of the huge issues that you're seeing with COVID is that the real-time surveillance of infectious diseases is a lot like sushi. It looks quite simple, and from afar it doesn’t look like there’s many moving parts to it, but to actually get it to work nicely is a beautiful art and requires a lot of very tiny things that people wouldn't really think about.

I think the best example is probably EuroMOMO (EuroMOMO is a collaborative network of European countries and regions that monitor all-cause excess mortality). Norway participates in EuroMOMO - every Tuesday we send our results to the EuroMOMO hub in Denmark, where it is aggregated and displayed on their website. It doesn't sound that difficult to find out “how many extra deaths happened this week”, but there's a few little tweaks that make is quite complicated. The major issue is registration delay. From the moment you die to the moment that we actually receive the notification of your death, it takes between one to three weeks. Part of EUROmomo’s work is to try and correct for that delay. If we were to blindly use the data without correcting for the registration delay, we would always say “looks like there’s fewer deaths this week than expected!”

When the nuances behind the data (e.g. how the numbers are received, which populations are over/underrepresented) are misunderstood, this can can unfortunately lead to biases in the data (such as a registration delay) being misattributed to a “meaningful trend.”

Can you tell us a little about your background, and how you have come to be based in Norway?

I was born in Wollongong, and grew up in Gerringong, where I attended primary and high school. I was accelerated when I was in primary school, so I started University at 15. At the end of my honour’s degree at UOW, I really wanted to be a firefighter as I had spent a lot of time in the RFS and really enjoyed it. But my parents thought I should do a PhD - so we came to an agreement that I would apply for a PhD and see if I got in. And if I didn't, then I would apply to become a firefighter.

So at that time, I was 18, or 19, and I applied to Oxford, Harvard, Princeton, and Johns Hopkins. I got into Harvard and Johns Hopkins, so then I thought, “Okay, well, I can always do a PhD and then see about becoming a firefighter afterwards.”

I moved to Boston for my PhD in the Harvard School of Public Health, and I met a Norwegian girl studying a masters in political science. She graduated and moved back to Norway, so I took a leave of absence from my PhD and started working with the World Health Organisation in Geneva because that was closer to Norway.

I then moved to Norway and finished my PhD remotely by doing research at the Norwegian Institute of Public Health, which is a government research institute that is similar to the US Centers for Disease Control and Prevention (CDC) and US National Institutes of Health (NIH) combined.

My PhD ended up being a combination of clinical trials (which I did in Boston), gut microbiome research (my job in Norway), and air pollution research (the statistical aspects of air pollution tied in nicely to my other work). It was a very strange combination. But I think most PhDs end up being very strange, because you know, towards the end everyone gets a bit desperate: “What can I write about that hasn’t been written before? Or that isn’t going to change by the time I’m ready to publish.”

You are currently working as an infectious diseases statistician at the Norwegian Institute of Public Health – what does a normal day look like for you? And an extraordinary day?

There aren't that many ordinary days anymore. I mean, I think so far [March 2021], I'm at 800 hours of overtime since the pandemic started and just keeps on going up. Yesterday I worked 14 hours.

Basically, before COVID-19: on a normal day in Norway, the day would mostly be about tweaking various surveillance systems. I work in real-time surveillance, which basically means “what is happening right now?” A health authority needs to know what they should care about and prioritize. Through my work I try to highlight that (maybe) “influenza is looking worse than normal”, or (maybe) “meningococcal seems to be an issue”, things like that.

I work at the national level, and for each of these diseases there is one or two people (or more!) that are responsible for a particular disease. These disease experts know a lot of what's going on with their particular disease, and they have close contacts with the local municipal health authorities and their counterparts in other countries. So they follow the trends. These disease experts are good at doing analyses to confirm or deny things that they suspect, but they're not good at sustaining these analyses long term (because that's not really in their job description). So I'd go talk to them, and ask “what have you been working on lately?”. And they'd reply “Well, I think that disease X is an issue in children now, because our counterparts have seen it in Sweden, and I did this analysis on Norwegian data in Excel, and I saw this trend and it was worrying.” I would take that and I'd implement it in a more robust fashion using my team’s infrastructure, possibly employing some fancier statistics. So now they can get an automatic report every day at 6am for the next year, instead of it taking a week to create their single report. This allows them to spend more of their time interpreting and responding to the reports, instead of creating them.

And then since COVID has happened: the day starts at 6am with monitoring the data coming into the system and the production of the morning reports. My team gets our data at 6.30am and needs to produce 368 reports (1 national, 11 county, and 356 municipality) by 7am.

There's lots of groups, and each group is responsible for their own data source. There's the lab data people, the vaccination people, the deaths people, the cases people, the hospital people, the ICU people, and my team is the octopus at the top that sucks up all of their results, runs some analyses in addition, and then produces the final reports. Unfortunately this means that my team is more likely to get yelled at if something goes wrong, because we are the final deliverers of the reports.

If something breaks, we need to (quickly!) figure out if it's our fault, if it's the fault of the teams who are delivering us the data, or if it's the IT departments fault. If it’s our fault, then we need to identify the problem and fix it as quickly as possible. If it’s someone else’s fault then we notify them and ask them to fix it as quickly as possible.

We also get random requests because our institute is an advisory authority, and we have a lot of data. If the ministry of health (MOH) wants to know something there’s a good chance the question is coming to my institute. Depending on how data-specific the question is, the request can make its way to my team. We then send data/results/a report back up the chain to the people who actually make decisions.

