Containment Under Uncertainty
Act on asymmetric risk before certainty eliminates the chance to contain it
- Difficulty
- Expert
- Time to result
- ~ongoing to results
- Steps
- 6
- Confidence
- 95%
Containment Under Uncertainty treats waiting as a consequential choice rather than a neutral default. Leaders first estimate the speed, scale, and downside of a possible threat, then identify how long meaningful containment remains possible. They compare the expected cost of proportionate early intervention with the cost of delay if the threat proves real. Because hospitalizations, deaths, and other definitive outcomes are lagging indicators, the response must use leading signals and explicit assumptions. Initial measures should be targeted and reversible where possible, while also generating better information. As evidence improves, leaders escalate, maintain, or relax the response. The central mechanism is asymmetric timing: acting slightly early may impose a manageable cost, while acting after certainty can make containment impossible.
Origin
Extracted from Masters in Business, where Charity Dean contrasted the need for early containment with institutions designed to act only after obtaining extensive data.
Core principles
- 01Fast-moving threats require decisions before certainty
- 02Inaction changes the outcome and is therefore an action
- 03Containment has a limited window
- 04Delayed false negatives can be costlier than proportionate early action
- 05Update decisions as evidence improves
How to run it
- 1
Characterize the threat
Estimate plausible speed, reach, severity, and uncertainty using the best available leading indicators.
Pro tip Use optimistic, central, and severe scenarios to expose asymmetry.
Watch out Do not present scenario ranges as predictions.
- 2
Locate the containment window
Determine how long targeted action can still prevent or materially limit spread and damage.
Pro tip Anchor the window to operational realities such as detection, mobilization, and transmission time.
Watch out Lagging outcome data may describe conditions from days or weeks earlier.
- 3
Price action against delay
Compare the downside of a proportionate intervention with the expected consequences of waiting under each scenario.
Pro tip Include second-order health, economic, and operational effects.
Watch out Treating inaction as costless biases the decision toward dangerous delay.
- 4
Choose a proportionate response
Select measures strong enough to preserve containment while limiting unnecessary harm and maintaining reversibility where possible.
Pro tip Combine protective action with accelerated testing or investigation.
Watch out Uncertainty does not justify measures unrelated to the plausible threat.
- 5
Act before confirmation
Implement the selected measures while there is still time for them to change the outcome.
Pro tip Document the assumptions and decision threshold used.
Watch out Waiting for definitive hospitalizations or deaths may mean the operational window has already closed.
- 6
Revise continuously
Gather new evidence and tighten, maintain, or relax the response as the risk estimate changes.
Pro tip Set explicit review times and predefined escalation triggers.
Watch out Do not let an early decision become permanent merely through institutional inertia.
In the wild
Testing restrictions and failed federal tests left the United States unable to detect community spread during a critical early period. Institutions waited for stronger evidence while transmission continued invisibly, making later choices far more damaging.
→ The country lost the opportunity for early containment and faced widespread health and economic disruption.
During the 1918 influenza pandemic, St. Louis introduced distancing measures sooner, while Philadelphia proceeded with a mass parade. The faster intervention reduced bad outcomes compared with the delayed response.
→ Earlier action limited deaths, while delay was followed by a severe mortality surge in Philadelphia.
Common mistakes
Demanding certainty
A system designed to act only when evidence is conclusive will consistently respond too late to rapidly compounding threats.
Treating inaction as neutral
Waiting allows the underlying threat to evolve and can eliminate safer response options.
Using only lagging indicators
Hospitalizations and deaths reveal infections that occurred earlier and therefore cannot serve as the sole trigger for containment.
Is it for you?
Best for
It is best for rapidly compounding threats whose early interventions are cheaper than their uncontrolled consequences.
Not ideal for
It is not ideal when intervention itself is irreversible, highly damaging, and unsupported by even preliminary evidence.
From the transcript
“by design, a system or a thinker a human that can contain a fast moving novel pathogen by design has to operate in the setting…”
“The problem is, in an uncertain circumstance, not taking action is action.”
“if you wait until you're certain, you've waited too long, you've lost your chance.”
From the episode
Charity Dean on the Lessons of Covid-19
Charity Dean