Exponential growth and epidemics

Exponential growth in epidemics isn’t just a curve that looks steep—it’s a process where the number of new cases each day is proportional to the number of existing cases, so the daily change multiplies by a constant. In the COVID-19 data discussed here, daily case counts often behave like a factor of roughly 1.15 to 1.25 times the previous day’s total. That “multiply by a constant” mechanism is why small-looking daily increases can suddenly become enormous, and why comparing countries by raw case counts can mislead: a tenfold difference can correspond to only about a month of time if both are on the same exponential trajectory.

On a logarithmic scale, exponential growth appears as a straight line, making it easier to estimate the growth rate. Using the reported case data, the analysis suggests a pattern like “multiply by 10 every 16 days” on average, with a regression fit described as extremely close to the exponential model. The practical implication is stark: if the trend continued uninterrupted from the recording date (March 6), projections would reach about a million cases in 30 days, 10 million in 47 days, 100 million in 64 days, and 1 billion in 81 days. The key warning is that such straight-line extrapolations can’t last forever—real epidemics must eventually slow.

The slowdown isn’t arbitrary; it follows from the same math that creates exponential growth. In a simple model, each infected person exposes others at an average rate (E), and each exposure has a probability (p) of becoming a new infection. New cases scale like E·p·n, where n is the current number of cases—so growth accelerates as n grows. That acceleration can only stop if either E or p declines. Even in a worst-case “random mixing” scenario, growth eventually saturates because people being exposed are increasingly likely to already be infected; mathematically, this introduces a factor like (1 − fraction already infected), turning the early exponential into a logistic curve.

Logistic growth is exponential at first, then levels off as the epidemic approaches a limiting size. The transition point is the inflection point, where the number of new cases stops increasing and begins to fall. A practical way to detect this is the growth factor: the ratio of new cases on one day to new cases on the previous day. During the exponential phase, this ratio stays clearly above 1; when it approaches 1, the epidemic is nearing the inflection point. That creates a counterintuitive interpretation of “small” changes: a growth factor near 1 can mean the total cases will cap at roughly about twice the current level, while a growth factor still meaningfully above 1 implies orders of magnitude of growth may still be ahead.

Finally, the model’s assumptions can be relaxed. Real populations are clustered, not randomly mixed, but simulations with travel between clusters still produce similar exponential-inducing dynamics, often yielding a fractal-like structure where communities act like interacting units. Importantly, the two levers that reduce growth—lower exposure (E) and lower infection probability (p)—can change through behavior and policy: reduced gathering and travel, or improved hygiene like handwashing. The closing takeaway is both sobering and optimistic: if the growth rate drops, projections collapse dramatically; if it doesn’t, concern is warranted rather than complacency.