Pandemics are big and complex problems. There are too many variables to estimate and their interactions are far beyond our capacity to model. We have limitless combinations of tools and options available to us in any moment, and we can’t consider them all. But pandemics are very dangerous too so, despite these obstacles, we have to find a way to use those tools and options to keep ourselves alive.
Since we cannot forecast in these environments, it behoves us to identify the primary drivers of the system – the variables with the greatest impact. If we can manage these variables, then we give ourselves the greatest influence over the outbreak and our best chance of survival.
The most influential variable in pandemic risk management is speed. The quicker we react to the outbreak, the smaller the problem will be at the time and the easier it will be to solve. In contrast, the slower we react, the bigger the problem and the greater the cost of solving it – if we even can.
In this post, I will look at speed from three perspectives: reaction time, resource requirements, and complexity. By the end, we should be able to agree that all three perspectives point us to the same conclusions: that speed is our greatest weapon against pandemic-potential pathogens (‘PPPs’) and that building more speed into our response plans should be the priority for pandemic policy-makers everywhere.
Reaction Time
At time = 0, a PPP infects the first human being. At time = t, the humans discover the PPP.
This 't' – the time it takes us to find the PPP; the reaction time of our response – is the single most important variable in pandemic risk management. The speed of our response on its own can determine whether our plans will succeed or fail because it determines the size of the problem we’ll need to solve. If we find the pathogen quickly, then the outbreak will be small and local, so eliminating it will be routine. Pandemic prevented; problem solved.
In a perfect world, we would react immediately to the initial infection (ie reduce 't' to 0) and isolate Patient 0 before the pathogen has had a chance to spread to anyone else. The outbreak would be over after just one case, and at minimal cost to society. This is the ideal scenario.
In the real world, we know that we cannot get 't' all the way to 0, but that should still be our ultimate goal (long-term problem-solving means aiming high) because the smaller this 't', the easier our decisions from that point on.
A Numerical Example
If an outbreak is doubling in size every week, and it takes us 10 weeks to discover it, then there will be about 1,000 cases by the time we do. If it only takes us 9 weeks, then there will be 500 cases. If it takes us 8 weeks, then it’s down to 250, etc.
Every week saved in finding the PPP, cuts the size of the outbreak in half (ie by 1/weekly growth rate).
These are massive improvements. There aren’t many scenarios in which a public health intervention can cut an outbreak in half in just one week. (Can you think of any?)
Better yet, as the number of cases decreases, so too does the geographical spread of the outbreak. Fewer countries will be infected and fewer regions within those countries will be infected too, which further simplifies the challenge for local public health functions. It also means there will be more spare resources elsewhere to redistribute where they are most needed.
Ideally again, we would find the PPP while it is still restricted to one region within one country as this would be the simplest scenario of all to solve. A faster reaction time makes that a possibility. Faster still, and it will become a probability.
Get your reaction time right, and everything else will take care of itself. Pretty simple, right? Let’s apply that logic to Covid-19.
The Covid-19 Outbreak
In the early stages of the outbreak, Covid-19 was roughly tripling every week (R0*7/generation time) and it could have had a 10-15 week head start on us. Even at the low end of these estimates, there could easily have been 60,000 cases by the 1st of January 2020 and who knows how many countries infected. A week saved would have cut the number of cases by a factor of three. A fortnight saved would have cut it by a factor of 10!
Think about that: a faster response to the Covid-19 outbreak could have reduced the size of the problem by orders of magnitude. What other public health intervention can promise the same impact?
Had we discovered SARS-CoV-2 at the start of December 2019 instead of the end – which was possible – then the outbreak would have been less than 1% of the size. It would have been more concentrated in Hubei province, making it easier for the Chinese public health officials to eliminate it, and massively reducing the disruption to the rest of the world.
The Covid-19 epidemic really could have been over by Spring 2020. All we needed was more speed. Or, in other words, better disease surveillance*.
Disease Surveillance
Surveillance is our radar for novel pathogens: it is always active, constantly scanning the planet for new threats. Once we find a sample we cannot identify, it triggers our public health alarm bells and our emergency responses will be initiated (or, should be). The better our surveillance programmes, the sooner we’ll find the anomalies and the novel PPPs, and the easier it will be to eliminate them.
