How to read a prevalence figure

Three numbers get used interchangeably in this field and they mean different things. Getting them straight is not pedantry; it changes what an intervention should look like.

Prevalence

How many people have the condition now. For chronic pain: an estimated 20.4% of US adults, about 50.0 million people, with 8.0% — 19.6 million — having high-impact chronic pain that frequently limits life or work activities. That is 2016 National Health Interview Survey data.

Prevalence tells you how big a service needs to be. It does not tell you whether the problem is growing.

Incidence

How many new cases arise per unit of time. For chronic pain among previously pain-free US adults: 52.4 per 1,000 person-years, with high-impact chronic pain at 12.0 per 1,000 person-years. That is higher than the published incidence of high blood pressure (45.3), depression (15.9) and diabetes (7.1).

Incidence tells you the trajectory. A service sized to current prevalence, with an incidence like that, falls behind every year.

Persistence

Whether people stay in the condition. Among adults with chronic pain in 2019, 61.4% still had it a year later, and 10.4% were pain free. Both halves matter: this is mostly not a passing state, and it is not always permanent either.

The mistake to avoid

Quoting a confidence interval bound as the estimate. Our own earlier material said chronic pain incidence was “60 in 1,000” — that is the upper bound of the 95% confidence interval around 52.4, not the estimate. It is an easy error and it inflates a figure by about 15%.

The other one: quoting a rate without naming the population it was measured in. “52.4 per 1,000” is meaningless without “among adults who were pain free at baseline.” The same study reports 116.2 per 1,000 for adults who had non-chronic pain at baseline and 462.0 for those who already had chronic pain. Same study, same year, three very different numbers.

More: the scale of chronic pain.