A data-driven approach in combating COVID19

Using operational command centres at governmental, regional and hospital levels

While the rising COVID19 numbers are shocking the world, I couldn’t help but question what the best way of fighting this disease could be. While many countries are opting for border closures, travel bans and lockdowns, I surprisingly found these measures are not recommended by the WHO. Rather, they recommend testing, isolation and contact-tracing. Of course, next to making sure the general public is adopting hygiene and social distancing principles. But, for a government to organise all of this for millions of its population, I can’t imagine how important access to and use of data would be in defining a successful strategy. Similarly, the healthcare system has to make daily choices on how to organise its limited capacity for the hoards of Corona-patients that come in daily. Currently, the length-of-stay of a Corona patient on an ICU-bed is at least 2-3 weeks. Just imagine the tense on a doctor when he/she has to make a choice to admit a frail old patient with limited survival chances to an already scarce IC-bed, knowing it might be the only available bed for a fit 40-something patient a week later. So many choices have to be made daily by policy-makers, managers, doctors and patients around Corona. My question is, could this be made easier? Could data help to streamline these processes? 

If we take a look at the past, data has actually always been at the forefront of fighting infectious diseases. One of the very first practical use of data in healthcare was done by Semmelweis in 1846, who is also believed to be the founder of infectious disease control. Through statistical analysis, Semmelweis found out the importance of washing hands at clinical wards in lowering the mortality rate at hospitals. During the Cholera outbreak in London in 1854, John Snow (this one actually knew something) mapped the cases of the outbreak on a map. Through this data-visualisation, he found that water pumps were the source of contamination, a big thing as until then it was believed Cholera was spread through the air. Even today, NGO’s and governments are heavily dependent on the availability of data to fight infectious diseases like Polio in Pakistan and Cholera in West-Africa. 

Cholera map of London (1854) by John Snow

So, the answer to the earlier asked question on the importance of data is obviously yes. Therefore, data is almost certainly already playing a huge role in this Corona crisis worldwide. But I want to zoom in on one specific data-driven solution. One which I have been working on in the past and it proved to be successful for governments and hospitals to optimally organise scarce capacity around difficult times. I am talking about Operational Command Centres. In this article, I am going to explain how I believe this could work out for the Dutch healthcare system. 

First, I’ll give a brief explanation of the Dutch healthcare system. All hospitals (around 85) in the Netherlands are private, except for 8 Academic hospitals. The Dutch government plays a limited role in organising care as they try to adopt a free-market principle. A big task is laid out for healthcare insurers as they contract the type of care and price with hospitals and practices. The government specifically sets the ground rules for this yearly play. This is quite different to, for instance, the United Kingdom and Australia where the government directly owns and controls health services. Furthermore, the Netherlands has 12 provinces, 17 million citizens and it has land borders with Germany and Belgium. The RIVM is the national Public Health agency, in some way similar to the American CDC. The GGD is the regional public health service. From my understanding, the Dutch government, hospitals, GP practices, RIVM and GGD are the main players involved around managing the Corona outbreak. It might be clear that a central approach in such a diverse organised landscape could be a challenge in a country like the Netherlands.

What is an Operational Command Centre? You could compare it to NASA’s Mission Control Room (Houston, we have a problem) or the Air Traffic Control Room Centres at airports controlling the airspace. It’s a central location from where a situation is monitored, steered, organised and controlled. While there probably are committees or executive boards in place currently who have central command in managing this disease, I rather want to emphasise on how they could make use of technology and data. Imagine rooms with big screens where all the data is presented in a timely matter so managers can make the right choices at the right time. Which data, I’ll come back on that later. 

Operational Command Centre at Johns Hopkins Hospital

For the Netherlands, I would imagine one Central Operational Command Centre in either The Hague (where the government is seated) or preferably Utrecht (most central point) and 4 to 5 Regional Command Centres spread over the country. In these Command Centres, executive managers representing the Government, RIVM, GGD, Hospitals and Primary Care practices could take place. They could centrally monitor the Corona spread, define strategies, start interventions and command lower branches that are involved as well. Logically, these command centres could have other relevant parties involved like communication specialists, military, police and technologists. Until now, all of this might not seem mind-blowing and we could (hopefully) expect this or something similar is already a reality. But now let’s go to what data could mean in such an organisation. 

If we look at the main problems, currently, it’s that (1) infections are spreading faster than they can be controlled, (2) high-risk groups are difficult to be protected, (3) GP-practices can’t cope with the surging demand, (4) hospitals don’t have enough bed-capacity and (5) the general public is unwilling to follow instructions. With the exponential growth of the infections, we could only expect worse. 

Therefore, these Operational Command Centres could have realtime access to the following data sources:

1. Realtime and on zipcode mapped confirmed cases of COVID-19 (GGD)
2. A geographical overview of high-risk groups mapped on zipcode (Gemeentelijke Basisregistratie)
3. Realtime overview of patients presenting themselves at GP-practices and hospital emergency departments mapped geographically (HIS/ZIS)
4. Realtime overview of bed-capacity at all hospitals (ZIS)
5. Realtime movement of the general public based on mobile connectivity (Telecom providers). 

If we could combine all of this realtime data and map it spatially, we could look for correlations between them. 

Specifically, by combining all this real-time data and analysing it, we could get insight into the following:

• Find clusters of infections (based on zip code)
• Find high-risk areas for high-risk groups 
• Find new trends in rising infections based on patient presentations at General Practitioner (GP) and Emergency Department (ED)
• Predict new high-risk clusters based on existing infections and general public movements
• Find opportunities to steer demand towards capacity-available hospitals in the region
• Predict new trends in regions for both GP/ED presentations as hospital bed-demand

Just imagine the new opportunities that arise to monitor, control and steer the spread of this disease. Just a few examples:

• Preventively install micro lockdowns on high-risk areas (zipcodes or cities)
• Preventively evacuate high-risk citizens from high-risk areas
• Spread the demand for IC-beds evenly over a region and/or the country
• Preventively install mobile IC-beds and staff at hospitals and regions where infection surge
• Target general public communications on a micro-level to steer behaviour
• Monitor in realtime the effect of lockdowns on a micro level

Now we haven’t even been talking about other potential sources of information. For instance using Google search trends, social media behaviour or a ton of other publicly available information sources. In my opinion, there are many ways to generate new insights by combining data sources that would normally be locked in silos of one specific organisation. By installing Operational Command Centres at different levels, both nationally as regionally, with access to combined sets of data, they could combat this virus far more effectively. 

After all, Command Centres are a known phenomenon in the military, air travel and space missions. As many governments are already claiming we’re fighting a war against COVID-19, let's start using the right measures to combat it.