In practice, better disease surveillance means more testing: more voluntary testing in communities and in health care settings, more environmental sampling, and more sewerage and wastewater testing, especially at key nodes in the global transport system. This also requires greater data storage and improved data sharing protocols too.
The great news is that we already know how to do this. And we are doing it. So long-term pandemic prevention does not rely on major innovation or a spark of genius, nor does it need significant sums of money. It’s just a case of scaling up our current operations, improving our methods where possible, and applying them in new domains – bog standard policy progress, in other words.
Summary: more speed means a shorter reaction time, a smaller outbreak when we find it, and a greater overall likelihood of success.
The Resource Ratio
The Resource Ratio is the size of your resources relative to the size of your problem. It is a simple, conceptual measure informing you about the strength of your position and your chances of success: the higher your Resource Ratio, the more likely you are to solve the problem, and vice-versa.
Think about a battle.
If your army is 1,000 times bigger than your enemy’s army, then you have a Resource Ratio of 1,000x and you are going to win the battle. If you have 1,000 times as many soldiers, guns, ammo, and other military resources, then there is no way you can lose. No matter how smart their tactics or how brave their warriors (or how dumb and cowardly your own), your resource advantage will be insurmountable. The Resource Ratio will win the day*.
The key insight here is that as the Resource Ratio increases (2x, 3x, 5x, 10x…) it will eventually pass a threshold beyond which failure is all but impossible. We don’t know where that line is drawn and we have no way to estimate it, but we do know that it exists in some meaningful sense, and that a sufficiently large resource advantage will take us past it.
The same logic applies to pandemic risk management, with the PPP as the enemy. Our resources are our health care workers, tests, medicines, hospitals, etc, and the total number of cases is the size of our problem. The more tests we can administer, the more cases we can identify. The bigger our contact tracing capacity, the faster we can track the rest down. The greater our total resource advantage, the quicker we can suppress transmission and the easier all of this will be.
For example, if there are 10 cases in an outbreak, they can easily be handled within the health system and the general public will be unaffected and unaware. If there are 10,000 cases, then the public health infrastructure will be tested, the quality of health care will suffer and there will be difficult decisions or trade-offs to be made. If there are 10,000,000 cases, well, forget about it. You’ve already lost.
So, the goal is to overwhelm the outbreak with our superior resources and to eliminate the PPP, but we can only do that if our resources truly are overwhelming. So the conceptual Resource Ratio threshold beyond which failure is all but impossible, that becomes our target.
How do we hit a conceptual target?
We can't hit an invisible target, admittedly, but we can aim to maximise our Resource Ratio. And if you want to increase a ratio, as any mathematician will tell you, you’ll get more bang for your buck by reducing the denominator rather than by increasing the numerator. So rather than trying to maximise our health care resources (which is also expensive and time-consuming), we should look to minimise the number of infections we are fighting.
How do we do that?
By finding the outbreak sooner and reacting quicker. More speed, in other words.
A faster reaction means fewer cases and a higher Resource Ratio, and the higher our Resource Ratio, the greater our chance of success. I don’t know where the magic line is, but I do know that if we can react quickly enough, then the Resource Ratio will pass that threshold and it will be decisive.
Thinking of Covid-19 again, if we had discovered the outbreak one week sooner it would have had the same effect on our Resource Ratio as tripling the amount of health care resources available to fight the outbreak. One week saved would have been equivalent to tripling our stock of PPE, our ICU capacity, and our infectious disease HCWs, and every other medical resource.
I ask you again: can you think of a single intervention which can have the same impact?
I can’t, which is why I argue that investments in disease surveillance offer the highest returns and should be the priority for policy-makers everywhere.
Summary: shorter reaction times increase our resource advantage over the outbreak, and eventually that advantage will become decisive.