For example, at 9am the MOH might say “with a deadline of 10am, we need to know how many municipalities have an increasing trend in infections.” This kind of request would probably make its way to my team. Then there is a decent chance you will see those results on the evening press conference, which is fun, but kind of stressful. It’s also rewarding to have successfully developed an analytics/surveillance infrastructure that is able to flexibly respond to random questions like this with such tight deadlines in a routine manner.

Richard sitting inside an oral rehydration point (ORP) in Beira, Mozambique, during a cholera outbreak in the aftermath of Cyclone Idai (2019).

You were a scholarship recipient whilst you studied at UOW - how did that help you during your studies?

I was given an undergraduate and an honours scholarship. Basically it helped me move out. If it wasn't for that, I would have been living with my parents the entire time at University, so it gave me the opportunity to move out from home and to also focus primarily on studying, which I think gave me a really nice experience, and I'm really grateful for it.

You have also worked on other large outbreaks, including Ebola – how have you been involved in health response teams when these outbreaks occur?

I’ve had many roles, but I can give a simple example that clearly explains how statisticians can help in these kinds of settings.

I was in rural Sierra Leone for a period during the 2014-2016 Ebola outbreak in West Africa. As a part of this I was working at the response’s district headquarters, which housed the “000” call centre. As I watched the person receiving the Ebola phone calls, I saw that they were just having a conversation with the caller and writing down notes manually. This was not the best way to record standardised data.

I listened to three or so calls, and recognised certain things kept on popping up. I also talked to the doctor that responded to the 000 calls by performing case investigations. I asked him, “when you do a case investigation, what do you actually want to know before you get there?”

I then combined this information and printed out 100 copies of a simple conversation template with tick boxes. It was kind of like, “Hello, you have reached the Ebola hotline. Are you calling about your [X] family member, [ ] neighbour, or [ ] someone else. Are they [ ] healthy, [X] sick, or [ ] dead” and items like that.

Compared to being in the office as a normal analyst, being useful in the field requires a lot more people skills and understanding of the context. Basically, no one is going to hold your hand in the field. You need to have the confidence and knowledge to identify areas where you can make a positive contribution to the team.

Richard in rural Kambia district, Sierra Leone (2015) during the 2014-2016 Ebola outbreak in West Africa. Here, Richard is part of a field team that is investigating rumors of Ebola cases in a particular area.

Is there a particular area that you personally want to make a difference in?

Yes, I think infrastructure regarding the real-time surveillance of infectious diseases is really cool. I don't think there's enough work that's been done on it. When I talk to my colleagues in different countries, we all struggle with the same problems.

Since roughly 2015, the team that I work on has been working on developing its own analytics platform/infrastructure for real-time surveillance of infectious diseases (it’s called “Sykdomspulsen” or “the disease pulse” in Norwegian). We open-source everything we do, because we're funded by the taxpayers. Our development of our platform has advanced roughly 10 years due to the sheer amount of work put in during COVID. Unfortunately this has resulted in our documentation being a little outdated, but this is something we are working hard on and we hope to properly open-source everything with great documentation within a year. The taxpayers have already paid for us to do this work, so why shouldn’t they get it open-sourced in return? In addition, it would be nice if someone else thought that what we’ve done is cool and used our work!

Your career has a distinct focus on maths/stats in a health setting. Has this been a personal passion of yours that you consciously pursued, or did circumstances present you these opportunities?

I always thought that infectious diseases were interesting - the original aim of my PhD was to do infectious disease work and I did some infectious disease work while I was doing my PhD. For example I worked closely with one professor to contribute to the revision the WHO guidelines for TB monitoring. It was meant to be a summer project, but it ended up lasting seven years. Literally, I would talk to her one day and then the next day I’d email her and get an auto reply: “I'm sorry I'm currently working in the Democratic Republic of the Congo, I will have no internet for the next nine months.” And then nine months come around, and she's like “Richard, how's the project going?” Rinse and repeat for seven years.

Broadly speaking, I’ve always wanted to work in infectious diseases, and I was lucky enough to find opportunities that helped me get there.

What advice would you have for any students aspiring to the type of work you do?

Three pieces of advice:

1. Learn to program well. One of the most useful skills from my time at UOW was that I learned computer programming. I was one subject away from getting a major in computer science, and it was very, very fundamental to the way my career has progressed. Statistics is great, because it teaches your brain, “This is how you should think statistically, these are the things that you should be aware of”. But if you can't actually execute on your vision, then you've got a problem.
2. Learn French, Spanish, or Arabic. Language skills can be the difference between being hireable and not. If you don't speak the language of the region/area you want to work in, you are hamstrung with regards to what you can do. Will you need a translator? Regardless, you are disadvantaged compared to others who do speak the language. If you do want to work internationally and you don’t speak the language, you have to be many times better than someone who does speak the language.
3. Work with an organization that has the same values as you, and that values security. Never downplay the importance of your own security. My personal opinion is that it's really important to find an organisation that has the same values of you, and values security. Security is crucial, but unfortunately a lot of organizations don't place enough emphasis on it. You should never downplay the importance of your own health and safety. If your organisation does not prioritise this, then I would recommend walking away and finding an organization that values you and takes its duty of care as an employer seriously. From what I have experienced, my personal values align well with the Norwegian Red Cross’ values, and they have a big focus on security.

Richard White
Bachelor of Mathematics (Applied Statistics), 2009