Hidden Risk aka Complexity
When we first discover a novel pathogen, we won’t know anything about its characteristics (R, IFR, attack rate etc) and we won’t know how long it has been transmitting or how far it has spread. The only data we will have is that single case. If we are anything but incredibly naïve, we will reason that the pathogen has had a head start on us, so there will be other undiscovered cases and they will be fuelling further unobservable spread.
These unobserved cases are the ‘hidden risk’ in the system.
Hidden risk is a particularly tricky because you don’t know where it is, or in what amount, but you do know it’s out there, accumulating silently in the shadows. Observable risks can be hard enough to manage, but they have the distinct advantage of being observable, and therefore measurable and finite. Once you can put bounds on the problem, the task of solving it is simplified, and you can plan accordingly.
But when the same risk is hidden, it adds a layer of uncertainty which makes the ‘outcome space’ explode, and with it, the number of questions you need to answer, the number of risks you need to manage, the work that needs to be done, and the resources needed to do it.
Good news is good news, and bad news is bad, but no news is worst of all, because without any information, I must prepare for every possibility.
What is the source of the outbreak? How long has this pathogen been spreading? How many cases are there in total? Where are they? Has it crossed international borders? What is the maximum possible spread? Are we importing cases? Or are we exporting them? How much ‘resource’ will be required to trace and isolate it at home? How long will that take, under various scenarios? Will we need to implement measures which could impact civilian life? If so, which ones and for how long? How much international co-ordination will be required? Will we need a global response? Who should lead it?
Each unknown adds complications as there are more variables to be estimated, more risks to be managed, and more territories and organisations are drawn in, tying down resources and depleting decision-making time. This layer of uncertainty make the problem bigger and more complex, and complexity will soon exhaust your resources – if you let it.
Successful pandemic management needs a way to tame this risk.
The key to making this problem tractable is the understanding that hidden risk is proportional to the pathogen’s head start. The more time the pathogen has had to spread before we discover it, the greater the hidden risk will be when we eventually do, the more unknowns we'll have to deal with and the more complex our task from that point on.
The solution to all of those problems – as it has been throughout this thread – is more speed. A faster response reduces the hidden risk and collapses the outcome space. When we discover that first case, we will won’t know how far the pathogen has spread, but we do know that the better our surveillance system, we can be confident that there will be less hidden risk.
Summary: more speed and faster reaction times make the complex, simple.
Conclusion
The speed of your response is the single most important variable in pandemic risk management, and nothing comes close.
Your reaction time determines the size of the problem when you find it, the cost of solving it, the probability that you will, and the amount of time and effort it will take to do so. Speed influences every other factor in the equation, to the point where it can make them all but irrelevant. Building greater speed into our pandemic responses should be THE priority for all pandemic risk managers everywhere.
This may sound simple or obvious but one should not underestimate the real-world value of this insight. There are many people working in pandemic risk management who genuinely believe that vaccination and other forms of medication are the long-term solution to pandemics. I trust that these people are sincere, but they could hardly be more wrong. And, given their prominence in policy-making positions, they risk taking the world in a very dangerous direction. Giving a PPP a 100-day head start is not a solution; it’s more like playing Russian Roulette with 8 billion lives.
Most of the challenges we face in pandemic risk management are either caused by or exacerbated by a lack of speed – not a lack of medication. Lockdowns, arguments over masks, the complexities of vaccine production and distribution, an overwhelmed health service, painful decisions over limited ICU capacity, political dysfunction, civil unrest… these problems only exist downstream of a slow response. They are the burdens we force upon ourselves when we fail to address the problem at its source.
Problems like these multiply and cascade, so you simply cannot afford to develop solutions to be administered after the event; you must have plans in place to prevent them. A faster response is central to the preventative approach and a faster response can always be achieved with better disease surveillance.
If humanity is to thrive (not just survive) then we need to permanently reduce the risk of PPPs. Building speed into our plans and responses, especially in our immediate reactions, will be essential if we are to achieve that goal. Investments in better disease surveillance are therefore the best use of scarce resources, and that is where our attention should be focused – NOT VACCINES.
You are aways better off reducing the size of the problem rather than increasing your capacity to deal with it.
Or, as we have known for millennia, prevention is better than cure.